Advances in Agronomy v.82

VOLUME 82
Advisory Board
John S. Boyer Paul Bertsch

University of Delaware University of Georgia
Ronald Phillips Kate M. Scow

University of Minnesota University of California, Davis
Larry P. Wilding
Texas A&M University
Emeritus Advisory Board Members
Kenneth J. Frey Eugene J. Kamprath

Iowa State University North Carolina State University
Martin Alexander
Cornell University
Prepared in cooperation with the

American Society of Agronomy Monographs Committee
Diane E. Stott, Chair
Lisa K. Al-Almoodi David M. Kral Diane H. Rickerl

David D. Baltensperger Jennifer W. MacAdam Wayne F. Robarge
Warren A. Dick Matthew J. Morra Richard Shibles
Jerry L. Hatfield Gary A. Pederson Jeffrey Volenec
John L. Kovar John E. Recheigl Richard E. Zartman
Edited by
Donald L. Sparks
Department of Plant and Soil Sciences
University of Delaware
Newark, Delaware
Amsterdam · Boston · Heidelberg · London · New York · Oxford · Paris · San Diego
San Francisco · Singapore · Sydney · Tokyo
ELSEVIER Inc.
360 Park Avenue South
New York, NY 10010-1710, USA
q 2004 Elsevier Inc. All rights reserved.
This work is protected under copyright by Elsevier, and the following terms and conditions apply to its use:
Photocopying
Single photocopies of single chapters may be made for personal use as allowed by national copyright laws.
Permission of the Publisher and payment of a fee is required for all other photocopying, including multiple or
systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery.
Special rates are available for educational institutions that wish to make photocopies for non-profit educational
classroom use.
Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK:
phone: (+44) 1865 843830, fax: (+44) 1865 853333, e-mail: permissions@elsevier.com. You may also complete
your request on-line via the Elsevier homepage (http://www.elsevier.com), by selecting ‘Customer Support’ and
then ‘Obtaining Permissions’.
In the USA, users may clear permissions and make payments through the Copyright Clearance Center, Inc., 222
Rosewood Drive, Danvers, MA 01923, USA; phone: (978) 7508400, fax: (978) 7504744, and in the UK through
the Copyright Licensing Agency Rapid Clearance Service (CLARCS), 90 Tottenham Court Road, London W1P
0LP, UK; phone: (+44) 207 631 5555; fax: (+44) 207 631 5500. Other countries may have a local reprographic
rights agency for payments.
Derivative Works
Tables of contents may be reproduced for internal circulation, but permission of Elsevier is required for external
resale or distribution of such material.
Permission of the Publisher is required for all other derivative works, including compilations and translations.
Electronic Storage or Usage

Permission of the Publisher is required to store or use electronically any material contained in this work, including
any chapter or part of a chapter.
Except as outlined above, no part of this work may be reproduced, stored in a retrieval system or transmitted in
any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written
permission of the Publisher.
Address permissions requests to: Elsevier Science & Technology Rights Department, at the phone, fax and e-mail
addresses noted above.
Notice
No responsibility is assumed by the Publisher for any injury and/or damage to persons or property as a matter of
products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or
ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular,
independent verification of diagnoses and drug dosages should be made.
First edition 2004
Library of Congress Cataloging in Publication Data

A catalog record from the Library of Congress has been applied for.
British Library Cataloging in Publication Data

A catalogue record from the British Library has been applied for.
Academic Press
An Imprint of Elsevier
525 B Street, Suite 1900, San Diego, California 92101-4495, USA
http://www.academicpress.com
ISBN: 0 12 000780 0
Typeset by Alden Press, Oxford

Printed in Great Britain by MPG Books Ltd, Bodmin, Cornwall
W
1 The paper used in this publication meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper).
Contents
CONTRIBUTORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii
PREFACE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv
FORCES BETWEEN MINERALS AND BIOLOGICAL

SURFACES IN AQUEOUS SOLUTION
Treavor A. Kendall and Steven K. Lower
I. Introduction — Forces in Nature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

II. Fundamental Forces at the Interface of Biological Particles
and Inorganic Surfaces. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
A. The van der Waals Force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
B. The Electrostatic Force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
C. The Solvation Force. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
D. The Steric Force and Bridging Polymers . . . . . . . . . . . . . . . . . . . . . . . . 8
III. Force Curve Theory and Collecting Force Data. . . . . . . . . . . . . . . . . . . . . 11
A. Force – Distance Curves, Capturing a Potential Force
Versus Separation Plot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
B. Hysteresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
C. Tip Shape . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
D. Spring Constant Determination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
E. Artifacts in Force Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
F. Data Processing and Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
G. Advanced Algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
H. Relating Bond Chemistry and Energies to
Force Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
I. Relevance of Dynamic Force Spectroscopy to
Biological – Inorganic Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
IV. Forces at the Biomolecule – Mineral Interface . . . . . . . . . . . . . . . . . . . . . . 27
A. Ligand Linkage Schemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
B. Siderophores and Oxide Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
V. Forces at the Bacterium – Mineral Interface . . . . . . . . . . . . . . . . . . . . . . . . 35
A. Force Microscopy Technique Using Whole Cells . . . . . . . . . . . . . . . . 35
B. Forces Between Escherichia coli and Muscovite . . . . . . . . . . . . . . . . . 39
C. Forces Between Shewanella oneidensis and Goethite
or Diaspore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
VI. Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
v
vi CONTENTS
RICE FUNCTIONAL GENOMICS : LARGE -SCALE GENE

DISCOVERY AND APPLICATIONS TO CROP IMPROVEMENT
Hei Leung and Gynleung An
I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
II. Rice as a Model Genetic System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
A. Diversity of Oryza . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
B. Grass as a Single Genetic System . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
III. Rice Genome . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
A. Genome Sequencing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
B. Salient Features of the Rice Genome . . . . . . . . . . . . . . . . . . . . . . . . . 66
C. Comparison Between Indica and Japonica Rice . . . . . . . . . . . . . . . . . 68
IV. Key Ingredients for Gene Discovery in Rice . . . . . . . . . . . . . . . . . . . . . . . 69
A. Genetic Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
B. High-throughput Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
C. Biological Evaluation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
D. Bioinformatics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
V. Forward and Reverse Genetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
A. Insertional Mutants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
B. Chemical- and Irradiation-induced Mutants . . . . . . . . . . . . . . . . . . . . . 83
C. Natural Genetic Variation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
VI. Functional Validation of Rice Genes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
A. Gene Expression Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
B. Gene Silencing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
C. Heterologous Bioassays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
D. Gene Replacement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
E. Allelic Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
VII. Applications to Crop Improvement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
A. Candidate Gene Approach and Allele Mining . . . . . . . . . . . . . . . . . . . 96
B. Pathways and Genetic Regulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
C. Cross-species Inference of Gene Function . . . . . . . . . . . . . . . . . . . . . . 97
VIII. International Collaboration and the Role
of Developing Countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
IX. Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
THE NATURE , PROPERTIES AND MANAGEMENT OF

VOLCANIC SOILS
R. A. Dahlgren, M. Saigusa and F. C. Ugolini
I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
II. Volcanic Soil Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
CONTENTS vii
III. Soil Genesis in Volcanic Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118

A. Allophanic and Nonallophanic Andisols . . . . . . . . . . . . . . . . . . . . . . 119
B. General Trends in Soil Development on Volcanic Materials. . . . . . . . 122
IV. Soil Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
A. World Reference Base for Soil Resources — Andosols . . . . . . . . . . . 124
B. Soil Taxonomy — Andisols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
V. Chemical Weathering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
VI. Mineralogical Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
A. Common Colloidal Constituents of Volcanic Soils. . . . . . . . . . . . . . . 135
B. Formation and Transformation of Colloids in Volcanic Soils . . . . . . . 141
VII. Selected Chemical Characteristics of Volcanic Soils . . . . . . . . . . . . . . . . 145
A. Organic Matter Accumulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
B. Aluminum Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
VIII. Productivity and Management of Volcanic Soils . . . . . . . . . . . . . . . . . . . 150
A. Charge Characteristics and Chemical Fertility . . . . . . . . . . . . . . . . . . 150
B. Physical Properties and Fertility . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
C. Soil Management and Conservation of Volcanic Soils . . . . . . . . . . . . 166
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
MICROARRAY TECHNOLOGY AND APPLICATIONS IN

ENVIRONMENTAL MICROBIOLOGY
Jizhong Zhou and Dorothea K. Thompson
I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
II. Microarray Types and Advantages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
A. Types of Microarrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
B. Advantages of Microarrays. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
III. Microarray Fabrication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
A. Microarray Substrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
B. Surface Modification for the Attachment of Nucleic Acids. . . . . . . . . 191
C. Arraying Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
D. Critical Issues for Microarray Fabrication . . . . . . . . . . . . . . . . . . . . . 201
IV. Microarray Hybridization and Detection . . . . . . . . . . . . . . . . . . . . . . . . . 205
A. Probe Design and Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
B. Target Labeling and Quality. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
C. Hybridization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
D. Detection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
E. Critical Issues in Hybridization and Detection . . . . . . . . . . . . . . . . . . 211
V. Microarray Image Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
A. Data Acquisition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
B. Assessment of Spot Quality and Background Subtraction. . . . . . . . . . 215
VI. Microarray Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
A. Data Normalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
viii CONTENTS
B. Data Transformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225

C. Methods for Identifying Differentially Expressed Genes . . . . . . . . . . 225
D. Microarray Data Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
VII. Using Microarrays to Monitor Genomic Expression . . . . . . . . . . . . . . . . 231
A. General Approaches to Revealing Differences in Gene
Expression. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
B. Experimental Design for Microarray-based Monitoring of Gene
Expression. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
C. Microarray-based Functional Analysis of Environmental
Microorganisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
VIII. Application of Microarrays to Environmental Studies . . . . . . . . . . . . . . . 242
A. Functional Gene Arrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
B. Phylogenetic Oligonucleotide Arrays . . . . . . . . . . . . . . . . . . . . . . . . 250
C. Community Genome Arrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
D. Whole-genome Open Reading Frame Arrays for Revealing Genome
Differences and Relatedness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
E. Other Types of Microarrays for Microbial Detection and
Characterization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258
IX. Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259
THE AGRONOMY AND ECONOMY OF BLACK PEPPER

(PIPER NIGRUM L.) — THE “KING OF SPICES ”
K. P. Prabhakaran Nair
I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
II. The Pepper Plant — Its Botany and Chemistry . . . . . . . . . . . . . . . . . . . . 279
A. Pepper Botany . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279
B. Pepper Chemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
III. Pepper Agronomy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
A. The Pepper Soils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
B. Nutrition of Black Pepper. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290
C. Evolution of Pepper Manuring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291
D. Response of Pepper to Mineral Nutrients . . . . . . . . . . . . . . . . . . . . . 291
IV. The Role of “The Nutrient Buffer Power Concept” in Pepper
Nutrition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296
A. The Buffer Power and Effect on Nutrient Availability . . . . . . . . . . . 296
B. Basic Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296
C. Measuring the Nutrient Buffer Power and its Importance in Affecting
Nutrient Concentration on Root Surfaces . . . . . . . . . . . . . . . . . . . . . 298
D. Background Informaton on the Importance of Measuring Zn
Buffer Power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301
E. Quantifying Zn Buffer Power of Pepper Growing Soils . . . . . . . . . . . 303
CONTENTS ix
V. Establishing a Pepper Plantation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306

A. The Indian Experience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306
B. The Indonesian Experience. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
VI. Pepper Pests and Their Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311
VII. The Processing of Black Pepper On-Farm. . . . . . . . . . . . . . . . . . . . . . . . 322
A. Sun Drying of Pepper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324
B. Solar Drying of Pepper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324
C. Garbling, Cleaning, and Fractionation . . . . . . . . . . . . . . . . . . . . . . . . 325
D. Packaging and Storing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327
VIII. An Account of Indonesian Pepper Processing . . . . . . . . . . . . . . . . . . . . . 328
IX. Industrial Processing of Black Pepper. . . . . . . . . . . . . . . . . . . . . . . . . . . 330
A. White Pepper. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333
B. Cryoground Pepper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
C. Pepper Oil and Oleoresin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
X. The Future of the Global Pepper Economy . . . . . . . . . . . . . . . . . . . . . . . 339
A. The Supply Side of the Pepper Economy . . . . . . . . . . . . . . . . . . . . . 339
B. The Demand Side of the Pepper Economy . . . . . . . . . . . . . . . . . . . . 343
C. Prices and World Market . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
D. The Pepper Price Outlook by 2020 . . . . . . . . . . . . . . . . . . . . . . . . . . 345
E. The Pepper Supply Outlook by 2020 . . . . . . . . . . . . . . . . . . . . . . . . 346
F. The Pepper Demand Outlook by 2020 . . . . . . . . . . . . . . . . . . . . . . . 346
G. Country-wise Economic Growth Impacting Production and
Consumption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346
XI. Pepper Economy in India. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347
A. Pepper Production Scenario in India . . . . . . . . . . . . . . . . . . . . . . . . . 349
B. Bush Pepper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 351
C. Economics of Pepper Production in the State of Kerala . . . . . . . . . . . 352
D. Marketing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 352
E. Pepper Futures Market . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353
XII. Pepper Pharmacopoeia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355
A. Antiinflammatory and Central Nervous System (CNS) Depressant
Activity of Pepper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357
B. Effect on Hepatic Enzymes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 358
C. Carcinogenic and Mutagenic Effects of Black Pepper . . . . . . . . . . . . 359
D. Pepper as an Antioxidant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 360
E. Pepper as an Antimicrobial Agent . . . . . . . . . . . . . . . . . . . . . . . . . . 361
F. The Pharmacological Effect of Pepper on Human Health . . . . . . . . . . 362
G. Clinical Applications of Pepper . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363
H. Toxicological Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363
I. The Insecticidal Activity of Pepper. . . . . . . . . . . . . . . . . . . . . . . . . . 364
XIII. Consumer Products Out of Black Pepper . . . . . . . . . . . . . . . . . . . . . . . . 365
XIV. Value Addition in Pepper. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365
XV. Conclusions and a Peep Into Pepper’s Future . . . . . . . . . . . . . . . . . . . . . 369
Acknowledgment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 372
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373
x CONTENTS
SOIL MINERAL – ORGANIC MATTER –MICROORGANISM

INTERACTIONS : FUNDAMENTALS AND IMPACTS
P. M. Huang
I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393
II. Mechanisms of Binding of Nonhumic Organics and Humic Substances
by Soil Mineral Colloids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 394
A. Nature of Mineral Colloid Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . 394
B. Binding of Nonhumic Organics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395
C. Binding of Humic Substances. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 397
III. Influence of Organic Substances on the Formation, Transformation,
and Surface Properties of Metal Oxides . . . . . . . . . . . . . . . . . . . . . . . . . 400
A. Aluminum Oxides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 400
B. Iron Oxides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 404
IV. Role of Soil Minerals in Abiotic Catalysis of the Formation of Humic
Substances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 406
A. Polyphenol Pathway. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 406
B. Maillard Reaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 410
V. Interactions of Soil Mineral Colloids with Microorganisms . . . . . . . . . . . 410
A. Surface Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 410
B. Influence on Microbial Activity and Survival . . . . . . . . . . . . . . . . . . 413
VI. Interactions of Soil Colloids with Enzymes. . . . . . . . . . . . . . . . . . . . . . . 415
A. Mineral Colloid – and Humic– enzyme Complexes . . . . . . . . . . . . . . 415
B. Effects on Enzymatic Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 417
VII. Microbial Mediation of Soil Mineral Weathering and Transformation . . . 418
A. Mineral Weathering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 418
B. Fine-grained Mineral Development. . . . . . . . . . . . . . . . . . . . . . . . . . 419
VIII. Interactions of Soil Mineral Colloids with Organic Substances and
Microorganisms in Relation to Soil Structure Stability . . . . . . . . . . . . . . 421
A. Organo-mineral Complexes in Relation to Soil Structure . . . . . . . . . . 421
B. Dynamics of Aggregate Turnover . . . . . . . . . . . . . . . . . . . . . . . . . . . 423
IX. Influence of Mineral Colloids on Biogeochemical Cycling of C, N, P,
and S in Soil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 426
A. Role of Mineral Colloids in Soil Organic Matter Storage and Turnover 426
B. Decomposition and Stabilization of C, N, P, and S in Relation to
Primary and Secondary Particles. . . . . . . . . . . . . . . . . . . . . . . . . . . . 429
X. Effects of Interactions between Microorganisms and Soil Colloids
on the Transformation of Organic Pollutants . . . . . . . . . . . . . . . . . . . . . . 431
A. Catalytic Transformations of Organic Pollutants . . . . . . . . . . . . . . . . 431
B. Binding of Organic Pollutants, Enzymes, and Microorganisms on
Mineral and Humic Surfaces and the Effects on Pollutant
Bioavailability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 436
XI. Impact of Interactions of Physicochemical, Biochemical, and Biological
Processes on Metal Transformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 440
A. Redox Reactions of Metals. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 440
CONTENTS xi
B.Complexation Reactions of Metals . . . . . . . . . . . . . . . . . . . . . . . . . . 441

C.Adsorption –Desorption Reactions of Metals . . . . . . . . . . . . . . . . . . . 442
D.Precipitation – Dissolution Reactions of Metals . . . . . . . . . . . . . . . . . 444
E.Metal Sorption and Uptake by Microorganisms and Biomineralization 445
F.Metal Transformation by Microbial Excretions and Mycorrhizal
Infection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 447
XII. Foreseeable Impacts of Soil Mineral – Organic Matter – Microorganism
Interactions on the Ecosystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 447
A. Global Ion Cycling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 448
B. Global Climate Change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 449
C. Biodiversity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 451
D. Biological Productivity and Human Nutrition . . . . . . . . . . . . . . . . . . 451
E. Geomedicine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 452
F. Ecotoxicology and Human Health. . . . . . . . . . . . . . . . . . . . . . . . . . . 453
G. Biotechnology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 454
H. Risk Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 456
I. Risk Management, Remediation, and Restoration . . . . . . . . . . . . . . . 457
XIII. Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 459
Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 459
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 460
LOW EXTERNAL INPUT TECHNOLOGIES FOR

LIVELIHOOD IMPROVEMENT IN
SUBSISTENCE AGRICULTURE
Anil Graves, Robin Matthews and Kevin Waldie
I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 473
II. The Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 476
A. Intercropping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 476
B. Alley Cropping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 479
C. Cover Crops and Green Manures . . . . . . . . . . . . . . . . . . . . . . . . . . . 484
D. Biomass Transfer Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 490
E. Compost . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 493
F. Animal Manure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 496
G. Improved Fallows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 499
III. Generic Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 504
A. Soil Fertility Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 504
B. Socio-economic Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 516
IV. Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 526
A. Integrated Nutrient Management. . . . . . . . . . . . . . . . . . . . . . . . . . . . 526
B. A Systems Perspective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 530
C. Modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 534
V. Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 539
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 541
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 541
xii CONTENTS
AMMONIA EMISSION FROM MINERAL FERTILIZERS

AND FERTILIZED CROPS
Sven G. Sommer, Jan K. Schjoerring and O. T. Denmead
I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 558
II. Mineral Fertilizer Consumption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 560
III. Ammonia Volatilization from Mineral Fertilizers . . . . . . . . . . . . . . . . . . 561
A. Production and Transport of NH3 in the Soil– Fertilizer – Atmosphere
Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 561
B. Temporal NH3 Loss Pattern . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 565
C. Hydrolysis of Urea. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 568
D. Soil Hþ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 570
E. Soil CEC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 574
F. Solid Phase Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 575
G. Climate and Infiltration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 576
H. Microbial Processes (Nitrification/Immobilization) . . . . . . . . . . . . . . 578
IV. Ammonia Emission from Crop Foliage. . . . . . . . . . . . . . . . . . . . . . . . . . 581
A. Transport of NH3 Between Leaves and the Atmosphere. . . . . . . . . . . 581
B. Magnitude of NH3 Losses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 582
C. Physiological Processes Involved in NH3 Emission from Crops . . . . . 584
V. Management Strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 586
A. Techniques for Reduction of NH3 Emission . . . . . . . . . . . . . . . . . . . 587
B. Fertilizer Composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 589
C. Flooded Fields (Rice Paddies) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 591
D. Injection of Anhydrous Ammonia. . . . . . . . . . . . . . . . . . . . . . . . . . . 594
E. Crop NH3 Emissions as Affected by Fertilizer Application . . . . . . . . 594
F. Ammonia Emission from Decomposing Plant Material . . . . . . . . . . . 596
G. Absorption by Crops . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 598
VI. Measurement Techniques. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 598
A. Tracer Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 598
B. Enclosures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 599
C. Micrometeorological Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 601
D. Gradient Diffusion Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 603
E. Eddy Correlation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 605
F. Relaxed Eddy Accumulation or Conditional Sampling. . . . . . . . . . . . 605
G. Lagrangian Dispersion Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 606
VII. Conclusions and Perspectives. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 607
Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 609
Acknowledgment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 611
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 611
INDEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 623
Contributors
Numbers in parentheses indicate the pages on which the authors’ contributions begin.
G. AN (55), Department of Life Science, Pohang University of Science and

Technology, San 31 Hyoja-dong, Nam-gu, Pohang, Korea 790-784
R. A. DAHLGREN (113), Land, Air and Water Resources, University of
California, Davis, California 95616, USA
O. T. DENMEAD (557), CSIRO Land and Water, GPO Box 1666, Canberra
ACT 2601, Australia
A. GRAVES (473), Institute of Water and Environment, Cranfield University,
Silsoe, Bedfordshire MK45 4DT, UK
P. M. HUANG (391), Department of Soil Science, University of Saskatchewan,
51 Campus Drive, Saskatoon, SK, S7N 5A8 Canada
T. A. KENDALL (1), Division of Engineering and Applied Sciences, Harvard
University, 40 Oxford St., Cambridge, MA 02138, USA
H. LEUNG (55), International Rice Research Institute, DAPO Box 7777, Metro
Manila, Philippines
S. K. LOWER (1), Department of Geological Sciences & School of Natural
Resources, The Ohio State University, 275 Mendenhall Laboratory, 125 South
Oval Mall, Columbus, Ohio 43210, USA
R. MATTHEWS (473), Institute of Water and Environment, Cranfield
University, Silsoe, Bedfordshire MK45 4DT, UK
K. P. PRABHAKARAN NAIR (271), Akshaya, East Hill, Calicut 673 005,
Kerala State, India
M. SAIGUSA (113), Experimental Farm of Tohoku University, Kawatabi,
Naruko, Tamatsukuri, Miyagi 989-6711, Japan
J. K. SCHJOERRING (557), Plant Nutrition Laboratory, Department of
Agricultural Sciences, The Royal Veterinary and Agricultural University,
Thorvaldsensvej 40, DK-1871 Frederiksberg C, Denmark
S. G. SOMMER (557), Department of Agricultural Engineering, Danish
Institute of Agricultural Sciences, Research Centre Bygholm, PO Box 536,
DK-8700 Horsens, Denmark
D. K. THOMPSON (183), Environmental Sciences Division, Oak Ridge
National Laboratory, Oak Ridge, Tennessee 37831, USA
F. C. UGOLINI (113), Dipartimento di Scienza del Suolo e Nutrizione della
Pianta, Università Degli Studi, Piazzale delle Cascine 15, 50144 Firenze, Italy
K. WALDIE (473), International and Rural Development Department,
University of Reading, Reading RG6 6AH, UK
J. ZHOU (183), Environmental Sciences Division, Oak Ridge National
Laboratory, Oak Ridge, Tennessee 37831, USA
xiii
This Page Intentionally Left Blank
Preface
Volume 82 continues the tradition and excellence of “Advances in Agronomy” by
containing eight state-of-the-art reviews on topics of interest in the plant and soil
sciences. Three of the reviews present cutting-edge molecular scale techniques
and approaches that directly impact food production, crop improvement, and
environmental quality and sustainability. Chapter 1 is an excellent review on the
use of force microscopy to investigate reactions at the mineral/microbe interface.
Chapter 2 is a definitive review on rice functional genomics including large-scale
gene discovery and applications to crop improvement. Chapter 3 presents a
comprehensive overview of volcanic soils including their genesis, chemical,
mineralogical, and physical properties and the productivity and management of
volcanic soils. Chapter 4 is a seminal review on microarray technology and
applications to environmental microbiology. Microarray types, fabrication,
hybridization and detection, image processing, data analysis, the use of
microarrays to monitor genomic expression, and applications of microarrays to
environmental studies are covered. Chapter 5 is a definitive overview on the
agronomy and economy of black pepper. Topics that are discussed include: the
botany and chemistry of black pepper, pepper agronomy, production and
management, and economic aspects of global black pepper production. Chapter
6 is an outstanding overview of soil mineral-organic matter-microorganism
interactions. Topics that are discussed include: mechanisms of binding of
nonhumic organics and humic substances by soil mineral colloids, the influence of
organic substances on the surface chemistry and transformations of metal oxides,
interactions of minerals with microorganisms and enzymes, the role of microbes
on mineral weathering, and impacts of mineral-humic-microbe interactions on
nutrient cycling and transport of pollutants and ecosystems. Chapter 7 reviews low
external input technologies for livelihood improvement in subsistence agriculture.
The technologies are discussed along with economic and nutrient management
considerations. Chapter 8 is a comprehensive review on ammonia emission from
mineral fertilizers and fertilized crops. Topics that are covered include: ammonia
volatilization from mineral fertilizers, ammonia emission from crop foliage, and
management strategies and techniques.
Many thanks to all the authors for their excellent reviews.
DONALD L. SPARKS
University of Delaware
xv
FORCES BETWEEN MINERALS AND
BIOLOGICAL SURFACES IN AQUEOUS
SOLUTION
Treavor A. Kendall1 and Steven K. Lower2
1
Harvard University, Division of Engineering and Applied Sciences, 40 Oxford St.,
Cambridge, MA 02138, USA
2
Department of Geological Sciences & School of Natural Resources, The Ohio State
University, 275 Mendenhall Laboratory, 125 South Oval Mall, Columbus,
Ohio 43210, USA, Email: lower.9@osu.edu
I. Introduction — Forces in Nature

II. Fundamental Forces at the Interface of Biological Particles
and Inorganic Surfaces
A. The van der Waals Force
B. The Electrostatic Force
C. The Solvation Force
D. The Steric Force and Bridging Polymers
III. Force Curve Theory and Collecting Force Data
A. Force– distance Curves, Capturing a Potential Force
Versus Separation Plot
B. Hysteresis
C. Tip Shape
D. Spring Constant Determination
E. Artifacts in Force Measurements
F. Data Processing and Statistics
G. Advanced Algorithms
H. Relating Bond Chemistry and Energies to
Force Measurements
I. Relevance of Dynamic Force Spectroscopy to
Biological – Inorganic Interface
IV. Forces at the Biomolecule– Mineral Interface
A. Ligand Linkage Schemes
B. Siderophores and Oxide Surfaces
V. Forces at the Bacterium– Mineral Interface
A. Force Microscopy Technique Using Whole Cells
B. Forces Between Escherichia coli and Muscovite
C. Forces Between Shewanella oneidensis and Goethite
or Diaspore
VI. Future Work
Acknowledgments
References
1
Advances in Agronomy, Volume 82
Copyright q 2004 by Academic Press. All rights of reproduction in any form reserved.
0065-2113/03 $35.00
2 T. A. KENDALL AND S. K. LOWER
At the most fundamental level, intermolecular forces (e.g., van der Waals,
electrostatic, solvation, steric) control interactions between biological
molecules and mineral surfaces. These are forces with magnitudes of
piconewtons to nanonewtons, which operate in a space that is on the order of
nanometers. We have used force microscopy to quantitatively probe forces,
energies, and distances between crystal surfaces and living microbial cells
or biological molecules in their native state. The systems we have studied
include those involving: Escherichia coli, Shewanella oneidensis, side-
rophores, muscovite, goethite, and/or diaspore, in aqueous solutions of
varying composition. Direct force measurements at the organic –
inorganic interface have been interpreted with theoretical models
describing interfacial forces, adhesion, and molecular dynamic calcu-
lations. A new perspective on bacterium – mineral interactions is
emerging from these studies. We have discovered a world that operates
under a very different set of principles than macroscopic bodies. A world
where the intermolecular force, rather than gravitational attraction, is
the preeminent force controlling the evolution of processes at the
bacterium – mineral interface. q 2004 Academic Press.
I. INTRODUCTION — FORCES IN NATURE
The bacterium – mineral interface is ubiquitous near the surface of the Earth.
As many as 97% of the , 1030 prokaryotes on Earth live in close proximity to
minerals in soil, marine, and terrestrial subsurface environments (Whitman et al.,
1998). As we will show in this manuscript, the fundamental forces at this interface
are very small, seemingly insignificant. This review will provide evidence that
forces on the order of nanonewtons (1029 N) to piconewtons (10212 N) dominate
the properties/processes at bacterium –mineral and biomolecule – mineral inter-
faces. For comparison, there is , 0.2 nN of gravitational attraction between a
person (50 kg) and the paper (5 g) upon which these words are written. Despite
their small magnitude, these forces are at the heart of all interactions between
biologically produced polymers and mineral surfaces in nature.
It is now well established that there are four fundamental forces in nature: the
strong and weak nuclear forces, the gravitational interaction, and electromagnetic
forces, which are the source of all intermolecular forces (Israelachvili, 1992).
Because the first two (i.e., nuclear forces) have a range of action that is less than
1025 nm (Israelachvili, 1992), we need not consider these for interactions
between biological molecules, microbial cells, and/or mineral surfaces. The
question then becomes, under what conditions do gravitational forces or
electromagnetic forces (more specifically, intermolecular forces) dominate
bacteria – mineral or biomolecule – mineral interactions?
FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES 3
In nature, living organisms exist in communities that are in contact with one
another, in contact with mineral surfaces, and they are also in contact with the
surface of the Earth (i.e., the upper crust). For simplicity, let us define a particular
species of organism as a spherical particle (having a density of water) with a
unique size or radius. Each species may interact with one another and/or the
Earth. In both instances, there is a force of gravitational attraction at each
interface. Figure 1 reveals that the gravitational attraction is much greater
between the Earth and a particle of a given size (e.g., , 4 £ 1025 N for a 1 mm
particle) relative to the gravitational attraction between two particles of the same
given size (e.g., , 3 £ 10216 N between two 1 mm particles). Also shown on this
figure is a theoretical prediction for another type of attractive force, the so-called
van der Waals force. This intermolecular force was determined using Eq. (1) (see
below) to describe the attraction between two similar objects of equal size in
contact with one another. For example, two identical 1 mm (radius) particles are
Figure 1 Log–log plot of the theoretical forces describing (1) gravitational attraction between a
particle and the Earth (solid “Earth-particle” line), (2) gravitational attraction between two particles of
the same size (dashed “particle-particle” line), and (3) van der Waals attraction between two particles
of the same size (dashed “vdw” lines). In all instances the particles are assumed to be in “contact” with
the Earth (for 1) or another particle (for 2 and 3). For gravitational attraction, mass was determined by
assuming each particle was a solid homogeneous sphere with a density of 1 g cm23, and contact was
defined as the radius of the Earth (~6.4 £ 106 m radius; “Earth-particle” interaction) or the sum of the
radii of two interacting particles (“particle-particle” interaction). The shaded region outlines the
boundaries of the expected van der Waals force using values for Hamaker constant of 10220 to
10221 J, which is appropriate for biological and inorganic phases (Israelachvili, 1992; Leckband and
Israelachvili, 2001; Vigeant et al., 2002), and defining “contact” as an effective separation between
particles of ~0.2 (for one hydration layer) to 2 nm, according to Israelachvili (1992) and Leckband and
Israelachvili (2001). Only the magnitudes of the forces are shown. By convention, attractive forces
(shown here) are negative. For reference, the three diamond symbols represent gravitational forces
between the Earth (~1024 kg) and each of three bodies (from left to right): a bacterium (10215 kg), a
human (50 kg), or the moon (1022 kg).
expected to have an attractive, adhesion force at contact (due solely to the van
der Waals force) equal to , 3 £ 1025 N (Hamaker constant ¼ 10220 J; effective
separation ¼ 0.165 nm, i.e., the “universal” cut-off separation, Israelachvili,
1992). This force magnitude is approximately the same as the gravitational force
between the Earth’s surface and one of these 1 mm particles. While it is
debatable whether the van der Waals force applies in the same manner to both a
particle of the size of an atom and an object of the size of the moon, the
predictions shown in Fig. 1 for objects smaller than , 1 cm are in agreement
with others (e.g., Israelachvili, 1992). Consequently, the force of gravity may
dominate the interactions between macroscopic bodies (e.g., plants and animals),
but intermolecular forces (e.g., van der Waals and others, see below) are the
prevailing forces with which microscopic bacteria must contend. This is
particularly true when one considers that the van der Waals force is significantly
weaker and shorter range than other intermolecular forces, such as electrostatic
and hydrophobic interactions as discussed below.
II. FUNDAMENTAL FORCES AT THE INTERFACE

OF BIOLOGICAL PARTICLES AND
INORGANIC SURFACES
“All intermolecular forces are essentially electrostatic in origin” (page 11,

Israelachvili, 1992). In theory, classical electrostatics could be used to calculate
intermolecular forces if one could determine the spatial distribution of the
electron cloud by solving the Schrödinger equation (Israelachvili, 1992).
Unfortunately this is challenging for even simple atomic interactions in vacuum,
never mind molecular or organism scale interactions between different functional
groups on bacteria and minerals in water. For this reason, it is useful to classify
four types of intermolecular forces that are expected to dominate the bacterium –
mineral and biomolecule –mineral interfaces. These include the van der Waals
force, electrostatic forces, solvation interactions, and steric or entropic forces
(Israelachvili and McGuiggan, 1988). The reader is referred to a number of
excellent reviews on these types of forces (e.g., Israelachvili and McGuiggan,
1988; Butt et al., 1995; Leckband and Israelachvili, 2001). This review will touch
on all four types of intermolecular forces, although the van der Waals and
electrostatic forces will be explored in more detail.
A. THE VAN DER WAALS FORCE
The van der Waals force, like the force of gravity, acts between all particles
(Israelachvili, 1992). It is quantum mechanical in origin and arises because of
the time dependent fluctuations in the electric dipole moment of a particle as it

comes into contact with other particles nearby. Even nonpolar particles, which
have a time averaged dipole moment of zero, have instantaneous dipoles due to
the movement of electrons relative to protons in a nucleus. Dipoles generate an
electric field that polarizes adjacent particles and gives rise to an instantaneous
force between neighboring particles. Two terms describe the van der Waals force:
the first polarization potential, which represents the energy necessary to ionize an
atom (i.e., a dipole moment due to interactions between electrons and protons
within a single particle); and the so-called dispersion term, which describes the
dipole induced interactions between two or more atoms (Israelachvili, 1992).
Because the dispersion term dominates the van der Waals force, it is sometimes
referred to as (London) dispersion forces (Butt et al., 1995).
The van der Waals force has an inverse power law dependence on the
separation between two particles. For atoms and small molecules the van der
Waals force is , D 27, where D is the separation distance between particles. It
can be attractive or repulsive (e.g., it is always attractive between two similar
particles immersed in a third liquid) and is described in terms of the Hamaker
constant (Ha), which depends upon the refractive indices and dielectric constants
of the interacting particles and intervening media (see Israelachvili, 1992).
Hamaker constants are in the order of 10220 to 10221 J, for biological cells or
molecules interacting with themselves or minerals across an aqueous solution
(Ducker et al., 1991; Butt et al., 1995; Ong et al., 1999; Bhattacharjee et al.,
2000; Leckband and Israelachvili, 2001; Vigeant et al., 2002).
For simple geometries, the forces between atoms or molecules can be assumed
to be additive (Israelachvili, 1992; Butt et al., 1995) such that equations can be
derived for larger particles (e.g., organic and inorganic surfaces). Two commonly
encountered geometric configurations include interactions between two spheres
or a sphere and a flat surface, both of which are given by Israelachvili (1992),
Butt et al. (1995) and Leckband and Israelachvili (2001):
2Ha Rx
FðDÞ ¼ ð1Þ
6D2
where Ha is the Hamaker constant (J), D is the separation distance (m) between
the two spheres or a sphere and a plane, and Rx (m) equals the radius of the sphere
for the sphere– plane configuration, or it is equal to (R1R2/(R1þ R2)) for the
interaction between two spheres of radius R1 and R2. A positive Hamaker
constant indicates attraction (negative force sign).
B. THE ELECTROSTATIC FORCE
The electrostatic force arises through a variety of mechanisms leading to the

development of surface charge (e.g., see Sposito, 1989). Water, which has a high
dielectric constant, causes the dissociation of surface functional groups. These

functional groups display protonation/deprotonation reactions that are dependent
upon pH. For example, in water, silanol groups on a silica surface undergo the
following reaction: . Si OH ¼ . Si O2 þ Hþ. Similar acid – base reactions take
place on carboxylic groups, amine groups, and other reactive moieties on
biological molecules and inorganic surfaces. Hence, many inorganic and
biological surfaces develop a charge that is dependent upon pH. Other factors
such as the adsorption of charged ions and presence of permanent structural charge
(e.g., for clays) are additional contributors to surface charge. The overall charge on
a surface is balanced by the dissolved counterions in solution, which are attracted
to the surface by its electric field and dispersed such that they (i.e., the counterions)
increase entropy (Butt et al., 1995). This creates the so-called electric double-layer
around surfaces immersed in aqueous solution (Stumm, 1992). When two charged
surfaces approach one another, the electric double-layers are perturbed resulting
in an electrostatic interaction. This interaction may be attractive (if surfaces are of
opposite charge) or repulsive (if surfaces are similarly charged).
The electrostatic force varies exponentially with the distance between
particles. It depends strongly upon the surface charge densities of the interacting
particles and the ionic strength of the intervening solution. Similar to the van der
Waals force (see above) equations can be derived to describe the electrostatic
force for various geometric configurations. The model for electrostatic forces
between two spheres or a sphere and flat surface is (Butt et al., 1995; Muller and
Engel, 1997; Leckband and Israelachvili, 2001):
4ps1 s2 Rx 2kD
FðDÞ ¼ e ð2Þ
110 k
where s is the surface charge density (C m22) of particles 1 and 2, 1 is the dielectric
constant of water (78.54 at 298 K), 10 is the permittivity of free space
(8.854 £ 10212 C2 J21 m21), Rx and D are defined as above. The Debye length
(1/k) describes the thickness of the diffuse double-layer of counterions that
surrounds charged particles in solution. The Debye length depends upon the
valence and concentration (c, mol L21) of the electrolyte. For monovalent electro-
lytes (e.g., NaCl) at a temperature of 298 K, the Debye length (in nm) ¼ 0.304/
(c)1/2; for 1:2 or 2:1 electrolytes (e.g., CaCl2) it is 0.174/(c)1/2; for 2:2 electrolytes it
is 0.152/(c)1/2 (Muller and Engel, 1997). In many instances, it is easier to determine
a particle’s surface potential as opposed to surface charge. The Graham equation
can be used to relate these two parameters according to (Stumm, 1992),
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
z cF
s ¼ 8RT110 c £ 103 £ sinh ð3Þ
2RT
where R is the gas constant (8.314 J mol21 K21), T is the temperature (K), z is
the valence of ions in solution, c is the surface potential (V), and F is the
Faraday constant (96,490 C mol21). A potential measured across an interface
contains contributions from at least two layers, the so-called Stern layer and the
“diffuse” layer (see Stumm, 1992). Techniques such as streaming potential and
electrophoresis are commonly used to determine a particle’s zeta potential,
which is used as a proxy for surface potential. However, the zeta potential
probably represents only the “diffuse” double-layer, which is lower than the true
surface potential (Stumm, 1992). Leckband and Israelachvili (2001) describe
the differences for surfaces that are assumed to have a constant surface charge
versus those that are assumed to have a constant surface potential. Interactions
at constant surface charge are expected to occur when surface ionizable groups
are fully dissociated and remain as such for all separations (D). This may be true
when the pH of a solution is much greater than the pK value(s) of a particular
protonation/deprotonation reaction(s). In instances where surface functional
groups are not fully ionized but in equilibrium with solution ions, interactions at
constant potential are expected to occur. In this latter case, as two surfaces come
together (i.e., very small D) the intervening concentration of solution ions
increases locally such that some solution ions bind to the surface thereby
reducing that surface’s density of charged sites (Leckband and Israelachvili,
2001). For many instances, this distinction influences the interaction only at
small separations where these two conditions define the boundaries of the
expected electrostatic force.
C. THE SOLVATION FORCE
The origin, theory, and force – distance relationships of the remaining two
force classes — solvation and steric — are indefinite compared to the forces
discussed above. Much work remains to be done before solvation and steric
forces can be appreciated to the same extent as the van der Waals and
electrostatic forces. However, it is well established that the models developed
for the van der Waals and electrostatic forces, which treat the intervening
solution as a continuum, break down when two particles or surfaces are within a
few nanometers (Butt et al., 1995; Leckband and Israelachvili, 2001). At such
close separations, solvation forces may dominate because the solvent (e.g., water)
takes on a more ordered structure. Steric forces may also come into play for
surfaces with polymers (e.g., biological cells or particles). Our discussion of
solvation and steric forces will be more qualitative however, because general
force laws (such as those described above) are relatively sparse for these latter
two force classes.
Solvation forces (also called hydration or structural forces when the solvent is
water) seem to be the result of interactions of solvent molecules with themselves
(e.g., in a confined space between two surfaces) or interactions between solvent
molecules and a surface (e.g., the orientation of water molecules at the interface
of a strongly hydrophilic surface). As two surfaces approach one another the
intervening liquid ceases to behave as a structureless media resulting in a force

that can be attractive, repulsive, or oscillatory (Butt et al., 1995). These forces
can be further subdivided into those that result from solvent –solvent, solvent –
surface, and surface – surface interactions (Israelachvili and McGuiggan, 1988).
For two rigid crystalline surfaces at short-range (, 2 nm for water), water
molecules interact with themselves such that they take on a semi-discrete
layering or structure, which causes the “structural” forces between the interacting
surfaces to oscillate between attraction and repulsion with a periodicity equal to
the molecular dimension of water (Leckband and Israelachvili, 2001). Between
surfaces with polymers, water cannot form well-defined layers because
headgroups on lipids, for example, are “rough” on the scale of a water molecule
(Israelachvili, 1992), and macromolecules in surfaces are thermally mobile
(Beveridge, 1999). Consequently, any repulsion is smeared out and takes on a
monotonic component (Israelachvili, 1992). For strongly hydrophilic surfaces in
aqueous solution, there is a strong solvent –surface interaction that leads to the
formation of hydration shells. These ordered water molecules within the “shell”
generate an electric field that impinges upon another surface as two particles
approach to within a few nanometers of one another (Israelachvili and
McGuiggan, 1988). For example, water molecules may associate with two,
adjacent hydrophillic surfaces such that the water’s hydrogens are oriented
towards each surface (attracted via hydrogen bonds) and the water oxygens are
exposed to the solution. This confers a negative character (from the lone pairs of
the water’s oxygens) to each surface, thereby generating a repulsive force.
Conversely, the dipoles may complement one another forming an attractive force
if water molecules are staggered on the two surfaces. This hydration force may
extend outwards more than the oscillatory force discussed previously (Leckband
and Israelachvili, 2001). Finally, for nonpolar surfaces that cannot bind to water
molecules — the so-called hydrophobic surfaces (defined as those surfaces
having a contact angle of 75– 1158 with water) — there is often a strong
attractive force that extends to separations of tens of nanometers or greater
(Leckband and Israelachvili, 2001). Hydrophobic forces can be significantly
greater than the van der Waals force and may play an important role in
interactions involving hydrophobic molecules and/or surfaces (Israelachvili and
McGuiggan, 1988; Israelachvili, 1992).
D. THE STERIC FORCE AND BRIDGING POLYMERS
The steric force affects surfaces that have flexible polymers extending out into
solution (e.g., polysaccharides on biological cells). As two surfaces approach
one another, the polymer chains become confined such that they are not free to
move at random. This entropic confinement results in a repulsive force whose
length scale is approximately equal to the radius of gyration of the polymer
(Butt et al., 1995), where the radius of gyration is proportional to the number of
monomer segments raised to some power between 0.33 (for poor solvents) and
0.6 (for good solvents) (Leckband and Israelachvili, 2001). Approximations
derived for the interaction between two flat surfaces reveal that this force depends
on the surface coverage of the polymer and may take an exponential form
(Israelachvili, 1992; Leckband and Israelachvili, 2001). At close separation, the
magnitude of the steric force can be similar to that of the electrostatic force
(Leckband and Israelachvili, 2001).
In some instances, surface-bound polymers may form an attractive interaction
at close separation as the polymer forms a “bridge” between two particles or
surfaces (Leckband and Israelachvili, 2001). The resulting adhesive bond may be
very long range (i.e., extend well beyond the radius of gyration of the polymer)
and resist separation when the surfaces are pulled apart (Jeppesen et al., 2001).
While there is no general description for attractive bridging forces by polymers,
the linkage of surfaces via a polymeric tether has been described by the so-called
freely jointed chain (see e.g., Leckband and Israelachvili, 2001), or worm-like
chain models (see e.g., Flory, 1989; Bustamante et al., 1994). In the case of the
latter, the polymer is viewed as an elastic element and the force (F) needed to
stretch the tethered polymer to a length x is:
FðxÞ ¼ ðkB T=bÞ½0:25ð1 2 x=LÞ22 2 0:25 þ x=L ð4Þ
where kB is the Boltzmann’s constant, T is the temperature, b is the persistence

length (i.e., length of the stiff segment or monomer of the chain), and L is the
contour length (i.e., length of the completely stretched chain).
Polymer bridging is a phenomenon that crosses between the disciplines of
colloidal science — which has historically tended to investigate intermolecular
forces that dominate the interface between two rigid surfaces that are
approaching one another — and adhesion science — which is interested in
describing the contact between two surfaces and the forces necessary to pull them
apart. While attractive intermolecular and intersurface forces (i.e., the four force
classes discussed above) are responsible for adhesion events, real particles (e.g.,
bacteria and minerals) that make contact will also adhere to one another due to
elastic or fluid-like deformation, which is an intrinsic and natural part of contact.
There is a wealth of information on adhesion processes and theories including the
Johnson – Kendall –Roberts (Johnson et al., 1971) and Derjaguin – Muller –
Toporov (Derjaguin et al., 1975) theories, which relate the force required to
pull two surfaces apart (i.e., the “pull off” force) to the surface energy, surface
tension, or work of adhesion. Suffice it to say that surface energy (or tension or
work) is determined from intermolecular forces between surfaces. For particles or
surfaces that are incapable of forming hydrogen bonding (e.g., nonmetallic
compounds), the surface energy can be related directly to the van der Waals
force, where surface tension < Ha/2.1 £ 10221 (Israelachvili, 1992). The surface
energies of more polar surfaces, which tend to be larger, are dependent upon van
der Waals interactions, as well as an additional electrostatic-like term that relates
surface energy to Lewis acid/base reactions (van Oss, 1993).
These four forces — van der Waals, electrostatic, solvation (hydration and
hydrophobic), and steric — operate concurrently at the interfaces between
microorganisms, biological molecules, and/or mineral surfaces (see Table 1).
In some instances, one force may dominate at all separations. In other
instances, there is a delicate balance such that each force dominates at its own
length scale. These four force classes are often invoked to describe interactions
as two surfaces approach one another. Two particles that are pulled apart may
experience the same sign, magnitude, and range of forces that existed upon
approach. However, there is often a notable hysteresis between the forces
measured upon approach versus those that are observed upon retraction for
soft biological particles and surfaces. This is due to the formation of adhesive
bonds (e.g., see discussion of polymer bridges and adhesion, above) once
contact has been established between surfaces. This review will provide
examples that illustrate the various forces and force models discussed above as
they pertain to interactions between biological and inorganic particles. Further,
we will discuss the differences between those forces measured as surfaces
Table I
Summary of Physical Forces of Interaction Between Particles and/or Surfacesp
Type of interaction Description
van der Waals Force between all particles due to

polarization; usually attractive; short-range
Electrostatic Force between charged particles; attractive (for particles of
opposite sign) or repulsive (for particles of similar sign);
depends upon ionic strength of solution; short to
long range
Solvation Structural or hydration forces are typically repulsive
due to sorbed water layers; short-range
Hydrophobic force is attractive between nonpolar
surfaces; short to long range
Steric Typically a short-range, repulsive force associated
with polymers; may be longer range, attractive
force for “bridging” polymers.
p
This review has followed the force characterization of Israelachvili (1988). Other force classes,
such as hydrogen bonding or thermal fluctuations, may dominate when two particles or surfaces are
very close. However, these other classes can often be described as a subset of electrostatic (for
hydrogen bonding) or steric (for thermal fluctuations) interactions, according to Leckband and
Israelachvili (2001). So-called specific interactions (e.g., ligand–receptor interactions) are typically a
result of unique combinations of these four “non-specific” physical forces (Israelachvili, 1992).
come together relative to surfaces that are pulled apart. As a final point to this
section, it should be noted that a force of interaction is related to energy (E)
according to F ¼ 2 dE/dD.
III. FORCE CURVE THEORY AND COLLECTING

FORCE DATA
Force measurements attempt to capture interactions representing the

electrostatic interplay between single molecules and atoms that are bound to a
solid surface or exist as components of a solvated environment. Given the
extraordinarily small dimensional (nanometer to angstrom) scale over which they
operate, many challenges exist in capturing molecular level forces. This section
reviews how force microscopy (or atomic force microscopy, AFM; also know as
scanning probe microscopy) addresses these challenges, describes its operation
and assesses how accurately the interactions are captured. Highlighted are some
of the basic assumptions associated with force microscopy, while noting some of
its advantages and limitations.
An AFM force probe consists of a tip attached to a flexible cantilever, which is
modeled mechanically as a single harmonic oscillator. Forces exerted on the tip
are registered as a spring-like deflection in the cantilever, which may be recorded
with various detection systems, including electron tunneling (Binnig et al., 1986),
interferomotery (Erlandsson et al., 1988; Rugar et al., 1989) and capacitance
(Goddenhenrich et al., 1990). The following summarizes an optical lever
collection system (Meyer and Amer, 1990) that is most commonly found in
commercially available AFMs, including the widely used Veeco/Digital
Instrument system. Here the deflection is typically recorded as a change in
voltage resulting from the displacement of a laser spot that is reflected off the top
of the cantilever and into a photodiode. Voltage (V) is translated into cantilever
deflection (nm) using a detector sensitivity value (V nm21) that is equal to the
slope of the line when the tip and sample are in contact (see region of contact in
Fig. 2). Provided the sample stiffness is significantly higher than the cantilever
(which is the case when probing mineral surfaces), there should be a 1:1
correlation between piezo movement and cantilever deflection once the (V to nm)
conversion is made. Small deviations from an absolute slope equal to one may be
an indication of detector drift, and can be corrected by dividing the deflection
values by the slope (H. Skulason, personal communication). A slope less than one
may also be an indication of a sample compliance that is less than the cantilever
(which may be the case when making measurements on a cell), in which case
alternative sensitivity determinations, such as the photodiode shift voltage
method may be employed (D’Costa and Hoh, 1995; Lower et al., 2001b).
Figure 2 A typical force versus piezo movement plot showing three general regions — contact,
interaction, and no contact. For clarity a single trace is shown (e.g. an approach curve); however, force
plots with both approach and retraction traces are also common. In the region of no contact the tip and
sample are separated at distances large enough that no interaction occurs. Hysteresis between the
approach and retraction curve in the region of no contact may be a function of solution viscosity, or
inelastic deformation of the cantilever. As the piezo advances the sample closer, the tip begins to
“feel” the surface. In the example plot we see an initial repulsion followed by an attraction recorded as
a sharp jump to contact that generates a minimum in the curve. Once in contact, the slope trace is
typically constant as the cantilever is moving with the piezo. Information from this region may be used
to determine detector sensitivity or elastic properties of the sample or tip.
Hooke’s law, F ¼ 2 ksd, then allows conversion of cantilever deflection, d,

into force, F, using the spring constant of the cantilever, ks. Note that sign
convention dictates that negative forces reflect attractive interactions and positive
forces are repulsive.
A. FORCE – DISTANCE CURVES, CAPTURING A POTENTIAL

FORCE VERSUS SEPARATION PLOT
Figure 2 shows a typical plot of force versus piezo movement. Note the x-axis
represents relative piezo movement or an indexing of a sample’s position relative
to the cantilever (tip). It does not reflect tip-sample separation (discussed below).
Three main components of the plot are identified: the regions of no contact,
interaction and contact. Several sub-features are contained within each region
including oscillations, subtle slope changes, linear and non-linear extensions,
jumps to and from contact (Ducker et al., 1992; Cappella and Dietler, 1999;
Gergely et al., 2001), which, in addition to providing reference points to register
the force curve to an origin (discussed below), contain valuable information on
the interaction between the tip and the surface, the nature of the intervening
solution, tribology, adhesion, and elastic properties of the system.
The focus is now turned to the region of interaction, which is of primary

interest when studying intermolecular forces at cell or biomolecule –mineral
interfaces. Here a wealth of information on the charge character of a mineral
surface or biomolecule; the nature and contour length of a polymer extending
from a bacterium; DLVO forces (see below) and Debye lengths associated with a
colloidal particle or cell; and the energy landscape and activation barriers of a
bond are found. But first, to draw both qualitative and quantitative conclusions
from forces of interaction, it is imperative to have an understanding of the
mechanical constraints of what is recorded in this region using force microscopy
using an AFM. To illustrate this we show a simple, short-range interaction
potential for atomic scale particles (note, Part II concentrated mainly on larger
particles and/or surfaces) described by the Lennard – Jones equation:
EðDÞ ¼ 2A=D6 þ B=D12 ð5Þ
Energy, E, has an inverse power law dependence on distance, D, with the 2 1/D 6
term representing the attractive component of the van der Waals force. The
absolute value of this term is maximized at a distance De where the fluctuations in
charge density coincide to result in a potential well. At separations less than De, the
potential rises rapidly with distance, 1/D 12, as the interaction is repulsive in nature
due to electronic overlap and nuclear interaction (Israelachvili, 1992; Cygan,
2001). Force microscopy (or AFM), however, does not record energy values
directly, but instead measures force. To compare the Lennard– Jones potential
with an AFM data set, we take its derivative, such that graphing the relationship
dE=dD ¼ FðDÞ ¼ 26A=D7 þ 12B=D13 ð6Þ
produces a theoretical force –separation distance curve similar to the one in Fig. 3a.
To further facilitate comparison with the theoretical curve, an origin is defined for
the force microscopy data set as follows. A force equal to zero can be defined as the
average force value within the region of no contact, while the point at which the tip
and sample come into (for approach) and out (for retraction) of contact can be
defined as the zero point on the x-axis. Determining the point of contact is clear
when a distinct attractive or adhesive component (e.g., a jump to contact) is
present, but ambiguous when such features are absent. In the latter case, the
intersection between the slope of the region of no contact and constant compliance
can be used as a guide (Cappella and Dietler, 1999). In a final important step, the x-
axis in the force microscopy data set is adjusted to reflect tip sample separation
instead of piezo movement, by adding the cantilever deflection values to the piezo
movement distances (Ducker et al., 1992; Butt et al., 1995). Here the selection of
the sign convention for the forces becomes intuitive. Addition of positive repulsive
deflections to the piezo movement results in larger tip-sample separation, while
adding negative attractive deflections result in a decrease in the separation. Unlike
the Lennard – Jones curve, note that the values in Fig. 3b, left of the point of contact
are essentially meaningless in terms of interaction force because the tip and sample
are in direct contact. The end result is a force versus tip-sample separation plot with
a region of interaction that can be compared to the theoretical curve (see Fig. 3a
and b; also see Section V, below).
Two main differences exist between the force microscopy data and the
potential: (1) the slope of the attractive component of each curve, and (2) the
hysteresis that exists between the approach and retraction forces in the force
microscopy plot. With the force microscope, it is not uncommon to record
Figure 3 (a) Differentiated Lennard– Jones potential provided as an example interaction to be

captured with force microscopy or AFM. During AFM operation the forces associated with the
potential are recorded as deflections in the cantilever. If the force gradient (tangent to the solid trace)
exceeds the spring constant, ks, the cantilever becomes mechanically unstable and will jump along a
slope equal to kg (dashed line). (b) Force– tip sample separation curve showing jumps to and from
contact along slope ¼ kg. Unlike Figure 2, the x-axis represents separation distance between the tip
and the sample. Here, both the approach (open circles) and the retraction (closed circles) traces are
shown. Note that the hysteresis between the two traces is absent in the Lennard–Jones curve where the
solid line represents both approach and retraction forces. Points to the left of zero separation (i.e.,
lowest most point on the approach or retraction curves) represent movement of the piezo while the tip
and sample are in contact. (c) Increasing the spring constant (e.g., using a stiffer cantilever) from ksA to
ksB will capture more of the potential (region A2 –B2), however force resolution is lost and smaller
magnitude forces will go undetected.
an attractive force as a characteristic jump to contact on approach. These “jumps”

represent mechanical instability in the cantilever due to a force gradient that
exceeds its spring constant, ks. Clearly, interaction information is lost as the
cantilever encounters a force gradient (tangent to the theoretical curve) at point
A1 that is greater than its stiffness and consequently jumps to point A2 along a
slope equal to ksA (see Fig. 3c). More of the attractive potential can be sampled
with a stiffer cantilever (e.g., ksB), however, force resolution is lost, and the
region along the theoretical curve between B2 and A1/B1 remains unsampled. A
similar situation may be encountered upon retraction, which, in part, contributes
to the hysteresis observed in the force data. Specifically, cantilevers with smaller
spring constants generate larger amounts of hysteresis. However, hysteresis
between approach and retraction curves is also due to the formation of adhesive
bonds once surfaces are in contact. This is common for soft samples such as
biological cells (see below). For some investigations excessive hysteresis is
undesirable and several techniques have been developed to reduce it thereby
recovering the “lost” information (i.e., region A2 – B2). These methods employ
an opposing force that is external to the system in an attempt to increase the
effective stiffness of the cantilever, while retaining force resolution. Electrostatic
force (Joyce and Houston, 1991), magnetic feedback (Jarvis et al., 1996;
Yamamoto et al., 1997; Jarvis et al., 1998; Ashby et al., 2000) and radiation
pressure from a laser (Aoki et al., 1997; Tokunaga et al., 1997) have all been used
to supply the steadying force to the cantilever.
B. HYSTERESIS
Certainly other sources besides the instability of the cantilever contribute to

approach –retraction hysteresis. In the theoretical Lennard –Jones relationship
given as a potential example, no adhesive reaction between the tip and sample is
modeled and the retraction curve retraces the approach curve (Fig. 3a). However,
this is not an appropriate model for soft biological cells, which have biopolymers,
designed over millions of years of evolutionary selection, for the express purpose
of adhesion. When making force microscopy measurements, the tip comes into
contact with the surface allowing for reaction and deformation between the two.
The bonds and coordinations that result can then be explored and characterized
using the associated adhesion forces and approach-retraction hysteresis
(Burnham et al., 1990; Cappella et al., 1997). In some systems, the number of
bonds that form (and, thus the level of hysteresis) is correlated with the amount of
pressure that is applied on the sample by the tip (Weisenhorn et al., 1992).
Specifically, increased pressure leads to sample and tip deformation resulting in
increased contact area (Israelachvili, 1992; Cappella et al., 1997), and in the case
of functionalized tips (e.g., those coated with self-assembling organic monolayers),
a possible rearrangement of functional groups terminating the monolayer.
Both phenomena facilitate additional bonding, larger hysteresis and higher

adhesion values, as documented by several workers (Weisenhorn et al., 1992;
Hutter and Bechhoefer, 1993; Toikka et al., 1996; Ashby et al., 2000). Therefore,
the amount of indentation must be carefully documented to facilitate the
comparison of adhesion data from one study to another. One way of controlling
the amount of indentation using commercially available AFMs is by adjusting the
scan start position, setpoint or the trigger settings. Varying these parameters can
be especially useful when probing many biological systems, where pressure and/
or contact time may be a natural mechanism of inducing adhesion (e.g., see
Leckband and Israelachvili, 2001; Lower et al., 2001a).
C. TIP SHAPE
Tip shape is a critical AFM parameter that can dictate the force values and
contact geometry between the sample and force probe (Hartmann, 1991; Butt
et al., 1995). Constraining this value is essential if experimental force traces are
going to be compared theoretical models such as “DLVO” (see Section V below).
Yet, tip shape can be difficult to determine, in part due to the surface roughness,
irregularities and asperities that are associated with traditional silicon or silicon
nitride tips (Cappella et al., 1997). Moreover, tip shape can change over time as
continued use promotes wear (Cappella et al., 1997). Solutions to this problem
include careful, periodic characterization of the tip with electron microscopy,
better constraint of tip geometry by attaching a spherical colloidal probe
(Ducker et al., 1991; Butt et al., 1995), or, as described in more recent work, by
attaching a carbon nanotube (Wong et al., 1998a; Hafner et al., 1999; Cheung
et al., 2000).
D. SPRING CONSTANT DETERMINATION
If a quantitative analysis of absolute force values is desired, determination of

the spring constant (ks) is critical and nominal values provided by the
manufacturer generally cannot be relied upon (Lower et al., 2001b). Many
factors affect the spring constant including primary characteristics such as
cantilever dimensions, geometry and substrate material; as well as, additional
modifications common in force spectroscopy such as gold coating, the addition of
organic monolayers, the attachment of colloidal spheres or cells, and even ion
adsorption (Sader et al., 1995; Craig and Neto, 2001; Cherian and Thundat,
2002). As a result, a large body of literature detailing several methods of directly
determining ks exists. A procedure commonly used because of its simplicity,
non-destructive nature, and applicability to common cantilever geometries
(e.g., V-shaped, rectangular) is provided by Cleveland et al. (1993). This method
derives the spring constant by measuring changes in the cantilever’s resonance

frequency after small masses (e.g. W microspheres) are loaded onto the end of the
tip. The Cleveland method is further optimized when corrected for errors
introduced by off-end loading of the mass (Sader et al., 1995). More recent
methods measure hydrodynamic drag of the cantilever through a fluid of known
viscosity to determine ks for bare rectangular cantilevers (Sader, 1998; Sader
et al., 1999; Maeda and Senden, 2000). This concept has also been applied to
determine ks for cantilevers activated with a colloidal probe (i.e., a silica or
polystyrene microsphere) (Craig and Neto, 2001). In the latter method, it is useful
to directly measure ks for a cantilever with an attached sphere because it accounts
for changes in the spring constant due to the shifts in the point of load associated
with the position of the colloid sphere and the change in the stiffness associated
with the adhesive used for microsphere attachment. Other methods measure ks
using thermal oscillations and the equipartition theorem (Hutter and Bechhoefer,
1993; Butt and Jaschke, 1995), a finite element analysis of the static deflection of
a cantilever for which the Young’s modulus is known (Sader and White, 1993);
unloaded resonant frequency of a cantilever of known mass (which is commonly
not the case) (Sader et al., 1995), radiation pressure from an acoustic transducer
(Degertekin et al., 2001); microscopic and macroscopic reference cantilevers of
known stiffness (Rabinovich and Yoon, 1994; Torii et al., 1996; Jericho, 2002);
and the change in resonant frequency due to gold coating (Gibson et al., 2001).
E. ARTIFACTS IN FORCE MEASUREMENTS
Several artifacts can arise during force measurements with the AFM. The
inverse path effect represented as an upward, hysteretic shift in the retraction
trace in the region of contact arises from nonlinearities of the piezoelectric
actuator that positions the sample (or tip) (Cappella et al., 1997; Cappella and
Dietler, 1999; Heinz and Hoh, 1999). A shift in the contact portion of the
retraction trace such that it is parallel with the extension trace reflects friction as
the tip plows or slides along the surface (Heinz and Hoh, 1999). A sinusoidal
oscillation in the region of no contact may also be present, representing the
interference of stray laser light bouncing off the sample and interfering with the
laser light reflected off the top of the cantilever (Weisenhorn et al., 1992;
Cappella et al., 1997). This oscillation can be distinguished from other artifacts,
such as noise due to mechanical vibrations, because its wavelength should
roughly be equal to , l/2n, where l is the wavelength of the laser source and n is
the index of refraction of the fluid between the tip and sample (Weisenhorn et al.,
1992; Craig and Neto, 2001). Thermal oscillation in the region of contact can be
recognized by deflection fluctuations whose standard deviation is roughly equal
to (kskBT)0.5; ks, the spring constant; kB, Boltzmann’s constant and T, temperature
(Gergely et al., 2001). Operational artifacts may include a large slope that is
present in the region of no contact, common when making measurements in a

fluid cell. The origin of this slope is unclear but can often be remedied by
eliminating air bubbles in the system or insuring a flat, even orientation of the
gasket used to seal the system. Large plateaus at the extremities of both traces
often represent a saturation of the detector, requiring an adjustment of the scale of
the plot, the deflection limit of the detector or the starting position of the scan.
F. DATA PROCESSING AND STATISTICS
With standard AFMs, the one-click ease with which a single force curve is
collected allows hundreds of curves to be recorded at a sample point in only a few
minutes. Considering the fact that a typical curve can contain 2048 data points, a
single experiment can produce an enormous volume of data. Further, the
variability between force curves collected at a single location can often be quite
high. This raises several questions regarding data processing and interpretation
that are often neglected. What is the most efficient way to process these data?
What is the minimum number of curves necessary to characterize each sample
point or a particular interaction? What level of error and variability is associated
with the force measurements? How is force data distributed about its mean? What
measurable parameters or features of a force curve are the most important in
characterizing the interactions (e.g., adhesion force)? What is the best way to
identify trends or correlations in these parameters? Clearly, answering these
questions requires statistical techniques, tests and models that determine
appropriate, significant average values of force curve parameters and facilitate
the identification of meaningful force curve features.
This process begins by collecting summary statistics for each data set,
including calculation of means, standard deviations, error values (e.g. confidence
limits) and by plotting histograms for multiple parameters derived from the
curves, including adhesion force and distance of jump to contact. When
comparing parameters from the curves, statistical tests (e.g. ANOVA, t-tests;
correlative tests) may be performed using a standard statistical and data
processing package (e.g., Igo Pro, Wavemetrics, Inc.). Simple regression models
may also be employed to determine important variables that contribute to the
shape of a force curve. To this end, a routine has been written (Kendall and
Hochella, 2003) that rapidly processes force curve data to produce plots of force
versus piezo movement and force versus tip sample separation using the
procedures discussed above. Automated parameter determination, statistical
calculations, whole force curve averaging, autocorrelation calculations (for
identifying quantized force values) and histogram generation are incorporated in
this customized routine written using Igor Pro’s internal programming
environment. The simple parameter extraction module quickly and consistently
identifies features and selects the values using basic, objective criteria such as
maxima thresholds in the differentiated force data and tolerance limits for specific
changes in slope (see Fig. 4).
G. ADVANCED ALGORITHMS
These criteria are appropriate when the bond ruptures and snaps to contact are
large and/or distinct. However, other more advanced algorithms (Baumgartner
et al., 2000; Kasas et al., 2000; Gergely et al., 2001) are required when features are
small, numerous, less distinct (e.g., multiple ligand –receptor interactions) and/or
have the potential to be masked by vibrational and thermal noise. For example
Kasas et al. (2000) employ a fuzzy logic algorithm that enables automated
discrimination of specific, significant adhesions in a retraction curve that might
otherwise be overlooked. The routine assigns a grade to each potential rupture
event, ranking it somewhere between non-specific (0) and specific (1). Assignment
of the grade relies on a priori knowledge of the interaction event morphology, and
uses criteria such as the angle between the jump and the background trace, or
whether or not the jump is U-shaped or V-shaped. This means that the procedure
is operationally defined and first has to be “taught” what the features of interest
look like in order to calibrate it to the system/features being studied.
Gergely et al. (2001) present an algorithm that identifies ruptures based on a
comparison of the minima with neighboring peaks. Selection is controlled by
adjusting an appropriate noise level, m, such that the difference between a feature
and its nearest neighbors must be greater than 2m times the standard deviation of
the force values. Additional smoothing of the force curve is also achieved by
fitting a second order polynomial to a designated amount, p, of consecutive
points. Using this routine, forces measured between human fibrinogen interacting
with a silica surface were processed. By monitoring histograms of inter-rupture
distances selected with successively more rigorous (higher) m values, the authors
were able to detect a significant peak at 20 – 25 nm, a value that corresponds
nicely to the known spacing between two domains in the protein.
H. RELATING BOND CHEMISTRY AND ENERGIES

TO FORCE MEASUREMENTS
Force microscopy measurements intuitively have the potential to describe

energies, E(D), associated with an interaction
R at a small separation, D, by
integrating force over distance, E(D) ¼ F dD. As discussed above, however,
differences in spring constants can produce variable hysteresis and, therefore, can
lead to drastically different energy values. Without fine control of the effective
spring constant, it is difficult to accurately capture a potential in a quantitative
fashion, which is critical for single molecular work. Moreover, if reaction occurs
upon contact (and provided single interactions can be identified in the force
spectra); simple prediction of bond/interaction energies based on rupture forces is
non-trivial. Specifically, it might be postulated that the maximum gradient in the
potential, [dE/dD ]max, is equal to the adhesion or rupture force from the
retraction trace; however, in a seminar paper, Evans and Ritchie (1997) showed
that such a simple correlation is not valid for single molecule interactions, and
more sophisticated theory is required for quantitative comment on the absolute
energetics of a bond using force data.
Before continuing, it must be noted that these findings do not preclude valuable
quantitative and qualitative comparison of force measurements and bond ruptures
to energy parameters. Indeed, early force experiments with various ligand
receptors (e.g., biotin, iminobiotin, avidin, streptavidin combinations) revealed a
correlation between the rupture forces and enthalpy values associated with each
complex (Moy et al., 1994a). This information together with the lack of
correlation between the rupture forces and total free energy suggested the
unbinding was adiabatic and that any entropic contributions to the system (e.g.,
solvation forces) occurred outside of the binding pocket, and were not recorded
with the AFM. Other studies followed, relating thermodynamic parameters to
interaction forces (Chilkoti et al., 1995), as well as many force experiments that
employed “elementary” or averaged rupture forces to compare two or more
systems in a relative fashion (Florin et al., 1994; Frisbie et al., 1994; Dammer
et al., 1996; Noy et al., 1997; Ito et al., 1999; Schmitt et al., 2000; Fiorini et al.,
2001; Lower et al., 2001a; Kreller et al., 2002; Kendall and Hochella, 2003;). The
true value of these studies is their relative quantitative and qualitative comparisons
of force data. These characterize the nature of forces at an interface, demonstrate
surface and molecule recognition, and define relative affinity between two
molecules or between a molecule and a surface. However, Evans demonstrated
that these rupture forces, as absolute values, represent one point in a continuum of
bond strengths (Merkel et al., 1999); and that the detachment force recorded with
the AFM (and other force measuring techniques) is not a singular fundamental
Figure 4 Screen shot of one module (Sensitivity Tweaks) in the force curve processing routine
AFM 4.4 written in Igor Pro, 4.04, WaveMetrics, Inc. (Kendall and Hochella, 2003; some of the base
code was provided by H. Skulason). It is designed primarily for handling force data produced Digital
Instrument’s Nanoscope IIIa MultiMode system. The Sensitivity Tweaks module is designed to
rapidly review and assess how well the normalizaiton routine automatically registers and normalizes
force data to an origin. The normalization procedure includes the identification of a baseline in the
region of no contact, calculating the detector sensitivity from the region of constant compliance and
detecting when the tip and the sample are in contact. The latter is determined using peaks in the
differentiated force wave that are selected based on threshold/sensitivity settings shown in the panel in
the lower right. If initial normalization is unsatisfactory, these settings may be optimized and an auto-
normalization may be run again; or features can be identified manually.
property of the molecular interaction being probed (Evans and Ritchie, 1997;
Evans, 1998; Merkel et al., 1999). Instead, apparent bond strength as estimated by
rupture force is a function of the loading rate (Evans and Ritchie, 1997).
This relationship represents a refinement of a model proposed by Bell (1978)
that predicts an exponential amplification of dissociation kinetics in the presence
of an external force (Merkel et al., 1999). Dissociation of relatively weak
associations can be conceptualized as a particle moving out of a potential well
(bond), over single (simple interaction) or multiple (complex interaction)
activation barriers representing transition states (Fig. 5). Under a zero force
condition the particle will migrate out of the well, through the transition states,
and ultimately to complete dissociation on a time scale that is dictated by thermal
agitation (kBT). A constant external force on the bond, however, expedites the
thermally mediated kinetics and decreases the lifetime of the bond by lowering
the activation barriers in the energy landscape along a projection that is
proportional to the amount of force (Fig. 5) (Evans and Ritchie, 1997). Under a
dynamic load (e.g. a retracting cantilever) where the force, F, increases over time,
t, as loading rate, Rf ¼ dF/dt ¼ ksvc (ks is the spring constant of the system and vc
is the velocity of the cantilever), inner activation barriers are revealed as outer
activation barriers are progressively lowered by the accumulating force. This
Figure 5 The effect of an external force on the energy landscape of a bond. (Modified from
Evans and Ritchie, 1997; Merkel et al., 1999). The minimum of the traces represents a bond or
potential well found along a reaction coordinate, x. Two activation barriers (local maxima) exist
representing transition states that a system must go through during dissociation of the bond. External
force, F, is represented as a mechanical potential, 2(F cos q)x oriented at an angle u to the reaction
coordinate. Increasing force lowers outer activation barriers to reveal the inner maxima. Eventually all
barriers are lowered allowing free diffusion from the initial minimum/bond (Evans and Ritchie, 1997).
phenomenon leads to an intriguing positive correlation between the rupture force,

Fr, and loge(Rf), that is best conceptualized in terms of thermally mediated nature
of the bond rupture kinetics. At small loading rates, activation barriers are
lowered at a rate slow enough for thermal contributions from the medium to be
effective in helping the molecule diffuse out of the well and over the barrier
before higher forces are reached (Gergely et al., 2002). Thus, with an effective
thermal contribution, a lower rupture force is recorded at the time of dissociation.
Under large loading rates (e.g., those typical of many AFM experiments),
activation barriers are lowered fast enough such that dissociation proceeds with
minimal thermal contribution, resulting in a higher rupture force (Gergely et al.,
2002). At ultrafast loading rates on the time scale typical of molecular dynamic
(MD) simulations (1012 s), the entire bonding potential is compromised quick
enough that only frictional drag is recorded as the molecule traverses a
completely “stretched” and coarsened energy landscape (Evans and Ritchie,
1997). Here, loading rates commonly exceed the time scale of unencumbered,
diffusive passage of a molecule from its bonded state, leaving the complex
kinetically trapped as force continues to rise. This was observed during molecular
dynamic simulations of the biotin – streptavidin complex (Grubmuller et al.,
1996). For the biotin system, the time for diffusive passage, tD (e.g., the lifetime
of the bond) is estimated to be 500 ps under a constant force of 280 pN (Evans
and Ritchie, 1997). All activation barriers are lowered at this force such that the
initial minimum (e.g., the original potential well representing the bond) is
exposed allowing direct diffusion out of the well. However, the ultrafast
molecular dynamic simulation loading rate (1.3 £ 1012 pN s21) exceeds 280 pN/
tD, therefore, leaving the complex kinetically trapped in the bound state as rupture
force rises well above 280 pN (Evans and Ritchie, 1997).
With these observations and extensive theory development, Evans recognized
that measurements of rupture forces over a large range of loading rates effectively
probes the lifetime of the interaction under different levels of force while
mapping out energy barrier position and heights in a technique now known as
dynamic force spectroscopy (Evans, 1998). A dynamic force spectra is
constructed by plotting the most probable rupture force of a single interaction,
Fr, versus loge[Rf], where Rf values span several orders of magnitude. Regions of
constant slope defined as fb ¼ kBT/xb represent activation barriers at a distance
of xb from the potential well. Barrier heights Eb can be derived from the intercept
of the slopes at zero force defined as (Evans and Ritchie, 1997):
loge ðRof Þ ¼ 2Eb =kB T þ loge ð fb =tD Þ ð7Þ
where, again tD is the time of diffusive passage, and thus 1/tD represents an
attempt frequency. The attempt frequency is generally not known, but can be
estimated from the damping phenomenon (Evans, 1998). Activation barrier
positions derived from experimental dynamic force spectra of the biotin –avidin
interaction compared well with barriers predicted by molecular dynamic

simulations (Evans and Ritchie, 1997; Izrailev et al., 1997), further emphasizing
the value of coupling force measurements with computer simulations. However,
it must be noted that bridging the gap between the orders of magnitude in the
loading rate of experimental systems versus loading rates used in computer
simulations is not straightforward and extrapolation of calculated rupture forces
to experimentally determined forces must be done with caution (Grubmuller
et al., 1996; Izrailev et al., 1997; Wong et al., 1999). While this was attempted by
Grubmuller (1996), Izrailev (1997) and Evans (1997, 1998) indicate that the real
value of molecular dynamic simulations, in this context, is their potential to
provide clues as to which structural determinants of the interactions contribute to
the activation barriers, thereby providing a qualitative mechanism to account for
dynamic force spectra features. For example, Izrailev et al. (1997) used a
collection of molecular dynamic simulation packages to demonstrate a ground
state avidin – biotin complex that was stabilized by hydrogen bonds between the
biotin head group and polar amino acids (e.g., Tyr33) within the binding pocket.
With the application of an external force, two intermediate states stabilized
by H-bonds with amino acids at different positions in and near the pocket
(e.g., Ser16 and Arg114) are revealed as the ligand is removed — an observation
that was consistent with dynamic force spectroscopy experiments.
Lo et al. (2002) also used a variation of the Bell model to explore the
relationship between rupture force and ambient temperature in the biotin – avidin
system. The experiments were conducted on an AFM with a constant,
millisecond time scale loading rate that was slow enough, compared to the
nano to picosecond time scale of molecular dynamic simulations, to neglect any
frictional energy loss due to viscous drag. The slow loading rates also allowed the
key assumption that thermal equilibrium is achieved at any moment during the
unbinding process. This validated the use of a Maxwell –Boltzmann energy
distribution to describe the thermal energy being supplied to the complex. The
end result is a relationship (Lo et al., 2002):

t
Fi2 ¼ 2DEC kbond 2 2kB Tkbond loge R ð8Þ
tD
that can be fitted to an experimental AFM data set of adhesion forces (Fi), to
derive bond stiffness (kbond), and critical binding energy (DEC). Both the derived
values reflect a summation of the different types of forces that make up the
biotin – avidin interaction (e.g., H-bonds, van der Waals and polar interactions).
Other variables include thermal energy (kBT), and the ratio of the rupture time
(tR, determined from the AFM data) and time required for the ligand to diffuse
out of the binding pocket (tD, estimated independently). Critical binding energy,
DEC, may be related to a dissociation energy (De) by defining a potential to
describe the interaction — in this case a Morse potential was used. Their De
value based on force measurements, 28.4 kcal mol21, compared favorably with
the enthalpy change of the dissociation determined by independent means

(DH ¼ 23.4 kcal/mol) (Swamy, 1995; Lo et al., 2002). Moreover, the enthalpy
value was combined with bond stiffness in additional calculations to determine a
critical displacement magnitude (0.1 nm) that was close to inner barrier position,
xb, value determined with dynamic force spectroscopy (xb ¼ 0.12 nm).
Both temperature dependent and load dependent dynamic force spectroscopy
rely on force measurements of the interaction of a single ligand – receptor pair.
This is accomplished by reducing the density of the sites that are present and
available for bonding (similar to protocols outlined by Florin et al. (1994), Rief
et al. (1997a, b), Marszalek et al. (1998) and Grandbois et al. (1999)), such that 1
in 7 –10 touches results in attachment (Evans, 1998). Governed by Poisson
statistics (Evans and Ipsen, 1991; Williams et al., 1996), 90– 95% of the
attachments are predicted to be single bonds. In addition to the biotin –avidin
linkage, Poisson distributions are common in force measurements associated with
several other systems (Han et al., 1995; Williams et al., 1996; Wenzler et al.,
1997; Lo et al., 1999; Stevens et al., 1999). The probability, P(n), of an attachment
representing n linkages follows a defined, Poisson distribution, making it possible
to extract the n ¼ 1 case from a large number of rupture force measurements (Lo
et al., 2002). Feedback mechanisms are also employed to control impingement on
the sample thereby insuring each approach and retraction cycle has the same
magnitude and history of contact force (Evans, 2001). This is especially important
when making measurements with biomolecules secured to monolayers that can
easily deform on contact to produce various contact areas and configurations, and
ultimately different numbers of attachment (Evans, 2001).
Although successful dynamic force spectroscopy experiments have been
carried out on a single force measuring instrument (AFM) with a range of loading
rates from 100 –5000 pN s21 (Yuan, 2000), due to the exponential relationship
between kinetic rates and barrier energies (discussed above), dynamic force
spectroscopy is optimized when collecting force measurements over a range of
loading rates that are different by orders of magnitude. This can require the use of
several force measuring techniques including laser/optical tweezers for slow
loading rates 1– 10 pN s21, a biomembrane force probe (BFP) for intermediate
rates (10 – 1000 pN s21) and AFM for fast loading rates (104 –106 pN s21)
(Evans and Ritchie, 1997).
I. RELEVANCE OF DYNAMIC FORCE SPECTROSCOPY

TO BIOLOGICAL – INORGANIC INTERFACE
Techniques to extract energy information from force spectroscopy were

developed primarily using the biotin –(strept)avidin system (Moy et al., 1994a;
Evans and Ritchie, 1997) due to its relevance to biological systems, non-covalent
nature, high affinity and extensive history of experimentation and study. Since
then, dynamic force spectroscopy has been applied to other systems, primarily
non-covalent and biological in nature. These include DNA base pair interactions
(Strunz et al., 1999), unfolding of muscle protein domains (Rief et al., 1997a),
antibody – antigen interactions (Schwesinger et al., 2000) and lipid anchoring in
membranes (Evans and Ludwig, 2000) to name a few. Grandbois et al. (1999)
also made an attempt at using the dynamic force spectroscopy concept to measure
the strength of a single covalent bond. Although they produced values for only
one loading rate due to the difficulty of collecting individual covalent interactions
(H. Gaub, personal communication). Their calculation of rupture force
probabilities based on dynamic force spectroscopy methods allowed them to
identify the covalent attachment being terminated as a Si –C bond.
Applying dynamic force spectroscopy concepts to AFM data collected on
environmental systems has great potential to provide new insight on the
interaction energetics and bond chemistry associated with biogeochemical
interfaces. This is, in part, because experimental data collected at the molecular
level to describe surface reactions between single biomolecules and mineral
surfaces is lacking. Traditionally confined to computer simulation (Cygan, 2001),
dynamic force spectroscopy now affords a unique, direct examination of energy
landscapes associated with some of the non-covalent mechanisms (e.g.,
H-bonding) assumed to initiate sorption reactions between minerals and ligands
(Holmen and Casey, 1996), the possible ionic or covalent binding of a metal in a
mineral surface associated with dissolution (Stumm, 1992), reversible and
non-reversible adhesion states of colloids or cells to a surface (Absolom et al.,
1983; Ryan and Elimelech, 1996; Ryan and Gschwend, 1994), mineral and
or metal recognition of a mineral structure by membrane bound proteins
(Lower et al., 2001a).
Perhaps, the true advantage of using dynamic force spectroscopy is realized
when used in a comparative framework, for example, dynamic force spectra of
cell or biomolecule– mineral interaction before and after structural and functional
changes in either (1) the cell surface (e.g., via altered gene expression due to
imposed environmental conditions) or the biomolecule (e.g., via point mutations
in proteins, functional group substitutions/inactivation in ligands) or (2) the
mineral via metal substitution or by comparing isostructural mineral equivalents
or different crystal growth faces. Changes in slopes of the dynamic force spectra
resulting from structural modifications can provide clues as to which proteins,
functional groups or even crystallographic constraints contribute to surface
complex stability or specific activation barriers to binding or detachment.
Concomitant correlation of force values with independently determined
thermodynamic parameters can also provide insight as to whether a surface
attachment or detachment or metal extraction is enthalpy driven or entropically
dominated. And as seen above, a common theme when using dynamic force
spectroscopy is to supplement and validate characteristics of a particular energy
landscape with mechanistic information derived from computer simulations.
The role of these simulations is anticipated to be just as important when applying

these techniques to biogeochemical systems.
IV. FORCES AT THE BIOMOLECULE–MINERAL

INTERFACE
Organic ligands produced by microorganisms such as low molecular weight

organic acids and siderophores have the potential to greatly impact the
geochemistry and ecology of soil environments (Stone, 1997; Hersman, 2000).
These ligands interact with mineral surfaces to form a critical interface that has
implications on biological nutrient/metals acquisition, control of metal toxicity or
even ecological competition (Bossier et al., 1988; Stone, 1997; Neubauer et al.,
2000; Brantley et al., 2001; Kraemer et al., 1999, 2002). Ligands enter into
sorption and desorption reactions with minerals that enhance dissolution or
surface passivation, mediate contaminant mobility, or alter the charge character
of the mineral surface (Barker et al., 1997; Stumm, 1992). As a result this
interface has been studied extensively with bulk experiments and sorption studies
(e.g., Kummert and Stumm, 1980; Ludwig et al., 1995; Yao and Yeh, 1996), and
with surface sensitive techniques such as XPS (Kalinowski et al., 2000), and
Fourier Transform Infrared (FTIR) Spectroscopy (Hansen et al., 1995; Holmen
et al., 1997). Key to sorption and desorption reactions between ligands and
minerals, however, are the forces that bring the ligand into and out of contact with
the surface. Such forces are dependent on the charge character, structure and
reactivity of the ligand, the mineral surface and the intervening solution.
Characterization of these forces using force microscopy holds great potential to
complement information from the existing methodologies listed above in
addition to providing new insight on how ligands interact and coordinate with
mineral bound metals.
In pioneering work, activation of an AFM tip with a specified chemistry
was carried out to examine the biotin –avidin interaction (Florin et al., 1994;
Lee et al., 1994b). Florin et al. (1994) sorbed biotin (an organic ligand) to an
AFM tip, and probed a surface coated with the protein receptor avidin. Force
measurements of this high affinity ligand – receptor system showed a positive
correlation between the elementary quantized adhesion forces detected with
an autocorrelation analysis and the thermodynamic binding affinities.
Specifically, biotin adhesion to the avidin substrate measured 160 pN,
while iminobiotin, which contains a nitrogen substitution in place of an
oxygen and has a lower binding affinity, exhibited adhesions closer to 85 pN.
Several force spectroscopy studies of the biotin system followed (Moy et al.,
1994a, b; Chilkoti et al., 1995; Wong et al., 1999; Lo et al., 2002), along with
other force investigations of biomolecules, including examination of: interactions
between antibodies and antigens (Dammer et al., 1996; Hinterdorfer et al., 1996;
Schwesinger et al., 2000); enzyme activity (Fiorini et al., 2001); proteoglycans
(Dammer et al., 1995); observations on the stretching of polysaccharides (Rief
et al., 1997b; Marszalek et al., 1998) and muscle proteins (Rief et al., 1997a); and
the hybridization of oligonucleotides (Lee et al., 1994a; Mazzola et al., 1999).
Simple functional groups have also been covalently attached to AFM tips in order
to explore more fundamental interactions, such as the forces between methyl,
carboxyl or methyl –carboxyl pairs (Frisbie et al., 1994; Noy et al., 1997).
Specifically, this technique, termed chemical force microscopy (CFM), was used
to identify the nature of the interacting force (H bond, van der Waals,
electrostatic), characterize surface energies and charge distributions, and
generate force maps that showed the spatial arrangement of simple functional
groups or hydrophobic regions on a monolayer or surface, sometimes with
nanometer resolution (Noy et al., 1997).
Collectively, these studies provide a foundation, which allows the application
of force spectroscopy to additional, more complex, natural systems, such as the
ligand/biomolecule– mineral interface that is characteristic of soil environments.
Indeed the same forces (e.g., H-bonding, hydrophobic/hydrophilic forces, the van
der Waals force, steric forces, non-specific and specific interactions) that allow
molecular recognition between biomolecules are also present in ligand mineral
interaction (Israelachvili, 1992; Stumm, 1992). However, to our knowledge, only
two studies, one of which is summarized below, have probed ligand interaction
with a mineral surface using force microscopy (Kendall and Hochella, 2003;
Kreller et al., 2002). A discussion of this burgeoning application begins with a
description of protocols enabling linkage of a ligand to an AFM tip.
A. LIGAND LINKAGE SCHEMES
Devising a suitable linkage scheme to attach the ligand of interest to the AFM
probe can present a significant challenge. Each scheme should be appropriately
tailored to the relevant experimental goal; however, the following summarizes
general considerations. Successful linkage will provide a strong (e.g., covalent or
stronger than the interaction of interest), reproducible bond between the ligand
and the tip while avoiding non-specific interactions associated with the cantilever
material, tip or linker molecule (Wagner, 1998; Fiorini et al., 2001).
Simple ligands such as carboxylate and phosphate groups are commonly
linked as terminations of alkylthiol monolayers that coat the tip (Noy et al., 1997;
Kreller et al., 2002). The ampiphilic molecules of the monolayer not only
provides an anchor for the ligand but also serves as a spacer, providing separation
between the ligand and the tip material thereby reducing non-specific
interactions. Larger ligands and proteins that contain either a free amino or
carboxyl group may be attached using an active ester technique commonly used
to couple two proteins (Cheung et al., 2000; Fiorini et al., 2001; Hinterdorfer
et al., 2002; Kendall and Hochella, 2003). In the presence of a carboxyl group,
1-Ethyl-3-(3-Dimethylaminopropyl) carbodiimide (EDC) together with N-hydro-
yxsuccinimide (NHS) will form a stable, hydrolysis resistant, active succinimidyl
ester that readily forms a peptide bond with an available amino group (Grabarek
and Gergely, 1990). Note that the position of the amino and carboxyl groups can
vary with one being supplied as a self-assembling monolayer (SAM) terminal
group on the tip and the other contributed by the molecule to be attached, or vice
versa. Other linkage protocols employ polyethylene glycol (PEG) as a cross-
linking spacer that is terminated with various functional groups such as
pyridildithiopropionate (PDP). PDP coordinates with thiol groups and nitrilo-
triacetic acid (NTA) which, in combination with various divalent metals, binds to
consecutive histidine residues (Kienberger et al., 2000; Schmitt et al., 2000;
Hinterdorfer et al., 2002). One advantage in using PEG-NTA-Me2þ-His linkage
system is that selection of the divalent metal (Cu2þ, Co2þ, Ni2þ) permits control
of the binding force, and, to a certain extent, the probability of the linkage. In
addition, the NTA-Me2þ-His bond is easily reversible, such that it can be
terminated with the use of EDTA, and then regenerated with the reintroduction of
the free metal (Schmitt et al., 2000). Other workers propose attaching ligands or
molecules via carbon nanotubes that extend from the AFM tip (Wong et al.,
1998a, b, Hafner et al., 2001). This provides ideal spacing between the molecule
and the tip, but more importantly, drastically increases the resolution of the force
spectroscopy (and imaging) due to the nanotube’s extremely small radius of
curvature compared to a traditional Si3N4 tip. Because nanotubes can only be
functionalized at the end termination of the carbon lattice this also places an
important constraint on the orientation and localization of the molecules being
linked. As a result, the probability of capturing a single molecule interaction is
increased, especially when working with lower molecular weight molecules.
It is important that the linkage must not directly interfere with the activity of
the ligand (Fiorini et al., 2001), and thus, electron donor functional groups should
be protected during the linkage reaction. Kendall and Hochella (2003)
accomplished this by inserting a metal (Al3þ) into the ligand (azotobactin)
structure to occupy and protect the chelating groups, while carrying out the
linkage reaction. Once attached to the tip, the azotobactin was reactivated by
removing the Al with high concentrations of a competing ligand (EDTA), a
process that was monitored in a test solution with UV – vis spectroscopy (Fig. 6).
Unfortunately, inherent to fixing a molecule to a surface is a reduction in the
degree of freedom afforded to the molecule’s conformation. This can result in an
alteration or loss of chelation or ligand activity and should be considered. To this
end, control activations are often run in parallel to tip activations, where
monolayers, linker molecules and the biomolecule of interest are reacted with
a flat, Si3N4 or SiOH substrate (Fiorini et al., 2001; Hinterdorfer et al., 2002).
Similar in composition to the tip, the flat test substrates serve as a proxy for
Figure 6 UV –vis spectra showing the transition of Al into and out of the azotobactin (Azb)
structure; corrected for dilution. Upon the addition of Al to the system a characteristic shoulder
appears in the spectra. This shoulder could be eliminated with high concentrations of EDTA. A similar
process was employed to protect and then regenerate the azotobactin chelating groups during linkage
of the siderophore to a hydrazide terminated AFM tip (see Kendall and Hochella, 2003).
tip that are readily probed with AFM imaging, fluorescence and confocal
microscopy, surface plasmon resonance (SPR) and various enzyme and ligand
assays in an effort to assess the success of the linkage reaction; estimate coverage,
density and footprint area of the monolayer-biomolecule construct; and evaluate
activity retention in the immobilized biomolecule (Fiorini et al., 2001).
B. SIDEROPHORES AND OXIDE SURFACES
Kendall and Hochella (2003) collected force signatures between a ligand

(siderophore) and two mineral bound metals (Fe(III) and Al(III)) in an attempt to
examine the mechanism of siderophore-mediated dissolution of oxide surfaces.
Siderophores are ligands produced by microorganisms to assimilate the essential
nutrient ferric iron, in spite of its extreme insolubility in near surface,
circumneutral environments. The aqueous chemistry of siderophores has long
been studied (Winkelmann, 1991), and it is recognized that their effectiveness in
acquiring iron, can, in part, be attributed to a thermodynamic binding affinity for
Fe(III) (aq) that has a magnitude above that for other metals, including Al(III).
Only recently, however, it was recognized that, in addition to the formation of
stable, aqueous iron complexes, siderophores can release iron from minerals
(Seaman et al., 1992; Watteau and Berthelin, 1994; Hersman et al., 1995;
Holmen and Casey, 1996; Liermann et al., 2000; Maurice et al., 2000).
The mechanism of this release, however, is not clearly defined.
Force microscopy of the pyoverdin type siderophore azotobactin interacting
with iron and aluminum oxide surfaces showed a unique relationship between
ligand – metal affinity and adhesion forces (Kendall and Hochella, 2003).
Average adhesion forces between azotobactin and goethite (a-FeOOH) at pH 7
were 2 –3 times the value between azotobactin and goethite’s isostructural
Al-equivalent, diaspore (a-AlOOH) (Fig. 7a). A similar force relationship was
also observed between the trihydroxamate siderophore deferoxamine (DFO) and
each oxide surface (Fig. 7b). Control experiments where each mineral surface
was probed with a SAM coated tip lacking the azotobactin molecule produced
force signatures that were almost identical, indicating the distinction in the force
signature between diaspore and goethite could be attributed to the presence of the
azotobactin on the tip.
Force measurements collected under various solution conditions (e.g., pH,
ionic strength and soluble iron concentrations) and at different sample locations
on the mineral extended the characterization of the ligand –mineral interaction
and helped identify the source of discrepancy in adhesion values associated with
each oxide. As a first guess, it could be hypothesized that the forces of interaction
Figure 7 Force spectra showing the interaction of two siderophores (a) azotobactin and (b)
deferoxamine (DFO) with goethite (FeOOH) and diaspore (AlOOH) surfaces. Note the large increase in
the adhesion force between each siderophore and goethite and versus the adhesion value for diaspore.
are dominated by an electrostatic component; and that the difference in the

adhesion values between diaspore and goethite, simply reflects variation in the
charge character of each mineral. Although point-of-zero-charge (pzc) literature
values for goethite (pzc 7 –9) and diaspore (pzc 7 –8) are similar, suggesting both
should be neutral or slightly positively charged (Cornell and Schwertmann, 1996;
Kosmulski, 2001), it is possible that our model system deviates from pristine
charge conditions, such that the goethite is positive and the diaspore is slightly
negative. With a net negative charge predicted for azotobactin at pH 7 (pKa
hydroxycarboxylate ¼ 4 –5; Telford and Raymond, 1996), this could result in a
larger adhesion force for goethite presumably due to a stronger electrostatic
interaction. Measurements at lower pH and different sample locations, however,
suggest otherwise. At pH 3.5, far from the pzc value of each mineral, and where
the azotobactin is anticipated to be neutral, the same 2 –3-fold increase in
adhesion values is observed. Moreover, the azotobactin-goethite/azotobactin-
diaspore force relationship remained intact when comparing adhesion distri-
butions representative of different sample locations on each mineral surface,
where anomalous charge distributions and changes in microtopography are
expected. Overall, similar to observations made with the biotin ligand system
(Moy et al., 1994a; Chilkoti et al., 1995; Izrailev et al., 1997), adhesion values
upon retraction appear to be relatively independent of protonation equilibria, and
may reflect a specific interaction between the siderophore oxygens and the metal
contained in each mineral. The discrepancy in adhesion for goethite versus
diaspore can then be explained by differences in the electronic character of each
metal (e.g., Fe(III) versus Al(III)), where the more electronegative ferric iron will
behave as a harder acid with a higher affinity for the oxygens. In additional
experiments with goethite only, this surface affinity was readily disrupted with
the addition of soluble iron (Fig. 8). Here, increased [FeCl3(6H2O] (pH 3.5) led to
a saturation of the ligand as the soluble iron out competed the mineral iron for the
siderophore oxygens, resulting in lowered adhesion values.
This does not discount an electrostatic component to the azotobactin –oxide
interaction. Indeed, decrease in the jump to contact distance with increasing ionic
strength thought to reflect a collapse in the double-layer associated with the
mineral surface (Noy et al., 1997; Lower et al., 2000), confirms the effect of
charge, especially upon approach. Instead, force evidence suggests a balance
between electrostatics dominating the approach and more specific interactions
directing surface adhesion; a scheme that is embodied in the following
observation — goethite force signatures at pH 7 often show a long range,
electrostatic repulsion on approach that was equal to or significantly lower in the
diaspore signatures; yet, the goethite adhesion force averaged 3.81 nN compared
to 1.38 nN for diaspore.
Force data also provided information on which azotobactin functional groups
might be important in the interaction. Distinctive plateaus were commonly
observed after retracting the tip ~6 – 7 nm from the surface (Fig. 9). These features
FORCES BETWEEN MINERALS AND BIOLOGICAL SURFACES
Figure 8 Plot showing decrease in azotobactin– geothite adhesion forces with increasing concentrations of added soluble iron (FeCl3 6H2O). Measurements
were collected at pH 3.5 to minimize the precipitation of solid iron phases upon the addition of the iron chloride. Taken from Kendall and Hochella (2003).
33
34
T. A. KENDALL AND S. K. LOWER
Figure 9 Plateau feature common in many retraction curves while probing oxide surfaces with an azotobactin activated AFM. It is suggested that this feature
may represents the extension of the azotobactin and linker molecule during separation from the mineral surface as shown in the inset (not to scale). Also shown in
the inset is the geometry of the linkage of the siderophore to the tip. Modified from Kendall and Hochella (2003).
are thought to represent the energy absorption associated with the combined
extension and stretching of azotobactin’s polypeptide chain and the molecule used
to link the ligand to the tip. Using an approximation of 0.38 nm amino acid, a quick
calculation shows that azotobactin’s fully outstretched length of , 3.8 nm,
together with an additional 3 nm from the linker molecule gives a value that is
close to the 6 – 7 nm observed in the force signatures. This distance, then, requires
that azotobactin’s terminal homoserine group serves as an anchor to the surface,
providing a strong, persistent link in the interaction. This coincides with other
reports that, in aqueous systems, the adjacent hydroxamate group initiates
chelation (Telford and Raymond, 1996; Albrecht-Gary and Crumbliss, 1998).
Additionally, considering its terminal position on the molecule, it is feasible that
the homoserine group is a dominant component during surface interaction.
Finally, these force microscopy results give cause to reassess the role of large
ligands, such as azotobactin, in dissolving minerals. Instead of serving as an Fe
shuttle between smaller ligands that interact with the surface and the cell, the
force evidence demonstrating azotobactin’s strong surface affinity presents a
distinct possibility of the relatively large molecule entering into a strong, stable
complex with the mineral. As seen above, the force data also allow comment on
the nature of the association with the surface. Steric constraints imposed by
ligand size, structure and conformation, together with the limited access to an
iron atom contained on a mineral surface, would certainly preclude the
hexadentate coordination characteristic of the siderophore-aqueous complex
(Holmen and Casey, 1996; Hersman, 2000; Cocozza et al., 2002). Instead,
plateau features in the retraction curves suggest a strong coordination formed
by a single oxygen pair that terminates the azotobactin molecule as one
possibility. Recent, ongoing MD simulations, in collaboration with U. Becker
(University of Michigan), confirm this possibility, as well as the extended
dimensions of the azotobactin-linker construct. Interestingly enough, however,
simulations reveal that the spacing between the two, chelating hydroxamate
oxygens is sufficient to allow individual coordination with neighboring irons in
the goethite structure (Fig. 10). Siderophore– oxide interaction continues to be
examined with molecular dynamic simulation as well as dynamic force
spectroscopy.
V. FORCES AT THE BACTERIUM–MINERAL

INTERFACE
A. FORCE MICROSCOPY TECHNIQUE USING WHOLE CELLS
The fundamental forces between a bacterium and mineral surface are central
to understanding the intricacies of interfacial phenomena such as bacterial
Figure 10 Molecular model of azotobactin (with linker molecule) interacting with a goethite
surface. Simulations were completed using Cerius2, Accelrys, Inc. Arrows point to terminal
hydroxamate group oxygens interacting and coordinating with irons (balls) in the lattice. Note the
spacing of the siderophore oxygens allow for “bonds” (i.e., Fe–O distances ,2.1 Angstroms) with
neighboring irons. With this coordination, the cross-distance between a siderophore oxygen and an
iron diagonally across is over 3 Angstroms.
adhesion to minerals and dispersal in the environment (van Loosdrccht et al., 1989;
Fletcher, 1996), mineral growth and dissolution (Myers and Nealson, 1988;
Hiebert and Bennett, 1992; Schultze-Lam et al., 1992; Roden and Zachara, 1996;
Fortin et al., 1997), biofilm formation (Lawrence et al., 1991; Davies et al., 1998),
and bacterial affinity for or recognition of specific mineral surfaces (Ohmura et al.,
1993; Fleminger and Shabtai, 1995; Bhosle et al., 1998; Dziurla et al., 1998;
Edwards et al., 1998). A myriad of physicochemical interactions occur at
biological – mineral interfaces in nature, due to (1) the mosaic of spatially discrete
macromolecular cell wall structures on bacteria, (2) the dynamic nature of these
structures, and (3) the diversity of mineral surface functionality, topography, and
crystallography (Lower et al., 2000). As discussed above in section II, these
interactions are expected to be governed by the cumulative effects of
intermolecular forces (Israelachvili and McGuiggan, 1988; Israelachvili, 1992;
Kendall, 1994; Butt et al., 1995; Fletcher, 1996; Gay and Leibler, 1999).
However, acquiring even an elementary appreciation of these forces presents a
daunting challenge, primarily due to the minute scale at which these interfaces
must be probed, and the difficulty in developing a technique that preserves the
natural intricacies of the bacterial surface (Lower et al., 2000).
Measurement of fundamental forces between whole bacterial cells and
inorganic phases can be conducted in one of two ways with force microscopy.
The first involves “fixing” cells to a solid substrate (e.g., a glass slide) and
probing these cells with a force-sensing cantilever. The simplest setup makes use
of the sharp tip that is integrated into most force microscopy cantilevers (see
above). In many instances, however, this is not ideal because these tips are not
well constrained with respect to their geometry and/or area of contact. As shown
in section II, this greatly influences force measurements thereby making it
difficult to compare measured data to theoretical force models, and impedes the
comparison of data collected with different tips. To overcome the limitations
imposed by using a sharp tip, Ducker et al. (1991) devised a simple yet ingenious
solution. They created a “colloidal tip” by attaching a glass bead to the end of a
force-sensing cantilever. This bead was then used to probe a flat silicon surface
(Ducker et al., 1991, 1992), although such a “colloidal tip” could also be used to
probe microbial cells on a glass surface. A number of companies, such as Duke
Scientific Corporation, Polysciences Incorporated, and Bangs Laboratories
Incorporated, sell spheres ranging in size from nanometers to micrometers.
A major drawback to this scenario, however, is that it limits the inorganic phases
that can be utilized to those materials commonly used to make tips (e.g., silicon
and silicon nitride) or beads (e.g., plastic and glass). With the exception of silica
(e.g., glass beads), minerals or other inorganic phases cannot be attached to a
force-sensing cantilever. Therefore, interactions between bacteria and minerals
much employ another technique. That is, the cells must be linked to the force-
sensing cantilever, which is then used to probe a particular face on a mineral
crystal or other surface.
The first cell to be linked to a force-sensing cantilever was a large mammalian
cell (Antonik et al., 1997). This cell was not actually “attached”, rather it was
induced to grow on the cantilever. The researchers conducting this experiment
were not interested in measuring forces, which was fortunate because cells grew
on both the top and bottom surfaces of the cantilever. Hence, the cell growth
would have affected not only the spring constant of the lever, but it would also
alter the optical lever detection system. Nonetheless, this opened the door to a
number of other protocols of linking cells to a force microscope cantilever.
It is a difficult challenge to link microbial cells on the order of 1 mm to the end
of a cantilever. Early attempts relied on the attachment of cells that had been
chemically fixed or treated with harsh chemicals (e.g., gluteraldehyde) (Razatos
et al., 1998a, b). While these investigations produced some very intriguing force
measurements, this type of linkage protocol is often undesirable because the cells
are killed in the process. Further, chemicals such as gluteraldehyde are known to
denature proteins and other macromolecules. Another method was developed that
allowed the force-sensing cantilever to support bacterial cells in a living, native,
fully functional state — thereby creating “biologically active force probes”
(Lower et al., 2000, 2001b). A polycationic linker molecule (e.g., aminopropyl-
triethoxysilane or polylysine) can be used to link living bacteria to a small glass
bead that is then attached to the cantilever, or the bacteria can be linked directly
onto the cantilever itself (Lower et al., 2001b). Polycationic linkers work well
because many bacteria are negatively charged over a wide range of pH
conditions. Hydrophobic molecules (e.g., octadecyltrichlorosilane) are also
attractive linkers because many microorganisms have hydrophobic surfaces.
Techniques similar to affinity chromatography (e.g., see Pleuddemann, 1991;
Egger et al., 1992; Rezania et al., 1999) may be employed to design tailor-made
linker molecules (e.g., ligand –receptor or antibody – antigen) that work on
virtually any bacterial species. The use of polycationic linkers, or similar
molecules, preserves the natural conformation, structure, and function of the
macromolecules on the microbial surface. When live cells are used (i.e., a
biologically active force probe), force measurements may be collected under
different physiological or environmental conditions in real time (Lower et al.,
2000, 2001a, b). Finally, for larger microbial cells such as yeast or fungal cells,
the “colloidal tip” technique (see above) can be used to glue a single cell to the
end of a cantilever (Bowen et al., 1998b).
Using these techniques, a number of groups have used force microscopy to
measure intermolecular forces at the bacterium – mineral interface (Ong et al.,
1999; Bowen et al., 1998a; Razatos et al., 1998a, b; Bowen et al., 2000a, b;
Camesano and Logan, 2000; Lower et al., 2000, 2001a, b). In our laboratories, we
have used biological force microscopy (Lower et al., 2000) to measure
intermolecular forces between living bacteria (e.g., E. coli, Burkholderia
cepecia, and S. oneidensis) and inorganic phases (e.g., muscovite, goethite,
diaspore, graphite, and glass) in solutions of varying ionic strength, pH, and
oxygen concentration (Lower et al., 2000, 2001a, b). Below we will examine the
force– distance relationships at the E. coli – muscovite and S. oneidensis– goethite
interfaces. We will concentrate on the forces measured upon approach of a
bacterium towards a mineral in the case of the former. For the latter system
(i.e., S. oneidensis – goethite), we will explore forces measured as the two

surfaces are pulled apart or retracted from one another.
B. FORCES BETWEEN ESCHERICHIA COLI AND MUSCOVITE
Figure 11 shows the interaction between E. coli and the (001) surface of
muscovite as the sodium chloride (NaCl) solution was exchanged five times
between low (, 1025 M) and high (~102l M) ionic strength. While both
approach and retraction forces were measured, shown in Fig. 11 are only the
forces detected upon approach of the mineral towards living cells on a
biologically active force probe. At low ionic strength, repulsive (positive sign)
forces were detected at a separation of approximately 100 run. This repulsive
interaction increased exponentially (see below) to a maximum value of , 30 –
35 nN at contact. At high ionic strength, the magnitude of repulsion was
significantly less as was the range of separation over which force interactions
took place. The two surfaces did not “feel” one another until they were within
15– 20 nm of separation. As with the measurements at low ionic strength, an
exponential force appears to dominate at high ionic strength. It is important to
note that these data shown in Fig. 11 span the entire range of measurements for
literally hundreds of force – distance curves taken as a solution was exchanged
Figure 11 Force–distance relationship between the basal plane surface of muscovite and E. coli
in solutions of high (open symbols across lower portion of figure) or low (closed symbols across upper
portion of figure) ionic strength. Shown for each solution condition are five data curves that span the
entire range of measurements for literally hundreds of force-distance curves. The lines correspond to
the DLVO model prediction at high (dotted) or low (solid) ionic strength. Repulsive forces have a
positive sign; whereas attractive forces have a negative sign. Only those forces measured upon
approach of the mineral towards the bacteria are shown. See text for discussion.
several times between high and low ionic strength. Hence, the measurements are
reproducible.
Results can be interpreted with the so-called DLVO theory (Derjaguin and
Landau, 1941; Verwey and Overbeek, 1948). This theory describes forces (F) as
a function of the distance (D) (e.g., between a bacterium, treated as a sphere and a
mineral, treated as a flat plane) as the sum of the electrostatic and van der Waals
forces (Ducker et al., 1991; Israelachvili, 1992; Butt et al., 1995; Muller and
Engel, 1997):
4psbacteria smineral R 2kD HR
FDLVO ðDÞ ¼ Felectrostatic ðDÞ 2 Fvdw ðDÞ ¼ e 2 a2
110 k 6D
where s is the surface charge density (C m22), R is the radius of a cell (or in this
case the radius of the bacteria coated bead attached to the cantilever), 1 is the
dielectric constant of water (78.54 at 298 K), 10 is the permittivity of free space
(8.854 £ 10212 C2 J21 m2l), k is the inverse Debye length (Debye length
, 1 nm at 102l M and , 100 nm at 1025 M; see above), and Ha is the Hamaker
constant. For the model results plotted in Fig. 11, Hamaker’s constant was
10221 J (Vigeant et al., 2002); surface charge density of the bacterium was
estimated using Eq. (3) as 2 0.001 or 2 0.04 C m22 at low or high IS,
respectively (surface potential measurements were taken from Camesano and
Logan (2000), Ong et al. (1999) and Vigeant et al. (2002), and the surface charge
density of the mineral was estimated using Eq. (3) as 2 0.004 or 2 0.2 C m22 at
low or high ionic strength, respectively (surface potential measurements were
taken from Ducker et al. (1992) and Ong et al. (1999)).
Figure 11 compares the measured forces with those predicted by DLVO theory.
Ionic strength (approx. 102l versus 1025 M) appears to have a strong effect on the
interactions between E. coli and muscovite. This is because higher salt
concentrations cause the electrostatic double-layer to become thinner (i.e.,
surfaces cannot “feel” one another until they get very close). Further, the increased
concentration of counter ions at high ionic strength effectively screens the
negative charges on both surfaces, thereby resulting in smaller magnitude forces
of repulsion. These particular measurements are fairly consistent with DLVO
theory. However, there are some important discrepancies. For example, at low
ionic strength the attractive van der Waals force is expected to dominate the
interaction at separations less than 5 nm. However, measurements reveal that E.
coli and muscovite do not exhibit attraction even at the closest approach. Indeed,
adhesion forces were not detected when E. coli and muscovite were forced
together and subsequently pulled apart at low ionic strength (Lower et al., 2000).
This suggests that electrostatic and/or other repulsive forces (e.g., solvation
interactions) dominate this particular interaction.
Many other force measurements conducted in our laboratories, suggest that
electrostatic and van der Waals forces are not the only intermolecular forces at
the bacterium – mineral interface (S. Lower, unpublished results). Others have
attempted to invoke extended-DLVO models to explain deviations from purely

van der Waals and electrostatic forces and fit model predictions to measurements
(see e.g., Ong et al., 1999; Camesano and Logan, 2000). While these
investigations may be valid, one needs to remember that force models are
sensitive to the geometric shapes of interacting particles as well as the roughness
of surfaces and contact area (Israelachvili, 1992) all of which are difficult to
rigorously define or control for minerals and cells with biopolymers. Further,
DLVO was developed to describe the phenomena between inanimate particles
rather than living cells that have exquisite control over the expression of surface
macromolecules. Seemingly simple concepts such as “contact” become difficult
to define for cells having polymers of varying length, which extend for some
distance beyond the cell wall. Further, living cells and/or surfaces with polymers
are expected to show a time dependent adhesion (measured by Lower et al.,
2001a) as biopolymers diffuse into the cell wall and reorient with respect to
another surface (Beveridge, 1999; Leckband and Israelachvili, 2001). The true
value may not be in whether a model perfectly fits data, but the most definite
answer comes when the measurements contradict the theory, thereby disproving a
particular construct and suggesting that other forces are responsible for a
particular bacteria –mineral interaction. As stated by Oreskes et al. (1994)
scientific investigations are at their best when one combines experimental
measurements and model predictions to challenge existing formulations. Hence,
there is a great need to test such models by comparing theories to precise force
measurements using “model” microorganisms and minerals. Only then will we be
able to understand how all of the various intermolecular forces (e.g., electrostatic,
van der Waals, hydration, hydrophobic, and steric interactions) govern
interactions at the bacterium –mineral interface.
C. FORCES BETWEEN SHEWANELLA ONEIDENSIS AND GOETHITE

OR DIASPORE
The forces required to pull the mineral and bacteria apart (i.e., retraction data)
are not shown in Fig. 11. In fact, a very strong adhesion force was detected
between E. coli and muscovite, but only at high ionic strength (Lower et al.,
2000). Aside from being a notable example of a situation that DLVO theory often
cannot explain, retraction data provide an immense amount of information about
the adhesive strength and structural properties of specific biopolymers on a cell’s
surface. Recently, we interpreted these retraction data for studies of bacterial
adhesion and electron transfer reactions between S. oneidensis (a dissimilatory-
iron-reducing-bacteria) and the minerals goethite (FeOOH) or diaspore
(AIOOH) under aerobic and anaerobic solution conditions (Lower et al.,
2001a; Lower et al., 2002). S. oneidensis is capable of using either oxygen or
ferric iron in the crystal structure of iron oxyhydroxides as a terminal electron
acceptor (Nealson and Saffarini, 1994). We used force microscopy to determine

whether a microorganism could discriminate between two very similar minerals
(diaspore and goethite) under anaerobic conditions, when electron transfer is
expected to occur between S. oneidensis and iron containing minerals.
A mineral crystal, mounted on a piezoelectric scanner, approached live bacteria
on a biologically active force probe at a rate that was comparable to the natural
velocity of motile bacteria. Once contact was established, the two surfaces were
pulled apart resulting in retraction data. Figure 12 illustrates the retraction profile
for S. oneidensis and goethite versus diaspore under anaerobic or aerobic
conditions (Lower et al., 2001a). The intricate details of these curves and the entire
data set provide valuable information about intermolecular forces and structures
at the bacterium –mineral interface. Initially the entire retraction data were
characterized by integrating the force with respect to distance. This provided
quantitative energy values associated with adhesion. The retraction curves were
further analyzed by the worm-like chain model (see above) to establish a
correlation between specific bridging polymers and unique signatures in the
retraction curves.
Energy values determined from retraction curves similar to those shown in
Lower et al. (2001a) revealed that S. oneidensis had a higher affinity for diaspore
(versus goethite) under aerobic conditions (Lower et al., 2001a). However, under
anaerobic conditions the bacteria exhibited a significant increase in affinity for
goethite (see Lower et al., 2001a); whereas the adhesion energy for diaspore was
indifferent to oxygen concentrations (Lower et al., 2001a). The attractive energy
between S. oneidensis and goethite was 30 aJ (aJ ¼ 10218 J) and 130 aJ under
aerobic and anaerobic conditions, respectively. Further, the energetic affinity
between goethite and S. oneidensis also increased with contact time under
Figure 12 Measured force-distance (or force-extension) relationship between living cells of S.

oneidensis and diaspore (AlOOH) under anaerobic conditions (solid circles). Interactions with
goethite (FeOOH) under aerobic (“X” symbols), or anaerobic conditions (open squares). Also shown
is the theoretical prediction for the unravelling of a 150 kDa protein that may tether the bacteria to the
surface of goethite (see text). Attractive forces, shown here, have a negative sign. Shown are only
those forces measured as the mineral is pulled away from the bacteria (retraction data). Modified from
Lower et al. (2001a).
anaerobic conditions. This provided quantitative evidence suggesting that this

microorganism recognized a particular inorganic surface such that it localized
and/or produced biopolymers to mediate contact with goethite under anaerobic
conditions.
This idea was further tested by using a “bridging polymer” model to interrogate
the intricate details (e.g., the saw-tooth pattern) in the retraction data. As shown
above, for the worm-like chain model, linear polymers such as proteins are
expected to unravel according to Eq. (4). According to this model, one needs only
the persistence length and the contour length to describe the force as a function of
the distance a polymer is extended. This model describes a physical process
similar to that which is recorded in the retraction data of force microscopy. A
protein, for example, in the cell wall of a bacterium makes contact with a mineral.
The mineral is then pulled away from the bacterium causing the protein to unravel
until it is completely extended at which point the protein is either ripped from the
cell wall or it breaks free of the mineral surface and recoils into the cell surface.
The outer surface proteins of Shewanella are well characterized. Shewanella
is known to have proteins on its outer membrane that mediate contact with iron
oxyhydroxides (Caccavo, 1999; Caccavo et al., 1997; Das and Caccavo, 2000).
Several of these proteins are putative iron reductases, which are expected to
make physical contact with goethite such that they can transfer electrons across
the organic – inorganic interface (Arnold et al., 1990; Myers and Nealson, 1990;
Myers and Myers, 1992, 1993, 1997; DiChristina and Delong, 1994; Nealson
and Saffarini, 1994; Roden and Zachara, 1996; Myers and Myers, 1998, 2000,
2001). In fact, four putative iron reductase proteins have been characterized
according to their mass and/or genetic sequence (Myers and Myers, 1997, 1998,
2000, 2001, 2002). The worm-like chain model was used to predict the way in
which each of these four proteins (ranging in size from , 50 to 150 kDa;
(Myers and Myers, 1997; Myers and Myers, 1998) would unfold. The molecular
mass of each protein was used to estimate its overall length according to the
following conversion, , 110 Da per amino acid residue (Voet and Voet, 1995).
The persistence length of each amino acid, defined as the distance between two
adjacent Ca, is equal to 0.38 nm (Karlsson et al., 1996; Muller et al., 1999;
Myers and Myers, 2001). As shown in Fig. 12, the saw-tooth pattern at
approximately 500 nm corresponds to the force-extension profile of one of the
four putative iron reductase proteins. This profile was reproducible suggesting
that the cell wall protein was not ripped from the bacterium, but rather retained
its native conformation after multiple extensions. This unique signature was
present only for goethite under anaerobic conditions where it was detected in
, 80% of the retraction curves, but only when the bacterium was given some
period of time to make contact with the surface of goethite (Lower et al.,
2001a). This suggests that the bacterium required time to “recognize” the
mineral surface and subsequently create and/or mobilize a specific protein to the
area of contact with goethite.
VI. FUTURE WORK
On Earth, literally millions of different species of prokaryotes may interact

with any of the thousands of different minerals. An interface is formed at the
junction of a bacterium and a mineral surface that is complex, dynamic, and by its
very nature, nanoscale in size. This is because bacteria are living cells that have
mastered the art of synthesizing fully functional structures (e.g., lipids, proteins,
polysaccharides) and utilizing properties that exist only at the nanometer scale.
The study of the interface between minerals and microorganisms requires a
unique fusion of geomicrobiology and nanoscale science. What are the
fundamental forces that control the binding of a silanol group on a mineral
surface to a carboxylic group in a bacterium’s cell wall? How does the density
and distribution of functional groups on a crystal face influence the way
microorganisms sense mineral surfaces? Do bacteria express specific outer
surface proteins to interact with certain minerals? How do bacteria modulate
forces of interaction between themselves and minerals (or other bacteria) to either
enhance or inhibit adhesion and subsequent biofilm formation? Researchers must
be able to thoroughly explore both sides of the interface (i.e., the bacterium and
the mineral) and the fundamental nanoscale forces in the intervening region to
discover phenomena that exist only in the nanospace between a microorganism
(or microbially produced polymers) and a mineral surface.
As mentioned earlier, application of force microscopy to the biogeosciences is
in its infancy, and there exists many other uses and unexplored possibilities of
force experiments with ligands, microorganisms, and minerals. Structural
elements within a particular biomolecule/ligand may contribute to its ability to
bind to a surface or promote dissolution, or chelate dissolved or mineral bound
metals (Stumm, 1992; Ludwig et al., 1995; Nubel et al., 1996). One can envision
collecting a force signature for a large ligand interacting with a mineral, followed
by collection of spectra associated with several, individual cognate functional
groups associated with the ligand. Comparison of the whole ligand, baseline
spectrum with the individual component spectra could reveal which functional
groups are dominating the interaction with the mineral. Or, a similar process
could be achieved by making force measurements after successive chemical
modification of the original ligand structure. Such modification might include
inactivation of a specific functional group with a residue-specific reactive reagent
(Voet and Voet, 1995), or an amino acid substitution resulting from alteration of
the genes associated with the biosynthesis of the molecule. Again changes in the
force signature with each modification might help determine the critical moieties
contributing to the interaction.
Force maps (Noy et al., 1997) are also possible using ligand activated tips.
Here, the contrast in the map may be supplied by the differential adhesion
between the ligand and various metals that are associated with a surface.
For example, a map constructed with a siderophore activated tip might show
large adhesions in areas of high concentrations of trivalent metals such as
Fe(III) or Al(III) and lower adhesions for divalent metals such as Cu(II),
Zn(II) or Fe(II). Given the spatial resolution of the AFM, such images could
be useful for identifying contaminant distribution on a surface or pinpointing
impurity concentrations on a mineral growth face, both on a nanometer scale.
CFM is traditionally carried out in a fluid cell (Digital Instruments) that
allows direct observation of ligand –surface interaction under environmentally
relevant conditions with pico- to nanonewton force resolution and a spatial
resolution of tens of nanometers down to potentially the atomic level.
Changes in the forces of interaction with solution composition provide
important information about the structure and charge character of the ligand
and mineral surface, and the nature of the interaction between the two. While
the effect of solution composition (e.g. pH) on ligand sorption can be
monitored with force measurements using a force titration (Kreller et al.,
2002). The sensitivity of this technique also allows small changes in mineral
solubility and associated metal concentrations, pH, and ionic strength to be
detected (Kendall and Hochella, 2003). Given the spatial resolution
mentioned above, this opens up the possibility of using this technique to
detect localized solution micro- or even nanoenvironments associated with a
surface.
Finally, force investigations with living microorganisms are rich with
possibilities. For example, one could measure forces of adhesion using wild-
type stains versus mutants that are incapable of producing specific cell wall
macromolecules. These data may result in unique force signatures characteristic
of particular biomolecules. Force measurements could also be coupled to other
techniques such as confocal scanning laser microscopy. This provides the
potential to collect force measurements concurrent with fluorescence obser-
vations of the distribution and localization of cell wall macromolecules.
ACKNOWLEDGMENTS
SKL acknowledges the support of the National Science Foundation, the

Department of Energy, the American Chemical Society, and the General
Research Board of the University of Maryland. SKL would also like to thank
J. Tak for support. Funding was provided to TAK by a GAAN Fellowship (U.S.
Dept. of Education), the NSF’s Nanoscale Science and Engineering (NSE)
Program (EAR 01-03053), and the Department of Energy’s OBES Geosciences
Program (DE-FG02-99ER 15002). TAK acknowledges the support of Michael
F. Hochella, Jr. (Virginia Tech).
REFERENCES
Absolom, D. R., Lamberti, F. V., Policova, Z., Zingg, W., van Oss, C. J., and Neumann, A. W. (1983).
Surface thermodynamics of bacterial adhesion. Appl. Environ. Microbiol. 46, 90–97.
Albrecht-Gary, A.-M., and Crumbliss, A. L. (1998). Coordination chemistry of siderophores:
Thermodynamics and kinetics of iron chelation and release. In “Iron transport and storage in
microorganisms, plants, and animals” (H. Sigel, Ed.), Vol. 35, pp. 239–327. Marcel Dekker,
New York.
Antonik, M. D., D’Costa, N. P., and Hoh, J. H. (1997). A biosensor based on micromechanical
interrogation of living cells. IEEE Engng. Med. Biol. 16, 66 –72.
Aoki, T., Hiroshima, M., Kitamura, K., Tokunaga, M., and Yanagida, T. (1997). Non-contact scanning
probe microscopy with sub-piconewton force sensitivity. Ultramicroscopy 70, 45– 55.
Arnold, R. G., Hoffmann, M. R., DiChristina, T. J., and Picardal, F. W. (1990). Regulation of
dissimilatory Fe(III) reductase activity in Shewanella putrefaciens. Appl. Environ. Microbiol. 56,
2811–2817.
Ashby, P. D., Chen, L., and Lieber, C. M. (2000). Probing intermolecular forces and potentials with
magnetic feedback chemical force microscopy. J. Am. Chem. Soc. 122, 9467–9472.
Barker, W. W., Welch, S. A., and Banfield, J. F. (1997). Biogeochemical weathering of silicate
minerals. In “Geomicrobiology: Interactions Between Microbes and Minerals” (K. H. Nealson,
Ed.), Vol. 35, p. 448. Mineralogical Society of America, Washington DC.
Baumgartner, W., Hinterdorfer, P., and Schindler, H. (2000). Data analysis of interaction forces
measured with the atomic force microscope. Ultramicroscopy 82, 85 –95.
Bell, G. I. (1978). Models for specific adhesion of cell to cells. Science 200, 618 –627.
Beveridge, T. J. (1999). Structures of Gram-negative cell walls and their derived membrane vesicles.
J. Bacteriol. 181, 4725–4733.
Bhattacharjee, S., Chen, J. Y., and Elimelech, M. (2000). DLVO interaction energy between
spheroidal particles and a flat surface. Colloid. Surf. A: Physicochem. Engng. Aspects 165,
143–156.
Bhosle, N., Suci, P. A., Baty, A. M., Weiner, R. M., and Geesey, G. G. (1998). Influence of divalent
cations and pH on adsorption of a bacterial polysaccharide adhesion. J. Colloid. Interf. Sci. 205,
89–96.
Binnig, G., Quate, C. F., and Gerber, C. (1986). Atomic force microscope. Phys. Rev. Lett. 56,
930–933.
Bossier, P., Hofte, M., and Verstraete, W. (1988). Ecological Significance of Siderophores in Soil. In
“Advances in Microbial Ecology” (K. C. Marshall, Ed.), Vol. 10, pp. 385–414. Plenum Press,
New York.
Bowen, R. W., Hilal, N., Lovitt, R. W., and Wright, C. W. (1998a). Direct measurement of interactions
between adsorbed protein layers using an atomic force microscope. J. Colloid Interf. Sci. 197,
348–352.
Bowen, R. W., Hilal, N., Lovitt, R. W., and Wright, C. W. (1998b). Direct measurement of the force of
adhesion of a single biological cell using an atomic force microscopy. Colloid Surf. A:
Physicochem. Engng. Aspects 136, 231–234.
Bowen, W. R., Lovitt, R. W., and Wright, C. J. (2000a). Direct quantification of Aspergillus niger
spore adhesion in liquid using an atomic force microscope. J. Colloid Interf. Sci. 228, 428– 433.
Bowen, W. R., Lovitt, R. W., and Wright, C. J. (2000b). Direct quantification of Aspergillus niger
spore adhesion to mica in air using an atomic force microscope. Colloid Surf. A-Physicochem.
Engng. Aspects 173, 205–210.
Brantley, S. L., Liermann, L., and Bau, M. (2001). Uptake of trace metals and rare earth elements from
hornblende by a soil bacterium. Geomicrobiol. J. 18, 37–61.
Burnham, N. A., Dominguez, D. D., Mowery, R. L., and Colton, R. J. (1990). Probing the
surface forces of monolayer films with an atomic-force microscope. Phys. Rev. Lett. 64,
1931–1934.
Bustamante, C., Marko, J. F., Siggia, E. D., and Smith, S. (1994). Entropic elasticity of lambda-phage
DNA. Science 265, 1599–1600.
Butt, H. J., and Jaschke, M. (1995). Calculation of Thermal Noise in Atomic-Force Microscopy.
Nanotechnology 6, 1– 7.
Butt, H. J., Jaschke, M., and Ducker, W. (1995). Measuring surface forces in aqueous electrolyte
solution with the atomic force microscope. Bioelectrochem. Bioenerg. 38, 191 –201.
Caccavo, F. (1999). Protein-mediated adhesion of the dissimilatory Fe(III)-reducing bacterium
Shewanella alga BrY to hydrous ferric oxide. Appl. Environ. Microbiol. 65, 5017–5022.
Caccavo, F. Jr., Schamberger, P. C., Keiding, K., and Nielsen, P. H. (1997). Role of hydrophobicity in
adhesion of the dissimilatory Fe(III)-reducing bacterium Shewanella alga to amorphous Fe(III)
oxide. Appl. Environ. Microbiol. 62, 3837–3843.
Camesano, T. A., and Logan, B. E. (2000). Probing bacterial electrosteric interactions using atomic
force microscopy. Environ. Sci. Technol. 34, 3354–3362.
Cappella, B., Baschieri, P., Frediani, C., Miccoli, P., and Ascoli, C. (1997). Force–distance curves by
AFM: A powerful technique for studying surface interactions. IEEE Engng. Med Biol. Mag. 16,
58 –65.
Cappella, B., and Dietler, G. (1999). Force–distance curves by atomic force microscopy. Surf. Sci.
Rep. 34, 1–104.
Cherian, S., and Thundat, T. (2002). Determination of adsorption-induced variation in the spring
constant of a microcantilever. Appl. Phys. Lett. 80, 2219–2221.
Cheung, C. L., Hafner, J. H., Chen, L. W., and Lieber, C. M. (2000). Direct growth of single-walled
carbon nanotube scanning probe microscopy tips and application to ultrahigh-resolution chemical
force microscopy and structural imaging. Abstr. Pap. Am. Chem. Soc. 219, 182.
Chilkoti, A., Boland, T., Ratner, B. D., and Stayton, P. S. (1995). The relationship between ligand-
binding thermodynamics and protein–ligand interaction forces measured by atomic force
microscopy. Biophys. J. 69, 2125–2130.
Cleveland, J. P., Manne, S., Bocek, D., and Hansma, P. K. (1993). A non-destructive method for
determining the spring constant of cantilevers for scanning force microscopy. Rev. Sci. Instrum.
64, 403–405.
Cocozza, C., Tsao, C. C. G., Cheah, S. F., Kraemer, S. M., Raymond, K. N., Miano, T. M., and
Sposito, G. (2002). Temperature dependence of goethite dissolution promoted by trihydroxamate
siderophores. Geochimica et Cosmochimica Acta 66, 431– 438.
Cornell, R. M., and Schwertmann, U. (1996). “The Iron Oxides: Structure, Properties, Reactions,
Occurrence and Uses”, VCH.
Craig, V. S. J., and Neto, C. (2001). In Situ calibration of colloid probe cantilevers in force
microscopy: hydrodynamic drag on a sphere approaching a wall. Langmuir 17, 6018–6022.
Cygan, R. T. (2001). Molecular Modeling in Mineralogy and Geochemistry. In “Molecular Modeling
Theory: Applications in the Geosciences” (J. D. Kubicki, Ed.), Vol. 42, p. 531. Mineralogical
Society of America Geochemical Society, Washington DC.
Dammer, U., Hegner, M., Anselmetti, D., Wagner, P., Dreier, M., Huber, W., and Guentherodt, H. J.
(1996). Specific antigen/antibody interactions measured by force microscopy. Biophys. J. 70,
2437–2441.
Dammer, U., Popescu, O., Wagner, P., Anselmetti, D., Guntherodt, H. J., and Misevic, G. N. (1995).
Binding Strength Between Cell-Adhesion Proteoglycans Measured By Atomic-Force
Microscopy. Science 267, 1173–1175.
Das, A., and Caccavo, F. (2000). Dissimilatory Fe(III) oxide reduction by Shewnnella alga BrY
requires adhesion. Curr. Microbiol. 40, 344 –347.
Davies, D. G., Parsek, M. R., Pearson, J. P., Iglewski, B. H., Costerton, J. W., and Greenberg, E. P.
(1998). The involvement of cell-to-cell signals in the development of a bacterial biofilm. Science
280, 295 –298.
D’Costa, N. P., and Hoh, J. H. (1995). Calibration of optical lever sensitivity for atomic force
microscopy. Rev. Sci. Instrum. 66, 5096– 5097.
Degertekin, F. L., Hadimioglu, B., Sulchek, T., and Quate, C. F. (2001). Actuation and
characterization of atomic force microscope cantilevers in fluids by acoustic radiation pressure.
Appl. Phys. Lett. 78, 1628–1630.
Derjaguin, B. V., and Landau, L. (1941). Acta Physiochim. URSS 4, 633.
Derjaguin, B. V., Muller, V. M., and Toporov, Y. P. (1975). J. Colloid Interf. Sci. 53.
DiChristina, T. J., and Delong, E. F. (1994). Isolation of anaerobic respiratory mutants of Shewanella
putrefaciens and genetic analysis of mutants deficient in anaerobic growth on Fe3þ. J. Bacteriol.
176, 1468–1474.
Ducker, W. A., Senden, T. J., and Pashley, R. M. (1991). Direct measurement of colloidal forces using
an atomic force microscope. Nature 353, 239–241.
Ducker, W. A., Senden, T. J., and Pashley, R. M. (1992). Measurement of forces in liquids using a
force microscope. Langmuir 8, 1831–1836.
Dziurla, M.-A., Achouak, W., Lam, B.-T., Heulin, T., and Berthelin, J. (1998). Enzyme-linked
immunofiltration assay to estimate attachment of Thiobacilli to pyrite. Appl. Environ. Microbiol.
64, 2937–2942.
Edwards, K. J., Schrenk, M. O., Hamers, R., and Banfield, J. F. (1998). Microbial oxidation of pyrite:
Experiments using microorganisms from an extreme acidic environment. Am. Mineral. 83,
1444–1453.
Egger, M., Heyn, S.-P., and Gaub, H. E. (1992). Synthetic lipid-anchored receptors based on the
binding site of a monoclonal antibody. Biochim. Biophys. Acta 1104, 45–54.
Erlandsson, R., McClelland, G. M., Mate, C. M., and Chiang, S. (1988). Atomic force
microscopy using optical interferometry. J. Vacuum Sci. Technol., A: Vacuum, Surf. Films 6,
266–270.
Evans, E. (1999). Energy landscapes of biomolecular adhesion and receptor anchoring at interfaces
explored with dynamic force spectroscopy. Faraday Discussions 111, 1 –16.
Evans, E. (2001). Probing the relation between force-lifetime and chemistry in single molecular
bonds. Annu. Rev. Biophy. Biomol. Struct. 30, 105–128.
Evans, E., and Ipsen, J. (1991). Entropy-driven expansion of electric double layer repulsion between
highly flexible membranes. Electrochim. Acta 36, 1735– 1741.
Evans, E., and Ludwig, F. (2000). Dynamic strengths of molecular anchoring and material cohesion in
fluid biomembranes. J. Phys.: Condensed Matter 12, A315–A320.
Evans, E., and Ritchie, K. (1997). Dynamic strength of molecular adhesion bonds. Biophysical
Journal 72, 1541– 1555.
Fiorini, M., McKendry, R., Cooper, M. A., Rayment, T., and Abell, C. (2001). Chemical force
microscopy with active enzymes. Biophys. J. 80, 2471–2476.
Fleminger, G., and Shabtai, Y. (1995). Direct and rapid analysis of the adhesion of bacteria to solid
surfaces: Interaction of fluorescently labeled Rhodococcus strain GIN-1 (NCIMB 40340) cells
with titanium-rich particles. Appl. Environ. Microbiol. 61, 4357–4361.
Fletcher, M., (Ed.) (1996). “Bacterial Adhesion: Molecular and Ecological Diversity”. Wiley-Liss,
New York.
Florin, E. L., Moy, V. T., and Gaub, H. E. (1994). Adhesion Forces Between Individual Ligand–
Receptor Pairs. Science 264, 415 –417.
Flory, P. J. (1989). “Statistical Mechanics of Chain Molecules”. Hanser Publishers, Munich.
Fortin, D., Ferris, F. G., and Beveridge, T. J. (1997). Surface-mediated mineral development by
bacteria. In “Geomicrobiology: Interactions between Microbes and Minerals” (K. H. Nealson,
Ed.), Vol. 35, pp. 161 –180. Mineralogical Society of America, Washington DC.
Frisbie, C. D., Rozsnyai, L. F., Noy, A., Wrighton, M. S., and Lieber, C. M. (1994). Functional-group
imaging by chemical force microscopy. Science 265, 2071–2074.
Gay, C., and Leibler, L. (1999). On stickiness. Phys. Today 52, 48– 52.
Gergely, C., Hemmerle, J., Schaaf, P., Horber, J. K. H., Voegel, J.-C., and Senger, B. (2002). Multi-
bead-and-spring model to interpret protein detachment studied by AFM force spectroscopy.
Biophys. J. 83, 706–722.
Gergely, C., Senger, B., Voegel, J. C., Horber, J. K. H., Schaaf, P., and Hemmerle, J. (2001). Semi-
automatized processing of AFM force-spectroscopy data. Ultramicroscopy 87, 67–78.
Gibson, C. T., Weeks, B. L., Lee, J. R. I., Abell, C., and Rayment, T. (2001). A nondestructive
technique for determining the spring constant of atomic force microscope cantilevers. Rev. Sci.
Instrum. 72, 2340–2343.
Goddenhenrich, T., Lemke, H., Hartmann, U., and Heiden, C. (1990). Force microscope with
capacitive displacement detection. J. Vacuum Sci. Technol., A: Vacuum, Surf., Films 8, 383–387.
Grabarek, Z., and Gergely, J. (1990). Zero-length crosslinking procedure with the use of active esters.
Anal. Biochem. 185, 131– 135.
Grandbois, M., Beyer, M., Rief, M., Clausen-Schaumann, H., and Gaub, H. E. (1999). How strong is a
covalent bond? Science 283, 1727–1730.
Grubmuller, H., Heymann, B., and Tavan, P. (1996). Ligand binding: molecular mechanics
calculation of the streptavidin –biotin rupture force. Science 271, 997 –999.
Hafner, J. H., Cheung, C. L., and Lieber, C. M. (1999). Growth of nanotubes for probe microscopy
tips. Nature 398, 761– 762.
Hafner, J. H., Cheung, C. L., Woolley, A. T., and Lieber, C. M. (2001). Structural and functional
imaging with carbon nanotube AFM probes. Progress in Biophys. & Molec. Biol. 77, 73– 110.
Han, T., Williams, J. M., and Beebe, T. P. (1995). Chemical-bonds studied with functionalized
atomic-force microscopy tips. Analytical Chemica Acta 307, 365–376.
Hansen, D. C., McCafferty, E., Lins, C. W., and Fitzpatrick, J. J. (1995). An FT-IR investigation of
parabactin adsorbed onto aluminium. Appl. Surf. Sci. 84, 85–90.
Hartmann, U. (1991). van der Waals interactions between sharp probes and flat sample surfaces. Phys.
Rev. B 43, 2404– 2407.
Heinz, W. F., and Hoh, J. H. (1999). Spatially resolved force spectroscopy of biological surfaces using
the atomic force microscope. Trends Biotechnol. 17, 143–150.
Hersman, L. E. (2000). The role of siderophores in iron oxide dissolution. In “Environmental microbe-
metal interactions” (D. R. Lovley, Ed.), p. 395. ASM Press, Washington DC.
Hersman, L., Lloyd, T., and Sposito, G. (1995). Siderophore-promoted dissolution of hematite.
Geochim. Cosmochim. Acta 59, 3327– 3330.
Hiebert, F. K., and Bennett, P. C. (1992). Microbial control of silicate weathering in organic-rich
ground water. Science 258, 278–281.
Hinterdorfer, P., Baumgartner, W., Gruber, H. J., Schilcher, K., and Schindler, H. (1996). Detection
and localization of individual antibody-antigen recognition events by atomic force microscopy.
Proc. Natl. Acad. Sci. USA 93, 3477–3481.
Hinterdorfer, P., Gruber, H. J., Kienberger, F., Kada, G., Riener, C., Borken, C., and Schindler, H.
(2002). Surface attachment of ligands and receptors for molecular recognition force microscopy.
Colloid Surf., B: Biointerf. 23, 115–123.
Holmen, B. A., and Casey, W. H. (1996). Hydroxamate ligands, surface chemistry, and the mechanism
of ligand-promoted dissolution of goethite [a-FeOOH(s)]. Geochim. Cosmochim. Acta 22,
4403–4416.
Holmen, B. A., Tejedor-Tejedor, M. I., and Casey, W. H. (1997). Hydroxamate complexes in solution
and at the goethite-water interface: a cylindrical internal reflection fourier transform infrared
spectroscopy study. Langmuir 13, 2197– 2206.
Hutter, J. L., and Bechhoefer, J. (1993). Calibration of atomic-force microscope tips. Rev. Sci.
Instrum. 64, 1868–1873.
Israelachvili, J. N. (1992). “Intermolecular and Surface Forces”, 2nd edn. Academic Press, San
Diego.
Israelachvili, J. N., and McGuiggan, P. M. (1988). Forces between surfaces in liquids. Science 241,
795–800.
Ito, T., Citterio, D., Buehlmann, P., and Umezawa, Y. (1999). Observation of silver and hydrogen ion
binding to self-assembled monolayers using chemically modified AFM tips. Langmuir 15,
2788–2793.
Izrailev, S., Stepaniants, S., Balsera, M., Oono, Y., and Schulten, K. (1997). Molecular dynamics
study of unbinding of the avidin–biotin complex. Biophys. J. 72, 1568– 1581.
Jarvis, S. P., Duerig, U., Lantz, M. A., Yamada, H., and Tokumoto, H. (1998). Feedback-stabilized
force sensors. A gateway to the direct measurement of interaction potentials. Appl. Phys. A:
Mater. Sci. Process. A66, S211–S213.
Jarvis, S. P., Yamada, H., Yamamoto, S. I., Tokumoto, H., and Pethica, J. B. (1996). Direct
mechanical measurement of interatomic potentials. Nature (London) 384, 247–249.
Jeppesen, C., Wong, J. Y., Kuhl, T. L., Israelachvili, J., Mullah, J. N., Zalipsky, S., and Marques, C. M.
(2001). Impact of polymer tether length on multiple ligand–receptor bond formation. Science
293, 465 –468.
Jericho, S. K., and Jericho, M. H. (2002). Device for the determination of spring constants of atomic
force microscope cantilever and micromachined springs. Rev. Sci. Instrum. 73, 2483–2485.
Johnson, K. L., Kendall, K., and Roberts, A. D. (1971). Surface energy and the contact of elastic
solids. Proc. R. Soc. Lond. A 324, 301–313.
Joyce, S. A., and Houston, J. E. (1991). A new force sensor incorporating force-feedback control for
interfacial force microscopy. Rev. Sci. Instrum. 62, 710–715.
Kalinowski, B. E., Liermann, L. J., Brantley, S. L., Barnes, A., and Pantano, C. G. (2000). X-ray
photoelectron evidence for bacteria-enhanced dissolution of hornblende. Geochim. Cosmochim.
Acta 64, 1331–1343.
Karlsson, J.-J., Kadziola, A., Rasmussen, A., Rostrup, T. E., and Ulstrup, J. (1996). Electron transfer
of the di-heme protein: Pseudomonas stutzeri cytochrome c4. In “Protein Folds: A Distance-
Based Approach” (S. Brunak, Ed.), pp. 56–67. CRC Press, Boca Raton.
Kasas, S., Riederer, B. M., Catsicas, S., Cappella, B., and Dietler, G. (2000). Fuzzy logic algorithm to
extract specific interaction forces from atomic force microscopy data. Rev. Sci. Instrum. 71,
2082–2086.
Kendall, K. (1994). Adhesion: Molecules and mechanics. Science 263, 1720–1725.
Kendall, T. A., and Hochella, M. F. Jr. (2003). The measurement and interpretation of molecular level
forces of interaction between the siderophore azotobactin and mineral surfaces. Geochim.
Cosmochim. Acta.
Kienberger, F., Kada, G., Gruber, H. J., Pastushenko, V. P., Riener, C., Trieb, M., Knaus, H.-G.,
Schindler, H., and Hinterdorfer, P. (2000). Recognition force spectroscopy studies of the NTA-
His6 bond. Single Mol. 1, 59–65.
Kosmulski, M. (2001). “Chemical properties of material surfaces”. Marcel Dekker, New York.
Kraemer, S. M., Cheah, S. F., Zapf, R., Xu, J. D., Raymond, K. N., and Sposito, G. (1999). Effect of
hydroxamate siderophores on Fe release and Pb(II) adsorption by goethite. Geochim.
Cosmochim. Acta 63, 3003–3008.
Kraemer, S. M., Xu, J. D., Raymond, K. N., and Sposito, G. (2002). Adsorption of Pb(II) and Eu(III)
by oxide minerals in the presence of natural and synthetic hydroxamate siderophores. Environ.
Sci. Technol. 36, 1287–1291.
Kreller, D. I., Gibson, G., van Loon, G. W., and Horton, J. H. (2002). Chemical Force Microscopy
Investigation of Phosphate Adsorption on the Surfaces of Iron(III) Oxyhydroxide Particles.
J. Colloid Interf. Sci. 254, 205–213.
Kummert, R., and Stumm, W. (1980). The surface complexation of organic acids on hydrous.gamma.
alumina. J. Colloid Interf. Sci. 75, 373–385.
Lawrence, J. R., Korber, D. R., Hoyle, B. D., Costerton, J. W., and Caldwell, D. E. (1991). Optical
sectioning of microbial biofilms. J. Bacteriol. 173, 6558–6567.
Leckband, D., and Israelachvili, J. (2001). Intermolecular forces in biology. Q. Rev. Biophy. 34,
105 –267.
Lee, G. U., Chrisey, L. A., and Colton, R. J. (1994a). Direct Measurement of the Forces between
Complementary Strands of DNA. Science 266, 771–773.
Lee, G. U., Kidwell, D. A., and Colton, R. J. (1994b). Sensing discrete streptavidin–biotin interactions
with atomic force microscopy. Langmuir 10, 354– 357.
Liermann, L. J., Kalinowski, B. E., Brantley, S. L., and Ferry, J. G. (2000). Role of bacterial
siderophores in dissolution of hornblende. Geochim. Cosmochim. Acta 64, 587–602.
Lo, Y. S., Huefner, N. D., Chan, W. S., Stevens, F., Harris, J. M., and Beebe, T. P. (1999). Specific
interactions between biotin and avidin studied by atomic force microscopy using the Poisson
statistical analysis method. Langmuir 15, 1373–1382.
Lo, Y. S., Simons, J., and Beebe, T. P. (2002). Temperature dependence of the biotin– avidin bond-
rupture force studied by atomic force microscopy. Journal of Physical Chemistry B 106,
9847–9852.
Lower, S. K., Tadanier, C. J., and Hochella, M. F. (2000). Measuring interfacial and adhesion forces
between bacteria and mineral surfaces with biological force microscopy. Geochim. Cosmochim.
Acta 64, 3133–3139.
Lower, S. K., Hochella, M. F., and Beveridge, T. J. (2001a). Bacterial recognition of mineral surfaces:
Nanoscale interactions between Shewanella and a-FeOOH. Science 292, 1360–1363.
Lower, S. K., Tadanier, C. J., and Hochella, M. F. (2001b). Dynamics of the mineral-microbe
interface: Use of biological force microscopy in biogeochemistry and geomicrobiology.
Geomicrobiol. J. 18, 63–76.
Lower, S. K., Hochella, M. F. Jr., Banfield, J. F., and Rosso, K. (2002). Nanogeoscience:
From the movement of electrons to lithosphere plates. Eos: Trans. Am. Geophys. Union 83,
53 –56.
Ludwig, C., Casey, W. H., and Rock, P. A. (1995). Prediction of ligand-promoted dissolution rates
from the reactivities of aqueous complexes. Nature (London) 375, 44 –47.
Maeda, N., and Senden, T. J. (2000). A Method for the Calibration of Force Microscopy Cantilevers
via Hydrodynamic Drag. Langmuir 16, 9282–9286.
Marszalek, P. E., Oberhauser, A. F., Pang, Y. -P., and Fernandez, J. M. (1998). Polysaccharide
elasticity governed by chair-boat transitions of the glucopyranose ring. Nature (London) 396,
661 –664.
Maurice, P. A., Lee, Y. J., and Hersman, L. E. (2000). Dissolution of Al-substituted
goethites by an aerobic Pseudomonas mendocina var. bacteria. Geochim Cosmochim Acta 64,
1363–1374.
Mazzola, L. T., Frank, C. W., Fodor, S. P. A., Mosher, C., Lartius, R., and Henderson, E. (1999).
Discrimination of DNA hybridization using chemical force microscopy. Biophys. J. 76,
2922–2933.
Merkel, R., Nassoy, P., Leung, A., Ritchie, K., and Evans, E. (1999). Energy landscapes of receptor-
ligand bonds explored with dynamic force spectroscopy. Nature 397, 50– 53.
Meyer, G., and Amer, N. M. (1990). Simultaneous measurement of lateral and normal
forces with an optical-beam-deflection atomic force microscope. Appl. Phys. Lett. 57,
2089–2091.
Moy, V. T., Florin, E. L., and Gaub, H. E. (1994a). Intermolecular forces and energies between ligands
and receptors. Science 266, 257–259.
Moy, V. T., Florin, E.-L., and Gaub, H. E. (1994b). Adhesive forces between ligand and receptor
measured by AFM. Colloid Surf., A: Physicochem. Engng. Aspects 93, 343–348.
Muller, D. J., and Engel, A. (1997). The height of biomolecules measured with the atomic force
microscope depends on electrostatic interactions. Biophy. J. 73, 1633– 1644.
Muller, D. J., Baumeister, W., and Engel, A. (1999). Controlled unzipping of a bacterial surface layer
with atomic force microscopy. Proc. Nat. Acad. Sci. USA 96, 13170– 13174.
Myers, C. R., and Nealson, K. H. (1988). Bacterial manganese reduction and growth with manganese
oxide as the sole electron acceptor. Science 240, 1319–1321.
Myers, C. R., and Nealson, K. H. (1990). Respiration-linked proton translocation coupled to anaerobic
reduction of manganese(IV) and iron(III) in Shewanella putrefaciens MR-1. J. Bacteriol. 172,
6232–6238.
Myers, C. R., and Myers, J. M. (1992). Localization of cytochromes to the outer membrane of
anaerobically grown. Shewanella putrefaciens MR-1. J. Bacteriol. 174, 3429–3438.
Myers, C. R., and Myers, J. M. (1993). Ferric reductase is associated with the membranes of
anaerobically grown Shewanella putrefaciens MR-1. Fems Microbiol. Lett. 108, 15–22.
Myers, C. R., and Myers, J. M. (1997). Outer membrane cytochromes of Shewanella putrefaciens
MR-1: Spectral analysis, and purification of the 83-kDa c-type cytochrome. Biochim. Biophys.
Acta 1326, 307– 318.
Myers, J. M., and Myers, C. R. (1998). Isolation and sequence of omcA, a gene encoding a decaheme
outer membrane cytochrome c of Shewanella putrefaciens MR-1, and detection of omcA
homologs in other strains of S. putrefaciens. Biochim. Biophys. Acta 1373, 237–251.
Myers, J. M., and Myers, C. R. (2000). Role of the tetraheme cytochrome CymA in anaerobic electron
transport in cells of Shewanella putrefaciens MR-1 with normal levels of menaquinone.
J. Bacteriol. 182, 67–75.
Myers, J. M., and Myers, C. R. (2001). Role for outer membrane cytochromes OmcA and OmcB of
Shewanella putrefaciens MR-1 in reduction of manganese dioxide. Appl. Environ. Microbiol. 67,
260–269.
Myers, J. M., and Myers, C. R. (2002). Genetic complementation of an outer membrane cytochrome
omcB mutant of Shewanella putrefaciens MR-1 requires omcB plus downstream DNA. Appl.
Environ. Microbiol. 68, 2781–2793.
Nealson, K. H., and Saffarini, D. (1994). Iron and manganese in anaerobic respiration: Environmental
significance, physiology, and regulation. Ann. Rev. Microbiol. 48, 311– 343.
Neubauer, U., Nowack, B., Furrer, G., and Schulin, R. (2000). Heavy metal sorption on clay minerals
affected by the siderophore desferrioxamine B. Environ. Sci. Technol. 34, 2749–2755.
Noy, A., Vezenov, D. V., and Lieber, C. M. (1997). Chemical force microscopy. Ann. Rev. Mater. Sci.
27, 381 –421.
Nubel, U., Engelen, B., Felske, A., Snaidr, J., Wieshuber, A., Amann, R. I., Ludwig, W., and
Backhaus, H. (1996). Sequence heterogeneities of genes encoding 16S rRNAs in Paenibacillus
polymyxa detected by temperature gradient gel electrophoresis. J. Bacteriol. 178, 5636–5643.
Ohmura, N., Kitamura, K., and Saiki, H. (1993). Selective adhesion of Thiobacillus ferrooxidans to
pyrite. Appl. Environ. Microbiol. 59, 4044–4050.
Ong, Y.-L., Razatos, A., Georgiou, G., and Sharma, M. M. (1999). Adhesion forces between E. coli
bacteria and biomaterial surfaces. Langmuir 15, 2719–2725.
Oreskes, N., Shrader-Frechette, K., and Belitz, K. (1994). Verification, validation, and confirmation of
numerical models in the Earth Sciences. Science 263, 641– 646.
Pleuddemann, E. P. (1991). “Silane Coupling Agents”. Plenum Press, New York.
Rabinovich, Y. I., and Yoon, R. H. (1994). Use of atomic force microscope for the measurements
of hydrophobic forces between silanated silica plate and glass sphere. Langmuir 10,
1903–1909.
Razatos, A., Ong, Y. L., Sharma, M. M., and Georgiou, G. (1998a). Evaluating the interaction of
bacteria with biomaterials using atomic force microscopy. J. Biomater. Sci. Polym. Edu. 9,
1361–1373.
Razatos, A., Ong, Y.-L., Sharma, M. M., and Georgiou, G. (1998b). Molecular determinants of
bacterial adhesion monitored by atomic force microscopy. Proc. Nat. Acad. Sci. USA 95,
11059– 11064.
Rezania, A., Johnson, R., Lefkow, A., and Healy, K. E. (1999). Bioactivation of metal oxide surfaces.
1. Surface characterization and cell response. Langmuir 15, 6931–6939.
Rief, M., Gautel, M., Oesterhelt, F., Fernandez, J. M., and Gaub, H. E. (1997a). Reversible unfolding
of individual titin immunoglobulin domains by AFM. Science (Washington, D.C.) 276,
1109–1112.
Rief, M., Oesterhelt, F., Heymann, B., and Gaub, H. E. (1997b). Single molecule force
spectroscopy on polysaccharides by atomic force microscopy. Science (Washington, D.C.)
275, 1295–1297.
Roden, E. E., and Zachara, J. M. (1996). Microbial reduction of crystalline iron(III) oxides: Influence
of oxide surface area and potential for cell growth. Environ. Sci. Technol. 30, 1618–1628.
Rugar, D., Mamin, H. J., and Guethner, P. (1989). Improved fiber-optic interferometer for atomic
force microscopy. Appl. Phys. Lett. 55, 2588–2590.
Ryan, J. N., and Gschwend, P. M. (1994). Effects of ionic-strength and flow-rate on colloid
release — relating kinetics to intersurface potential-energy. J. of Colloid and Interface Science
164, 21–34.
Ryan, J. N., and Elimelech, M. (1996). Colloid mobilization and transport in groundwater. Colloid
Surf. a-Physicochem. Eng. Aspects 107, 1–56.
Sader, J. E., and White, L. (1993). Theoretical analysis of the static deflection of plates for atomic
force microscope applications. J. Appl. Phys. 74, 1–9.
Sader, J. E., Larson, I., Mulvaney, P., and White, L. R. (1995). Method for the calibration of atomic
force microscope cantilevers. Rev. Sci. Instrum. 66, 3789– 3798.
Sader, J. E. (1998). Frequency response of cantilever beams immersed in viscous fluids with
applications to the atomic force microscope. J. Appl. Phys. 84, 64–76.
Sader, J. E., Chon, J. W. M., and Mulvaney, P. (1999). Calibration of rectangular atomic force
microscope cantilevers. Rev. Sci. Instrum. 70, 3967–3969.
Schmitt, L., Ludwig, M., Gaub, H. E., and Tampe, R. (2000). A metal-chelating microscopy tip as a
new toolbox for single-molecule experiments by atomic force microscopy. Biophy. J. 78,
3275–3285.
Schultze-Lam, S., Harauz, G., and Beveridge, T. J. (1992). Participation of a cyanobacterial S-layer in
fine-grain mineral formation. J. Bacteriol. 174, 7971–7981.
Schwesinger, F., Ros, R., Strunz, T., Anselmetti, D., Guntherodt, H. J., Honegger, A., Jermutus, L.,
Tiefenauer, L., and Pluckthun, A. (2000). Unbinding forces of single antibody-antigen complexes
correlate with their thermal dissociation rates. Proc. Nat. Acad. Sci. USA 97, 9972–9977.
Seaman, J. C., Alexander, D. B., Loeppert, R. H., and Zuberer, D. A. (1992). The availability of iron
from various solid-phase iron sources to a siderophore producing pseudomonas strain. J. Plant
Nutr. 15, 2221– 2233.
Sposito, G. (1989). “The Chemistry of Soils”. Oxford University Press, New York.
Stevens, F., Lo, Y. S., Harris, J. M., and Beebe, T. P. (1999). Computer modeling of atomic force
microscopy force measurements: Comparisons of Poisson, histogram, and continuum methods.
Langmuir 15, 207–213.
Stone, A. T. (1997). Reactions of extracellular organic ligands with dissolved metal ions and mineral
surface. “Geomicrobiology: Interactions Between Microbes and Minerals” (K. H. Nealson, Ed.),
Vol. 35, pp. 309–341. Mineralogical Society of America, Washington D.C.
Strunz, T., Oroszlan, K., Schafer, R., and Guntherodt, H.-J. (1999). Dynamic force spectroscopy of
single DNA molecules. Proc. Nat. Acad. Sci. USA 96, 11277– 11282.
Stumm, W. (1992). “Chemistry of the Solid-Water Interface”. Wiley, Inc., New York.
Swamy, M. J. (1995). Thermodynamic analysis of biotin binding to avidin. A high sensitivity titration
calorimetric study. Biochem. Mol. Biol. Int. 36, 219 –225.
Telford, J. R., and Raymond, K. N. (1996). Siderophores. Compr. Supramol. Chem. 1, 245– 266.
Toikka, G., Hayes, R. A., and Ralston, J. (1996). Surface forces between spherical ZnS particles in
aqueous electrolyte. Langmuir 12, 3783– 3788.
Tokunaga, M., Aoki, T., Hiroshima, M., Kitamura, K., and Yanagida, T. (1997). Subpiconewton
intermolecular force microscopy. Biochem. Biophys. Res. Commun. 231, 566–569.
Torii, A., Sasaki, M., Hane, K., and Okuma, S. (1996). A method for determining the spring constant
of cantilevers for atomic force microscopy. Meas. Sci. Technol. 7, 179– 184.
van Loosdrecht, M. C. M., Lyklema, J., Norde, W., and Zehnder, A. J. B. (1989). Bacterial adhesion:
A physicochemical approach. Microbial. Ecol. 17, 1– 15.
van Oss, C. J. (1993). Acid-base interfacial interactions in aqueous media. Colloid Surf. A:
Physicochem. Engg. Aspects 78, 1 –49.
Verwey, E. J., and Overbeek, J. T. G. (1948). “Theory of the Stability of Lyophobic Colloids”.
Elsevier Publishing, Amsterdam.
Vigeant, M. A.-S., Ford, R. M., Wagner, M., and Tamm, L. K. (2002). Reversible and irreversible
adhesion of motile Escherichia coli cells analyzed by total internal reflection aqueous
fluorescence microscopy. Appl. Environ. Microbiol. 68, 2794–2801.
Voet, D., and Voet, J. (1995). “Biochemistry”, 2nd edn. John Wiley & Sons, New York.
Wagner, P. (1998). Immobilization strategies for biological scanning probe microscopy. FEBS Lett.
430, 112 –115.
Watteau, F., and Berthelin, J. (1994). Microbial dissolution of iron and aluminum from soil
minerals — efficiency and specificity of hydroxamate siderophores compared to aliphatic-acids.
Eur. J. Soil Biol. 30, 1– 9.
Weisenhorn, A. L., Maivald, P., Butt, H. J., and Hansma, P. K. (1992). Measuring Adhesion,
Attraction, and Repulsion Between Surfaces in Liquids With an Atomic-Force Microscope. Phys.
Rev. B 45, 11226–11232.
Wenzler, L. A., Moyes, G. L., Olson, L. G., Harris, J. M., and Beebe, T. P. (1997). Single molecule
bond-rupture force analysis of interactions between AFM tips and substrates modified with
organosilanes. Analytical Chemistry 69, 2855–2861.
Whitman, W. B., Coleman, D. C., and Wiebe, W. J. (1998). Prokaryotes: The unseen majority. Proc.
Nat. Acad. Sci. USA. 95, 6578–6583.
Williams, J. M., Han, T. J., and Beebe, T. P. (1996). Determination of single-bond forces from contact
force variances in atomic force microscopy. Langmuir 12, 1291–1295.
G. Winkelmann, (Ed.), (1991). In “CRC Handbook of Microbial Iron Chelates” CRC Press, Boca
Raton, Florida.
Wong, S. S., Joselevich, E., Wooley, A. T., Cheung, C. L., and Lieber, C. M. (1998a). Covalently
functionalized nanotubes as nanometre-sized probes in chemistry and biology. Nature 394,
52–55.
Wong, S. S., Woolley, A. T., Joselevich, E., Cheung, C. L., and Lieber, C. M. (1998b). Covalently-
functionalized single-walled carbon nanotube probe tips for chemical force microscopy. J. of
American Chemical Society 120, 8557–8558.
Wong, J., Chilkoti, A., and Moy, V. T. (1999). Direct force measurements of the streptavidin–biotin
interaction. Biomol. Engg. 16, 45 –55.
Yamamoto, S.-i., Yamada, H., and Tokumoto, H. (1997). Precise force curve detection system with a
cantilever controlled by magnetic force feedback. Rev. Sci. Instrum. 68, 4132–4136.
Yao, H.-L., and Yeh, H.-H. (1996). Fumarate, Maleate, and Succinate Adsorption on Hydrous.delta.-
A12O3. 1. Comparison of the Adsorption Maxima and Their Significance. Langmuir 12,
2981–2988.
RICE FUNCTIONAL GENOMICS:
LARGE-SCALE GENE DISCOVERY
AND APPLICATIONS TO CROP
IMPROVEMENT
Hei Leung1 and Gynleung An2
1
International Rice Research Institute, DAPO Box 7777, Metro Manila, Philippines
2
Pohang University of Science and Technology, San 31 Hyoja-dong, Nam-gu,
Pohang, Korea 790-784
I. Introduction
II. Rice as a Model Genetic System
A. Diversity of Oryza
B. Grass as a Single Genetic System
III. Rice Genome
A. Genome Sequencing
B. Salient Features of the Rice Genome
C. Comparison Between Indica and Japonica Rice
IV. Key Ingredients for Gene Discovery in Rice
A. Genetic Resources
B. High-throughput Technologies
C. Biological Evaluation
D. Bioinformatics
V. Forward and Reverse Genetics
A. Insertional Mutants
B. Chemical- and Irradiation-induced Mutants
C. Natural Genetic Variation
VI. Functional Validation of Rice Genes
A. Gene Expression Analysis
B. Gene Silencing
C. Heterologous Bioassays
D. Gene Replacement
E. Allelic Series
VII. Applications to Crop Improvement
A. Candidate Gene Approach and Allele Mining
B. Pathways and Genetic Regulation
C. Cross-species Inference of Gene Function
VIII. International Collaboration and the Role
of Developing Countries
IX. Concluding Remarks
Acknowledgments
References
55
0065-2113/03 $35.00
56 H. LEUNG AND G. AN
I. INTRODUCTION
Rice has an old and rich cultivation history. Archeological ruins found in
China and India suggest that domestication of rice might date back 7000
years (Watanabe, 1997). Several regions in Asia, including Assam in India,
Myanmar, northern Thailand, and southwest and southern China, have been
hypothesized as the origins of cultivated rice. Also, rice cultivation probably
originated independently in several locations (Matsuo, 1997).
Rice is more than just a food crop. As a key sustenance for people since
ancient times, rice has become part of the social and religious fabric of many
cultures (see White 1994 for an interesting popular article about rice culture in
National Geographic). Today, rice remains the staple food of more than half
the world population and supplies up to 60% of the calories in the diet of people
living in Asia. In many developing countries, consumers spend more than
20% of disposable income just to buy sufficient rice to sustain daily food needs.
Rice cultivation is also a major source of employment in rural areas. The rice
economy — production, distribution, and consumption — therefore plays an
important role in the livelihood of many people.
Rice is cultivated in diverse ways. Rice production is often classified
according to the ecosystem in which it is grown. In terms of area, a majority
(58%) of rice is grown in the irrigated environment, commonly called paddy rice.
This is a highly intensive rice production environment, which accounts for about
75% of the rice produced worldwide. Most investment in production technologies
has been applied to this environment. The remaining portion of rice is grown in
the rainfed environment, that is, on land that relies on natural rainfall. About half
of the rainfed rice is grown in areas with a reasonably reliable water supply (the
favorable rainfed area), but the other half is grown in very difficult conditions
with too little or too much water and poor soil conditions.
Figure 1 summarizes yield trends of rice over the course of the Green
Revolution that started more than four decades ago in response to the food crisis
in many parts of the developing world (Hossain et al., 2000). As a result of
concentrated efforts devoted to improving rice germplasm and agronomy, a
dramatic and steady increase in rice productivity has occurred across Asia
(Hossain, 1998). Rice yield in irrigated areas nearly doubled from 1960 to 1980.
This has resulted in lower rice prices that benefit consumers. The yield increase,
however, is low in unfavorable environments where the water supply is
unreliable and the soil is often infertile or problematic for rice production.
Recognizing this historical trend is relevant in the context of applying genomics
to crop improvement. The trend enables us to identify the key areas where
advances in science can help solve intractable problems. In the highly intensive
system, technologies are needed to increase production efficiency with less cost.
The system also needs improved resilience against pests and diseases to reduce
RICE FUNCTIONAL GENOMICS 57
Figure 1 Yield trend of rice from 1967 to 1997 in three rice production environments (modified
from Hossain et al., 2000).
reliance on chemical inputs. In the unfavorable rainfed environment, there is

plenty of room to improve productivity and the nutritional value of germplasm to
meet the need of the rural poor.
Although rice is such an important food crop, it has only a relatively short
history as a model organism for biological investigation. Before the 1980s, rice
research — from basic to applied — was mostly done by countries where rice is
the staple crop and where sufficient infrastructure existed to conduct research.
This limited the majority of rice research to a few countries or international
agricultural research centers where varietal improvement is a mandate or a
national necessity. There was hardly any private interest in rice research. In many
respects, rice was an “orphan” crop in terms of research investment. This picture
changed in the mid-80s when advances made in plant molecular biology and gene
transfer techniques were seen as promising avenues to convert biological
discoveries to applications. The Rockefeller Foundation, a long-time supporter
of international agricultural development, made a strategic decision to support
the application of molecular biology to solve rice production problems. The
beginning of the International Rice Biotechnology Program in 1985 launched
the “molecular revolution” in rice science. For the next 16 years (1985 – 2001),
we witnessed the blossoming of basic research in rice. The sustained support of
rice biotechnology produced a large cadre of scientists in the developing world,
well trained and ready to apply useful genes in production (O’Toole et al., 2001).
While gene identification and cloning were advanced toward the end of the
Rockefeller Biotechnology Program, there was a growing sense that many
biological problems limiting rice production required a comprehensive under-
standing of multiple genes and their interactions with the environment. In the last
Rice Biotechnology meeting held in Phuket, Thailand, in 1998, questions were
raised about the next quantum leap in rice productivity. What would the
investment in research capacity and infrastructure in the rice-growing world do to
future rice production? Around this time, the genome of the first eukaryote —
Saccharomyces cerevisiae — was completed (Williams, 1996), followed by
several initiatives to sequence other model organisms — Drosophila, nematode,
Arabidopsis, and human. These events marked a genomics decade that has put
rice at the center of plant science research and provided a logical extension of the
Rockefeller Foundation Rice Biotechnology Program.
Genomics can be broadly defined as the understanding of all genes and their
encoded biological function in an organism in totality. It has emerged as an
integrative discipline that provides the means to understand biological complex-
ity and to solve difficult practical problems. Rice, with the smallest genome size
among major cereals and crop plants, is seen as the obvious target for genomics
research. The Rockefeller Foundation sponsored a meeting in 1997 at the
International Congress of Plant Molecular Biology in Singapore to discuss the
merits and feasibility of sequencing the entire rice genome. At that time, physical
mapping using YAC and BAC libraries was progressing rapidly, producing the
critical resources for sequencing. However, with the available sequencing
technologies at that time, completing the rice genome would cost about $200
million over 10 years, a formidable undertaking for any individual institute or
program. Thus, a sensible way to achieve the task was to launch an international
collaboration. This led to the formation of an International Rice Genome
Sequencing Project (IRGSP) in 1998 under the coordination of the Rice Genome
Program of Japan. The IRGSP represents a consortium of multiple countries
collaborating to sequence the rice genome. In the initial stage, over 10 countries
pledged to take part to sequence individual chromosomes. The IRGSP was a
unique undertaking that has galvanized government commitment and it
subscribed to the principle of releasing all sequence information to the public.
Since then, rice genome sequencing has been the topic of scientific drama and
excitement involving the private and public sector (Normile and Pennisi, 2002).
To a large extent, the public initiative has been a key factor to leverage the
contribution of rice genome data from the private sector. Together, this has led to
the availability of genome sequence of two rice subspecies ( japonica and indica),
a favorable situation for public researchers working not only on rice but on other
plant species.
With a genetic blueprint in place, we ask, what are the new opportunities to
understand gene functions and their interactions? What are the most efficient
approaches to find genes of agronomic importance? Why is international
collaboration necessary for gene discovery in rice? How can we turn gene
discovery into practical products? This review is organized with these questions
in mind. In particular, we hope to lay out the framework for large-scale gene
discovery that will benefit the developing world, where access to knowledge and
tools is often limited. For more discussion of rice gene function analyses
and comparative genomics, readers are referred to a recent review by Shimamoto
and Kyozuka (2002).
In this chapter, we will begin with an introduction of the rice plant as a model
species. We will discuss the ingredients needed for large-scale gene discovery
and the approaches that will benefit the improvement of traits needed in rice
production. We will discuss the importance of producing public resources and the
need for international collaboration, and how gene discovery can be integrated
into plant improvement programs.
II. RICE AS A MODEL GENETIC SYSTEM
Rice has emerged as a model in plant science research because of its many
positive attributes for genetic studies. Rice has a simple genetic system (diploid
and disomic inheritance) and one of the most amenable transformation systems
among grasses (Hiei et al., 1994; Christou, 1997; Cheng et al., 1998). The small
genome size and relatively low amount of repetitive sequences make rice a
tractable system for whole-genome sequencing. Other than these technical
advantages, an important, but perhaps less appreciated attribute is the great store
of genetic diversity in the genus Oryza. With a long history of cultivation and
selection, and a large and heterogeneous geographical niche, the wealth of
phenotypic and genetic variation in all aspects of plant life, particularly genes
controlling agronomically important traits, is enormous in rice. Below, we
highlight the genetic diversity available in rice that is relevant for functional
genomics research.
A. DIVERSITY OF ORYZA
Taxonomically, rice is in the subfamily Ehrhartoideae (former Oryzoideae) of

the grass family Poaceae (Gramineae). Figure 2 shows the overall relationship of
rice to other cereals and the recognized genomes within the genus Oryza
(Katayama, 1997; Ge et al., 1999; Lu, 1999; Ge et al., 2001). The most widely
cultivated rice, O. sativa, has the AA genome. Another cultivated rice species,
O. glaberrima Steud., with the AA genome is grown only in West Africa (for an
interesting anthropological perspective of African rice, see Linares, 2002). Both
O. sativa and O. glaberrima have 12 chromosomes as their basic complement
Figure 2 Diversity of rice as part of the broad gene pool of the grass families. Dendrograms
depict the relationships among cereal crops in the grass family (Kellogg, 1998), domesticated and wild
rice with different genomes (from Ge et al. 1999), and cultivated rice species. Dendrograms are not
drawn to scale.
(2n ¼ 24). F1 hybrids between the two species can be produced, but advancing
beyond F2 generation is difficult because of sterility. Nonetheless, the small
amount of fertility has allowed the successful transfer of useful genes from O.
glaberrima into O. sativa (Jones et al., 1997).
The cultivated O. sativa has two subspecies — indica and japonica — that
provide the main gene pool for rice breeding. There is a third minor group called
“tropical japonica” (also called javanica) that is common in Southeast Asia.
Japonica rice is generally adapted to the temperate area, whereas indica rice is
planted in the tropics. Indica rice is the most widely grown type, constituting
about 80% of world rice production. Chang et al. (1991) summarized the main
agronomic characteristics of the main rice types (Table I). There is much overlap
in agronomic characteristics between the two subspecies. Whether indica and
japonica are monophyletic or diphyletic remains a research question (Morishima,
2001). Although divergence of these subspecies has occurred over time, sufficient
fertility exists to allow routine cross hybridization and gene flow between the two
subspecies. Also, because of their divergence, a high level of hybrid heterosis can
be harnessed from indica £ japonica crosses. Such extra hybrid vigor is being
exploited in hybrid rice production (Virmani and Ilyas-Ahmed, 2001). Now that
the genome sequence of these two subspecies is available, this will open up
opportunities to understand the molecular basis of japonica £ indica sterility as
well as hybrid heterosis.
There are more than 30 named Oryza species (Watanabe, 1997) with a broad
geographic distribution. Table II lists some of the species that have been used in
rice improvement programs or genetic studies (Brar and Khush, 2002). This is by
no means a comprehensive list of all known wild rice species but it illustrates the
diversity of the domesticated and wild germplasm that forms the basis for gene
Table I
Diverse Morphological and Physiological Features of indica, japonica, and Tropical japonica
( javanica) Rice Reflect a Broad Genetic Base for Gene Discovery
Rice type
Tropical japonica
Characteristics Indica Japonica (javanica)
Leaf morphology Broad to narrow, light green Narrow, dark green Broad, stiff, light green
Grain Long to short, slender, flat Short, roundish Long, broad, thick
Tillering High Medium Low
Plant stature Tall to intermediate Short to intermediate Tall
Awn Mostly awnless Awnless to long-awned Long-awned or awnless
Lemma and palea Thin, short hair Dense, long hair Long hair
Shattering Easy Low Low
Plant tissue Soft Hard Hard
Photosensitivity Varying Zero to low sensitivity Low sensitivity
Amylose 23–31% 10 –24% 20–25%
Gelatinization Varying Low Low
temperature
Adapted from Chang et al. (1991).
discovery. The wild species that have evolved in diverse, heterogeneous, and
often harsh environments are potentially a rich pool of useful genes for breeding.
Wild rice with a non-AA genome is not interferile with cultivated rice, thus
making gene transfer between these wild species difficult, though not impossible,
through intervention with embryo rescue and tissue culture techniques (Brar and
Khush, 1997).
There are examples of useful genes that have been extracted from wild
germplasm for breeding (Brar and Khush, 2002). One well-known example is the
grassy stunt resistance gene obtained from O. nivara. This resistance gene has
been used in nearly all rice cultivars developed for the tropical lowland and it
remains effective in conditioning resistance to the dominant strain of the grassy
stunt virus until now (Khush, 1977). The first cloned disease resistance gene,
Xa21, was also originally introgressed from the wild rice O. longistaminata from
an accession in Mali, Africa. After many years of genetic and breeding work,
Xa21 was put into an agronomically acceptable background that has facilitated its
molecular cloning and use in rice production (Song et al., 1995). Yet, one may
also argue that the number of genes extracted for use represents only a handful of
the potential useful genes in the wild species. Tanksley and McCouch (1999)
popularized the use of unadaptive wild germplasm through systematic
introgression into elite backgrounds. However, the large-scale introgression of
wild species alleles into domesticated rice is limited by the laborious process of
embryo rescue. This picture may change if we have a way to predetermine
Table II
Reservoir of Useful Genes and Alleles in Domesticated and Wild Rice Germplasm
Number of accessions at
Genome Speciesa Distributionb IRRI Gene Bankc
AA O. sativa Worldwide 91,963

AA O. glaberrima West Africa 1543
AA O. barthii Africa 218
AA O. nivara Tropical, Subtropical Asia 1258
AA O. rufipogon Tropical, Subtropical Asia, Tropical 1048
Australia
AA O. longistaminata Africa 206
AA O. glumaepatula South, Central America 54
AA O. meridionalis Tropical Australia 56
BB, BBCC O. punctata Africa 71
BBCC O. minuta Philippines, Papua New Guinea 64
CC O. officinalis Tropical, Subtropical Asia, Tropical 279
Australia
CC O. rhizomatis Sri Lanka 19
CC O. eichingeri South Asia, East Africa 30
CCDD O. alta South, Central America 6
CCDD O. latifolia South, Central America 58
CCDD O. grandiglumis South, Central America 10
EE O. australiensis Tropical Australia 36
FF O. brachyantha Africa 19
HHKK O. schlechteri Papua New Guinea 1
HHJJ O. longiglumis Indonesia (Irian Jaya), Papua 6
New Guinea
HHJJ O. ridleyi South Asia 15
GG O. granulata South, Southeast Asia 24
GG O. meyeriana Southeast Asia 11
GG O. neocaledonica New Caledonia 1
a
Lu (1999).
b
Brar and Khush (2002).
c
Information from IRGCIS database as of February 7, 2003, http://grcsvr4/irgc/main.htm.
the potential value of specific alleles in useful donors. The application of reverse
genetics to determine the allelic state of candidate genes can be an attractive
strategy to select wild germplasm for introgression. How to tap this genetic
potential is the core agenda of functional genomics.
B. GRASS AS A SINGLE GENETIC SYSTEM
The grass family diverged from the dicotyledon (dicot) lineage about 200
million years ago (Wolfe et al., 1989; Crane et al., 1995). Within the grass family,
an estimated . 10,000 species diverged from each other about 70 million years
ago (Kellogg, 2001). During this period of evolution, there has been an expansion
and contraction of the ancestral genome estimated to be about 3.5 pg/2C nucleus.
Kellogg (1998) examined the DNA contents of representative species in relation
to the phylogenetic tree of four subfamilies of grasses that contain the cereals and
showed that the genome size had expanded in some but contracted in others (see
Fig. 4 in Kellogg 1998). The Pooideae, which contains wheat, barley, rye, and
oats, has a large genome (10 pg DNA per 2C nucleus), whereas members of the
Ehrhartoideae have a generally smaller genome (3.3 pg). The large difference in
genome size is attributed, to a large extent, to the expansion of repetitive
sequences. The large genome size has been an impediment to the systematic
sequencing of certain cereals, although progress has been made in identifying
gene-rich regions in large genomes (Gill et al., 1996; Farris et al., 2000). That the
grasses belong to a monophyletic group and all share a set of common
“grass alleles” argues favorably for using a few reference species to avoid the
difficulties (and expenditure involved) in sequencing individual species
(Freeling, 2001).
The concept of grass as a single genetic system rests on two main premises.
First, the gene order on a chromosome is conserved in different related genomes.
Second, there is sufficient conservation of gene content across species, and that
function can be inferred by sequence structure. In dicots, Bonierbale et al. (1988)
first demonstrated that tomato cDNA probes could be used to establish an RFLP
linkage map in potato that is colinear with that in tomato. In monocots, Ahn and
Tanksley (1992) used a common set of probes to show conserved linkage in
maize and rice. Since then, comparative mapping in both dicots and monocots has
led to the general conclusion that the gene order of related plant species is
sufficiently conserved even after millions of years of evolution, making it
possible to infer the location of a particular gene based on linkage information
from another species (Devos and Gale, 2000). In most cases, colinearity between
plant species can be demonstrated at the macro chromosomal level. There are,
however, exceptions to colinearity due to the accumulation of chromosomal
changes over millions of years of evolution. Such exceptions are particularly
common at the micro level because of inversion, translocation, insertion, gene
duplications, and deletions.
A distinction between macro- and micro-colinearity is important because
success in finding a gene through a syntenic relationship depends on the
maintenance of micro-colinearity between related species. Freeling (2001) offers
a useful clarification of the terms “colinearity” and “synteny”. Macro-colinearity
is inferred when the same marker orders are observed in a typical mapping
experiment with 10 — 20 cM resolution. Occasional chromosomal rearrange-
ments may break up this gene order, but the pair of chromosomes in question
is considered to be syntenic to each other. Thus, even for related species
that do not have perfect macro-colinearity, they can still be syntenous.
However, macro-colinearity may or may not enable one to predict gene position
at a fine scale. For a syntenic relationship to be useful for finding a gene, we need
markers with a high probability of finding the same adjacent markers in another
species. Currently, the number of successful cases of gene isolation strictly based
on positional relationships is still too few. The predictive power of the syntenic
relationship has not yet been vigorously tested (Gaut, 2002).
The exceptions in micro-colinearity, however, by no means argue against
the use of grass as a single genetic system (Bennetzen and Freeling, 1997;
Freeling, 2001). With the whole-genome sequence available, the benefit of
using a model species to find a gene will be based more on gene sequence
conservation rather than on map locations. For example, a candidate gene for
disease resistance in barley can be used to find its orthologs in rice or vice
versa without knowledge of map location. Thus, map locations are useful data,
but they are not a prerequisite for finding orthologous genes to evaluate
function. In the future, the syntenic relationship will be more relevant for
mapping unique traits in a species with little sequence information. For
example, drought tolerant traits are common in many millets which do not have
detailed sequence information. In such a case, moderate investment in mapping
in the “trait-bearing” crops can lead to identification of candidate genes in a
species with a detailed sequence map.
In summary, the original discovery of a colinear relationship between markers
in different species triggered a large number of studies on the extent of gene order
conservation. While macrosynteny (10 – 20 cM range) has usually been
supported, evidence of microsynteny (within 1 cM, or in the 100-kb range) is
more difficult to obtain. Rearrangements, insertions, and deletions appear to have
disrupted the syntenic relationship, which potentially leads to futile pursuit
(Feuillet and Keller, 2002; Li and Gill, 2002). Notwithstanding these exceptions,
comparative analyses of genome organization and gene function within the
grasses will elevate the studies of genome evolution and practical applications to
a new level. With the completion of the rice genome sequence, a comparative
genomic approach can now be pursued with even greater intensity.
III. RICE GENOME
Not long ago, sequencing the genome of an organism was a daunting

proposition. However, with advances in sequencing technologies, the cost of
sequencing a genome has dropped dramatically. For the past decade, genome
sequencing projects, either completed or begun, have produced such excitement
in biological sciences that many young students have been attracted to the
discipline of genomics. The complete genome sequence represents a public
resource that opens up opportunities for new understanding of biological systems.
With sequence available, identification and isolation of a gene can be largely

done electronically or by simple PCR amplification, making gene cloning in the
conventional sense something of the past. This may prove to be one of the biggest
returns from the investment in sequencing projects.
A. GENOME SEQUENCING
As of January 2003, four versions of rice genome sequences exist: three draft
versions produced by Monsanto, Syngenta, and the Beijing Genomics Institute
(BGI) and a highly accurate version produced by the IRGSP (for a summary, see
Buell, 2002). The Monsanto and Syngenta sequences, originally produced by
independent efforts, have now been integrated with the IRGSP sequence. The
BGI sequence is from indica rice and has now been improved from the initial 4X
draft to a higher-quality 6X draft (Yu Jun, personal communication). Different
versions of the genome sequence serve different purposes and require different
levels of investment. The Syngenta and BGI genome sequences were produced
by a “shot gun” approach, whereas the IRGSP sequence used a map-based
approach. In shot-gun sequencing, the whole genome is sheared into small pieces
(1 – 2 kb), subcloned in plasmids, and sequenced. On average, each sequence is
analyzed 4 to 6 times. Then the sequences are assembled by a computer program.
In high eukaryotes, the high proportion of repetitive sequences can obscure the
alignment of truly contiguous sequences. Computationally demanding programs
are therefore needed to solve this problem to obtain a reliable assembly. The shot-
gun sequencing approach has been successfully applied to the complex genomes
of human and Drosophila. Depending on the objectives, the shot-gun approach
has the advantage of producing a draft sequence quickly and allowing researchers
to identify similar sequences in different species.
On the other hand, there are compelling reasons for developing a high-quality
sequence using a map-based approach. Here, each stretch of sequence is
physically anchored to a chromosome. At the outset, the IRGSP committed to
sequencing the rice genome at an accuracy of , 1 error per 10,000 bases to
provide the gold standard for accurate sequence annotation and for detecting
natural variation. On December 18, 2002, the IRGSP announced the completion
of the “phase 2” sequence data with all contigs anchored on the 12 rice
chromosomes. As of February 2003, 94% of the genome is covered, spanning a
genetic length of 1530 cM. By definition, finishing the rice genome means that:
(1) all sequences are physically aligned to the chromosome map, (2) there are no
gaps, and (3) all sequences are annotated. Currently, chromosomes 1, 3, 4, 10 are
considered to be finished (phase 3). It is expected that a completely finished
genome sequence will be accomplished by the end of 2004.
B. SALIENT FEATURES OF THE RICE GENOME
Although sequencing of some chromosomes remains to be finished, the

existing sequence information together with the completed Arabidopsis genome
(The Arabidopsis Genome Initiative, 2000) give us a good picture of the structure
and organization of the rice genome (Table III). For details, readers are referred
to the original papers by Goff et al. (2002), Feng et al. (2002), Sasaki et al.
(2002), and Yu et al. (2002).
The rice genome is predicted to contain a large number of genes with an upper
estimate as high as 62,000. On average, a gene can be found every 5– 7 kb, a
density similar to that observed in Arabidopsis. The functional classification of
rice genes is similar to the Arabidopsis sequence [see Fig. 9 in Yu et al. (2002)
and Fig. 1 in Goff et al. (2002)]. About 30% of the rice genes are plant-specific.
There are signs that rice experienced many duplication events at the level of
genes or chromosomal segments. To determine the extent of gene duplication,
Goff et al. (2002) examined homology among all predicted genes (longer than
300 bp) using the Syngenta draft sequence and found that 77% were homologous
to at least another gene, yielding about 15,000 gene families, a picture similar to
that observed in Arabidopsis and other eukaryotes. Local duplications (i.e.,
homologous genes mapped to a single BAC contig) range from 15.4 to 30.4%
depending on the chromosomes. The authors concluded that a genome
duplication event might have occurred in rice 40– 50 million years ago and
that chromosomes 11 and 12 were probably a result of duplication about 25
million years ago.
The extent of gene or chromosomal duplication has implications for the
strategy for assigning function to each gene. If gene redundancy is widespread,
simple knockouts at individual genes would not produce detectable phenotypes.
However, it is too early to tell the extent to which structural duplication is
translated into functional redundancy. Genes that are structurally similar may
have diverged to provide developmental and tissue specific expression. We have
yet to determine what proportion of the duplicated genes is indeed functionally
equivalent. Also, the number of gene families in tight clusters will affect the way
we confirm the function of a candidate gene. When we encounter a cluster of
genes with similar sequences, how can we determine whether the QTL effect is
due to a single member of the cluster or a collective effect of multiple members?
For these reasons, better gene annotation and understanding of the functional
relationship of paralogous and orthologous genes and members within gene
families are important.
While Arabidopsis and rice share common characteristics of plant genomes,
there are significant differences. The average size of rice genes is larger than
that of Arabidopsis. The larger gene size in rice is apparently due to larger
introns. Thus, the transcript size of a rice gene is not much larger than that of
Arabidopsis (Sasaki et al., 2002). The rice genome also has a high CG content.
Table III
Features and Composition of the Rice Genome Revealed by Different Versions of the Rice Genome Sequencea
Finished genome sequence Draft Sequence
Chromosome 1 Chromosome 4 Arabidopsis

Features of Nipponbare of Nipponbare Nipponbareb Indica 93.11c thaliana
Gene size (kb) 3.4 2.7 2.4 4.5 2.4
RICE FUNCTIONAL GENOMICS

Gene density (kb/gene) 6.4 7.4 5.7 4.5 4.5
No. exon/gene 4.8 4.4 4.2 3.3 5.2
Exon size (bp) 230 340 296 201 250
No. intron/gene 3.8 – 3.2 2.3 4.2
Intron size (bp) 605 370 371 356 170
CG content in coding 58.2 47.2 54.9 51.4 44.1
region (%)
CG content in non-coding 40.7 42.3 38.9 37.0 32.7
region (%)
CG content overall (%) 43.8 44.2 44.0 43.3 34.9
Physical distance/genetic 214 kb/cM for short arm; 208 kb/cM for euchromatic 253 kb/cMd – 213 kb/cMd
distance 288 kb/cM for long arm region; 636 kb/cM for
heterochromatic region
Estimated genome size (Mb) – – 420 466 125
Estimated gene number for 62,500 57,000 32,000– 50,000 46,000– 55,000 25,498
whole genome
a
Compiled from published data in Yu et al., 2002, Goff et al., 2002, Feng et al., 2002, Sasaki et al., 2002, and The Arabidopsis Genome Initiative (2000). – ¼ no
information provided in these papers.
b
TIGR IRGSP annotation statistics as of 23 Jan 2003 (www.tigr.org/tdb/e2k1/osa1/riceInfo/info.shtml#Genes)
c
Beijing Genomics Institute published draft (4X genome coverage), Wong et al. (2002).
d
For rice, assuming a genetic length of 1699 cM and a genome size of 430 Mb; for Arabidopsis, assuming a genetic length of 586 cM and genome size of 125 Mb.
67
Wong et al. (2002) observed a negative gradient of CG content starting from the
50 end to 30 end of the coding region. This CG gradient may reflect codon usage,
and hence may influence the accuracy of gene prediction as most gene prediction
programs do not take into account such a bias. It has been suggested that the large
difference between the number of predicted genes in rice and Arabidopsis may in
part be caused by artifacts of gene prediction.
Nearly all Arabidopsis genes can be found in rice but not vice versa. Based
on the draft sequences of indica and japonica and the completed chromosomes
1 and 4, nearly 50% of the predicted rice genes have no known function in
Arabidopsis. Yu et al. (2002) suggested that the rice genome represents a
“superset” of plant genes, only part of which is present in Arabidopsis.
Although part of the difference could be due to imperfect sequence annotation,
rice probably has many unique genes that are diverged from Arabidopsis genes.
Since a large sequence dataset is not yet available in other cereals, a pairwise
genome comparison with other cereals is not yet possible. Nonetheless, the
utility of rice sequences for other cereals has been demonstrated by using rice
sequences as substrates for investigating gene expression and genetic variation
in other species. Goff et al. (2002) reported that 90% of the rice sequences
arrayed as oligos on a chip can be used to detect expression in maize, barley,
and wheat.
C. COMPARISON BETWEEN INDICA AND JAPONICA RICE
Because genetic variation within a species represents the raw material for
adaptive changes, special attention is given to examining the occurrence of single
nucleotide polymorphisms (SNP), deletions and insertions between indica and
japonica rice on a genome-wide scale. At the whole-chromosome level, Sasaki
et al. (2002) used sequences from finished chromosome 1 of Nipponbare to query
the assembled contigs of the BGI draft indica genome data. They found about
78% of the japonica sequence in the indica database but failed to identify
matches to the remaining 22% of the sequences in chromosome 1. How much of
this discrepancy is due to genetic differences between the indica and japonica
sequences or to artifacts in the assembly of the indica sequence remains to be
determined.
At a finer scale, Feng et al. (2002) compared a 2.3-Mb region on chromosome
4 of japonica (Nippponbare) and indica (Guanglui 4) rice. While the overall
colinearity is maintained, there were many insertions and deletions. Insertions
occurred not only in the intergenic region but also in the coding regions.
Interestingly, there were more transposable elements in the japonica chromo-
some, contrary to the common perception that the total genome size of indica rice
is larger than that of japonica rice. Overall, there was one SNP per 268 bp
(0.37%) between the two sequenced varieties in this region of the chromosome.
Using genome sequence of indica variety 93-11 as a reference, Yu et al. (2002)

examined the polymorphisms with three varieties — Nipponbare, PA64s, and
Guanglui 4 — and found frequencies of SNP in unique sequences (relative to
93-11) at 0.50, 0.35, and 0.50%, respectively. Nasu et al. (2002) examined 417
chromosomal regions in three japonica, two indica varieties and a wild relative
O. rufipogon and found one SNP in 232 bp between two randomly chosen lines.
In comparison, about 1 SNP in 1000 bp was observed between Arabidopsis
ecotypes Columbia and Landsberg (Cho et al., 1999).
More recently, Katagiri (2003) presented preliminary data on the sequence
variation between chromosome 1 of indica variety Kasalath and Nipponbare.
Of 189,912 bp BAC-end sequences compared, one SNP/122 bp was observed.
About 0.15% of the SNP was found in expressed genes. Monna (2003)
analyzed six japnoica varieties (Nipponbare, Koshihikari, Kitaake, Akihikari,
Sensho, and Itadaki) and two indica varieties (Kasalath and Guang-lu-ai 4) and
found more SNP between the two indica varieties than between japonica and
indica varieties. It is too early to conclude that these statistics are indicative of
the degree of variation between the subspecies as these differences may simply
reflect differences between varieties (rather than between two subspecies).
A generalization of level of SNP can only be made after more sequence
data become available for multiple varieties within japonica and indica rice.
Such a determination of the within- and between-subspecies variation will
be important for applying nucleotide-level variation in genetic association
tests.
IV. KEY INGREDIENTS FOR GENE DISCOVERY IN RICE
The genetic blueprint of rice can be viewed as a dictionary of characters, but

without explanation of their function. To find the gene function encoded by each
gene or set of genes requires special kinds of genetic resources and
methodologies. Figure 3 depicts such a platform of resources and technologies
for large-scale gene discovery. The platform has three basic components: genetic
resources, phenotypic/biological evaluation, and genomic tools. These three
components are integrated by bioinformatics, which provides integration of data
produced from different activities. Furthermore, bioinformatics helps develop
hypotheses for further experimentation. This reiteration of hypothesis testing
and data production eventually leads to the discovery and validation of gene
functions. The framework brings multidisciplinary teams and different expertise
together. It also highlights the importance of delivering the products of gene
discovery to end-users through collaboration with breeding programs that can
incorporate genes into appropriate genetic backgrounds.
Figure 3 Public research platform for rice functional genomics: genetic resources, phenotyping,
high-throughput genomic tools, and bioinformatics. Linkage to breeding networks in rice growing
countries represents the practical dimension of rice gene discovery.
A. GENETIC RESOURCES
Genetic stocks harboring the traits of interest and associated variation are the
essential materials for revealing biological function of a sequence. A range of
genetic stocks can be used to identify or validate gene function. These include
mutants, segregating mapping populations, backcross lines, well-characterized
breeding lines and traditional germplasm. Among these genetic stocks, mutants
are the “work horse” of gene discovery because they reveal phenotypic changes
that correspond to a precise change in the genome. Furthermore, mutations often
lead to over- or under-production of biochemical intermediates that can lead to
identification of pathways. Induced mutations alone, however, are not sufficient
to understand phenotypic variation caused by allelic diversity. Therefore, it is
important to couple functional analysis with the diverse germplasm developed for
genetic mapping and breeding. Allelic variation present in conserved or selected
germplasm is particularly valuable because they represent products of natural
selection under various evolutionary forces.
B. HIGH-THROUGHPUT TECHNOLOGIES
A distinct feature of functional genomics is the application of high-throughput

technologies that can reveal the behavior of thousands of genes at one time
(Hieter and Boguski, 1999). Large-scale and genome-wide methodologies
become an essential part of experimental design, hypothesis testing, and data

analysis. We consider only briefly the current technologies that are sufficiently
robust for broad applications.
With available genome sequence information from the large-scale sequencing
of many organisms, it is possible to produce arrays carrying thousands of genes,
providing a high-throughput method for gene expression analysis (essentially
reverse northern analysis) and for the detection of genetic variation (reverse
Southern analysis) (DeRisi et al., 1997; Schena and Davis, 1999). Originally,
there were two main technologies for producing gene arrays: (1) spotting DNA
[cDNA clones, PCR products, oligonucleotides (oligos)] and (2) in situ synthesis
of oligos on slides by photolithography (Fodor et al., 1991). New and “hybrid”
methodologies have been developed: ink-jet deposition (Hughes et al., 2001) and
maskless array synthesis (Nuwaysir et al., 2002), which improve the flexibility
and utility of the arrays. Furthermore, some arrays (e.g., Diversity Array) assay
the genetic diversity of any organism without depending on available sequence
information (Jaccoud et al., 2001). While spotted cDNA arrays are currently the
most commonly used, increasing attention is given to oligo arrays (50 or 60 mers)
because of their increased specificity, quality control, and portability (Kane et al.,
2000). Depending on the design of the oligos (choice of gene sequence, oligo size
and number), oligo arrays can be used effectively to detect nucleotide
polymorphism or deletions in the target genomes. Currently, rice gene arrays-
cDNA or oligos- of varying sizes and coverage are produced by individual groups
or national programs. Progress is being made to develop large-scale oligo arrays
for rice that are publicly accessible (S. Kikuchi, personal communication; Beijing
Genomics Institute, personal communication).
Although gene arrays provide a picture of the transcription profile, they do not
reveal post-transcriptional regulation of the genes. As in transcript analysis,
proteomics provides a high-throughput platform to survey many proteins
simultaneously under different experimental conditions (Zivy and de Vienne,
2000). Proteins can be separated in two-dimensional polyacrylamine gel
electrophoresis (2D-PAGE), first by isoelectric focusing in the first dimension
followed by molecular weight in the second dimension. Protein spots can then be
extracted for microsequencing. In general, 2000 –3000 proteins can be resolved
and analyzed using the 2D gel technique coupled with mass spectrometry. New
technologies for proteomic analyses are being developed (Koller et al., 2002).
Readers are referred to an overview of proteomic technology and its applications
for gene discovery by Salekdeh et al. (2002a).
C. BIOLOGICAL EVALUATION
High-quality phenotypic data are central to a gene discovery program yet

infrastructure and expertise for phenotyping are often taken for granted. The lack
of adequate phenotyping capacity can be a bottleneck for gene discovery.

High-quality phenotyping means a precise, reproducible biological assay of the
trait in question. Phenotyping may entail different levels of sophistication from
simple visual evaluation to the measure of a detailed metabolic response under
highly controlled environmental conditions.
Overall, in a chemical- or irradiation-induced mutant population, 6 – 10% of
the mutant lines has visible phenotypes in either the vegetative or reproductive
stage. While the obvious phenotype (e.g., plant size and color) can be discerned
by inexperienced observers, subtle morphological changes often require trained
eyes. Plant breeders, for instance, are particularly keen to observe overall plant
architecture, panicle shape, and grain filling as a means to improving plant type
and yield. To maximize the use of any large collection of mutants, it would be
beneficial to have regular “grow-out” field days for expert observers to select
phenotypes of interest. For a lot of agronomic traits, the target phenotypes must
be observed through conditional screens in which mutants, isogenic lines and
breeding lines are subject to specific conditions to express the phenotypes. For
examples, in screening for drought-tolerance, the developmental stages
(vegetative or reproductive) at which the plants are subject to water stress
may determine the kinds of genes that can be identified. The nature and
parameters of conditional screens are therefore vital for successful detection of
the desired traits. Mutants with a uniform genetic background have the
advantage that any deviation in phenotype can be identified readily. As we
search for specific variants, innovative screening techniques using reporter gene
constructs will prove useful (Page and Grossniklaus, 2002). Furthermore,
screening for secondary mutations will be increasingly important for defining
the effect of the first mutations and revealing interaction between genes and
pathways.
D. BIOINFORMATICS
Bioinformatics brings together biological and computational tools that can

integrate data from structural and functional analyses and present new models
and hypotheses for further testing or validation. Bioinformatics helps us integrate
information to formulate new questions to be examined, hence bringing new
insights into the process of gene discovery. Because of the increasingly large
amount of information produced, it is no longer adequate to examine data in an
isolated manner. A key function of bioinformatics is to provide a system by
which information can be retrieved, analyzed, and distributed to a broad
community of researchers and users.
In summary, the four components — genetic resources, biological evaluation,
genomics tools, and bioinformatics — require diverse skills in biology and
modern technologies. They require infrastructure in modern science and
engineering (nanotechnology, computer science, etc.) and conventional

biological studies (phenotyping and innovative screening procedures).
While individual laboratories or institutions may be well suited to
pursue some of these activities, a single laboratory will unlikely have all
the resources and expertise to examine all questions. Thus, functional
genomics is necessarily a collaborative process and a public resource
platform provides a framework for the division of labor and formation of
partnerships.
V. FORWARD AND REVERSE GENETICS
Forward genetics refers to finding gene function by starting with a phenotype

and then identifying the genes responsible through inheritance studies, followed
by genetic mapping to locate the target gene on a chromosomal region.
By “walking” down the chromosome using overlapping markers, the DNA
sequence responsible for the trait can eventually be identified. The final step of
gene isolation can be hastened considerably with the whole-genome sequence
available.
Reverse genetics, in contrast, begins with a known DNA sequence and works
backward to identify the biological function ascribed to that sequence. Typically,
a sequence with a predictive function is used to search for knockout lines in a
pool of DNA or electronically from databases of insertion mutants (see below).
Once a mutant is found, it can be assayed for phenotypes based on the functional
prediction. For organisms with whole-genome sequences available, this provides
a new paradigm for gene discovery. With forward and reverse genetics, it is
within our reach to systematically assign a function to every gene of interest in
rice. For this reason, there have been many initiatives around the world to
produce rice mutants. Below, we will use several representative mutant stocks to
demonstrate the various strategies.
A. INSERTIONAL MUTANTS
Insertional mutagenesis has been the favored approach for tagging genes of
interest. The most advanced systems have been developed in maize and
Arabidopsis using both transposons and integrative plasmids as the mutagenic
agents. Although the initial investment in producing insertion mutants is high, the
pay off is great because the mutated genes are potentially tagged. Using a variety of
vectors originally designed and tested in Arabidopsis, rapid progress has been
made in the production and characterization of rice insertion mutants over the past
several years.
1. T-DNA Tagging
The Arabidopsis knockout facility at the University of Wisconsin recently

established a population of 60,480 T-DNA-tagged lines (Krysan et al., 1999).
This is a significant step toward the production of genome-wide mutations.
T-DNA insertion in Arabidopsis is probably a random event, with the inserted
sequences being stable through multiple generations (Azpiroz-Leehan and
Feldmann, 1997). Recently, this T-DNA tagging strategy was employed in rice
(Jeon et al., 2000). An’s laboratory at Pohang University of Science and
Technology in Korea has produced approximately 100,000 fertile rice lines that
have been tagged by T-DNA using binary vectors of Agrobacterium tumefaciens
(Jeon et al., 2000). Figure 4 illustrates the basic principle of insertional
mutagenesis using known DNA elements. Analysis of randomly selected
transgenic plants has indicated an average of 1.4 loci of T-DNA inserts per
line. Therefore, the T-DNA knockout stocks would provide a 63% probability of
finding an insertion within a given gene if T-DNA inserts randomly. It appears
that T-DNA prefers transcriptionally active regions in rice (Barakat et al., 2000).
Therefore, the chance of finding a knockout gene in the rice T-DNA insertional
stocks would be higher than the estimated value.
One difficulty in establishing a population of T-DNA-tagged lines is obtaining
enough seeds from the first generation. In experiments with 10,743 primary
transgenic plants, 39% bore fewer than 50 seeds and only 50% yielded more
than 100 (Jeon et al., 2000). Yin and Wang (2000) observed similar results from
their analysis of 2633 transgenic rice plants. Although the majority of the
transgenic plants became fertile in the T2 generation, 22% of the lines showed
Figure 4 Production of insertion lines using introduced T-DNA and endogenous retrotransposon
Tos17. T-DNA can be inserted into either a gene (line 1) or an intergenic region (line 2). During the
transformation procedure, the endogenous retrotransposon Tos17 can be activated and inserted into a
gene (line 3).
less than 50% fertility. Reduced fertility is probably due, in part, to the lethal
effects of T-DNA insertion. However, a considerable number of lethal mutations
are likely induced during tissue culture. Those lines that show low fertility must
be individually amplified before being used for further analysis.
2. Transposon Tagging
Transposon tagging has been a powerful tool for isolating new genes since
the controlling element was first recognized by McClintock (Fedoroff et al.,
1983). The first successful cloning of a plant gene was achieved via the
Ac/Ds (Activator – Dissociation) transposon system (Fedoroff et al., 1984).
Other transposon systems, such as En/Spm (Enhancer/Suppressor – mutator)
and Mu (Mutator), have been used for cloning several genes of maize
(Walbot, 2000). These elements are active not only in their natural hosts but
also in other plants, including Arabidopsis.
The maize Ac/Ds system has been tested for gene tagging in rice. First, the
autonomous Ac element was cloned between a promoter and the hygromycin
phosphotransferase coding region. The construct was then introduced into rice
chromosomes by direct transformation. Transposition of the Ac element was
proven by recovering hygromycin-resistant plants (Izawa et al., 1991; Murai
et al., 1991). Enoki et al. (1999) analyzed the behavior of 559 plants from four
transgenic rice families through three successive generations; 18.9% of the plants
contained newly transposed Ac insertions. DNA pools from 6000 Ac-inserted
plants were organized in a 3D matrix (row — column –plate pools) and subjected
to PCR screening. Of the 14 randomly selected genes, two knockouts were
identified, one of which encodes the rice cytochrome P450 (CYP86) gene.
Molecular analysis of Ac transmission to the progeny revealed that its germinal
transmission was detected in 9 of the 12 insertions, indicating that some Ac
transmissions occurred in somatic tissues that would not transmit to the progeny.
Sequence analysis of 99 flanking regions over the 50 region of Ac indicated that it
preferentially transposed into protein-coding regions.
The non-autonomous Ds element has been transposed in the presence of Ac
transposase via the direct gene transfer method (Shimamoto et al., 1993;
Sugimoto et al., 1994). Germinal transposition of Ds was observed at a high
frequency in the R2 progeny when a transgenic plant containing the Ds element
was crossed with a transgenic plant carrying Ac transposase under the control
of the cauliflower mosaic virus (CaMV) 35S promoter. A wide spectrum of
mutations, affecting growth, morphogenesis, flowering time, and disease
resistance, was observed in the Ds population (Izawa et al., 1997). Whether
these mutations are due to Ds must still be determined. The frequency of
Ds transposition declined significantly in subsequent generations, even in
the presence of Ac transposase. However, transposition could be recovered

in certain lines through protoplast regeneration.
Similar Ac/Ds gene tagging systems have been developed for rice by
employing the Agrobacterium-mediated gene delivery method (Chin et al., 1999;
Nakagawa et al., 2000). This method mediates transfer of one to a few copies of
T-DNA into plant cell chromosomes, thereby making the transgene more stable
compared with the direct gene transfer method (Dai et al., 2001). As many as
80% of the Ds elements have been excised from the original integration sites in
the presence of Ac transposase activity provided in trans. The germinal excision
frequency was no more than 40% (Nakagawa et al., 2000). Repetitive ratoon
culturing caused new transposition, at about 30% frequency, thereby demonstrat-
ing that the procedure is capable of producing a large population of mutants
(Chin et al., 1999).
Linkage analysis of the empty donor site and the transposed Ds insertion site
revealed that four of five insertion sites were linked to the donor site (Nakagawa
et al., 2000). This indicates that most of the transposition events occurred at a
tightly linked site, as had been reported for Arabidopsis (Machida et al., 1997;
Sundaresan, 1996). In that species, the physical map positions of 356 Ds
insertions showed significant preference for transposition to the regions adjacent
to the nucleolus organizer regions (Parinov et al., 1999). For any given gene,
insertions preferentially occurred at the 50 end of the gene. Therefore, the short-
range and highly preferential transposition system can be effectively used for the
targeted mutagenesis of closely linked genes.
In contrast, transpositions of another mobile element, En/spm, were well
distributed over different chromosomes and were not clustered at a few
genomic locations in Arabidopsis (Wisman et al., 1998). However, the tendency
of the En/Spm element to amplify can potentially complicate the interpretation of
phenotypes and molecular analyses. Ac/Ds in conjunction with Spm/dSpm has
been proposed for use in developing a four-element system for directed tagging
of crop-specific alleles (Phogat et al., 2000). In the first step, stocks carrying Ds
within dSpm would be crossed to a line with Spm to facilitate random
transposition of the transposon. In the second step, individual lines would be
crossed to a line with Ac to induce local transposition of the Ds element. This
four-element system would therefore exploit the natural tendency of unlinked
jumps of dSpm and linked jumps of Ds.
Unfortunately, the Ac element in this four-element system would undergo
frequent excision from the target gene, causing variegation. To overcome this, a
self-stabilizing Ac derivative has been developed that undergoes autonomous
transposition but is stable after integration (Schmitz and Theres, 1994). Charng
et al. (2000) have designed an inducible transposon system using an element that
contains a PR-1a promoter::transposase fusion (PR-1a::Tpase). Treatment with
salicylic acid (SA) induces the expression of transposase, resulting in the
transposition of Ds in somatic tissues. Germinal transposition events also have
been observed after SA treatment of flowers. The development of inducible

transposable element systems will also be valuable for rice. Recently, Qu et al.
(2003) explored the use of a GVG chemically inducible system to control
transposition of Ds. Preliminary data suggested that transposition of Ds could be
induced by dexamethasone (DEX).
3. Retrotransposon Tagging
Another effort toward gene tagging in rice involves introducing the tobacco
retrotransposon Tto1 and demonstrating its autonomous transposition through
reverse transcription (Hirochika et al., 1996). A transposable element, Tag1 from
Arabidopsis, has also been analyzed in rice (Liu et al., 1999b). The transcription
and excision behaviors of Tag1 were similar in both rice and Arabidopsis.
However, the excision was tissue-specific only in rice.
A survey of 73,000 sequence-tagged connectors (STC), corresponding to
nearly 50 Mb of the rice genome, showed that retroelements are randomly
distributed with respect to potential genes (Mao et al., 2002). Retrotransposons
are both functionally and structurally different from well-characterized
transposable elements such as Ac and Spm. These retrotransposons are believed
to be involved in gene duplication as well as in the regulation of gene expression.
Transposition of some retroelements can be induced by stresses caused by
pathogen infection, cell culture, and wounding (Takeda et al., 1998; Parinov
et al., 1999).
Such endogenous elements in the rice genome have provided excellent tools
for gene tagging. One such element is Tos17, which is activated during tissue
culture and amplified up to 30 copies (Parinov et al., 1999). A gene knockout
system using Tos17 has been developed for identifying insertional mutations in
several genes. After screening 550 plants that were mutagenized by Tos17,
Sato et al, (1999) identified a mutation in the homeobox gene OSH15. Rice
phytochrome A ( phyA) mutant lines have also been isolated from a Tos17-
induced mutant population using a 3D DNA-pooling system (Takano et al.,
2001). By screening for viviparous mutants, Agrawal et al. (2001) identified
Tos17 insertions in the rice zeaxanthin epoxidase gene (OsABA1) and in a novel
OsTATC gene, which shows a weak homology with bacterial Sec-independent
translocase TATC. More than 32,000 Tos17 insertion lines have been produced
and more than 8600 independent sequence-flanking insertion sites have been
determined (Miyao et al., 2001). The mutant stocks are available upon request
(see http://tos.nias.affrc.go.jp/~miyao/pub/tos17). Analysis of insertion points in
the rice genome sequence has revealed that insertions in the genic regions (exon
and intron) were 2-fold higher than in the intergenic region, suggesting that Tos17
prefers the genic regions for target sites. In addition, hot spots of Tos17 insertions
were observed, with some of them being clustered. This may limit the efficiency
of genome-wide gene tagging using Tos17 alone.
4. Activation Tagging
Conventional strategies, such as T-DNA mutagenesis or transposon tagging,

are not efficient for analyzing the function of redundant genes. Likewise, neither
T-DNA tagging nor transposon tagging will identify the genes that are required
during multiple stages of a life cycle and whose loss-of-function results in early
embryonic or gametophytic mortality (Weigel et al., 2000). Fewer than 10% of
the genes tagged in the Arabidopsis genome are likely to produce a visible
phenotypic change. Therefore, complementing technologies are needed to assess
the function of the remaining genes.
One such technique is the activation-tagging system developed in Arabidopsis
and successfully used for cloning several genes (Kardailsky et al., 1999; Ito and
Meyerowitz, 2000; Lee et al., 2000; Weigel et al., 2000). This system uses
T-DNA vectors that contain multimerized CaMV 35S transcriptional enhancers
positioned near the right T-DNA border (Hayashi et al., 1992). Gain-of-function
mutagenesis in rice by activation tagging requires a strong enhancer element.
Since the CaMV 35S enhancer element has enhancer activity in rice cells, they
can be used to produce activation tagging lines in rice (Fig. 5).
A transposon-mediated activation-tagging system has also been developed
using the Ds element that carries the tetramerized CaMV 35S enhancer (Mori et al.,
2000). From the T-DNA activation-tagging pools of Arabidopsis, Weigel et al.
(2000) characterized over 30 dominant mutants with various phenotypes.
Figure 5 Activation tagging using an insertion construct carrying a 4X 35S promoter to drive the
expression of endogenous genes. The multimerized 35S enhancer located at one end of the T-DNA
can enhance the expression of a nearby gene from upstream (top) or downstream (bottom) of the gene.
Analysis of a subset of the mutants has shown that the tagging vector causes
overexpression of the gene immediately adjacent to the inserted enhancer.
In rice, T-DNA activation-tagging vector pGA2715 was developed for both
trapping and activation tagging of rice genes (Jeong et al., 2002). The binary
vector pGA2715 contains the promoterless glucuronidase (GUS) reporter gene
next to the right border and the multimerized transcriptional enhancers from the
CaMV 35S promoter next to the left border. A total of 13,450 T-DNA
insertional lines were produced using the binary vector pGA2715. Histochem-
ical GUS assay revealed that the GUS-staining frequency from the pGA2715
lines was about two times higher than that from the lines transformed with
binary vector pGA2707, which lacks the enhancer element. This result suggests
that the enhancer sequence present in the T-DNA enhanced GUS tagging
efficiency. RT-PCR analysis of a subset of randomly selected pGA2715 lines
has shown that expression of the genes immediately adjacent to the inserted
enhancer was increased significantly. These results indicated that the large
population of T-DNA-tagged lines transformed with pGA2715 could be used
for trapping a gene using the GUS reporter as well as for isolation of gain-of-
function mutants in rice.
5. Entrapment tagging
An entrapment-tagging system allows for monitoring gene activity by creating

fusions between tagged genes and a reporter gene, such as GUS and green
fluorescent protein (GFP). Insertion of the promoterless reporter will not only
destroy normal gene function but also activate expression of the reporter gene.
Three entrapment systems are available: enhancer trap, promoter trap, and gene
trap (Springer, 2000). The devices for entrapment can be transferred into plant
cells as a part of T-DNA or transposons (Sundaresan, 1996; Martienssen, 1998;
Springer, 2000). This approach has been successfully applied to genes that are
difficult to identify by traditional methods. Some examples are regulatory
sequences that drive reporter gene expression in nematode-feeding structures
(Barthels et al., 1997), molecular markers for embryogenesis (Topping and
Lindsey, 1997) or regulatory sequences that mediate guard-cell-specific
expression (Plesch et al., 2000).
Activation of a reporter gene in a promoter trap vector can be as high as 30%
(Sundaresan, 1996). In rice, at least 5% of the T-DNAs and 10% of the transposed
Ds elements become activated in various tissues, for examples, roots, leaves,
flowers, and seeds (Chin et al., 1999; Jeon et al., 2000). If reporter gene activation
by certain environmental conditions or by chemicals such as growth substances is
included, total tagging efficiency can be higher. Some of the tags have displayed
tissue- or organ-specific reporter expression, while others have exhibited
ubiquitous expression patterns. One must study whether these lines will manifest
any mutant phenotypes in organs where the reporter has been activated, and
whether the phenotypes co-segregate with the inserted reporter gene.
The GUS gene has frequently been used for gene trapping in plants because
of the accurate detection of its gene product and tolerance for the N-terminal
translational fusions in its enzyme activity. However, one problem with the
GUS assay is the destructive nature of its staining and destaining procedures
(Jefferson et al., 1987). Non-invasive and non-destructive reporter genes, such
as GFP or luciferase, have not yet been widely used for gene trapping in
plants. Recently, a visualization system with a charge-coupled device camera,
band-pass filters, and a light source was used to demonstrate that green
fluorescence emitted from GFP could be visualized in the calli, dry seeds,
roots, and seedlings of transgenic rice plants (Chung et al., 2000). Such an
efficient visualization system should facilitate the use of a non-destructive
visual selection marker for entrapment tagging in rice. We are currently
investigating the efficiency of GFP expression in rice plants transformed with
the vector pGA2717 that carries the promoterless GFP next to the left T-DNA
border (Gyn An, unpublished data).
A new genomic DNA-based signal sequence trap method, the signal-exon trap
(SET), may help identify genes that encode secreted and membrane-bound
proteins (Peterfy et al., 2000). SET is based on the coupling of an exon trap to the
translation of captured exons, thus allowing the screening of exon-encoded
polypeptides for signal peptide function. This system may be helpful in the
discovery of novel members of known secretory gene clusters as well as for other
positional cloning approaches.
6. Forward Screening
Insertional mutant pools are useful resources for studying gene function. First,
they can be used to identify mutants whose function is altered by the insertion.
Because most mutations are recessive, their phenotypes are easily detected in a
segregating population. However, because rice plants are much larger than
Arabidopsis and need to be grown in specific conditions, large-scale phenotypic
screening is often limited by space and time. Systematic efforts by collaboration
with several research groups are therefore needed to screen a large number of
mutants. Because mutations could be induced by unknown endogenous
transposons during tissue culture, one must determine whether the phenotypes
induced by the rice transformation are due to the insertion. Once the mutant
phenotype has been shown to co-segregate with the insertional element, the
sequence flanking the insert can then be isolated by PCR-based methods, such as
thermal asymmetric interlaced (TAIL) PCR (Liu and Whittier, 1995),
adapter-ligated PCR (Balzergue et al., 2001), inverse PCR (Triglia et al., 1988),
a universal biotinylated adapter amplification procedure (Hanley et al., 2000), a
panhandle PCR (Walbot, 2000), or a plasmid rescue system (Weigel et al., 2000).
7. DNA Pool Screening
Insertional mutants can be identified in a given gene via PCR-based screening

(Krysan et al., 1999; Sato et al., 1999). Using a gene-specific primer and a primer
located near the end of the insert, the fragment flanking the insert can be
amplified and its sequence then determined. Because screening of individual
lines requires great effort, DNA pools of a large number of lines are commonly
used for identifying a knockout. In Arabidopsis, it was reported that the
maximum useful pool size is about 2350 lines per pool (Krysan et al., 1999).
Because the rice genome is about four times larger than the Arabidopsis genome,
a pool of 500 lines (less complexity) would be needed for PCR-based screening
of the knockout mutations in rice. DNA pool screening has been performed from
the Tos17 insertional pools. An’s laboratory has prepared DNA pools from
40,000 T-DNA lines and provided the screening service for the isolation of
mutants from the T-DNA insertional lines. Figure 6 shows the overall scheme of
detecting insertional rice mutants from the screening of DNA pool from T-DNA
tagged lines.
8. Insertion Sequence Database
Random sequence analysis of the fragments flanking the insertional element

can help identify gene function. Flanking sequences from Ds insertions of 931
independent transgenic lines in Arabidopsis have been characterized (Parinov
et al., 1999). In a similar study, Ito et al. (2002) determined the genomic
sequences flanking the Ds elements and mapped the elements’ insertion sites in
1173 transposed lines by comparison with the published genomic sequences.
One-half of the lines contained Ds on the same chromosome, in particular, the hot
spots near the three starting loci. In the other lines, the Ds elements were
transposed across chromosomes. The Ds elements tended to transpose near the
chromosome ends and rarely transposed near the centromeres. Likewise, direct-
sequence characterization of 761 Mutator-tagged fragments in maize has
shown that random sequencing of transposon-tagged fragments can produce
significant numbers of interesting transposon-tagged genes and mutant lines
(Hanley et al., 2000).
In rice, the flanking sequences of Tos17 were determined by TAIL- and
suppression-PCR (Hirochika, 2001). As of October 2002, 23,709 flanking
sequences from 2424 insertion lines had been determined (A Miyao et al.,
Figure 6 PCR screening of DNA pool obtained from T-DNA insertion lines. a) DNA from
individual plants pooled in different levels of complexity. Each pool contains DNA from 50 to 100
lines. A super pool consists of 500 lines. b) A rice gene OsMADS6 is used to illustrate PCR screening
to detect the tagged mutant. Four screens were conducted using two primers from the gene and two
unpublished data). These are classified into 8653 independent flanking

sequences, about 39% of which showed homology to known genes. The flanking
sequences of T-DNA insertional lines and Ds-tagged lines of rice are being
determined. This strategy should be useful for both the identification of mutants
and the discovery of new genes. Since the rice genome sequence is almost
completed, disrupted genes can be readily identified and it will be possible to map
the insertion sites on the chromosomes by comparing the flanking sequences with
the genomic sequence. An insertion mutant line of any gene can be identified
in silico and, thus, flanking-sequence databases will become the basis for reverse
genetics studies. Construction of such databases through collaboration with
laboratories producing rice insertional mutant lines will be valuable for
determining gene function on a large scale.
B. CHEMICAL- AND IRRADIATION-INDUCED MUTANTS
Although insertional mutagenesis represents the most direct approach for

tagging a particular gene, there are technical and logistical problems in achieving
complete genome coverage based on insertional mutagenesis alone. First, the
investment of producing a large collection of insertion mutants is high and not
necessarily feasible for many laboratories. Second, insertional mutagenesis based
on the introduction of exogenous transponsons and T-DNA is often genotype-
dependent. Currently, only certain japonica rice genotypes are amenable to
efficient large-scale transformation. Activation of retrotransposons such as Tos17
depends on the ability to perform tissue culture and the level of transposition also
varies among genotypes. Third, the transgenic nature of the mutants often
prevents large-scale field cultivation in certain parts of the world. Finally, parts of
the genome may not be accessible to insertional inactivation, thus preventing
complete genome coverage. In contrast, chemical and irradiation mutagenesis
offer several advantages. Production of the mutants is relatively inexpensive. Any
genotype can be mutagenized and the distribution of mutations is probably
random in the genome. The main shortcoming of the chemical- and irradiation-
induced mutants is that they are not tagged, making subsequent gene isolation
difficult. However, with the new high-throughput genotyping techniques, it is
now possible to detect mutations rapidly, hence increasing the utility of these
mutants.
primers located near the right or left border of the T-DNA. In the first step, super pool 7 was detected
to carry a T-DNA insertion in the OsMADS6 gene. To increase sensitivity and fidelity, a DNA-gel blot
analysis was performed with the radioactively labeled OsMADS6 probe. In step 2, pool 65 was
identified to carry the insertion. An individual line that carried the insertion was identified in step 3.
The exact insertion position can be determined by sequencing the PCR product from step 3.
1. Choice of Mutagens
To produce a comprehensive mutant library useful for forward and reverse

genetics, we need to consider the efficiency and specificity of the mutagens.
Efficiency means how much potency the mutagen has for producing mutations. It
also means that the mutagenesis procedure should be straightforward such that a
sufficient number of mutations in the genome can be produced in a few large-
scale experiments. In this respect, chemical mutagenesis is more convenient than
irradiation experiments as special radiation facilities are not needed. Specificity
refers to the predominant kinds of mutations (small or large deletions, point
mutations) produced by a particular mutagen (Koornneef et al., 1982).
Depending on the approaches to be used to identify mutations, specific mutagens
can be selected to produce the desired type and size of genetic lesions.
Ethyl methanesulfonate (EMS) is perhaps the most commonly used mutagen
in plants because of its potency and ease of use. The mutagen causes mostly
transition point mutations. In rice, EMS mutagenesis involves soaking the seeds
in an aqueous solution at a chosen concentration (from 0.2 – 2.0%) for a given
time. It should be emphasized that rice genotypes have a different sensitivity to
mutagens. It is therefore important to produce a kill curve to determine the
desirable level of survival. An advantage of EMS mutagenesis is that a relatively
small mutant population (, 10,000) is needed to provide near-genome coverage
because the genome can tolerate a large number of point mutations without
showing deleterious effects. In Arabidopsis, point mutation density can be as high
as four mutations per Mb (L. Comai, personal communication).
Deletion can be induced by a variety of chemical and irradiation treatments.
Ionizing radiation has been widely used to induce mutations for plant breeding
and classical genetic analysis, but the consequences of ionizing radiation have
been examined at the molecular level only in a few organisms. Shirley et al.
(1992) examined the effects of fast neutron (FN) and X-ray irradiation at two
genetic loci encoding chalcone flavonone isomerase and dihydroflavonol
4-reductase in Arabidopsis. They found not only deletions at the loci but also
distant chromosomal rearrangements. From a population of FN-induced M2
seedlings, Bruggemann et al. (1996) isolated 20 mutations at the HY4 locus in
Arabidopsis; 15 of these were deletions detectable by RFLP analysis. Okubara
et al. (1994) used FN to produce lettuce mutants with increased susceptibility to
downy mildew. Of 14 mutations at four Dm genes isolated from 2211 FN-treated
M2 lines, 10 were shown to be deletions by analysis of RAPD markers linked to
the Dm genes.
Besides ionizing irradiation, several chemicals have been shown to induce
predominantly small to kilobase-size deletions. Using diepoxybutane (DEB) to
induce mutations at the rosy locus in Drosophila, Reardon et al. (1987) found that
43% of the mutations were deletions ranging from 50 to 8 kb. Yandell et al.
(1994) found that trimethylpsoralen was effective in inducing deletions, ranging
from 0.1 to 15 kb in Caenorhabditis elegans. Aguirrezabalaga et al. (1995)

showed that hexamethylphosphoramide (HMPA), a cross-linking agent, caused
deletions in 96% of all mutants recovered in Drosophila.
How many mutants do we need to have sufficient coverage of the genome? In
Arabidopsis, Krysan et al. (1996) estimated that the probability of any gene being
tagged by T-DNA insertion depends on the size of the gene and the total genome
size. Dividing the total genome by an average gene size will give the number of
targets in the genome to be tagged or mutated. In Arabidopsis, to have a 95%
chance of finding an insertion in a 3.2-kb gene, one would need 95,000 T-DNA
lines (assuming 1.5 insertions per line). Extending the same reasoning to rice
(with approximately 4 £ the genome size of Arabidopsis), the population size
needed for genome coverage would be . 400,000 lines, assuming one deletion
per line. Since multiple mutations are induced by chemical or irradiation
mutagenesis, fewer independent lines are needed. For example, assuming an
average of 10 deletions per line, one would need approximately 40,000 lines to
have a 95% chance of mutating any gene in the genome. These are obviously
rough estimates as we do not have an accurate estimate of the number of induced
mutations per genome. Nonetheless, it appears that a population of 50,000 mutant
lines should provide reasonable genome coverage.
2. Deletion and Point Mutation Stocks
To our knowledge, there are currently two large collections of chemical- and
irradiation-induced rice mutants. The first one is an FN-induced population
originally developed by Pamela Ronald at the University of California-Davis
using japonica variety M202 and subsequently acquired by a private company.
The collection consists of 24,660 M2 lines that have been used for high-
throughput PCR screening (Li et al., 2001).
The second collection is produced and maintained at IRRI. Four mutagenic
agents — fast neutron, gamma ray, DEB, and EMS — were used to produce
mutants with different sizes of genetic lesions. Variety IR64 was chosen as the
standard genotype for producing a comprehensive mutant population in indica
rice. IR64 is the most widely grown rice variety in the tropics and it carries many
valuable agronomic traits related to yield, plant architecture, grain quality, and
tolerance for biotic and abiotic stresses. Creating mutations in such an elite
genetic background can facilitate the detection of phenotypic changes in
important agronomic traits (Leung et al., 2001). Currently, the population
consists of approximately 30,000 lines at the M4 stage. The goal is to produce
. 40,000 independent mutant lines toward the end of 2003 to give a high
probability of detecting a mutation in most genes. By producing a population of
mutants with an identical genotype using different mutagens, we aim to produce
a large allelic series for any locus as a tool for gene validation. Furthermore, the
population is advanced to individual M4 lines such that sufficient seeds are
available for distribution for the screening and evaluation of quantitative traits.
Deletions. The concept of using deletion mutations for forward and reverse
genetics is straightforward though it could be technically challenging. If sizable
DNA regions (e.g., 100 –1000 base pairs) are deleted in a genome, it should be
possible to detect such lesions by various genotyping approaches. As a proof of
concept, Chang et al. (2003) evaluated the utility of a genome wide chip to detect
deletions in known mutations. DNA from a gamma ray induced d1 mutation with
a known deletion in the locus encoding a G-protein (Fujisawa et al., 1999) was
used to hybridize to 21,000 predicted genes (each represented by a 16 25-mer
probe set) on a Rice GeneChip. A hybridization signal five times below the wild-
type was used as a criterion to indicate deletion. A stretch of deletions
encompassing four genes was detected on chromosome 5 precisely at the map
location containing the d1 gene. Results from experiments with a deletion at the
Xa21 locus, however, were less definitive due to cross hybridization between
similar sequences of the disease resistance gene family. Nonetheless, these
experiments open up a powerful approach to apply genome-wide oligo chip to
identify deletions in mutants that have been phenotypically well characterized.
A successful reverse genetic screen using deletion mutants was first
demonstrated in C. elegans (Jansen et al,. 1997; Westlund et al., 1999; Liu
et al., 1999a). Unlike screening for insertion lines where the sequence of one end
of the T-DNA or inserted transposon is known, screening for deletion requires
sequence information flanking the target region on both sides. In C. elegans, a
high-throughput approach was developed by growing the worms and extracting
genomic DNA in a 96-well microtiter plate format. DNA is pooled at different
levels of complexity (pooling of mutant lines, plates, and multiple plates) and
PCR primers are then designed to bracket a region of the gene for deletion
detection. The PCR amplicon of a deletion mutant, despite its low frequency,
would be preferentially amplified in a complex DNA pool as a small amplicon
(due to deletion). Differential amplification can be further enhanced by adjusting
the extension time in the PCR procedure. Liu et al. (1999a) described a
systematic PCR screen using C. elegans mutant libraries produced by four
mutagens: EMS, ethylnitrosourea, diepoxyoctane, and ultraviolet-activated
trimethylpsoralen. These mutagens produce deletion sizes from 700 to 2900 bp
(average 1400 bp). From screening a 3-kb window for more than 100 genes, this
method detected a deleted target gene per 600,000 mutagenized genomes.
The approach was successfully extended to plants by Li et al. (2001). Using a
population of 51,840 FN-induced Arabidopsis lines, they succeeded in detecting
deletion mutants in a DNA pool containing 2592 lines. The scanning window
chosen was from 3 to 17 kb, somewhat wider than that used in C. elegans.
Twenty-one out of 25 (84%) genes attempted were successful. The detected
deletions ranged from 0.8 to 12 kb. The authors emphasized the need to use PCR
conditions that favored the amplification and detection of the smaller amplicons.
The screening technique was applied to a rice mutant library of 24,660
FN-induced lines and a 2.5-kb deletion in a target gene, RG1, was found. We are
developing a modified PCR screening and detection protocol for the IR64
mutants (Fig. 7). More than 8000 M3 lines derived from FN and DEB were
pooled (10 lines per pool, with 10 plants per line) and amplicons were observed
on polyacrylamide gel to increase the sensitivity of detection. In a pilot
experiment, a deletion at the PAL1 gene (phenylammonia lyase) was identified
(P. Manosalva, unpublished data). By using a DNA pool of lesser complexity
(1/100 dilution), the need for preferential amplification of the small amplicons is
less, but the pooling efficiency is lower than that reported by Li et al. (2001).
Point mutations. In genetic analysis, it is often useful to have many point
mutations that give a range of allelic variants. The value of a point mutation stock
becomes even greater with the development of the high-throughput detection
system called Targeting Induced Local Lesions IN Genome (TILLING)
(McCallum et al., 2000a, 2000b).
TILLING is a reverse genetics strategy based on high-throughput detection of
point mutations in targeted genetic loci. This approach makes use of DNA strand
mismatches formed between mutant and wild-type DNA. EMS is the preferred
mutagen because it is known to produce predominantly A/T to G/C transitions,
allowing the prediction of nucleotide positions that would yield missense,
Figure 7 Reverse genetics in rice using populations of plants mutagenized with fast neutron and
diepoxybutane that create sizable deletions. An approach similar to that in screening insertion lines but
this protocol requires primer sequences on both side of the target gene. DNA is extracted from
mutagenized plants and grouped into pools. PCR reactions are performed on DNA pools using
oligonucleotide primer combinations from the gene of interest. PCR products are then loaded on an
agarose gel, transferred to a membrane, and hybridized to probes from the gene of interest. Detection
of a smaller amplicon suggests the presence of a deletion in the mutant DNA pool. Once the deletion
has been confirmed by PCR, the pools can be deconvoluted into individual lines.
nonsense, or splice site mutations. DNA from individual M2 plants is isolated,

pooled, and arrayed in 96-well plates. Primers are designed using CODDLE
(www.proweb.org/input/) to bracket a 1-kb region that most likely contains a
deleterious mutation in a target gene. The primers are then used to amplify the
gene of interest and, by denaturing and re-annealing DNA, homoduplexes and
heteroduplexes will be formed in the DNA pool. Originally, denatured HPLC was
used to detect the presence of a DNA mismatch; however, a high-throughput
system making use of enzymatic detection of heteroduplexes has proved to be
more efficient (Colbert et al., 2001).
For TILLING to be cost-effective, mutation density (number of mutations per
Mb) has to be high. It is therefore necessary to evaluate the relationship between
the mutagen dose and mutation density. We first produced a mutant population
using 0.8% EMS. DNA was isolated individually from approximately 2000 M2
plants, pooled (8 genotypes per pool), and subject to screening. Of 10 genes
screened, independent mutations were detected in two genes — pp2A4 encoding
the serine/threonine protein phosphatase catalytic subunit and a gene encoding
callose synthase. Sequencing of the mutated loci confirmed that they were A/T to
G/C transition mutations (B. Till, personal communication). Although the
frequency of the mutations is low, the preliminary data suggest that TILLING
can be a robust strategy for identifying an allelic series for any gene of interest.
High-dose mutagenesis is being evaluated to produce a high mutation density in
the mutant population.
C. NATURAL GENETIC VARIATION
Natural variants have been selected purposefully to produce favorable

agronomic traits in plant breeding. The widely used semidwarf gene sd1 in rice is
a mutation in a gene encoding gibberellic acid (GA)-20 oxidase, which controls a
step in gibberellin biosynthesis (Spielmeyer et al., 2002; Sasaki et al., 2002a).
Mutations in Os20ox2 (280-bp deletion and a point mutation in the conserved
motif of the dioxygenase) result in a semidwarf phenotype. Another example is
the use of allelic variation to adjust flowering time and maturity for local
adaptation. The gene Hd1 controlling heading date in rice encodes a zinc finger
transcription factor (homolog of CONSTANS (CO) of Arabidopsis). The “wild-
type” Hd1 allele in variety Nipponbare inhibits heading under long days (natural
field condition) but promotes heading under short days. Sequence analysis
showed that the Nipponbare allele has a 36-bp deletion in the zinc finger domain,
which probably affects the regulatory role of the transcription factor (Yano et al.,
2000). These examples suggest that many useful functional variants used in plant
breeding are in fact mutant forms of ancestral “wild-type” alleles. Therefore, the
natural variation present in traditional rice germplasm represents a reservoir of
useful alleles awaiting discovery.
The conventional way of investigating genes controlling quantitative variation

is to first map the QTL to a chromosomal region. For genes with large phenotypic
effects, one can produce a large mapping population with sufficient resolution to
fine-map the locus. For traits that require quantitative measurement, inbred
families (such as recombinant inbred lines or advanced backcross lines) are
needed for replicated trait measurement. With sequence information in hand, it is
now possible to begin with a gene sequence and look for sequence variation that
is likely to affect function. To some extent, genetic variation observed in
traditional or modern improved varieties can be viewed as a large collection of
mutations under the influence of natural or artificial selection. Many high-
throughput techniques developed for forward and reverse genetics are therefore
well suited for characterizing germplasm. For example, the same principle of
TILLING can be applied to detecting spontaneous mutations present in the
natural population (L. Comai and S. Henikoff, personal communication). In this
case, instead of pooling DNA from mutants, DNA of individual germplasm
accessions is annealed with a reference template DNA to detect mismatches.
Once a variant is detected, the PCR fragment can be sequenced to determine the
precise location of the variant.
VI. FUNCTIONAL VALIDATION OF RICE GENES
So far, we have discussed the kinds of genetic materials available for

discovering gene function. Without these genetic stocks, it is not possible to
assign biological function. However, the utility of these genetic materials can be
maximized when used with genome-wide analytical tools. It is important to note
that genome-wide approaches can help identify candidate genes and pathways in
a broad sense, but the data produced are often correlative in nature. Thus,
standard genetic and biochemical techniques are necessary to pinpoint the
functions and to dissect the pathways. Ultimately, the convergence of multiple
lines of experimental evidence holds promise in understanding gene function and
gene interactions. We highlight below several approaches that will probably be
common tools for gene function validation in the next few years.
A. GENE EXPRESSION ANALYSIS
1. Transcription Profiling
Transcription profiling refers to the display of a large set of genes that are
expressed under specified conditions. Genetic messages from genetic stocks with
different characteristics (e.g., salinity-tolerant versus sensitive) can be used to
assay which genes are expressed and the quantity of messages produced. Gene
expression profiling also reveals what genes are regulated in coordination. If
genome sequence information is available, one can look for a common sequence
upstream of co-regulated genes to find important motifs involved in gene
regulation. A parallel analysis of many genes at one time can reveal which genes
are collectively responsible for complex traits.
While the literature on the use of microarrays and gene chips is growing, few
published papers are available on genome-wide expression analysis in rice. The
first analysis was conducted by Kawasaki et al. (2000) who applied a set of 1728
cDNA libraries of salt-stressed roots obtained from salt-tolerant variety Pokkali.
The expression patterns of Pokkali and sensitive variety IR29 under salt-induced
conditions were compared in this set of genes. The data suggested that the most
significant difference in gene expression occurred within the initial phase of salt
stress. Within 15 min after salt stress (NaCl, 150 mM), about 10% of the genes
were either up-regulated or down-regulated in Pokkali. This expression
pattern was delayed in sensitive variety IR29. The results suggest that the salt-
tolerant property observed in Pokkali may in part be attributed to the rapid
expression of a small set of genes, transmitted as signals leading to a downstream
response. Failure to mount this rapid response, as in the case of IR29, could
be responsible for the down-regulation of transcription and eventual death
in 24 h.
A second example of gene expression analysis focuses on nutrient partitioning
during rice grain filling (Zhu et al., 2003). The objective of this study was to
identify sets of co-regulated genes involved in grain filling. In this case, the array
platform is an oligo chip designed to contain 16 probes (25-mer per probe)
covering the 30 end of each of 21,000 predicted genes. Thirty-three samples of
rice RNA representing different stages of grain filling were used as targets. Based
on sequence annotation, the authors narrowed down from 21,000 genes to 491
candidate genes that are potentially involved in the synthesis and transport of
carbohydrates, proteins, and fatty acids. Expression analysis led to the
identification of 269 “grain-filling genes” potentially involved in grain
development. By examining the upstream regions of these genes, Zhu et al.
(2003) identified a cis-element AACA that is over-represented in the promoters
of 103 genes presumably involved in grain filling. The AACA element is
hypothesized to interact with certain transcription factors to regulate nutrient
partitioning in the grain.
Both the stress tolerance and nutrient partitioning studies illustrate the power
of using gene arrays to identify a core set of genes that are potentially important
in regulating or being regulated in the traits of interest. The power of microarray
analysis will be much enhanced when expression data are compiled and made
available for data mining by the public. Efforts are underway to make available
the expression data of a variety of experiments from more than 60 laboratories in
Japan (http://red.dna.affrc.go.jp/RED; S. Kikuchi, personal communication).
2. Protein Profiles
Proteomic analysis of rice is still in its infancy but the few studies published so
far have suggested its potential to gain a deeper understanding of gene expression
in rice under various environmental and developmental conditions. Koller et al.
(2002) reported a comprehensive proteomic analysis of rice using tissues
from leaf, root, and seed. Complementary protein separation and analytical
techniques were applied (2D PAGE followed by HPLC-tandem mass
spectrometry and multidimensional protein identification technology). A total
of 2528 unique proteins was identified. About 32% of all the proteins identified in
leaves, roots, and seeds could not be functionally classified based on known
databases.
Salekdeh et al. (2002a,b) investigated the protein profiles of rice plants in
response to drought and salt stresses. In their studies, the 2D PAGE system can
reproducibly display about 2000 proteins. Salt-tolerant variety Pokkali and
sensitive IR29 were subjected to 50 and 100 mM NaCl at 14 and 21 days,
respectively, and root proteins were compared. Two proteins — ascorbate
peroxidase and caffeoyl-CoA O-methyltransferase — showed particularly inter-
esting patterns in Pokkali that are distinct from those of the salt-sensitive IR29.
Ascorbate peroxidase was 4.4 times more abundant in Pokkali than in IR29 in the
absence of salt stress. The constitutively high level of enzyme may confer a
greater antioxidant capacity to tolerate salt stress. Caffeoyl-CoA O-methyltrans-
ferase, an enzyme involved in lignin biosynthesis, was induced upon salt stress,
and it may be responsible for enhancing lignification as a mechanism to counter
salt stress. In their drought-stress studies, Salekdeh et al. (2002b) compared
protein profiles of seedlings of two genotypes, CT9993 and IR62266, under
different water regimes. CT9993 is an upland variety known for its drought-
tolerance attributes, whereas IR62266 is a drought-sensitive lowland variety.
Of more than 1000 protein spots analyzed, 42 showed a significant change in
abundance under drought stress. Four proteins — S-like RNase homologue,
actin depolymerizing factor, Rubisco activase, and isoflavone reductase-like
protein — showed distinct behavior under drought stress, suggesting that these
proteins may play a role in different drought-response mechanisms not
considered previously.
As of today, transcript and protein profiling represent accessible techniques
for examining gene expression on a genome scale. Undoubtedly, new and
improved high-throughput techniques will become available in the coming years,
but the principles of designing informative experiments remain the same. By
carefully choosing the appropriate experimental conditions as well as genetic
stocks such as mutants and genetically well-defined near-isogenic lines, gene
expression experiments can help us narrow down genes from thousands to a
manageable number, whose function can be confirmed by genetic and physio-
logical experiments.
B. GENE SILENCING
Post-transcriptional gene silencing (PTGS) was discovered in plants and

other organisms as a means to “defend” the host cell against foreign or
unwanted transcripts. This mechanism involves degradation of mRNA
mediated by dsRNA sharing homology with the target sequence, resulting in
the suppression of the expression of the gene (Chuang and Meyerowitz, 2000;
Baulcombe, 2002). Since the process is highly sequence-specific, it can be
used to selectively degrade the transcript of a chosen sequence. This targeted
mRNA destruction technique, called RNA inference (RNAi), has been used to
suppress gene expression in a wide range of organisms, including fungi,
animals, and plants. Details of the mechanisms and genetics of RNAi can be
found in several recent reviews (Waterhouse et al., 2001; Dernburg and
Karpen, 2002; Hannon 2002).
RNAi has been demonstrated as a powerful tool to systematically suppress
gene expression in C. elegans (Fire et al., 1998). RNAi can be induced by
injecting nematodes or feeding bacteria expressing dsRNA. The technology has
been used to suppress genes in chromosomes I and III of C. elegans (Fraser et al.,
2000; Gonczy et al., 2000). Because RNAi is sequence homology-dependent, it
can be used to silence the expression of individual genes or members of a gene
family by designing a gene-silencing sequence carrying a 26-bp region identical
to that of the target (Parrish et al., 2000). Compared to gene knockout, gene
silencing via RNAi is not always complete. This may be an advantage because
partial silencing may allow the analysis of genes whose knockout mutations
would cause lethality. In plants, Wesley et al. (2001) produced a series of vector
constructs that encode self-complementary “hairpin” RNA (hpRNA). This
modification enhances the efficiency of silencing over a dsRNA construct.
Furthermore, a new vector (pHELLSGATE) has been made to produce a gene
library for high-throughput silencing. In rice, there is increasing interest in
applying gene silencing as a complementary tool to gene knockout for
investigating gene function. Shimamoto et al. (2002) applied RNAi to investigate
the seven members of the OsRac gene family. By specifically targeting the
30 UTR region that is less conserved among the family members, it is possible to
selectively suppress each gene member of the family. We expect that RNAi will
be an important tool kit for gene validation as we encounter more and more gene
families with multiple members.
C. HETEROLOGOUS BIOASSAYS
A standard way to validate gene function is to transform a plant deficient in

a phenotype with a candidate gene to look for restored function in the
transformant. Technologically, genetic complementation is no longer a problem
in rice as various useful transformation vectors are available for performing

complementation tests. The problem is more on the robustness of the system to
assay a large collection of candidate genes in a high-throughput manner. Also,
technical problems exist in cases where the candidate allele is recessive and
where the recipient genotypes (e.g., a recessive mutant) could be compromised
in vigor which makes them unsuitable for transformation. Alternative and
efficient ways to obtain data equivalent to genetic complementation are
therefore needed.
One possibility is to use a heterologous expression system in which rice genes
are introduced into an alternative host and assayed for function. This is similar to
the use of shuttle vectors between yeast and bacteria for testing the gene function
of S. cerevisiae, a method widely used in fungal genetics. For plants, Arabidopsis
probably represents the most versatile system for transformation. The fact that
Arabidopsis has a comprehensive mutant collection offers a good chance that an
Arabidopsis mutant can be used for complementation with a rice candidate gene.
The same approach can be used to test rice genes in yeast mutants for phenotypes
that can be assayed in yeast cells.
The use of a virus vector to express foreign genes has been practiced for
many years in dicots, but much less has been done in monocots (Scholthof et al.,
1996; Karrer et al., 1998). Recently, a gene expression system has been
developed for monocots using the wheat streak mosaic virus (WSMV) (Choi
et al., 2000). WSMV has a broad host range, including important cereals such as
wheat, maize, barley and oats. In this system, the WSMV is engineered to accept
foreign genes as cDNA cassettes (e.g., candidate genes from rice). Production of
in-frame transcripts is done in vitro and the infectious transcripts are used to
infect wheat plants mechanically. The foreign gene carried by the virus will be
expressed in the cytoplasm and transmitted systemically throughout the plant,
enabling observation of the phenotypes in nearly all tissues and stages of plant
growth. Phenotypes conferred by the foreign genes can be assayed under
screening conditions that may last for more than a month (Fig. 8). The system
has been successfully tested using the GUS gene and antibiotic resistance.
Application of this approach to evaluate rice gene function is being explored (R.
Choi, personal communication). Using wheat as a heterologous host for assaying
rice genes can be attractive because of the similar phenology between rice and
wheat. This approach may be useful for testing a large number of candidate
genes with predicted functions based on sequence information or gene
expression data. The use of heterologous systems for gene functional assay
obviously has its limitation as not all rice genes will be properly expressed in
the alternative host. Also, some genes may not give the correct phenotypes.
Nonetheless, a heterologous expression system with high throughput potential
can offer an efficient way to produce supporting functional evidence for
rice genes.
Figure 8 Wheat streak mosaic virus (WSMV) as a vehicle to assay rice gene function in wheat
plants. Individual candidate genes as cDNA clones are inserted into the virus vector. Viral transcripts
carrying the foreign genes are obtained by in vitro transcription and used to infect wheat plants by
mechanical inoculation. The foreign genes are expressed and spread systemically along with the
WSMV. Wheat plants are assayed for phenotypes within one to two months.
D. GENE REPLACEMENT
When integrative vectors are introduced into plants, the vector-borne

sequences nearly always integrate into the host chromosome by non-homologous
recombination, a process called ectopic integration. In lower eukaryotes
(e.g., fungi), however, there is a high frequency of homologous recombination
when the introduced vector has a stretch of sequence identical to a sequence in
the host genome. The ability to conduct site-specific homologous recombination
in higher plants has been a long-sought goal because it offers perhaps the most
precise form of genetic engineering by replacing one allele with another carrying
the desirable effects. The problem of positional effect caused by integration into
different chromosomes can also be eliminated. The recent breakthrough by
Terada et al. (2002) holds great promise for adding gene replacement to the tool
kit of rice functional genomics.
To test the gene replacement technique, Terada et al. (2002) used the Waxy
locus as the target gene because the phenotype can be easily assayed. The
targeting vector pRW1 was engineered to carry the Waxy locus of rice plus
selective markers (hygromycin B resistance and diphtheria toxin) for positive and
negative selection of homologous recombination events. Out of about
600 hygromycin-resistant transformants obtained, six showed precise homolo-

gous recombination between the vector and the Waxy locus, yielding a success
rate of 1%. Since the vector construct is not designed for specific genes, the
authors believe that the method is of general applicability to other rice genes. If
the system is robust enough, any gene sequence, native or modified, can be used
to replace the resident gene. Thus, the technology is not only suitable for
validating the function of a gene but also suitable for creating the desirable allelic
state at a specified locus.
E. ALLELIC SERIES
Direct functional tests by genetic complementation can be substituted if there

is a large series of allelic mutations at the candidate gene locus. A perfect
correlation between phenotypic changes and multiple alleles at the locus can
provide convincing evidence that the candidate gene is responsible for a certain
phenotype. This approach was used in proving the identity and function of the
Mlo locus in barley (Buschges et al., 1997). With the large-scale production of
mutant lines worldwide, allelic series will be increasingly useful for confirming
the function of gene sequences.
VII. APPLICATIONS TO CROP IMPROVEMENT
Given the large number of traits amenable for improvement using genomic
tools, it will not be possible to describe all of them in meaningful detail.
Instead, it is more relevant to discuss what tools and approaches will become
common in plant improvement programs when the functions of most rice
genes are known. We can think of three important ways in which crop
improvement programs will benefit from having a functional dictionary of
rice genes. First, with the identification of actual genes controlling certain
traits, we will shift from using linked markers to actual genes as units of
selection. Second, we will make use of the knowledge of genetic regulation
and biochemical pathways to create genotypes that confer large phenotypic
effects. We will increasingly be able to recognize genetic switches or
regulatory circuitry that could lead to a significant change in phenotype. This
will drastically improve selection theories that are primarily based on
quantitative genetics models with little integration of genetic knowledge in
biochemical and metabolic terms. Third, sequence-guided comparative
biology will maximize the use of knowledge of gene function across related
or distant plant species. These approaches can be used in combination to
solve problems of a varying degree of complexity.
A. CANDIDATE GENE APPROACH AND ALLELE MINING
Although marker-aided selection (MAS) has been used successfully to

accumulate genes with large effects, its use for selecting quantitative traits has not
been widespread. The main reason is that most of the markers available are
loosely linked to the target gene and are easily uncoupled by recombination. As
functional annotation of the genome sequence continues to improve, we expect
that candidate gene approach to be widely adopted by breeding programs.
Immediate gain will come from the selection of traits such as tolerance for biotic
stresses where there is a good knowledge of host defense genes (Ramalingam
et al., 2003).
Conventionally, determining the phenotypic contribution of a particular allele
is based on setting up a cross between lines with high and low trait values and
examining the segregating progeny. Because genetic mapping experiments often
take 1 – 2 years to complete, functional association through genetic mapping is
not satisfactory for examining multiple traits at the same time. The ability to
establish a function– structure relationship at the level of nucleotide variation for
target alleles across a large number of genetic loci would be of great value in
breeding (Drenkard et al., 2000; Rafalski, 2002). Several studies in maize have
demonstrated the potential of association genetics by examining sequence
variation to establish structure –function relationships (Remington et al., 2001;
Thornsberry et al., 2001; Vigouroux et al., 2002; Whitt et al., 2002).
If candidate gene loci are known, it will be possible to identify sequence
variants at the loci and observe patterns of structure –function association — an
emerging area called allele mining. Implicit in allele mining is that the gene pool
is so large that it is not feasible to conduct controlled genetic crosses to determine
the function of each allele. The gene pool could be represented by
uncharacterized raw germplasm or advanced breeding lines in a breeding
program. One may consider both forward and reverse genetics approaches to
search for useful alleles. Forward genetics is applicable to advanced germplasm
that is known to be associated with an improved trait. The advanced germplasm
represents an enriched pool of useful alleles and, by examining the shift in allele
frequency of certain genes under selection, one can detect a pattern of association
that gives meaningful prediction of allele function. If breeding records and
pedigree information are well kept, it is possible to trace the transmission of
favorable alleles from the donor germplasm that gives rise to improved varieties.
The second set of materials represents the “raw” germplasm that has not been
widely used and whose phenotypes are largely unknown. In this case, the reverse
genetics approach can be used to examine a stratified collection of germplasm
grouped by geographic origin or historical performance data. Using a known
candidate gene sequence, the appropriate region of the gene sequence can be
amplified from germplasm and examined for association with phenotypes.
Several association mapping methods have been developed for such analyses
(Rigoutsos and Floratos, 1998; Pritchard and Rosenberg, 1999; Pritchard et al.,
2000). As more knowledge is accumulated on the function of various sequence
motifs and gene structures, functional inference based on association genetics
will continue to improve.
B. PATHWAYS AND GENETIC REGULATION
We envision that a significant outcome of functional genomics research will be

the discovery of genes that can contribute a large change in the value of breeding
targets. These include new pathways or biochemical mechanisms not obvious in
current rice genotypes. In this context, genes that determine the hierarchical
regulation of pathways are potential targets. A major class of such genes are those
encoding transcription factors that control genetic switch points affecting
multiple downstream genes.
In Arabidopsis, about 1700 genes belong to members of known classes of
transcription factors and there are 1300 such genes in rice (The Arabidopsis
Genomics Initiative, 2000; Goff et al., 2002). Manipulation of regulatory
elements can bring about dramatic changes in phenotypes that are often viewed as
being controlled by many genes with minor effects. The dramatic effect of
transcriptional factors is well illustrated by the CBFs (C-repeat binding factors)
and the related DREB (drought responsive element binding) factors. These
factors control multiple downstream genes that confer tolerance for cold, salt, and
drought stress (Stockinger et al., 1997; Jaglo-Ottosen et al.,1998; Kasuga et al.,
1999; Thomashaw 1999). Sequences with specific domains/motifs can be
extracted from public plant genomics databases, and families of transcription
factors can be categorized phylogenetically and putative gene functions inferred
based on sequence similarity and experimental data. Chen et al. (2002) examined
the expression patterns of 402 transcription factor genes in Arabidopsis using a
variety of defense-response mutants and plants under different stress conditions.
The results revealed coordinated regulation of specific “regulons”. Transcription
factors with putative roles on target traits can be further assayed by
overexpression and knockout mutations. Expansion of these studies will open
up opportunities for the manipulation of transcription control in plant breeding
(Zhu et al., 2003). We can expect rapid progress in this area as all ingredients for
conducting these experiments are already in place in rice.
C. CROSS-SPECIES INFERENCE OF GENE FUNCTION
We expect comparative genetics to become an integral part of future plant

improvement programs. Comparative analysis between genomes will broaden
the “gene pool” for trait manipulation in rice. Here, “gene pool” refers to the pool
of genetic knowledge. Kaplinsky et al. (2002) showed that comparison of
conserved and diverged sequences in the non-coding regions of genes in different
grass genomes can lead to the identification of critical regions of genes that
control function. Similarity in function will enable us to extend knowledge about
one species to another. A main objective of comparative genetic analyses is to
test whether orthologous genes (of common descent) function similarly in
different species.
With more knowledge accumulated in allied species, we will increasingly
look beyond rice for a solution to genetic and breeding problems in rice. For
example, the introgression of wild species alleles from wild to domesticated
rice is limited by a lack of chromosome pairing between AA and alien
genomes. In wheat, the Ph1 locus controls the pairing of homoeologous
chromosomes of different wheat genomes. The wild-type Ph1 gene, located on
chromosome 5B, prevents the pairing of homoeologous chromosomes so that
only homologous chromosomes of each genome can pair in meiosis. In the
absence of the Ph1 gene, homoeologous chromosomes can pair, allowing
genetic recombination between different genomes. By comparative mapping,
Foote et al. (1997) found a region on rice chromosome 9 syntenic to the wheat
5B region carrying Ph1. Using deletion mapping, Robert et al. (1999)
narrowed down the region to 4 Mb that potentially carry an ortholog of the
Ph1 gene in rice. We can now ask whether an orthologous Ph1 gene indeed
exists in rice, and, if so, can we manipulate the gene to facilitate gene transfer
between wild and domesticated rice (D. Brar, personal communication)? This
sort of cross-species inquiry will likely be routinely used in future plant-
improvement programs.
Even if the hypothesis of similar function is not supported, the finding of
distinct functions in genes of high sequence similarity points to exciting new
avenues for investigation. Shimamoto and Kyozuka (2002) recently reviewed the
genes controlling disease-response signaling and reproductive development in
rice and their counterparts in related species. They concluded that, between rice
and Arabidopsis, sequence similarities do not always imply functional similarity.
Orthologous genes in different species can play a divergent role. Changes in
function and regulation of orthologous genes could be the basis for divergence of
plant species.
VIII. INTERNATIONAL COLLABORATION AND

THE ROLE OF DEVELOPING COUNTRIES
Given the resources already available or being developed by different national

programs, the scenario of finding all agronomically important rice genes by 2010
seems realistic. But the agenda is large and it requires diverse inputs. Unlike
sequencing projects that can be carried out by a few laboratories with high
capacity in sequencing and bioinformatics, functional genomics requires a wide
range of experience from the cell to the whole plant to crop biology. To assign
functions to gene sequences, we must have the means and skills to examine a
wide range of phenotypes. Collaboration in rice functional genomics is therefore
a necessity if we are to achieve the goal of finding the function of all important
rice genes in a timely manner.
Besides technical considerations, there is an impetus to ensure that modern
science is put to use in solving difficult production problems, particularly in the
developing world, where food security and access are a major concern. Despite
the impressive gains in rice productivity over the past four decades, yield in
marginal agricultural areas has remained low. The development of adaptive,
high-yielding, and nutritious varieties must therefore be part of a strategy for
improving the livelihood of people living in these unfavorable regions (Hossain
et al., 2000). To solve these problems, there must be open access to genetic
knowledge and empowering tools. Collaboration in gene discovery ensures such
access.
A useful collaborative model in plant science has been provided by
the Multinational Arabidopsis Consortium (www.Arabidopsis.org/info/
2010_projects/MASC_info.html). The Arabidopsis consortium has participation
by 21 members from 13 countries where significant programs in Arabidopsis
research are taking place. It serves to promote communication and the exchange
of data and to coordinate research activities to maximize efficiency. The goal of
the Arabidopsis community is to find out the function of all genes in the plant by
2010. A similar goal is envisioned by the rice research community (Fig. 9). But
rice is different from Arabidopsis; many traits have to be measured under realistic
agronomic conditions. A rice consortium therefore needs to incorporate the
practical aspects of plant improvement and at the same time make use of the
expertise and experience from rice-breeding institutions around the world. In this
context, the roles of developing countries become important although they often
lack the infrastructure for genomics research. Rice-breeding institutions have the
genetic resources, phenotyping capacity, and breeding and evaluation networks
for testing gene function and genotype £ environment interactions. Breeding
institutions also have the capacity to assemble genes in varieties that will be
accepted by farmers and consumers. Finally, genomics, because of the
excitement and publicity involved, provides a fertile ground to attract
young students in the developing world to enter into plant science research,
hence building the human resources for future agricultural research and
development.
Efforts are already under way to forge international collaboration in rice
functional genomics. Since the beginning of the international sequencing project,
a consortium approach to accelerate rice gene discovery is being promoted by
Figure 9 Functional characterization of rice genes will be the challenge after completion of the
rice genome sequence. Delivery of research results for the development of adaptive and nutritious
varieties can be accelerated by international collaboration within this decade.
the International Rice Functional Genomics Working Group organized by IRRI

(Fischer et al., 2000). With the completion of the rice genome sequence, there is
consensus that a more structured consortium is needed to facilitate collaboration
and sharing of information and resources. In January 2003, the International Rice
Functional Genomics Consortium was formed with scientists from 17 institutions
of 12 countries exploring ways to consolidate resources and build common
strategies (http://www.IRIS.IRRI.org/IRFGC).
IX. CONCLUDING REMARKS
Recognizing the rapid progress in plant genomics, this chapter does not
attempt to provide a comprehensive review of the past and on-going work related
to rice genomics. Rather, we provide a snapshot of the current status of the field
and focus on the framework needed to make gene discovery relevant to plant
improvement. Rice is special because of its botanical history and biological
attributes. The completion of the rice genome sequence is a major milestone in
plant science; however, for the sequence information to be useful, it must be
converted into a format that allows investigation with a variety of genetic
materials. Functional genomics concerns the development of a resource platform
by which the function of every gene and its interaction with other genes can be
characterized and understood in an efficient manner. The excitement in functional
genomics lies in the creative application of high-throughput tools to probe into
the biological complexity of genetic networks and biological processes. In the
post-genome-sequencing era, we may save efforts in gene cloning in the
conventional sense, but there is no substitute for detailed studies involving
genetics, physiology, and biochemistry if we are to understand the biological
functions of genes.
Because rice is a major food crop in many societies and cultures, discovering
rice genes and having access to the knowledge inevitably has a social impact.
Starting in the 1960s, the Green Revolution has dramatically increased the food
supply in the developing world (Evenson et al., 1996; Hossain 1996). The yield
increase in the major cereals has largely come from genetic improvement of the
plants supported by improved agronomic practices. The higher efficiency in
production has reduced the rice price to about 50% in real terms over the past
three decades (Khush, 2001; M. Hossain, IRRI, personal communication).
Despite this impact, the problem of access to technology and resources-improved
seeds, fertilizer-has fueled debates about the achievements of the Green
Revolution. As we move forwards with the large-scale discovery of rice
genes — the ingredients for plant improvement — we must be mindful of public
access to the empowering knowledge and tools.
In this chapter, we have highlighted the resources available and being
developed to accelerate gene discovery and we also speculated on the future
adoption of new knowledge and tools in plant improvement programs.
Developing countries have a unique role to play in rice functional genomics
because they are rich in genetic resources and agronomic knowledge of the rice
crop. The benefits of engaging developing countries in gene discovery are
obvious but this requires the will and efforts to support the partnerships. The
power of bringing together a full range of expertise in basic and applied research
is enormous. The move toward more international collaboration in rice functional
genomics is very encouraging. Success in this effort will rest upon the continuing
goodwill and determination of the research community to sustain the spirit of
collaboration to channel gene discovery into practical crop improvement.
ACKNOWLEDGMENTS
We thank Alice Bordeos and Ramil Mauleon for assistance in the preparation of
this manuscript, Il Ryong Choi for providing Fig. 8 and valuable discussion, and
Bill Hardy for technical editing. We are grateful to Rebecca Nelson and Jan
Leach for their helpful comments and to many colleagues who share their
knowledge on the topic.
REFERENCES
Agrawal, G. K., Yamazaki, M., Kobayashi, M., Hirochika, R., Miyao, A., and Hirochika, H. (2001).
Screening of the rice viviparous mutants generated by endogenous retrotransposon Tos17
insertion. Tagging of a zeaxanthin epoxidase gene and a novel OsTATC gene. Plant Physiol. 125,
1248–1257.
Aguirrezabalaga, I., Nivard, M. J. M., Comendador, M. A., and Vogel, E. W. (1995). The cross-
linking agent hexamethylphosphoramide predominantly induces intra-locus and multi-locus
deletions in postmeiotic germ cells of Drosophila. Genetics 139, 649–658.
Ahn, S., and Tanksley, S. D. (1992). Comparative linkage maps of the rice and maize genomes. Proc.
Natl Acad. Sci. USA 90, 7980–7984.
Azpiroz-Leehan, R., and Feldmann, K. A. (1997). T-DNA insertion mutagenesis in Arabidopsis:
going back and forth. Trends Genet. 13, 152 –156.
Balzergue, S., Dubreucq, B., Chauvin, S., Le-Clainche, I., Le Boulaire, F., de Rose, R., Samson, F.,
Biaudet, V., Lecharny, A., Cruaud, C., Weissenbach, J., Caboche, M., and Lepiniec, L. (2001).
Improved PCR-walking for large-scale isolation of plant T-DNA borders. BioTechniques 30,
496–504.
Barakat, A., Gallois, P., Raynal, M., Mestre-Ortega, D., Sallaud, C., Guiderdoni, E., Delseny, M., and
Bernardi, G. (2000). The distribution of T-DNA in the genomes of transgenic Arabidopsis and
rice. FEBS Lett. 471, 161 –164.
Barthels, N., van der Lee, F. M., Klap, J., Goddijn, O. J., Karimi, M., Puzio, P., Grundler, F. M., Ohl,
S. A., Lindsey, K., Robertson, L., Robertson, W. M., van Montagu, M., Gheysen, G., and
Sijmons, P. C. (1997). Regulatory sequences of Arabidopsis drive reporter gene expression in
nematode feeding structures. Plant Cell 9, 2119– 2134.
Baulcombe, D. (2002). RNA silencing. Curr. Biol. 12, R82–R84.
Bennetzen, J., and Freeling, M. (1997). Review: the unified grass genome: synergy in synteny.
Genome Res. 7, 301–306.
Bonierbale, M. W., Plaisted, R. L., and Tanksley, S. D. (1988). RFLP maps based on a common set of
clones reveal modes of chromosomal evolution in potato and tomato. Genetics 120, 1095– 1103.
Brar, D. S., and Khush, G. S. (1997). Alien introgression in rice. Plant Mol. Biol. 35, 35–47.
Brar, D. S., and Khush, G. S. (2002). Transferring genes from wild species into rice. In “Quantitative
Genetics, Genomics and Plant Breeding” (M. S. Kang, Ed.), pp. 197–217. CAB International,
London.
Brown, P. O., and Botstein, D. (1999). Exploring the new world of the genome with DNA microarrays.
Nat. Genet. 21, 33–37.
Bruggemann, E., Handwerger, K., Essex, C., and Storz, G. (1996). Analysis of fast neutron-generated
mutants at the Arabidopsis thaliana HY4 locus. Plant J. 10, 755– 760.
Buell, R. (2002). Current status of the sequence of the rice genome and prospects for finishing the first
monocot genome. Plant Physiol. 130, 1585–1586.
Buschges, R., Holricher, K., Panstruga, R., Simons, G., Wolter, M., Frijters, A., van Daelen, R., van
der Lee, T., Diergaarde, P., Groenendijk, J., Topsch, S., Vos, P., Salamini, F., and Schulze-Lefert,
P. (1997). The barley Mlo gene: a novel control element of plant pathogen resistance. Cell 88,
695–705.
Chang, T. T., Pan, Y., Chu, Q., Peiris, R., and Loresto, G. C. (1991). Cytogenetic, electrophoretic, and
root studies of javanica rices. In “Rice Genetics II” (G. S. Khush, D. S. Brar, and B. Hardy, Eds.),
pp. 21– 32. International Rice Research Institute, Los Baños, Philippines.
Chang, H. S., Wu, C., Zeng, L., Dunn, M., Wang, G. L., Leung, H., Goff, S., Wang, X., Zhu, T., and
Leach, J. E. (2003). Detection of deleted genes in rice mutants using the Rice GeneChip genome
array, Page 100 in Abstracts of Plant and Animal Genomes XI. January 11 –15, 2003, San Diego,
California.
Charng, Y.-C., Pfitzner, A. J. P., Pfitzer, U. M., Charng-Chang, K.-F., Chen, C., Tu, J., and Kuo, T.-T.
(2000). Construction of an inducible transposon, INAc, to develop a gene tagging system in higher
plants. Mol. Breed. 6, 353 –367.
Chen, W., Provart, N., Glazebrook, J., Katagiri, F., Chang, H. S., Eulgem, T., Mauch, F., Luan, S.,
Zou, G., Whitnam., S., Budworth, P. R., Tao, Y., Xie, Y., Chen, X., Lam, S., Kreps, J., Harper, J.,
Si-Ammour, A., Mauch-Mani, B., Heinlein, M., Kobayashi, K., Hohn, T., Dangl, J., Wang, X.,
and Zhu, T. (2002). Expression profile matrix of Arabidopsis transcription factor genes suggests
their putative functions in response to environmental stresses. Plant Cell 14, 559– 574.
Cheng, X., Sardana, R., Kaplan, H., and Altosaar, I. (1998). Agrobacterium-transformed rice plants
expressing synthetic cry1A(b) and cry1A(c) genes are highly toxic to striped stemborer and
yellow stemborer. Proc. Natl Acad. Sci. USA 95, 2767–2772.
Chin, H. G., Choe, M. S., Lee, S. H., Park, S. H., Koo, J. C., Kim, N. Y., Lee, J. J., Oh, B. G., Yi, G. H.,
Kim, S. C., Choi, H. C., Cho, M. J., and Han, C. (1999). Molecular analysis of rice plants
harboring an Ac/Ds transposable element-mediated gene trapping system. Plant J. 19, 615–623.
Cho, R. J., Mindrinos, M., Richards, D. R., Sapolsky, R. J., Anderson, M., Drenkard, E., Dewdney, J.,
Beuber, T. L., Stemmers, M., Federspiel, N., Theologis, A., Yang, W.-H., Hubbell, E., Au, M.,
Chung, E. Y., Lashkari, D., Lemieux, B., Dean, C., Lipshutz, R. J., Ausubel, F. M., Davis, R. W.,
and Oefner, P. J. (1999). Genome-wide mapping with biallelic markers in Arabidopsis thaliana.
Nature Gen. 23, 203–207.
Choi, I.-R., Stenger, D. C., Morris, J., and French, R. (2000). A plant virus vector for systemic
expression of foreign genes in cereals. Plant J. 23, 547 –555.
Christou, P. (1997). Rice transformation: bombardment. Plant Mol. Biol. 35, 197 –203.
Chuang, C. F., and Meyerowitz, E. M. (2000). Specific and heritable genetic interference by double-
stranded RNA in Arabidopsis thaliana. Proc. Natl Acad. Sci. USA 97, 4985–4990.
Chung, B. C., Kim, J. K., Nahm, B. H., and Lee, C. H. (2000). In planta visual monitoring of green
fluorescent protein in transgenic rice plants. Mol. Cells 10, 411–414.
Colbert, T., Till, B. J., Tompa, R., Reynolds, S., Steine, M. N., Yeung, A. T., McCallum, C. M.,
Comai, L., and Henikoff, S. (2001). High-throughput screening for induced point mutations.
Plant Physiol. 126, 480 –484.
Crane, P. R., Friis, E. M., and Raunsgaard-Pedersen, K. (1995). The origin and early diversification of
angiosperms. Nature 374, 27–33.
Dai, S., Zheng, P., Marmey, P., Zhang, S., Tian, W., Chen, S., Beachy, R. N., and Fauquet, C. (2001).
Comparative analysis of transgenic rice plants obtained by Agrobacterium-mediated transform-
ation and particle bombardment. Mol. Breed. 7, 25–33.
DeRisi, J. L., Iyer, V. R., and Brown, P. O. (1997). Exploring the metabolic and genetic control of
gene expression on a genomic scale. Science 278, 680 –686.
Dernburg, A. F., and Karpen, G. H. (2002). A chromosome RNAissance. Cell 111, 159 –162.
Devos, K. M., and Gale, M. D. (2000). Genome relationships: the grass model in current research.
Plant Cell 12, 637 –646.
Drenkard, E., Richter, B. G., Rozen, S., Stutius, L. M., Angell, N. A., Mindrinos, M., Cho, R. J.,
Oefner, P. J., Davis, R. W., and Ausubel, F. M. (2000). A simple procedure for the analysis of
single nucleotide polymorphisms facilitates map-based cloning in Arabidopsis. Plant Physiol.
124, 1483–1492.
Enoki, H., Izawa, T., Kawahara, M., Komatsu, M., Koh, S., Kyozuka, J., and Shimamoto, K. (1999).
Ac as a tool for the functional genomics for rice. Plant J. 19, 605 –613.
Evenson, R. E., Herdt, R. W., and Hossain, M. (1996). “Rice research in Asia: progress and priorities”,
p. 418, CAB International, Wallingford, UK.
Farris, J. D., Haen, K. M., and Gill, B. S. (2000). Saturation mapping of a gene-rich recombinant hot
spot region in wheat. Genetics 154, 823– 835.
Fedoroff, N. V., Wessler, S., and Shure, M. (1983). Isolation of the transposable maize controlling
elements Ac and Ds. Cell 35, 235 –242.
Fedoroff, N. V., Furtek, D. B., and Nelson, O. E. (1984). Cloning of the bronze locus in maize by a
simple and generalizable procedure using the transposable element Activator (Ac). Proc. Natl
Acad. Sci. USA 81, 3825–3829.
Feng, Q., Zhang, Y., Hao, P., Wang, S., Fu, G., Huang, Y., Li, Y., Zhu, J., Liu, Y., Hu, X., Jia, P.,
Zhang, Y., Zhao, Q., Ying, K., Yu, S., Tang, Y., Weng, Q., Zhang, L., Lu, Y., Mu, J., Lu, Y.,
Zhang, L. S., Yu, Z., Fan, D., Liu, X., Lu, T., Li, C., Wu, Y., Sun, T., Lei, H., Li, T., Hu, H., Guan,
J., Wu, M., Zhang, R., Zhou, B., Chen, Z., Chen, L., Jin, Z., Wang, R., Yin, H., Cai, Z., Ren, S.,
Lv, G., Gu, W., Zhu, G., Tu, Y., Jia, J., Zhang, Y., Chen, J., Kang, H., Chen, X., Shao, C., Sun, Y.,
Hu, Q., Zhang, X., Zhang, W., Wang, L., Ding, C., Sheng, H., Gu, J., Chen, S., Ni, L., Zhu, F.,
Chen, W., Lan, L., Lai, Y., Cheng, Z., Gu, M., Jiang, J., Li, J., Hong, G., Xue, Y., and Han, B.
(2002). Sequence and analysis of rice chromosome 4. Nature 420, 316–320.
Feuillet, C., and Keller, B. (2002). Comparative genomics in the grass family: molecular
characterization of grass genome structure and evolution. Ann. Bot. 89, 3– 10.
Fire, A., Xu, S., Montgomery, M. K., Kostas, S. A., Driver, S. E., and Mello, C. C. (1998). Potent and
specific genetic interference by double-stranded RNA in Caenorhabditis elegans. Nature 391,
806–811.
Fischer, K. S., Barton, J., Khush, G. S., Leung, H., and Cantrell, R. (2000). Collaboration in rice.
Science 290, 279 –280.
Fodor, S. P. A., Read, J. L., Pirrung, M. C., Styryer, L., Lu, A. T., and Solas, D. (1991). Light-directed,
spatially addressable parallel chemical synthesis. Science 251, 767 –773.
Foote, T., Roberts, M., Kurata, N., Sasaki, T., and Moore, G. (1997). Detailed comparative mapping of
cereal chromosome regions corresponding to the Ph 1 locus in wheat. Genetics 147, 801– 807.
Fraser, A. G., Kamath, R. S., Zipperlen, Martinez-Campos, M., Sohrmann, M., and Ahringer, J.
(2000). Functional genomic analysis of C. elegans chromosome 1 by systematic RNA
interference. Nature 408, 325–330.
Freeling, M. (2001). Grasses as a single genetic system: reassessment. Plant Physiol. 15, 1191– 1197.
Fujisawa, Y., Kato, T., Ohki, S., Ishikawa, A., Kitano, H., Sasaki, T., Asahi, T., and Iwasaki, Y.
(1999). Suppression of the heterotrimetric G protein causes abnormal morphology, including
dwarfism in rice. Proc. Natl Acad. Sci. USA 96, 7575–7580.
Gaut, B. (2002). Evolutionary dynamics of grass genomes. New Physiol. 154, 15–28.
Ge, S., Sang, T., Lu, B. R., and Hong, D. U. (1999). Phylogeny of rice genomes with emphasis on
origins of allotetraploid species. Proc. Natl Acad. Sci. USA 96, 14400–14405.
Ge, S., San, T., Lu, B-R., and Hong, D.-Y. (2001). Phylogeny of the genus Oryza as revealed by
molecular approaches. In “Rice Genetics IV” (G. S. Khush, D. S. Brar, and B. Hardy, Eds.),
pp. 89–105. New Delhi Science Publishers Inc. and International Rice Research Institute,
Los Baños, Philippines.
Gill, K. S., Gill, B. S., Endo, T. R., and Boiko, E. V. (1996). Identification and high density mapping of
gene-rich regions in chromosome group 5 of wheat. Genetics 143, 1001–1012.
Goff, S. A., Ricke, D., Lan, T. H., Presting, G., Wang, R., Dunn, M., Glazebrook, J., Sessions, A.,
Oeller, P., Varma, H., Hadley, D., Hutchison, D., Martin, C., Katagiri, F., Lange, B. M.,
Moughamer, T., Xia, Y., Budworth, P., Zhong, J., Miguel, T., Paszkowski, U., Zhang, S., Colbert,
M., Sun, W., Chen, L., and Cooper, B. (2002). A draft sequence of the rice genome (Oryza sativa
L. ssp. japonica). Science 296, 92–100.
Gonczy, P., Echeverri, C., Oegema, K., Coulson, A., Jones, S. J. M., Copley, R. R., Duperon, J.,
Oegema, J., Brehm, M., Etienne, C., Hannak, E., Kirkham, M., Pichler, S., Flohrs, K., Goessen,
A., Leidel, S., Alleaume, A.-M., Martin, C., Ozlu, N., Bork, P., and Hyman, A. A. (2000). Nature
408, 331 –336.
Hanley, S., Edwards, D., Stevenson, D., Haines, S., Hegarty, M., Schuch, W., and Edwards, K. J.
(2000). Identification of transposon-tagged genes by the random sequencing of mutator-tagged
DNA fragments from Zea mays. Plant J. 23, 557–566.
Hannon, G. J. (2002). RNA interference. Nature 418, 244–251.
Hayashi, H., Czaja, I., Lubenow, H., Schell, J., and Walden, R. (1992). Activation of a plant gene by
T-DNA tagging: auxin-independent growth in vitro. Science 258, 1350–1353.
Hieter, P., and Boguski, M. (1997). Functional genomics: it’s all how you read it. Science 278,
601 –602.
Hiei, Y., Ohta, S., Komari, T., and Kumashiro, T. (1994). Efficient transformation of rice (Oryza
sativa L.) mediated by Agrobacterium and sequence analysis of the boundaries of the T-DNA.
Plant J. 6, 271– 282.
Hirochika, H., Otsuki, H., Yoshikawa, M., Otsuki, Y., Sugimoto, K., and Takeda, S. (1996).
Autonomous transposition of the tobacco retrotransposon Tto1 in rice. Plant Cell 8, 725–734.
Hirochika, H. (2001). Contribution of the Tos17 retrotransposon to rice functional genomics. Curr.
Opin. Plant Biol. 4, 118–122.
Hossain, M. (1996). Recent developments in the Asian rice economy: challenges for rice research. In
“Rice Research in Asia: Progress and Priorities” (R. E. Evenson, R. W. Herdt, and M. Hossain,
Eds.), pp. 17 –33. CAB International, London and International Rice Research Institute, Los
Baños, Philippines.
Hossain, M. (1998). Sustaining food security in Asia: economic, social, and political aspects. In
“Sustainability of Rice in the Global Food System” (N. G. Dowling, S. M. Greenfield, and K. S.
Fischer, Eds.), pp. 19 –43. Pacific Basin Center, USA, and International Rice Research Institute,
Los Baños, Philippines.
Hossain, M., Bennett, J., Datta, S., Leung, H., and Khush, G. S. (2000). Biotechnology research in rice
for Asia: priorities, focus and directions. In “Agricultural Biotechnology in Developing
Countries: Towards Optimizing the Benefits of the Poor” (M. Qaim, et al., Eds.), pp. 99–120.
Kluwer Academic Publishers, Netherlands.
Hughes, T. R., Mao, M., Jones, A. R, Burchard, J., Marton, M. J., Shannon, K. W., Lefkowitz,
S. M., Ziman, M., Schelter, J. M., Meyer, M. R., Kobayashi, S., Davis, C., Dai, H., He, Y. D.,
Stephaniants, S. B., Cavet, G., Walker, W. L., West, A., Coffey, E., Shoemaker, D. D.,
Stoughton, R., Blanchard, A. P., Friend, S. H., and Linsley, P. S. (2001). Expression profiling
using microarrays fabricated by an ink-jet oligonucleotide synthesizer. Nat. Biotechnol. 19,
342 –347.
Ito, T., and Meyerowitz, E. M. (2000). Overexpression of a gene encoding a cytochrome P450,
CYP78A9, induces large and seedless fruit in Arabidopsis. Plant Cell 12, 1541–1550.
Ito, T., Motohashi, R., Kuromori, T., Mizukado, S., Sakurai, T., Kanahara, H., Seki, M., and
Shinozaki, K. (2002). A new resource of locally transposed dissociation elements for screening
gene-knockout lines in silico on the Arabidopsis genome. Plant Physiol. 129, 1695– 1699.
Izawa, T., Miyazaki, C., Yamamoto, M., Terada, R., Iida, S., and Shimamoto, K. (1991). Introduction
and transposition of the maize transposable element Ac in rice (Oryza sativa L.). Mol. Gen. Genet.
227, 391–396.
Izawa, T., Ohnishi, T., Nakano, T., Ishida, N., Enoki, H., Hashimoto, H., Itoh, K., Terada, R., Wu, C.,
Miyazaki, C., Endo, T., Iida, S., and Shimamoto, K. (1997). Transposon tagging in rice. Plant
Mol. Biol. 35, 219–229.
Jaccoud, D., Peng, K., Feinstein, D., and Kilian, A. (2001). Diversity arrays: a solid state technology
for sequence information independent genotyping. Nucleic Acids Res. 29(4), e25.
Jaglo-Ottosen, K. R., Gilmour, S. J., Zarka, D. G., Schabenberger, O., and Thomashaw, M. F. (1998).
Arabidopsis CBF1 overexpression induces COR genes and enhances freezing tolerance. Science
280, 104–106.
Jansen, G., Hazendonk, E., Thijssen, K. L., and Plasterk, R. H. (1997). Reverse genetics by chemical
mutagenesis in Caenorhabditis elegans. Nature Gen. 17, 119–121.
Jefferson, R. A., Kavanagh, T. A., and Bevan, M. W. (1987). GUS fusions: b-glucuronidase as a
sensitive and versatile gene fusion marker in higher plants. EMBO J. 6, 3901–3907.
Jeon, J.-S., Lee, S., Jung, K.-H., Jun, S.-H., Jeong, D.-H., Lee, J., Kim, C., Jang, S., Yang, K., Nam,
J., An, K., Han, M.-J., Sung, R.-J., Choi, H.-S., Yu, J.-H., Choi, J.-H., Cho, S.-Y., Cha, S.-S.,
Kim, S.-I., and An, G. (2000). T-DNA insertional mutagenesis for functional genomics in rice.
Plant J. 22, 561–570.
Jeong, D.-H., An, S., Kang, H.-G., Moon, S., Han, J., Park, S., Lee, H. S., An, K., and An, G. (2002).
T-DNA insertional mutagenesis for activation tagging in rice. Plant Physiol. 130, 1636– 1644.
Jones, M. P., Dingkuhn, M., Aluko, G. K., and Semon, M. (1997). Interspecific Oryza sativa L. £ O.
glaberrima Steud. progenies in upland rice improvement. Euphytica 92, 237–246.
Kane, M. D., Jatkoe, T. A., Stumpf, C. R., Lu, J., Thomas, J. D., and Madore, S. J. (2000). Assessment
of the sensitivity and specificity of oligonucleotide (50 mer) microarrays. Nucleic Acids Res. 28,
4552–4557.
Kaplinsky, N. J., Braun, D. M., Penterman, J., Goff, S. A., and Freeling, M. (2002). Utility and
distribution of conserved non-coding sequences in the grasses. Proc. Natl Acad. Sci. USA 99,
6147–6151.
Kardailsky, I., Shukla, V. K., Ahn, J. H., Dagenais, N., Christensen, S. K., Nguyen, J. T., Chory, J.,
Harrison, M. J., and Weigel, D. (1999). Activation tagging of the floral inducer FT. Science 286,
1962–1965.
Karrer, E. E., Beachy, R. N., and Holt, C. A. (1998). Cloning of tobacco genes that elicit the
hypersensitive response. Plant Mol. Biol. 36, 681 –690.
Kasuga, M., Liu, Q., Miura, S., Yamaguchi-Shinozaki, K., and Shinozaki, K. (1999). Improving plant
drought, salt, and freezing tolerance by gene transfer of a single stress-inducible transcription
factor. Nat. Biotechnol. 17, 287–291.
Katagiri, S. (2003). Comparative genomics of rice chromosome 1: map construction and preliminary
genome sequencing of an indica variety, Kasalath, In “Abstracts of Rice Genome Workshop.
February 5 – 7, 2003”, p. 13. Rice Genome Research, Program, National Institute of
Agrobiological Resources/STAFF Institute, Tsukuba, Japan.
Katayama, T. (1997). Relationships between chromosome numbers and genomic constitutions in
genus Oryza. In “Science of the Rice Plant” (T. Matsuo, Y. Futsuhara, F. Kikuchi, and
H. Yamaguchi, Eds.), pp. 39–48. Food and Agriculture Policy Research Center, Tokyo.
Kellogg, E. A. (1998). Relationships of cereal crops and other grasses. Proc. Natl Acad. Sci. USA 95,
2005–2010.
Kellogg, E. A. (2001). Evolutionary history of the grasses. Plant Physiol. 125, 1198–1205.
Kawasaki, S., Borchert, C., Deyholos, M., Wang, H., Brazille, S., Kawai, K., Galbraith, D., and
Bohnert, H. (2000). Gene expression profiles during the initial phase of salt stress in rice. Plant
Cell 13, 889–905.
Khush, G. S. (1977). Disease and insect resistance in rice. Adv. Agron. 29, 265–341.
Khush, G. S. (2001). Green revolution: the way forward. Nat. Rev. Genet. 2, 815 –822.
Koller, A., Washburn, M. P., Lange, B. M., Andon, N. L., Deciu, C., Haynes, P. A., Hays, L., Schieltz,
D., Ulaszek, R., Wei, J., Wolters, D., and Yates, J. R. (2002). Proteomic survey of metabolic
pathways in rice. Proc. Natl Acad. Sci. USA 99, 11968–11974.
Koornneef, M., Dellaert, L. W., and van der Veen, J. H. (1982). EMS- and radiation-induced mutation
frequencies at individual loci in Arabidopsis thaliana (L.) Heynh. Mutat. Res. 93, 109–123.
Krysan, P. J., Young, J. C., Tax, F., and Sussman, M. R. (1996). Identification of transferred DNA
insertions within Arabidopsis genes involved in signal transduction and ion transport. Proc. Natl
Acad. Sci. USA 93, 8145–8150.
Krysan, P. J., Young, J. C., and Sussman, M. R. (1999). T-DNA as an insertional mutagen in
Arabidopsis. Plant Cell 11, 2283–2290.
Lee, H., Suh, S. S., Park, E., Cho, E., Ahn, J. H., Kim, S. G., Lee, J. S., Kwon, Y. M., and Lee, I.
(2000). The AGAMOUS-LIKE 20 MADS domain protein integrates floral inductive pathways in
Arabidopsis. Genes Dev. 14, 2366–2376.
Leung, H., Wu, C., Baraoidan, M. R., Bordeos, A., Ramos, M., Madamba, S., Cabauatan, P., Vera
Cruz, C. M., Portugal., A., Reves, G., Bruskiewich, R., McLaren, G., Gregorio, G., Bennett, J.,
Brar, D., Khush, G., Schnable, P., Wang, G. L., and Leach, J. E. (2001). Deletion mutants for
functional genomics: progress in phenotyping, sequence assignment, and database development.

In “Rice Genetics IV” (G. S. Khush, D. S. Brar, and B. Hardy, Eds.), pp. 239– 251. New Delhi
Science Publishers, Inc.
Li, W., and Gill, B. S. (2002). The colinearity of the Sh2/A1 orthologous region in rice, sorghum and
maize is interrupted and accompanied by genome expansion in the Triticeae. Genetics 160,
1153–1162.
Li, X., Song, Y., Century, K., Straight, S., Ronald, P., Dong, X., Lassner, M., and Zhang, Y. (2001). A
fast neutron deletion mutagenesis-based reverse genetics system for plants. Plant J. 27, 235–242.
Linares, O. F. (2002). African rice (Oryza glaberrima): history and future potential. Proc. Natl Acad.
Sci. USA 99, 16360–16365.
Liu, Y.-G., and Whittier, R. (1995). Thermal asymmetric interlaced PCR: automatable amplification
and sequencing of insert end fragments from P1 and YAC clones for chromosome walking.
Genomics 25, 674–681.
Liu, L. X., Spoerke, J. M, Mulligan, E. L., Chen, J., Reardon, B., Westlund, B., Sun, L., Abel, K.,
Armstrong, B., Hardiman, G., King, J., McCague, L., Basson, M., Clover, R., and Johnson, C. D.
(1999a). High-throughput isolation of Caenorhabditis elegans deletion mutants. Genome Res. 9,
859 –867.
Liu, D., Zhang, S., Fauquet, C., and Crawford, N. M. (1999b). The Arabidopsis transposon Tag1 is
active in rice, undergoing germinal transposition and restricted, late somatic excision. Mol. Gen.
Genet. 262, 413–420.
Lu, B. R. (1999). Taxonomy of the genus Oryza (Poaceae): historical perspective and current status.
Intl. Rice Res. Notes 24, 4–8.
Machida, C., Onouchi, H., Koizumi, J., Hamada, S., Semiarti, E., Torikai, S., and Machida, Y. (1997).
Characterization of the transposition pattern of the Ac element in Arabidopsis thaliana using
endonuclease I-SceI. Proc. Natl Acad. Sci. USA 94, 8675–8680.
Mao, L., Wood, T. C., Yu, Y., Budiman, M. A., Tomkins, J., Woo, S., Sasinowski, M., Presting, G.,
Frisch, D., Goff, S., Dean, R. A., and Wing, R. A. (2002). Rice transposable elements: a survey of
73,000 sequence-tagged-connectors. Genome Res. 10, 982 –990.
Martienssen, R. A. (1998). Functional genomics: probing plant gene function and expression with
transposons. Proc. Natl Acad. Sci. USA 95, 2021–2026.
Matsuo, T. (1997). Origin and distribution of cultivated rice. In “Science of the Rice Plant” (T. Matsuo,
Y. Futsuhara, F. Kikuchi, and H. Yamaguchi, Eds.), pp. 69–88. Food and Agriculture Policy
Research Center, Tokyo.
McCallum, C. M., Comai, L., Greene, E. A., and Henikoff, S. (2000a). Targeted screening for induced
mutations. Nat. Biotechnol. 18, 455–457.
McCallum, C. M., Comai, L., Greene, E. A., and Henikoff, S. (2000b). Targeting induced local lesions
in genomes (TILLING) for plant functional genomics. Plant Physiol. 123, 439–442.
Miyao, A., Murata, K., Tanaka, K., Onosato, K., Yamashita, Y., and Hirochika, H. (2001). Profile of
retrotransposon Tos17 insertions in rice genome sequence, OsCesA7 disrupted by insertion of the
retrotransposon Tos17, In “9th Conference of Plant and Animal Genome P5b-1”.
Monna, L. (2003). Single nucleotide polymorphisms (SNPs) in rice, In “Abstracts of Rice Genome
Workshop. February 5 –7, 2003.” Rice Genome Research Program, National Institute of
Agrobiological Resources/STAFF Institute, Tsukuba, Japan, p. 15.
Mori, M., Sugimoto, K., Satoh, K., Okabe, K., Toki, S., Ugaki, M., Hirochika, H., and Kikuchi, S.
(2000). Toward activation tagging in rice “6th International Congress of Plant Molecular
Biology”, p. S1-57.
Morishima, H. (2001). Evolution and domestication of rice. In “Rice Genetics IV” (G. S. Khush, D. S.
Brar, and B. Hardy, Eds.), pp. 63– 77. New Delhi Science Publishers Inc., and International Rice
Research Institute, Los Baños, Philippines.
Murai, N., Li, Z., Kawagoe, Y., and Hayashimamoto, A. (1991). Transposition of the maize activator
element in transgenic rice plants. Nucleic Acids Res. 19, 617–622.
Nasu, S., Suzuki, J., Ohta, R., Hasegawa, K., Yui, R., Kitazawa, N., Monna, L., and Minobe, Y.
(2002). Search for and analysis of single nucleotide polymorphisms (SNPs) in rice (Oryza sativa,
Oryza rufipogon) and establishment of SNP markers. DNA Research 9, 163–171.
Nakagawa, Y., Machida, C., Machida, Y., and Toriyama, K. (2000). Frequency and pattern of
transposition of the maize transposable element Ds in transgenic rice plants. Plant Cell Physiol.
41, 733 –742.
Normile, D., and Pennisi, E. (2002). Rice: boiled down to bare essentials. Science 296, 32 –36.
Nuwaysir, E. F., Huang, W., Albert, T. J., Singh, J., Nuwaysir, K., Pitas, A., Richmond, T., Gorski, T.,
Berg, J., Ballin, J., McCormick, M., Norton, J., Pollock, T., Sumwalt, T., Butcher, L., Porter,
D. A., Molla, M., Hall, C., Blattner, F., Sussman, M., Wallace, R., Cerrina, F., and Green, R.
(2002). Gene expression analysis using oligonucleotide arrays produced by maskless
photolithography. Genome Res. 12, 1749–1755.
Okubara, P. A., Anderson, P. A., Ochoa, O. E., and Michelmore, R. W. (1994). Mutants of downy
mildew resistance in Lactuca sativa (lettuce). Genetics 137, 867–874.
O’Toole, J. C., Toenniessen, G. H., Murashige, T., Harris, R. R., and Herdt, W. W. (2001). The
Rockefeller Foundation’s International Program on Rice Biotechnology. In “Rice Genetics IV”
(G. S. Khush, D. S. Brar, and B. Hardy, Eds.), pp. 39–59. New Delhi Science Publishers Inc., and
International Rice Research Institute, Los Baños, Philippines.
Page, D. R., and Grossniklaus, U. (2002). The art and design of genetic screen: Arabidopsis thaliana.
Nat. Rev. Genet. 3, 124 –135.
Parinov, S., Sevugan, M., De, Y., Yang, W. C., Kumaran, M., and Sundaresan, V. (1999). Analysis of
flanking sequences from dissociation insertion lines: A database for reverse genetics in
Arabidopsis. Plant Cell 11, 2263–2270.
Parrish, S., Fleenor, J., Xu, S., Mello, C., and Fire, A. (2000). Functional anatomy of a dsRNA trigger:
differential requirement for the two trigger strands in RNA interference. Mol. Cell 6, 1077– 1087.
Peterfy, M., Gyuris, T., and Takacs, L. (2000). Signal-exon trap: a novel method for the identification
of signal sequences from genomic DNA. Nucleic Acids Res. 28, E26.
Phogat, S., Burma, P. K., and Pental, D. (2000). A four-element based transposon system for allele
specific tagging in plants: theoretical considerations. J. Biosci. 25, 57– 63.
Plesch, G., Kamann, E., and Mueller-Roeber, B. (2000). Cloning of regulatory sequences mediating
guard-cell-specific gene expression. Gene 249, 83 –89.
Pritchard, J. K., and Rosenberg, NA. (1999). Use of unlinked genetic markers to detect population
stratification in association studies. Am. J. Hum. Genet. 65, 220 –228.
Pritchard, J. K., Stephens, M., Rosenberg, N. A., and Donnelly, P. (2000). Association mapping in
structured populations. Am. J. Hum. Genet. 67, 170 –181.
Qu, S., Jeon, J.-S., Richter, T., Leach, J. E., Ronald, P., and Wang, G.-L. (2003). Development of
chemically inducible Ac/Ds elements for gene mutagenesis in rice. In “Abstracts of Plant and
Animal Genome XI. January 11–15, 2003, San Diego, California, USA”. p. 151, p. 515,
Abstract P306.
Rafalski, A. (2002). Applications of single nucleotide polymorphisms in crop genetics. Curr. Opin.
Plant Biol. 5, 94–100.
Ramalingam, J., Vera Cruz, C. M., Kukreja, K., Chittoor, J. M., Wu, J. L., Lee, S. W., Baraoidan,
M. R., George, M. L., Cohen, M., Hulbert, S, Leach, J. E., and Leung, H. (2003). Candidate
resistance genes from rice, barley, and maize and their association with qualitative and
quantitative resistance in rice. Mol. Plant-Microbe Interact. 16, 14– 24.
Reardon, J. T., Liljestrand-Golden, C. A., Dusenbery, R. L., and Smith, P. D. (1987). Molecular
analysis of diepoxybutane-induced mutations at the rosy locus of Drosophila melanogaster.
Genetics 115, 323–331.
Remington, D. L., Thornsberry, J. M., Matsuoka, Y., Wilson, L. M., Whitt, S. R., Doebley, J.,
Kresovich, S, Goodman, M. M., and Buckler, E. S. (2001). Structure of linkage disequilibrium
and phenotypic associations in the maize genome. Proc. Natl Acad. Sci. USA 98, 11479–11484.
Rigoutsos, I., and Floratos, A. (1998). Combinatorial pattern discovery in biological sequences: the
TEIRESIAS algorithym. Bioinformatics 14, 55– 67.
Robert, M. A., Reader, S. M., Dalgliesh, C., Miller, T. E., Foote, T. N., Fish, L. J., Snape, J. W., and
Moore, G. (1999). Induction and characterization of Ph1 wheat mutants. Genetics 153,
1909–1918.
Salekdeh, Gh. H., Siopongco, J., Wade, L. J., Ghareyazie, B., and Bennett, J. (2002a). A proteomic
approach to analyzing drought- and salt-responsiveness in rice. Field Crop Res. 76, 199–219.
Salekdeh, Gh. H., Siopongco, J., Wade, L. J., Ghareyazie, B., and Bennett, J. (2002b). Proteomics
analysis of rice leaves during drought stress and recovery. Proteomics 2, 1131–1145.
Sasaki, A., Ashikari, M., Ueguchi-tanaka, M., Itoh, H., Nishimura, A., Swapan, D., Ishiyama, K.,
Saito, T., Kobayashi, M., Khush, G. S., Kitano, H., and Matsuoka, M. (2002). Green revolution: a
mutant gibberellin-synthesis gene in rice. Nature 416, 701 –702.
Sasaki, T., Matsumoto, T., Yamamoto, K., Sakata, K., Baba, T., Katayose, Y., Wu, J., Niimura, Y.,
Cheng, Z., Nagamura, Y., Antonio, B. A., Kanamori, H., Hosokawa, S., Masukawa, M., Arikawa,
K., Chiden, Y., Hayashi, M., Okamoto, M., Ando, T., Aoki, H., Arita, K., Hamada, M., Harada,
C., Hijishita, S., Honda, M., Ichikawa, Y., Idonuma, A., Iijima, M., Ikeda, M., Ikeno, M., Ito, S.,
Ito, T., Ito, Y., Ito, Y., Iwabuchi, A., Kamiya, K., Karasawa, W., Katagiri, S., Kikuta, A.,
Kobayashi, N., Kono, I., Machita, K., Maehara, T., Mizuno, H., Mizubayashi, T., Mukai, Y.,
Nagasaki, H., Nakashima, M., Nakama, Y., Nakamichi, Y., Nakamura, M., Namiki, N., Negishi,
M., Ohta, I., Ono, N., Saji, S., Sakai, K., Shibata, M., Shimokawa, T., Shomura, A., Song, J.,
Takazaki, Y., Terasawa, K., Tsuji, K., Waki, K., Yamagata, H., Yamane, H., Yoshiki, S.,
Yoshihara, R., Yukawa, K., Zhong, H., Iwama, H., Endo, T., Ito, H., Hahn, J. H., Kim, H-I., Eun,
M-Y., Yano, M., Jiang, J., and Gojobori, T. (2002). The genome sequence and structure of rice
chromosome 1. Nature 420, 312– 316.
Sato, Y., Sentoku, N., Miura, Y., Hirochika, H., Kitano, H., and Matsuoka, M. (1999). Loss-of-
function mutations in the rice homeobox gene OSH15 affect the architecture of internodes
resulting in dwarf plants. EMBO J. 18, 992– 1002.
Schena, M., and Davis, R. W. (1999). Parallel analysis with biological chips. In “PCR Applications:
Protocols for Functional Genomics” (M. A. Innis, D. H. Gelfand, and J. J. Sninsky, Eds.),
Academic Press.
Schmitz, G., and Theres, K. (1994). A self-stabilizing Ac derivative and its potential for transposon
tagging. Plant J. 6, 781 –786.
Scholthof, H. B., Scholthof, K.-B. G., and Jackson, A. O. (1996). Plant virus gene vectors for transient
expression of foreign proteins in plants. Annu. Rev. Phytopathol. 34, 299– 323.
Shimamoto, K., Miyazaki, C., Hashimoto, H., Izawa, T., Itoh, K., Terada, R., Inagaki, Y., and Iida, S.
(1993). Trans-activation and stable integration of the maize transposable element Ds
cotransfected with the Ac transposase gene in transgenic rice plants. Mol. Gen. Genet. 239,
354 –360.
Shimamoto, K., and Kyozuka, J. (2002). Rice as a model for comparative genomics in plants. Annu.
Rev. Plant Biol. 53, 399 –419.
Shimamoto, K., Moritoh, S., and Miki Daisuke (2002). RNAi as a complementary method for
functional genomics of rice. In “Abstracts of International workshop on towards building a
global rice gene machine, November 11–12, 2002”, p. 27, CSIRO Plant Industry, Canberra,
Australia.
Shirley, B. S., Hanley, S., and Goodman, H. M. (1992). Effects of ionizing radiation on a plant
genome: analysis of two Arabidopsis transparent testa mutations. Plant Cell 4, 333 –347.
Song, W. H., Wang, G. L., Chen., L. L., Kim, H. S., Pi, L. Y., Holsten, T., Gardner, J., Wang, B., Zhai,
W. X., Zhu, L. H., Fauquet, C., and Ronald, P. (1995). A receptor kinase-like protein encoded by
the rice disease resistance gene, Xa21. Science 270, 1804–1806.
Spielmeyer, W., Ellis, M. H., and Chandler, P. M. (2002). Semidwarf (sd-1), “green revolution” rice,
contains a defective gibberellin 20-oxidase gene. Proc. Natl Acad. Sci., USA 99, 9043–9048.
Springer, P. S. (2000). Gene traps: tools for plant development and genomics. Plant Cell 12,
1007–1020.
Stockinger, E. J., Gilmour, S. J., and Thomashaw, M. F. (1997). Arabidopsis thaliana CBF1 encodes
an AP2 domain-containing transcriptional activator that binds the C-repeat/DRE, a cis-acting
DNA regulatory element that stimulates transcription in response to low temperature and water
deficit. Proc. Natl Acad. Sci. USA 94, 1035–1040.
Sugimoto, K., Otsuki, Y., Saji, S., and Hirochika, H. (1994). Transposition of the maize Ds element
from a viral vector to the rice genome. Plant J. 5, 863– 871.
Sundaresan, V. (1996). Horizontal spread of transposon mutagenesis: new uses of old elements.
Trends Plant Sci. 1, 184–191.
Takano, M., Kanegae, H., Shinomura, T., Miyao, A., Hirochika, H., and Furuya, M. (2001). Isolation
and characterization of rice phytochrome A mutants. Plant Cell 13, 521–534.
Takeda, S., Sugimoto, K., Otsuki, H., and Hirochika, H. (1998). Transcriptional activation of the
tobacco retrotransposon Tto1 by wounding and methyl jasmonate. Plant Mol. Biol. 36,
365–376.
Tanksley, S. D., and McCouch, S. R. (1997). Seed banks and molecular maps: unlocking genetic
potential from the wild. Science 277, 1063–1066.
Terada, R., Urawa, H., Inagaki, Y., Tsugane, K., and Iida, S. (2002). Efficient gene targeting by
homologous recombination in rice. Nat. Biotechnol. 20, 1030– 1034.
Thornsberry, J. M., Goodman, M. M., Doebley, J., Kresovich, S., Nielsen, D., and Buckler, E. S. IV
(2001). Dwarf8 polymorphisms associate with variation in flowering time. Nat. Genet. 28,
286–289.
The Arabidopsis Genome Initiative, (2000). Analysis of the genome sequence of the flowering plant
Arabidopsis thaliana. Nature 408, 796– 815.
Thomashaw, M. F. (1999). Plant cold acclimation: freezing tolerance genes and regulatory
mechanisms. Annu. Rev. Plant Physiol. Plant Mol. Biol. 50, 571–599.
Topping, J. F., and Lindsey, K. (1997). Promoter trap markers differentiate structural and positional
components of polar development in Arabidopsis. Plant Cell 9, 1713– 1725.
Triglia, T., Peterson, M. G., and Kemp, D. J. (1988). A procedure for in vitro amplification of DNA
segments that lie outside the boundaries of known sequences. Nucleic Acids Res. 16, 81–86.
Vigouroux, Y., McMullen, M., Hittinger, C. T., Houchins, K., Schulz, L., Kresovich, S., Matsuoka, Y.,
and Doebley, J. (2002). Identifying genes of agronomic importance in maize by screening
microsatellites for evidence of selection during domestication. Proc. Natl Acad. Sci. USA 99,
9650–9655.
Virmani, S. S., and Ilyas-Ahmed, M. (2001). Environment-sensitive genic male sterility (EGMS) in
crops. Adv. Agron. 72, 139– 195.
Walbot, V. (2000). Saturation mutagenesis using maize transposons. Curr. Opin. Plant Biol. 3,
103–107.
Watanabe, Y. (1997). Phylogeny and geographical distribution of genus Oryza. In “Science of the
Rice Plant” (T. Matsuo, Y. Futsuhara, F. Kikuchi, and H. Yamaguchi, Eds.), pp. 29–39. Food and
Agriculture Policy Research Center, Tokyo.
Waterhouse, P. M., Wang, M. B., and Lough, T. (2001). Gene silencing as an adaptive defence against
viruses. Nature 411, 834–842.
Weigel, D., Ahn, J. H., Blazquez, M. A., Borevitz, J. O., Christensen, S. K., Fankhauser, C., Ferrandiz,
C., Kardailsky, I., Malancharuvil, E. J., Neff, M. M., Nguyen, J. T., Sato, S., Wang, Z. Y., Xia, Y.,
Dixon, R. A., Harrison, M. J., Lamb, C. J., Yanofsky, M. F., and Chory, J. (2000). Activation
tagging in Arabidopsis. Plant Physiol. 122, 1003–1013.
Wesley, S. V., Helliwell, C. A., Smith, N. A., Wang, M. B., Rouse, D. T., Liu, Q., Gooding, P. S.,
Singh, S. P., Abbott, D., Stoutjesdijk, P. A., Robinson, S. P., Gleave, A. P., Green, A. G., and
Waterhouse, P. M. (2001). Construct design for efficient, effective and high-throughput gene
silencing in plants. Plant J. 27, 581–590.
Westlund, B., Parry, D., Clover, R., Basson, M., and Johnson, C. D. (1999). Reverse genetic analysis
of Caenorhabditis elegans presenilins reveals redundant but unequal roles for se-12 and hop-1 in
Notch-pathway signaling. Proc. Natl Acad. Sci. USA 86, 2497–2502.
White, P. T. (1994). Rice: the essential harvest. National Geographic, 48–79 (May issue).
Whitt, S. R., Wilson, L. M., Tenaillon, M. I., Gaut, B. S., and Buckler, E. S. (2002). Genetic diversity
and selection in maize starch pathway. Proc. Natl Acad. Sci. USA 99, 12959–12962.
Williams, N. (1996). Genome projects: yeast genome sequence ferments new research. Science 272,
481.
Wisman, E., Cardon, G. H., Fransz, P., and Saedler, H. (1998). The behaviour of the autonomous
maize transposable element En/Spm in Arabidopsis thaliana allows efficient mutagenesis. Plant
Mol. Biol. 37, 989–999.
Wolfe, K. H., Gouy, M., Yang, Y. W., Sharp, P. M., and Li, W. H. (1989). Date of the monocot-dicot
divergence estimated from chloroplast DNA sequence data. Proc. Natl Acad. Sci. USA 86,
6201–6205.
Wong, G. K.-S., Wang, J., Tao, L., Tan, J., Zhang, J., Passey, D. A., and Yu, J. (2002). Compositional
gradients in Gramineae genes. Genome Res. 12, 851–856.
Yandell, M. D., Edgar, L. G., and Wood, W. B. (1994). Trimethylpsoralen induces small deletion
mutations in Caenorhabditis elegans. Proc. Natl Acad. Sci. USA 91, 1381–1385.
Yano, M., Katayose, Y., Ashikari, M., Yamanouchi, U., Monna, L., Fuse, T., Baba, T., Yamamoto, K.,
Umehara, Y., Nagamura, Y., and Sasaki, T. (2000). Hd1, a major photoperiod sensitivity
quantitative trait locus in rice, is closely related to the Arabidopsis flowering time gene
CONSTANS. Plant Cell 12, 2473– 2483.
Yin, Z., and Wang, G.-L. (2000). Evidence of multiple complex patterns of T-DNA integration into
rice genome. Theor. Appl. Genet. 100, 461–470.
Yu, J., Hu, S., Wang, J., Wong, G. K., Li, S., Liu, B., Deng, Y., Dai, L., Zhou, Y., Zhang, X., Cao, M.,
Liu, J., Sun, J., Tang, J., Chen, Y., Huang, X., Lin, W., Ye, C., Tong, W., Cong, L., Geng, J., Han,
Y., Li, L., Li, W., Hu, G., Huang, X., Li, W., Li, J., Liu, Z., Li, L., Liu, J., Qi, Q., Liu, J., Li, L., Li,
T., Wang, X., Lu, H., Wu, T., Zhu, M., Ni, P., Han, H., Dong, W., Ren, X., Feng, X., Cui, P., Li,
X., Wang, H., Xu, X., Zhai, W., Xu, Z., Zhang, J., He, S., Zhang, J., Xu, J., Zhang, K., Zheng, X.,
Dong, J., Zeng, W., Tao, L., Ye, J., Tan, J., Ren, X., Chen, X., He, J., Liu, D., Tian, W., Tian, C.,
Xia, H., Bao, Q., Li, G., Gao, H., Cao, T., Wang, J., Zhao, W., Li, P., Chen, W., Wang, X., Zhang,
Y., Hu, J., Wang, J., Liu, S., Yang, J., Zhang, G., Xiong, Y., Li, Z., Mao, L., Zhou, C., Zhu, Z.,
Chen, R., Hao, B., Zheng, W., Chen, S., Guo, W., Li, G., Liu, S., Tao, M., Wang, J., Zhu, L.,
Yuan, L., and Yang, H. (2002). A draft sequence of the rice genome (Oryza sativa L. ssp. indica).
Science 296, 79–92.
Zhu, T., Budworth, P., Chen, W., Provart, N., Chang, H.-S., Guimil, S., Su, W., Estes, B., Zou, G., and
Wang, X. (2003). Transcriptional control of nutrient partitioning during rice grain filling. Plant
Biotechnol. 1, 59 –70.
Zivy, M., and de Vienne, D. (2000). Proteomics: a link between genomics, genetics and physiology.
Plant Mol. Biol. 44, 575–580.
This page is intentionally left blank
THE NATURE, PROPERTIES AND
MANAGEMENT OF VOLCANIC SOILS
R.A. Dahlgren,1 M. Saigusa2 and F.C. Ugolini3
1
Land, Air and Water Resources, University of California, Davis,
California 95616, USA
2
Experimental Farm of Tohoku University, Kawatabi, Naruko,
Tamatsukuri, Miyagi, 989-6711 Japan
3
Dipartimento di Scienza del Suolo e Nutrizione della Pianta,
Università Degli Studi, Piazzale delle Cascine 15, 50144, Firenze, Italy
I. Introduction
II. Volcanic Soil Distribution
III. Soil Genesis in Volcanic Materials
A. Allophanic and Nonallophanic Andisols
B. General Trends in Soil Development on Volcanic Materials
IV. Soil Classification
A. World Reference Base for Soil Resources—Andosols
B. Soil Taxonomy—Andisols
V. Chemical Weathering
VI. Mineralogical Characteristics
A. Common Colloidal Constituents of Volcanic Soils
B. Formation and Transformation of Colloids in Volcanic Soils
VII. Selected Chemical Characteristics of Volcanic Soils
A. Organic Matter Accumulation
B. Aluminum Dynamics
VIII. Productivity and Management of Volcanic Soils
A. Charge Characteristics and Chemical Fertility
B. Physical Properties and Fertility
C. Soil Management and Conservation of Volcanic Soils
References
Soils formed in volcanic ejecta have many distinctive physical, chemical,

and mineralogical properties that are rarely found in soils derived from
other parent materials. These distinctive properties are largely attributable to
the formation of noncrystalline materials (e.g., allophane, imogolite,
ferrihydrite) containing variable charge surfaces, and the accumulation of
organic matter. Formation of noncrystalline materials is directly related to
the properties of volcanic ejecta as a parent material, namely the rapid
weathering of glassy particles. The composition of the colloidal fraction
forms a continuum between pure Al – humus complexes and pure
allophane/imogolite, depending on the pH and organic matter characteristics
113
0065-2113/03 $35.00
114 R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI
of the weathering environment. For soil management purposes, volcanic soils

are often divided into two groups based on the colloidal composition of the
surface horizons: allophanic soils dominated by allophane and imogolite, and
nonallophanic soils dominated by Al –humus complexes and 2:1 layer
silicates. Volcanic soils exhibit a wide range of agricultural productivity,
depending on the degree or intensity of pedogenic development and the
colloidal composition of the rooting zone. The different charge character-
istics of allophanic and nonallophanic soils is the most important factor
regulating chemical fertility attributes. Phosphorus fixation, strong acidity,
and aluminum toxicity are the primary chemical limitations to agricultural
productivity. Volcanic soils generally have high physical fertility (tilth) and
mature soils are relatively resilient to erosion and compaction. To maximize
the productivity of volcanic soils, proper management based on an
understanding of the unique physical, chemical, and mineralogical properties
of these soils must be practiced. q 2004 Academic Press.
I. INTRODUCTION
Volcanoes are revered and feared for their awesome and devastating eruptions
that obliterate terrestrial ecosystems and often cause tremendous casualties
to humans and wildlife (Fig. 1). There have been about 1500 active volcanoes
during the last 10,000 years with approximately 60 eruptions each year (Simkin
and Siebert, 1994; Global Volcanism Program, 2003). In terms of human
casualties, the most catastrophic destruction is caused by pyroclastic flows
(e.g., 1902 Mont Pelée, Martinique 29,025 casualties) and mudflows (e.g., 1985
Nevado del Ruiz, Colombia 25,000 casualties), but tsunamis (e.g., 1883
Krakatau, Indonesia 36,417 casualties), starvation (e.g., 1815 Tambora,
Indonesia 92,000 casualties), and carbon dioxide degassing of volcanic lakes
(e.g., 1986 Lake Nyos, Cameroon 1700 casualties) have also resulted in the loss
of many human lives (Blong, 1984). Yet from these ashes of devastation arise
some of the most productive soils in the world with the capacity to sustain high
human population densities. Volcanism plays an important beneficial role in
sustaining the productivity of terrestrial ecosystems through soil rejuvenation.
Volcanism is probably the most obvious mechanism of recycling large amounts
of geological material and gases (e.g., CO2, SO2) on the planet Earth. Further,
deposition of volcanic ejecta is not limited to the immediate vicinity of
volcanoes; volcanic ash may be transported to great distances once injected into
the atmospheric circulation. Because volcanism is responsible for bringing new
material to the earth’s surface, volcanism is counteracting the effects of physical
and chemical erosion. By providing a source of easily weatherable materials, this
natural phenomenon represents an effective sink for carbon dioxide through
NATURE, PROPERTIES, MANAGEMENT OF VOLCANIC SOILS 115
Figure 1 The 1980 eruptions of Mt. St. Helens, Washington, created a disturbance gradient that
ranged from complete obliteration of the landscape near the mountain to a mere dusting of volcanic
ash hundreds of kilometers from the vent. These eruptions rejuvenate soil-forming processes and
renew the long-term fertility status of terrestrial ecosystems.
carbonic acid weathering ðCO2 ðgÞ þ H2 O ¼ Hþ þ HCO2 3 Þ. Soils formed in

volcanic deposits also contain the largest accumulations of organic carbon among
the mineral soil orders (Eswaran et al., 1993). Thus, volcanism plays an
important role in the global carbon cycle, representing a primary source and sink
for carbon.
Soils formed in volcanic deposits have many distinctive properties that are
rarely found in soils derived from other parent materials. Some unique properties
of volcanic soils include variable charge, high water retention, high phosphate
retention, low bulk density, high friability, highly stable soil aggregates, and
excellent tilth (Shoji et al., 1993a). These distinctive properties are largely due to
the formation of noncrystalline materials (e.g., active Al and Fe – allophane,
imogolite, ferrihydrite, Al/Fe –humus complexes) and the accumulation of organic
carbon, the dominant pedogenic processes occurring in volcanic soils. Formation
of noncrystalline materials is directly related to the properties of volcanic ejecta as
a parent material, namely the rapid weathering of the glassy particles.
Volcanic soils exhibit a wide range of agricultural productivity, depending
primarily on the degree or intensity of pedogenic development. The high potential
of volcanic soils for agricultural production is illustrated by the fact that many of
the most productive agricultural regions of the world are located near active or
dormant volcanoes (Shoji et al., 1993a). In volcanic regions such as Indonesia, the
most densely populated areas are found near volcanoes. The intermittent additions
of volcanic ash renew the long-term fertility status of terrestrial ecosystems by
providing a source of nutrients from the rapid weathering of volcanic deposits.
However, as soils become progressively more weathered, negative attributes such
as P fixation, low exchangeable base concentrations, and Al toxicity may severely
limit agricultural production. To maximize the productivity of volcanic soils,
proper management based on an understanding of the unique physical, chemical,
and mineralogical properties of these soils must be practiced.
The purpose of this chapter is to present a brief overview of soils derived from
volcanic ejecta in terms of their distribution, genesis, classification, and their
mineralogical, chemical, physical, and agronomic characteristics. In this context,
volcanic ejecta include ash, tephra, cinders, lahars, lapilli, tuff, pumice, volcanic
bombs, and lava. These materials may also be reworked and mixed with materials
from other geologic sources (e.g., loess, alluvium, and glacial drift). Research on
volcanic soils would fill volumes and several compilations on volcanic soils have
been previously published (e.g.,Wada and Harward, 1974; Ugolini and Zasoski,
1979; Theng, 1980; Wada, 1980, 1985, 1986; Yoshinaga, 1983; Mizota and van
Reeuwijk, 1989; Shoji et al. 1993a; Kimble et al., 2000; Ugolini and Dahlgren,
2002). We focus this review on recent developments and concepts affecting their
properties and management as a soil resource.
II. VOLCANIC SOIL DISTRIBUTION
The distribution of soils derived from volcanic materials closely parallels the
global distribution of active and recently active volcanoes (Fig. 2). Volcanism is
associated with interactions along tectonic plate boundaries (compressional and
extensional margins) and with “hot spots” that occur within tectonic plates. The
most extensive chain of volcanoes is associated with subduction of the Pacific
oceanic plate with continental masses along its boundary. This circum-Pacific
belt, known as the “Ring of Fire”, contains about 75% of the world’s approxi-
mately 520 currently active volcanoes (Harris, 1988). Other prominent chains of
volcanoes associated with plate boundaries include subduction zones in the
Mediterranean Sea, Indonesia (Java Trench), and the Caribbean, rift zones in
northeast Africa (East Africa rift system), and spreading centers in Iceland and
the Azores (Mid-Atlantic Ridge). Hot spot volcanoes occur where fractures in the
continental plate or hot plumes of magma rise close to the earth’s surface
resulting in magma ejection at the earth’s surface. About 5% of the world’s active
volcanoes are associated with hot spots, the most famous including the Hawaiian
Islands –Emperor Seamounts chain and Galapagos Islands on the Pacific Plate
and Yellowstone Park on the North American Plate.
Figure 2 Global distribution of Andisols (Courtesy of USDA-NRCS, Soil Survey Division, World Soil Resources, 1998). Andisols cover about 120 million
hectares or nearly 1% of the world’s land surface.
While it is difficult to establish the area influenced by volcanic ejecta, soils

with andic soil properties that are characterized as Andosols or Andisols cover
from 110 to 124 million hectares, about 0.84% of the world’s land surface
(Leamy, 1984; FAO, ISRIC and ISSS, 1998; Kimble et al., 2000). In addition,
large land areas receive periodic additions of volcanic materials that influence
soil properties, but are not sufficient to meet the minimum requirements for
Andosols or Andisols. For example, the cataclysmic eruption of Long Valley
volcano in eastern California about 730,000 years ago ejected 600 km3 of
pyroclastic material. These materials inundated 1500 km2 of central California
and southwestern Nevada with several meters of pyroclastic materials, and
deposited recognizable ash layers as far away as central Nebraska, a distance of
over 1700 km (Harris, 1988). Volcanic soils occur across a wide range of
temperature regimes: tropical (49%) . boreal (28%) . temperate (23%); and
soil moisture regimes: udic (63%) . ustic (30%) . aridic (4%) . xeric (3%)
(udic moisture regime includes perudic and aquic conditions — Wilding, 2000).
III. SOIL GENESIS IN VOLCANIC MATERIALS
Formation of noncrystalline materials (active Al and Fe compounds) and

accumulation of organic matter are the dominant pedogenic processes
occurring in most soils formed in volcanic materials (Shoji et al., 1993a).
This combination of processes, occurring preferentially in soils formed in
volcanic materials, is termed “andosolization” (Duchaufour, 1977; Ugolini
et al., 1988). Succinctly, andosolization is the darkening of the soil due to an
accumulation of stable humic substances for considerable depth occurring in
subacid conditions. Andosolization is a case of melanization in which Al3þ is
the dominant cation; the other case of melanization occurs in the presence of
Ca2þ and produces mollic epipedons. This relatively thick A horizon, that
may or may not contain poorly crystalline secondary aluminosilicate material,
is underlain by a Bw horizon. The colloidal composition of B horizons is
dominated by allophane/imogolite and other noncrystalline materials includ-
ing ferrihydrite (Childs et al., 1991; Shoji et al., 1993a). These soils lack
distinct eluvial horizons, and illuvial horizons containing enrichment of clay
or humus. Soil solution studies in volcanic deposits documented that
andosolization is characterized by accumulation of Fe, Al, and dissolved
organic carbon in A horizons with little translocation of these components to
B horizons (Ugolini et al., 1988; Dahlgren et al., 1991). Thus, noncrystalline
materials (e.g., allophane, imogolite, and ferrihydrite) in B horizons have
formed in situ, rather than by translocation. Formation of noncrystalline
materials is promoted by the carbonic acid weathering regime and hindered
by the organic acid weathering regime. In areas near intermittently active
Figure 3 A polygenetic soil profile in the vicinity of Mt. St. Helens consisting of four distinct
ash deposits. Burial of A horizons contributes to organic matter preservation in volcanic soils
experiencing intermittent ash deposition.
volcanoes, it is common to find polygenetic profiles resulting from soil burial

by successive depositional layers (Fig. 3).
A. ALLOPHANIC AND NONALLOPHANIC ANDISOLS 1
Under humid weathering conditions, the composition of the colloidal fraction

forms a continuum between pure Al –humus complexes and pure allophane/
imogolite, depending on the pH and organic matter characteristics of
1
The classical expression of soils developed in volcanic materials is termed Andisols in Soil
Taxonomy and Andosols in the WRB classification system. For convenience, we will use the term
Andisols to generically describe volcanic soils meeting Andisol and/or Andosol criteria in the text.
Andosols will only be used when specifically referring to volcanic soils meeting WRB criteria.
the weathering environment (Mizota and van Reeuwijk, 1989). For soil
management purposes, Andisols are often divided into two groups based on
the mineralogical composition of the surficial horizons (A horizons): allophanic
Andisols dominated by allophane and imogolite, and nonallophanic Andisols
dominated by Al –humus complexes and 2:1 layer silicates (Shoji and Ono, 1978;
Shoji, 1985; Shoji et al., 1985). In Japan, allophanic Andisols are mainly
distributed in areas having thick deposition of Holocene and/or late-Pleistocene
tephras, while nonallophanic Andisols are found in areas with older tephra
deposits. Older tephra deposits are generally more acidified and have been
exposed to deposition of exogenous materials, such as loess from China
dominated by 2:1 clay minerals (Inoue and Naruse, 1987; Mizota and Inoue,
1988; Mizota et al., 1990; Bautista-Tulin and Inoue, 1997; Saigusa and
Matsuyama, 1998). Nonallophanic Andisols represent about 30% of all Andisols
in Japan and are distributed throughout the world (e.g., Japan, USA, Indonesia,
Spain, Italy, Portugal, Chile, New Zealand, West Samoa, Taiwan) (Shoji et al.,
1985, 1987; Leamy et al., 1988; Madeira et al., 1994; Johnson-Maynard et al.,
1997; Chen et al., 1999).
The availability of Al3þ appears to be the critical factor regulating the
formation of nonallophanic versus allophanic Andisols. Nonallophanic
Andisols form preferentially in pedogenic environments that are rich in
organic matter and have pH values of 5 or less (Fig. 4; Shoji and Fujiwara,
1984). Base-poor volcanic deposits (e.g., rhyolitic, dacitic, or andesitic) having
noncolored volcanic glass and precipitation greater than about 1000 mm are
two factors that contribute to soil acidification. At pH values less than 5,
organic acids are the dominant proton donor lowering pH and aqueous Al3þ
activities through formation of Al– humus complexes. Aqueous Al3þ is also
incorporated into the interlayer of 2:1 layer silicates when they are present
(Dahlgren and Ugolini, 1989b). Under these conditions, humus and 2:1 layer
silicates effectively compete for dissolved Al, leaving little Al available for
co-precipitation with silica to form aluminosilicate materials, such as
allophane/imogolite. Preferential incorporation of Al into Al-humus complexes
and hydroxy-Al interlayers of 2:1 layer silicates has been termed the anti-
allophanic effect (Shoji et al., 1993a).
Allophanic Andisols dominated by allophane/imogolite form preferentially in
weathering environments with pH values in the range of 5 –7 and a low content of
complexing organic compounds (Fig. 4; Ugolini and Dahlgren, 1991). As
revealed by Shoji et al. (1982) and Shoji and Fujiwara (1984) allophanic Andisols
are favored in basic parent materials (e.g., andesitic basalt, basalt) having colored
volcanic glass, and climates having less than 1000 mm of precipitation. These
conditions favor higher pH values (pH . 5), which promotes formation
of Al-polymers relative to Al –humus complexes (Jackson, 1963a, b). The
Al-polymers are able to react with silica and form allophane/imogolite.
While Al– humus complexes also form in A horizons of allophanic Andisols,
Figure 4 Examples of Al–humus complex and allophane/imogolite distribution in allophanic and nonallo-
phanic Andisols. Allophane/imogolite is most prominent at soil pH values greater than 5. Al –humus complexes
and allophane/imogolite show an inverse relationship in nonallophanic Andisols. [Nonallophanic Andisol ¼
Mukaiyama (Shoji et al., 1993a); Allophanic Andisol ¼ Sirrah (Takahashi et al., 1993)].
there is an excess of Al relative to the complexing capacity of the humic sub-

stances. Parfitt and Saigusa (1985) found that when (Alp þ Fep)/Cp (p ¼
pyrophosphate extractable) was greater than about 0.1, allophane/imogolite was
present. This implies that once the complexation capacity of humic substances is
saturated, the excess Al can react with Si to form allophanic materials.
B. GENERAL TRENDS SOIL DEVELOPMENT

IN ON VOLCANIC
MATERIALS
Soils displaying a wide range of characteristics form in volcanic deposits

(Table I). Climate, in conjunction with its influence on vegetation, and time of
exposure to weathering are the two primary factors regulating soil developmental
pathways in volcanic materials. Andisols generally form rapidly in humid cli-
mates and alter to other soil orders as soil age and degree of weathering increase.
Table I
Examples of Soil Genesis Pathways in Contrasting Climates Extrapolated from the Literature
Climatic regime Literature source
Warm/Arid and Semi-arid

Arid (400 mm): Entisols Dubroeucq et al., 1998
Semi-arid (700 mm): Ustivitrands Dubroeucq et al., 1998
Cold/Dry
Entisols ! Vitricyrands Arnalds and Kimble, 2001
Cold/Humid
Entisols ! Spodosols Ping et al., 1988; 1989; Shoji et al., 1988a
Entisols ! Andisols ! Spodosols Ping et al., 1988; 1989; Shoji et al., 1988a
Humid/Temperate
Warm/dry: Entisols ! Mollisols Otsuka et al., 1988; Shoji et al., 1990a
Warm/moist: Entisols ! Andisols Numerous reports
Warm/moist: Entisols ! Andisols Southard and Southard, 1987; Takahashi et al., 1993;
! Inceptisols ! Alfisols/Ultisols Chen et al., 2001
Cool/moist: Entisols ! Andisols Shoji et al., 1988b; Takahashi et al. 1989
! Spodosols
Cold/moist: Entisols ! Spodosols Shoji et al., 1988b; Ugolini et al., 1977; Parfitt and
Saigusa, 1985
Tropical
Warm dry/moist: Entisols ! Mollisols Wielemaker and Wakatsuki, 1984; Yerima et al., 1987;
Chadwick et al., 1994
Warm dry: Entisols ! Vertisols Yerima et al., 1987
Warm/moist: Entisols ! Andisols Martini, 1976; Delvaux et al., 1989
! Inceptisols ! Alfisols/Ultisols
Warm/moist: Entisols ! Andisols Chadwick et al., 1994; Martini, 1976; Quantin, 1992;
! Inceptisols ! Oxisols Van Ranst et al., 2002; Beinroth, 1982; Kimble and
Eswaran, 1988
However, in regions with intermittent deposition of volcanic ash each addition of

new material rejuvenates soil developmental processes so that Andisols may be
maintained as a relatively stable soil condition. Under humid weathering
conditions, criteria for andic soil properties (e.g., active Al and Fe) necessary for
classification as Andisols may be achieved in 200– 300 years (Dahlgren et al.,
1997). In dry climates, soil development is stalled in the Entisols/Vitrands stage
as weathering is limited by moisture (Dubroeucq et al., 1998; Arnalds and
Kimble, 2001). Mollisols form in volcanic materials under dry conditions in
temperate and tropical environments and under wetter conditions in the presence
of basic volcanic ash (Shoji et al., 1990a; Chadwick et al., 1994). Warm/dry
conditions promote formation of crystalline layer silicates rather than noncrystal-
line materials and leaching is limited leading to higher base saturation. Vertisols
may also form in warm regions having a distinct dry season as volcanic ash
weathers preferentially to smectite and vermiculite rather than noncrystalline
materials (Yerima et al., 1987). In the semi-arid (xeric) moisture regime of
California, Andisols are transformed to Inceptisols as noncrystalline materials are
transformed to kaolins resulting in a decrease of the active Al and Fe fraction
(Takahashi et al., 1993). Clay translocation is not a prevalent process in Andisols
due to the difficulty in dispersing noncrystalline materials. However, once
noncrystalline materials are transformed to layer silicate clays, translocation is
greatly enhanced, especially in the xeric moisture regime where repeated wetting
and drying leads to slaking and mobilization of clays. Formation of an argillic
(clay-enriched) horizon results in transformation of Inceptisols to Alfisols/
Ultisols (Southard and Southard, 1987; Takahashi et al., 1993).
Andisols are the typical product of weathering in both temperate and tropical
environments with sufficient moisture. As the degree of weathering increases,
metastable noncrystalline materials are gradually consumed by transformation to
more stable crystalline minerals (e.g., halloysite, kaolinite, gibbsite). This often
leads to the alteration of Andisols to Inceptisols, Alfisols or Ultisols (Martini,
1976; Delvaux et al., 1989; Chen et al., 2001). Formation of Oxisols occurs in the
perhumid tropics on highly weathered stable landscapes that have escaped
rejuvenation by intermittent ashfall or erosion (Beinroth, 1982). Under cool to
cold-humid conditions, Spodosols are often the dominant soil as cooler
temperatures favor coniferous vegetation and predominance of the organic acid
weathering/transport regime (Ugolini et al., 1977). Spodosols may form directly
from fresh volcanic deposits or as an alteration product from Andisols.
IV. SOIL CLASSIFICATION
At the international level, volcanic soils were first recognized as a distinct

category of soils, Andosols (from Japanese an, black and do, soil), in Japan in
1947 (Simonson, 1979). Andosols are currently defined as a major soil group
in the World Reference Base for Soil Resources (WRB) classification system
(FAO, ISRIC and ISSS, 1998). Volcanic soils were introduced as the Andept
suborder of Inceptisols in the Seventh Approximation (Soil Survey Staff, 1960)
and Soil Taxonomy (Soil Survey Staff, 1975). In 1978, Guy Smith proposed the
reclassification of Andepts into a new soil order of Soil Taxonomy termed
Andisols (Soil Survey Staff, 1998). The classical expression of volcanic soils is
roughly equivalent between Andosols (WRB) and Andisols (Soil Taxonomy) and
both of these systems have been rigorously tested for the international
classification of volcanic soils. In addition to these international classification
systems, some national soil classification systems [e.g., Classification of
Cultivated Soils in Japan, 3rd Approximation (Classification Committee of
Cultivated Soils, 1996); New Zealand Soil Classification — Version 3.0 (Hewitt,
1998)], provide a more detailed separation of soils for specific conditions within a
given country.
Because of their international applicability and wide spread utilization, we will
provide a brief discussion of Andosol (WRB) and Andisol (Soil Taxonomy)
classification criteria. The central concept of both Andosols and Andisols is soils,
dominated by volcanic ejecta (having a large component of volcanic glass) that
have a colloidal fraction dominated by amorphous and short-range-order materials
(allophane, imogolite, ferrihydrite, and Al/Fe – humus complexes) resulting in pH-
dependent variable charge cation exchange capacity (CEC) and anion exchange
capacity (AEC), high phosphate retention and low bulk density. Classification is
based on selected chemical, physical, and mineralogical properties acquired
through rapid weathering of volcanic glass and is not based exclusively on parent
material. As such, soils formed on volcanic ejecta may be classified in several other
soil groups/orders in WRB (e.g., Regosols, Nitisols, Cambisols, Luvisols, and
Vertisols) and Soil Taxonomy (e.g., Entisols, Inceptisols, Spodosols, Mollisols,
Alfisols, Ultisols, and Oxisols). Similarly, there are a few reports of Andosol-
s/Andisols formed in nonvolcanic parent materials in Spain (Garcia-Rodeja et al.,
1987), USA (Hunter et al., 1987; Wilson et al., 1996), Nepal (Bäumler and Zech,
1994), and India (Caner et al., 2000). These latter soils have an active iron
and aluminum fraction that is generally dominated by Al/Fe – humus complexes
rather than noncrystalline inorganic materials, such as allophane and imogolite.
A. WORLD REFERENCE BASE FOR SOIL RESOURCES — ANDOSOLS
WRB classification divides soils into 30 soil groups based on soil properties
defined in terms of reference horizons (FAO, ISRIC and ISSS, 1998). Andosols-
are defined taxonomically by the presence of either an andic or vitric horizon.
Andic horizons are characterized by the presence of high concentrations of active
Al and Fe, defined as those components extracted by acid oxalate (Alox, Feox, and
Siox from allophane, imogolite, ferrihydrite, and Al/Fe –humus complexes).

Vitric horizons are distinguished from andic horizons by their lesser degree of
weathering. As such, they are weakly weathered volcanic materials dominated by
volcanic glass and other primary minerals derived from volcanic ejecta that
display a weaker expression of andic properties.
Andic horizons must have the following properties:
1. bulk density at field-moisture water content , 0.9 g cm23,

2. 10% or more clay and an Alox þ 12Feox value in the fine-earth fraction $ 2%,
3. phosphate retention $ 70%,
4. a thickness of at least 30 cm.
Andic horizons are further subdivided into sil-andic horizons (dominated by

allophane and/or imogolite) having Siox $ 0.6%, and alu-andic horizons
(dominated by Al/Fe – humus complexes) having a Siox , 0.6%. These horizons
may be alternatively defined, as sil-andic horizons having an Alp/Alox , 0.5 and alu-
andic horizons having an Alp/Alox $0.5 (p ¼ pyrophosphate extractable). This
subdivision is important for differentiating whether the andic properties
result primarily from allophane/imogolite (sil-andic ¼ allophanic Andosols) or
Al–humus complexes (alu-andic ¼ nonallophanic Andosols).
Vitric horizons must have the following properties:
1. $ 10% volcanic glass and other primary minerals in the fine-earth fraction;
and either:
2. , 10% clay in the fine-earth fraction; or
3. a bulk density . 0.9 g cm23; or
4. Alox þ 12Feox $ 0.4%; or
5. phosphate retention . 25%; and
6. a thickness of at least 30 cm.
There are 25 lower-level soil units within Andosols that describe a wide range
of soil properties, such as humus accumulation, forms of active Al/Fe, degree of
weathering and leaching, drainage condition, presence of cemented layers,
presence of buried horizons, and accumulation of carbonates or salts. These soil
units provide important information regarding specific soil properties that
strongly affect soil productivity, utilization, and management.
B. SOIL TAXONOMY — ANDISOLS
Andisols are one of 12 soil orders in Soil Taxonomy and are defined
taxonomically by the presence of andic soil properties (Soil Survey Staff, 1998).
Andic soil properties result primarily from the presence of allophane, imogolite,
ferrihydrite, and Al/Fe –humus complexes. Criteria for andic soil properties in
Soil Taxonomy are nearly identical to andic horizons in WRB. Andic soil
properties are defined as soil materials with , 25% organic carbon that meet one
or both of the following requirements:
1. In the fine-earth fraction, all of the following:

1.1. Alox þ 12Feox $ 2.0; and
1.2. a bulk density, measured at 33 kPa water retention # 0.90 g cm23; and
1.3. a phosphate retention $ 85%; or
2. In the fine-earth fraction, a phosphate retention $ 25%, $ 30% particles 0.02
to 2.0 mm in size, and one of the following:
2.1. Alox þ 12Feox $ 0.4% and, in the 0.02 –2.0 mm fraction, $ 30% volcanic
glass; or
2.2. Alox þ 12Feox $ 2.0% and, $ 5% volcanic glass; or
2.3. Alox þ 12Feox totaling between 0.4 and 2.0%; and there is at least a
proportional content of volcanic glass in the 0.02 to 2.0 mm fraction
between 30 and 5% (this latter volcanic glass requirement is displayed
graphical in Soil Taxonomy).
Definition #1 is equivalent to the andic horizon in WRB while definition #2 is

similar to the vitric horizon in WRB. Definition #2 assumes that there is sufficient
volcanic glass remaining in the 0.02– 2.0 mm fraction that continued weathering
would result in Alox þ 12Feox values greater than 2%.
There are seven subgroups of Andisols that are based on soil moisture regimes
(aquic, aridic, xeric, ustic, udic), soil temperature regime (cyric), or low water
retention characteristics due to a low degree of weathering (vitric). The global
distribution of Andisols in these subgroups follows: vitrands (30.8%) < udands
(30.6%) . cryands (28.0%) . ustands (6.9%) . xerands (3.5%) . torrands
(0.2%) . aquands (, 0.1%) (Wilding, 2000). There are 12 unique elements
identified at the great group level that are based on several factors including soil
temperature and moisture regimes, cemented layers, degree of weathering,
organic matter characteristics and accumulation, water retention properties, and
drainage characteristics. Similarly, there are 24 unique elements identified at the
subgroup level that identify various soil properties.
In summary, Andosols (WRB) and Andisols (Soil Taxonomy) are differ-
entiated on the basis of selected chemical, physical, and mineralogical properties
rather than parent material alone (i.e., volcanic ejecta). As such, soils formed on
volcanic materials may be classified in several other soil groups/orders in WRB
and Soil Taxonomy and it is possible for soils formed on nonvolcanic parent
materials to be classified as Andosols/Andisols. Both classification systems have
nearly identical requirements defining andic/vitric horizons (WRB) and andic
soil properties (Soil Taxonomy): acid oxalate extractable Al/Fe, phosphate
retention, bulk density, and volcanic glass content. Both systems place strong
importance on the degree of weathering as reflected by vitric horizons in WRB

and the vitric suborder in Soil Taxonomy for weakly weathered volcanic
materials. WRB classification places an emphasis on distinguishing between
nonallophanic (alu-andic) and allophanic (sil-andic) Andosols as a high priority.
Soil Taxonomy does not specifically distinguish between these two subdivisions
of soils even though they indicate important differences in soil forming processes
and soil properties. In contrast, Soil Taxonomy places much greater importance
on soil temperature and moisture regimes by emphasizing these factors at the
suborder and great group levels. Most of the lower-level units of Andosols are
consistent with great group and subgroup units in Soil Taxonomy. Both systems
have received rigorous international testing and review during their development.
As a result, both classification systems provide excellent classification criteria for
soils formed from volcanic materials. These classification systems provide
important information regarding soil properties, productivity, utilization, and
management.
V. CHEMICAL WEATHERING
A distinctive feature of soils derived from volcanic material is the dominance

of noncrystalline and poorly crystalline components in the colloidal fraction.
Preferential formation of these materials results, in part, from rapid weathering of
volcanic glass, which shows low resistance to chemical weathering. In addition,
the fine particle size and vesicular nature (i.e., high surface area) and high
porosity and permeability (i.e., effective leaching) of volcanic materials enhance
weathering rates. Rapid weathering results in soil solutions becoming over
saturated with respect to several noncrystalline components. It is presumed that
rapid precipitation kinetics favor formation of metastable, noncrystalline
constituents compared to their more stable crystalline mineral counterparts.
Thus, rapid weathering is a key factor leading to the formation of andic soil
properties in pyroclastic deposits.
Volcanic materials are generally classified based upon their total silica content
(SiO2%): rhyolite (70 – 100%), dacite (62 –70%), andesite (58 –62%), basaltic
andesite (53.5 – 58%) and basalt (45 – 53.5%) (Shoji et al., 1975). The
mineralogical composition of volcanic materials varies according to rock types;
however, volcanic glass or glassy aggregates (i.e., glassy coated phenocrysts)
are the dominant component of all types of volcanic ejecta (Fig. 5; Shoji, 1983).
The chemical composition of volcanic glass is determined by the composition of
the magma from which it forms. Because of the rapid cooling of extrusive
magma, there is little segregation of crystals during the solidification process.
Volcanic glass is divided into colored and noncolored categories based on
refractive indices (colored . 1.52, noncolored # 1.52). Colored glass is found in
Figure 5 Relationship between rock types of volcanic materials and their mineralogy (adapted
from Shoji, 1983; copyright Hakuyusha, with permission).
basaltic andesite and basalt rock types while noncolored glass is found in
rhyolite, dacite, and andesite rock types. Colored glass has a low silica content
and a high concentration of cations, such as Al, Fe, Ca, and Mg, which leads to
substitution of these cations for silica in the glass structure. As a result, colored
glass is more susceptible to weathering due to destabilization of the Si –O –Si
framework in the glass by cations. The relative stability of mineralogical
components with regard to chemical weathering in volcanic ejecta has been
shown to follow (Aomine and Wada, 1962; Loughnan, 1969; Shoji et al., 1974;
Mitchell, 1975; Yamada et al., 1978):
Colored volcanic glass , noncolored glass < olivine , plagioclase ,
augite , hyperthene , hornblende , ferromagnetic minerals.
Laboratory weathering of volcanic glass shows an initial period of rapid
hydration and release of cations through surface exchange with aqueous
hydrogen ions (White and Claassen, 1980; White, 1983). Incongruent dissolution
forms a cation-depleted leached layer at the surface of the glass particle. The
extent of cation depletion and the thickness of the leached layer both increase as
the pH decreases (White, 1983). As weathering progresses, elemental release
rates are controlled by surface dissolution with concurrent diffusion of cations

through the leached layer. A shift from incongruent to congruent dissolution
presumably occurs when the increase in the diffusion length equals the rate of
retreat of the solution– solid interface.
Short-term (, 60 d) laboratory glass dissolution studies in the pH range of
5– 7 exhibit parabolic rate constants between 10214 and 10213 mol cm22 s20.5
for base cations, such as sodium, and linear rate constants between 10216 and
10215 mol cm22 s21 for silicon and aluminum (White and Claassen, 1980;
White, 1983). Shoji et al. (1993b) measured rates of aluminum release from
colored (basaltic-andesite composition) and noncolored (rhyolite composition)
glass in the order of 10212 mol g21 s21 at pH 4.0. Release rates were 1.5 times
greater for the colored glass reflecting its lower stability compared to noncolored
glass. Weathering rates increased about 1.5 times for each 108C increase in
temperature between 0 and 308C (Shoji et al., 1993b).
Airfall tephra often undergo considerable chemical weathering within the
eruption plume as the tephra interacts with acidic aerosols (H2SO4, HCl, HF,
HNO3) during the eruption and subsequent transport. Once condensed on tephra
particles, these acids are neutralized by hydrolytic reactions, which consume
protons and liberate cations. These cations combine with the strong acid anions to
form soluble salts that are quickly leached from the tephra. As a result, the initial
leachates from the 1980 eruption of Mt. St. Helens tephra had a near neutral pH
value with high concentrations of soluble salts composed primarily of base
cations (Ca2þ . Naþ .. Mg2þ . Kþ) and strong acid anions (SO22 4 .. Cl .
2
2 2
F .. NO3 ) (Dethier et al., 1981; Hinkley and Smith, 1987; Dahlgren and
Ugolini, 1989a).
In the immediate vicinity of volcanoes, where complete obliteration of the
vegetation has occurred, a primary proton donor driving chemical weathering is
carbonic acid (Ugolini et al., 1991). The primary source of carbonic acid in these
systems is from the atmosphere because production of CO2 by soil organism
respiration is low due to low organic matter concentrations in volcanic deposits.
However, in those cases in which volcanic gas emanations continue from the
volcanic vent, H2SO4 can be a prominent proton donor driving weathering
reactions. In the immediate vicinity of Mt. St. Helens, the pH of precipitation
ranged from 3.6 to 5.2 in the six years following the 1980 eruption with the
acidity originating primarily from H2SO4. Emission of sulfur dioxide (SO2)
averaged more than 1000 Mg d21 following the 1980 eruption (Casadevall et al.,
1983) providing a considerable proton load to the atmosphere upon oxidation to
H2SO4. Acidity originating from the precipitation in the vicinity of Mt. St. Helens
was largely consumed by weathering reactions in the upper 5 cm of the
pyroclastic flow (Nuhn, 1987).
Another example of chemical weathering on the barren landscape of the
Mt. St. Helens pyroclastic flows occurs in conjunction with lupine (Lupinus
species, a nitrogen fixer) establishment during primary succession. Protons are
generated as nitrogen fixed by lupine becomes oxidized to nitrate and leaches

along with base cations. Root and microbial respiration associated with lupine
also enhance soil CO2 concentrations leading to higher carbonic acid concentra-
tions in the soil solution. As a result, chemical weathering is enhanced in soils
occupying the “islands of soil fertility” in the vicinity of the individual lupine
plants (Nuhn, 1987).
In areas where airfall tephra deposition does not kill the vegetation, chemical
weathering is immediately influenced by biota that generate organic ligands and
protons from organic acids, carbonic acid, and through uptake of cations in
excess of anions. To examine the initial stages of chemical weathering in airfall
tephra deposits, Dahlgren and coworkers (1989a; 1997; 1999) applied 5 and
15 cm of previously nonleached tephra from Mt. St. Helens to the surface of
Spodosols in a subalpine forest ecosystem in the Washington Cascades. Soil
solutions from the tephra layer were collected continuously over a four-year
period and the elemental fluxes were used to calculate chemical weathering rates.
The primary proton donor in the fresh tephra layer was carbonic acid with a
minor contribution from organic acids originating from leaching of the tree
canopy and the incipient litter layer accumulating at the surface of the tephra
layer (Dahlgren et al., 1999). The dominance of carbonic acid weathering was
demonstrated by the near stoichiometric balance between base cations and the
bicarbonate anion in tephra leachates. Solutions leached from the tephra layer
indicated incongruent dissolution resulting in formation of a cation-depleted,
silica-rich leached layer on the glass and mineral surfaces. Due to the near neutral
pH values and low concentrations of complexing organic ligands, aluminum and
iron were relatively insoluble and accumulated in the tephra layer rather than
being leached. Calcium and sodium were released about 3– 4 times faster than
silica based on their stoichiometry in the tephra. Field weathering rates,
calculated as yearly averages, ranged between 10218 and 10217 mol cm22 s21
for sodium, calcium and silicon (Fig. 6). These rates are 1– 3 orders of magnitude
less than those determined for glass and plagioclase minerals at steady state in
laboratory dissolution experiments at pH 5 –7 and 258C. The principal factors
contributing to the difference between field and laboratory weathering rates are
availability of water, temperature, and the extent of solution –solid mixing that
serves to reduce the influence of solute diffusion on the rates. The field
weathering rates at this study site were substantially hindered by low soil
temperatures (5.58C mean annual air temperature), which for at least six months
each year are near freezing when the soils are buried by a snowpack. Similarly,
when soil temperatures are higher during summer, there is a water deficit that
hinders chemical weathering. Thus, field weathering rates are greatly attenuated
because of the inverse relationship between optimal temperature and moisture
conditions: (1) when moisture is most abundant, the soil temperatures are near
freezing; and (2) when soil temperatures are highest, soil water availability and
leaching are at their lowest.
Figure 6 Weathering rates for 5 cm ash layer reported as a yearly average for the first four years
following addition of ash to the surface of a Spodosol in the western Cascade Range, Washington
(Dahlgren et al., 1999).
Given these unfavorable field conditions for chemical weathering, it is

remarkable how fast dissolution rates are for the fresh tephra. Cation denuda-
tion rates from the 5 and 15 cm tephra layers (1.2 – 9.4 kmolc ha21 yr21)
generally exceed cation denudation rates determined for entire watersheds
(0.1 – 1.5 kmolc ha21 yr21) as reviewed by Sverdrup and Warfvinge (1988) and
Sverdrup (1990). This comparison implies that the initial rates of tephra
weathering are much more rapid compared to those of other silicate parent
materials. Over the 4-year study period, weathering rates decreased by a factor of
3– 5 times. The largest fraction of the decrease occurred between the first and the
second year, while weathering rates were more similar in the third and the fourth
year. The large initial decrease is probably attributable to the formation of the
cation-depleted leached layer and to the rapid dissolution of easily weatherable
materials (e.g., glass) and high surface area microparticles. Electron micrographs
comparing the original, nonweathered tephra with the same tephra collected after
10 years of weathering show that there is a distinct loss of the sharp edges and
vesicularity (high-energy sites) in the weathered tephra (Fig. 7). In a separate
study, White et al. (1986) demonstrated the formation of a cation-depleted
leached layer (. 10 nm) on tephra particles after two years of weathering in the
soil environment using X-ray photoelectron spectroscopy (XPS).
Weathering rates for the 5 cm tephra layer were about a factor of two greater
than for the 15 cm tephra layer. The primary reason for this difference is believed
to be due to a greater proton flux of carbonic acid per unit of tephra in the 5 cm
layer. The majority of the CO2 in the tephra layer originates from upward
Figure 7 Volcanic ash from the 18 May 1980 eruption of Mt. St. Helens, Washington.
Nonweathered ash (top) and the same ash weathered for 10 years (bottom) in a cool–humid
environment in the Washington Cascades. Weathering results in a distinct loss of the sharp edges and
vesicularity (Dahlgren et al., 1997; 1999).
transport of CO2 from the organic-rich soil horizons (~15 cm of O horizon) of the
buried soil. Elevated concentrations of CO2 beneath the tephra layer originate
from biological respiration (e.g., roots and microorganisms) and from
protonation of HCO2 3 leaching from the overlying tephra layer. Cations released
by weathering in the tephra layer (pH < 62 7) leach downward with bicarbonate
to the acidic organic horizons (pH < 4) where H2CO3 reequilibrates with the
high pCO2 (Fig. 8). At pH 4, H2CO3 decomposes to CO2 and H2O and the
gaseous CO2 diffuses upwards to take part in another cycle of weathering and
transport. If the flux of CO2 transported upwards from the buried soil is similar
Figure 8 Schematic representation of the CO2 –H2CO3 –HCO2 3 weathering/transport cycle

occurring between the ash layer at the surface of the soil and the buried organic soil layer (Dahlgren
et al., 1999).
beneath the 5 and 15 cm tephra layers, the amount of H2CO3 available for
weathering on a per unit volume or mass basis of tephra would be three times
greater for the 5 cm layer. The fact that the measured weathering rates in the
5 cm treatment are only a factor of 2 greater may be related to kinetic limitations
of weathering reactions or to more rapid consumption of easily weatherable
minerals in the 5 cm layer.
This example of weathering in airfall tephra deposits demonstrates a unique
weathering pathway in which the buried organic-rich soil pumps protons upward
to the tephra layer that acts as an alkaline trap for CO2. Thus, the overall process
of tephra weathering appears to be controlled largely by solute/gas transport of the
carbonate system (CO2 –H2CO3 – HCO3). In addition, the base cations leached
from the tephra layer exchange with Hþ on the cation-exchange complex of the
organic horizons resulting in an increased base saturation, somewhat higher pH
levels, and enhanced plant availability of nutrient cations.
An important variable controlling the mineralogy of the colloidal fraction in
volcanic soils is the soil solution silicon activity. The initial rapid weathering of
glassy volcanic materials provides an abundant source of silica to support silica-
rich minerals, such as opaline silica and halloysite. As weathering proceeds, the
concentrations of highly weatherable materials, most importantly volcanic glass
and glassy aggregates, are depleted and soil solution silica activities decrease.
Studies examining long-term weathering of volcanic glass in tephra deposits
indicate half-lives for volcanic glass ranging from 1650 to 7000 years (Kirkman
and McHardy, 1980; Ruxton, 1988; Shoji et al., 1993b). It is important to
consider that weathering rates and volcanic glass half-lives will vary considerably
depending on factors such as climate, vegetation, tephra depth, particle size and
chemistry.
In summary, rapid chemical weathering of volcanic ejecta is an important
factor leading to the formation of a colloidal fraction dominated by noncrystalline
and poorly crystalline constituents. These materials are in turn responsible for
andic horizons/soil properties, which are the primary features distinguishing
Andosols/Andisols as a distinctive soil category. Carbonic acid is an important
proton donor driving weathering reactions in tephra. Carbonic acid weathering is
an appreciable sink for carbon dioxide as gaseous carbon dioxide is sequestered
as aqueous bicarbonate/carbonate. The rapid release of nutrient elements by
weathering is also an important factor for rejuvenating soil forming processes and
enhancing soil mineral nutrition.
VI. MINERALOGICAL CHARACTERISTICS
Rapid weathering of volcanic materials often leads to a colloidal fraction

dominated by noncrystalline or poorly crystalline materials. The rapid kinetics
of nucleation for noncrystalline materials favor formation of these metastable
solid phases relative to their more thermodynamically stable crystalline
mineral equivalents (Ostwald step rule or the rule of steps; Stumm, 1992).
Preferential precipitation of noncrystalline materials occurs because nucleation
of a more soluble phase (i.e., noncrystalline materials) is kinetically favored
over that of a less soluble phase (i.e., crystalline mineral) owing to the lower
solid –solution interfacial tension of the more soluble phase (Stumm, 1992).
Thus, nucleation of a noncrystalline phase will occur at a lower degree of
supersaturation resulting in a higher nucleation rate than for crystalline min-
erals. Once nucleation is initiated, the rapid expansion in noncrystalline surface
area that is nucleation-controlled will act to lower the degree of super-
saturation, further reducing the probability of nucleation of the more crystalline
phase (Robarge, 1999).
Over time, crystalline minerals form at the expense of their metastable,
noncrystalline precursors. The “aging” or “Ostwald Ripening” process (i.e.,
transformation of noncrystalline to crystalline phases via dissolution and
reprecipitation) is typically a very slow process due to the small differences in
free energy between the two phases. Climatic conditions play an important role in
the formation of crystalline minerals as crystallization is promoted as the soil
climate becomes warmer and dryer (Talibudeen, 1981; Schwertmann, 1985).
Noncrystalline materials are more persistent under cooler soil conditions because
crystallization is hindered by the low input of thermal energy. Seasonal soil
desiccation has been shown to enhance transformation of noncrystalline
materials to crystalline minerals (Quantin, 1992). Fieldes (1966) proposed that a

pronounced dry period led to crystallization via formation of cross-linkages
within aluminosilicate gels during the period of dehydration. Thus, periods of
wetting and drying promote formation of crystalline minerals relative to non-
crystalline materials.
The dominant colloidal assemblage of soils formed in volcanic materials varies
widely depending on factors such as the composition of the parent material, stage
of soil formation (i.e., degree of weathering), pH, soil temperature and moisture
regimes, and the accumulation of organic matter (Lowe, 1986; Shoji et al., 1993a).
The weathering environment defined by these factors regulates Al and Si activities,
which along with soil climate are the primary determinants of soil colloid genesis
and transformation. Three colloidal assemblages tend to predominate in weakly
to moderately weathered soils derived from volcanic materials: (1) Al– humus
complexes, often in association with hydroxy-Al interlayered 2:1 layer silicates
and opaline silica, (2) allophane and imogolite, and (3) halloysite. Ferrihydrite
is often the predominant iron (hydr)oxide formed during the early stages of
weathering in all volcanic materials (Childs et al., 1991). Authigenic formation of
2:1 layer silicates is limited in volcanic materials, however, formation of smectites
and 1:1 –2:1 mixed-layer clays (e.g., halloysite –smectite) appears to be common
in soils derived from volcanic materials of basaltic composition in tropical and
subtropical areas (Quantin et al., 1988; Delvaux et al., 1990b, 1992; Quantin,
1992; Delvaux and Herbillon, 1995). Noncrystalline and poorly crystalline
materials contribute to many of the unique physical and chemical properties of
volcanic soils due to their high surface area, variable charge surfaces, high
solubility, rapid reaction kinetics and unusual physical properties. In the following
sections, we briefly describe the major colloidal constituents found in soils derived
from volcanic materials and discuss genesis and transformation pathways. For a
more detailed discussion, refer to reviews by Wada and Harward (1974), Wada
(1980; 1985; 1986; 1989), Yoshinaga (1988), Parfitt and Kimble (1989), Parfitt
(1990), and Shoji et al. (1993a).
A. COMMON COLLOIDAL CONSTITUENTS OF VOLCANIC SOILS
1. Allophane
Allophane is a group name given to a series of naturally occurring,

noncrystalline, hydrous aluminosilicates with a varying chemical composition
(van Olphen, 1971). Allophane consists of hollow, irregularly spherical particles
with outside diameters of 3.5– 5.0 nm and a wall thickness of 0.7 –1 nm. There
is evidence that micopores exist in the allophane spherule with diameters
ranging between 0.3 and 2.0 nm (Paterson, 1977). The chemical composition
generally ranges from an Al:Si atomic ratio of 1:1 – 2:1. Based on the structural
and chemical characterization of natural allophanes, three major types of

allophanes have been identified: Al-rich (proto-imogolite allophane), Si-rich
(halloysite-like or defect kaolin), and hydrous feldspathoid allophanes (Farmer
and Russell, 1990).
Al-rich allophane (proto-imogolite) has the same local atomic arrangement
(short-range-order) and chemical composition (Al:Si ¼ 2:1) as imogolite;
however, they differ morphologically (Parfitt and Henmi, 1980). The structure
consists of a gibbsite-like sheet of octahedrally coordinated Al with Si bound to
three Al– O of the gibbsite sheet and the remaining Si – OH group pointing toward
the center of the sphere. It appears to form as fragments of the imogolite structure,
depending on solution conditions they either form tubular (imogolite) or spherical
morphology (Farmer and Russell, 1990).
Si-rich allophanes have an Al:Si atomic ratio close to 1:1 with Al primarily in
octahedral coordination (gibbsite sheet) and silica occurring as both polymerized
silicate and orthosilicate groups (Parfitt, 1990). Spectroscopic studies reveal that
Si-rich allophane contains both defect kaolin- and imogolite-structural units with
the portion of polymerized Si increasing as the Al:Si atomic ratio decreases
(Parfitt et al., 1980; Goodman et al., 1985; Shimizu et al., 1988). The isolated
orthosilicate groups are associated with the gibbsite sheet of the 1:1 kaolin layer
and the orthosilicate groups penetrate the inner sheet of silica tetrahedra through
the holes formed by defects (MacKenzie et al., 1991). Thus, its structure is
similar to a kaolin mineral with defects in the tetrahedral sheet. Reports of
allophane with Al:Si atomic ratios between 2:1 and 1:1, based on selective
dissolution analyses, may result from mixtures of the Al-rich and Si-rich
allophane end-members. However, laboratory synthesis experiments utilizing a
range of Al:Si atomic ratios in the matrix solutions are able to produce allophanes
with a range of Al:Si atomic ratios from 1 to 2. Attempting to understand the
chemical composition of allophanes based on selective dissolution analysis is not
very reliable. Further research is needed to verify the nature of allophanes with
apparent Al:Si atomic ratios between 1 and 2.
Hydrous feldspathoid allophane contains no imogolite unit structures and
has a significant amount of tetrahedrally coordinated Al. The silicate layer is
substantially polymerized with 1:3 Al for Si substitution (Childs et al., 1990).
This is the same degree of substitution as in muscovite and suggests that this
allophane phase has a structure based on a similar condensed silicate sheet.
The proposed structure consists of a more or less complete tetrahedral sheet on
the outer surface with an incomplete octahedral Al sheet on the inside.
It preferentially forms at higher concentrations of silicon and at slightly acidic
to neutral pH values, consistent with the formation of tetrahedrally
coordinated Al species in solution. The most notable occurrence of natural
feldspathoid allophane is as stream channel deposits below Silica
Springs on the flanks of Mt. Ruapehu, New Zealand (Wells et al., 1977; Childs
et al., 1990).
2. Imogolite
Imogolite was first described by Yoshinaga and Aomine (1962) in a soil

derived from glassy volcanic ash, known as “imogo.” It has a distinct tubular
morphology consisting of bundles of well-defined fine tubes with inner and outer
diameters of about 1.0 and 2.0 nm, respectively. Imogolite displays long-range
order only along the length of the tubes, which may extend several microns. It has
a fixed chemical composition [(OH)3Al2O3SiOH] with an Al:Si atomic ratio of
2:1. The external surface consists of a gibbsite-like sheet with orthosilicate
groups coordinated through oxygen with three aluminum atoms and one free OH
group pointing toward the center of the tube (Cradwick et al., 1972). Aluminum
is found exclusively in octahedral coordination.
The precursor to imogolite is a sol consisting of small fragments of proto-
imogolite allophane sheets that are too small to show tubular morphology.
The formation of imogolite from proto-imogolite sheets is a time dependent
process akin to crystallization (Farmer and Fraser, 1979; Wilson et al., 2001).
Crystallization is hindered by the presence of iron and complexing anions, such
as weak organic acids. Thus, imogolite formation is favored under conditions of
carbonic acid weathering (generally B horizons) as opposed to the organic acid
weathering regime (often A horizons), which produces conjugate bases with
complexing capacity.
3. Halloysite
Halloysite is a common constituent in volcanic soils and is often the dominant

colloidal constituent in Si-rich environments (Parfitt and Wilson, 1985).
Halloysite is a 1:1 aluminosilicate mineral, which is characterized by a diversity
of morphologies (e.g., spheroidal, tubular), chemical composition, specific
surface area, structural disorder, and physiochemical properties (e.g., CEC, ion
selectivity). Halloysite is hydrated (1.0 nm basal spacing) by two interlayer water
molecules; however, they are susceptible to dehydration (0.7 nm basal spacing)
under climatic conditions showing a distinct seasonal moisture deficit (Church-
man and Gilkes, 1989; Takahashi et al., 2001). In terms of nomenclature, the
various forms of halloysite are often described as hydrated halloysite for those
having a 1.0 nm basal reflection and dehydrated halloysite for those having a
0.7 nm basal reflection that is expandable to 1.0 nm by formamide treatment. In
the xeric moisture regime of California, halloysite (1.0 nm with or without
formamide treatment) and kaolinite (0.7 nm with formamide) concentrations
show an inverse relationship with depth, with the content of halloysite increasing
down the profile (Takahashi et al., 2001). The kaolinite in the surface horizons has
the same tubular morphology as the halloysite at depth suggesting that hydrated
halloysite transforms to kaolinite upon dehydration. The term “tubular kaolinite”
has been used to describe kaolins with tubular morphology and a 0.7 nm basal
reflection that does not expand with formamide (Churchman and Gilkes, 1989).
Halloysite displays a wide range of structural disorder, which is primarily due
to Al-vacancy displacements in the octahedral sheet and to random stacking of
structural layers (Churchman and Theng, 1984; Soma et al., 1992; Newman et al.,
1994). These vacancies may originate from nonstoichiometric substitution of
Fe3þ for Al3þ in the octahedral sheet (Soma et al., 1992). A number of authors
(e.g., Wada and Mizota, 1982; Delvaux et al., 1990b; Malucelli et al., 1999) have
shown instances of halloysite incorporating appreciable concentrations of iron (up
to 17% Fe). Variable effects of hydration might also add to the degree of disorder
in halloysite. Vacancies in the octahedral sheet and isomorphous substitution of
Al3þ for Si4þ in the tetrahedral position have been suggested as mechanisms by
which halloysite may acquire permanent negative charge. The CEC values for
halloysite generally range between 2 and 10 cmolc kg21, although values in
excess of 20 cmolc kg21 clay have been reported (Bailey, 1990; Norrish, 1995).
4. 2:1 Layer Silicate Clays
Layer silicate minerals of the 2:1 type are often found in soils derived from
volcanic ejecta. Their occurrence in young volcanic materials is variously
ascribed to in situ pedogenic origin, eolian addition, or inheritance from
hydrothermally altered materials in the parent material. These layer silicates are
the dominant clay mineral in the nonallophanic group of Andosols/Andisols.
During the early stages of weathering, mica, vermiculite, and smectite are present
as the dominant 2:1 layer silicates. As weathering advances, hydroxy-Al
interlayering occurs in vermiculite and smectite resulting in a sink for Al released
by weathering (Shoji and Fujiwara, 1984). Hydroxy-Al interlayering of 2:1 layer
silicates reduces their susceptibility to alteration and increases their thermo-
dynamic stability (Karathanasis et al., 1983).
The authigenic formation of 2:1 layer silicates in volcanic soils has long been a
topic of debate. Inheritance of 2:1 layer silicates formed by hydrothermal
alteration in the volcanic cone prior to eruption and deposited with volcanic ejecta
has been documented (e.g., Kondo et al., 1979; Pevear et al., 1982; LaManna and
Ugolini, 1987; Jongmans et al., 1994). Recent isotopic evidence demonstrates the
importance of eolian transport as a major source of 2:1 layer silicates in areas such
as Japan, Canary Islands, Tanzania, and Hawaii (e.g., Dymond et al., 1974; Inoue,
1981; Inoue and Naruse, 1987; Mizota et al., 1988; Mizota and Matsuhisa, 1995).
Eolian deposition explains the preferential accumulation of 2:1 layer silicates in
surficial horizons of older landscapes and those having higher precipitation.
While the formation of 2:1 layer silicates in volcanic materials in the udic soil
moisture regime cannot be ruled out, current evidence strongly suggests that
eolian deposition is the major source of 2:1 layer silicates, with lesser amounts
inherited from pyroclastic deposits.
Weathering of basalt results in the formation of smectite as the initial
crystalline weathering product in seasonally dry environments (Glassmann and
Simonson, 1985; Quantin, 1992). Under a tropical environment with a distinct
dry season, Quantin (1992) identified a beidellite-phase with a crumpled
morphology after 500 years and it transformed to a more crystalline form after
2000 years. In contrast, no 2:1 layer silicates were detected on similar parent
materials under perudic conditions in the same region.
5. 1:1 – 2:1 Mixed-layer Clays (Halloysite-Smectite,

Kaolinite-Smectite)
The presence of 1:1 – 2:1 mixed-layer clays as weathering products in volcanic

materials has been demonstrated in basaltic volcanic ash, primarily in tropical
and subtropical areas (Quantin et al., 1988; Delvaux et al., 1990 a, b, 1992;
Delvaux and Herbillon, 1995). The occurrence of such clays explains the high
CEC and Kþ selectivity of soils containing these clay minerals. Halloysite –
smectite mixed-layer clays form in silica- and base-rich environments and are
highly sensitive to the microenvironment. Weathering of basalt leads to smectite
formation on the fresh surfaces of the pyroclastic materials while halloysite forms
in the more open microenvironment subject to more leaching (Delvaux and
Herbillon, 1995). It is therefore expected that 1:1 – 2:1 mixed-layer clays would
form from primary minerals via a syngenetic (co-neoformation) pathway under
’intermediate’ microenvironmental conditions. In contrast, the formation path-
way of kaolinite-smectite is a desilication process in which iron-rich 2:1 clays
transform to 1:1 clays and iron (hydr)oxides. The intermediate step of such a
transformation consists of 1:1 –2:1 mixed-layer clays in which the smectite
content and layer charge both decrease as weathering proceeds (Delvaux et al.,
1990b). Delvaux et al. (1990b) estimated a smectite content of 14 to 31% in
halloysite – smectite mixed-layer clays from Western Cameroon. The smectite
component contained octahedral iron and appeared to be a high-charge beidellite
phase. Due to the difficulty of detecting 1:1 – 2:1 mixed-layer clays by standard
mineralogical methods, the occurrence of these minerals may be appreciably
underestimate in volcanic soils. Thus, special efforts should be made to verify the
presence/absence of 1:1 – 2:1 mixed-layer clays in soils displaying halloysite-rich
clays with high CEC and Kþ selectivity.
6. Opaline Silica
Two types of opaline silica are common in young volcanic soils: pedogenic
(commonly known as laminar opaline silica) and biogenic (plant opal and
diatoms). Laminar opaline silica is predominantly found in surface horizons

during the early stages of weathering of volcanic materials (Shoji and Masui,
1971). It was also found in the E horizon of Spodosols derived from recent
volcanic ash in the Washington Cascades (Ugolini et al., 1991) and as a cementing
agent in indurated subsoil horizons (tepetate horizons) of volcanic soils in Mexico
(Hidalgo et al., 1995). Laminar opaline silica is most often circular or elliptical in
shape and is most abundant in the 0.4 to 2 mm fraction.
Opaline silica forms in Si-rich environments from supersaturated solutions
resulting from surface evaporation and possibly freezing. Preferential incorpor-
ation of Al into humus complexes and hydroxy-Al interlayers rather than
aluminosilicate compounds may result in high silica activities that favor the
formation of opaline silica. Opaline silica may persist in volcanic soils for more
than 5000 years; however, it is progressively lost as chemical weathering rates
decrease over time resulting in lower availability of silica (Shoji and Masui,
1971). Opaline silica does not have a significant influence on soil properties other
than buffering soil solution silica activities at a relatively high level during the
early stages of soil development.
7. Ferrihydrite
Iron-bearing weathering products occur primarily as noncrystalline (hydr)-

oxides in young volcanic soils. The dominant noncrystalline (hydr)oxide is
ferrihydrite, a short-range-order Fe oxyhydroxide with a bulk composition of
5Fe2O3(9H2O) (Schwertmann and Taylor, 1989). Ferrihydrite resembles hematite
structurally, except that some Fe positions are vacant and some O and OH groups
are replaced by water molecules (Towe and Bradley, 1967). The degree of
crystalline order varies widely, with “2-line” and “6-line” ferrihydrite
differentiated by the number of diffraction peaks observed by X-ray diffraction
(Schwertmann and Cornell, 1991). Ferrihydrite appears as individual spherical
particles ranging in size between 2 and 5 nm. These particles become highly
aggregated forming aggregates 100 – 300 nm in diameter. Ferrihydrite is
metastable and with time converts to stable Fe (hydr)oxides, usually to goethite
under temperate or cool, humid climate, and to hematite under warmer, dryer
climate. The high affinity of ferrihydrite for sorption of silica and humic
substances may contribute to its formation by impeding the crystallization
process (Schwertmann, 1988). Formation of siliceous iron (hydr)oxides from
peralkaline volcanic ash was promoted by drier climatic condition in the Great
Rift Valley of Kenya (Wakatsuki and Wielemaker, 1985). The SiO2/Fe2O3
molecular ratios of citrate –dithionite –bicarbonate extracts for these siliceous
iron (hydr)oxides ranged from 1.8 to 4.2.
8. Hisingerite
An iron silicate mineral termed hisingerite has been proposed as an early

product of weathering in iron-rich volcanic deposits (Sudo and Nakamura, 1952;
Quantin, 1992). Hisingerite has been variously described as a noncrystalline iron
silicate (silica-rich ferrihydrite), as a ferric allophane, as ferric halloysite, and as a
poor crystalline nontronite (Eggleton and Tilley, 1998). A rigorous evaluation of
hisingerite specimens by Eggleton and Tilley (1998) showed that pure specimens
were a ferric analogue of spherical halloysite [Fe2Si2O5(OH)4(2H2O)]. Trans-
mission electron microscopy shows a fabric of concentric spheres, with a diameter
of about 14 nm, and shell walls 0.7 or 1.0 nm thick. The controversy concerning
the nature of hisingerite is most likely the result of several impurities in soil
samples.
9. Aluminum – and Iron –Humus Complexes
Aluminum – and iron– humus complexes are often the dominant form of active
Al and Fe in acidic (pH , 5), organic-rich horizons, especially in nonallophanic
Andisols (Shoji and Fujiwara, 1984). This fraction is typically defined as the Al
and Fe fraction extracted by pyrophosphate reagent (McKeague, 1967). While the
Al– humus fraction can be large in some soils, Fe– humus complexes are generally
much lower because iron has a greater stability as Fe (hydr)oxides compared to
humus complexes (Wada and Higashi, 1976). Metal –humus complexes are
believed to form primarily by the interaction of metals with carboxylic functional
groups. The complexing capacity of humus increases as the degree of humification
increases. The degree of metal complexation by humic substances can be
evaluated by examining pyrophosphate extractable Al, Fe, and C using the ratio:
(Alp þ Fep)/Cp. For most Andosols/Andisols, the ratio ranges between 0.1 and
0.2 (Inoue and Higashi, 1988). The degree of complexation is determined by the
stability constants for metals with complexing functional groups, solution pH,
the concentration and speciation of aqueous Al and Fe, and the concentration of
competing aqueous species. It is suggested that the accumulation and stabilization
of humus in volcanic soils is, in part, due to the formation of metal – humus
complexes (Percival et al., 2000).
B. FORMATION AND TRANSFORMATION OF COLLOIDS IN

VOLCANIC SOILS
The composition of the colloidal fraction in volcanic soils is regulated by

elemental activities, pH, soil temperature, and seasonal desiccation of the soil
profile (Table II). During the early stages of weathering, elements are rapidly
142
R. A. DAHLGREN, M. SAIGUSA AND F. C. UGOLINI
Table II
Soil Environmental Conditions Favoring the Formation of Colloidal Constituents in Soils Formed in Volcanic Materials
Humus Base Seasonal Soil

Colloidal constituent pH Al activity Si activity content saturation desiccation temperature Weathering regime
Al–humus complexes ,5 high Organic acid

Hydroxy-Al interlayered 2:1 5– 6 Organic acid
layer silicates or carbonic acid
Opaline silica 4– 6 Low High High Promotes Freezing Organic acid
Al-rich allophane 5– 6 Moderate Moderate Low Cool Carbonic acid
Imogolite 5– 6 Moderate Moderate Low Cool Carbonic acid
Si-rich allophane 6– 7 Moderate High Low Cool Carbonic acid
Halloysite 5– 6 Moderate High Low Low Promotes Warm Carbonic acid
2:1 layer silicates 6– 8 Low High Low Moderate Promotes Warm Carbonic acid
to high
Ferrihydrite 4– 7 Moderate Moderate Hinders Cool Organic acid
to high or carbonic acid
Goethite Low Low Promotes Cool Carbonic acid
released from volcanic glass and primary minerals resulting in high elemental
activities. Elemental activities are subsequently changed by complexation with
organic matter, incorporation into colloidal constituents, consumption of easily
weatherable constituents, and leaching or desiccation. Wetting and drying cycles
coupled with soil temperature regimes also have a major influence on the colloidal
fraction through their influence on crystallization. In addition, complexing and
sorbed anions (e.g., organic acids, sulfate, silicate) can hinder the crystallization
processes by blocking lattice sites (Huang and Violante, 1986).
Humus and 2:1 layer silicates have a strong affinity for aluminum resulting in
their preferential formation until they become effectively saturated. Al– humus
complexes with or without hydroxy-Al interlayered 2:1 layer silicates are the
dominant colloidal constituents in nonallophanic Andosols. They form in areas
with higher precipitation resulting in greater leaching and lower pH values
(pH , 5). At low pH values, aluminum exists primarily as Al3þ which is strongly
complexed by organic acids and does not undergo appreciable hydrolysis
and polymerization, processes necessary for formation of aluminosilicate
constituents.
In addition to humus, 2:1 layer silicates act as a sink for Al released by
weathering. Soil pH values in the range of 5 – 6 have been shown to favor
polymerization of hydroxy-Al polymers and their incorporation into the
interlayer of 2:1 layer silicates (Jackson, 1963b). Nonallophanic Andosols
often have a considerable amount of 2:1 minerals showing a wide range in
the degree of hydroxy-Al interlayer filling. Preferential incorporation of Al into
Al– humus complexes and hydroxy-Al interlayers effectively competes for
soluble Al so that little Al is available for formation of allophanic materials. This
inhibition mechanism has been termed the “anti-allophanic effect ” (Shoji et al.
1993) following the “anti-gibbsitic effect ” coined by Jackson (1963b). As a
result, there is an inverse relationship between the concentrations of Al– humus
complexes plus 2:1 layer silicates and concentrations of allophane/imogolite
(Fig. 4).
Once humus and 2:1 layer silicates are saturated with respect to aqueous Al3þ
activities, excess Al3þ is available for combining with silica to form allophane
and imogolite. Formation of allophane/imogolite occurs at (Alp þ Fep)/Cp values
greater than about 0.1 –0.2, corresponding to maximum Al-complexation by
humus (Parfitt and Saigusa, 1985; Dahlgren and Ugolini, 1991). Allophane and
imogolite formation is favored in the pH range 5 – 7, coinciding with carbonic
acid as the major proton donor. Within this pH range, soluble Al undergoes
hydrolysis and polymerization reactions. The polymerized Al may then combine
with silica to form allophane and/or imogolite structures. Al-rich allophane is
composed of proto-imogolite subunits and appears to form preferentially under
conditions of greater supersaturation and in the presence of complexing or sorbed
anions. More rapid precipitation kinetics and blockage of precipitation sites by
sorbing anions lead to a more disordered proto-imogolite allophane structure in
which the proto-imogolite fragments form irregular edge-to-edge associations

and a hollow spheroidal morphology. Slower precipitation kinetics coupled with
the relative absence of complexing/sorbing anions allows formation of a more
ordered arrangement of the proto-imogolite subunits into the thread-like
imogolite strands. Proto-imogolite allophane cannot transform to imogolite
without dissolution and reprecipitation (Farmer and Russell, 1990). In silica-rich
environments, the Al:Si atomic ratio of the allophane approaches 1:1 and consists
of a disordered spheroidal morphology. In the case of the hydrous feldspathoid
allophane from Silica Springs, New Zealand, rapid neutralization of acidity by
CO2 degassing in the stream channel results in the precipitation of allophane
within a few hundred meters of the spring outlet. Thus, allophane precipitation
kinetics can be extremely fast in nature.
In the soil environment, allophane and imogolite are believed to form in situ
through carbonic acid weathering in A and B horizons of Andisols (Ugolini et al.,
1988) and in the B horizons of tephritic Spodosols (Ugolini and Dahlgren, 1987;
Dahlgren and Ugolini, 1989b). The carbonic acid weathering regime favors
allophane/imogolite formation because it occurs in the preferred pH range of
5 –7 (promoting hydrolysis and polymerization of Al) and the bicarbonate anion
does not complex Al or appreciably sorbed to mineral surfaces to interfere with
crystallization. Cooler soil temperatures and the lack of a distinct dry season also
appear to favor formation of poorly ordered allophane and imogolite relative to
their more crystalline counterparts (e.g., halloysite).
Halloysite forms in silica-rich environments, such as regions with low rainfall
and leaching, in buried soil layers receiving silica from above, and in zones with
restricted drainage that prevent leaching. Halloysite may occur in the clay, silt, and
sand fractions (Simonett and Bauleke, 1963; Takahashi et al., 2001). Halloysite
may crystallize directly from the dissolution products of volcanic glass and
primary minerals, alteration of feldspars, or as a transformation product of
allophane/imogolite via dissolution and reprecipitation. Halloysite formation is
favored by high silica activities (. 250 – 350 mM) and by warmer and
drier conditions that tend to increase dissolved silica concentrations and
promote crystallization (Parfitt and Wilson, 1985). Halloysite is found as the
dominant mineral in volcanic materials where precipitation is generally less
than about 1500 mm (Parfitt et al., 1983; Lowe, 1986; Mizota and van Reeuwijk,
1989; Takahashi et al., 1993). Seasonal desiccation of the upper soil horizons leads
to dehydration of 1.0 nm halloysite to 0.7 nm halloysite/kaolinite that may or may
not be expandable to 1.0 nm with formamide. Halloysite is also favored by
high silica concentrations associated with poorly drained subsoil horizons (Lowe,
1986) and buried soil environments (Saigusa et al., 1978).
There has been considerable controversy concerning the origin of 2:1 layer
silicates in soils derived from volcanoclastic materials. Initially, 2:1 layer silicates
were thought to form by an in situ synthesis process starting with amorphous
precursors. More recently, this theory has been largely discarded in favor of eolian
sources and inheritance. However, the formation of 2:1 layer silicates in soils
formed on basaltic parent materials in tropical environments having a pronounced
dry season is well documented (Quantin, 1992). Smectite formation is favored by
high pH, silica activities, and base saturation, along with warm soil temperatures
and seasonal drying that promote crystallization. The leaching environment
(a function of precipitation) strongly affects the stability of smectite in the tropical
areas. Smectite is stable for only a short time (, 10,000 years) under humid
conditions, while no smectite persists under perhumid conditions (Quantin, 1992).
In contrast, smectite (beidellite) was a stable clay mineral along with halloysite
(1.0 dehydrating to 0.7 nm) in tropical regions with a pronounced dry season.
Similarly, the presence of 1:1 – 2:1 mixed-layer clays (halloysite-smectite and
kaolinite-smectite) has been shown in basaltic volcanic ash of the tropical
and subtropical areas (Quantin et al., 1988; Delvaux et al., 1990a, b, 1992; Delvaux
and Herbillon, 1995).
VII. SELECTED CHEMICAL CHARACTERISTICS

OF VOLCANIC SOILS
A. ORGANIC MATTER ACCUMULATION
Accumulation of organic matter is a characteristic property of Andisols. Soils

formed in volcanic materials contain the largest accumulations of organic
carbon among the mineral soil orders (Eswaran et al., 1993). Due to the high
productivity of volcanic soils, there is typically a large annual input of detritus
to the soil organic matter pool. In the case of grassland vegetation, a large
component of detritus is incorporated directly into the mineral soil horizons
(Shoji et al., 1990b). Once in the soil environment, organic matter preservation
results from burial of soils by repeated additions of volcanic ash, chemical
interactions with polyvalent cations and noncrystalline inorganic materials (e.g.,
allophane, imogolite, ferrihydrite), and physical protection from the micro-
aggregation that these materials impart to soil structure. The 14C ages of humic
acids extracted from A horizons of Andisols ranged from modern to 30,000
YBP, with the majority in the range 1000– 5000 YBP (Inoue and Higashi,
1988). It was concluded that mean residence times of organic carbon in
Andisols were appreciably greater than for Mollisol A horizons and Spodosol
Bh horizons.
Burial of soils by addition of volcanic ash is an important factor contributing
to subsurface accumulation of organic matter (Fig. 3). Burial by tephra brings
organic materials into intimate association with mineral components, reduces
aeration and decreases interactions with soil macro- and mesofauna whose
processing greatly enhances decomposition rates. Organic-rich buried soil
horizons may be distinctly evident in polygenetic soils for several thousand years
following burial (Saigusa et al., 1978; Shoji et al., 1993a).
Organic matter stabilization may occur through the formation of Al/Fe –
humus complexes and sorption to allophane, imogolite, and ferrihydrite. Torn
et al. (1997) found that soil organic carbon accumulation was related to
concentrations of noncrystalline materials along a 4-million year soil
chronosequence formed on basaltic lava in Hawaii. Soil organic matter pools
reached a maximum after 150,000 years and then decreased as noncrystalline
minerals declined and more stable crystalline minerals accumulated. Specifically,
allophane was shown to have a protective effect on soil organic matter as
demonstrated by decreased organic C mineralization rates in volcanic soils
(Zunino et al., 1982; Boudot et al., 1988, 1989). Noncrystalline inorganic
components have a strong affinity for sorption of degradation enzymes and
organic matter substrate (Wada, 1977; Tate and Theng, 1980). It is generally
accepted that sorption reactions provide a mechanism for stabilization of organic
matter against biodegradation. The protective capacity of soil clays appears to be
more a function of the surface area available for sorption of the organic matter
than the actual amount of clay present (Baldock and Nelson, 2000). Thus, the
high specific surface areas of noncrystalline constituents in volcanic soils would
be expected to have a high capacity to stabilize soil organic matter. These
constituents are also responsible for the high phosphorus sorption capacity that
may further hinder decomposition by contributing to phosphorus deficiency of
microorganisms (Brahim, 1987).
Other studies show that organic carbon concentrations in Andisols are more
strongly associated with metal – humus complexes than with concentrations of
noncrystalline materials. Accumulation of organic C in the upper 35 cm of
Andisols displays a strong linear relationship with pyrophosphate-extractable
Al þ Fe (Inoue and Higashi, 1988). Similarly in New Zealand soils, allophane
concentrations were unrelated to soil carbon, however, the content of pyropho-
sphate extractable Al correlated strongly with soil carbon concentrations (Percival
et al., 2000). Complexation of multivalent cations (e.g., Al3þ and Fe3þ) by humic
substances results in organic bearing functional groups becoming more con-
densed and less susceptible to biological attack (Baldock and Nelson, 2000).
With increasing acidity, organic matter may be protected against biodegradation
by Al toxicity to microorganisms (Tokashiki and Wada, 1975).
Organic matter may be physically protected from microbial and enzyme
attack by soil structural conditions that limit the accessibility of soil organic
matter to decomposers (Baldock and Nelson, 2000; Gregorich and Janzen, 2000).
Volcanic soils have an abundance of microaggregates resulting from the occur-
rence of noncrystalline materials with variable charge surfaces (Warkentin et al.,
1988). In a review of the literature, Baldock and Nelson (2000) reported that
physical protection might result from the ability of clays to encapsulate organic
materials (Tisdall and Oades, 1982), the burial of organic C within aggregates
(Golchin et al., 1994, 1997), and the entrapment of organic C within small pores
(Elliott and Coleman, 1988). For example, Kilbertus (1980) indicated that
bacteria can only enter into pores with diameters greater than about 3 mm. Thus,
a significant fraction of soil organic matter in volcanic soils may be inaccessible
to decomposing organisms. Due to the high water retention by microaggregates
in volcanic soils, it is also plausible that anaerobic conditions may hinder
decomposers in the aggregate interiors, even under otherwise well-aerated
conditions. Breakdown of microaggregates, such as by tillage operations, may
expose a large fraction of the physically protected organic matter to
microorganisms resulting in a rapid loss of organic matter.
B. ALUMINUM DYNAMICS
Processes regulating aluminum solubility and release/retention kinetics play

an important role in determining soil mineralogy (i.e., formation of metastable
noncrystalline materials) and crop productivity (i.e., Al toxicity). Studies
examining Al solubility and Al dissolution kinetics in volcanic soils with a
large fraction of active Al (i.e., oxalate-extractable Al) show fast reaction kinetics
and rapid attainment of an apparent equilibrium (Dahlgren and Walker, 1993;
Dahlgren and Saigusa, 1994; Takahashi et al., 1995; Takahashi and Dahlgren,
1998). In soils containing hydroxy-Al polymers (interlayered 2:1 layer silicates),
allophane/imogolite, Al – humus complexes, and exchangeable Al, there appears
to be a simultaneous equilibrium existing between all four phases (Fig.
9;Dahlgren and Walker, 1993). The rapid Al release/retention associated with
exchangeable Al and Al –humus complexes results in apparent equilibrium from
over- and under-saturation within a 20 min residence time (Dahlgren et al.,
1989). It appears that hydroxy-Al polymers ultimately regulate Al solubility in
this case and that they have a solubility similar to that of synthetic gibbsite,
Figure 9 The active Al fractions in many volcanic soils appear to form a simultaneous equili-
brium between the various solid-phase constituents. Rapid Al dissolution kinetics for exchangeable Al
and Al–humus complexes allows soil solutions to quickly attain an apparent equilibrium.
which has a log K *SO of 8.1 for the reaction: Al(OH)3 þ 3Hþ ¼ Al3þ þ 3H2O
(May et al., 1979). Experimental data showed a slope of about 2.7, compared to
the theoretical value of 3.0 for gibbsite. A slope less than 3.0 may indicate that the
formula for hydroxy-Al polymers is AlðOHÞ0:3þ 2:7 , the positive charge satisfying
the negative charge of the 2:1 layer silicate (Huang, 1988).
Based on the assumption that hydroxy-Al polymers control Al3þ activities as a
function of Hþ activities, the silica activities of soil solutions will be buffered by
allophane/imogolite precipitation and dissolution:
hskip26pt2AlðOHÞ3 þ H4 SiO4 ¼ Al2 SiO3 ðOHÞ4 þ 3H2 O Log K ¼ 4:1
hydroxy-Al imogolite
polymer
In fact, soil solutions collected from horizons having both hydroxy-Al inter-
layered 2:1 layer silicates and imogolite show remarkably consistent pH4SiO4
values between 4.0 and 4.1 (Farmer, 1987; Ugolini et al., 1988; Dahlgren and
Ugolini, 1989b; Dahlgren et al., 1991). This suggests that imogolite formation
may be limited by the supply of H4SiO4 in these soils because Al for synthesis of
imogolite would be available from dissolution of interlayer Al(OH)3.
In humus-rich horizons, the Al3þ activities are significantly lower than that
predicted for hydroxy-Al polymers (Dahlgren and Ugolini, 1989b; Takahashi
et al., 1995). These studies suggest that exchange of Al3þ with humic substances
(Al –humus complexes) controls the relationship between Al3þ and pH. In this
case, the degree of Al3þ saturation of carboxyl groups on humic substances will
determine the pAl versus pH solubility relationship (Cronin et al., 1986).
The kinetics of Al release from soils containing a variety of active Al forms
shows that Al release rates are rapid from both allophanic and nonallophanic
Andisols (Fig. 10; Dahlgren and Saigusa, 1994; Takahashi et al., 1995). To
determine the source of Al released, soils were chemically treated with KCl,
pyrophosphate and acid oxalate to extract various pools of active Al. Removal of
KCl extractable Al resulted in a large decrease in Al release rates for
nonallophanic Andisols, but an increase in release rates for allophanic Andisols.
Decreased release rates observed in the nonallophanic Andisols are explained by
the removal of a large pool of exchangeable Al. In contrast, the increased Al
release rates observed in the allophanic Andisols are believed to be due to the
formation of an easily dissolvable, polymeric-Al surface precipitate following the
mechanism of “induced hydrolysis” outlined by Wada (1987a, b). This
mechanism suggests that adsorbed forms of Al are released to solution in the
exchange process with Kþ. In the case of the allophanic Andosols, the most
probable source of exchangeable Al is from weakly held Al – humus complexes
because these soils do not contain detectable levels of 2:1 layer silicates.
Following the release, Al undergoes hydrolysis and releases Hþ that is adsorbed
on the surface of variable charge materials. The increased ionic strength due to
Figure 10 Aluminum release rates at pH 3.3 for nontreated soil samples and residues following
selective dissolution treatment for a Bw horizon from allophanic and nonallophanic Andosols.
Samples were treated with KCl, pyrophosphate, and acid oxalate to remove various Al pools for
determining their affect on the overall Al dissolution rates (Dahlgren and Saigusa, 1994).
the 1 M extracting solution drives the reaction by increasing surface charge on

variable charge constituents resulting in a proton sink for the Hþ released by Al
hydrolysis, thereby buffering solution pH. Saturation indices were near
equilibrium to slightly supersaturated with respect to synthetic gibbsite indicating
that the precipitation of displaced Al could have occurred. These precipitates are
most probably adsorbed to the soil surfaces making them readily available for
dissolution by acidic solutions.
Removal of Al – humus complexes with pyrophosphate reagent resulted in a
substantial decrease in Al release rates for both allophanic and nonallophanic
Andosols. This suggests that Al –humus complexes are a labile source of dissolved
Al, especially for the nonallophanic Andosols. Further treatment with acid oxalate
had little effect on Al release from nonallophanic Andisols; however, Al release
from allophanic Andisols decreased to very low rates. This suggests that allophane
and imogolite are an important source of dissolved Al in allophanic Andosols.
VIII. PRODUCTIVITY AND MANAGEMENT

OF VOLCANIC SOILS
Volcanic soils are among the most productive soils in the world as suggested
by their high human carrying capacity and high agronomic potential (Leamy,
1984; Shoji et al., 1993a). In areas surrounding Mt. Merapi volcano, central Java,
Indonesia, the most fertile soils are juvenile volcanic soils (800 to more than 1000
people per km2), followed by the youngest ash deposits (more than 400 people
per km2), and then areas without ash deposits (245 people per km2) (Mohr, 1938).
In contrast, volcanic soils in Japan are sometimes regarded as one of the most
infertile agricultural soils. Differences in the agricultural productivity among
Andisols are largely attributed to the colloidal composition in the rooting zone,
namely allophanic versus nonallophanic. While nonallophanic Andisols share
many common properties with allophanic Andisols, they display some distinctive
properties due to the presence of 2:1 layer silicates (Saigusa et al., 1991; Saigusa,
1992). In particular, the strong acidity, high exchangeable Al content, and low
exchangeable base concentrations of nonallophanic Andisols lead to severe
aluminum toxicity in sensitive crops (Saigusa et al., 1980; Shoji et al., 1980a, b).
Therefore, from a soil management perspective, Andisols should be divided into
allophanic and nonallophanic categories.
A. CHARGE CHARACTERISTICS AND CHEMICAL FERTILITY
The different charge characteristics between allophanic and nonallophanic

Andisols is the most important factor regulating chemical fertility attributes
(Table III) (Shoji et al., 1985; Saigusa, 1991; 1992). Allophanic Andisols are
dominated by variable charge constituents (allophane/imogolite), while non-
allophanic Andisols display both variable charge and constant charge
components (2:1 layer silicates and Al –humus complexes). The CEC and
AEC of variable charge components depend on pH and the ionic strength of the
soil solution (Okamura and Wada, 1983). The variable positive charge results
from protonation of surface Al– OH, Fe –O and Fe – OH groups, while the
variable negative charge results from dissociation of surface Si-OH and organic
functional groups (e.g., carboxylic). Charge characteristics of Andisols in
northern Japan were fully investigated with special reference to their colloidal
Table III
Comparison of Selected Soil Properties Between Allophanic Andisols and Nonallophanic
Andisols
Property Allophanic Andisols Nonallophanic Andisols
Major clays Allophane and imogolite Hydroxy-Al interlayered

2:1 layer silicates
Source of active Al Allophane and Al–humus Al –humus complexes
complexes
Constant negative charge None High
Variable positive charge Some None
Soil acidity Weakly acid Strongly acid
KCl-extractable Al Low High
Al-toxicity Rare Common
Consistency–LL, PL, PIa High Low
Stickiness None-slightly Moderately
Plasticity None-slightly Moderately
Compaction potential Slight Moderate
a
LL is liquid limit, PL is plastic limit; PI is plasticity index.
composition (allophanic and nonallophanic Andisols, with or without organic

matter enrichment) (Saigusa et al., 1992b). The development of negative charge
with increasing soil pH was common to all Andisols and was strongly related to
the amount of soil organic matter. Soils with a large variable charge component
required large additions of lime for pH amendment, and were susceptible to
leaching of cations when the soil pH decreased. Some constant negative charge
was observed in nonallophanic Andisols with or without organic matter and
contributed to aluminum toxicity due to the presence of exchangeable Al.
Positive variable charge increased with decreasing pH in Andisols having low
organic matter, while Andisols with higher organic matter concentrations showed
virtually no positive charge, even when soil pH decreased. AEC values for
Andisols were generally less than 3 cmolc kg21 at pH values greater than 5.
1. Nitrogen Dynamics
Nitrogen accumulation and mineralization. Nitrogen is the major nutrient

limiting establishment of plants in thick deposits of volcanic materials following
an eruption. Plants are able to rapidly regenerate in locations where their roots
can tap nitrogen pools in the buried soil or where erosion exposes the buried soil.
The thick pyroclastic flows surrounding Mt. St. Helens were colonized by
nitrogen fixing lupine (Lupinus spp.). Individual lupine plants created “islands of
soil fertility” by fixing nitrogen, incorporating organic matter into the soil, and
trapping nutrient-rich eolian materials. Alder (Alnus spp.), another common

nitrogen fixer and pioneer species, has also been shown to contribute large
amounts of nitrogen to disturbed ecosystems (Ugolini, 1968). These ecosystem
pioneering species play an important role in revegetation of terrestrial ecosystems
through their role in fixing nitrogen, enhancing chemical weathering and
incorporating organic matter into the soil to initiate nutrient cycling.
Nitrogen is generally the most important nutrient determining yield and quality
in agricultural ecosystems. Andisols often accumulate large amounts of nutrients
in their organic-rich A horizons. For example, a 50 cm thick A horizon has about
15 Mg N ha21 and 5.5 Mg P ha21 (Shoji et al., 1993a). While the N pool greatly
exceeds the requirement of commonly cultivated crops (~200 kg ha21), it is the
mineralization rate during the growing season that determines the amount of soil
nitrogen available to plants. Under intensive agricultural production in Japanese
Andisols, about 80 to 110 kg N ha21 was made available to crops through
nitrogen mineralization (Saigusa et al., 1983; Shoji et al., 1986). Therefore,
mineralized nitrogen plays an important role in supplying N to crops and may be a
source of nitrogen leaching to subsoil and ground water.
Incubation methods are commonly used to estimate potentially mineralizable
N in Andisols (Sugihara et al., 1986; Fridrik et al., 1996; Parfitt et al., 2001).
These studies find that the amount of nitrogen released by mineralization is
highly correlated with the amounts of soil organic matter, especially the fraction
defined as the easily decomposable organic matter pool (Okano, 1990; Sakamoto
and Oba, 1993). The following three equations, representing first-order kinetic
reactions, were proposed for estimating nitrogen release patterns in Andisols
under aerobic (simple model) and anaerobic conditions (two source model)
(Sugihara et al., 1986):
1. Simple type model (for upland soils):

N ¼ N0 [1 2 exp(2 kt)]
2. Two-source type model (for paddy soils):
N ¼ N01 [1 2 exp(2 k1t)] þ N02 [1 2 exp(2 k2t)]
3. Simple type model combined with immobilization:
N ¼ Nim [1 2 exp(2 kimt)] þ N0 [1 2 exp(2 kt)] þ C
where, N is the amount of mineralized nitrogen (mg N kg21 soil); N0 is the

mineralization potential; N01 and N02 are the nitrogen mineralization potentials
for easily and slowly decomposable fractions, respectively (mg N kg21 soil);
Nim is the amount of immobilized nitrogen (mg N kg21 soil); k, k1, and k2 are
the mineralization rate constants; kim is the immobilization rate constant
(day21) and t is time (days).
Stanford and Smith’s (1972) original kinetic equations were modified into
Equation (1) for practical use under field conditions, Equation (2) for the existence
of two different nitrogen sources (easily and slowly decomposable nitrogen
sources), and Equation (3) for the coexistence of both mineralization and
immobilization.
N mineralization parameters for Andisols and nonandic soils from north-
eastern Japan were determined by Saito (1990) using Equation (1). The nitrogen
mineralization potential (N0) in Andisols was considerably greater than for
nonandic soils (176 versus 100 mg N kg21 soil); however, mineralization rate
constants (k) were smaller in Andisols than in nonandic soils (0.0043 versus
0.0059 d21) incubated under aerobic conditions. As a result, the percentage of
mineralizable N (N0/Nt values) from Andisols was less than 50% that of nonandic
soils (3.5 versus 8.2%). Retardation of N mineralization by allophane and Al3þ
ions in Andisols was suggested in comparison with nonandic soils (Parfitt et al.,
2001). The incubation approach was very effective in predicting nitrogen
mineralization under field conditions of corn cultivation (Saito and Ishi, 1987). A
similar kinetic analysis conducted for Typic Hapludands showed that the k value
of reduced tillage was greater than that of conventional tillage (Nira and
Nishimune, 1993).
In the two-source model (paddy field), the N mineralization potential for the
easily decomposable component (N01) was influenced strongly by soil drying
before puddling. The easily decomposable pool was important for supplying soil
N for rice during the early growing stages. On the other hand, N mineralization
from the slowly decomposable component (N02) was the primary source of N
during the middle- and late-growth stages. The easily decomposable pool was
much smaller than the slowly decomposable pool in Andisols compared to
alluvial soils (Ando and Shoji, 1986; Shoji et al., 1993a).
Nitrogen leaching in Andisols. Several investigations on adsorption and leaching

of nitrate in Andisols conclude that nitrate mobility is affected by soil pH, organic
matter, anion exchange capacity, competing anions, and the point of zero charge
(Reynolds-Vargas et al., 1994; Babbar and Zak, 1995; Kamewada, 1996; Katou
et al., 1996; Qafoku et al., 2000; Vaje et al., 2000; Ryan et al., 2001). The
development of positive charge in both allophanic and nonallophanic Andisols is
depressed by the accumulation of organic matter in surface horizons (Saigusa
et al., 1992b). In addition, the pH of surface horizons is generally increased by
addition of lime to obtain a recommended pH of 6 – 7 for most crops.
Consequently, the nitrate adsorption capacity in the Ap layer of Andisols is
very low and thus NO2 3 is freely leached to the subsoil and eventually to ground
water after heavy rain (Saigusa, 1988; 1991). Exchangeable NO2 3 concentrations
reported for volcanic soils receiving heavy nitrogen fertilization can be as high as
1– 2 cmol kg21 (Reynolds-Vargas et al., 1994; Kamewada, 1996); however,
values less than 0.1 cmol kg21 are more common for soils receiving moderate N
fertilization.
A critical factor regulating N-use efficiency between allophanic and non-
allophanic Andisols is the depth of the rooting zone. In fields of winter-planted
barley and wheat, nitrogen applied as (NH4)2SO4 was easily converted to

NO2 3 within two weeks in both allophanic and nonallophanic Andisols. The
downward movement of NO2 3 was 16 – 19 cm per 100 mm of rainfall in autumn
when evapotranspiration was low (Saigusa et al., 1983; 1991). Because
nonallophanic Andisols have strongly acidic subsoil horizons, roots of Al
sensitive crops, such as barley, are restricted to the topsoil layer that has received
lime. In contrast, rooting depth in allophanic Andisols was not restricted by Al
toxicity leading to a much greater depth of rooting. Consequently, N fertilizer
recovery was less than 10% on nonallophanic Andisols compared to 30– 50% on
allophanic Andisols. Barley grown on nonallophanic Andisols showed nitrogen
deficiency symptoms and yields were less than half of that obtained in allophanic
Andisols.
Evaluation of nitrate sorption potential of soils at local and landscape scales is
useful for nitrogen management to reduce ground water pollution (Ryan et al.,
2001). In the Central Valley of Costa Rica, the fertilizer used for coffee plantation
was suspected to be a cause of ground water nitrate pollution. Differences in the
nitrate adsorption characteristics and leaching were related to clay mineralogy.
Andisols dominated by Al-rich allophane (Al:Si, 2:1) showed significant nitrate
adsorption (about 15 mg N g21) and thus retarded nitrate leaching compared to
negligible nitrate adsorption in Andisols containing Si-rich allophane (Al:Si, 1:1)
21
and halloysite. They estimated that about 300 kg NO2 3 -N ha was retained on
the AEC in the 4 m soil profile containing Al-rich allophane. Significant nitrate
adsorption in allophanic Andisols retards leaching of nitrate to ground water by
retaining NO2 3 within the rooting zone allowing a greater time for plant uptake.
However, the effectiveness of NO2 3 sorption in coffee plantations is limited
because of high inputs of nitrogen fertilizer and liming (i.e., reduction of positive
charges) (Reynolds-Vargas et al., 1994).
Improvement of fertilizer nitrogen-use efficiency in Andisols. Because of the low

NO2 3 retention characteristics of Andisols, N-use efficiencies of readily soluble
fertilizers (e.g., (NH4)2SO4) are considered to be low, especially in nonallophanic
Andisols because of their strongly acidic subsoil. To improve the nitrogen-use
efficiency of applied fertilizer nitrogen, several approaches were examined in
nonallophanic Andisols: top-dressing versus split application of basal nitrogen
fertilizers, use of controlled release nitrogen fertilizers, and band versus nested
placement of fertilizer (Saigusa et al., 1991; Saigusa, 1999).
Among these methods, a split application of basal rapidly available nitrogen
fertilizer and both basal and top dressings of controlled release nitrogen fertilizers
were verified to be the most effective in improving the productivity of barley in
strongly acidic nonallophanic Andisols (Saigusa et al., 1983; 1991). The N-use
efficiencies of readily available nitrogen applied as top dressing, especially that
applied at later growing periods were much higher than those applied as basal
fertilizer nitrogen because of the intensive root development in the surface soil
and increased nutrient absorbing ability with age of the crops. Thus split
application of basal readily available nitrogen fertilizer increased the N uptake of
winter barley by 28% and grain yields by 38% in nonallophanic Andisols
(Saigusa et al., 1991).
The nutrient release pattern of some controlled release fertilizers may be
synchronized with the growth rates of the crops to enhance N-use efficiencies. For
example, polyolefin coated ureas can be formulated to release N over a specified
time period: POCU-70 releases . 80% of its nitrogen within 70 days at 25 8C.
POCU-70 applied as a single basal application to silage-corn resulted in 49 – 66%
N-use efficiency of applied N compared to 30– 44% for (NH4)2SO4 (Gandeza
et al., 1991; Shoji et al., 1991). Greatly enhanced N-use efficiencies were also
reported for sorghum (Saigusa et al., 1990), oats and barley (Ito et al., 1998), and
tender green mustard (Ombodi and Saigusa, 2000). Similarly, a single basal
application of POCU-100 (sigmoid-type) applied directly to nursery boxes with
rice seeds (co-situs application) improved N-use efficiency by three times while
also reducing the labor cost associated with fertilization (Saigusa et al., 1996).
Mixtures of rapid and slower release POCU fertilizers may be formulated to
provide a pulse of nitrogen during the early growth stages and sustained nitrogen
availability throughout the later growth stages of crops. Therefore, the use of
controlled release fertilizers for crop production in Andisols is an advantageous
method that saves labor and maintains nutrient availability to plants during their
late-growth stages. Additionally, the increased N-use efficiency reduces
environmental impacts associated with NO2 3 leaching to ground water and
N2O emissions, a potent greenhouse gas.
2. Phosphorus Dynamics
Phosphorus is often a growth-limiting nutrient for agricultural crops grown on

Andisols (Alvarez et al., 1999; Manske et al., 2000). Therefore, heavy application
of phosphorus fertilizer to Andisols is commonly practiced in developed
countries to maximize yields. Phosphorus availability decreases with increasing
soil development due to incorporation of P into organic forms and P fixation by
active Al and Fe components (e.g., allophane, imogolite, Al –humus complexes).
In contrast, P availability in young volcanic materials is relatively high due to
rapid weathering of apatite, and low phosphate retention due to low
concentrations of active Al and Fe.
Fresh volcanic ash contains an appreciable amount of P extractable by dilute
acid, Troug or Bray methods. Troug-extractable P in rhyolite and andesite tephras
(0.1 – 0.6 g P2O5 kg21) was much higher compared to basaltic andesite and basalt
tephras (, 0.1 g P2O5 kg21) (Nanzyo et al., 1997). The source of extractable P
was attributed to apatite, the primary phosphorus-bearing mineral in volcanic
materials. Dissolution rates of apatite increased with decreasing pH and

increasing concentrations of chelates (Nanzyo et al., 1999; Nakamaru et al.,
2000a). Although, the content of apatite is usually less than 0.3% in fresh
volcanic materials, it can provide an important source of P in natural ecosystems
and agricultural production. For example, pigeon pea and chick pea were grown
in fresh Mt. Pinatubo, Philippines volcanic ash containing 1.7 g P2O5 kg21,
primarily as apatite. Dry matter production of pigeon pea was similar in
plots with and without added P, and both pigeon pea and chick pea
enhanced P availability through production of organic acids and chelates
(Nakamaru et al., 2000b).
While mature Andisols commonly have large amounts of active Al, the form
of active Al is different between allophanic and nonallophanic Andisols (Table
III). The major components responsible for phosphorus sorption in allophanic
Andisols are allophane/imogolite and Al– humus, while Al –humus complexes
are the dominant components in nonallophanic Andisols. The amount of
phosphate adsorbed per mole of active Al was higher in nonallophanic Andisols
(0.13 mol) compared to the allophanic Andisols (0.09 mol) (Saigusa et al., 1991;
Matsuyama et al., 1999a). As a result, P availability is generally higher for
allophanic Andisols. In fact, the amount of P absorbed by barley, buckwheat,
sorghum, and dent corn was larger in allophanic than in nonallophanic Andisols,
even though their active Al and Fe concentrations were similar (Mastuyama et al.,
1994; Ito et al., 2000a).
Phosphorus fertilizer applied to Andisols is rapidly sorbed to active Al and Fe
components and changes to less available forms with time (Nanzyo, 1987). Many
soil tests for estimating available soil P have been evaluated for volcanic soils.
The Truog method is most popular in Japan, followed by Bray II, Mehlich No. 3,
Olsen, and anion exchange resin methods. Each method has proved useful for
estimating the amount of plant-available P, but only when comparisons are made
among soils with similar colloidal composition (i.e., allophanic versus
nonallophanic). Therefore, it is difficult to compare results between allophanic
and nonallophanic Andisols.
Due to the large accumulation of organic matter in many volcanic soils,
organically bound P is often a large pool (about 5.5 Mg ha21 in Andisols with
thick A horizons; Shoji et al., 1993a). Organic P in uncultivated Andisols in Chile
was strongly correlated with organic carbon content (Org. P [mg kg21] ¼ 59
organic C[%] þ 133; r 2 ¼ 0.83) (Escudey et al., 2001). Organic P is considered
an important P source for crops through mineralization. However, measuring P
mineralization rates by simple incubation methods is difficult in Andisols because
PO4 released by mineralization is rapidly sorbed by active Al and Fe. Adding
anion exchange resins or membranes to incubating soil was shown to effectively
capture P released from mineralization, preventing it from being sorbed to active
Al and Fe components (Zou et al., 1992; Parfitt et al., 1994). P mineralization
studies using anion exchange methods showed rates of 1.3 mg P kg21 d21 for
Hawaiian forest Andisols (Zou et al., 1992) and 0.2 –0.5 mg P kg21 d21 for
Japanese Andisols (Ito et al., 2000b). These results confirm that P mineralization
is an important P source for plants growing in Andisols.
Improvement of P uptake and plant growth by arbuscular mycorrhizae (AM)
under conditions of low available phosphorus has been well documented.
Improved P uptake and plant growth were demonstrated in Andisols following
AM inoculation of onion (Botello et al., 1987; Palacios et al., 1987), onion and
white clover (Tawaraya et al., 1995), and kidney bean (Isobe and Tsuboki, 1997).
Inoculation increased both spore number and the grain weight of kidney bean at a
soil-available P (Bray No. 2) level of 10 mg P kg21, while these effects were
suppressed when soil-available P was 25 mg P kg21.
Indigenous AM fungi predominate over inoculated species in most soils. The
effectiveness of using indigenous AM fungi was investigated in several cropping
systems (Arihara et al., 2000; Karasawa et al., 2000a). Soils contained more
AM spores following cultivation of mycorrhizal crops (e.g., sunflower, maize,
soybean, potato kidney bean, adzuki bean, and wheat) than after cultivation of
nonmycorrhizal crops (mustard, radish, sugar beet and buckwheat). As a result
there was increased growth of succeeding crops following mycorrhizal crops. This
effect on maize growth and phosphorus uptake was evident in dry soil conditions
and was less pronounced with increasing soil moisture (Karasawa et al., 2000b).
3. Other Mineral Nutrients
Generally, volcanic ash soils are formed by intermittent tephra deposition and
thus surface soils are often regenerated by addition of fresh volcanic ash. This
means that Andisols are relatively young soils and the soil fertility is closely
related to the properties of the parent tephra. The chemical composition of tephra
varies depending on the chemical composition of the magma. There is generally a
close relationship between the SiO2 content and concentrations of other elements
(Shoji et al., 1975; Kobayashi et al., 1976; Yamada, 1988). In general, Al, Fe,
Mg, Ca, Mn, Ti, Cu, and Co are most abundant in mafic (basic) tephras, whereas
Na and K tend to be most abundant in felsic (acidic) tephras. Greenhouse studies
examining fertility of Andisols formed from different tephra chemistries showed
that the amounts of N, Ca, Mg, and Cu absorbed by plants were greater in mafic
than felsic materials, and only K uptake was greater in felsic materials (Saigusa
et al., 1976; Yamada and Shoji, 1980). Thus, the nutrient supplying power of
elements, except K, was stronger in the Andisols derived from mafic materials.
Intermittent deposition of tephra on soils has been shown to affect soil fertility
(Dahlgren and Ugolini, 1989a; Cronin et al., 1997; 1998; Dahlgren et al., 1999).
Under cool – humid weathering conditions in the Cascade Mountains, Washington,
a 5 cm layer of fresh Mt. St. Helens tephra (1980 eruption) applied to the surface
of a Spodosol released appreciable concentrations of nutrient cations and sulfur

over a 4 year period of investigation (Ca ¼ 217; Mg ¼ 31; K ¼ 50 and
S ¼ 69 kg ha21 over 4 years; Dahlgren et al., 1999). Release rates showed a
decrease with time as soluble salts and easily weatherable materials were depleted
(Fig. 11). The base cations were preferentially retained on cation-exchange sites in
the buried organic horizons as the HCO3 anion was lost to protonation (HCO2 3 þ
Hþ ¼ H2CO3) (Fig. 8). Sulfur (as sulfate) was retained by sorption within
B horizons (Fig. 12). The high anion sorption by noncrystalline materials in the
B horizons resulted in virtually no enhanced losses of sulfate from the C horizons.
Thus, release of nutrients from intermittent airfall tephra additions provides an
important source of slowly available nutrients that helps to sustain forest
productivity in these strongly acidic soils. The large pulse of base cations from the
leaching of soluble salts contained in the fresh tephra resulted in displacement of
high concentrations of Hþ and Al3þ that may result in short-term Al toxicity. The
high Hþ and Al3þ concentrations returned to ambient levels within about 6 months
of the tephra addition.
Similarly, eruptions from Mt. Ruapehu, New Zealand in 1995 –1996 added
between 30 and 1500 kg S ha21 (Cronin et al., 1997; 1998). As little as 0.25 mm
of ash caused pale yellow S-deficient ryegrass to “green” within two days of
tephra application. Smaller, but agronomically significant amounts of K, Mg,
Na, Se, B, and Co were also leached from the ash deposits. Ash addition caused
a short-term depression (, 50 d) of soil pH (, 4.5) due to the oxidation
of elemental sulfur, which represented 55 –77% of the total sulfur content
Figure 11 Nutrient release rates for a 5 cm ash layer for the first four years following addition of
ash to the surface of a Spodosol in the western Cascade Range, Washington (Dahlgren et al., 1999).
Figure 12 Soil solution charge balance for soils receiving 0, 5, or 15 cm of Mt. St. Helens ash.
Soil solutions represent the first month of leaching of the fresh ash (September) after application to the
surface of a Spodosol in the western Cascades Washington. The width of each compartment is in
proportion to the ion’s charge contribution. The anion deficit was assumed to be the contribution of
dissociated organic ligands (Dahlgren and Ugolini, 1989a).
(Cronin et al., 1998). With respect to animal health, the most common cause of
livestock poisoning results from fluorosis (Oskarsson, 1980; Arya et al., 1990).
Ingestion of tephra particles and/or forage with high fluoride levels may result in
acute or chronic fluorosis. Concentrations of F, S, and Se from Mt. Ruapehu were
high enough to reach potentially harmful levels within water and in the foliage of
plants that accumulate these elements (Cronin et al., 1998).
4. Silicon Nutrition
Silicon is generally not considered one of the essential elements for higher
plants, but is a beneficial element for certain plants such as rice, sugarcane,
cucumber, and turf. In rice production, Si is considered an “agronomically
essential element” because of its multiple functions in rice growth, such as
improving the resistance to lodging, pest and disease, reducing excessive
transpiration, and keeping leaves erect. Through these functions, silicon enhances
water-use efficiency and the photosynthesis rate (Savant et al., 1997; Saigusa
et al., 1998; 1999; Saigusa, 2000).
The silicon content of sugarcane is strongly correlated with soluble Si
concentrations (Fox et al., 1967), which in turn is strongly related to soil
mineralogy. Silicon is the most abundant element in fresh tephra ranging from
about 48 to 73% SiO2; it is highest in felsic ash and lowest in basaltic ash (Shoji
et al., 1975). Generally, volcanic soils have a high potential for supplying Si
when they are young. However, Si concentrations in Andisols decrease as
weathering proceeds in the udic soil moisture regime because volcanic glass is
consumed by weathering and Si is easily leached due to its high relative mobility
(i.e., compared to Al and Fe). Silicon is incorporated in allophane/imogolite,
opaline silica, and halloysite, and is easily adsorbed by noncrystalline materials
in Andisols. As a result, the amount of Si available for some Si accumulating
plants is not sufficient. Therefore, the application of silicate fertilizer is a
common practice to maintain the silicon supply in paddy fields, especially
Andisols and peat soils (Saigusa et al., 1998; 1999; Saigusa, 2000).
5. Soil Productivity and Acidity
Aluminum toxicity in Andisols. In acidic soils, Hþ, Mn2þ, and Al toxicities, P, Ca,
Mg and micronutrient deficiencies, and suppression of some microorganisms
have been shown to affect crop productivity. Among these factors, Al toxicity is
the most important constraint on crop growth in acidic Andisols. Among
Andisols, a significant difference in Al toxicity potential was observed between
allophanic and nonallophanic soils. Severe Al toxicity was observed in
nonallophanic Andisols containing high amounts of KCl-extractable Al derived
primarily from 2:1 layer silicates (strongly acidic constant charge), but not in
allophanic Andisols whose major clay mineral was allophane (weakly acidic
variable charge) (Saigusa et al., 1980; Shoji et al., 1980a, b). The primary
manifestation of Al toxicity is impaired root development resulting in the
formation of a shallow rooting system. The inhibition of root growth in Andisols
is closely related to the amount of KCl-extractable Al rather than soil pH(H2O).
Therefore, KCl-extractable Al is the most reliable measure of Al toxicity
potential in Andisols. Threshold values for Al toxicity were 2 cmolc kg21 for
common agricultural crops and 1 cmolc kg21 for Al-sensitive crops. A value of
2 cmolc kg21 of KCl-extractable Al was adopted for the allic subgroup of
Andisols to reflect their potential for Al toxicity (Soil Survey Staff, 1998). The
origin of KCl-extractable Al has been attributed primarily to Al3þ adsorbed to 2:1
minerals, and to a lesser extent release from Al –humus complexes (Dahlgren and
Saigusa, 1994; Takahashi et al., 1995). Allophanic Andisols generally contain
very low KCl-extractable Al concentrations; however, these values may be
underestimated due to “induced hydrolysis” of displaced Al and subsequent
adsorption of polymeric Al to allophanic materials (Wada, 1987a, b).
Plant nutritionists generally explain plant uptake of elements by active
absorption from soil solution rather than by contact exchange. Therefore, it has
been difficult to explain Al toxicity in soils with pH . 4.5, where low
concentrations of monomeric Al3þ should not contribute to Al toxicity. At
pH . 4.5, Al3þ becomes polymerized. Polymerized Al was shown to be
extremely toxic to higher plants compared to monomeric Al3þ in solution culture.
Consequently, many plant nutritionists believed that the cause of Al toxicity at
pH . 4.5 was attributably to polymerized Al, such as the Al13 polymer
[AlO4Al12 (OH)24 (H2O)12]7þ (Bartlett and Riego, 1972; Wagatsuma and Ezoe,
1985; Shann and Bertsch, 1993). The presence of Al13 polymer was suggested in
Spodosols with extremely low pH (pH ¼ 3.5) (Hunter and Ross, 1991) and in
Andisols based on Al speciation calculations (Kato et al., 1995). A detailed study
of phytotoxic forms of Al in an Andisol examined both root growth of rice plants
under submerged soil culture and monomeric and polymeric sorption reactions to
soil materials. Polymeric Al was selectively and irreversibly adsorbed by soil
colloids and thus monomer Al3þ was concluded to be the main cause of Al
toxicity (Saigusa et al., 1995). Aluminum speciation of KCl extracts from a
variety of acidic soils using 27Al nuclear magnetic resonance (NMR) did not
detected the presence of Al13 polymers (Hiradate et al., 1998). These results
suggest that Al13 polymers were not formed in these soils, or alternatively Al13
polymers were removed from solution by adsorption and/or precipitation
reactions. As a result, most of the recent work on the mechanism of Al toxicity
and tolerance in higher plants, and the creation of Al tolerant plants by gene
manipulation have focused on monomeric Al3þ.
Allophanic Andisols generally have only trace levels of KCl-extractable Al and
pH(H2O) values near 5. Aluminumphilic plants, such as tea, grow best under
extremely acidic soil conditions. Therefore, farmers tended to apply huge amounts
of nitrogen fertilizer (. 1500 kg N ha21 added as (NH4)2SO4) to maintain high
quality of tea leaves (Okada et al., 1986; Tachibana et al., 1995; Pansombat et al.,
1997). Similar acidification of allophanic Andisols was observed in mulberry
garden (Inamatsu et al., 1991) and some upland soils in Japan (Matsuyama et al.,
1999b). Under heavy application of chemical fertilizer, allophanic Andisols
become acidified to pH(H2O) , 4.0 and large amounts of KCl-extractable Al
develop. The extreme acidity breaks down allophane and contributes monomeric
Al for uptake by tea plants (Pansombat et al., 1997). The origin of KCl-extractable
Al in highly acidified allophanic Andisols was attributed to the dissolution of
aluminum-sulfate compounds and/or Al –humus complexes.
Subsoil acidity and acidity amelioration. Subsoil horizons play an important role
in supplying both water and nutrients to crops. If subsoil acidity is severe enough
to limit root development, as in many nonallophanic Andisols, crops cannot use
water and nutrients stored in the subsoil and thus water and nutrient stress results
in decreased yield and quality of crops (Saigusa et al., 1995; Shoji et al., 1986).
Both yields and nitrogen uptake of Al-sensitive crops such as barley, sorghum,
and alfalfa were remarkably reduced in nonallophanic Andisols with strongly
acidic subsoils. In addition, nitrogen-use efficiency was decreased and nitrate
leaching was increased due to the inability of the shallow rooted crops to extract
nitrogen from subsoil horizons.
The amount of KCl-extractable Al (phytotoxic Al) decreased with increasing
soil pH(H2O), to less than 1 cmolc kg21 at pH(H2O) ¼ 6. Therefore, liming of
the plow layer is commonly conducted using various materials, such as calcitic
limestone and dolomite (Simono, 1990; Borie and Rubio, 1999; Mora et al.,
1999a; 1999b; Toda et al., 1999). However, liming of acidic subsoil horizons is
uncommon due to the difficulty and cost of delivering lime materials to subsoil
horizons. Given the importance of subsoil horizons to crop production,
amelioration of subsoil acidity in nonallophanic Andisols has been explored
through surface application of gypsum, phosphogypsum, and organic calcium
salts (Saigusa et al., 1994; Toma and Saigusa, 1997; Inoue et al., 2001). Surface
application of gypsum was much more effective at ameliorating subsoil Al
toxicity compared to application of CaCl2 or CaCO3, and this effect was
especially prominent in subsoil horizons with low humus content. The
mechanism responsible for reduction of exchangeable Al by gypsum additions
was explained as follow: (1) release of exchangeable Al into soil solution from
cation-exchange sites, (2) polymerization of monomeric Al in soil solution, and
(3) selective and irreversible adsorption of polymerized Al on 2:1 clay minerals
(Saigusa and Toma, 1997). The form of polymeric Al was not identified, but
27
Al-NMR showed that it was not the Al13 polymer (Toma et al., 1999).
Suppression of soil borne plant pathogens by aluminum. Root rot of kidney beans
(Phaseolus vurgaris L ) caused by Fusarium solani f. sp. phaseoli and potato scab
caused by Streptomyces scabies have been devastating in Hokkaido, Japan where
allophanic Andisols are widely distributed (Mizuno and Yoshida, 1993; Furuya
et al., 1996). Suppression of these soil borne plant pathogens is reported in soils
whose clay mineralogy is dominated by hydroxy-Al interlayered 2:1 minerals
(i.e., nonallophanic Andisols). In the case of root rot of kidney bean, inhibition of
macroconidal germination and disease incidence was observed in soils with
exchangeable aluminum contents of at least 0.4 cmol kg21 (Furuya et al., 1999).
Common potato scab was suppressed in soils having relatively high amounts of
exchangeable Al, or soluble Al concentrations of 0.3 mg kg21. As a result, a
single basal band application of ammonium sulfate was recommended to acidify
soils and suppress disease incidence in allophanic Andisols (Mizuno and
Yoshida, 1993; Mizuno et al., 1998b; 2000). Rice seedlings are highly tolerant to
Al toxicity and have been grown in soils having a pH range of 4.5 –5.5 to control
damping-off disease and sudden withering of young seedlings (physiological
disease) (Saigusa et al., 1992a). These examples indicate that the disease
incidence of some soil borne plant pathogens is closely related to the clay
mineralogy of Andisols, which regulates the contents of exchangeable and water
soluble Al (Mizuno et al., 1998a; Furuya et al., 1999).
B. PHYSICAL PROPERTIES AND FERTILITY
Compared to chemical properties, Andisols generally display excellent

physical properties for crop production. They typically have high water retention,
high porosity, high permeability, high friability, low bulk density, and stable
aggregates. These characteristics were originally attributed to the presence of
allophane; however, the discovery of nonallophanic Andisols forced a
reexamination of physical properties in Andisols with special reference to their
colloidal composition. Consequently, the unique physical properties common to
Andisols are attributed to the abundance of noncrystalline materials and/or soil
organic matter. The favorable physical properties of Andisols contribute to a high
rate of seedling emergence, excellent rooting characteristics, and strong
resistance to drought stress. The fundamental physical properties of Andisols
relating to soil productivity are well documented by Shoji et al. (1993a). Reviews
concerning the physical and engineering properties of Andisols have been
published by Yamanaka (1964), Maeda et al. (1977), Warkentin and Maeda
(1980), Wada (1985), Maeda and Soma (1986), Warkentin et al. (1988) and
Kimble et al. (2000).
1. Soil Texture and Structure
Soil texture is an important index to describe soil physical conditions.

However, particle-size analysis of volcanic soils is challenging due to poor
colloidal dispersion resulting from the presence of extremely stable aggregates.
Consequently, the clay content determined by standard methods is generally
much smaller than those estimated by field methods (Ping et al., 1988). Soil
Taxonomy (Soil Survey Staff, 1998) uses 1500 kPa water retention to determine
particle-size class modifiers rather than mechanical analysis. There is a near

linear relationship between the clay content determined by mechanical analysis
with careful pretreatment for dispersion (ultra-sonication and pH adjustment) and
their 1500 kPa water retention values (Shoji et al., 1993a).
Generally speaking, Andisols have high physical fertility (tilth), such as high
air and water retention, good drainage, and ease of root penetration. These
physical properties are closely related to the development of soil structure. Soil
structure in Andisols is strongly affected by soil organic matter and noncrystal-
line materials, which are generally abundant in Andisols (Wada, 1985). There is a
strong correlation between dithionite –citrate –bicarbonate extractable Al and
the degree of macroaggregation suggesting that active Al is important in the
formation of water-stable macroaggregates (Ibrahim et al., 1998). The surface
soils of uncultivated Andisols generally have moderate to strong granular
structure. However, cultivation tends to degrade granular aggregates and
promotes the formation of subangular blocky or angular blocky structure. This
conversion is due to both compaction by machinery and reduction of soil organic
matter. Subsurface horizons of Andisols tend to have mostly subangular and
angular blocky structure (Shoji et al., 1993a). In contrast, young Andisols
generally do not have any soil structure (i.e., single grained) due to low
concentrations of noncrystalline materials and organic matter to bind particles.
2. Bulk Density and Porosity
One of the most outstanding characteristics of Andisols is extremely low bulk

density (, 0.9 Mg m23). Low bulk density is attributable to the large amounts of
noncrystalline minerals and organic matter. Thus, bulk density of Andisols
decreases with increasing concentrations of active Al and Fe. Andisols showing
Alo þ 1/2Feo . 2% generally have bulk density values ranging from 0.2 to
0.9 Mg m23 (Shoji et al., 1996). Bulk density values decreased to , 0.9 Mg m23
when allophane concentrations were greater than about 5% (Shoji et al., 1993a).
Organic matter also contributes to the low soil bulk density, and the bulk density
of nonallophanic Andisols decreases with increasing organic carbon content.
Bulk densities , 0.9 Mg m23 generally occur when organic C concentrations
exceed approximately 3%.
The low bulk density of Andisols mainly reflects the development of porous
soil structure, while relatively low particle density of humus (1.4 –1.8 Mg m23)
and volcanic glass (2.4 Mg m23) further contribute (Wada, 1985). The large pore
space of allophanic Andisols is partially attributed to the intra- and interparticle
micropores of allophane, while that of nonallophanic Andisols is primarily
attributable to humus. The aggregate size of allophane clusters in perhumid Costa
Rica Andisols was studied by laser diffraction grain-sizing and was between 2
and 20 mm (Buurman et al., 1997). The size of allophane aggregates increases

with increasing allophane content. The positive charges displayed by many
noncrystalline materials are believed to play an important role in aggregate
formation (Wada, 1985). Macropores contain gravitational water or air, and
the plant root distribution corresponds strongly with the distribution of the
macropores (Hatano et al., 1988). These physical characteristics lead to soils that
are very friable when moist resulting in ease of root penetration. While Andisols
are considered relatively resilient to compaction, the pore structure can be broken
down and bulk density increased by excessive compaction associated with crop
production (Ito et al., 1997a; Dorel et al., 2000), forest harvest practices
(Allbrook, 1986; Lenhard, 1986; Meurisse, 1988; Geist et al., 1989; Cullen et al.,
1991), and grazing (Hofstede, 1995).
3. Water Retention and Plant-Available Water
Andisols generally have high porosity and thus can retain large amounts of
water. Micropores in noncrystalline materials and humus hold hygroscopic water,
mesopores retain capillary water, and macropores contain gravitational water or
air. The high water holding capacity of Andisols is attributed to large volumes of
both mesopores and micropores. The 1500 kPa water content of allophanic
Andisols was related to both the humus and noncrystalline material (N.M.)
concentrations (Ito et al., 1991):
1500 kPa water ð%Þ ¼ 1:93 £ humus ð%Þ £ 1:47 £ N:M: ð%Þ 2 0:92
ðr ¼ 0:944; n ¼ 37Þ:
Plants can absorb the water retained in mesopores and available water content
is generally expressed as either 10 kPa or 33 kPa water content minus 1500 kPa
water content. A matric potential of about 10 kPa is often considered as field
capacity because the natural moisture content of many Andisols is equivalent to
this matric potential. Plant-available water contents can be effectively estimated
from the 33 kPa water content (Saigusa et al., 1987):
Surface soil: Available H2 O ð%Þ ¼ 0:586 £ 33 kPa water content ð%Þ þ 1:66
ðr ¼ 0:902; n ¼ 28Þ
Subsoil: Available H2 O ð%Þ ¼ 0:486 £ 33 kPa water content ð%Þ þ 0:85

ðr ¼ 0:890; n ¼ 32Þ
Thus, it may be concluded that Andisols retain large amounts of plant-

available water, about 60 and 50% of the 33 kPa water in surface and subsoil
horizons, respectively.
Following the 1980 eruption of Mt. St. Helens, wheat fields in eastern
Washington (USA) recorded a bumper crop after deposition of a few centimeters
of tephra that acted as a mulch by killing weeds and conserving moisture.
Similarly, forest productivity throughout the semi-arid western United States is
enhanced by surficial deposits of volcanic ash that act as a mulch and weather
to soils with high plant-available water retention (Meurisse, 1988). In areas
receiving intermittent additions of pyroclastic materials, contrasting particle sizes
(e.g., fine ash over coarse pumice) within the profile can result in textural dis-
continuities that impede water transmission. This condition often results in
imperfect drainage that may limit root growth into subsoil horizons. Mechanical
mixing of these layers by deep plowing is often employed to improve water
movement and enhance productivity.
C. SOIL MANAGEMENT AND CONSERVATION OF VOLCANIC

SOILS
Mature Andisols are typically considered to be resistant to water erosion

because of their high infiltration rates and stable soil aggregates (Shoji et al.,
1993a). However, immature soils developed on recent pyroclastic materials do
not have stable aggregates and are highly susceptible to water erosion. Even well-
developed Andisols are highly susceptible to both wind and water erosion when
surface vegetation is removed because of their low bulk density (Kimble et al.,
2000). Due to the serious consequences of soil degradation from erosion, several
investigations have examined methods to suppress erosion in Andisols (Perret
et al., 1996; Perret and Dorel, 1999; Poudel et al., 1999; Poulenard et al., 2001).
Runoff and soil erosion on Andisols of northern Ecuador were investigated
using rainfall simulations with special reference to soil maturity and land use.
Noncultivated, mature Andisols showed high infiltration rates and low sediment
loss, whereas cultivated, mature Andisols, Andisols after burning, and young
Andisols were prone to crusting, surface runoff and erosion (Poulenard et al.,
2001). Compared to conventional tillage practices, contouring, strip cropping,
and contour hedgerow cultivation were effective in suppressing erosion under
intensive farming systems on sloping topography in the Philippine (Poudel et al.,
1999). Agro-ecological practices, such as maintaining plant cover and hedging
along field boundaries, were effective for soil rehabilitation and erosion control
on Andisols in Reunion Island, Indian Ocean (Perret et al., 1996; Perret and
Dorel, 1999). Thus, with proper soil conservation practices, it is possible to
minimize soil erosion in Andisols under intensive cropping systems.
No-tillage or reduced tillage practices are highly recommended to maintain
surface vegetation or crop residues to suppress soil erosion in Andisols. In
addition to erosion control, these cropping systems have many beneficial aspects
that contribute to environmental protection and sustainable agriculture:
† Suppressed soil erosion by maintaining surface cover (vegetation and residues).

† Decreased cost for tillage, energy savings, and reduced CO2 emission.
† Increased water holding capacity and infiltration, contributing to improved
germination and drought resistance (Tsuji et al., 2000).
† Recycles plant biomass, which enhances nutrients for future crops.
† Accumulates organic matter in surface horizons, which maintains a stable
temperature (mulching effect) and reduces bird damage and weed infestations.
† Increased lodging resistance by improved rooting, and thus reducing harvest
loss and improving crop yield and quality.
† Maintains a firm soil surface, and thus allows timely field operations, such as
transplanting or harvesting even after rain.
† Improves oxidative conditions (i.e., gas exchange) in paddy soils, and thus
reduces hydrogen sulfide damage to rice roots.
Some negative attributes of no-tillage and reduced tillage practices on

Andisols are increased herbicide requirements, soil compaction, and decreased
fertilizer-use efficiency. The most serious problem arises from low fertilizer-use
efficiency, especially in paddy soils. Controlled release fertilizers with nutrient
release patterns synchronized to crop nutrient demands resulted in greatly
improved fertilizer-use efficiency for no-tillage cultivation of silage-corn and rice
in Andisols (Saigusa et al., 1996; Ito et al., 1997b; Inoue et al., 2000). After five
successive years of silage-corn production, Ito et al. (1997b) concluded that a no-
tillage system with a single basal application of polyolefin coated urea was highly
effective in terms of fertilizer-use efficiency and yield optimization in
nonallophanic Andisols. Similarly, no-tillage transplanting of rice in nonallo-
phanic Andisols with a single basal application of polyolefin-coated urea directly
in the nursery box resulted in 80% nitrogen-use efficiency (Saigusa et al., 1996).
This compares to less than 30% nitrogen-use efficiency using basal ammonium
sulfate applied in a conventional tillage transplanting system. The recycling of
rice straw under no-till greatly suppressed CH4 emissions from paddy soils and
did not result in any reductive damage of rice (Saigusa et al., 1996). Therefore,
no-tillage systems in combination with a single basal application of controlled
release fertilizer in Andisols are very promising management strategies for both
rice and upland crops.
REFERENCES
Allbrook, R. F. (1986). Effect of skid trail compaction on a volcanic soils in central Oregon, USA. Soil
Sci. Soc. Am. J. 50, 1344–1346.
Alvarez, S. E., Etchevers, J. D., Ortiz, J., Nunez, R., Volke, V., Tijerina, L., and Martinez, A. (1999).
Biomass production and phosphorus accumulation of potato as affected by phosphorus nutrition.
J. Plant. Nutr. 22, 205 –217.
Ando, H., and Shoji, S. (1986). Kinetic study on soil nitrogen mineralization and nitrogen
immobilization of paddy soils. Jpn. J. Soil Sci Plant Nutr. 57, 1–7.
Aomine, S., and Wada, K. (1962). Differential weathering of volcanic ash and pumice, resulting in
formation of hydrated halloysite. Am. Miner. 47, 1024–1048.
Arihara, J., and Karasawa, T. (2000). Effects of previous crops on arbuscular mycorrhizal formation
and growth of succeeding maize. Soil Sci. Plant Nutr. 46, 43–51.
Arnalds, O., and Kimble, J. (2001). Andisols of deserts in Iceland. Soil Sci. Soc. Am. J. 65,
1778–1786.
Arya, O., Wittwer, F., Villa, A., and Ducon, C. (1990). Bovine fluorosis following volcanic activity in
the Southern Andes. Vet. Rec. 26, 641– 642.
Babbar, L. I., and Zak, D. R. (1995). Nitrogen loss from coffee agroecosystems in Costa Rica — leaching
and denitrification in the presence and absence of shade trees. J. Environ. Qual. 24, 227–233.
Bailey, S. W. (1990). Halloysite — a critical assessment. In “Proceedings 9th Int. Clay Conf.,
Strasbourg, France 1989” (V. C. Farmer and Y. Tardy, Eds.), pp. 89 –98. Sciences Géologiques,
Mémoires No. 86.
Baldock, J. A., and Nelson, P. N. (2000). Soil organic matter. In “Handbook of Soil Science” (M. E.
Sumner, Ed.), pp. B25–84. CRC Press, Boca Raton, FL.
Bartlett, R. J., and Riego, D. C. (1972). Toxicity of hydroxy aluminum in relation to pH and
phosphorous. Soil Sci. 114, 194–200.
Bäumler, R., and Zech, W. (1994). Characterization of Andisols developed from nonvolcanic material
in eastern Nepal. Soil Sci. 158, 211–217.
Bautista-Tulin, A. T., and Inoue, K. (1997). Hydroxy-interlayered minerals in Japanese soils
influenced by eolian deposition. Soil Sci. Soc. Am. J. 61, 631–640.
Beinroth, R. H. (1982). Some highly weathered soils of Puerto Rico. 1. Morphology, formation and
classification. Geoderma. 27, 1–73.
Blong, R. J. (1984). “Volcanic Hazards: A Sourcebook on the Effects of Eruptions”, p. 424. Academic
Press, Orlando, FL.
Borie, F., and Rubio, R. (1999). Effects of arbuscular mycorrhizae and liming on growth and mineral
acquisition of aluminum-tolerant and aluminum-sensitive barley cultivars. J. Plant Nutr. 22,
121–137.
Botello, G. J., and Ferrera, C. R. (1987). Effect of VA mycorrhiza on onion allium strain in a Mexican
Andosol. Rev. Latinoam. Microbiol. 29, 97–102.
Boudot, J. P., Bel Hadi Brahim, A., and Chone, T. (1988). Dependence of carbon and nitrogen
mineralization rates upon amorphous metallic constituents and allophanes in highland soils.
Geoderma. 42, 245–260.
Boudot, J. P., Bel Hadi Brahim, A., Steiman, R., and Seigle-Murandi, F. (1989). Biodegradation of
synthetic organo-metallic complexes of iron and aluminum with selected metal to carbon ratios.
Soil Biol Biochem. 21, 961– 966.
Brahim, B. H. (1987). “Influence des constituants alumineux et ferriques non cristallins sur les cycles
du carbone et de l’azote dans les sols montegands acids”, p. 99, These Universite de Nancy I.
Buurman, P., deBoer, K., and Pape, T. (1997). Laser diffraction grain-size characteristics of Andisols
in perhumid Costa Rica: the aggregate size of allophane. Geoderma. 78, 71–91.
Caner, L., Bourgeon, G., Toutain, F., and Herbillon, A. J. (2000). Characteristics of non-allophanic
Andisols derived from low-activity clay regoliths in the Nilgiri Hills (Southern India). Eur. J. Soil
Sci. 51, 553–563.
Casadevall, T., Rose, W., Gerlach, T., Greenland, L. P., Ewert, J., Wunderman, R., and Symonds, R.
(1983). Gas emissions and the eruptions of Mount St. Helens through 1982. Science. 221,
1383–1385.
Chadwick, O. A., Olson, C. G., Hendricks, D. M., Kelly, E. F., and Gavenda, R. T. (1994). Quantifying
climatic effects on mineral weathering and neoformation in Hawaii “Transactions of the 15th
World Congress of Soil Science, Acapulco, Mexico, 1994”, pp. 94–105.
Chen, Z. S., Asio, V. B., and Yi, D. F. (1999). Characteristics and genesis of volcanic soils along a
toposequence under a subtropical climate in Taiwan. Soil Sci. 164, 510 –525.
Chen, Z. S., Tsou, T. C., Asio, V. B., and Tsai, C. C. (2001). Genesis of Inceptisols on a volcanic
landscape in Taiwan. Soil Sci. 166, 255–266.
Childs, C. W., Parfitt, R. L., and Newman, R. H. (1990). Structural studies of Silica Springs allophane.
Clay Min. 25, 329 –341.
Childs, C. W., Matsue, N., and Yoshinaga, N. (1991). Ferrihydrite in volcanic ash soils of Japan. Soil
Sci. Plant Nutr. 37, 299–311.
Churchman, G. J., and Gilkes, R. J. (1989). Recognition of intermediates in the possible
transformation of halloysite to kaolinite in weathering profiles. Clay Miner. 24, 579–590.
Churchman, G. J., and Theng, B. K. G. (1984). Interactions of halloysites with amides: mineralogical
factors affecting complex formation. Clay Miner. 19, 161–175.
Cradwick, P. D. G., Farmer, V. C., Russell, J. D., Masson, C. R., Wada, K., and Yoshinaga, N. (1972).
Imogolite, a hydrated aluminum silicate of tubular structure. Nature (London) Phys. Sci. 240,
187 –189.
Cronan, C. S., Walker, W. J., and Bloom, P. R. (1986). Predicting aqueous aluminum concentrations
in natural waters. Nature. 324, 140–143.
Cronin, S. J., Hedley, M. J., Neall, V. E., and Smith, R. G. (1998). Agronomic impact of tephra fallout
from the 1995 and 1996 Ruapehu volcano eruptions, New Zealand. Environ. Geol. 34, 21–30.
Cronin, S. J., Hedley, M. J., Smith, R. G., and Neall, V. E. (1997). Impact of Ruapehu ash fall on soil
and pasture nutrient status. 1. October 1995 eruptions. N.Z. J. Agric. Res. 40, 383–395.
Cullen, S. J., Montagne, C., and Ferguson, H. (1991). Timber harvest trafficking and soil compaction
in western Montana. Soil Sci. Soc. Am. J. 55, 1416–1421.
Classification Committee of Cultivated Soils. (1996). “Classification of Cultivated Soils in Japan,
Third Approximation”, p. 82. National Institute of Agro-Environmental Sciences, Tsukuba,
Japan.
Dahlgren, R. A., and Saigusa, M. (1994). Aluminum release rates from allophanic and nonallophanic
Andosols. Soil Sci. Plant Nutr. 40, 125–136.
Dahlgren, R. A., and Ugolini, F. C. (1989a). Effects of tephra addition on soil processes in Spodosols
in the Cascade Range, Washington, U.S.A. Geoderma. 45, 331 –355.
Dahlgren, R. A., and Ugolini, F. C. (1989b). Formation and stability of imogolite in a tephritic
Spodosol, Cascade Range, Washington, USA. Geochim. Cosmochim. Acta. 53, 1897–1904.
Dahlgren, R. A., and Ugolini, F. C. (1991). Distribution and characterization of short-range-order
minerals in Spodosols from the Washington Cascades. Geoderma. 48, 391 –413.
Dahlgren, R. A., and Walker, W. J. (1993). Aluminum release rates from selected Spodosol Bs
horizons: effect of pH and solid-phase aluminum pools. Geochim. Cosmochim. Acta. 57, 57–66.
Dahlgren, R. A., Driscoll, C. T., and McAvoy, D. C. (1989). Aluminum precipitation and dissolution
rates in Spodosol Bs horizons in the northeastern USA. Soil Sci. Soc. Am. J. 53, 1045–1052.
Dahlgren, R. A., Ugolini, F. C., Shoji, S., Ito, T., and Sletten, R. S. (1991). Soil-forming processes in
Alic Melanudands under Japanese pampas grass and oak. Soil Sci. Soc. Am. J. 55, 1049–1056.
Dahlgren, R. A., Dragoo, J. P., and Ugolini, F. C. (1997). Weathering of Mt. St. Helens tephra under a
cryic-udic climatic regime. Soil Sci. Soc. Am. J. 61, 1519–1525.
Dahlgren, R. A., Ugolini, F. C., and Casey, W. H. (1999). Field weathering rates of Mt. St. Helens
tephra. Geochim. Cosmochim. Acta. 63, 587 –598.
Delvaux, B., and Herbillon, A. J. (1995). Pathways of mixed-layer kaolin–smectite formation in soils.
In “Clays – Controlling the Environment” (G. J. Churchman, R. W. Fitzpatrick, and
R. A. Eggleton, Eds.), pp. 457– 461. Proceedings 10th Int. Clay Conf., Adelaide, Australia,
1993, CSIRO Publ., Melbourne, Australia.
Delvaux, B., Herbillon, A. J., and Vielvoye, L. (1989). Characterization of a weathering sequence of
soils derived from volcanic ash in Cameroon– taxonomic, mineralogical and agronomic
implications. Geoderma. 45, 375– 388.
Delvaux, B., Herbillon, A. J., Dufey, J. E., and Vielvoye, L. (1990a). Surface properties and clay
mineralogy of hydrated halloysitic soil clays. I. Existence of interlayer Kþ specific sites. Clay
Miner. 25, 129–139.
Delvaux, B., Herbillon, A. J., Vielvoye, L., and Mestdagh, M. M. (1990b). Surface properties and clay
mineralogy of hydrated halloysitic soil clays. II. Evidence for the presence of halloysite/smectite
(H/SM) mixed-layer clays. Clay Miner. 25, 141–160.
Delvaux, B., Tessier, D., Herbillon, A. J., Gurtin, G., Jaunet, A. M., and Vielvoye, L. (1992).
Morphology, texture, and microstructure of halloysitic soil clays as related to weathering and
exchangeable cation. Clays Clay Miner. 40, 446– 456.
Dethier, D. P., Pevear, D. R., and Frank, D. (1981). Alteration of new volcanic deposits. In “The 1980
Eruption of Mount Mt. Helens, Washington”. (P. W. Lipman and D. R. Mullineaux, Eds.). U.S.
Geol. Surv. Prof. Pap, 1250, pp. 649 –665.
Dorel, M., Roger-Estrade, J., Manichon, H., and Delvaux, B. (2000). Porosity and soil water properties
of Caribbean volcanic ash soils. Soil Use Manage. 16, 133–140.
Dubroeucq, D., Geissert, D., and Quantin, P. (1998). Weathering and soil forming processes under
semi-arid conditions in two Mexican volcanic ash soils. Geoderma. 86, 99–122.
Duchaufour, P. (1977). “Pédogenése et Classification”, p. 477. Masson, Paris.
Dymond, J., Biscaye, P. E., and Rex, R. W. (1974). Eolian origin of mica in Hawaiian soils. Geol. Soc.
Am. Bull. 85, 37–40.
Eggleton, R. A., and Tilley, D. B. (1998). Hisingerite: A ferric kaolin mineral with curved
morphology. Clays Clay Miner. 46, 400– 413.
Elliot, E. T., and Coleman, D. C. (1988). Let the soil do the work for us. Ecol. Bull. 39, 23–32.
Escudey, M., Galindo, G., Forster, J. E., Briceno, M., Diaz, P., and Chang, A. (2001). Chemical forms
of phosphorus of volcanic ash-derived soils in Chile. Commun. Soil Sci. Plant Nutr. 32, 601– 616.
Eswaran, H., Vandenberg, E., and Reich, P. (1993). Organic carbon in soils of the world. Soil Sci. Soc.
Am. J. 57, 192 –194.
FAO, ISRIC, and ISSS. (1998). “World Reference Base for Soil Resources” FAO, Rome.
Farmer, V. C. (1987). The role of inorganic species in the transport of aluminum in Podzols.
In “Podzols et Podzolization” (D. Righi and A. Chauvel, Eds.), pp. 187 –194. AFES et INRA,
Paris.
Farmer, V. C., and Fraser, A. R. (1979). Synthetic imogolite, a tubular hydroxy-aluminium silicate.
In “Proceedings, International Clay Conference (Oxford 1978)” (M. M. Mortland and V. C.
Farmer, Eds.), pp. 547–553. Elsevier, Amsterdam.
Farmer, V. C., and Russell, J. D. (1990). The structure and genesis of allophanes and imogolite: their
distribution in nonvolcanic soils. In “Soil Colloids and their Association in Soil Aggregates”
(M. F. DeBoot, M. H. B. Hayes, and A. Herbillon, Eds.), pp. 165–178. Proc. NATO Advanced
Studies Workshop, Ghent 1985, Plenum, New York.
Fieldes, M. (1966). The nature of allophane in soils. Part 1. Significance of structural randomness in
pedogenesis. N.Z. J. Sci. 9, 599 –607.
Fox, R. L., Silva, J. A., Younge, O. R., Plucknett, D. L., and Sherman, G. D. (1967). Soil and plant
silicon and silicate response by sugarcane. Soil Sci. Soc. Am. Proc. 31, 775–779.
Fridrik, P., Halldor, T., Holmfridur, S., Holmgeir, B., and Olafur, A. (1996). Nitrogen mineralization
in soil [Icelandic]. Buvisindi. 10, 185 –208.
Furuya, H., Wakai, A., Takahashi, T., and Ui, T. (1996). Inhibition of macroconidial germina-
tion of Fusarium solani f. sp. Phaseoli by soil aluminum. Ann. Phytopathol. Soc. Jpn. 62,
69–74.
Furuya, H., Takahashi, T., and Matsumoto, T. (1999). Suppression of Fusarium solani f. sp. Phaseoli
on bean by aluminum in acid soils. Phytopathology. 89, 47– 52.
Gandeza, A. T., Shoji, S., and Yamada, I. (1991). Simulation of crop response to polyolefin-coated
urea. I. Field dissolution. Soil Sci. Soc. Am. J. 55, 1462–1467.
Garcia-Rodeja, E., Silva, B. M., and Macı́as, F. (1987). Andosols developed from non-volcanic
materials in Galicia, NW Spain. J. Soil Sci. 38, 573–591.
Geist, J. M., Hazard, J. W., and Seidel, K. W. (1989). Assessing physical conditions of some Pacific
Northwest volcanic ash soils after forest harvest. Soil Sci. Soc. Am. J. 53, 946–950.
Glassmann, J. R., and Simonson, G. M. (1985). Alteration of basalt in soils of Western Oregon. Soil
Sci. Soc. Am. J. 49, 262–273.
Global Volcanism Program. (2003). Dept. of Mineral Sciences, Smithsonian Institution, Washington
DC. http://www.volcano.si.edu/gvp/.
Goldchin, A., Oades, J. M., Skjemstad, J. O., and Clarke, P. (1994). Soil structure and carbon cycling.
Aust. J. Soil Res. 32, 1043–1068.
Goldchin, A., Baldock, J. A., Clarke, P., Higashi, T., and Oades, J. M. (1997). The effects of vegetation
and burning on the chemical composition of soil organic matter in a volcanic ash soil. II. Density
fractions. Geoderma. 76, 175–192.
Goodman, B. A., Russell, J. D., Montez, B., Oldfield, E., and Kirkpatrick, R. J. (1985). Structural
studies of imogolite and allophane by aluminum-27 and silicon-29 nuclear magnetic resonance
spectroscopy. Phys. Chem. Miner. 12, 342–346.
Gregorich, E. G., and Janzen, H. H. (2000). Decomposition. In “Handbook of Soil Science” (M. E.
Sumner, Ed.), pp. C107–120. CRC Press, Boca Raton, FL.
Harris, S. L. (1988). “Fire Mountains of the West — The Cascade and Mono Lake Volcanoes”, p. 379.
Mountain Press Pub., Missoula, MT.
Hatano, R., Iwanaga, K., Okajima, H., and Sakuma, T. (1988). Relationship between the distribution
soil macropores and root elongation. Soil Sci. Plant Nutr. 34, 535 –546.
Hewitt, A. E. (1998). “New Zealand Soil Classification”, 2nd edition. “Landcare Research Science
Series No. 1”. Manaaki Whenua Press, Lincoln, New Zealand.
Hidalgo, C., Bertaux, J., and Quantin, P. (1995). Forms of silica in indurated volcanic soils of the
Mexico Valley. In “Clays–Controlling the Environment” (G. J. Churchman, R. W. Fitzpatrick,
and R. A. Eggleton, Eds.), pp. 487– 493. Proceedings 10th Int. Clay Conf., Adelaide, Australia.
CSIRO Publ., Melbourne, Australia.
Hinkley, T. K., and Smith, K. S. (1987). Leachate chemistry of ash from the May 18, 1980 eruption of
Mount St. Helens. In “Chemistry of Ash and Leachates from May 18, 1980 Eruption of Mount St.
Helens, Washington” (T. K. Hinkley, Ed.), U.S. Geological Survey Professional Paper, 1397-B,
pp. 27–55.
Hiradate, S., Taniguchi, S., and Sakurai, K. (1998). Aluminum speciation in aluminum-silica solutions
and potassium chloride extracts of acidic soils. Soil Sci. Soc. Am. J. 62, 630 –636.
Hofstede, R. G. M. (1995). The effects of grazing and burning on soil and plant nutrient concentrations
in Colombian paramo grasslands. Plant Soil. 173, 111–132.
Huang, P. M. (1988). Ionic factors affecting aluminum transformations and the impact on soil and
environmental sciences. Adv. Soil Sci. 8, 1–78.
Huang, P. M., and Violante, A. (1986). Influence of organic acids on crystallization and surface
properties of precipitation products of aluminum. In “The Interactions of Soil Minerals with
Natural Organics and Microbes” (P. M. Huang and M. Schnitzer, Eds.), pp. 159–221. Soil
Science Society of America Special Publication 17, Madison, WI.
Hunter, C. R., Frazier, B. E., and Busacca, A. J. (1987). Lytell series: a nonvolcanic Andisol. Soil Sci.
Soc. Am. J. 51, 376– 383.
Hunter, D., and Ross, D. S. (1991). Evidence for a phytotoxic hydroxyaluminum polymer in organic
soil horizons. Science. 251, 1056–1058.
Ibrahim, S. M., Inoue, Y., and Shindo, H. (1998). Role of active aluminum in the formation of water-
stable macroaggregates. Soil Sci. Plant Nutr. 44, 685–689.
Inamatsu, K., Kiuchi, M., and Shibuya, K. (1991). Acidification of mulberry fields soil in case of
volcanic ash soils on Tsukuba upland. Jpn. J. Soil Sci. Plant Nutr. 62, 351–356.
Inoue, K. (1981). Implications of eolian dusts to 14Å minerals in the volcanic ash soils in Japan.
Pedologist. 25, 97–118.
Inoue, K., and Naruse, T. (1987). Physical chemical and mineralogical characteristics of modern
eolian dust in Japan and rate of dust deposition. Soil Sci. Plant Nutr. 33, 327 –345.
Inoue, K., and Higashi, T. (1988). Al– and Fe–humus complexes in Andisols. In “Proceedings 9th Int.
Soil Classification Workshop, Japan” (D. I. Kinloch, S. Shoji, F. M. Beinroth, and H. Eswaran,
Eds.), pp. 535– 546. Japanese Committee for 9th International Soil Classification Workshop and
Soil Management Support Services, Washington, D.C.
Inoue, H., Ito, T., and Saigusa, M. (2000). Yield and nutrient uptake of dent corn (Zea mays L.) in no-
tillage cultivation with single basal co-situs fertilization. Jpn. J. Soil Sci. Plant Nutr. 71,
674–681.
Inoue, K., Kondo, S., Tamano, Y., and Yokota, H. (2001). Amelioration of subsoil acidity in nonallo-
phanic Andosol by surface application of organic calcium salts. Soil Sci. Plant Nutr. 47, 113 –122.
Isobe, K., and Tsuboki, Y. (1997). Study on utilization of arbuscular mycorrhizal fungi in kidney bean
cultivation — effects of available phosphorus content on the infection of arbuscular mycorrhizal
fungi and two species of arbuscular mycorrhizal on the growth of kidney bean (Phaseolus
vulgaris L.). Jpn. J. Crop Sci. 66, 374– 380.
Ito, T., Shoji, S., and Saigusa, M. (1991). Classification of volcanic ash soils from Konsen District,
Hokkaido, according to the last Keys to Soil Taxonomy (1990). Jpn. J. Soil Sci. Plant Nutr. 62,
237–247.
Ito, T., Inoue, H., and Saigusa, M. (1997a). Effects of no-tillage system on silage-corn yields and
characteristics in Andisol. In “Plant Nutrition for Sustainable Food Production and Environment”
(T. Ando, K. Fujita, T. Mae, H. Matsumoto, S. Mori, and J. Sekiya, Eds.), pp. 653–654. Kluwer
Academic Publishers, Dordrecht.
Ito, T., Ide, T., and Saigusa, M. (1997b). Change of physical properties with agricultural machine
traffic in cultivated Andisols with different colloidal composition. Bull. Kawatabi Exp. Farm. 13,
17–21.
Ito, T., Yamada, D., and Saigusa, M. (1998). Effects of subsoil and polyolefin-coated urea application
on growth and nitrogen uptake of oats and barley in Andisols. Tohoku J. Agric. Res. 49, 25 –32.
Ito, T., Kikawa, N., and Saigusa, M. (2000a). Phosphorus availability difference between allophanic
and nonallophanic Andisols. In “Abstracts — Int. Symp. Productivity and Charge Characteristics
of Soils Derived from Volcanic Ash”, p. 18. Meirin Printing Co, Sendai, Japan.
Ito, T., Yokota, T., and Saigusa, M. (2000b). Measurement of organic phosphorus mineralization in
Andisols using anion exchange resin In “Abstracts — Int. Symp. Productivity and Charge
Characteristics of Soils Derived from Volcanic Ash”, p. 19. Meirin Printing Co, Sendai, Japan.
Jackson, M. L. (1963a). Aluminum bonding in soils: a unifying principle in soil science. Soil Sci. Soc.
Am. Proc. 37, 1–10.
Jackson, M. L. (1963b). Interlayering of expansible layer silicates in soils by chemical weathering.
Clays Clay Mineral. 11, 29–46.
Johnson-Maynard, J. L., McDaniel, P. A., Ferguson, D. E., and Falen, A. L. (1997). Chemical and
mineralogical conversion of Andisols following invasion by bracken fern. Soil Sci. Soc. Am. J. 61,
549–555.
Jongmans, A. G., van Oort, F., Nieuwenhuyse, A., Buurman, P., Jaunet, A. M., and van Doesburg, J. D. J.
(1994). Inheritance of 2:1 phyllosilicates in Costa Rican Andisols. Soil Sci. Soc. Am. J. 58,
494–501.
Kamewada, K. (1996). Application of “four-plane model” to the adsorption of Kþ, NO2 22
3 , and SO4
from a mixed solution of KNO3 and K2SO4 on Andisols. Soil Sci. Plant Nutr. 42, 801–808.
Karasawa, T., Arihara, J., and Kasahara, Y. (2000a). Effects of previous crops on arbuscular
mycorrhizal formation and growth of maize under various soil moisture conditions. Soil Sci.
Plant Nutr. 46, 53–60.
Karasawa, T., Takebe, M., and Kasahara, Y. (2000b). Arbuscular mycorrhizal (AM) effects on maize
growth and AM colonization of root under various soil conditions. Soil Sci. Plant Nutr. 46, 61–67.
Karathanasis, A. D., Adams, F., and Hajek, B. F. (1983). Stability relationships in kaolinite, gibbsite,
and Al-hydroxy interlayered vermiculite soil systems. Soil Sci. Soc. Am. J. 47, 1247–1251.
Kato, H., Shirai, M., and Matsukawa, S. (1995). Ionic species of aluminium in acid soils in terms of
chemical equilibrium. Jpn. J. Soil Sci. Plant Nutr. 66, 39 –47.
Katou, H., Clothier, B. E., and Green, S. R. (1996). Anion transport involving competitive adsorption
during transient water flow in an Andisol. Soil Sci. Soc. Am. J. 60, 1368–1375.
Kilbertus, G. (1980). Microhabitats in soil aggregates. Their relationship with bacterial biomass and
the size of procaryotes present. Rev. Ecol. Biol. Sol. 17, 543 –557.
Kimble, J. M., and Eswaran, H. (1988). The Andisol-Oxisol transition. In “Proceedings 9th Int. Soil
Classification Workshop, Japan” (D. I. Kinloch, S. Shoji, F. M. Beinroth, and H. Eswaran, Eds.),
pp. 219–232. Japanese Committee for 9th International Soil Classification Workshop and Soil
Management Support Services, Washington, DC.
Kimble, J. M., Ping, C. L., Sumner, M. E., and Wilding, L. P. (2000). Andisols. In “Handbook of Soil
Science” (M. E. Sumner, Ed.), pp. E209–224. CRC Press, Boca Raton, FL.
Kirkman, J. H., and McHardy, W. J. (1980). A comparative study of the morphology, chemical
composition and weathering of rhyolitic and andesitic glass. Clay Miner. 15, 165–173.
Kobayashi, S., Shoji, S., Yamada, I., and Masui, J. (1976). Chemical and mineralogical studies on
volcanic ashes. III. Some mineralogical and chemical properties of volcanic glasses with special
reference to the rock types of volcanic ashes. Soil Sci. Plant Nutr. 22, 7–13.
Kondo, Y., Fujitani, T., Katsui, Y., and Niida, K. (1979). Nature of the 1977–1978 volcanic ash from
Usu volcano, Hokkaido, Japan. Kazan. 24, 223– 238.
LaManna, J. M., and Ugolini, F. C. (1987). Trioctahedral vermiculite in the 1980 pyroclastic flow,
Mt. St. Helens, Washington. Soil Sci. 143, 162–167.
Leamy, M. L. (1984). “International Committee on the Classification of Andisols (ICOMAND)
Circular letter No. 6.” New Zealand Soil Bureau, DSIR, Lower Hutt.
Leamy, M. L., Clayden, B., Parfitt, RL., Kinloch, D. I., and Childs, C. W. (1988). “The Andisol
Proposal 1988, Final proposal of the International Committee on the Classification of Andisols
(ICOMAND).” New Zealand Soil Bureau, Lower Hutt, New Zealand.
Lenhard, R. J. (1986). Changes in void distribution and volume during compaction of a forest soil. Soil
Sci. Soc. Am J. 50, 462–464.
Loughnan, F. C. (1969). “Chemical Weathering of the Silicate Minerals”, p. 154. Elsevier, New York.
Lowe, D. J. (1986). Controls on the rates of weathering and clay mineral genesis in airfall tephras: a
review and New Zealand case study. In “Rates of Chemical Weathering of Rocks and Minerals”
(S. M. Colman and D. P. Dethier, Eds.), pp. 265 –330. Academic Press, Orlando.
MacKenzie, K. J. D., Bowden, M. E., and Meinhold, R. H. (1991). The structure and thermal
transformations of allophanes studied by 29Si and 27Al high-resolution solid-state NMR. Clays
Clay Miner. 39, 337– 346.
Madeira, M., Furtado, A., Jeanroy, E., and Herbillon, A. J. (1994). Andisols of Madeira Island
(Portugal) — characteristics and classification. Geoderma. 62, 363–383.
Maeda, T., and Soma, K. (1986). Physical properties. In “Ando Soils in Japan” (K. Wada, Ed.),
pp. 99–111. Kyushu University Press, Fukuoka, Japan.
Maeda, T., Takenaka, H., and Warkentin, B. P. (1977). Physical properties of allophanic soils. Adv.
Agron. 29, 229 –264.
Malucelli, F., Terribile, F., and Colombo, C. (1999). Mineralogy, micromorphology and chemical
analysis of Andosols on the Island of São Miguel (Azores). Geoderma. 88, 73–98.
Manske, G. G. B., Ortiz-Monasterio, J. I., Van Ginkel, M., Gonzalez, R. M., Rajaram, S., Molina, E.,
and Vlek, P. L. G. (2000). Trait associated with improved P-uptake efficiency in CIMMYT’s
semidwarf spring bread wheat grown on an acid Andisol in Mexico. Plant Soil. 221, 189–204.
Martini, J. A. (1976). The evolution of soil properties as it relates to the genesis of volcanic ash soils in
Costa Rica. Soil Sci. Soc. Am. J. 40, 895–900.
Matsuyama, N., Saigusa, M., and Kudo, K. (1994). Phosphorus response of dent corn young plants and
evaluation of phosphorus availability in Andosols. Tohoku J. Crop Sci. 37, 57–58.
Matsuyama, N., Kudo, K., and Saigusa, M. (1999a). Active aluminum of cultivated Andosols and
related soil chemical properties in Japan. Bull. Fac. Agric. Life Sci. Hirosaki Univ. 1, 30 –36.
Matsuyama, N., Saigusa, M., and Kudo, K. (1999b). Acidity of Japanese cultivated Andosols and
significance of exchange acidity Y1 in their classification. J. Soil Sci. Plant Nutr. 70, 754– 761.
May, H. M., Helmke, P., and Jackson, M. L. (1979). Gibbsite solubility and thermodynamic properties
of hydroxyaluminum ions in aqueous solutions at 258C. Geochim. Cosmochim. Acta. 50,
1667–1677.
McKeague, J. A. (1967). An evaluation of 0.1 M pyrophosphate and pyrophosphate-dithionite in
comparison with oxalate as extractants of the accumulation products in podzols and some other
soils. Can. J. Soil Sci. 47, 95–99.
Meurisse, R. T. (1988). Forest productivity and the management of U.S. Andisols. In “Proceedings
9th Int. Soil Classification Workshop, Japan” (D. I. Kinloch, S. Shoji, F. M. Beinroth, and
H. Eswaran, Eds.), pp. 245 –257. Japanese Committee for 9th International Soil Classification
Workshop and Soil Management Support Services, Washington, DC.
Mitchell, W. A. (1975). Heavy minerals. In “Soil Components, Vol. 2. Inorganic Components” (J. E.
Gieseking, Ed.), pp. 449 –480. Springer-Verlag, Berlin-Heidelberg-New York.
Mizota, C., and Inoue, K. (1988). Eolian dust contribution to soil development on volcanic ashes
in Japan. In “Proceedings 9th Int. Soil Classification Workshop, Japan” (D. I. Kinloch, S.
Shoji, F. M. Beinroth, and H. Eswaran, Eds.), pp. 547–557. Japanese Committee for 9th
International Soil Classification Workshop and Soil Management Support Services,
Washington, DC.
Mizota, C., and Matsuhisa, Y. (1995). Isotopic evidence for the eolian origin of quartz and mica in
soils developed on volcanic materials in the Canary Archipelago. Geoderma. 66, 313– 320.
Mizota, C., and van Reeuwijk, L. P. (1989). “Clay Mineralogy and Chemistry of Soils formed in
Volcanic Material in Diverse Climatic Regions”, p. 186. Soil Monograph 2, ISRIC, Wageningen,
the Netherlands.
Mizota, C., Kawasaki, I., and Wakatsuki, T. (1988). Clay mineralogy and chemistry of seven pedons
formed in volcanic ash, Tanzania. Geoderma. 43, 131–141.
Mizota, C., Kusakabe, M., and Noto, M. (1990). Oxygen isotope composition of quartz in soils
developed on late Quaternary volcanic ashes in Japan. Geoderma. 46, 319–327.
Mizuno, N., and Yoshida, H. (1993). Effect of exchangeable aluminum on the reduction of potato
scab. Plant Soil. 156, 505–508.
Mizuno, N., Nishimidin, K., Yoshida, H., Nanzyo, M., and Tadano, T. (1998a). Effects of deep
plowing and “cutting and banking” practice on the concentration of water-soluble aluminum and
allophane content in Andosols: Implication for recent incidence of common potato scab in
Abashiri area. Soil Sci. Plant Nutr. 44, 571 –578.
Mizuno, N., Yoshida, H., Nanzyo, M., and Tadano, T. (1998b). Chemical characterization of
conducive and suppressive soils for potato scab in Hokkaido, Japan. Soil Sci. Plant Nutr. 44,
289–295.
Mizuno, N. K., Yoshida, H., and Tadano, T. (2000). Efficacy of single application of ammonium
sulfate in suppressing potato scab. Soil Sci. Plant Nutr. 46, 611 –616.
Mohr, E. C. J. (1938). The relation between soil and population density in the Netherlands Indies.
In “Compies Rendus du Congres International de Geogrphic”, pp. 478 –493. Amsterdam. Tome
Deuxiem, Section IIIc.
Mora, M. L., Baeza, G., Pizarro, C., and Demanet, R. (1999a). Effect of calcitic and dolomitic lime
on physicochemical properties of a Chilean Andisol. Commun. Soil Sci. Plant Anal. 30,
427–439.
Mora, M. L., Schnettler, B., and Demanet, R. (1999b). Effect of liming and gypsum on soil chemistry,
yield, and mineral composition of rye grass grown in an acidic Andisol. Commun. Soil Sci. Plant
Anal. 30, 1251–1266.
Nakamaru, Y., Nanzyo, M., and Yamasaki, S. (2000a). Short-term reaction of apatite in fresh Pinatubo
volcanic ash with reactive components in the soil-plant system. Jpn. J. Soil Sci. Plant Nutr. 71,
55 –62.
Nakamaru, Y., Nanzyo, M., and Yamasaki, S. (2000b). Utilization of apatite in fresh volcanic ash by
pigeonpea and chickpea. Soil Sci. Plant Nutr. 46, 591–600.
Nanzyo, M. (1987). Formation of noncrystalline aluminum phosphate through phosphate sorption on
allophanic Andosols. Commun. Soil Sci. Plant Anal. 18, 735 –742.
Nanzyo, M., Takahashi, T., Sato, A., Shoji, S., and Yamada, I. (1997). Dilute acid-soluble phosphorus
in fresh air-borne tephras and fixation with an increase in active aluminum and iron. Soil Sci.
Plant Nutr. 43, 839 –848.
Nanzyo, M., Nakamaru, Y., and Yamasaki, S. (1999). Inhibition of apatite dissolution due to
formation of calcium oxalate coating. Phosphorus Res. Bull. 9, 17–22.
Newman, R. H., Childs, C. W., and Churchman, G. J. (1994). Aluminum coordination and
structural disorder in halloysite and kaolinite by 27Al NMR spectroscopy. Clay Miner. 29,
305 –312.
Nira, R., and Nishimune, A. (1993). Studies on nitrogen mineralization properties of Tokachi soils by
kinetic analysis. Soil Sci. Plant Nutr. 39, 321 –329.
Norrish, K. (1995). An unusual fibrous halloysite. In “Clays–Controlling the Environment” (G. J.
Churchman, R. W. Fitzpatrick, and R. A. Eggleton, Eds.), pp. 275– 284. Proceedings 10th Int. Clay
Conf., Adelaide, Australia. 1993. CSIRO Publ., Melbourne, Australia.
Nuhn, W. W. (1987). “Soil Genesis on the 1980 Pyroclastic Flow of Mount St. Helens”, p. 179.
M.S. thesis, College of Forest Resources, University of Washington, Seattle, WA.
Okada, T., Hosokawa, K., and Kawaguchi, K. (1986). Influences of excessive fertilizing on tea plants
and soils. Bull. Tokushima Exp. Agric. Instit. 23, 23–31.
Okamura, Y., and Wada, K. (1983). Electrical charge characteristics of Ando(B) and Red-Yellow B
soils and weathered pumices. J. Soil Sci. 34, 287 –295.
Okano, S. (1990). Availability of mineralized nitrogen from microbial biomass and organic matter
after drying and heating of grassland soils. Plant Soil. 129, 219 –225.
Ombodi, A., and Saigusa, M. (2000). Improvement of nitrogen use efficiency of tendergreen mustard
grown in the field during rainy seasons by using polyolefin-coated urea. Tohoku J. Agric. Res. 51,
1 –6.
Oskarsson, N. (1980). The interaction between volcanic gases and tephra: fluorine adhering to tephra
of the 1970 Hekla eruption. J. Volcanol. Geotherm. Res. 8, 251 –256.
Otsuka, H., Briones, A. A., Daquiado, N. P., and Evangelio, F. A. (1988). Characteristics and genesis
of volcanic ash soils in the Philippines. Technical Bulletin of Tropical Agricultural Research
Center, Japan. 24, 1–122.
Palacios, M. S., Shimada, M. K., and Salinas, C. C. (1987). Effect of inoculation in two varieties of
onion (Allium cepa L.) with four VA endomycorrhizal fungi in phosphate deficient soil. Revista
Latinoamericana de Microbiologia. 29, 329–336.
Pansombat, K., Kanazawa, S., and Horiguchi, T. (1997). Microbial ecology in tea soils. I. Soil
properties and microbial populations. Soil Sci. Plant Nutr. 43, 317–327.
Parfitt, R. L. (1990). Allophane in New Zealand — A review. Aust. J. Soil Res. 28, 343–360.
Parfitt, R. L., and Henmi, T. (1980). Structure of some allophanes from New Zealand. Clays Clay
Miner. 28, 285– 294.
Parfitt, R. L., and Kimble, J. M. (1989). Conditions for formation of allophane in soils. Soil Sci. Soc.
Am. J. 53, 971–977.
Parfitt, R. L., and Saigusa, M. (1985). Allophane and humus–aluminum in Spodosols and Andepts
formed from the same volcanic ash beds in New Zealand. Soil Sci. 139, 149– 155.
Parfitt, R. L., and Wilson, A. D. (1985). Estimation of allophane and halloysite in three sequences of
volcanic soils, New Zealand. In “Volcanic Soils” (E. Fernandez Caldas and D. H. Yaalon, Eds.),
Catena Suppl. 7, pp. 1–8.
Parfitt, R. L., Furkert, F. J., and Henmi, T. (1980). Identification and structure of two types of
allophane from volcanic ash soils and tephra. Clays Clay Min. 28, 328–334.
Parfitt, R. L., Russell, M., and Orbell, G. E. (1983). Weathering sequence of soils from volcanic ash
involving allophane and halloysite. Geoderma. 29, 41–57.
Parfitt, R. L., Tate, K. R., and McKercher, R. B. (1994). Measurement of phosphorus mineralisation
using an anion exchange membrane. Commun. Soil Sci. Plant Anal. 25, 3209–3219.
Parfitt, R. L., Salt, G. J., and Saggar, S. (2001). Effect of leaching and clay content on carbon and
nitrogen mineralisation in maize and pasture soils. Aust. J. Soil Res. 39, 535 –542.
Paterson, E. (1977). Specific surface area and pore structure of allophanic soil clays. Clay Miner. 12,
1–9.
Percival, H. J., Parfitt, R. L., and Scott, N. A. (2000). Factors controlling soil carbon levels in New
Zealand grasslands: is clay content important? Soil Sci. Soc. Am. J. 64, 1623–1630.
Perret, S., and Dorel, M. (1999). Relationship between land use, fertility and Andisol behaviour:
examples from volcanic islands. Soil Use Manage. 15, 144– 149.
Perret, S., Michellon, R., Boyer, J., and Tassin, J. (1996). Soil rehabilitation and erosion control
through agro-ecological practices on Reunion Island. Agric. Ecosyst. Environ. 59, 149 –157.
Pevear, D. R., Dethier, D. P., and Frank, D. (1982). Clay minerals in the 1980 deposits from Mount St.
Helens. Clays Clay Miner. 30, 241–252.
Ping, C. L., Shoji, S., and Ito, T. (1988). Properties and classification of three volcanic ash-derived
pedons from Aleutian Islands and Alaska Peninsula, Alaska. Soil Sci. Soc. Am. J. 52, 455– 462.
Ping, C. L., Shoji, S., Ito, T., Takahashi, T., and Moore, J. P. (1989). Characteristics and classification
of volcanic-ash-derived soils in Alaska. Soil Sci. 148, 8–28.
Poudel, D. D., Midmore, D. J., and West, L. T. (1999). Erosion and productivity of vegetable systems
on sloping volcanic ash-derived Philippine soils. Soil Sci. Soc. Am. J. 63, 1366–1376.
Poulenard, J., Podwojewski, P., Janeau, J. L., and Collinet, J. (2001). Runoff and soil erosion under
rainfall simulation of Andisols from the Ecuadorian Paramo: effect of tillage and burning.
Catena. 45, 185 –207.
Qafoku, N. P., Sumner, M. E., and Radcliffe, D. E. (2000). Anion transport in columns of variable
charge subsoils: nitrate and chloride. J. Environ. Qual. 29, 484–493.
Quantin, P. (1992). “Les Soils de l’ Archipel Volcanique des Nouvelles-Hebrides (Vanuatu)”, p. 498.
ORSTOM, Paris.
Quantin, P., Gautheyrou, J., and Lorenzoni, P. (1988). Halloysite formation through in situ weathering
of volcanic glass from trachytic pumices, Vico’s Volcano, Italy. Clay Miner. 23, 423–437.
Reynolds-Vargas, J. S., Richter, D. D., and Bornemisza, E. (1994). Environmental impacts of
nitrification and nitrate adsorption in fertilized Andisols in the Valle-Central of Costa Rica. Soil
Sci. 157, 289–299.
Robarge, W. P. (1999). Precipitation/dissolution reactions in soils. In “Soil Physical Chemistry”, 2nd
edn. (D. L. Sparks, Ed.), pp. 193– 238. CRC Press, Boca Raton.
Ruxton, B. P. (1988). Towards a weathering model of Mount Lamington Ash, Papua New Guinea.
Earth Sci. Rev. 25, 387–397.
Ryan, M. C., Graham, G. R., and Rudolph, D. L. (2001). Contrasting nitrate adsorption in Andisols of
two coffee plantation in Costa Rica. J. Environ. Qual. 30, 1848–1852.
Saigusa, M. (1988). Agronomy of Aludands in northeastern Japan. In “Proceedings 9th Int. Soil
Classification Workshop, Japan” (D. I. Kinloch, S. Shoji, F. M. Beinroth, and H. Eswaran, Eds.),
pp. 593 –602. Japanese Committee for 9th International Soil Classification Workshop and Soil
Management Support Services, Washington, DC.
Saigusa, M. (1991). Plant growth on acid soils with special reference to phytotoxic Al and subsoil
acidity (review). Jpn. J. Soil Sci. Plant Nutr. 62, 451 –459.
Saigusa, M. (1992). Characterization and soil management of acid Andisols. In “Plant-Soil Interaction at
Low pH” (T. Sakuma, Ed.), pp. 36–41. Faculty of Agric., Hokkaido Univ, Sapporo, Japan.
Saigusa, M. (1999). Slow-release nitrogen fertilizers and plant nutrition. In “Nitrogen Nutrition and
Plant Growth” (H. S. Srivastava and R. P. Singh, Eds.), pp. 305–335. Science Publishers, Enfield,
N.H.
Saigusa, M. (2000). Growth of rice and its silicon nutrition. Kagaku To Seibutsu. 38, 514–519.
Saigusa, M., and Matsuyama, N. (1998). Distribution of allophanic Andosols and non-allophanic
Andosols in Japan. Tohoku J. Agric. Sci. 48, 75– 83.
Saigusa, M., and Toma, M. (1997). Mechanism of reduction of exchangeable aluminum by gypsum
application in acid Andosols. Soil Sci. Plant Nutr. 43, 343 –349.
Saigusa, M., Saito, K., and Shoji, S. (1976). Occurrence of Cu and Zn within the profiles of soils
formed from different parent ashes and corn response to these elements. Tohoku J. Agric. Res. 27,
12 –19.
Saigusa, M., Shoji, S., and Kato, T. (1978). Origin and nature of halloysite in Ando soils from Towada
tephra, Japan. Geoderma. 20, 115–129.
Saigusa, M., Shoji, S., and Takahashi, T. (1980). Plant root growth in acid Andosols from northeastern
Japan. Exchange acidity Y1 as a realistic measure of aluminum toxicity potential. Soil Sci. 130,
242 –250.
Saigusa, M., Shoji, S., and Sakai, H. (1983). The effects of subsoil acidity of Andosols on the growth
and nitrogen uptake of barley and wheat. Jpn. J. Soil Sci. Plant Nutr. 54, 460– 466.
Saigusa, M., Shoji, S., and Nakaminami, H. (1987). Measurement of water retention at 15 bar tension
by pressure membrane method and available moisture of Andosols. Jpn. J. Soil Sci. Plant Nutr.
58, 374–377.
Saigusa, M., Shoji, S., Goto, J., Kodama, H., and Abe, T. (1990). Nitrogen fertility of acid Andisols
and nitrogen application to upland crops. Trans. 14th Inter. Cong. Soil Sci. IV, 638– 639.
Saigusa, M., Matsuyama, N., Honnna, T., and Abe, T. (1991). Chemistry and fertility of acid Andisols
with special reference to subsoil acidity. In “Plant– Soil Interactions at Low pH” (R. J. Wright,
V. C. Baligar, and R. P. Murrmann, Eds.), pp. 73–80. Kluwer Academic Publishers, Dordrecht.
Saigusa, M., Hossain, M. Z., Toma, M., Shibuya, K., and Abe, T. (1992a). Utilization of nonallophanic
Andosols as nursery bed soil of rice seedlings. Bull. Kawatabi Farm. 8, 1–5.
Saigusa, M., Matsuyama, N., and Abe, T. (1992b). Electric charge characteristics of Andosols and its
problem on soil management. Jpn. J. Soil Sci. Plant Nutr. 63, 196–201.
Saigusa, M., Toma, M., and Abe, T. (1994). Effects of phosphogypsum application in topsoil on
amelioration of subsoil acidity of nonallophanic Andosols. J. Jpn. Soc. Grass Sci. 39,
397 –404.
Saigusa, M., Matsumoto, T., and Abe, T. (1995). Phytotoxicity of monomer aluminum ions and
hydroxy-aluminum polymer ions in an Andosol. In “Plant–Soil Interaction at Low pH” (R. A.
Date, N. J. Grundon, G. E. Rayment, and M. E. Probert, Eds.), pp. 367 –370. Kluwer Academic
Publishers, Dordrecht.
Saigusa, M., Hossain, M. Z., Tashiro, T., and Shibuya, K. (1996). Maximizing rice yield with
controlled availability fertilizer and no-tillage culture in controlling environmental degradation.
In “Proc. Int. Symp. Maximizing Sustainable Rice Yields through Improved Soil and
Environmental Management” Khonkaen, Thailand”, 1, pp. 75 –85.
Saigusa, M., Yamamoto, A., and Shibuya, K. (1998). Agricultural use of porous hydrate calcium
silicate. I. Evaluation of porous hydrate calcium silicate as a silicate fertilizer. Jpn. J. Soil Sci.
Plant Nutr. 69, 576 –581.
Saigusa, M., Yamamoto, A., and Shibuya, K. (1999). Agricultural use of porous hydrate calcium
silicate-Effect of porous hydrated calcium silicate on resistance of rice plant (Oryza sativa L.) to
rice blast (Pyricularia olyzae). Plant Prod. Sci. 3, 51 –54.
Saito, M. (1990). Nitrogen mineralization parameters and its availability indices of soils in Tohoku
District, Japan: their relationship. Jpn. J. Soil Sci. Plant Nutr. 61, 265– 272.
Saito, M., and Ishi, K. (1987). Estimation of soil nitrogen mineralization in corn-grown fields based on
mineralization parameters. Soil Sci. Plant Nutr. 33, 555 –566.
Sakamoto, K., and Oba, Y. (1993). Relationship between available N and soil biomass in upland field
soils. Jpn. J. Soil Sci. Plant Nutr. 64, 42–48.
Savant, N. K., Snyder, G. H., and Datnoff, L. E. (1997). Silicon management and sustainable rice
production. Adv. Agron. 58, 151 –199.
Schwertmann, U. (1985). The effect of pedogenic environments on iron oxide minerals. Adv. Soil Sci.
1, 171– 200.
Schwertmann, U. (1988). Occurrence and formation of iron oxides in various pedoenvironments.
In “Iron in Soils and Clay Minerals” (J. W. Stucki, B. A. Goodman, and U. Schwertmann, Eds.),
pp. 267 –308. Reidel, Bordrecht.
Schwertmann, U., and Cornell, R. M. (1991). “Iron Oxides in the Laboratory — Preparation and
Characterization.”, p. 137. Weinheim, New York. VCH.
Schwertmann, U., and Taylor, R. M. (1989). Iron oxides. In “Minerals in Soil Environments”, 2nd edn.
(J. B. Dixon and S. B. Weed, Eds.), pp. 379–438. Soil Sci. Soc. Am., Madison, WI.
Shann, J. R., and Bertsch, P. M. (1993). Differential cultivar response to polynuclear hydroxyo-
aluminum complexes. Soil Sci. Soc. Am. J. 57, 116 –120.
Shimizu, H., Watanabe, T., Henmi, T., Masuda, A., and Saito, H. (1988). Studies on allophane
and imogolite by high-resolution solid-state 29Si- and 27Al-NMR and ESR. Geochem. J. 22, 23 –31.
Shimono, K. (1990). Effects of soil pH on the growth and yield of field crops in wet Andosols. Jpn.
J. Soil Sci. Plant Nutr. 61, 8–15.
Shoji, S. (1983). Mineralogical properties of volcanic ash soils. In “Volcanic Ash Soil — Genesis,
Properties, Classification” (N. Yoshinaga, Ed.), pp. 31 –72. Hakuyusha, Tokyo.
Shoji, S. (1985). Genesis and properties of non-allophanic Andisols in Japan. Applied Clay Sci. 1,
83–88.
Shoji, S., and Fujiwara, T. (1984). Active aluminum and iron in the humus horizons of Andosols from
northeastern Japan: their forms, properties, and significance in clay weathering. Soil Sci. 137,
216–226.
Shoji, S., and Masui, J. (1971). Opaline silica of recent volcanic ash soils in Japan. J. Soil Sci. 22,
101–112.
Shoji, S., and Ono, T. (1978). Physical and chemical properties and clay mineralogy of Andosols from
Kitakami, Japan. Soil Sci. 126, 297–312.
Shoji, S., Yamada, I., and Masui, J. (1974). Soils formed from the andesitic and basaltic ashes. I. The
nature of the parent ashes and soil formation. Tohoku, J. Agric. Res. 25, 104–112.
Shoji, S., Kobayashi, S., Yamada, I., and Masui, J. (1975). Chemical and mineralogical studies on
volcanic ashes. I. Chemical composition of volcanic ashes and their classification. Soil Sci. Plant
Nutr. 21, 311 –318.
Shoji, S., Saigusa, M., and Ebihara, M. (1980a). Cobalt content in volcanic ashes. Jpn. J. Soil Sci.
Plant Nutr. 51, 335–336.
Shoji, S., Saigusa, M., and Takahashi, T. (1980b). Plant root growth in acid Andosols from
northeastern Japan. I. Soil properties and root growth of burdock, barley, and orchard grass. Soil
Sci. 130, 124–130.
Shoji, S., Fujiwara, Y., Yamada, I., and Saigusa, M. (1982). Chemistry and clay mineralogy of ando
soils, brown forest soils, and podzolic soils formed from recent Towada ashes, N.E. Japan. Soil
Sci. 133, 69–86.
Shoji, S., Ito, T., Saigusa, M., and Yamada, I. (1985). Properties of nonallophanic Andosols from
Japan. Soil Sci. 140, 264– 277.
Shoji, S., Saigusa, M., and Goto, J. (1986). Acidity to subsoil of Andisols and nitrogen uptake and
growth of sorghum. Jpn. J. Soil Sci. Plant Nutr. 57, 264–271.
Shoji, S., Suzuki, Y., and Saigusa, M. (1987). Clay mineralogical and chemical properties of
nonallophanic Andepts (Andisols) from Oregon, USA. Soil Sci. Soc. Am. J. 51, 986–990.
Shoji, S., Hakamada, T., and Tomioka, E. (1990a). Properties and classification of selected volcanic
ash soils from Abashiri, northern Japan–transition of Andisols to Mollisols. Soil Sci. Plant Nutr.
36, 409–423.
Shoji, S., Kurebayashi, T., and Yamada, I. (1990b). Growth and chemical composition of Japanese
pampas grass (Miscanthus sinensis) with special reference to the formation of dark-colored
Andisols in northeastern Japan. Soil Sci. Plant Nutr. 36, 105– 120.
Shoji, S., Gandeza, A. T., and Kimura, K. (1991). Simulation of crop response to polyolefin-coated
urea. II. Nitrogen uptake by corn. Soil Sci. Soc. Am. J. 55, 1468–1473.
Shoji, S., Nanzyo, M., and Dahlgren, R. A. (1993a). “Volcanic Ash Soils — Genesis, Properties and
Utilization.”, p. 288. Elsevier, Amsterdam, the Netherlands.
Shoji, S., Nanzyo, M., Shirato, Y., and Ito, T. (1993b). Chemical kinetics of weathering in young
Andisols from northeastern Japan using 108C-normalized soil age. Soil Sci. 155, 53–60.
Shoji, S., Nanzyo, M., Dahlgren, R. A., and Quantin, P. (1996). Evaluation and proposed revisions of
criteria for Andosols in the World Reference Base for Soil Resources. Soil Sci. 161, 604–615.
Shoji, S., Takahashi, T., Ito, T., and Ping, C. L. (1988a). Properties and classification of selected
volcanic ash soils from Kenai Peninsula, Alaska. Soil Sci. 145, 395 –413.
Shoji, S., Takahashi, T., Saigusa, M., Yamada, I., and Ugolini, F. C. (1988b). Properties of Spodosols
and Andisols showing climosequential and biosequential relations in southern Hakkoda,
northeastern Japan. Soil Sci. 145, 135–150.
Simkin, T., and Siebert, L. (1994). “Volcanoes of the world: a regional directory, gazetteer, and
chronology of volcanism during the last 10,000 years.” 2nd edn. p. 349. Geoscience Press,
Tucson, AZ.
Simonett, D. A., and Bauleke, M. P. (1963). Mineralogy of soils on basalt in North Queensland. Soil
Sci. Soc. Am. Proc. 27, 205–212.
Simonson, R. W. (1979). Origin of the name “Ando soils”. Geoderma. 22, 333– 335.
Soil Survey Staff, (1960). “Soil Classification, a Comprehensive System, 7th Approximation.” U.S.
Dept. Agric., Soil Conservation Service, Washington D.C.
Soil Survey Staff, (1975). “Soil Taxonomy: A Basic System of Soil Classification for Making and
Interpreting Soil Surveys,”. USDA Agric. Handb. No. 436. U.S. Govt. Printing Office,
Washington, DC.
Soil Survey Staff, (1998). “Keys to Soil Taxonomy”, 8th ed. USDA Natural Resources Conservation
Service, U.S. Govt. Printing Office, Washington DC.
Soma, M., Churchman, G. J., and Theng, B. K. G. (1992). X-ray photoelectron spectroscopic analysis
of halloysites with different composition and particle morphology. Clay Miner. 27, 413–421.
Southard, R. J., and Southard, S. B. (1987). Sand-sized kaolinized feldspar pseudomorphs in a
California Humult. Soil Sci. Soc. Am. J. 51, 1666–1672.
Stanford, G., and Smith, S. J. (1972). Nitrogen mineralization potentials of soils. Soil Sci. Soc. Am.
Proc. 36, 465– 472.
Stumm, W. (1992). “Chemistry of the Solid–Water Interface.”, p. 428. Wiley, New York.
Sudo, T., and Nakamura, T. (1952). Hisingerite from Japan. Am. Mineral. 37, 618 –621.
Sugihara, S., Konno, T., and Ishii, K. (1986). Kinetics of mineralization of organic nitrogen in soil.
Bull. Nat. Inst. Agro-Environ. Sci. 1, 127 –166.
Sverdrup, H. U. (1990). “The Kinetics of Base Cation Release Due To Chemical Weathering”, p. 246.
Lund University Press, Lund, Sweden.
Sverdrup, H., and Warfvinge, P. (1988). Weathering of primary silicate minerals in the natural soil
environment in relation to a chemical weathering model. Water, Air, Soil Pollut. 38, 387–408.
Tachibana, N., Yoshikawa, S., and Ikeda, K. (1995). Influences of heavy application of nitrogen on
soil acidification and root growth in tea fields. Jpn. J. Crop Sci. 64, 516–522.
Takahashi, T., and Dahlgren, R. A. (1998). Possible control of aluminum solubility by 1 M KCl
treatment in some soils dominated by aluminum– humus complexes. Soil Sci. Plant Nutr. 44,
43 –51.
Takahashi, T., Shoji, S., and Sato, A. (1989). Clayey Spodosols and Andisols showing a biosequential
relation from Shimokita Peninsula, northeastern Japan. Soil Sci. 148, 204 –218.
Takahashi, T., Dahlgren, R. A., and van Susteren, P. (1993). Clay mineralogy and chemistry of soils
formed in volcanic materials in the xeric moisture regime of northern California. Geoderma. 59,
131–150.
Takahashi, T., Fukuoka, T., and Dahlgren, R. A. (1995). Aluminum solubility and release rates from
soil horizons dominated by aluminum–humus complexes. Soil Sci. Plant Nutr. 41, 119–131.
Takahashi, T., Dahlgren, R. A., Theng, B. K. G., Whitton, J. S., and Soma, M. (2001). Potassium-
selective, halloysite-rich soils formed in volcanic materials from northern California. Soil Sci.
Soc. Am. J. 65, 516–526.
Talibudeen, O. (1981). Precipitation. In “The Chemistry of Soil Processes” (D. J. Greenland and M. H.
B. Hayes, Eds.), pp. 81–114. Wiley, New York.
Tate, K. R., and Theng, B. K. G. (1980). Organic matter and its interactions with inorganic soil
constituents. In “Soils with Variable Charge” (B. K. G. Theng, Ed.), pp. 225 –249. Soil Bureau,
Lower Hutt, New Zealand.
Tawaraya, K., Kunii, Y., and Wagatsuma, T. (1995). Effects of arbuscular mycorrhizal inocula-
tion and phosphate application on phosphorus uptake and growth of white clover (Trifolium
repens L.) and onion (Allium cepa L.) in Andosols. Jpn. J. Soil Sci. Plant Nutr. 66, 48–53.
Theng, B. K. G. (1980). “Soils with Variable Charge”, p. 448. Soil Bureau, Lower Hutt, New Zealand.
Tisdall, J. M., and Oades, J. M. (1982). Organic matter and water-stable aggregates in soils. J. Soil Sci.
33, 141 –163.
Toda, T., Koyama, H., Hori, T., and Hara, T. (1999). Aluminum tolerance of Arabidopsis thaliana
under hydroponic and soil culture conditions. Soil Sci. Plant Nutr. 45, 419–425.
Tokashiki, T., and Wada, K. (1975). Weathering implications of the mineralogy of clay fractions of
two Ando soils, Kyushu. Geoderma. 14, 47–62.
Toma, M., and Saigusa, M. (1997). Effects of phosphogypsum on amelioration of strongly acid
nonallophanic Andosols. Plant Soil. 192, 49– 55.
Toma, M., Hiradate, S., and Saigusa, M. (1999). Chemical species of Al in a gypsum-treated Kitakami
Andosol. Soil Sci. Plant Nutr. 45, 279 –285.
Torn, M. S., Trumbore, S. E., Chadwick, O. A., Vitousek, P. M., and Hendricks, D. M. (1997). Mineral
control of soil organic carbon storage and turnover. Nature. 389, 170–173.
Towe, K. M., and Bradley, W. F. (1967). Mineralogical constitution of colloidal “hydrous ferric
oxides.” J. Colloid Interface Sci. 24, 384–392.
Tsuji, H., Yamamoto, H., Matsuo, K., and Usuki, K. (2000). Effect of nontilled cropping on drought
injury of upland rice. Jpn. J. Crop Sci. 69, 20–23.
Ugolini, F. C. (1968). Soil development and alder invasion in a recently deglaciated area of Glacier
Bay, Alaska. In “Biology of the Alder” (G. M. Trappe, J. F. Franklin, R. F. Tarrant, and G. M.
Hansen, Eds.), pp. 115–140. U.S. Forest Service, Pacific Northwest Forest and Range
Experiment Station, Portland, Oregon.
Ugolini, F. C., and Dahlgren, R. A. (1987). The mechanism of podzolization as revealed by soil
solution studies. In “Podzols et Podzolization” (D. Righi and A. Chauvel, Eds.), pp. 195– 203.
Assoc. Fr. Estude Sol, Plasisier, France.
Ugolini, F. C., and Dahlgren, R. A. (1991). Weathering environments and occurrence of imogolite/
allophane in selected Andisols and Spodosols. Soil Sci. Soc. Am. J. 55, 1166–1171.
Ugolini, F. C., and Dahlgren, R. A. (2002). Soil development in volcanic ash. Global Environ. Res. 6,
69–81.
Ugolini, F. C., and Zasoski, R. J. (1979). Soils derived from tephra. In “Volcanic Activity and Human
Ecology” (P. D. Sheets and K. K. Grayson, Eds.), pp. 83–124. Academic Press, New York.
Ugolini, F. C., Dahlgren, R. A., Shoji, S., and Ito, T. (1988). An example of andosolization
and podzolization as revealed by soil solution studies, S. Hakkoda, N.E. Japan. Soil Sci. 145,
111–125.
Ugolini, F. C., Dahlgren, R. A., LaManna, J., Nuhn, W., and Zachara, J. (1991). Mineralogy and
weathering processes in recent and Holocene tephra deposits of the Pacific Northwest, USA.
Geoderma. 51, 277 –299.
Ugolini, F. C., Minden, R., Dawson, H., and Zachara, J. (1977). An example of soil processes in the
Abies amabilis zone of central Cascades, Washington. Soil Sci. 124, 291– 302.
Vaje, P. I., Singh, B. R., and Lal, R. (2000). Leaching and plant uptake of nitrogen from a volcanic ash
soil in Kilimanjaro Region, Tanzania. J. Sustain. Agric. 16, 95–112.
Van Olphen, H. (1971). Amorphous clay materials. Science. 171, 90–91.
Van Ranst, E., Utami, S. R., and Shamshuddin, J. (2002). Andisols on volcanic ash from Java Island,
Indonesia: physico-chemical properties and classification. Soil Sci. 167, 68–79.
Wada, K. (1977). Active aluminum in Kuroboku soils and non- and para-crystalline clay minerals.
Nendo Kagaku. 17, 143 –151.
Wada, K. (1980). Mineralogical characteristics of Andisols. In “Soils with Variable Charge” (B. K. G.
Theng, Ed.), pp. 87 –107. New Zealand Society of Soil Science, Lower Hutt.
Wada, K. (1985). The distinctive properties of Andosols. Adv. Soil Sci. 2, 173 –229.
Wada, K. (1986). “Ando Soils in Japan”, p. 276. Kyushu University Press, Fukuoka, Japan.
Wada, K. (1989). Allophane and imogolite. In “Minerals in Soil Environments, 2nd edn.” (J. B. Dixon
and S. B. Weed, Eds.), pp. 1051–1087. Soil Sci. Soc. Am., Madison, WI.
Wada, K., and Harward, M. E. (1974). Amorphous clay constituents of soils. Adv. Agron. 26,
211 –260.
Wada, K., and Higashi, T. (1976). The categories of aluminum– and iron –humus complexes in Ando
soils determined by selective dissolution. J. Soil Sci. 27, 357– 368.
Wada, S.-I. (1987a). Critical evaluation of 1 M KCl-extraction method for determining exchangeable
Al ions in variable charge soils. Soil Sci. Plant Nutr. 33, 153–160.
Wada, S.-I. (1987b). Adsorption of Al(III) on allophane, imogolite, goethite, and noncrystalline silica
and the extractability of the adsorbed Al(III) in 1 M KCl solution. Soil Sci. Plant Nutr. 33,
487 –491.
Wada, S.-I., and Mizota, C. (1982). Iron-rich halloysite (10Å) with crumpled lamellar morphology
from Hokkaido, Japan. Clays Clay Miner. 30, 315–317.
Wagatsuma, T., and Ezoe, Y. (1985). Effect of pH on ionic species of aluminum in medium and on
aluminum toxicity under solution culture. Soil Sci. Plant Nutr. 31, 547 –561.
Wakatsuki, T., and Wielemaker, W. G. (1985). Clay mineralogy of a soil formed in peralkaline
volcanic ash from the Great Rift Valley in Kenya. Soil Sci. Plant Nutr. 31, 475 –480.
Warkentin, B. P., and Maeda, T. (1980). Physical and mechanical characteristics of Andosols.
In “Soils with Variable Charge” (B. K. G. Theng, Ed.), pp. 97 –107. N.Z. Soil Bureau, Lower
Hutt, New Zealand.
Warkentin, B. P., Maeda, T., and Soma, K. (1988). Physical characteristics for classification of
Andisols. In “Proceedings 9th Int. Soil Classification Workshop, Japan” (D. I. Kinloch, S. Shoji,
F. M. Beinroth, and H. Eswaran, Eds.), pp. 97 –107. Japanese Committee for 9th International
Soil Classification Workshop and Soil Management Support Services, Washington, DC.
Wells, N., Childs, C. W., and Downes, C. J. (1977). Silica Springs, Tongariro National Park,
New Zealand: analyses of the spring water and characterization of the alumio-silicate deposit.
Geochim. Cosmochim. Acta. 41, 1497– 1506.
White, A. F. (1983). Surface chemistry and dissolution kinetics of glassy rocks at 258C. Geochim.
Cosmochim. Acta. 47, 805– 815.
White, A. F., and Claassen, H. C. (1980). Kinetic model for short-term dissolution of a rhyolitic glass.
Chem. Geol. 29, 91 –109.
White, A. F., Benson, L. V., and Yee, A. (1986). Chemical weathering of the May 18, 1980 Mount St.
Helens ash fall and the effect on the Iron Creek Watershed, Washington. In “Rates of Chemical
Weathering of Rocks and Minerals” (S. M. Colman and D. P. Detheir, Eds.), pp. 351–375.
Academic Press, Orlando.
Wielemaker, W. G., and Wakatsuki, T. (1984). Properties, weathering and classification of some soils
formed in peralkaline volcanic ash in Kenya. Geoderma. 32, 21 –33.
Wilding, L. P. (2000). Classification of soils. In “Handbook of Soil Science” (M. E. Sumner, Ed.),
pp. E175–183. CRC Press, Boca Raton, FL.
Wilson, M. A., Burt, R., Sobecki, T. M., Engel, R. J., and Hipple, K. (1996). Soil properties and
genesis of pans in till-derived Andisols, Olympic Peninsula, Washington. Soil Sci. Soc. Am. J. 60,
206–218.
Wilson, M. A., Lee, G. S. H., and Taylor, R. C. (2001). Tetrahedral rehydration during imogolite
formation. J. Non-Crystall. Solids. 296, 172– 181.
Yamada, I. (1988). Tephra as parent material. In “Proceedings 9th Int. Soil Classification Workshop,
Japan” (D. I. Kinloch, S. Shoji, F. M. Beinroth, and H. Eswaran, Eds.), pp. 509–519. Japanese
Committee for 9th International Soil Classification Workshop and Soil Management Support
Services, Washington, DC.
Yamada, I., and Shoji, S. (1980). Fertility of Andosols formed from basaltic and dacitic volcanic
ashes. Tohoku J. Agri. Res. 31, 149– 155.
Yamada, I., Saigusa, M., and Shoji, S. (1978). Clay mineralogy of Hijiori and Numazawa Ando soils.
Soil Sci. Plant Nutr. 24, 75–89.
Yamanaka, K. (1964). Physical properties. In “Volcanic Ash Soils in Japan” (Ministry of Agriculture
and Forestry, Ed.), pp. 69–91. Japanese Govt, Tokyo.
Yerima, B. P. K., Wilding, L. P., Calhoun, F. G., and Hallmark, C. T. (1987). Volcanic ash-influenced
Vertisols and associated Mollisols of El Salvador–physical, chemical, and morphological
properties. Soil Sci. Soc. Am. J. 51, 699–708.
Yoshinaga, N. (1983). “Volcanic Ash Soil — Genesis, Properties, Classification”, Hakuyuisha, Tokyo.
Yoshinaga, N. (1988). Mineralogy of Andisols. In “Proceedings 9th Int. Soil Classification Workshop,
Japan” (D. I. Kinloch, S. Shoji, F. M. Beinroth, and H. Eswaran, Eds.), pp. 45–59. Japanese
Committee for 9th International Soil Classification Workshop and Soil Management Support
Services, Washington, DC.
Yoshinaga, N., and Aomine, S. (1962). Imogolite in some Ando soils. Soil Sci. Plant Nutr. 8, 22 –29.
Zou, X. M., Binkley, D., and Doxtader, K. G. (1992). A new method for estimating gross phosphorus
mineralization and immobilization rates in soils. Plant Soil. 147, 243–250.
Zunino, H., Borie, F., Aguilera, S., Martin, J. P., and Haider, K. (1982). Decomposition of 14C-labeled
glucose, plant and microbial products and phenols in volcanic ash-derived soils of Chile. Soil
Biol. Biochem. 14, 37–43.
MICROARRAY TECHNOLOGY AND
APPLICATIONS IN ENVIRONMENTAL
MICROBIOLOGY
Jizhong Zhou and Dorothea K. Thompson
Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge,
Tennessee 37831, USA
I. Introduction
II. Microarray Types and Advantages
A. Types of Microarrays
B. Advantages of Microarrays
III. Microarray Fabrication
A. Microarray Substrates
B. Surface Modification for the Attachment of Nucleic Acids
C. Arraying Technology
D. Critical Issues for Microarray Fabrication
IV. Microarray Hybridization and Detection
A. Probe Design and Synthesis
B. Target Labeling and Quality
C. Hybridization
D. Detection
E. Critical Issues in Hybridization and Detection
V. Microarray Image Processing
A. Data Acquisition
B. Assessment of Spot Quality and Background Subtraction
VI. Microarray Data Analysis
A. Data Normalization
B. Data Transformation
C. Methods for Identifying Differentially Expressed Genes
D. Microarray Data Analysis
VII. Using Microarrays to Monitor Genomic Expression
A. General Approaches to Revealing Differences in Gene
Expression
B. Experimental Design for Microarray-based Monitoring of
Gene Expression
C. Microarray-based Functional Analysis of Environmental
Microorganisms
VIII. Application of Microarrays to Environmental Studies
A. Functional Gene Arrays
B. Phylogenetic Oligonucleotide Arrays
C. Community Genome Arrays
183
0065-2113/03 $35.00
184 J. ZHOU AND D. K. THOMPSON
D. Whole-genome Open Reading Frame Arrays for Revealing

Genome Differences and Relatedness
E. Other Types of Microarrays for Microbial Detection and
Characterization
IX. Concluding Remarks
References
I. INTRODUCTION
Microarrays are miniaturized arrays of hundreds to thousands of discrete

DNA fragments or synthetic oligonucleotides that have been attached to a solid
substrate (e.g., glass) using automated printing equipment such that each spot
(element) in a fixed position on the array corresponds to a unique DNA (Schena
et al., 1998; Schena, 2003). Microarrays are variously referred to as microchips,
biochips, DNA chips, or gene chips and have emerged as a widely accepted
functional genomics technology for large-scale genomic analysis. In particular,
DNA or oligonucleotide arrays have been used to monitor messenger RNA
(mRNA or transcript) abundance levels of differentially expressed genes under
different cell growth conditions or in response to environmental perturbations or
genetic mutations (c.f., Lockhart et al., 1996; Schena et al., 1996; DeRisi et al.,
1997; Wodicka et al., 1997; Richmond et al., 1999; Ye et al., 2000; Thompson
et al., 2002; Liu et al., 2003) and to detect specific mutations in DNA sequences
(Hacia, 1999; Broude et al., 2001). Recently, the potential research applications
of microarray technology to studies in microbial ecology have been explored
(Zhou and Thompson, 2002; Zhou, 2003).
In principle and practice, microarrays are extensions of conventional
membrane-based Southern and Northern hybridization blots, which have been
used for decades to detect and characterize nucleic acids in diverse biological
samples. Microarray hybridization is based on the association of a single-
stranded molecule labeled with a fluorescent tag, or fluorescein, with its
complementary molecule, which is covalently attached or immobilized to a solid
support, usually glass. In such an assay, the specific hybridization pattern or gene
expression profile generated by an unknown (experimental) sample is typically
compared with a control (reference) pattern. In microarray terminology, the
fluorescein-labeled DNA in solution is generally termed the target, and the DNA
strand immobilized on the microarray surface is referred to as the probe. Because
the sequence of the arrayed molecule is usually known, it is used to ‘probe’ or
investigate the unknown target molecule in solution. This is directly opposite to
the convention established with the development of Southern blot hybridization,
MICROARRAY TECHNOLOGY 185
in which target molecules fixed to a porous membrane are interrogated by known

solution-phase probes.
The concept of microarrays was first proposed in the late 1980s. One of the first
descriptions of DNA microarrays in the literature was provided by Augenlicht
and his colleagues, who spotted 4000 complementary DNA (cDNA) sequences
on nitrocellulose and used radioactive labeling to analyze differences in gene
expression patterns among different types of colon tumors in various stages of
malignancy (Augenlicht et al., 1987, 1991). At the same time, four separate
research groups simultaneously developed the concept of determining a DNA
sequence by hybridization to a comprehensive set of oligonucleotides, i.e.,
sequencing by hybridization or SBH (Bains and Smith, 1988; Drmanac et al.,
1989; Khrapko et al., 1989; Southern et al., 1992). Although SBH is an extremely
elegant alternative to conventional DNA sequencing, various inherent problems
associated with repeated sequences and the imperfect specificity of hybridization
limit the practicality of using SBH for routine sequence determination. Such
technical challenges, therefore, have led researchers to focus on the more readily
addressable applications of microarray technology, such as gene expression
profiling. By the mid-1990s, the reverse dot-blot scheme for monitoring genomic
expression was reformulated by several different groups. Both DNA fragments
and synthetic oligonucleotides were arrayed on various substrates, including
nylon membranes, plastic and glass (Schena et al., 1995; Lockhart et al., 1996).
All of them depended on sequence-specific hybridization between the arrayed
DNA and the labeled nucleic acids from cellular mRNA. Later studies using the
simple eukaryote, yeast, clearly demonstrated that DNA and oligonucleotide
arrays are powerful tools for monitoring global gene expression (DeRisi et al.,
1997; Wodicka et al., 1997).
Microarray-based genomic technology has greatly benefited from many
parallel advances in other fields. Without such advancements, the development of
high-density microarrays and the various applications that we see today would
not be possible (Eisen and Brown, 1999; Schena and Davis, 2000). For example,
large-scale genome sequencing projects have produced the raw sequence
information needed for microarray expression profiling, and the development
of robotic printers or arrayers has made it possible to fabricate high-density
microarrays in a very small area. In addition, recent advances in methods of
fluorescent labeling and detection offer significant advantages in speed, data
quality, and user safety for microarray-based assays. Together, these technical
advancements have enabled microarray-based genomic technologies to revolu-
tionize genetic analyses of biological systems. The widespread, routine use of
such genomic technologies will shed light on a wide range of important research
questions associated with the genetic programs controlling cell growth
and differentiation, bacterial pathogenesis and the host response to infection,
antibiotic resistance, specialized metabolic capabilities of microorganisms of
bioremediation potential, as well as agricultural and pharmaceutical applications.
In this chapter, we review the technical underpinnings of microarray

hybridization and its applications in environmental microbiology, with emphasis
on the most important issues of microarray-based assays as outlined in Fig. 1. We
first describe the technologies used for microarray fabrication, followed by a
discussion on microarray hybridization, fluorescence detection technologies,
image processing, and data analysis. In addition, we provide an overview of the
recent applications of microarray technologies to study gene expression patterns
in environmentally important microorganisms. Finally, we describe various types
of microarrays specifically developed for analyzing microbial community
composition and function in natural environments. Because glass-based DNA
microarrays are currently preferred by most basic research laboratories, our
discussion of microarray technologies will focus on this type of array, while other
types will be mentioned only briefly. It should be noted that the goal of this
chapter is to provide an in-depth description of the basis and principles of
Figure 1 A flow chart of a microarray experiment.

microarray-based technologies rather than an exhaustive review of current

microarray technology.
II. MICROARRAY TYPES AND ADVANTAGES
A. TYPES OF MICROARRAYS
Microarrays can be divided into two major formats based on the type of
immobilized probes: (1) DNA microarrays constructed with DNA fragments
typically generated using the polymerase chain reaction (PCR) (Schena et al.,
1995; DeRisi et al., 1997; Marshall and Hodgson, 1998); and (2) oligonucleotide
microarrays constructed with shorter (10- to 40-mer) or longer (up to 120-mer)
oligonucleotide sequences that are designed to be complementary to specific
coding regions of interest.
DNA microarrays have certain advantages over oligonucleotide microarrays,
especially for monitoring gene expression patterns. While oligonucleotide
microarrays are limited to array elements of low sequence complexity, the
specificity of hybridization for a complex probe is improved with arrays
containing DNA fragments that are substantially longer than oligonucleotides
(Shalon et al., 1996). In addition, oligonucleotide synthesis requires prior
sequence knowledge, whereas DNA arrays do not because DNA fragments of
unknown sequences can be amplified from clones using vector-specific primers.
For microarrays constructed with PCR-amplified DNA elements, nucleic acids of
virtually any length, composition, or origin can be arrayed (Shalon et al., 1996).
However, oligonucleotide-based microarrays have the advantage of minimizing
the potentially confounding effects of occasional cross-hybridization (Wodicka
et al., 1997) and are uniquely suited for detecting genetic mutations and
polymorphisms. Since oligonucleotide probes can be commercially synthesized,
the handling and tracking of oligonucleotide array elements, unlike PCR
products, is generally easier. Amplifying all of the probes with a desired
minimum quantity for printing is labor intensive and time-consuming.
Based on probe immobilization and fabrication strategies, there are two
general types of oligonucleotide microarrays:
1. Direct parallel synthesis on solid substrates by light-directed or photoactiva-

table chemistries (Pease et al., 1994; Lipshultz et al., 1999) or standard
phosphoramidite chemistries (Southern et al., 1994); or
2. Chemical attachment of pre-made oligonucleotides to solid supports (Khrapko
et al., 1989; Beattie et al., 1992, 1995; Eggers et al., 1994; Lamture et al.,
1994; Fotin et al., 1998; Guschin et al., 1997a, b; Rehman et al., 1999; Rogers
et al., 1999).
Each strategy for oligonucleotide immobilization has its own set of specific
advantages and disadvantages (Schena et al., 1998; Hoheisel, 1997). The direct
or in situ synthesis approach has two major advantages. First, the
photoprotected versions of the four DNA bases allow microarrays to be manu-
factured directly from DNA sequence databases, thereby removing the uncertain
and burdensome aspects of sample handling and tracking. Second, the use of
synthetic reagents minimizes variations between arrays by ensuring a high
degree of precision in each coupling cycle. Costliness is a major disadvantage
of the photolithographic approach; photomasks, which direct light to specific
areas on the array for localized chemical synthesis, are very expensive and time-
consuming to design and build. Also, the yield and length of the synthesized
oligonucleotides are subject to wide variation and uncertainty, which could lead
to unpredictable effects on hybridization across the microarray. A major
advantage of the attachment of pre-synthesized probes is that the concentration
and length of each oligonucleotide on the array can be controlled prior to
immobilization. Standard synthesis chemistry is also well established for many
nucleotide derivatives for which no light-inducible monomer equivalents are
available. In addition, the post-synthesis approach is less complicated and can
be customized according to the specific needs of the laboratory. The critical
drawback of the post-synthesis approach, however, is still the need for the
external synthesis and storage of different oligonucleotides prior to array
fabrication.
B. ADVANTAGES OF MICROARRAYS
Microarrays offer the following advantages over conventional nucleic acid-

based approaches.
1. High-throughput and Parallel Analysis
Microarray technology allows thousands to hundreds of thousands of

array elements or probes to be uniformly deposited in a very small area
on the surface of a non-porous substrate. Consequently, the high-density
capacity of microarrays permits parallel analysis, in which the expression of
the entire gene content of a genome of interest can be monitored, or many
constituents of a microbial community can be simultaneously assessed in a
single assay using the same microarray. Genomic expression data allows
researchers to begin to build a comprehensive, integrated view of a complex
biological system.
2. High Sensitivity
High sensitivity can be achieved in probe-target hybridization, because

microarray hybridization uses a very small volume of probe and the target
nucleic acid is restricted to a small area (Shalon et al., 1996; Guschin
et al., 1997a). This feature enables high sample concentrations and rapid
hybridization kinetics.
3. Differential Display
Multi-fluorescence detection schemes allow differential display of different

biological samples. Different target samples, for example, can be labeled with
different fluorescent tags and then hybridized in parallel to the same microarray,
allowing the simultaneous analysis of two or more biological samples in a single
assay. Multi-color hybridization detection minimizes variations resulting from
inconsistent experimental conditions and allows direct and quantitative
comparison of target sequence abundance among different biological samples
(Shalon et al., 1996; Ramsay, 1998).
4. Low Background Signal Noise
Non-porous surfaces substantially reduce the amount of non-specific

hybridization; as a result, organic and fluorescent compounds that attach to
microarrays during fabrication and hybridization procedures can be rapidly
removed by post-hybridization washing, resulting in considerably less back-
ground signal noise than is typically encountered with porous membranes
(Shalon et al., 1996).
5. Real-time Data Analysis
Once the microarrays are constructed, hybridization and detection are

relatively simple and rapid, allowing real-time data analysis in field-scale
heterogeneous environments.
6. Automation
Microarray technology is amenable to automation and therefore, has the

potential of being cost-effective compared to traditional hybridization methods
(Shalon et al., 1996).
III. MICROARRAY FABRICATION
A. MICROARRAY SUBSTRATES
The substrate used for printing microarrays has a large impact ultimately on
the quality of the data obtained from microarray hybridizations. Poor surface
treatment may result in poor attachment of the DNA probes to the slide, and a
non-uniform surface will cause variations in the amount of the attached DNA.
Furthermore, residual substances deposited on the slide surface during a
microarray experiment can lead to high background fluorescence or noise.
Thus, the selection of appropriate substrates for microarray experiments is of
critical importance. The substrates used for fabricating microarrays fall into two
categories: porous and non-porous.
1. Non-porous Substrate
At present, a non-porous solid surface is the most common type of substrate

used for printing arrays. Several non-porous materials, such as glass and
polypropylene, are suitable for microarray fabrication. Glass slides are the most
widely used substrates, because they are inexpensive, possess physical
characteristics advantageous to hybridization, and are easily modified for nucleic
acid attachment and synthesis (Southern, 2001). In general, non-porous substrates
offer a number of advantages over porous substrates (Schena and Davis, 2000).
First, small amounts of molecules may be deposited at precise, predefined
positions on the substrate surface with little diffusion, thus enabling the high-
density capacity of microarrays. Second, hybridization between target and probe
molecules occurs at a much faster rate on non-porous solid surfaces than on
porous substrates. This is because molecules do not have to diffuse in and out of
the pores and as a result, steric inhibition of hybridization is not a problem
(Southern, 2001). Third, because small sample volumes can be applied to a non-
porous surface under a coverslip, high probe concentrations, rapid hybridization
kinetics, and high sensitivity can be achieved. Fourth, non-porous substrates
prevent the absorption of reagents and samples into pores, thus allowing unbound
labeled materials to be easily removed. This expedites the procedure, improves
reproducibility and reduces background. Fifth, a non-porous substrate has low
intrinsic fluorescence and thus allows the use of fluorescence detection. Finally, a
solid substrate offers a homogeneous attachment surface, and its inherent uniform
flatness permits true parallel analysis. However, a major drawback of non-porous
substrates, such as glass, is the susceptibility to dust and other airborne particle
contamination, which can cause a scanned slide to appear “dirty”. Poor
modification of microarray slides is another common cause of poor hybridization
results (Southern, 2001). In addition, because of the planar surface, the capacity
for immobilization is limited, and consequently, the sensitivity of the assay is
relatively low compared to that of the porous substrates (Afanassiev et al., 2000).
2. Porous Substrates
Porous substrates such as nitrocellulose and nylon membranes have also been
used for microarray fabrication (Englert, 2000). The principal advantage of
membranes is that larger volumes and concentrations of samples can be immo-
bilized on a small area, because the pores of the substrates provide a larger total
surface area for binding. As a result, a relatively higher sensitivity and better
dynamic range for quantitative comparison can be achieved. Homogeneous spots
are also more readily obtained, because deposited samples are able to distribute
immediately into the membrane through capillary flow. In addition, membrane-
based microarrays can be reused several times (Beier and Hoheisel, 1999).
However, there are some important disadvantages in using porous membranes
for microarray fabrication. The boundaries and shapes of the spots are poorly
defined, and membranes swell in solvent, and shrink and distort when dried. Such
fragility and flexibility make it difficult to precisely locate probe positions during
spotting and image analysis. Also, many membranes have high intrinsic
fluorescence and thus higher background noise compared to non-porous
substrates. In addition, because the spot sizes on a membrane cannot be reduced
to a level comparable with glass slides or other non-porous substrates, much more
DNA is required for producing a membrane-based microarray (Beier and
Hoheisel, 1999).
Overall, non-porous substrates are preferred for microarray experiments, even
though porous substrates have some advantages. This is because the unique
physical and chemical characteristics of glass slides (e.g., little diffusion, low
intrinsic fluorescence) allow miniaturization and use of fluorescent labeling
and detection, which are the most critical requirements for large-scale
genomic analysis. The miniaturized microarray format coupled with fluorescent
detection represents a fundamental revolution in biological analysis (Schena
and Davis, 2000).
B. SURFACE MODIFICATION FOR THE ATTACHMENT

OF NUCLEIC ACIDS
1. Attachment Strategies
Appropriate attachment and retainment of nucleic acid probes to an array

surface is very important for microarray analysis. For reliable microarray
hybridization, the attachment chemistry must meet the following criteria:

(i) nucleic acids must be tightly (covalently) bound to the array surface;
(ii) the surface-bound molecules must be accessible for hybridization;
and (iii) the attachment chemistry must be reproducible. Both ionic interaction
and covalent bonding, which are described in more detail below, are used for
attaching nucleic acids to solid surfaces, depending on the size of the nucleic acid
molecules.
Electrostatic interaction. Long DNA fragments (in the order of several hundred
bases in length) can be immobilized on a glass surface through ionic interaction
between the negatively charged phosphodiester backbone and the positively
charged slide surface (Fig. 2A). Recent studies showed that synthetic
oligonucleotides more than 70 bp in length can also be bound to glass surfaces
through ionic interaction (Hughes et al., 2001). Generally, an amine or lysine
coating is used to adsorb DNA to glass slides. Because amines have a positive
charge at neutral pH, they allow attachment of native DNA molecules through
the formation of ionic bonds with the negatively charged phosphate backbone.
Electrostatic attachment can be enhanced by exposing the fabricated arrays to
ultraviolet light or heat, which induces free radical-based coupling between
thymidine residues in the DNA and carbons on the alkyl amine. The combination
of electrostatic bonding and non-specific covalent attachment, links native DNA
to the substrate surface in a stable manner. Although ionic interaction-mediated
attachment is less expensive and more versatile than covalent bonding (Worley
et al., 2000), the immobilized DNA molecules are susceptible to removal under
high salt and/or high temperature conditions. Therefore, covalent binding
methods are preferred.
Covalent bonding. DNA can also be covalently attached to glass surfaces using
different attachment chemistries (Fig. 2B). Although long DNA molecules can be
attached covalently to the microarray surface by different methods, immobiliz-
ation of aminated DNA to an aldehyde-coated slide is the usual method of choice
(Zammatteo et al., 2000).
Because oligonucleotides are typically short, covalent bonding is generally
required for attachment of such molecules to a glass surface. Usually,
oligonucleotides are fixed covalently onto solid surfaces at one end of the
molecule using a variety of methods. The attachment of biomolecules to a solid
phase presents some problems that are unique to homogeneous solutions.
Because the bound probe is not free to diffuse, a lower reaction rate is expected.
In addition, target molecules in solution may not be able to effectively interact
with the bound probes due to steric hindrance from the solid support and the
close proximity of other bound probes (Shchepinov et al., 1997). Additional
molecules, termed linkers or spacers, are used to tether the probe to the
substrate surface, thereby providing a sufficient amount of distance between the
Figure 2 Attachment strategies. (A) Attachment of nucleic acids to solid surfaces through
electrostatic interactions. The microarray substrates contain primary amine groups (NHþ3 ) attached
covalently to the glass surface. The amines carry a positive charge at neutral pH, which permits
attachment of native DNA through the formation of ionic bonds with the negatively charged
phosphate backbone. Covalent attachment of DNA to the surface can be further achieved by treatment
with ultraviolet light or heat. (B) Attachment of nucleic acids to solid surfaces through covalent
bonding. The microarray substrates contain primary aldehyde groups attached covalently to the glass
surface. Primary amino linkers (NH2) on the DNA attack the aldehyde groups to form covalent bonds.
Such attachment is stabilized with a dehydration reaction by drying in low humidity, which leads to
Schiff base covalent bond formation.
oligonucleotide probe and the support to minimize steric interference. To serve as

an effective linker, the molecule must meet several criteria (Guo et al., 1994).
First, the linkage must be chemically stable under the hybridization conditions
used and must be sufficiently long to minimize steric interference. Second, the
linker should be sufficiently hydrophilic to be freely soluble in aqueous solution.
Third, there should be no non-specific binding of the linker to the support.
Shchepinov et al. (1997) showed that the optimal linker for immobilizing
oligonucleotides at either the 50 or 30 terminus should have low negative charge
density and a length of 30– 60 atoms.
Generally, a linker is coupled to the probe during oligonucleotide synthesis.

Various types of linkers have been used, including poly-dT (Guo et al., 1994;
Rehman et al., 1999; Afanassiev et al., 2000), poly carbon atoms (Afanassiev et
al., 2000), and oligodeoxyribonucleotides with hairpin stem-loop structures
(Zhao et al., 2001). Studies have shown that the length of the spacer has a
significant impact on the success of hybridization. While virtually no
hybridization signal was observed for poly-dT spacers (15 bp), about a 20-fold
enhancement of hybridization was obtained for a poly-dT spacer of 15 nucleotides
(Guo et al., 1994). The effect of linkers composed of multiple carbon atoms (e.g.,
C36, C18, C12, and C6) on microarray hybridization was also examined
(Afanassiev et al., 2000). Overall, the signal intensity was improved with a
longer C linker. A spacer can also be added by chemically modifying the slide
surface (Guo et al., 1994; Beier and Hoheisel, 1999).
Oligonucleotides can be immobilized onto solid supports through homo-
bifunctional or hetero-biofunctional cross-linkers. For example, amino-modified
oligonucleotides can be covalently attached to glass surfaces containing amine
functional groups through homo-biofunctional cross-linkers (Guo et al., 1994)
and to glass surfaces containing aldehyde and epoxide through hetero-
biofunctional cross-linkers (Lamture et al., 1994; Schena et al., 1996). Thiol-
modified or disulfide-modified oligonucleotides can be attached to the glass
surface containing amine functional groups via hetero-biofunctional cross-
linkers (Chrisey et al., 1996). A hetero-biofunctional cross-linker is
generally preferred over a homo-biofunctional cross-linker to prevent surface-
to-surface linkages and probe-to-probe linkages as opposed to the desired
surface-to-probe linkages (Steel et al., 2000). When using a hetero-biofunc-
tional cross-linker, the probe should have a different modification chemistry
from the array surface.
While microscope slides made from low-fluorescence glass are suitable
substrates for microarray construction, the glass surface must be modified with a
chemical coating and cleaned (e.g., free of dust particles) before use. The glass
surface must have suitable functional groups for the attachment of target DNA
molecules, because DNA does not inherently bind to untreated glass. A
hydrophobic surface is essential for achieving high-density spots, because
spotted hydrophilic samples will spread less on a hydrophobic surface than on an
untreated hydrophilic glass surface (Rose, 2000).
The quality of slide coating ultimately impacts the quality of the microarray
data. A poor surface coating can result in poor probe retention. For spot-to-spot
consistency, the coating must be uniform and homogeneous without untreated
patches. In addition, the coating must be non-fluorescent and capable of resisting
harsh physical conditions, such as boiling, baking and soaking. Silanization,
dendrimeric linker coating, gel coating, and nitrocellulose coating are types of
glass surface modifications that are discussed in greater detail in the following
sections.
2. Silanization
Silanes are most commonly used for slide surface modification to provide
organic functional groups for the covalent attachment of biomolecules
(Shriver-Lake, 1998), and silanized glass slides are commercially available
from a number of companies. Glass slides can be modified to contain surface
hydroxyls that react with methoxy or ethoxy residues of a silane molecule.
Many different commercially available silanes contain various functional
reactive groups such as amino, epoxide, carboxylic acid and aldehyde, which
are suitable for covalent bonding with appropriately modified biomolecules
(Schena, 2003). Most microarray analyses are performed with slide surfaces
that contain reactive amine and aldehyde groups (Schena, 2003), which
allow attachment of biomolecules via electrostatic interactions or covalent
bonding. Silanization can be accomplished simply by immersing the slides
into a silane-containing solution or by vapor deposition (Steel et al., 2000;
Worley et al., 2000). Vapor-phase coating is most effective at uniformly
depositing a monolayer of silane on the slide surface (Chrisey et al., 1996;
Worley et al., 2000).
3. Dendrimeric Linker Coating
To increase the binding capacity of arrayed probes, a more elaborate

chemistry was developed for synthesizing dendrimeric linkers on silanized glass
slides to allow covalent attachment of aminated DNA molecules (Beier and
Hoheisel, 1999). Such a linker system multiplies the coupling sites by introducing
additional reactive groups through branched linker molecules. There are several
advantages of this linker system. First, it allows covalent immobilization of both
pre-synthesized and in situ synthesized oligonucleotides on glass slides. Second,
the dendrimeric linker system increases the loading capacity by a factor of 10.
Third, it eliminates non-specific attachment of hybridization probes and provides
a low fluorescent background. Fourth, covalent bonding is stable, and the
microarrays can be reused many times. Finally, bonding through a terminal group
of the attached molecules produces no apparent negative effects on hybridization
efficiency.
4. Gel Coating
Recently developed attachment strategies that use polyacrylamide or agarose

for surface modification combine the advantages of porous and non-porous
substrates (Afanassiev et al., 2000; Zlatanova and Mirzabekov, 2001). In this
approach, polyacrylamide gel elements or pads, ranging in size from 10 £ 10 £ 5
to 100 £ 100 £ 20 mm3 with volumes varying from picoliters to nanoliters, are
affixed to the glass surface. Because each gel-pad is surrounded by a hydrophobic
surface that prevents solution diffusion among the elements, they can function
independently. The probe molecules are immobilized in the gel-pads by robotic
application. Compared to the direct attachment of probes to solid supports, the
use of polyacrylamide gel-pad as an immobilization support offers some
important advantages (Drobyshev et al., 1999). Three-dimensional immobiliz-
ation of probes in gel-pads provides a greater density capacity and a more
homogeneous environment than heterophase immobilization on glass, leading to
higher sensitivity and a faster hybridization rate (Vasiliskov et al., 1999).
However, like nylon membrane-based supports, gel-pads can yield higher
background levels (Beier and Hoheisel, 1999). In addition, there are restrictions
on the size of the molecules that can diffuse into the gel, such that fragmentation
of the probe and target DNA may be required to generate molecules of the
appropriate size (Englert, 2000).
The convenience of using the gel-pad attachment strategy is limited by the fact
that the method requires activation of gels and probes with labile reactive
chemicals (Rehman et al., 1999). A more flexible attachment method using co-
polymerization of 50 -terminal modified oligonucleotides with acrylamide
monomers was developed (Rehman et al., 1999). The advantages of this method
are that probes can be prepared easily using standard DNA synthesis chemistries
and probes can be specifically and efficiently immobilized in the absence of
highly reactive and unstable chemical crosslinking agents.
Agarose was also examined as a coating material for probe attachment
(Afanassiev et al., 2000). Agarose film is activated to produce reactive sites
that permit covalent immobilization of molecules with amino groups. Agarose
has a higher binding capacity compared to a glass-based planar surface and does
not interfere with fluorescent detection. In contrast to acrylamide gels
(Zlatanova and Mirzabekov, 2001) and dendrimeric branched systems (Beier
and Hoheisel, 1999), this method does not require complex preparation
technology.
5. Nitrocellulose Coating
Glass slides coated with a proprietary nitrocellulose-based polymer have also

been examined as an immobilization support (Stillman and Tonkinson, 2000).
The nitrocellulose-based polymer can bind biomolecules (both DNA and
proteins) noncovalently but irreversibly, providing better spot-to-spot consist-
ency, higher binding capacity, and greater dynamic range compared to other glass
slide modifications. Nitrocellulose-coated slides are also suitable for fluorescent
detection due to their relatively low-light scattering capacity. Although they
have a higher fluorescent background, the background-subtracted signal is

significantly higher on this support than the glass support coated with polylysine.
However, the potential for miniaturizing the array dimensions of such slides
remains to be determined.
C. ARRAYING TECHNOLOGY
Microarray fabrication involves the printing and stable attachment of

DNA probes on the array (Fig. 1). The microarray format is compatible with
many advanced printing technologies, of which the most widely used are
photolithography, mechanical microspotting, and ink-jet ejection. Each fabrica-
tion technology possesses advantages and disadvantages (Schena et al., 1998;
Schena and Davis, 2000). All three technologies allow the manufacture of
microarrays with sufficient density for genetic mutation detection and gene
expression profiling applications. The key considerations in selecting a fabri-
cation technology include microarray density and design, biochemical compo-
sition and versatility, reproducibility, high-throughput capacity, and cost.
Because of its versatility, affordability, and wide applications, microspotting is
likely to become the printing technology of choice for the basic research
laboratory. Thus, our discussion in this section will focus primarily on
microspotting technology.
1. Light-directed Synthesis
In photolithography, oligonucleotides are synthesized in situ on a solid surface

in a predefined spatial pattern by using a combination of chemistry and
photolithographic methods borrowed from the semiconductor industry (Fodor
et al., 1991; McGall and Fidanza, 2001) (Fig. 3). Briefly, a glass or fused silica
substrate is covalently modified with a silane reagent to obtain a surface
containing reactive amine groups, which are then modified with a specific
photoprotecting group, namely methylnitropoperonyloxycarbonyl (MeNPOC).
Then the specific regions of the surface are activated through exposure to
light, and a single base is added to the hydroxyl groups of these exposed
surface regions using a standard phosphoramidite DNA synthesis method.
The process of photodeprotection and nucleotide addition is iterated until the
desired sequences are generated. Typically, the probes synthesized in situ on
the arrays are 20– 25 bp in length. Since the average stepwise efficiency of
oligonucleotide synthesis ranges from 90 – 95%, the proportion of the full-length
sequences for 20-mer probes is approximately 10%. However, this should have a
relatively minor effect on the performance of microarray hybridization because
Figure 3 In situ light-directed oligonucleotide probe array synthesis. The solid surface contains
linkers with a photolabile protecting group X (black box) (e.g., MeNPOC). MeNPOC is resistant to
many chemical reagents but it can be removed selectively by using ultraviolet light for a short time.
When MeNPOC is removed, the deprotected region on the surface can form chemical bonds with
DNA bases containing a MeNPOC photoprotecting group at its 50 hydroxyl position. In this
illustration, light is directed through a photolithographic mask to specific areas of the array surface,
which are activated for chemical coupling. The first chemical building block A containing a
photolabile protecting group X is then attached. Next, light is directed to a different region of the array
surface through a new mask. The second chemical building block T containing a photolabile
protecting group X is also added. This process is repeated until the desired product is obtained.
of the high absolute amount of full-length probes on the support (McGall and
Fidanza, 2001).
Another emerging light-directed synthesis approach for constructing high-
throughput oligonucleotide arrays is to use a digital light processor (DLP), i.e.,
micromirror (Nuwaysir et al., 2002). This maskless array synthesizer (MAS)
technology uses DLP to create “virtual” masks that direct an ultraviolet light
beam to discrete locations on a glass substrate for DNA synthesis. Similar to the
Affymetrix photolithographic approach, MAS is capable of constructing high-
density microarrays containing any desired nucleotide sequence. In contrast,
MAS does not require photomasks, which are very expensive and time-
consuming to manufacture. The MAS technology makes photolithography much
more flexible and user-friendly, although it is still in the early development stages
(Nuwyasir et al., 2002).
Photolithographic parallel synthesis offers a very efficient approach to high-
density array fabrication, in which the maximum achievable density is ultimately
dependent on the spatial resolution of the photolithographic process. Due to steric
and/or electrostatic repulsive effects, there is an optimum probe density for
maximum hybridization signal. Affymetrix chips currently contain ~250,000
oligonucleotides in an area of 1 cm2. One of the main advantages of this

approach is that microarrays of extremely high-density can be constructed
(Ramsay, 1998), but only oligonucleotides can be used in photolithography.
2. Contact Printing
The most commonly used microarray fabrication technology is mechanical

microspotting, which uses direct contact of computer-controlled multiple pins,
tweezers, or capillaries to deliver picoliter volumes of pre-made biochemical
reagents (e.g., oligonucleotides, cDNA, genomic DNA, antibodies, or small
molecules) to a solid surface. Currently, more than 1,000 individual cDNA
molecules can be deposited in an area of 1 cm2 using this technology (Rose,
2000). The advantages of microspotting include ease of implementation, low
cost, and versatility, while a major disadvantage is that each sample to be
arrayed must be prepared, purified, and stored prior to microarray fabrication.
In addition, microspotting rarely produces the densities that can be achieved
with photolithography. The various pin technologies for microspotting are
described below.
Solid pins. Solid pins have either flat or concave tips. Because such tip can
accommodate only a relatively small volume of the sample, only one microarray
can be printed generally with a single sampling load. Consequently, the overall
printing process is slow, making this technology suitable only for constructing
low-density arrays. Additionally, loss of sample due to evaporation is a
significant problem due to the large surface-to-sample volume ratio of solid pins.
Under standard laboratory conditions, about half of a 250 pl volume is lost in 1 s
(Mace et al., 2000). To minimize evaporation loss, a highly humid environment is
absolutely necessary. However, high humidity may prevent the sample from
drying sufficiently on the slide, resulting in sample migration or spreading.
Split pins. Split pins have a fine slot at the end of the pin for sample holding.
When the slit pin is dipped into the sample solution, the sample is loaded into the
slot, which generally holds 0.2 –1.0 ml of sample solution. A small volume of
sample (0.5 – 2.5 nl) is deposited on the microarray by tapping the pins onto the
slide surface with sufficient force (Rose, 2000) or touching the pins lightly on the
surface like an ink stamp (Martinsky and Haje, 2000). The company, TeleChem
International, manufactures a split pin by using digital control, so that there are
virtually no variations in mechanical quality from pin to pin (Martinsky and Haje,
2000). Thus, TeleChem pins provide very high printing consistency under
conditions of good sample preparation, proper motion control and homogeneous
printing substrate. Because the split pins hold a larger sample volume than
the solid pins, more than one microarray can be printed from a single sampling
event. Although split pin technology has been used successfully to print
microarrays, one of the drawbacks is that dust, particulates, evaporated
buffer crystals and/or other contaminants can clog the pin slot. Tapping the
pins on the substrate surface, however, is not recommended for various reasons
(Martinsky and Haje, 2000). Physical tapping leads to bulk transfer of the
sample from the pins and hence causes non-uniformly large spots and spot
merging. Also, tapping on the slide surface may lead to deformation of the pin
tip, causing larger spots, irregular spot shapes, a larger amount of sample
deposition, and poor printing quality (Mace et al., 2000). In addition,
tapping may fracture the surface coating and cause irregular spot shapes such
as doughnut shapes, in which the center of the spot lacks probe material
(Martinsky and Haje, 2000).
Pin and Ring. This is a variation of the pin-based printing process. The sample is
taken by dipping the ring into the sample well and then a small volume of sample
solution is deposited onto the slide surface by pushing the sample captured in the
ring using solid pins (Mace et al., 2000). Different sized rings can be selected to
hold 0.5– 3.0 ml of sample. Many different spot sizes can be obtained by
simply using pins of different diameter. In addition, loss of sample due to
evaporation is alleviated by minimizing fluid exposure through specific ring
configuration.
The pin-and-ring arraying technology offers a number of advantages. Since
the pin is used only for spotting and the sample fluid captured by the ring is
relatively large, the deposition of samples on different slides is accomplished in
an identical manner for each printing cycle, yielding a microarray fabrication
quality that is consistent and reproducible (Mace et al., 2000). In addition, the
ring geometry is capable of handling a wide variety of volumes and fluid
viscosities. Unlike the split pin, the pin and ring configuration is not susceptible
to clogging by the accumulation of dust, particulate matter, high-viscosity fluids,
debris, buffer or salts, and other materials. Finally, the pin and ring can deposit
samples on soft substrates such as agar, gels and membranes. However, one
drawback of this technology is sample loss. The majority of samples captured by
the ring cannot be used for spotting and is lost through washing for the next
sampling cycle.
3. Non-contact Ink-jet Printing
Ink-jet ejection technologies provide another means of fabricating micro-

arrays. In this approach, the sample is taken from the source plate, and a droplet
of sample is ejected from the print head onto the surface of the substrate. Similar
to microspotting, ink-jet ejection allows the spotting of virtually any biological

molecule of interest, including cDNA, genomic DNA, antibodies, and small
molecules. In contrast to microspotting, ink-jets have the advantage of avoiding
direct surface contact but cannot be used to manufacture microarrays as dense as
those prepared by photolithography or microspotting approaches.
Currently two types of non-contact ink-jet print technologies, piezoelectric
pumps and syringe-solenoid, are used for printing microarrays.
Piezoelectric Pumps. This printing technology utilizes a piezoelectric crystal,

which contacts a glass capillary containing the sample liquid (Englert, 2000;
Mace et al., 2000; Rose, 2000) and is still in the early stages of development.
When the crystal is biased with a voltage and subsequently deformed, the
capillary is squeezed and a small volume (0.05 –10 nl) of fluid is ejected
through the tip from the reservoir. Piezoelectric printing has the advantages of
an extremely fast dispensing rate (on the order of several thousand drops per
second), very small print volumes, and consistency of droplet size. The main
problem with this technology is clogging by air bubbles and particulates, which
makes the system less reliable compared to other printing methods. In addition,
the void volume of sample solution contained in the capillary is very large
(100 – 500 ml) and not recoverable. It is also difficult to change samples using
piezoelectric printing.
Syringe-solenoid Printing Technology. This technology uses a syringe pump and

a microsolenoid valve for dispensing samples (Rose, 2000). The sample is taken
by a syringe, and sample droplets, ranging from 4 to 100 nl in volume, are ejected
by pressure onto the surface through the microsolenoid valve. The main
advantages of this technology are reliability and low cost. However, it is not
suitable for fabricating high-density microarrays because of the large printing
volume and spot size.
D. CRITICAL ISSUES FOR MICROARRAY FABRICATION
This section highlights some practical issues that are important for microarray
fabrication: microarray density, reproducibility, storage time, contamination and
printing quality.
1. Microarray Density
Microarray density is one of the most important parameters for microarray

fabrication. The number of DNA elements that can be deposited on a slide will
depend on spot size and pin configuration.
Spot Size. Microarray density is determined by the size of the DNA spot and
depends directly on the volume of sample deposited on the substrate surface.
The volume of sample deposited per position on the array generally ranges from
50 to 500 pl, with high and low extremes of 10 pl and 10 nl possible,
respectively (Mace et al., 2000). Several factors affect the volume of sample
that can be applied to an array surface, including the surface properties (e.g.,
surface energy) of the slides and pins, and the sample solution characteristics
(e.g., viscosity). For printing high-density microarrays, a hydrophobic glass
surface (e.g., aldehyde-modified slide) is preferred, because spotted hydrophilic
samples will spread less on a hydrophobic surface than on a hydrophilic one.
Because the pin contact surface area determines the initial contact between the
sample and slide, the spot size increases as the pin contact surface area
increases. In addition, pin velocity has a great effect on the spot size. The
loading sample volume for a split pin (e.g., ChipMaker and Stealth pins from
TeleChem) typically ranges from 0.2 to 0.6 ml. Thus, if the pins tap the surface
at high speed (. 20 mm/s), a large sample volume may be forced out of the pin
and large spots will be produced (Rose, 2000).
Pin Configuration. Printing pins are mounted in a print head, which can hold up
to 64 pins. The distance between pins on the print head is 4.5 mm and precisely
matches the well spacing of a 384-well microtitre plate. DNA samples are first
taken from 96-well or 384-well source plates by dipping the pins into the sample
wells with either a single pin or multiple pins, and then depositing the sample on
the slide surface by gently touching the pins to the surface. Fabrication of arrays
using a single pin is the most straightforward approach, but it is also time-
consuming. The main advantage of single pin printing is that the DNA samples in
the source plates can be positioned on the array in the same order as they occur in
the source plate, thus making sample tracking and post-hybridization analysis
easier. Another advantage is that pin-to-pin variations are not a problem when
using a single pin for microarray printing. Using a single pin and 250-mm spot-to-
spot spacing, for example, more than 20,000 spots can be deposited on
22 £ 72 mm2 printing area.
Multiple pins are generally used for printing high-density microarrays because
of the increased printing speed, even though different pins can cause variations in
array quality. To print with multiple pins, the pins are dipped into sample wells of
a 384-well plate and then touched to the slide surface simultaneously to create
separate spots at a 4.5-mm spacing in the first round. Later rounds of printing are
achieved by spotting with a predefined spot-to-spot offset distance from the
previous location. Each pin deposits samples in a sub-grid. Since some areas
within each sub-grid might not be completely filled with spots due to the
restriction of the layout, the density of microarrays will generally decrease as the
number of pins used increases. Printing microarrays with multiple pins requires
more time in designing the array layout, as well as sophisticated sample tracking
and deconvolution in the data analysis phase.
2. Reproducibility
Fabricating microarrays of reproducible quality is of the utmost importance in

microarray-based experimentation. For reliable and reproducible data, the
uniformity of individual spots across the entire array is very important for
simplifying image analysis and enhancing the accuracy of signal detection.
Several factors can affect the uniformity of spots, including array substrate,
pins, printing buffer, and environmental controls. As mentioned previously,
non-homogeneous surface treatment will cause variations in the amount of
attached DNA.
Variations in array quality can be caused by differences in pin geometry, pin
age and sample solutions. Movement of the pin across the surface in the XY
direction may cause the tip to bend (Rose, 2000). Tapping the pins on the surface
may result in deformation of the pin tips. In addition, dragging the pin tip across
the surface may cause clogging of the pin sample channel. Therefore, great care is
needed in handling pins, even though they are robust. Pins should be cleaned with
an ultrasonic bath after each printing (Rose, 2000).
Environmental conditions have significant effects on spot uniformity and size
(Hegde et al., 2000). Humidity control is absolutely necessary for preventing
sample evaporation from source plates and the pin channel during the printing
process. Sample evaporation can cause changes in DNA concentration and
viscosity. Reducing the extent of evaporation can help the small spotted volume
of DNA have more time to bind at equal rates across the entire spot. As a result,
DNA spots of high homogeneity will be obtained (Diehl et al., 2001). Generally,
the relative humidity is controlled between 65 and 75% (Rose, 2000).
Condensation could occur if the relative humidity is greater than 75%.
Producing homogeneous spots on arrays also depends on the printing or
deposition buffer. Saline sodium citrate is commonly used as a printing buffer in
microarray construction; however, spot homogeneity and binding efficiency with
this buffer can be poor. The addition of 1.5 M betaine to the printing buffer
can significantly improve spot homogeneity and binding efficiency (Diehl et al.,
2001), because betaine increases the viscosity of a solution and reduces the rate of
evaporation. More uniform spots can also be achieved with a deposition buffer that
contains 50% dimethyl sulfoxide (DMSO) (Hegde et al., 2000; Wu et al., 2001).
3. Storage Time
Another important practical issue concerns the shelf life of unused

microarrays. The maximum time that microarrays can be stored prior to use is
currently unknown. The shelf time could depend on the coating chemistry
of the slide and the storage conditions. Unprocessed microarrays can be
stored in a dessicator for many months without deterioration of performance
(Worley et al., 2000).
4. Contamination
To produce high-quality microarrays, the collection of airborne dust and

impurities on the slide surface must be eliminated or at least minimized during
array fabrication. Dust and particulate matter can settle on the slide surface
and cause printing inaccuracies as well as poor quality scanned image displays.
Enclosing the array device in a humidity chamber can minimize dust
contamination.
Because pins are generally reused for depositing different biological samples,
sample carryover during the printing process is a practical concern and can
complicate interpretation of hybridization results. Efficient cleaning of the pins is
therefore required for the printing process. Generally, the pins are cleaned by
dipping them into distilled water or detergent and then using a vacuum to remove
the wash solution from the pin channel. Repeating this process three times is
generally sufficient to eliminate sample carryover problems. Cross-contami-
nation during sample preparation and handling is another important concern in
the microarray printing process. For making microarrays, plasmids containing the
desired cDNA clones are generally extracted from bacterial cultures and the
desired genes are amplified from the plasmid DNA. Recent studies showed that
up to 30% of clones contained the wrong cDNA (Knight, 2001). This is most
likely due to bacterial contamination and handling errors during sample
preparation. Therefore, great care must be taken to eliminate or minimize such
errors. Errors in public sequence databases are also possible and can lead to
failures in microarray-based detection. For instance, some mouse sequences in
the public databases correspond to the wrong strand of the DNA double helix. As
a result, the designed oligonucleotide probes were not able to detect their target
mRNAs (Knight, 2001).
5. Evaluation of Printing Quality
After printing, it is important to assess the quality of the arrayed slides prior to
hybridization in terms of surface quality, integrity and homogeneity of each DNA
spot, the amount of deposited DNA, and consistency among replicated spots.
Staining prior to hybridization will identify any problems introduced during the
fabrication process. Microarrays can be stained with various fluorescent dyes,
such as PicoGreen, SYBR Green II and Topo Green (Battaglia et al., 2000),
followed by fluorescent scanning. Such information is critical in assessing the
quality of the printed arrays and the data output from those arrays.
IV. MICROARRAY HYBRIDIZATION AND DETECTION
A. PROBE DESIGN AND SYNTHESIS
Probe design and synthesis are critical steps in generating high-quality

microarrays for gene expression analysis. Three types of probes are used for
microarray fabrication: PCR products, cDNA clones, and oligonucleotides. For
the construction of cDNA microarrays, individual open reading frames (ORFs)
can be amplified using gene-specific primers. Because cross-hybridization among
homologous genes is a potential problem, full-length genes cannot be used for
microarray construction. Several computer programs are available that can
identify DNA fragments specific (, 75% sequence identity) to each ORF by
comparing the target gene with all other genes in a genome (Xu et al., 2002).
Once the specific fragments are identified, more than one set of primers can be
obtained based on the identified unique fragments using the PCR primer design
program Primer 3 (Whitehead Institute). The designed primers are generally
synthesized commercially.
Optimal forward and reverse primers are generally selected based on
the following considerations. First, for genes shorter than 1000 bp, the PCR-
amplified unique fragments should be as long as possible. For genes longer than
1000 bp, the optimal amplified fragments should be within 500 –1200 bp.
Second, each oligonucleotide primer should be 20 –28 bp in length and the set
of primer pairs (typically stored in 96-well plates) should have an annealing
temperature of approximately 65 8C to simplify PCR amplification. If the
desired target annealing temperature cannot be obtained, a lower annealing
temperature can be used. In the case where specific fragments cannot be
identified for some homologous genes, fragments with higher than 75% sequ-
ence identity will be selected and appropriate primers can be designed based on
these fragments. However, hybridization signals for these genes should be
carefully interpreted during microarray data analysis. One of the great practical
problems with probe amplification is that PCR product yields vary considerably
among different genes and some primers may fail to yield PCR products.
This may cause significant variation in the DNA concentration present on the
slide surface.
The cDNA clone-based probes are generally derived from whole genes or
fragments of genes that are amplified from clone libraries using vector-specific
primers. The size of clone probes generally ranges from a few hundred to a few
thousand base-pairs. Generally, since vector-specific primers are used for

amplifying the cloned inserts, clone-based probes cannot be specifically designed
for regions of low homology to other genes. As mentioned above, a substantial
portion of clones may contain the wrong cDNA due to bacterial contamination
and mishandling (Knight, 2001).
Oligonucleotide probes are different from other probes in that they can be
deposited by printing or synthesized in situ on a solid surface. Specific
oligonucleotide probes can be designed based on gene sequences. Generally, the
sizes of the oligonucleotide probes are shorter than 25 bp, and several different
oligonucleotide probes are used per gene for high-density oligonucleotide
microarrays. To discriminate mispriming, a probe is designed deliberately to
have a single mismatch at the central position (Lockhart et al., 1996;
Warrington et al., 2000). Recently, the utility and performance of oligonucleo-
tide microarrays containing 50- to 70-mer oligonucleotide probes were
evaluated (Kane et al., 2000). The results indicate that such oligonucleotide
microarrays can be used as a specific and sensitive tool for monitoring
gene expression.
B. TARGET LABELING AND QUALITY
Target labeling is another critical step in successful microarray-based

experimentation. The methods available for labeling nucleic acids for micro-
array hybridization can be classified into two categories: direct and indirect
labeling.
1. Direct Labeling
In direct labeling, fluorescent tags are directly incorporated into the nucleic
acid target mixture before hybridization by enzymatic synthesis in the
presence of either labeled nucleotides (e.g., Cy3- or Cy5-dCTP) or PCR primers
(Fig. 4A). The most commonly used method is to label the target mRNA or
total cellular RNA using reverse transcriptase. In a first-strand reverse
transcription reaction, fluorescently labeled cDNA copies of RNA are
synthesized by incorporation of a fluorescein-labeled nucleotide analog.
Random hexamers, oligo(dT) or gene-specific primers can serve as primers for
the initiation of reverse transcription. Since prokaryotic mRNA has no poly-(A)
tail, random hexamers are generally used for reverse transcription. In this case,
total cellular RNA is used as the template for cDNA synthesis, and hence a
greater degree of background fluorescence intensity can occur. Although gene-
specific primers can reduce such background levels by copying gene-specific
Figure 4 Labeling strategies. (A) Direct incorporation of fluorescent dyes into target sample
through reverse transcription. (B) Incorporation of fluorescent dyes into target samples through
reverse transcription in the presence of amino-allyl-dUTP, followed by chemical coupling with
fluorescent dyes. (C) Dendrimer-based indirect labeling. (Courtesy of Molecular Probes).
fragments, it requires reverse transcription with hundreds or thousands of

primers.
A variation of the direct labeling approach is that mRNA is amplified up to
1,000 –10,000-fold by T7 polymerase to obtain antisense mRNA (aRNA), and
then the aRNA is reverse transcribed to obtain labeled cDNA (Salunga et al.,
1999). One of the advantages of the T7 polymerase-based amplification method
over other amplification methods is that all mRNAs are almost equally amplified,
because amplification with T7 polymerase is a linear process. Another advantage
is that mRNA can be labeled easily with reverse transcriptase, which incorporates
fluorescent tags much more readily than DNA polymerase.
Figure 4 (continued )
One of the problems with the reverse transcriptase-based labeling approach

is that nucleotides tagged with structurally different fluorescent dyes are
differentially and non-uniformly incorporated into cDNA. To resolve this
problem, a two-step approach was proposed. The FairPlaye system developed
by Stratagene Corporation uses a two-step chemical coupling method to
fluorescently label cDNA. First, an amino allyl-dNTP is uniformly and efficiently
incorporated into cDNA by reverse transcriptase (Fig. 4B), because the amino
allyl-dNTP does not exhibit steric hindrance. Then, an amine-reactive cyanine is
chemically coupled to the amino-modified cDNA. The main advantage of this
approach is that this system efficiently produces uniformly labeled cDNA without
any dye bias. As a result, this system is highly sensitive (5-fold increase in
sensitivity), requires less RNA, and allows detection of low abundance genes.
Any labeling bias resulting from fluorescent dye incorporation also appears to be
negligible and thus the dual labeling experimental approach is not needed.
2. Indirect Labeling
In the indirect labeling approach, fluorescence is introduced into the

detection procedure following hybridization. Briefly, epitopes are incorporated
into the target samples during cDNA synthesis. After hybridization with the
epitope-tagged target samples, the microarray is incubated with a fluorescently
tagged protein that binds to the epitopes. The most common indirect labeling
method uses a biotin epitope and a fluorescent streptavidin – phycoerythrin
conjugate (Warrington et al., 2000). The biotinylated nucleotides are
incorporated into cDNA by reverse transcription and hybridized with the
microarrays. After hybridization, the array is stained with a streptavidin –
phycoerythrin conjugate, which binds to biotin tags and emits fluorescent light
when excited with a laser.
Another indirect labeling approach is known as Tyramide Signal Amplifica-
tion (TSA) (Adler et al., 2000). This approach uses biotin and dinitrophenol
(DNP) epitopes as well as streptavidin and antibody conjugates linked to
horseradish peroxidase (HRP). In the presence of hydrogen peroxide, HRP
catalyzes the deposition of Cy3- or Cy5-tyramide compounds on the microarray
surface. By this method, a DNP- or biotin-dCTP analog is first incorporated into
cDNA, and then the epitope-tagged cDNA is hybridized with the microarray.
Following hybridization, the microarray is incubated with anti-DNP-HRP, and
Cy3-tyramide is deposited on the microarray surface, followed by incubation
with streptavidin – HRP and deposition of Cy5-tyramide (Adler et al., 2000). The
main advantage of this approach is that it can provide10- to 100-fold signal
amplification over the direct labeling approach. Thus, this approach can be used
effectively to monitor the expression or abundance level of rare transcripts or to
analyze samples prepared from small numbers of cells. The main disadvantage of
this method is that it is generally less precise for comparative analysis due to
variations arising from differences in labeling efficiencies and protein-binding
affinities (Schena and Davis, 2000). In addition, the signal intensity is only semi-
quantitative because of the involvement of enzymatic signal amplification (Alder
et al., 2000).
The third indirect approach is to use DNA dendrimer technology (Stears et al.,
2000) (Fig. 4C). Dendrimers are stable, spherical complexes of partially double-
stranded oligonucleotides with a determined number of free ends, which are
tagged with fluorescent dyes, Cy3 or Cy5. In this technology, the cDNA is first
synthesized by reverse transcriptase with primers containing specific capture
sequences that can bind the Cy3- or Cy5-tagged dendrimers through sequence
complementarity. The synthesized cDNAs are then hybridized to microarrays,
and the bound cDNAs on the microarrays are detected by incubating the arrays
with the fluorescent dye-tagged dendrimers. The dendrimer detection approach is
highly sensitive, requiring up to 16-fold less RNA for probe synthesis. Since the
fluorescent dye is attached to the free end of the dendrimers, signal intensity is
independent of probe size and composition. In addition, this detection system has
a high signal-to-background ratio and can be used for multiple channel detection
on a single microarray.
C. HYBRIDIZATION
After microarray fabrication, the most important issue in microarray-based

analysis is probe –target hybridization. Conceptually, microarray hybridization
and detection are quite similar to the traditional membrane-based hybridization
(Eisen and Brown, 1999). Before hybridization, the free functional groups
(e.g., amine) on the slide should be blocked or inactivated to eliminate non-
specific binding, which causes high background and depletion of probes. Any
unbound DNA on the slides can be washed away during the pre-hybridization
process. Removal of unbound DNA in pre-hybridization is important, because
any DNA that washes from the surface during hybridization competes with
DNA bound to the slide. Since the rate of hybridization in solution is much
faster than that on surfaces, the presence of unbound probe DNA can lead to a
dramatic decrease in the measured signals obtained from microarrays.
After pre-hybridization, the microarray is hybridized with fluorescently
labeled target DNA or RNA for a certain period of time. Post-hybridization
washing then removes unbound labeled material. Regardless of the hybridization
format, the hybridization solution should be mixed well so that the labeled targets
are evenly distributed across the array surface to obtain the maximum number of
optimal target – probe interactions. In addition, the wash solutions should be
uniformly distributed to eliminate unbound probes, remove non-specific
hybridization, and minimize background signal.
D. DETECTION
Both the confocal scanning microscope and coupled charge device (CCD)
camera have been successfully used for the detection of microarray hybridization
signals, and many such devices are commercially available (Hegde et al., 2000).
Although the confocal scanning microscope and CCD camera systems both have
advantages and disadvantages (as described below), the former is more
commonly used.
Generally, a confocal scanner uses laser excitation of a small region of the
glass slide (~100 mm2), and the entire array image is acquired by moving the
glass slide, the confocal lens, or both across the slide in two directions
(Schermer, 1999). The fluorescence emitted from the hybridized target molecule
is gathered with an objective lens and converted to an electrical signal with a
photomultiplier (PMT) or an equivalent detector. The main drawbacks of using
a confocal scanning microscope for signal detection is that each excitation

wavelength must have its own laser, which can be expensive, and the device is
very sensitive to any non-uniformity of the glass slide surface.
The CCD camera exploits many of the same principles as a confocal scanner,
but the CCD camera utilizes substantially different excitation and detection
technologies (Schermer, 1999). CCD systems typically use broad-band xenon
bulb technology and spectral filtration (Basarsky et al., 2000). The key advantage
of the CCD camera-based imaging systems is that they allow simultaneous
acquisition of relatively large images of a slide (1 cm2) and hence do not require
moving stages and optics, which reduces cost and simplifies instrument design.
However, since the CCD camera does not move the optics or stages, several
images need to be captured from different fields of the microarray and then
stitched together to represent the entire information on the slide. Because most
commonly used fluoresceins have a small difference between excitation and
emission maxima, it is difficult to effectively separate excitation and emission
light in the spectral filtration process.
E. CRITICAL ISSUES IN HYBRIDIZATION AND DETECTION
This section highlights some important practical issues related to microarray

hybridization and detection, namely, probe retention and quantitative hybridiz-
ation, target labeling and availability, spatial resolution and cross-talk, and
photobleaching and scanning parameters.
1. Probe DNA Retention and Quantitative Hybridization
In solution-based hybridization, signal intensity depends on both target and

probe DNA concentrations. In gene expression profiling studies, it is assumed
that the concentrations of all probe DNAs deposited on the microarrays are much
higher than the mRNA concentrations in the fluorescently labeled target samples,
so that signal intensity is dependent exclusively on the mRNA concentration in
the target samples. Therefore, many factors causing probe deposition variations
will have negligible effects on hybridization signal intensity.
For the accurate quantitation of gene expression, it is essential to ensure that
the arrayed DNA probes are in excess relative to the labeled target cDNAs.
Generally, a DNA concentration of 100 – 200 ng/ml is used for spotting, which
corresponds to 100 –200 pg/spot for a 1-nl deposition. The retention is about
20– 30% on silanized glass surfaces (Worley et al., 2000). Thus, after boiling
and hybridization, this corresponds to approximately 20– 60 pg of double-
stranded DNA present in each spot for binding. Studies indicate that the arrayed
DNA appears to be in excess for all the protein-coding genes in Escherichia coli
(Worley et al., 2000). However, probe DNA retention depends on slide surface,
coating chemistry, post-processing, hybridization, and washing conditions.
Therefore, to ensure accurate quantitative results for highly expressed genes, it
is important to understand how much spotted DNA can actually be retained
after hybridization.
Whether the probe DNA concentration represented on the array substrate
is in excess obviously depends on the amount of target sample used. Typically,
10 – 20 mg of total cellular RNA is used for monitoring gene expression in
prokaryotes. However, for monitoring rare transcripts, higher RNA concen-
trations (e.g., 50 mg) are generally used. In this case, probes corresponding to
abundant transcripts may not be in excess relative to the target samples, resulting
in hybridization that is not quantitative. Hence, it is important to select the
appropriate amount of RNA to ensure that the microarray signal is within the
range of linear response for the system being used.
2. Target Labeling and Availability
The integrity and purity of the RNA are crucial for obtaining high-quality
microarray hybridization results. Impurities in RNA preparations could have an
adverse effect on both labeling efficiency and the stability of the fluorescent dyes.
Thus, the RNA must be free of contaminants such as polysaccharides, proteins
and DNA. Many commercial RNA purification kits are available for producing
RNA of sufficient purity for microarray studies. In addition, unincorporated
nucleotides present in the labeling reaction must be removed to reduce
background noise. Finally, both Cy3 and Cy5 are sensitive to light, and thus
great caution must be taken to minimize exposure to light during labeling,
hybridization, washing and scanning.
The most frequently encountered experimental problem is the variation in
hybridization signal between labeling reactions. In many cases, poor hybridiz-
ation signals result from poor dye incorporation. Very low dye incorporation
(, 1 dye molecule/100 nucleotides) gives unacceptably low hybridization signal
intensities. However, studies showed that very high dye incorporation (e.g., . 1
dye molecules/20 nucleotides) is also not desirable, because high dye
incorporation significantly destabilizes the hybridization duplex (Worley et al.,
2000). Thus, it is important to measure the dye incorporation efficiency prior to
hybridization. The specific activity of dye incorporation can be determined by
measuring the absorbance at wavelengths of 260 and 550 nm for Cy3 or 650 nm
for Cy5. A suitable labeling reaction should have 8– 15 A260/A550 ratio for Cy3
and 10 –20 A260/A650 for Cy5.
Another problem encountered routinely is the quality of fluorescent dyes. The

labeling efficiency and hybridization vary significantly sometimes from batch to
batch, especially for Cy5. Fresh reagents are very important in achieving a high
degree of detection sensitivity (Wu et al., 2001).
Microarray hybridization is generally performed in the absence of mixing.
Since the diffusion coefficient is very small for large labeled target DNA
molecules, the probe at each arrayed spot is in effect hybridizing with its labeled
counterpart from its immediate or nearly immediate local environment (Worley
et al., 2000). Thus, the target solution should be mixed well and uniformly
distributed over the microarray surface area. Otherwise, the availability of the
labeled target molecules to the arrayed spots could be significantly different
across the microarray surface. As a result, significant differences in signal
intensity can be observed.
3. Spatial Resolution and Cross-talk
The spatial resolution of microarray detection systems is usually expressed as

a pixel size, the physical “bin” in which a single datum is acquired. The spatial
resolution for many commercial systems usually ranges from 5 to 20 mm. The
selection of spatial resolution depends on spot size, and in general, the pixel
dimension should be less than 1/10 of the diameter of the smallest spot on the
array. For example, microarrays containing 100-mm spots require fluorescent
detectors with a spatial resolution of 10 mm pixel size.
Cross-talk refers to the phenomenon in which an emission signal from one
channel is detected in another channel, resulting in an elevated, erroneous
fluorescence reading. Cross-talk is most likely from the shorter wavelength
channel into the longer wavelength channel. For example, the fluorescence
intensity from the Cy3 channel can contaminate the Cy5 channel but not vice
versa. Cross-talk is the most common potential problem in the simultaneous
scanning approach, which acquires both images from two channels at the same
time (Basarsky et al., 2000). For gene expression experiments, cross-talk should
be kept to less than 0.1%. The most common and cost-effective way to minimize
cross-talk is to use emission filters that reject light outside the desired
wavelengths. Cross-talk can also be greatly minimized by selecting fluorescent
dyes and lasers with sufficient differences in wavelength (Schermer, 1999).
4. Photobleaching and Scanning Parameters
Light is emitted from a fluorescent dye when it is illuminated by a light source.

Generally, the emitted fluorescence is directly proportional to the power of
the excitation light. Therefore, to increase detection sensitivity, higher power of

excitation light is preferred. If the excitation light is excessive, however, the
incoming photons can damage the dyes and reduce the fluorescent signals during
successive scans, leading to photobleaching of the signal intensity. Photobleach-
ing depends on the duration of sample illumination. More powerful light sources
and/or a longer laser exposure time can result in significant photobleaching.
When acquiring an array image, it is best to keep photobleaching to less than 1%
per scan.
Different dyes have considerable differences in their photostabilities. For
example, Cy5 is more sensitive to photobleaching than Cy3. The differences
in photostability among different dyes could be a significant problem when
multiple dyes are used in an experiment, because the ratios measured can
lead to significant quantitative errors. To minimize photobleaching, the Cy5
channel is always scanned first, followed by the Cy3 channel.
Although Cy3 (0.15, no unit) has a lower quantum yield than Cy5 (0.28), Cy3
is more efficiently incorporated into cDNA during reverse transcription. Such dye
characteristics can cause variations in the signal intensity obtained in reverse
labeling experiments. The signal should be balanced during scanning by using a
higher PMT setting for the dye with the weaker signal to allow detection of more
spots of low signal intensity.
V. MICROARRAY IMAGE PROCESSING
A. DATA ACQUISITION
The fundamental aim of image processing is to measure the signal intensity of

arrayed spots and then quantify gene expression levels based on the signal
intensities acquired for each spot. Therefore, spots on the array image must be
correctly identified.
The spots on microarrays are arranged in grids. An ideal microarray image for
easy spot detection should have the following properties: (i) the location of spots
should be centered on the intersections between the row and column lines; (ii) the
spot size and shape should be circular and homogeneous; (iii) the location of the
grids on the images should be fixed; (iv) the slides should have no dust or other
contaminants; and (v) the background intensity should be very low and uniform
across the entire image. In practice however, it is difficult to obtain such ideal
images. First, the spot position variation occurs because of mechanical
limitations in the arraying process, including inaccuracies in robotic systems,
the printing apparatus and the platform for holding slides. Second, the shape and
size of the spots may fluctuate considerably across the array because of variations
in the size of the droplets of DNA solution, DNA and salt concentration in
the printing solution, and slide surface properties. In addition, contamination

from airborne dust and impurities on the slide surface is a major problem for
processing array images. To obtain accurate measurements of hybridization
signals, all of these potential problems should be taken into consideration.
Many methods are available for resolving spot location errors, spot size and
shape irregularities and contamination problems (Zhou et al., 2000) in order to
accurately estimate spot intensities. Commercial and free software, including
ImaGenee from BioDiscovery (Los Angeles, CA), QuantArraye from GSI
Lumonics, and the software on Axon GenePixe systems (Bassett Jr. et al., 1999),
can be used for microarray image processing. Typically, a user-defined gridding
pattern is overlaid on the image, and the areas defined by patterns of circles are
used for spot intensity quantification.
The data are extracted and generally expressed as the total (the sum of the
intensity values of all pixels in the signal region), mean (the average intensity of
pixels), and median (the signal intensity of the median pixel). Microarray output
corresponding to the total intensity is not the best measurement of hybridization
signal, because it is particularly sensitive to variations in the amount of DNA
deposited on the surface and the presence of contamination (Zhou et al., 2000).
The mean is probably the best measurement when using an advanced image
processing system that permits accurate segmentation of contaminated pixels,
because the mean measurement reduces variations caused by differences in the
amount of DNA deposited within a spot. However, the mean measurement is
vulnerable to outliers (Petrov et al., 2002). The median is a better choice than the
mean if the image processing software is not good enough for correctly
identifying signal, background and contaminated pixels. The median intensity
value is very stable and is close to the mean if the distribution profile of pixels is
uni-modal. The median is equal to the mean when the distribution is symmetric.
An alternative to the median measurement is to use a trimmed mean (the mean of
the pixel intensity after a certain percentage of the pixels are removed from both
tails of the distribution).
Some comparative studies indicate that the choice of measurements depends
on the segmentation techniques used. The mean is the best measurement if the
combined and trimmed segmentation techniques are used, whereas the median is
the best without trimming (Petrov et al., 2002).
B. ASSESSMENT OF SPOT QUALITY AND BACKGROUND

SUBTRACTION
For some spots, signal intensity data may not be reliable because of the
inherently high variation associated with array fabrication, hybridization, and
image processing. Thus, the first step in data processing is to assess the quality of
spots, with the removal or filtering of unreliable poor spots or outlying spots
(outliers) prior to data analysis (Heyer et al., 1999; Tseng et al., 2001). It is
critical to identify problematical slides, because without assessing the quality of
the spot signals, conclusions drawn from the analysis of such data could be
misleading.
1. Identification of Poor Slides
Due to the multiple steps involved in microarray experiments, it is important

to evaluate array slides as a whole and to eliminate unreliable hybridization
signals prior to rigorous data analysis. Two measures can be used to assess the
overall slide quality if replicate spots are present on the arrays (Worley et al.,
2000): one can calculate (i) the average coefficient of variation (CV) of replicates
in the spot pairs and (ii) the r 2 value of the regression line from a scatter plot of
duplicate spots. Although there is no general consensus on the appropriate
threshold value for rejecting slides, slides are generally accepted if the average
CV is less than 50%. If there are no replicate spots on the microarrays, slide
quality can be assessed by determining the number of spots that are of poor
quality. Generally, microarray experiments should be repeated if more than 30%
of the spots on the microarray are flagged as poor spots.
2. Identification of Poor Quality Spots
There are no rigorously defined rules for identifying poor spots from a
biological or statistical perspective. The spot quality and integrity are generally
assessed based on the following criteria:
Spot size and shape. Spots with excessively large or small diameters compared to
the majority of spots should be discarded. Discarding such low-quality spots
significantly improves the reliability of the data (Zhou et al., 2000).
Spot homogeneity. The distribution of pixels within the spots can be used to
assess spot homogeneity. Generally, spots with less than a certain percentage
(e.g., 55 – 60%) of all pixels having intensities greater than the average
background intensities (Khodursky et al., 2000) or one standard deviation (SD)
above local background are flagged as poor quality spots (Murray et al., 2001).
Spot intensity. Spots with signals not significantly above background should be
identified using various standards. For example, spots with median or mean
signals less than one to three SDs above background in both channels (Chen et al.,
1997; Basarsky et al., 2000; Hegde et al., 2000) are flagged as poor quality spots.
In addition, spots whose signal is not at least 2.5 times higher than the
background signal in both channels are excluded (Evertsz et al., 2000).
Another way to define poor spots is based on signal-to-noise ratio (SNR),
which is often defined as the ratio of the difference between signal and
background divided by SD of background intensity (Verdnik et al., 2002). This
ratio indicates how well one can resolve a true signal from the instrumental
noises. A commonly used criterion for the minimum signal that can be accurately
determined is an SNR value equal to 3. Below that value, the signal cannot be
accurately quantified, and such spots are treated as poor spots.
The commercially available software, ImaGene from Biodiscovery, is able to
automatically flag poor spots. Spots identified as poor quality are not included in
the data analysis. Although the criteria for defining poor spots are based on
subjective thresholds rather than statistically robust tests, they take into account
the major factors affecting the quality of data and are likely to be very effective in
reducing the amount of noise.
3. Removal of Outlying Spots
Outliers represent extreme values in a distribution of replicates. Outlying spots

can be caused by uncorrected image artifacts such as dust or by factors
undetectable by image analysis such as cross-hybridization. Outliers significantly
affect the estimation of expression values and its associated random errors. Thus,
removal of outlying spots is an important step in data filtering. However,
distinguishing outliers from differentially expressed genes is very challenging,
because there is no general definition describing outliers. In this section, we
briefly describe several commonly used methods for identifying outliers.
Simple threshold cutoff. A gene whose CV is greater than a certain threshold

(Murray et al., 2001), e.g., 30 –50%, is excluded from the data analyses.
Intensity-dependent threshold cutoff by windowing procedure. (Tseng et al.,

2001). The CV values for individual genes are plotted against the average signal
intensity of the two channels [(Cy3 þ Cy5)/2]. For each gene, a windowing
subset is constructed by selecting a certain number of genes (e.g., 50) whose
mean intensities are closest to this gene. If the CV of this gene is within a top
certain percentage (e.g., 10%) among genes in its windowing subset, then data on
this gene are regarded as unreliable, and hence all replicate data for this gene are
discarded. However, in many cases, not all replicate spot data for this gene are
unreliable. To salvage some information for this gene, the most outlying spots
can be eliminated, and the CV of the intensity ratios of the remaining spots
corresponding to this gene can be recalculated. If the CV is significantly reduced
below the threshold level, the data for the remaining spots can be used in
subsequent analyses. The CV can also be used for assessing the quality of
different slides and different experiments (Tseng et al., 2001).
Removal of outliers with jackknife procedure. The jackknife correlation can

be used to remove outliers for expression data obtained from time-series
microarray experiments (Heyer et al., 1999). In this statistical approach, the
correlation coefficient is calculated for each pair of genes using all of the time-
series data points. Then the data at one time point are deleted, and the correlation
coefficient is recalculated respectively for each pair of genes with all of the time-
series data points but one. The jackknife correlation is the minimum correlation
coefficients obtained above and can then be used for further cluster analysis.
Jackknife correlation is robust and insensitive to single outliers. Applying
jackknife correlation reduces false positives, while capturing the shape of an
expression pattern. Hyer et al. (1999) showed that the genes showing similar
expression patterns generally had a jackknife correlation of 0.7 or higher.
Identification of outliers based on pooled error methods. Several methods are

used for statistical detection of outliers, but they are generally less adequate for
typical microarray studies due to the small number of replicates (Nadon et al.,
2001). The random error estimation for each gene based on a small number of
replicates is imprecise, which makes statistical tests insensitive. As a result, many
replicate spots may be falsely identified as outliers or many true outliers may not
be identified (Nadon and Shoemaker, 2002). The potential violation of the
normality assumption makes inferences of outliers and gene differential
expression less reliable (Nadon et al., 2001).
The pooled error method assumes that all probes or probes of similar
intensities within a specific study have the same true random error. Variance
estimates therefore can be pooled together across many genes and the precision of
error estimation can be greatly improved. Furthermore, it is assumed that the
standardized residuals have a normal distribution if the pooled error model is
correct. Under these assumptions, the existence of outliers will cause the
distribution of the entire data set to deviate from normal. Removal of spots with
large residuals will improve the normality of the entire data set. Generally,
outliers are identified in an iterative fashion: spots with large absolute residuals
are removed from the data set; data are examined for normality and the residuals
are calculated again. The process is iterated until the index asymptotes approach a
stable value, which indicates that further removal of data values would not
improve the normality of the distribution of the remaining data set. Software is
available (ArrayState) to facilitate array-based statistical analysis (Nadon et al.,
2001). In this software package, outliers are automatically detected. The pooled
error method is a better, more sensitive method for outlier detection and can be
used appropriately for microarray experiments having as few as two replicates.
4. Background Subtraction
Subtraction of background fluorescence from hybridization signals is the

second step in microarray data processing. Background subtraction is necessary
to distinguish actual signals based on hybridization from noise and allows the
comparison of specific spots. There are two approaches to background
subtraction. The first approach is to take signal intensity levels from blank
areas on the array and use this for subtraction. The problem with this approach is
that the background varies across the array and thus the background noise among
spots can be significantly different. The second approach is to use a local
background for the area around each spot for background determination. Local
sampling of the background is generally used to specify a threshold that the true
signal must exceed. By doing this, it is possible to detect weak signals and extract
an average density above the background for each array element (Chee et al.,
1996). After removing poor slides, poor quality spots, outliers and background,
the microarray data are ready for further normalization and data analysis.
VI. MICROARRAY DATA ANALYSIS
A. DATA NORMALIZATION
Microarray hybridization possesses intrinsic variation, which can potentially

occur at every step in the microarray process. One key question prior to applying
statistical analyses is whether such variations represent true random variations
in expression values or are due to systematic variations arising from differences
in the experimental conditions. Before pursuing further statistical analyses
of microarray data, the systematic variations must be removed by normali-
zation to allow statistical comparisons among different slides and different
experiments.
1. Sources of Systematic Variations
Systematic variation stems from a number of sources during microarray

experiments. The major anticipated sources of variations include the following
(Tseng et al., 2001).
Variations within a Slide or Spatial Effect. Many studies show that substantial
signal variation occurs for the same gene within a slide (Dudoit et al., 2001).
Differences in pin geometry, slide homogeneity, hybridization and target fixation
could all contribute to variations observed among repeated spots within a slide.
Some systematic differences may occur between different pins due to differences
in the length or in the opening of the tips, pin deformation following multiple
rounds of printing, and slight differences in surface properties. All of these can
lead to differences in target DNA transfer and hence may cause systematic
variations in microarray signal intensity. The amount of deposited target DNA
also fluctuates for the same pins, while studies showed that the variation among
different pins was significantly higher (Dudoit et al., 2001). In addition, the
fraction of target DNA that is chemically fixed onto the slides is unknown and
could vary considerably within slides. For various reasons, the labeled targets
may also be distributed unevenly over the slide and/or the hybridization reaction
may occur unequally in different parts of the slides. Finally, some areas of a
slide may be contaminated and have a high background. The influence of these
factors on signal intensity measurement within a slide is generally referred to as
spatial effect.
Variation among slides or slide effect. Differences in surface properties,

microarray fabrication, hybridization and imaging could lead to systematic
variations in hybridization signals among different slides. The amount of probe
DNA immobilized on the slide during array printing and probe fixation can be
substantially different among different slides due to various factors such as
differences in slide surface properties and sample evaporation during printing.
Also, the amount of cDNA added to the slides, especially when different RNA
preparations are used, and the local environment and hybridization conditions,
such as temperature, buffer pH, target concentration, incubation and washing
time in each hybridization chamber, could be considerably different. Background
noise and the local curvature of the surface among different slides may have a
large impact on scanning, especially for confocal scanners which are sensitive to
focus. The influence of these factors on measuring signal intensity is defined as
slide effect. Tseng et al. (2001) showed that such effects are significant, and
normalization is slide-dependent.
Variation from probe labeling or label effect. The most commonly used
fluorescent dyes, Cy3 and Cy5, are not equally incorporated into DNA molecules
by reverse transcriptase and DNA polymerase. Cy3 is incorporated more
efficiently than Cy5 with the same preparation and amount of RNA. While both
Cy dyes are relatively unstable, Cy3 and Cy5 have different quantum efficiencies
and are detected by the array scanner with different efficiencies. While the
detection limit of Cy5 with the scanner is lower than that of Cy3, Cy5 is more
sensitive to photobleaching. The influence of these factors on intensity
measurements is referred to as label effect.
The use of two fluorescent dye labels may also introduce gene-label
interactions. For instance, fluorescent labeling may fluctuate systematically,
depending on the nucleotide composition of the target sequences, and one or the
other dye may be preferentially incorporated into specific gene sequences. Also,
the length of Cy3- and Cy5-labeled cDNA by reverse transcription with random
priming could be significantly different from sequence to sequence. This longer
labeled cDNA could potentially lead to higher intensity levels for certain arrayed
probes. If such interaction occurs, certain sequences will always show higher
intensities in one channel than the other channel even under non-differential
conditions and after normalization.
Variation in growth conditions and mRNA preparation or sample effect. In a

comparative microarray experiment, two RNA samples extracted from cells
grown under different conditions are labeled with different fluorescent dyes.
Because of the differences in genetic identity (e.g., wild type versus mutant
strains) and environmental growth conditions, cell biomass and mRNA
abundance could fluctuate significantly among different cultures. The RNA
purity also could be very different from sample to sample and this could lead
to different labeling and hybridization efficiencies. Furthermore, sensitivity to
mRNA degradation could be considerably different between preparations. All of
these factors affecting signal intensity are referred to as sample effects. Due to
experimental variations, hybridization signals from microarrays should be
normalized prior to comparing data from a single array or multiple arrays.
2. Genes Used for Normalization
Two critical issues in the analysis of microarray data, are how to eliminate
systematic variations and which genes should be selected as references for
normalization. Experimental design largely determines the strategy used for
normalization, three of which are described in detail below.
Using all genes on the array or global normalization. Under a certain condition,
only a small portion of the genes is expected to be differentially expressed. Thus,
the remaining genes should exhibit constant expression levels between two
channels and can be used for normalization to calibrate spatial effects (Dudoit
et al., 2001), slide effects and label effects (Tseng et al., 2001). The prerequisite
for using almost all genes on the array for normalization is that only a small
fraction of the genes are expressed, and the numbers of down- or up-regulated
genes are approximately equal.
Using constantly expressed housekeeping genes. The housekeeping genes that

are constantly expressed across a variety of conditions can be used for
normalization (Duggan et al., 1999). Although it can be difficult to identify a set
of housekeeping genes that do not change significantly under any condition, it
may be possible to identify small sets of temporary housekeeping genes for

particular experimental conditions. A limitation in using housekeeping genes for
normalization is that housekeeping genes are generally highly expressed and thus
may not be representative of other genes of interest.
Using controls. A third normalization approach is to use spiked controls or a

titration series of control sequences. In the spiked control method, DNA
sequences from organisms different from the ones being studied are printed on
the array and then the mRNAs of the control sequences are mixed with the two
different mRNA samples in equal amounts. These spotted controls should have
equal Cy5- and Cy3-derived intensities, and thus can be used for normalization.
One limitation is that the composition of the control sequences could be
considerably different from the target sequences, and as a result, they may not
be representative of the genes of interest. Another limitation is that it may be
difficult to determine how much mRNA to spike, because there are always
varying amounts of rRNA and tRNAs present, and the degree of RNA
degradation varies from sample to sample.
In the titration series method, a series of concentrations of the control
sequences are printed on the arrays. These control spots are expected to have
equal Cy5- and Cy3-derived intensities across a range of concentrations.
Genomic DNA could be used in the titration method, because it should have a
consistent expression level across various conditions.
3. Experimental Design and Normalization Strategies
Since microarray experiments have inherently high variation, careful

experimental design and execution are critical for accurately identifying
differentially expressed genes under different conditions. Appropriate normal-
ization is therefore necessary to eliminate different types of systematic variations.
Minimizing spatial effects. To minimize spatial effects, multiple spots for a gene
or control DNA should be fabricated on the microarrays. For control sequences,
various concentrations of sequences should be spotted on arrays. Multiple spots
of genes or control DNAs within the same slide are very useful for identifying
contaminated spots, spots having high background noise, and problematical
slides in each experiment (Tseng et al., 2001). To minimize spatial effects,
normalization can be performed for each sector of the microarray-based on all the
genes in that sector. Since DNAs in different sectors are deposited by different
pins, normalization is an effective way to eliminate pin-to-pin variations (Dudoit
et al., 2001). By comparing the normalization results for different genes and
control sequences among different sectors, one should be able to assess and
minimize the systematic variations associated with slide surface properties.
Minimizing labeling effects and label –gene interaction. To eliminate systematic

variations in probe labeling and gene-label interaction, a reverse labeling
experimental design is recommended (Kerr and Churchill, 2001a; Tseng et al.,
2001). For this, two aliquots of the two RNA samples (A and B) are labeled with
Cy3 or Cy5 separately, and then hybridized with two microarrays. The
hybridization solution for the first microarray consists of Cy3-labeled sample
A and Cy5-labeled sample B, whereas the labeling for the second microarray
hybridization is reversed for the two target samples. Then the signal intensity for
each microarray is normalized based on all genes on the microarray or on a set of
housekeeping genes using different normalization approaches (see below). After
normalization, the signal intensities from both channels for each sample are
averaged, and the intensity ratios of the two samples are calculated based on the
averaged signals. The reverse labeling experimental design is effective in
eliminating labeling effects and gene-label interaction (Tseng et al., 2001).
Minimizing slide and sample effects. In a typical comparative study, multiple

replicated treatments (e.g., 3) under each condition are used and mRNAs from
two different conditions are fluorescently labeled and co-hybridized to the same
single or multiple slides. Under such an experimental design, the signal intensity
is impacted by both slide and sample effects, and it will be difficult to eliminate
the resulting systematic variation based on using all arrayed genes for
normalization. In this situation, one should identify a sufficient number of non-
differentially expressed genes on each slide and use them to construct a
normalization curve, because the expression level of the non-differentially
expressed genes are expected to remain constant under the experimental
conditions tested. One challenge is to select the non-differentially expressed
genes. Although predetermined housekeeping genes are good candidates, they
may not provide a good fit for normalization due to the high level of expression
and natural variability of their expression level.
A rank invariant selection approach (Schadt et al., 2000; Tseng et al., 2001)
can be used for selecting non-differentially expressed genes. This method
presumes that for an up-regulated gene, the signal intensity rank for a channel
will be significantly higher than the rank in the other channel, and vice versa.
Briefly, the signal intensities of individual genes from both channels are ranked.
If the ranks of Cy3 and Cy5 intensities for a gene differ by less than a
certain threshold value, and the rank of the averaged intensity is not within
the known levels of the lowest and highest ranks, then this gene is classified
as a non-differentially expressed gene (Tseng et al., 2001). This method
works well if the majority of the genes are not differentially expressed. However,
this method may fail if majority of the genes are up- or down-regulated
(Tseng et al., 2001).
4. Normalization Approaches
Methods for the normalization of microarray hybridization data can generally

be categorized as linear or nonlinear. The major difference between these two
types is that linear methods multiply all values in one channel by a correction
factor, whereas nonlinear methods, which are preferred by most researchers
working with microarrays, take the channel intensity into account and therefore
are thought be more accurate. Here, we briefly describe the most commonly used
normalization methods.
Correction factor based on total intensity. This method calculates a correction

factor based on the total measured fluorescence intensity. The primary underlying
assumption is that the total amount of RNA labeled with Cy3 and Cy5 is equal
because the same amount of RNA from the same sample is used in separate
labeling reactions. Although the spot for any one gene in one channel may be
higher than that in the other, such variations should be averaged out over
thousands of spots on the array. Therefore, the total integrated intensity of all
spots should be equal in both channels, and a constant signal correction factor can
be derived to rescale the signal intensity of the other channel.
Linear regression method. For differential experiments, it is expected that many

genes will be expressed at a nearly constant level under two different growth
conditions or treatments. Thus, the slope of the intensity in a scatter plot of both
channels should be 1. Based on this assumption, the slope can be calculated by
linear regression to obtain a correction factor, and then all values in one channel
are multiplied by the correction factor to adjust the slope to 1.
Trimmed geometric mean (TGM). This nonlinear method was initially described
by Morrison et al. (1999) and is generally recommended for most normalization
needs. The method assumes that under a certain condition, only a small
proportion of the genes will be differentially expressed. Thus, the remaining
genes should display a constant level of expression and can be used for
normalization (Beliaev et al., 2002; Thompson et al., 2002). The signals from
each channel are log transformed and sorted based on the intensity, then 5% of
the extreme values (minimum and maximum) are discarded. The log-TGM and
the SD of the log-trimmed means are calculated. The normalized value for a gene
is obtained by dividing the difference between log intensity and log-trimmed
means by the SD of the log-trimmed means. The normalized values are then
converted back from log to normal values, which are then used to calculate
expression ratios.
Intensity-dependent nonlinear normalization method. In many cases, the dye bias

is dependent on spot density (Dudoit et al., 2001; Tseng et al., 2001). Thus, an
intensity-dependent normalization method may be preferable. Yang et al. (2001)
proposed an intensity-dependent nonlinear normalization method that utilizes
most of the genes on an array. Since this method is complicated, the reader is
referred to the original paper for details (Dudoit et al., 2001). Briefly, the log
intensity ratio and the mean log intensity of both channels are calculated. The
normalized intensity ratio is the difference between the actual log intensity ratio
and the intensity ratio estimated based on Lowess function. Theoretically, this
normalization method should be the most robust.
B. DATA TRANSFORMATION
Prior to statistically analyzing the microarray data, it is important to establish

whether the data meet the underlying assumptions of the particular statistical
model that will be used. The most common requirements for statistical techniques
are that the data have a normal distribution and homogeneous variance. If the data
do not meet these assumptions, they may be transformed and reevaluated to
determine if they meet the underlying assumptions. If the data do not meet the
assumptions, the statistical analyses will not be valid.
Although there are many different approaches to data transformation, the most
commonly used approach in microarray studies is taking the logarithm of the
quantified expression values. The rationale for this is three-fold. First, the
variation in logs of intensities and logs of ratios of intensities are less dependent
on absolute magnitude. Log transformation can equalize variability in microarray
data with high variability. Second, log transformation evens out highly skewed
distributions and thus brings the data closer to a normal distribution. Third,
normalization is additive for logs of intensities. Studies show that log
transformation is very effective in bringing the microarray data approximately
to a normal distribution and is the best approach for the analysis of microarray-
based gene expression data (Kalosai and Shams, 2001).
C. METHODS FOR IDENTIFYING DIFFERENTIALLY

EXPRESSED GENES
Generally, normalized intensity ratios under two different experimental

conditions are used to assess differentially expressed genes. Standard statistical
techniques cannot be easily used to determine which level of difference in gene
expression reflects an actual biological difference. This is because of the

inherently high variation associated with microarray experiments and low-level
replications.
Three basic approaches are currently used for identifying differentially
expressed genes. The first approach, which is commonly reported in the
literature, is based on arbitrarily assigned fold differences (Schena et al., 1996;
Heller et al., 1997). If the average expression level varies by more than a constant
factor (e.g., 2) between the treatment and control conditions, then this gene is
considered to have changed significantly in its expression. However, such a fixed
fold rule is unlikely to identify real biologically differences, because a factor of
two has a different significance, depending on the levels of gene expression and
variation. The fold rule method is applicable only when the variance among the
replicates within a treatment is identical for every gene so that the sample
variance can be ignored. However, in practice, the variance differs among genes,
and it is critical to incorporate such information into a statistical test.
The second approach is to use standard statistical t-test (Baldi and Long,
2001; Beliaev et al., 2001; Thompson et al., 2002) or paired t-tests (Rogge et al.,
2000) using the intensity ratio or log of the intensity ratio to test whether the
fold change is significantly different from 1 or 0. When the t value exceeds
a certain threshold, depending on the confidence level selected (typically the
95% confidence level or P , 0.05), the gene expression level is considered to
be significantly different between two conditions. The t-test incorporates
variance information and could potentially overcome the drawbacks of the fold
rule method. Application of the t-test requires that all microarray experiments
be highly replicated to obtain accurate estimates of the variance within
experimental treatments. However, the level of replication within experimental
treatments is often too low to permit t-tests, because the microarray experiments
are costly and time-consuming to repeat or the amount of biological samples is
very limited. A small number of replicates could lead to inaccurate estimation
of variance and a correspondingly poor performance of the t-test itself (Baldi
and Long, 2001).
The third approach is to apply Bayesian probabilistic model-based regularized
t-test to improve the confidence in interpreting DNA microarray data with a low
number of replicates (Baldi and Long, 2001). This method assumes that genes of
similar expression levels have similar measurement errors, and that data from all
of the genes with similar expression can serve as pseudo-replication of the
experiment. Thus, variance of any single gene can be estimated by the weighted
average of the variances from a number of genes with similar expression levels.
This method has been applied to identify global expression profiles in E. coli K12
(Long et al., 2001). The results showed that the Bayesian approach identified a
stronger set of genes that were significantly up- or down-regulated and required
less replication to achieve the same level of reliability as the t-test method. Since
this method is computationally demanding, a program for accommodating this
approach, Cyber-T, is available at the Web interface, www.genomics.uci.edu/

software.html. Various statistical methods are also available in ArrayState for
identifying differentially expressed genes.
D. MICROARRAY DATA ANALYSIS
A massive amount of data is generated by microarray hybridization, and the

great challenge is how to extract meaningful biological information. One of the
key goals for microarray data visualization and analysis is to identify statistically
significant up- and down-regulated genes and co-regulated genes exhibiting
similar expression patterns. Although many different statistical methods have
been used for analyzing microarray data, they are still in the early stages of
development. In this section, several current methods will be briefly reviewed.
1. Scatter Plot
Scatter plots are the simplest way to visualize microarray expression data. In a
comparative experiment, microarray hybridization is generally performed with
two samples from two different conditions. One can use a scatter plot to visualize
up- and down-regulated genes by assigning x- and y-axis values to represent
signal intensity under the two different conditions. In the scatter plot, genes with
equal expression values for two conditions fall along the diagonal identity line,
whereas genes that are differentially expressed fall-off the diagonal line; the
greater the deviation from the diagonal line, the greater the difference in the
expression of a given gene between two samples.
2. Similarity Measurement
In a typical microarray experimental design, multiple experimental conditions

at multiple time points are generally compared. In large experiments analyzing
thousands of genes, the increased data volume makes it very difficult to identify
gene expression patterns using scatter plots. More sophisticated multivariate
analysis techniques should therefore be used in such cases. To use different
multivariate analysis methods, the relationships among different genes should
first be quantified based on signal intensity using appropriate metrics.
Two approaches are generally used for quantifying the relationships among
different genes. One approach is to use Euclidean distance, which is defined as
the square root of the summation of the squares of the differences between all
pair-wise comparisons. This metric measures the absolute distance between two
points in space defined by the two gene expression profiles. In general, such
distance measures are suitable when the objective is to cluster genes with similar
expression patterns.
The other approach is to use Pearson correlation coefficient. For understanding
regulatory networks, it is biologically more interesting to search for genes
expressed at different levels but with similar overall profiles. Pearson correlation
coefficient is ideal for identifying profiles of similar shape. The values of
this correlation coefficient range from 21 (negative correlation) to 1 (positive
correlation), and the method can detect both negatively and positively correlated
genes. Several variations of the correlation metric have been used such as the
correlation coefficient with an offset of zero for specifically taking into account
the reference state (Eisen et al., 1998) and jackknife correlation to counter against
outlier effects (Heyer et al., 1999).
3. Principal Component Analysis
Principal component analysis (PCA) is an exploratory multivariate statistical

method for simplifying data sets that reduces the dimensionality of the variables
by finding new variables, which are independent of each other. A few of the new
variables, typically 2– 3, are selected to explain the majority of variance in the
original data. Since each principal component is a linear combination of the
original variables, it is often possible to assign meaning to what the principal
components represent. For microarray data analysis, genes or experiments can be
considered as variables. PCA has been used in a variety of biochemical studies,
including the analysis of microarray data in identifying outlier genes and/or
experiments (Hilsenbeck et al., 1999). The main advantage of PCA is that it
identifies outliers in the data or genes that behave differently than most of the
genes across a set of experiments. It can also be used to visualize clusters of genes
that behave similarly across different experiments. However, the number of
clusters in the data sets is arbitrary and dependent on the user’s intuition or
experience.
4. Cluster Analysis
One of the most commonly used methods is cluster analysis. Cluster analysis
is used to identify groups of genes, or clusters that have similar expression
profiles. Clusters and the genes within them can be subsequently examined for
commonalities in functions as well as sequences in order to gain a better
understanding of how and why they behave similarly. Cluster analysis can help
establish functionally related groups of genes and can predict the biochemical
and physiological roles of functionally unknown ORFs.
Despite the emergence of many methods for microarray data analysis, the
optimal way of classifying such data is still open to debate. Depending on the way
in which the data are clustered, cluster analysis can be divided into hierarchical
clustering and non-hierarchical clustering.
(i) Hierarchical clustering. This method attempts to group genes and/or

experiments in small clusters and then group these clusters into higher-level
clusters and so on. As a result of this grouping process, a tree structure called
a dendrogram is generated for visualization of the relationships between
genes and/or experiments. There are three common options for hierarchical
analysis based on the definition of the distance between two clusters: single
linkage, average linkage, and complete linkage (Heyer et al., 1999).
Although there are numerous versions of the basic algorithm, the most
common is known as average linkage. Applications of hierarchical
clustering to gene expression data have been described in recent studies
(Eisen et al., 1998).
Hierarchical clustering methods are very popular due to their simplicity
and analysis speed. However, there are several problems associated with
these methods (Heyer et al., 1999). First, decisions to group two elements
are based only on the distance between them and once elements are joined, it
is impossible for them to be separated. In addition, it is a local decision-
making method and does not consider the data as a whole. It suffers from a
lack of robustness and solutions may not be unique and dependent on the
data order, leading to incorrect clustering overall. Finally, the tree is
extremely complex for large data sets, with the performance decreasing with
the square of the number of genes requiring classification.
(ii) Non-hierarchical clustering. One of the typical non-hierarchical clustering
methods is k-means clustering, which identifies predetermined k points as
cluster centers. Each data point is assigned to one of these centers in a way
that minimizes the total of distance between all points and their centers. The
subsequent centers are chosen by identifying the data points farthest from
the centers already chosen, and this process is iterated until the cluster
memberships do not change appreciably (Tavazoie and Church, 1998).
The advantage of k-means clustering is that it provides sufficient clustering

without having to create the full distance and similarity matrix or scan the whole
dataset excessively (Zhou et al., 2000). This is particularly useful for microarray
data with large numbers of genes and many different experimental conditions.
The algorithm converges quickly for good initial choices of the cluster centers.
The main disadvantage of this method is that the number of clusters, k, must be
specified prior to running the algorithm, and the final clustering relies heavily on
the choice of k. Generally, the number of clusters is not known in advance.

In addition, the quality of the clusters identified by k-means is not guaranteed
(Heyer et al., 1999). Recently, a new version of k-means (progressive k-means)
was proposed to analyze gene expression data. This new procedure identifies the
number of different clusters from the data itself and is independent of a priori
specified number of clusters (Ben-Dor et al., 1999; Herwig et al., 1999). Despite
such limitations, the k-means methods appear to perform quite well with a large
number of genes (Dopazo et al., 2001).
To avoid limitations of hierarchical clustering and k-means clustering, another
non-hierarchical clustering procedure, quality cluster, was developed (Heyer
et al., 1999) that focuses on identifying large clusters with a quality guarantee.
The quality cluster allows each ORF to initiate a candidate cluster, which is
formed by starting each ORF and grouping the ORF with the greatest jackknife
correlation coefficient. Other ORFs are iteratively added in a way to minimize the
increase in cluster diameter without removing the ORFs, which previous clusters
included (Heyer et al., 1999). One characteristic of this procedure is that the
number of candidate clusters is equal to the ORF numbers and many candidate
clusters overlap, with the largest candidate cluster being retained. The ORFs it
contains are eliminated and the entire procedure is iterated on the smaller set of
ORFs until the largest remaining cluster has fewer than some pre-specified
number of elements.
There are several advantages of the quality cluster over both hierarchical and
k-means clustering. First, the total number of clusters is not required prior to
running the algorithm, and the quality of all clusters are guaranteed. Second,
although the quality cluster algorithm is similar to the complete linkage
hierarchical procedure, the clusters identified at a specified threshold are much
larger on average. Third, since each ORF is considered a potential cluster center,
local decisions do not have a great impact on the final clustering results. Thus, it
is expected that this method is less sensitive than hierarchical approaches to small
changes in the data such as removal of ORFs through filtering. Finally, this
method is not sensitive to the order in which the similarity or distance data
appear. Since this is a new clustering method, its value as an analysis tool remains
to be determined.
5. Neural Network Analysis
Since clustering methods have some serious drawbacks in dealing with data
with a significant amount of noise, a fundamentally different neural network-
based approach has been proposed for microarray data analysis (Tamayo et al.,
1999; Toronen et al., 1999; Herrero et al., 2001). Unsupervised neural networks,
and in particular self-organizing maps (SOMs), are a more robust and accurate
method for grouping large data sets. The algorithm for neural network analysis
works in the following way. First, a two-dimensional grid typically of hexagonal
or rectangular geometry is defined. Then, similar to k-means clustering, the
number of clusters (k) is specified to correspond to the representative points in
the specified geometrical configuration. Data points are mapped onto the grid and
the positions of the representative points are iteratively relocated in a way that
each center has one representative point. Clusters close to each other in the grid
will be more similar to each other than those further apart.
The main advantage of SOMs is that they are robust to noise. In other words,
they are able to handle large data sets containing noisy, poorly defined items with
irrelevant variables and outliers. This is particularly useful for analyzing
microarray data. SOMs are also reasonably fast and can be easily scaled up to
large data sets. One disadvantage of SOMs is that they require pre-determined
choices about geometry, like the k-means method. The number of clusters is
arbitrarily fixed from the beginning and consequently, it is difficult to recover the
natural cluster structure of the data set. SOMs also yield non-proportional
classification. If irrelevant data or some particular type of profile is abundant, the
most interesting profile will be mapped in few clusters and hence their resolution
could be low. In addition, it is very difficult to detect higher-order relationships
between clusters of profiles due to the lack of a tree structure (Herrero et al., 2001).
To overcome some of the limitations of SOMs, an unsupervised neural
network with a binary tree topology, termed the self-organizing tree algorithm
(SOTA), was proposed (Dopazo and Carazo, 1997). This new algorithm
combines the advantages of hierarchical clustering (tree topology) and neural
network (accuracy and robustness) and was used to analyze gene expression data
(Herrero et al., 2001).
VII. USING MICROARRAYS TO MONITOR GENOMIC

EXPRESSION
Microarrays have been used widely to quantify and compare global gene
expression in a high-throughput fashion. This section briefly reviews the
fundamental basis, general approaches to experimental design, and hybridization
performance of microarrays in monitoring gene expression levels.
A. GENERAL APPROACHES TO REVEALING DIFFERENCES IN

GENE EXPRESSION
Temporal and spatial information concerning gene expression, as well as

changes in mRNA abundance levels in response to different environmental
conditions, are important for understanding gene function and regulation. Three
comparative approaches have been used for the display of differential gene
expression.
1. Differential Display of mRNA Under Different Physiological

Conditions
Cells of interest are cultured under different physiological conditions, and the
differences in mRNA abundance between the test and reference samples are
compared using high-density microarrays. This is the most straightforward and
widely used approach for identifying gene expression patterns associated with
various physiological states (DeRisi et al., 1997; Tao et al., 1999; Ye et al., 2000;
Beliaev et al., 2002).
2. Differential Display of Temporal Gene Expression
Cells of interest are grown under a specific physiological condition and then
harvested at different time points during growth. Changes in mRNA levels are
revealed using microarrays. Information on the temporal dynamics of gene
expression is very useful in understanding when genes are turned on or off and
how genes interact with each other (DeRisi et al., 1997; Liu et al., 2003).
3. Comparison of Gene Expression Patterns Between Wild-type

and Mutant Cells
Differences in gene expression in response to changing environmental

conditions can be very complicated, and oftentimes the expression profiles of
many genes are altered as a result. Changes in the expression profiles for many
genes present a great challenge to understand the underlying molecular
mechanisms controlling these genes. The most effective approach to define the
contributions of individual regulatory genes in a complex metabolic process is to
use DNA microarrays to identify genes whose expression is affected by mutations
in putative regulatory genes (DeRisi et al., 1997; Beliaev et al., 2002; Thompson
et al., 2002).
The basic approach to microarray-based gene expression studies is outlined in
Fig. 5. In a typical microarray experiment for monitoring gene expression, gene-
specific PCR primers are designed based on whole-genome sequence information
and synthesized. Gene-specific fragments are then amplified with specific
primers, purified, and arrayed on solid substrates. Once the microarrays are ready,
total cellular RNA isolated from bacterial cells grown under two different
Figure 5 General approach for using microarrays for monitoring gene expression.
conditions (a control and experimental condition) is fluorescently labeled with

different dyes (Cy3 or Cy5) via the enzyme, reverse transcriptase. The microarray
is then simultaneously hybridized with fluorescently tagged cDNA from the test
and reference samples. The signal intensity of each fluorescent dye on the array is
then measured with a confocal laser scanning microscope or CCD camera. The
quantitative ratio of red (Cy5) to green (Cy3) signal for each spot reflects the
relative abundance of that particular gene in the two experimental samples. With
appropriate controls, the intensity can be converted into biologically relevant
outputs (e.g., the number of transcripts per cell). A series of samples can be
compared with each other through separate co-hybridizations with a common
reference sample, and the data can be analyzed with various statistical methods.
Eisen and Brown (1999) provide a detailed discussion of the technical aspects of
microarray experiments for monitoring gene expression.
B. EXPERIMENTAL DESIGN FOR MICROARRAY-BASED

MONITORING OF GENE EXPRESSION
Microarray experiments generate massive data sets, which must be analyzed

and interpreted in a rapid and meaningful way. To improve the efficiency and
reliability of experimental data, careful experimental design is needed. Without
this, the collected data may fail to answer the research question of interest or lead
to a biased, inadequate interpretation of the experimental results (Yang and
Speed, 2002).
The main objective of experimental design is to make the data analysis and
interpretation as simple and powerful as possible. For a competitive microarray
hybridization experiment in which two fluorescent dyes are used, the most
important experimental design issue is how the mRNAs are labeled and which
mRNAs are hybridized together on the same slide (Yang and Speed, 2002). In
most experiments, several designs can be devised. The selection of the most
appropriate design will depend on the particular research questions being asked,
the number of comparisons, the number of slides available for hybridization, the
amount of mRNAs available, and cost.
Various design schemes have been described in great detail by Yang and
Speed (2002) and several designs could be devised for a particular microarray
experiment. The microarray experiment design scheme can be classified into the
three categories (Fig. 6): reference design, all-pairs design, and loop design. In
reference design, all treatment samples are labeled with one dye and are
hybridized, respectively, with a common reference sample labeled with another
dye (Fig. 6A). This indirect design is used widely in gene expression studies. This
design is especially suitable when the amount of mRNA from treatment samples
is limited and when many treatment samples are compared. Another advantage of
this design is that data analyses and interpretation are easy and do not require
sophisticated statistical tools. However, the average variance for this indirect
reference design is considerably higher than that for the other designs. Since it is
straightforward, the reference design is used much more often than the other
designs.
Figure 6 Illustrations of basic types of microarray experimental design schemes with five
treatment samples. By convention, the green-labeled sample (Cy3) is placed at the tail while the red-
labeled sample (Cy5) is placed at the head of the arrow. (A) Reference design. The five treatment
samples (A –E) are labeled with one dye and hybridized, respectively, with the common reference
sample R, which is labeled with the other dye. Altogether five hybridizations are needed. (B) All-pair
design. Each sample is labeled twice with red and twice with green. Ten pair-wise hybridizations are
needed. (C) Loop design. Each sample is labeled once with red and once with green. Five successive
pair hybridizations are needed.
In the all-pairs design scheme, all of the treatment samples are labeled with
different fluorescent dyes and directly hybridized together in pair-wise fashion
(Fig. 6B). The main advantage of this design is that more precise comparisons
among different treatment samples can be obtained. However, this design is
unlikely to be feasible and desirable when a large number of comparisons are
made, due to the constraints on mRNA quantity and cost. Finally, in the loop
design, all of the treatments are successively connected as a loop (Fig. 6C) (Kerr
and Churchill, 2001a, b). Using the same number of microarrays as the reference
design, the loop design obtains twice as much data on the treatments of interest.
The loop design requires far fewer slides than the all-pairs design. However, long
paths between some pairs of treatment samples are needed in larger loops, and
thus some comparisons are much less precise than others (Yang and Speed,
2002). Another practical problem is that each sample should be labeled with both
Cy dyes, which doubles the number of labeling reactions. In addition, the failure
of microarray hybridization in one sample will affect the analysis of other
samples in the loop.
C. MICROARRAY-BASED FUNCTIONAL ANALYSIS OF

ENVIRONMENTAL MICROORGANISMS
Genome sequence information for a number of bacteria and archaea of

potential environmental or biotechnological relevance is accumulating rapidly
and includes representatives from such genera as dissimilatory metal-reducing
bacteria (Shewanella oneidensis [Heidelberg et al., 2002]), extreme radiation-
resistant bacteria (Deinococcus radiodurans [White et al., 1999]), photosynthetic
cyanobacteria (Anabaena sp. strain PCC 7120 [Kaneko et al., 2001],
Synechocystis sp. strain PCC6803 [Kaneko et al., 1996]), thermophilic and
hyperthermophilic archaea (Pyrococcus horikoshii [Kawarabayasi et al., 1998],
Aeropyrum pernix [Kawarabayasi et al., 1999], Thermotoga maritima [Nelson
et al., 1999], Thermoplasma volcanium [Kawashima et al., 2000], Pyrococcus
furiosus [Robb et al., 2001], Pyrobaculum aerophilum [Fitz-Gibbon et al.,
2002]), thermoacidophilic archaea (Sulfolobus tokodaii [Kawarabayasi et al.,
2001], Sulfolobus solfataricus [She et al., 2001]), methanogens (Methanococcus
jannaschii [Bult et al., 1996], Methanobacterium thermoautotrophicum [Smith
et al., 1997], Methanopyrus kandleri [Slesarev et al., 2002], Methanosarcina
acetivorans [Galagan et al., 2002]), sulfate-reducing archaea (Archaeoglobus
fulgidus [Klenk et al., 1997]), and halophilic archaea (Halobacterium species
NRC-1 [Ng et al., 2000]). However, to-date, very few studies have explored the
transcriptomes of these organisms using microarray technology. The large
majority of microarray-based genomic expression analyses have focused on
bacterial pathogens and such model organisms as E. coli, B. subtilis, and yeast.
In this section, we will briefly discuss microarray profiling of gene expression
in three organisms of environmental significance, namely, S. oneidensis,
D. radiodurans, and P. furiosus, as examples of the application of microarrays
to environmental microbiology.
1. Shewanella oneidensis, a Dissimilatory Metal-reducing Bacterium
S. oneidensis MR-1 (formerly Shewanella putrefaciens strain MR-1

[Venkateswaran et al., 1999]) is a facultatively anaerobic g-proteobacterium
that is noted for its remarkably diverse respiratory capacities. In addition to

utilizing oxygen as a terminal electron acceptor during aerobic respiration,
S. oneidensis can anaerobically respire various organic and inorganic substrates,
including oxidized metals (e.g., Mn(III) and (IV), Fe(III), Cr(VI), U(VI)), fuma-
rate, nitrate, nitrite, thiosulfate, elemental sulfur, trimethylamine N-oxide,
DMSO, and anthraquinone-2,6-disulphonate (Lovley, 1991; Nealson and
Saffarini, 1994; Moser and Nealson, 1996). This unusual versatility in the use
of alternative electron acceptors for anaerobic respiration is conferred in part by
complex electron transport networks, the components of which remain to be
elucidated (Richardson, 2000). The metal ion-reducing capabilities of this
bacterium, in particular, have important implications with regard to the potential
for in situ bioremediation of metal contaminants in the environment. However,
the effective prediction and assessment of bioremediation performance or activity
is complicated due to insufficient knowledge concerning the gene networks and
regulatory mechanisms enabling microbial metal reduction.
To expedite understanding of metal reduction by S. oneidensis MR-1, its
~5-Mb genome was determined recently by The Institute for Genomic Research
(TIGR) under the support of the U.S. Department of Energy (DOE) (Heidelberg
et al., 2002), making it feasible to apply microarray technology to the study
of energy metabolism in this bacterium. The transcriptional response of
S. oneidensis to different respiratory growth conditions (Beliaev et al., 2002b)
and to the disruption (inactivation) of genes encoding putative transcriptional
regulators (Beliaev et al., 2002a; Thompson et al., 2002) were examined using
DNA microarrays containing 691 arrayed genes. These partial genome
microarrays consisted of PCR-amplified MR-1 ORFs putatively involved in
energy metabolism, transcriptional regulation, adaptive responses to environ-
mental stress, iron acquisition, and transport systems according to the sequence
annotation. These arrays were constructed prior to the closure and publication of
the S. oneidensis genome sequence.
To identify genes specifically involved in anaerobic respiration, differential
mRNA expression profiles of S. oneidensis were monitored under aerobic and
fumarate-, Fe(III)-, or nitrate-reducing conditions using partial genome
microarrays (Beliaev et al., 2002b). Gene expression profiling indicated that
121 of the 691 arrayed ORFs showed at least a 2-fold difference in mRNA
abundance in response to changes in growth conditions (Beliaev et al., 2002b),
with a number of genes required for aerobic growth being repressed in the
transition from aerobic to anaerobic respiration. Genes induced in a general
response to anaerobic respiration, irrespective of the terminal electron acceptor,
belonged to several different categories of cellular function: cofactor biosynthesis
and assembly, substrate transport, and anaerobic energy metabolism. Of
particular importance was the observation that certain genes preferentially
displayed increased transcript levels in response to specific electron acceptors.
For example, the expression of genes encoding a periplasmic nitrate reductase
(napBHGA operon), cytochrome c552, and prismane was elevated 8- to 56-fold

specifically in response to the presence of nitrate, while genes encoding a tetra-
heme cytochrome c (cymA), a flavocytochrome c (ifcA), and a fumarate reductase
( frdA) were preferentially induced 3- to 8-fold under conditions of fumarate
reduction. In addition, the mRNA abundance levels for two oxidoreductase-
related genes of unknown function and several cell envelope genes involved in
multidrug resistance increased specifically under Fe(III)-reducing conditions.
This work represented the first attempt to characterize a complex system in
S. oneidensis on a genome scale. Other microarray-based transcriptomic studies
have focused on defining the functions of putative S. oneidensis regulatory genes
encoding a ferric uptake regulator ( fur; Thompson et al., 2002) and an electron
transport regulator (etrA; Beliaev et al., 2002a).
2. Deinococcus radiodurans, an Extreme Radiation-resistant Bacterium
D. radiodurans strain R1 is the most characterized member of the DNA-

damage resistant bacterial family Deinococcaceae, which is comprised of at least
seven different species that form a distinct eubacterial phylogenetic lineage
(Makarova et al., 2001). D. radiodurans is a Gram-positive, non-sporulating
bacterium that was originally isolated in 1956 from canned meat that had spoiled
following exposure to X-ray sterilization (Anderson et al., 1956). Species in the
genus Deinococcus, particularly D. radiodurans, are extremely resistant to a
number of physicochemical agents and environmental conditions that damage
DNA, including ionizing and ultraviolet radiation, desiccation, heavy metals, and
oxidative stress (reviewed in Minton, 1996; Battista, 1997; Battista et al., 1999).
Studies have demonstrated that D. radiodurans can survive acute exposures to
gamma radiation that exceed 15,000 Gy without lethality or induced mutation
(Daly et al., 1994; Daly and Minton, 1995) and flourish in the presence of high-
level chronic irradiation (60 Gy/h) (Lange et al., 1998; Venkateswaran et al.,
2000). D. radiodurans also expresses an intrinsic ability to reduce metals and
radionuclides (Fredrickson et al., 2000) and thus has potential applications for the
bioremediation of metal- and radionuclide-contaminated sites where the presence
of radioactivity prohibitively restricts the activity of more sensitive dissimilatory
metal-reducing bacteria such as Shewanella.
To enhance the understanding of the molecular basis of extreme DNA damage
resistance, the complete genome of D. radiodurans R1 was sequenced by TIGR
(White et al., 1999) under DOE support. Sequence analysis of this organism’s
multigenomic content indicates that essentially the entire repertoire of
recombinational DNA repair genes identified in D. radiodurans has functional
homologs in other prokaryotes (White et al., 1999; Makarova et al., 2001),
suggesting that the extreme radioresistance of R1 may be attributable
to novel genes, repair pathways, and mechanisms yet to be described.
Detailed computational genomic analyses alone, therefore, are unlikely to

uncover the fundamental answers underlying the remarkable ability of
D. radiodurans to withstand DNA-damaging conditions.
The transcriptome dynamics of D. radiodurans in response to cellular
recovery from acute ionizing radiation was examined using DNA microarrays
covering ~94% of the organism’s predicted protein-encoding genes (Liu et al.,
2003). In this time-series study, D. radiodurans cells exposed to acute irradiation
(15 kGy) were allowed to recover at 378C for time intervals ranging from 0 to
24 h. Deinococcus transcriptome dynamics were monitored in cells representing
early (0 – 3 h), middle (3 –9 h), and late (9 –24 h) phases of recovery from
ionizing radiation and compared to non-irradiated control cells. Microarray analy-
sis of genomic expression patterns revealed a large number of D. radiodurans
genes responding to acute irradiation: 832 genes (28% of the genome) were
induced and 451 genes (15% of the genome) were repressed two-fold or greater at
one point during D. radiodurans recovery (Liu et al., 2003). Genes exhibiting
increased transcription in the early phase of cell recovery belonged to a number
of broad functional groups, including DNA replication, DNA repair, recombina-
tion, cell wall metabolism, cellular transport, and uncharacterized proteins.
Hierarchical clustering of genes showing differential expression revealed
similar expression patterns for groups of genes and clusters of presumably
co-regulated genes (Fig. 7). Genes responding to recovery from irradiation
clustered into three distinct groups: (1) recA-like activation pattern (based on the
expression profile of recA, which is critical for D. radiodurans recovery and is
substantially upregulated during early-phase recovery and down-regulated before
the onset of late phase), (2) growth-related activation pattern, and (3) repressed
patterns. Unexpectedly, genes encoding tricarboxylic acid (TCA) cycle
components were repressed in the early and middle phases of recovery, whereas
genes encoding the glyoxylate shunt pathway were induced during this interval
(Liu et al., 2003). In addition, a number of poorly characterized genes showed
high induction folds in expression during at least one phase of recovery, thus
implicating their encoded proteins in the functional role of cell recovery. The
response of metabolic gene systems, however, is not immediately clear and will
require further, more focused experimentation. The study by Liu et al. (2003)
represents the first published description of the application of DNA microarrays
to the functional analysis of D. radiodurans and suggests that the recovery
process for this organism involves the complicated coordination of DNA repair
and metabolic functions as well as other cellular functions.
3. Pyrococcus furiosus, a Hyperthermophilic Archaeon
P. furiosus is a member of a phylogenetically distinct group of prokaryotes

called the Archaea, which constitutes a primary, separate domain in the universal
tree of life (Woese et al., 1990; Olsen and Woese, 1997). The Archaea domain is
composed of organisms with diverse phenotypes, such as methane-producing
methanogens, extreme halophiles, and extremely thermophilic sulfur-metaboliz-
ing species (Woese, 1987). Typically, archaeal genes involved in energy
production, cell division, cell wall biosynthesis, and metabolism have homologs
in bacteria, whereas genes encoding proteins that function in the informational
processes of DNA replication, transcription, and translation are more similar to
their eucaryal counterparts (Bult et al., 1996). Archaea also share certain RNA
processing components with Eucarya, such as fibrillarin (a pre-rRNA processing
protein) (Bult et al., 1996; Belfort and Weiner, 1997) and tRNA splicing
endonucleases (Belfort and Weiner, 1997; Kleman-Leyer et al., 1997). The
mosaic nature of archaea makes this group of organisms extremely interesting
from an evolutionary perspective. The sequencing and analysis of archaeal
genomes should provide valuable insights into the origin or evolution of
eukaryotes, as well as the molecular mechanisms enabling their adaptation to
extreme environments.
The hyperthermophilic archaeon P. furiosus is able to grow optimally at a
temperature of 1008C (Fiala and Stetter, 1986). Studies support a highly regulated
fermentative-based metabolism in P. furiosus (Adams et al., 2001), which can
utilize the disaccharide maltose in the presence or absence of elemental sulfur
(S0). In addition, P. furiosus can couple the reduction of S0 to the oxidation of
catabolism-generated, reduced ferredoxin, but the molecular mechanism of this
metabolic coupling is not presently known (Schut et al., 2001). The availability
of the complete genome sequence of P. furiosus (Robb et al., 2001) permits the
global analysis of gene function and expression using high-density DNA
microarray technology. To investigate the molecular basis of S8 metabolism,
Schut et al. (2001) used DNA microarrays containing 271 ORFs (of the ca. 2200
total ORFs predicted) from the P. furiosus genome (1.9 Mb) to measure
Figure 7 Hierarchical clustering analyses of expression profile patterns. Gene expression

patterns are displayed graphically. Three distinct patterns are sorted according to the hierarchical
clustering analyses, i.e., (A) recA-like activation pattern, (B) growth-related activation pattern, and
(C) repressed patterns. The top row represents the general pattern of the selected group where a
Pearson correlation coefficient (r) is shown on the left side. All displayed graphs are organized in a
row/column format. Each row of colored strips represents a single gene whose expression levels are
color-recorded sequentially in every column of the same row that represents recovery time intervals.
Red color denotes up-regulation, whereas green indicates down-regulation. Black indicates the control
level. The variation in transcript abundance is positively correlated with the color darkness. (a) Gene
numbers are offered for tracking the primary information of the gene of interest. (b) The maximum
(for recA-like and growth-related activation pattern) or minimum (for the repressed pattern)
expression level for each of the exhibited genes over the 24-h recovery period is presented as the dye
intensity ratio of the irradiated sample to the non-irradiated control at (c) the indicated time interval.
Values in parentheses show the SD for each mean expression ratio (Courtesy of PNAS).
differential gene expression in cells grown at 958C on maltose in the presence or

absence of S8. The arrayed PCR products represented ORFs with proposed
functions in sugar and peptide catabolism, metal utilization, and the biosynthesis
of cofactors, amino acids, and nucleotides. This study by Schut et al. (2001)
represents the first and, to-date, only account describing the application of DNA
microarray analysis to a member of the Archaea. Currently published genomic
analyses of archaea have been almost exclusively limited to the sequencing,
annotation, and in silico comparative analysis of archaeal genomes.
DNA microarray analysis revealed a number of ORFs whose expression was
dramatically down-regulated (. 5-fold decrease) by S8, including 18 genes
encoding various subunits associated with three different hydrogenase systems
(Schut et al., 2001). Other genes displaying decreased transcription, when
P. furiosus cells were grown with S8, encoded a hypothetical protein and two
homologs (ornithine carbamoyltransferase and HypF) involved in hydrogenase
biosynthesis. In the presence of S8, the expression of two previously
uncharacterized ORFs (encoding products designated SipA and SipB for
“sulfur-induced proteins”) increased by a striking . 25-fold. The encoded
proteins of these ORFs were proposed by the authors to be part of a novel
S8-reducing, membrane-associated, iron-sulfur cluster-containing complex in
P. furiosus (Schut et al., 2001). The research reported by Schut et al. (2001)
clearly illustrates the power of DNA microarray analysis in generating new lines
of experimentation and in implicating previously uncharacterized ORFs identified
by genome sequencing in biological processes. There is little doubt that the
continuing determination of archaeal genomes will spawn more microarray-based
functional studies of extremophiles.
VIII. APPLICATION OF MICROARRAYS

TO ENVIRONMENTAL STUDIES
In addition to monitoring transcription patterns on a genomic scale,

microarray-based technology is well suited for detecting microorganisms in
natural environments. Many target functional genes involved in biogeochemical
cycling in environments are highly diverse, and it is difficult or impossible to
identify conserved regions for designing PCR primers or oligonucleotide probes.
The microarray-based approach does not require such sequence conservation,
because all of the diverse gene sequences from different populations of the same
functional group can be fabricated on arrays and used as probes to monitor their
corresponding populations.
In contrast to studies using pure cultures, microarray analysis of environmen-
tal nucleic acids presents a number of technical challenges that must be
overcome. First, target and probe sequences in environmental samples can be
very diverse, and it is not clear whether the performance of microarrays with
diverse environmental samples is similar to that with pure culture samples and
how sequence divergence is reflected in hybridization signal intensity. Second,
environmental samples are generally contaminated with substances such as
humic matter, organic contaminants, and metals, which may interfere with the
hybridization reaction on microarrays. Third, in contrast to pure cultures, the
recoverable biomass in environmental samples is generally low; consequently, it
is not clear whether microarray hybridization is sensitive enough to detect
microorganisms in all types of environmental samples. Finally, it is uncertain
whether microarray-based detection can be quantitative. Environmental and
ecological studies require experimental tools that not only detect the presence or
absence of particular groups of microorganisms but also provide quantitative data
on their in situ biological activities.
In the following sections, we discuss three different types of microarray
formats that have been developed for use in environmental studies: functional
gene arrays (FGAs), phylogenetic oligonucleotide arrays (POAs), and commu-
nity genome arrays (CGAs).
A. FUNCTIONAL GENE ARRAYS
Genes encoding functional enzymes involved in various biogeochemical

cycling processes (e.g., carbon, nitrogen, sulfate and metals) are very useful as
molecular signatures for assessing the physiological status and functional
activities of microbial populations and communities in natural environments.
Microarrays containing functional gene sequence information are referred to
as FGAs, because they are primarily used for the functional analysis of microbial
community activities in environments (Wu et al., 2001). Similar to gene
expression profiling arrays, both oligonucleotides and PCR-amplified
DNA fragments corresponding to functional genes can be used for fabricating
FGAs.
1. Selection of Gene Probes
FGAs are designed for studying functional gene diversity in natural

environments. To construct FGAs, the gene probes should be carefully defined
and selected based on the specific research questions to be addressed. As an
example, microarrays can consist of gene probes that are involved in such
biogeochemical processes as nitrification (ammonia monooxygenase, amoA),
denitrification (nitrite reductases, nirS and nirK), nitrogen fixation (nitro-
genases, nifH), sulfite reduction (sulfite reductase, dsvA/B), methanogenesis
(methyl coenzyme M reductase genes, mcrA), methane oxidation (methane

mono-oxygenases, mmo), and plant and fungal polymer degradation (cellulases,
xylanases, ligin peroxidases, chitinases).
Probes for the construction of FGAs can be generated in three ways. The desired
gene fragment can be amplified from genomic DNA extracted from pure bacterial
cultures using specific primers or from cloned plasmids containing the desired
gene insert using vector-specific primers. However, the availability of pure
cultures and plasmid clones can be limited. In the second approach, desired
gene fragments are recovered from natural environments using PCR-based
cloning methods (Zhou et al., 1997). Generally, sequences that show . 85%
identity can be used as specific probes for FGAs. These two approaches were used
to construct FGAs containing nitrite reductase genes and ammonia monooxygen-
ase genes for monitoring bacteria involved in nitrification and denitrification,
respectively (Wu et al., 2001). Finally, in the third strategy, oligonucleotides,
usually 50– 70-mers, are designed based on the functional sequences available in
public databases and synthesized for microarray fabrication (Tiquia et al.,
unpublished).
2. Specificity
Hybridization specificity is an important parameter that impacts any detection

method. It is influenced by many factors, such as G þ C content, degree of
sequence divergence, sequence length, secondary structure of the probe,
temperature, and salt concentrations. To determine the specificity of DNA
microarray hybridization, we have constructed and used FGAs consisting of
heme- and copper-containing nitrite reductase genes, ammonia monooxygenase
(amoA), and methane monooxygenase genes [ pmoA] (Wu et al., 2001). Small
subunit (SSU) rRNA genes and yeast genes were used as positive and negative
controls, respectively, on the FGAs. Cross-hybridization among different gene
groups was not observed at either low (458C) or high (658C) stringency.
Furthermore, no hybridization was observed with any of the five yeast genes,
which served as negative controls for hybridization on the array (Fig. 8A).
Based on the sequence similarities, it was estimated that microarray hybridi-
zation can differentiate between sequences exhibiting a dissimilarity of
approximately 15% at 658C and 10% at 758C (Wu et al., 2001). In addition, at
low stringency, most nirS, nirK or amoA genes hybridized well with their
respective homologous target DNA, suggesting that a broad range of detection
can be achieved by adjusting the conditions for microarray hybridization.
These results indicate that specific hybridization can be achieved using the
glass slide-based microarray format with bulk community DNA extracted from
environmental samples.
Figure 8 Specificity and sensitivity of DNA fragments-based FGAs. (A) Fluorescence images
showing the specificity of nirS in DNA microarray hybridization. Target DNA was labeled with either
Cy3 from a pure culture using the method of PCR amplification and hybridized separately at high
stringency (658C) to FGAs containing nirS, nir K, and amoA gene probes from both pure bacterial
cultures and environmental clones. The 16S rRNA and yeast genes served as positive and negative
controls, respectively. (B) Array hybridization images showing the detection sensitivity with labeled
pure genomic DNA Genomic DNA from a pure culture of nirS-containing P. stutzeri E4-2 was labeled
with Cy5 using the random primer labeling method. The target DNA was hybridized to the
microarrays at total concentrations of 0.5, 1, and 5 ng.
To determine the potential performance of oligonucleotide microarrays for

environmental studies, an FGA consisting of 50-mer oligonucleotide probes
was constructed and evaluated using 1033 genes involved in nitrogen cycling
(nirS, nir K, nif H, amoA, and pmoA) and sulfite reduction (dsrA/B) from
public databases and our own sequence collections (Tiquia et al, unpublished).
Under hybridization conditions of 508C and 50% formamide, genes having
, 86 –90% sequence identity were clearly distinguished. As expected, the
oligonucleotide-based FGAs showed a higher degree of hybridization
specificity than the DNA-based FGAs. Comparison of probe sequences from
pure cultures of bacteria involved in nitrification, denitrification, nitrogen
fixation, methane oxidation and sulfate reduction indicated that the average
similarity of these functional genes at the species level ranged from 74 to
84%. These results suggest that the 50-mer FGAs could provide species-level
resolution for analyzing microorganisms involved in these biogeochemical
processes.
Compared to DNA-based FGAs, the 50-mer oligonucleotide arrays offer the

following main advantages. First, higher hybridization specificity can be achi-
eved because the probe sizes in oligo arrays are much smaller than those used in
the DNA-based FGAs. Thus, this type of environmental array could provide a
higher level of resolution in differentiating microbial populations. Since the
probes can be directly designed and synthesized based on sequence information
from public databases, construction of oligonucleotide arrays is much easier than
that of the DNA-based FGAs. To construct microarrays containing large DNA
fragments, the probes used for array fabrication are generally amplified by PCR
from environmental clones or from pure genomic DNA. However, obtaining all
the diverse environmental clones and bacterial strains from various sources as
templates for amplification can be a considerable challenge. As a result, the
construction of comprehensive microarrays representing all functional genes of
interest is not practically feasible. With oligo arrays, a great number of genes can
easily be arrayed for comprehensive survey of the populations and activities of
diverse microbial communities in the environment. In addition, since no PCR
amplification is involved in oligonucleotide microarray fabrication, potential
cross-contamination due to PCR amplification is minimized.
3. Sensitivity
Sensitivity is another critical parameter that impacts the effectiveness of

microarray-based detection of microorganisms. The detection sensitivity of
hybridization with a prototype DNA-based FGA was determined using genomic
DNA from both pure cultures and soil community samples. At high stringency,
strong hybridization signals were observed with 5 ng of DNA for both nirS and
SSU rRNA genes, whereas hybridization signals were weaker but detectable with
1 ng of DNA (Fig. 8B). The hybridization signals at low DNA concentrations
were stronger for SSU rRNA genes than for nirS genes. Hybridization signals
derived from 0.5 ng of genomic DNA were measurable, but the fluorescence
intensity was poor. As a result, the detection limit was estimated to be
approximately 1 ng with randomly labeled pure genomic DNA under the tested
hybridization conditions.
The detection sensitivity of FGA hybridization was also evaluated using
community genomic DNA isolated from surface soil that contained a high level
of chromium and organic matter. All of the arrayed genes, with the exception of
the five yeast genes, showed hybridization with 50 and 25 ng of labeled
community DNA. Only the SSU rRNA genes could be detected when as little as
10 ng of the soil community DNA was used in the hybridization reaction. Thus,
the detection sensitivity of nirS and SSU rRNA genes in this soil sample was
considered to be approximately 25 and 10 ng of the total environmental DNA,
respectively. These approximate levels of detection sensitivity should be
sufficient for many studies in microbial ecology and suggest that microarray
hybridization can be used as a sensitive tool for analyzing microbial community
composition in environmental samples.
The detection limit with 50-mer FGAs was approximately 8 ng of pure
genomic DNA. As expected, the sensitivity of the 50-mer FGAs is 10 times lower
than the DNA-based FGAs and 100 times lower than CGAs (Tiquia et al.,
unpublished; Zhou, 2003), which are discussed below. One of the main reasons
for the lower sensitivity of the 50-mer FGAs is that the oligonucleotide probes are
much shorter than the probes used in DNA-based FGAs and CGAs, which have
more binding sites available for capturing the labeled target DNAs. In addition,
good hybridizations were obtained with the 50-mer FGAs using 2 mg of bulk
community DNA from marine sediments. These results suggest that the amount
of DNA sample should not be a major limiting factor in using this type of
microarray for environmental studies, because the average DNA yields generally
range from 10 to 400 mg of DNA per g (dry weight) for many surface soil and
sediment samples.
Although sensitive detection can be obtained with microarray hybridization,
the detection sensitivity is dependent on reagents, especially the fluorescent dyes.
We found that the sensitivity varies greatly with different batches of fluorescent
dyes. In addition, the sensitivity with direct microarray hybridization may still be
100 to 10,000-fold less than with PCR amplification. Microarray hybridization is
still not sensitive enough for some environmental studies where the amount of
recoverable biomass is very low, thus requiring the development of more
sensitive methods.
4. Quantitation
Many environmental and ecological studies require quantitative data on

the in situ abundance and biological activities of microbial communities. The
accuracy of microarray-based quantitative assessment is still uncertain because
of the inherently high variation associated with array fabrication, probe labeling,
hybridization, and image processing. Comparison of microarray hybridization
results with previously known results suggested that microarray hybridization
appears to be quantitative enough for detecting differences in gene expression
patterns under various conditions (DeRisi et al., 1997; Lockhart et al., 1996;
Taniguchi et al., 2001). DNA microarrays have also been used to measure
differences in DNA copy number in breast tumors (Pinkel et al., 1998; Pollack
et al., 1999) and to detect single-copy deletions or additions (Pollack et al., 1999),
suggesting that microarray-based detection is potentially quantitative. A recent
study in which lambda (l) DNA was co-spotted with DNA from reference
bacterial strains also indicated that microarrays could accurately quantify genes
in DNA samples (Cho and Tiedje, 2001).
To evaluate whether microarray hybridization can be used as a quantitative
tool to analyze environmental samples, the relationship between target DNA
concentration and hybridization signal was examined with DNA-based FGAs
(Wu et al., 2001). A strong linear relationship (r 2 ¼ 0.96) was observed between
signal intensity and target DNA concentration with DNA from a pure bacterial
culture within 1 to 100 ng. Similar to the DNA-based FGAs, a strong linear
relationship was observed using the 50-mer oligonucleotide FGAs between signal
intensity and target DNA concentrations from 8 to 1000 ng for all six different
functional gene groups (r 2 ¼ 0.96 – 0.98) (Tiquia et al., unpublished). These
results suggest that microarray hybridization is quantitative for pure bacterial
cultures within a limited range of DNA concentration. With our optimized
protocol, experimental variation between array slides can be reduced to below
15% with environmental samples (Wu et al., 2001). This is consistent with the
findings of microarray studies on gene expression (Bartosiewicz et al., 2000).
Since environmental samples contain a mixture of target and non-target
templates, the presence of other non-target templates could affect microarray-
based quantification. To determine whether microarray hybridization is quan-
titative for targeted templates within the context of environmental samples, 11
different genes, exhibiting less than 80% sequence identity, were labeled and
hybridized with the DNA-based FGAs. For this mixed DNA population, a linear
relationship (r 2 ¼ 0.94) was observed between signal intensity and target DNA
concentration (Fig. 9), further suggesting that microarray hybridization holds
promise as a quantitative tool for studies in environmental microbiology.
The target genes within functional groups present in environmental samples
may have different degrees of sequence divergence. Such sequence differences
will affect microarray hybridization signal intensities and hence its quantitative
power. Although it was shown that microarray hybridization could be used to
quantify mixed DNA templates, the difficult challenge in quantifying the
abundance of microbial populations in natural environments, based on
hybridization signal intensity, is how to distinguish differences in hybridization
intensity due to population abundance from those due to sequence divergence.
One possible solution is to carry out microarray hybridization under conditions of
varying stringency. Based on the relationships among signal intensity, sequence
divergence, hybridization temperature, and washing conditions, it should be
possible to distinguish, to some extent, the contributions of population abundance
and sequence divergence to hybridization intensity (Wu et al., 2001). For
instance, Wu et al. (2001) showed that at about 55 – 608C, sequence divergence
had little or no effect on signal intensity for amoA genes with greater than 80%
identity to the labeled target DNA. This suggests that under such hybridization
conditions the effect of sequence divergence on signal intensity is negligible for
genes with . 80% sequence identity; therefore, any significant differences in
Figure 9 Relationship of hybridization signal intensity to DNA target concentration using a

mixture of target DNAs. The PCR products from the following nine strains were mixed together in
different quantities (pg): E4-2 (nirS), 1000; G179 (nirK), 500; wc301-37 (amoA), 250; ps-47 (amoA),
125; pB49 (nirS), 62.5; Y32K (nirK), 31.3; wA15 (nirS), 15.6; ps-80 (amoA), 7.8; wB54 (nirK), 3.9.
All of these genes are less than 80% identical. The mixed templates were labeled with Cy5. The plot
shows the log-transformed average hybridization intensity versus the log-transformed target DNA
concentration for each strain. The target DNA was prepared by labeling MR-1 genomic DNA with
Cy5 using Klenow fragment with random hexamer primers. The data points are mean values derived
from three independent microarray slides, with three replicates on each slide (nine data points). Error
bars showing the SD are presented.
signal intensity are most likely due to differences in population abundance.

Another possible solution to this problem is to use microarrays containing probes
that are extremely specific to the target population of interest, such as those used
in oligonucleotide microarrays.
5. Applications
FGAs for microbial detection are still in the developmental stages, and thus
their applications are still being explored. To demonstrate the applicability of
DNA microarrays for microbial community analysis, Wu et al. (2001) used FGAs
to analyze the distribution of denitrifying and nitrifying microbial populations in
marine sediment and soil samples. The prototype FGA revealed differences in
the apparent distribution of nirS, nir K and amoA/pmoA gene families in

sediment and soil samples. Recently, a 70-mer oligonucleotide microarray
containing 64 nirS genes (14 from cultured microorganisms and 50 from
environmental clones) was evaluated for studying functional gene diversity in the
Choptank River-Chesapeake Bay system (Taroncher-Oldenburg et al., 2003).
Significant differences in the hybridization patterns were observed between the
sediment samples from two stations in the Choptank River. The changes in the
nirS-containing denitrifier population could have been caused by differences in
salinity, inorganic nitrogen and dissolved organic carbon between these two
stations.
So far, very limited studies have been carried out to evaluate specificity,
sensitivity, sequence divergence and quantitation of DNA microarrays for
environmental applications. While this tool is potentially valuable for
environmental studies, more development is needed, especially for improved
sensitivity, quantitation, and the biological meaning of a detectable specificity
before it can be used broadly and interpreted meaningfully within the context of
microbial ecology.
B. PHYLOGENETIC OLIGONUCLEOTIDE ARRAYS
Ribosomal RNA genes are powerful molecules for studying phylogenetic

relationships among different organisms and for analyzing microbial community
structure in natural environments, because these genes exist in all organisms and
contain both highly conserved and highly variable regions, which are useful for
differentiating microorganisms at different taxonomic levels (e.g., kingdom,
phyla, family, genus, species, and strain). A very large database of ribosomal
RNA genes exists (http://www.cme.msu.edu), making them ideal molecules for
developing microarray-based detection tools. In addition, cells generally have
multiple copies of rRNA genes, and the majority (. 95%) of total RNA isolated
from samples is rRNA. Consequently, the detection sensitivity will be higher for
rRNA genes than for functional genes. Therefore, rRNA genes are very useful
targets for developing microarray-based detection approaches.
Oligonucleotide microarrays containing information from rRNA genes are
referred to as phylogenetic oligonucleotide microarrays (POAs), because such
microarrays are used primarily for phylogenetic analysis of microbial commu-
nities. The POAs can be constructed for different phylogenetic taxa and used in
microbial community analysis studies. The oligonucleotide probes can be
designed in a phylogenetic framework to survey different levels of sequence
conservation, from highly conserved sequences giving broad taxonomic group-
ings to hypervariable sequences giving genus- and potentially species- level
groupings. Because highly conserved universal primers for amplifying rRNA
genes are available, POA-based hybridization can be easily coupled with PCR
amplification, thus enabling the implementation of highly sensitive assays.
1. Challenges of Plylogenetic Oligonucleotide Arrays
Non-rRNA gene-based oligonucleotide microarrays have been used success-

fully for monitoring genome-wide gene expression (e.g., Lockhart et al., 1996; de
Saizieu et al., 1998) and detecting genetic polymorphisms (e.g., Wang et al.,
1998). In contrast, rRNA gene-based oligonucleotide arrays present some unique
technical challenges concerning hybridization specificity and sensitivity (Zhou
and Thompson, 2002; Zhou, 2003).
Specificity. Since rRNA genes are highly conserved and present in all
microorganisms, specific detection with rRNA-targeted oligonucleotide micro-
arrays can be difficult. First, the probe length and G þ C content can significantly
impact microarray hybridization (Guschin et al., 1997a). Second, probe selection
is limited by the sequence differences among target genes, and cross-
hybridization can be a problem. Oligonucleotide microarrays typically contain
many probes. Ideally, all of the oligonucleotides should have similar or identical
melting kinetics, so that all of the probes on an array element can be subjected to
the same hybridization conditions at once. This can be difficult to achieve,
because the melting temperature depends on the length and composition of the
oligonucleotide probe as well as the target 16S rRNA molecules in the samples.
Secondary structure. The hybridization of oligonucleotide probes to target
nucleic acids possessing stable secondary structures can be particularly
challenging, since low stringency conditions (i.e., hybridization temperatures
between 0 –308C) are required for stable association of a long target nucleic acid
with a short immobilized oligonucleotide probe (Drobyshev et al., 1997; Guschin
et al., 1997a, b; Southern et al., 1999). Any stable secondary structure of the
target DNA or RNA must be overcome in order to make complementary
sequence regions available for duplex formation. The stable secondary structure
of SSU rRNA will have serious effects on hybridization specificity and detection
sensitivity.
2. Specificity and Sensitivity
In a study by Guschin et al. (1997a), gel-pad oligonucleotide microarrays were

constructed using oligonucleotides complementary to SSU rRNA sequences from
key genera of nitrifying bacteria. The results showed that specific detection could
be achieved with this type of microarray. However, the probe specificity depends
on various factors, such as probe length. Guschin et al. (1997a) showed that, as
the length of the oligonucleotide probe increases, mismatch discrimination is

lost; conversely, as the length of the probe decreases, hybridization signal
intensity (i.e., sensitivity) is sacrificed. A recent study showed that gel-pad-based
oligonucleotide microarrays could also be used to distinguish between
B. thuringiensis and B. subtilis (Bavykin et al., 2001). Using glass-based two-
dimensional microarrays, Small et al. (2001) detected such metal-reducing
bacteria as Geobacter chapellei and Desulfovibrio desulfuricans.
The potential advantage of oligonucleotide probes is that target sequences
containing single-base mismatches can be differentiated by microarray
hybridization. However, this has not been fully demonstrated with SSU rRNA
gene-based probes. To systematically determine whether single mismatch
discrimination can be achieved for SSU rRNA genes using microarray
hybridization, we constructed a model oligonucleotide microarray consisting of
probes derived from three different regions of the SSU rRNA molecule
corresponding to different bacterial taxa (X. Zhou and J. Zhou, unpublished
data). The probes had 1 – 5 mismatches in different combinations along the length
of the oligonucleotide probe with at least one mismatch at the central position.
Hybridization signal intensity with a single-base mismatch was decreased by 10
to 30%, depending on the type of mismatched nucleotide base. The signal
intensity of probes with two base mismatches was 5 to 25% of that of the perfect
match probes. Probes with three or four base-pair mismatches yielded signal
intensities that were 5% of that of the perfect match probes. Maximum
discrimination and signal intensity was achieved with 19-base probes. These
results indicated that single base discrimination for SSU rRNA genes can be
achieved with glass slide-based array hybridization, but complete discrimination
appears to be problematic with SSU rRNA genes (Bavykin et al., 2001; Small
et al., 2001; Urakawa et al., 2002). Urakawa et al. (2002) demonstrated that the
single-base-pair near-terminal and terminal mismatches have a significant effect
on hybridization signal intensity. With SSU rRNA gene-based oligonucleotide
microarrays, the level of detection sensitivity obtained using the G. chapellei 16S
rRNA gene is about 0.5 mg of total RNA extracted from soils (Small et al., 2001).
3. Applications
As with all the arrays developed for environmental applications, SSU rRNA
gene-based oligonucleotide arrays are still in the early stages of development, and
therefore, only a few studies have applied POAs to the analysis of microbial
structure within the context of environmental samples. Using photolithography-
based Affymetrix technology, Wilson et al. (2002) designed a gene chip
(microarray) containing 31,179 and 20-mer oligonucleotide probes specific for
SSU rRNA genes. All of the probes are derived from a small SSU rRNA gene
region (i.e., E. coli positions 1409 to 1491), which is bound on both ends by
universally conserved segments. The gene chip also contained control sequences,
which were paired with the probe sequences. A control sequence was identical to
the paired probe sequence except that there was a mismatch nucleotide at the 11th
position. Thus, the gene chip contained a total of 62,358 features. The number of
probes for individual sequences contained in the Ribosomal Database Project
(RDP version 5.0, with about 3200 sequences) ranges from 0 to 70. A total of 17
pure bacterial cultures were used to assess the performance of this gene chip, and
15 bacterial species were identified correctly. However, it failed to resolve the
individual sequences comprising complex mixed samples (Wilson et al., 2002).
Rudi et al. (2000) constructed a small microarray containing 10 SSU rRNA
probes derived from cyanobacteria, and used it to analyze the presence and
abundance of these organisms in lakes with both low and high biomass. The
probes were specific to the cultures analyzed, and reproducible abundance
profiles were obtained with these lake samples. Relatively good qualitative
correlations were observed between the community diversity and standard
hydrochemical data, but the levels of correlation were lower for the quantitative
data.
Loy et al. (2002) developed a microarray containing 132 SSU rRNA-targeted
oligonucleotide probes, which represented all recognized groups of sulfate-
reducing prokaryotes. Microarray hybridizations with 41 reference strains
showed that, under the hybridization conditions used, clear discrimination
between perfectly matched and mismatched probes were obtained for most, but
not all of the 132 probes. This microarray was used to determine the diversity of
sulfate-reducing prokaryotes in periodontal tooth pockets and a hypersaline
cyanobacterial mat. The microarray hybridization results were consistent
with those obtained using well-established conventional molecular methods.
These results suggest that microarray hybridization is a powerful tool in
analyzing community structure but great caution is needed in data interpretation
because of the potential for cross-hybridization.
C. COMMUNITY GENOME ARRAYS
Decades of scientific investigations have led to the isolation of many

microorganisms from a variety of natural habitats. However, little or nothing is
known about the genomic sequences for the majority of these microorganisms.
Such a large collection of pure cultures should be very useful for monitoring
microbial community structure and composition in natural environments. To
exploit such a resource, a novel prototype microarray containing whole genomic
DNA, termed community genome array (CGA), was developed and evaluated in
my (Zhou’s) laboratory.
The CGA is conceptually analogous to membrane-based reverse sample

genome probing (RSGP) (Voordouw, 1998), but CGA hybridization is distinctly
different from RSGP in terms of the arraying substrate and signal detection
strategies. In contrast to RSGP, the CGA uses a non-porous (i.e., glass) surface
for fabrication and fluorescence-based detection. The capability of accurate and
precise miniaturization with robots on non-porous substrates is one of the two key
advances of microarray-based genomic technologies. The miniaturized micro-
array format coupled with fluorescent detection represents a fundamental
revolution in biological analysis. Like RSGP, the main disadvantage of the CGA
is that only the cultured components of a community can be monitored, because
the construction requires the availability of individual pure isolates, even though
CGA-based hybridization itself does not require culturing (Voordouw, 1998).
With the recent advances in environmental genomics, high-molecular-weight
DNA from uncultivated microorganisms could be accessed through bacterial
artificial chromosomes (BACs). BAC clones could also be used to fabricate
CGAs, thus allowing the investigation of uncultivated components of a complex
microbial community. In the following sections, we will briefly describe the
performance of CGA-based hybridization in terms of specificity, sensitivity and
quantitation.
1. Specificity
To examine hybridization specificity under varying experimental conditions

and to determine the threshold levels of genomic differentiation, a prototype
microarray was fabricated that contained genomic DNA isolated from 67
different representative environmental microorganisms classified as a-, b-, and
g-proteobacteria and Gram-positive bacteria. Many of the selected species are
closely related to each other based on SSU rRNA and gyr B gene phylogenies and
belong primarily to three major bacterial genera (Pseudomonas, Shewanella, and
Azoarcus). The G þ C content of the genomes varies from 37 to 69.3%. By
adjusting hybridization temperature and the concentration of additives such as
formamide (which increases hybridization stringency), different threshold levels
of phylogenetic differentiation could be achieved using the CGAs. For instance,
under hybridization conditions of 558C and 50% formamide, strong signals were
obtained for genomic DNAs of corresponding species to the labeled target. Little
or no cross-hybridization (~0 –4%) was observed for non-target species as well as
for negative controls (yeast genes), thus indicating that species-specific
differentiation can be achieved with CGAs under the hybridization conditions
used. However, different strains of Pseuodomonas stutzeri, Azoarcus tolulyticus,
Bacillus methanolicus, and Shewanella algae could not be clearly distinguished
under these conditions (Wu et al., unpublished). By further increasing
hybridization temperature (65 and 758C), strain-level differentiation was

obtained for closely related Azoarcus strains (Wu et al., unpublished).
Due to the complicated nature of microarray hybridization, it is unlikely that
such assays will completely eliminate some degree of hybridization to non-target
strains. The central question is how to distinguish true hybridization signals from
non-specific background noise. One common approach is to determine SNRs and
discard values below certain threshold value. Our studies showed that the average
SNR for hybridizations with different species within a genus is about
3.35 ^ 0.32, which is substantially lower than hybridizations with different
strains from the same species. This value is very close to the commonly used
threshold value (SNR ¼ 3.0).
CGAs could be used to determine the genetic distance between different
bacteria at the taxonomic levels of species and strain. Significant linear
relationships were observed between CGA hybridization ratios and sequence
similarity values derived from SSU rRNA and gyr B genes, DNA-DNA
reassociation, or REP- and BOX-PCR fingerprinting profiles (r 2 ¼ 0.80 2 0.95)
(Wu et al., unpublished), suggesting that CGAs could provide meaningful insights
into relationships between closely related strains. Because of its high capacity, one
can construct CGAs containing bacterial type strains plus appropriately related
strains. By hybridizing genomic DNA from unknown strains with this type of
microarray, one should be able to quickly and reliably identify unknown strains
provided a suitably related probe is on the array. When using CGAs for strain
identification, less stringent hybridization conditions (e.g., 458C and 50%
formamide) should be used first to ensure that good hybridization signals can be
obtained for distantly related target species. If multiple probes have significant
hybridization with the unknown target strains, highly stringent hybridization
conditions should then be used.
Compared to the traditional DNA –DNA reassociation approach, CGAs have
several advantages for determining species relatedness. Since many bacterial
genomes can be deposited on microarray slides, the tedious and laborious pair-
wise hybridizations associated with the traditional DNA – DNA reassociation
approach among different species are not needed with CGAs. In contrast to the
traditional DNA – DNA reassociation approach, which generally requires about
100 mg DNA, CGA-based hybridization requires only about 2 mg of genomic
DNA. This is important for determining the relationships between bacterial
species that are recalcitrant to cultivation or grow very slowly.
2. Sensitivity and Quantitative Potential
To determine the detection sensitivity of CGAs, genomic DNA from a pure

bacterial culture was fluorescently labeled and hybridized with the CGA at
different concentrations. Under stringent hybridization conditions (i.e., 658C), the

detection limit with randomly labeled pure genomic DNA was estimated to be
approximately 0.2 ng, whereas genomic DNA concentrations of 0.1 ng were
barely detectable above background levels (Wul et al., unpublished). The level of
CGA detection sensitivity should be sufficient for many studies in microbial
ecology. The detection sensitivity was approximately 10-fold higher than that of
DNA-based FGAs and about 100 times higher than that of the 50-mer FGAs.
These results were expected, because the CGA probes represent entire genomes
rather than a single gene.
The capacity of CGA hybridization to serve as a quantitative tool was
explored by examining the relationship between the concentration of labeled
target DNA and hybridization signal intensity. Quantitative potential was
determined using labeled genomic DNA from a single pure culture and from 16
targeted bacteria representing different genera and species. In both cases, strong
linear relationships between fluorescence intensity and DNA concentration were
observed within a certain range of concentrations (r 2 ¼ 0.92 2 0.95) (Wu et al.,
unpublished). The results indicate that CGAs can be used for quantitative
analysis of microorganisms in environmental samples. The quantitative feature of
CGA is similar to those of the DNA- and oligonucleotides-based FGAs (Wu et al.,
2001; Tiquia et al., unpublished).
D. WHOLE-GENOME OPEN READING FRAME ARRAYS FOR

REVEALING GENOME DIFFERENCES AND RELATEDNESS
Many microorganisms that are closely related based on SSU rRNA gene
sequences show dramatic differences in phenotypic characteristics. One way to
understand the genetic basis for such phenotypic differences is to obtain whole-
genome sequence information for all closely related species of interest. Patterns
of sequence similarity and variability will provide insights on the conservation of
gene functions, physiological plasticity and evolutionary processes. However,
sequencing the entire genomes of all closely related species is expensive and
time-consuming. In addition, it may not be necessary to sequence all closely
related genomes once the complete genome sequence for one representative
microorganism is available, because substantial portions of the genomic
sequence will be common among closely related species. One way to circumvent
the need for sequencing multiple genomes of closely related species is to
use DNA microarrays containing individual ORFs of a sequenced microorganism
to view genome diversity and relatedness of other closely related
microorganisms.
The whole-genome ORF array-based hybridization approach has been used
to reveal genome diversity and relatedness among closely related organisms in
several studies. Murray et al. (2001) used this approach to evaluate the
genome diversity and relatedness of several related metal-reducing bacteria
within the Shewanella genus using partial ORF microarrays for the sequenced
metal-reducing bacterium, S. oneidensis MR-1. Both conserved and poorly
conserved genes were identified among the nine species tested. Under the
conditions used in this study, the hybridization results were most informative
for the closely related organisms with SSU rRNA sequence similarities greater
than 93% and gyrB sequence similarities greater than 80%. Above this level
of homology, the similarities of microarray hybridization profiles were
strongly correlated with gyrB sequence divergence. In addition, most genes
in operons had high levels of DNA relatedness, suggesting that this approach
can be used to identify genes or operons that were horizontally transferred
(Murray et al., 2001).
Using the ORF arrays for E. coli K-12, Dong et al. (2001) identified the genes
in a common endophyte of maize, Klebsiella pneumoniae 342, which is closely
related to E. coli. About 3000 (70%) of E. coli genes were found in strain 342
with greater than 55% identity, whereas about 24% of the E. coli genes were
absent in strain 342. The genes with high sequence identity were those involved
in cell division, DNA replication, transcription, translation, transport, regulatory
proteins, energy, amino acid and fatty acid metabolism, and cofactor synthesis,
whereas the genes that are less conserved were involved in carbon compound
metabolism, membrane proteins, structural proteins, central intermediary
metabolism, and proteins involved in adaptation and protection. Genes that
were not identified in strain 342 included putative regulatory proteins, putative
chaperones, surface structure proteins, mobility proteins, putative enzymes and
hypothetical proteins. These results on genomic diversity are consistent with the
physiological properties of these two strains, suggesting that the microarray-
based whole-genome comparison is a powerful approach to revealing the
genomic diversity and relatedness of closely related organisms.
The whole-genome ORF array approach was also successfully used to
identify genome differences among 15 Helicobacter pylori strains with more and
less virulence (Salama et al., 2000) and to detect the deletions existing in other
strains of Mycobacterium tuberculosis and M. bovis (Behr et al., 1999). All of
these studies suggest that whole-genome ORF arrays will be useful for revealing
genome difference and relatedness. Whole-genome ORF arrays are available
from many microorganisms and they will be valuable for studying genome
diversity and relatedness of closely related microorganisms. For example, the
whole-genome arrays for six environmentally important microorganisms,
including S. oneidensis MR-1, D. radiodurans R1, Rhodopseudomonas palustris,
Nitrosomonas europaea, Desulfovibrio vulgaris, and Geobacter metallireducens
are available at Oak Ridge National Laboratory, and we are also currently using
these whole-genome ORF arrays to understand the genome diversity and
relatedness of some important environmental isolates.
E. OTHER TYPES OF MICROARRAYS FOR MICROBIAL

DETECTION AND CHARACTERIZATION
DNA microarrays containing random genomic fragments have been used to

determine species relatedness in instances where genome sequence information
is not available. In this approach, 60– 96 genomic fragments of about 1 kb were
randomly selected from four fluorescent Pseudomonas species as reference
genomes for microarray fabrication (Cho and Tiedje 2001). Cluster analysis of
hybridization profiles from 12 well-characterized fluorescent Pseudomonas
species indicated that such types of microarray hybridization could provide
species to strain level resolution. This approach could have higher resolution than
CGA because extensive component information is obtained rather than an
average for the whole-genome. However, this approach is more time-consuming
and costly to develop than CGA and such an array would be more limited in
scope since many of the array positions would be used for each reference
microorganism (L. Wu, personal communication).
Recently, a random nonamer oligonucleotide microarray was developed and
evaluated for obtaining fingerprinting profiles among closely related strains
instead of using a gel electrophoresis-based method (Kingsley et al., 2002). A
prototype array containing 47 randomly selected nonamer oligonucleotides was
constructed and used to differentiate 14 closely related Xanthomonas strains. The
REP-PCR was first carried out to obtain the fingerprints from different strains,
then the amplified REP-PCR products were hybridized with the nonamer array,
and fingerprinting profiles for each strain were obtained based on microarray
hybridization. The results showed that the microarray-based fingerprinting
methods provide clear resolution among all strains examined, including two
strains (X. oryzae 43836 and 49072) which could not be resolved using traditional
gel electrophoresis of REP-PCR amplification methods. This suggests that the
microarray hybridization-based approach could provide higher resolution in
strain differentiation than the conventional gel electrophoresis-based fingerprint-
ing approach. This approach is attractive because a universal nonamer array can
be developed to generate fingerprints from any microorganisms.
IX. CONCLUDING REMARKS
Microarray is a recently developed functional genomics technology that has

powerful applications in a wide array of biological research areas, including the
medical sciences, agriculture, biotechnology and environmental studies. Since
many universities, research institutions and industries have established micro-
array-based core facilities and services, microarrays have become a readily
accessible, widely used technology for investigating biological systems. As the
technologies for array instrumentation are relatively mature, major trends are
emerging in such issues as novel array platforms, attachment strategies and
substrates, miniaturization with higher density, novel labeling strategies, scanning
technologies and automation (Constans, 2003; Stears et al., 2003). Besides the
DNA-based array assay, the microarray platform is also being rapidly expanded to
include the analysis of other biomolecules such as proteins and carbohydrates
(Stears et al., 2003). Along with exploration in microarray technology appli-
cations, novel strategies and approaches for experimental controls and design are
needed to ensure that microarray hybridization data from different samples are
comparable, interpretable and biologically significant because of the inherent
variability in microarray hybridization signals. Finally, more advanced automatic
mathematical and computational tools, such as multivariate analysis, time-series
analysis, neural network, artificial intelligence, and differential equation-based
modeling approaches, should be extremely useful for rapid pattern recognition,
visualization, data mining, cellular modeling, simulation and prediction.
The development and application of microarray-based genomic technology for
environmental studies has received a great deal of attention. Because of its high-
density and high-throughput capacity, it is expected that microarray-based
genomic technologies will revolutionize the analyses of microbial community
structure, function and dynamics. Microarray-based assays have great potential as
specific, sensitive, quantitative, parallel, and high-throughput tools for microbial
detection, identification and characterization in natural environments. However,
more rigorous and systematic assessment and development are needed to realize
the full potential of microarrays for microbial ecology studies. Several key issues
need to be addressed, including novel experimental designs and strategies for
minimizing inherent high hybridization variations to improve microarray-based
quantitative accuracy, novel approaches for increasing hybridization sensitivity to
detect extremely low biomass in natural environments, novel computational tools
for microarray data extraction and interpretation, and broad integration and
application of microarray technologies with environmental studies to address
ecological and environmental questions and hypotheses.
REFERENCES
Adams, M. W. W., Holden, J. F., Menon, A. L., Schut, G. J., Grunden, A. M., Hou, C., Hutchins, A. M.,
Jenney, F. E. Jr., Kim, C., Ma, K., Pan, G., Roy, R., Sapra, R., Story, S. V., and Verhagen, M. F.
(2001). Key role for sulfur in peptide metabolism and in regulation of three hydrogenases in the
hyperthermophilic archaeon Pyrococcus furiosus. J. Bacteriol. 183, 716–724.
Adler, K., Broadbent, J., Garlick, R., Joseph, R., Khimani, A., Mikulskis, A., Rapiejko, P., and Killian,
J. (2000). MICROMAXe: A highly sensitive system for differential gene expression on
microarrays. “In Microarray Biochip Technology” (M. Schena, Ed.), pp. 221–230. Eaton
Publishing, Natick, MA.
Afanassiev, V., Hanemann, V., and Wolfl, S. (2000). Preparation of DNA and protein micro arrays on
glass slides coated with an agarose film. Nucl. Acids Res. 28, E66.
Anderson, A., Nordan, H., Cain, R., Parrish, G., and Duggan, D. (1956). Studies on a radioresistant
micrococcus, I. Isolation, morphology, cultural characteristics, and resistance to gamma
radiation. Food Technol. 10, 575–578.
Augenlicht, L., Taylor, J., Anderson, L., and Lipkin, M. (1991). Patterns of gene expression that
characterize the colonic mucosa in patients at genetic risk for colonic cancer. Proc. Nat. Acad.
Sci. 88, 3286–3289.
Augenlicht, L., Wahrman, M., Halsey, H., Anderson, L., Taylor, J., and Lipkin, M. (1987). Expression
of cloned sequences in biopsies of human colonic tissue and in colonic–carcinoma cells induced
to differentiate invitro. Cancer Res. 47, 6017–6021.
Bains, W., and Smith, G. (1988). A novel method for nucleic acid sequence determination. J. Theoret.
Biol. 135, 303–307.
Baldi, P., and Long, A. D. (2001). A Bayesian framework for the analysis of microarray expression
data: regularized t-test and statistical inferences of gene changes. Bioinformatics 17, 509–519.
Bartosiewicz, M., Trounstine, M., Barker, D., Johnston, R., and Buckpitt, A. (2000). Development of a
toxicological gene array and quantitative assessment of this technology. Arch. Biochem. Biophys.
376, 66 –73.
Basarsky, T., Verdnik, D., Zhai, J. Y., and Wellis, D. (2000). Overview of a microarray scanner:
Design essentials for an integrated acquisition and analysis platform. In “Microarray Biochip
Technology” (M. Schena, Ed.), pp. 265 –284. Eaton Publishing, Natick, MA.
Bassett, D. Jr., Eisen, M., and Boguski, M. (1999). Gene expression informatics — it’s all in your
mine. Nat. Genet. Sup. 21, 51–55.
Battaglia, C., Salani, G., Consolandi, C., Bernardi, L. R., and De, Bellis, G. (2000). Analysis of DNA
microarrays by non-destructive fluorescent staining using SYBR green II. Biotechniques 29,
78–81.
Battista, J. R. (1997). Against all odds: the survival strategies of Deinococcus radiodurans. Annu. Rev.
Microbiol. 51, 203–224.
Battista, J. R., Earl, A. M., and Park, M.-J. (1999). Why is Deinococcus radiodurans so resistant to
ionizing radiation? Trends Microbiol. 7, 362–365.
Bavykin, S. G., Akowski, J. P., Zakhariev, V. M., Barsky, V. E., Perov, A. N., and Mirzabekov, A. D.
(2001). Portable system for microbial sample preparation and oligonucleotide microarray
analysis. Appl. Environ. Microbiol. 67, 922 –928.
Beattie, K. L., Eggers, M. D., Shumaker, J. M., Hogan, M. E., Varma, R. S., Lamture, J. B., Hollis,
M. A., Ehrlich, D. J., and Rathman, D. (1992). Genosensor technology. Clin. Chem. 39, 719– 722.
Beattie, W. G., Meng, L., Turner, S., Varma, R. S., Dao, D. D., and Beattie, K. L. (1995). Hybridization
of DNA targets to glass-tethered oligonucleotide probes. Mol. Biotechnol. 4, 213 –225.
Behr, M. A., Wilson, M. A., Gill, W. P., Salamon, H., Schoolnik, G. K., Rane, S., and Small, P. M.
(1999). Comparative genomics of BCG vaccines by whole-genome DNA microarray. Science
284, 1520–1523.
Beier, M., and Hoheisel, J. (1999). Versatile derivatisation of solid support media for covalent
bonding on DNA-microchips. Nucl. Acids Res. 27, 1970–1977.
Belfort, M., and Weiner, A. (1997). Another bridge between kingdoms: tRNA splicing in Archaea and
Eukaryotes. Cell 89, 1003– 1006.
Beliaev, A. S., Thompson, D. K., Fields, M. W., Wu, L., Lies, D. P., Nealson, K. H., and Zhou, J.
(2002a). Microarray transcription profiling of a Shewanella oneidensis etrA mutant. J. Bacteriol.
184, 4612–4616.
Beliaev, A. S., Thompson, D. K., Khare, T., Lim, H., Brandt, C. C., Li, G., Murray, A. E., Heidelberg,
J. F., Giometti, C. S., Yates, III, J., Nealson, K. H., Tiedje, T. M., and Zhou, J. (2002b). Gene and
protein expression profiles of Shewanella oneidensis during anaerobic growth with different
electron acceptors. OMICS 6, 39– 60.
Ben-Dor, A., Shamir, R., and Yakhini, Z. (1999). Clustering gene expression patterns. J. Comput.
Biol. 6, 281–297.
Broude, N. E., Woodward, K., Cavallo, R., Cantor, C. R., and Englert, D. (2001). DNA microarrays
with stem-loop DNA probes: preparation and applications. Nucl. Acids Res. 29, E92.
Bult, C. J., White, O., Olsen, G. J., Zhou, L. X., Fleischmann, R. D., Sutton, G. G., Blake, J. A.,
FitzGerald, M., Clayton, R. A., Gocayne, J. D., Kerlavage, A. R., Dougherty, B. A., Tomb, J. F.,
Adams, M. D., Reich, C. I., Overbeek, R., Kirkness, E. F., Weinstock, K. G., Merrick, J. M.,
Glodek, A., Scott, J. L., Geoghagen, N. S. M., Weidman, J. F., Fuhrmann, J. L., Nguyen, D.,
Utterback, T. R., Kelley, J. M., Peterson, J. D., Sadow, P. W., Hanna, M. C., Cotton, M. D.,
Roberts, K. M., Hurst, M. A., Kaine, B. P., Borodovsky, M., Klenk, H. P., Fraser, C. M., Smith,
H. O., Woese, C. R., and Venter, J. C. (1996). Complete genome sequence of the methanogenic
archaeon, Methanococcus jannaschii. Science 273, 1058–1073.
Chee, M., Yang, R., Hubbell, E., Berno, A., Huange, X. C., Stern, D., Winkler, J., Lockhart, D. J.,
Morris, M. S., and Fodor, S. P. A. (1996). Accessing genetic information with high-density DNA
arrays. Science 274, 610 –614.
Chen, Y., Dougherty, E. R., and Bittner, M. L. (1997). Ratio-based decisions and the quantitative
analysis of cDNA microarray images. J. Biomed. Optics. 2, 364–374.
Cho, J. C., and Tiedje, J. M. (2001). Bacterial species determination from DNA –DNA hybridi-
zation by using genome fragments and DNA microarrays. Appl. Environ. Microbiol. 67,
3677–3682.
Chrisey, L. A., Lee, G. U., and O’Ferrall, C. E. (1996). Covalent attachment of synthetic DNA to
self-assembled monolayer films. Nucl. Acids Res. 24, 3031–3039.
Clewell, D. R. (1993). Sex pheromones and the plasmid encoded mating response in Enterococcus
faecalis. In “Bacterial conjugation” (D. B. Clewell, Ed.), pp. 349–367. Plenum Press, New York,
NY.
Constans, A. (2003). Microarray instrumentation. The Scientist 17, 37–38.
Daly, M. J., Ling, O., and Minton, K. W. (1994). Interplasmidic recombination following irradiation
of the radioresistant bacterium Deinococcus radiodurans. J. Bacteriol. 176, 7506–7515.
Daly, M. J., and Minton, K. W. (1995). Interchromosomal recombination in the extremely
radioresistant bacterium Deinococcus radiodurans. J. Bacteriol. 177, 5495–5505.
Dassy, B., and Fournier, J. M. (1996). Respiratory activity is essential for post-exponential-phase
production of type 5 capsular polysaccharide by Staphylococcus aureus. Infect. Immun. 64,
2408–2414.
de Saizieu, A., Certa, U., Warrington, J., Gray, C., Keck, W., and Mous, J. (1998). Bacterial transcript
imaging by hybridization of total RNA to oligonucleotide arrays. Nat. Biotechnol. 16, 45–48.
DeRisi, J. L., Iyer, V. R., and Brown, P. O. (1997). Exploring the metabolic and genetic control of
gene expression on a genomic scale. Science 278, 680 –686.
Diehl, F., Grahlmann, S., Beier, M., and Hoheisel, J. D. (2001). Manufacturing DNA microarrays of
high spot homogeneity and reduced background signal. Nucleic Acids Res. 29, E38.
Dong, S., Wang, E., Hsie, L., Cao, Y., Chen, X., and Gingeras, T. R. (2001). Flexible use of high-
density oligonucleotide arrays for single-nucleotide polymorphism discovery and validation.
Genome Res. 11, 1418–1424.
Dopazo, J., and Carazo, J. M. (1997). Phylogenetic reconstruction using an unsupervised growing
neural network that adopts the topology of a phylogenetic tree. J. Molecul. Evol. 44, 226 –233.
Dopazo, J., Zanders, E., Dragoni, I., Amphlett, G., and Falciani, F. (2001). Methods and approaches in
the analysis of gene expression data. J. Immunol. Meth. 250, 93– 112.
Drmanac, R., Labat, I., Brukner, I., and Crkvenjakov, R. (1989). Sequencing of megabase-plus DNA
by hybridization: Theory of the method. Genomics 4, 114–128.
Drobyshev, A., Mologina, N., Shik, V., Pobedimskaya, D., Yershov, G., and Mirzabekov, A. (1997).
Sequence analysis by hybridization with oligonucleotide microchip: identification of beta-
thalassemia mutations. Gene 188, 45–52.
Drobyshev, A. L., Zasedatelev, A. S., Yershov, G. M., and Mirzabekov, A. D. (1999). Massive parallel
analysis of DNA-Hoechst 33258 binding specificity with a generic oligodeoxyribonucleotide
microchip. Nucl. Acids Res. 27, 4100–4105.
Dudoit, S. Y., Yang, H., Gallow, M. J., and Speed, T. P. (2001). Statistical methods for identifying
genes with differential expression in replicated cDNA microarray experiments. Statist Sincia 12,
111–139.
Duggan, D. J., Bittner, M., Chen, Y., Meltzer, P., and Trent, J. M. (1999). Expression profiling using
cDNA microarrays. Nat. Genet. 21, 10–14.
Dunman, P. M., Murphy, E., Haney, S., Palacios, D., Tucker-Kellogg, G., Wu, S., Brown, E. L.,
Zagursky, R. J., Shlaes, D., and Projan, S. J. (2001). Transcription profiling-based identification of
Staphylococcus aureus genes regulated by the agr and/or sarA loci. J. Bacteriol. 183, 7341– 7353.
Eckmann, L., Smith, J. R., Housley, M. P., Dwinell, M. B., and Kagnoff, M. F. (2000). Analysis of
high density cDNA arrays of altered gene expression in human intestinal epithelial cells in
response to infection with the invasive enteric bacteria Salmonella. J. Biol. Chem. 275,
14084– 14094.
Eggers, M. D., Hogan, M. E, Reich, R. K., Lamture, J., Ehrlich, D., Hollis, M., Kosicki, B., Powdrill,
T., Beattie, K., Smith, S., Varma, R., Gangadharan, R., Mallik, A., Burke, B., and Wallace, D.
(1994). A microchip for quantitative detection of molecular utilizing luminescent and
radioisotope reporter groups. Biotechniques. 14, 516–525.
Eisen, M., and Brown, P. (1999). DNA microarrays for analysis of gene expression. Meth. Enzymol.
303, 179 –205.
Englert, D. (2000). Production of microarrays on porous substrates using noncontact piezoelectric
dispensing. In “Microarray Biochip Technology” (M. Schena, Ed.), pp. 231–246. Eaton
Evertsz, E., Starink, P., Gupta, R., and Watson, D. (2000). Technology and applications of gene
expression microarrays. In “Microarray Biochip Technology” (M. Schena, Ed.), pp. 149– 166.
Eaton Publishing, Natick, MA.
Fiala, G., and Stetter, K. O. (1986). Pyrococcus furiosus sp. nov. represents a novel genus of marine
heterotrophic archaebacteria growing optimally at 1008C. Arch. Microbiol. 145, 56–61.
Fitz-Gibbon, S. T., Ladner, H., Kim, U.-J., Stetter, K. O., Simon, M. I., and Miller, J. H. (2002).
Genome sequence of the hyperthermophilic crenarchaeon Pyrobaculum aerophilum. Proc. Natl.
Acad. Sci. USA 99, 984 –989.
Fodor, S. P. A., Read, J., Pirrung, M., Stryer, L., Lu, A., and Solas, D. (1991). Light-directed, spatially
addressable parallel chemical synthesis. Science 251, 767–773.
Fotin, A., Drobyshev, A., Proudnikov, D., Perov, A., and Mirzabekov, A. (1998). Parallel
thermodynamic analysis of duplexes on oligodeoxyribonucleotide microchips. Nucl. Acids Res.
26, 1515–1521.
Fredrickson, J. K., Kostandarithes, H. M., Li, S. W., Plymale, A. E., and Daly, M. J. (2000). Reduction
of Fe(III), Cr(VI), U(VI), and Tc(VII) by Deinococcus radiodurans R1. Appl. Environ.
Microbiol. 66, 2006–2011.
Galagan, J. E., Nusbaum, C., Roy, A., Endrizzi, M. G., Macdonald, P., FitzHugh, W., Calvo, S., Engels,
R., Smirnov, S., Atnoor, D., Brown, A., Allen, N., Naylor, J., Stange-Thomann, N., DeArellano, K.,
Johnson, R., Linton, L., McEwan, P., McKernan, K., Talamas, J., Tirrell, A., Ye, W. J., Zimmer, A.,
Barber, R. D, Cann, I., Graham, D. E., Grahame, D. A., Guss, A. M., Hedderich, R., Ingram-Smith,
C., Kuettner, H. C., Krzycki, J. A., Leigh, J. A., Li, W. X., Liu, J. F., Mukhopadhyay, B., Reeve,
J. N., Smith, K., Springer, T. A., Umayam, L. A., White, O., White, R. H., de, Macario, E. C., Ferry,
J. G., Jarrell, K. F., Jing, H., Macario, A. J. L., Paulsen, I., Pritchett, M., Sowers, K. R., Swanson,
R. V., Zinder, S. H., Lander, E., Metcalf, W. W., and Birren, B. (2002). The genome of
M. acetivorans reveals extensive metabolic and physiological diversity. Genome Res. 12,
532–542.
Guo, Z., Guilfoyle, R. A., Thiel, A. J., Wang, R., and Smith, L. M. (1994). Direct fluorescence analysis
of genetic polymorphisms by hybridization with oligonucleotide arrays on glass supports. Nucl.
Acids Res. 22, 5456–5465.
Guschin, D. Y., Mobarry, B. K., Proudnikov, D., Stahl, D. A., Rittmann, B. E., and Mirzabekov, A.
(1997a). Oligonuclotide microchips as genosensors for determinative and environmental studies
in microbiology. Appl. Environ. Microbiol. 63, 2397–2402.
Guschin, D., Yershov, G., Zaslavsky, A., Gemmell, A., Shick, V., Proudnikov, D., Arenkov, P., and
Mirzabekov, A. (1997b). Manual manufacturing of oligonucleotide, DNA, and protein
microchips. Anal. Biochem. 250, 203 –11.
Hacia, J. G. (1999). Resequencing and mutational analysis using oligonucleotide microarrays. Nat.
Genet. 21, 42–47.
Hegde, P., Qi, R., Abernathy, K., Gay, C., Dharap, S., Gaspard, R., Hughes, J. E., Snesrud, E., Lee, N.,
and Quackenbush, J. (2000). A concise guide to cDNA microarray analysis. Biotechniques 29,
548 –560.
Heidelberg, J. F., Paulsen, I. T., Nelson, K. E., Gaidos, E. J., Nelson, W. C., Read, T. D., Eisen, J. A.,
Seshadri, R., Ward, N., Methe, B., Clayton, R. A., Meyer, T., Tsapin, A., Scott, J., Beanan, M.,
Brinkac, L., Daugherty, S., DeBoy, R. T., Dodson, R. J., Durkin, A. S., Haft, D. H., Kolonay, J. F.,
Madupu, R., Peterson, J. D., Umayam, L. A., White, O., Wolf, A. M., Vamathevan, J., Weidman,
J., Impraim, M., Lee, K., Berry, K., Lee, C., Mueller, J., Khouri, H., Gill, J., Utterback, T. R.,
McDonald, L. A., Feldblyum, T. V., Smith, H. O., Venter, J. C., Nealson, K. H., and Fraser, C. M.
(2002). Genome sequence of the dissimilatory metal ion-reducing bacterium Shewanella
oneidensis. Nature Biotechnol. 20, 1118–1123.
Heller, H. M., Schena, M., Chai, A., Shalon, D., Bedilion, T., and Gilmore, J. (1997). Discovery and
analysis of inflammatory disease-related genes using cDNA microarrays. Proc. Natl. Acad. Sci.
USA 94, 2150– 2155.
Herrero, J., Valencia, A., and Dopazo, J. (2001). A hierarchical unsupervised growing neural network
for clustering gene expression patterns. Bioinformatics 17, 126–136.
Herwig, R., Poustka, A. J., Muller, C., Bull, C., Lehrach, H., and O’Brien, J. (1999). Large-scale
clustering of cDNA-fingerprinting data. Genome Res. 9, 1093–1105.
Heyer, L. J., Kruglyak, S., and Yooseph, S. (1999). Exploring expression data: identification and
analysis of coexpressed genes. Genome Res. 9, 1106–1115.
Hilsenbeck, S. G., Friedrichs, W. E., Schiff, R., O’Connell, P., Hansen, R. K., Osborne, C. K., and
Fuqua, S. A. (1999). Statistical analysis of array expression data as applied to the problem of
tamoxifen resistance. J. Natl. Cancer Inst. 91, 453–459.
Hoheisel, J. (1997). Oligomer-chip technology. Trends Biotech. 15, 465–469.
Hughes, T. R., Marton, M. J., Jones, A. R., Roberts, C. J., Stoughton, R., Armour, C. D., Bennett, H. A.,
Coffey, E., Dai, H., He, Y. D., Kidd, M. J., King, A. M., Meyer, M. R., Slade, D., Lum, P. Y.,
Stepaniants, S. B., Shoemaker, D. D., Gachotte, D., Chakraburtty, K., Simon, J., Bard, M., and
Friend, S. H. (2000). Functional discovery via a compendium of expression profiles. Cell 102,
109 –126.
Kalocsai, P., and Shams, S. (2001). Use of bioinformatics in arrays. Meth. Mol. Biol. 170,
223 –236.
Kane, M. D., Jatkoe, T. A., Stumpf, C. R., Lu, J., Thomas, J. D., and Madore, S. J. (2000). Assessment
of the sensitivity and specificity of oligonucleotide (50mer) microarrays. Nucl. Acids Res. 28,
4552–4557.
Kaneko, T., Nakamura, Y., Wolk, C. P., Kuritz, T., Sasamoto, S., Watanabe, A., Iriguchi, M.,
Ishikawa, A., Kawashima, K., Kimura, T., Kishida, Y., Kohara, M., Matsumoto, M., Matsuno, A.,
Muraki, A., Nakazaki, N., Shimpo, S., Sugimoto, M., Takazawa, M., Yamada, M., Yasuda, M.,
and Tabata, S. (2001). Complete genomic sequence of the filamentous nitrogen-fixing
cyanobacterium Anabaena sp. strain PCC 7120. DNA Res. 8, 205–213.
Kaneko, T., Sato, S., Kotani, H., Tanaka, A., Asamizu, E., Nakamura, Y., Miyajima, N., Hirosawa, M.,
Sugiura, M., Sasamoto, S., Kimura, T., Hosouchi, T., Matsuno, A., Muraki, A., Nakazaki, N.,
Naruo, K., Okumura, S., Shimpo, S., Takeuchi, C., Wada, T., Watanabe, A., Yamada, M.,
Yasuda, M., and Tabata, S. (1996). Sequence analysis of the genome of the unicellular
cyanobacterium Synechocystis sp. strain PCC6803. II. Sequence determination of the entire
genome and assignment of potential protein-coding regions. DNA Res. 3, 109 –136.
Kawarabayasi, Y., Hino, Y., Horikawa, H., Yamazaki, S., Haikawa, Y., Jin-no, K., Takahashi, M.,
Sekine, M., Baba, S., Ankai, A., Kosugi, H., Hosoyama, A., Fukui, S., Nagai, Y., Nishijima, K.,
Nakazawa, H., Takamiya, M., Masuda, S., Funahashi, T., Tanaka, T., Kudoh, Y., Yamazaki, J.,
Kushida, N., Oguchi, A., Aoki, K., Kubota, K., Nakamura, Y., Nomura, N., Sako, Y., and
Kikuchi, H. (1999). Complete genome sequence of an aerobic hyper-thermophilic crenarchaeon,
Aeropyrum pernix K1. DNA Res. 6, 83–101.
Kawarabayasi, Y., Hino, Y., Horikawa, H., Jin-no, K., Takahashi, M., Sekine, M., Baba, S., Ankai, A.,
Kosugi, H., Hosoyama, A., Fukui, S., Nagai, Y., Nishijima, K., Otsuka, R., Nakazawa, H.,
Takamiya, M., Kato, Y., Yoshizawa, T., Tanaka, T., Kudoh, Y., Yamazaki, J., Kushida, N.,
Oguchi, A., Aoki, K., Masuda, S., Yanagii, M., Nishimura, M., Yamagishi, A., Oshima, T., and
Kikuchi, H. (2001). Complete genome sequence of an aerobic thermoacidophilic crenarchaeon,
Sulfolobus tokodaii strain7. DNA Res. 8, 123 –140.
Kawarabayasi, Y., Sawada, M., Horikawa, H., Haikawa, Y., Hino, Y., Yamamoto, S., Sekine, M.,
Baba, S., Kosugi, H., Hosoyama, A., Nagai, Y., Sakai, M., Ogura, K., Otsuka, R., Nakazawa, H.,
Takamiya, M., Ohfuku, Y., Funahashi, T., Tanaka, T., Kudoh, Y., Yamazaki, J., Kushida, N.,
Oguchi, A., Aoki, K., and Kikuchi, H. (1998). Complete sequence and gene organization of the
genome of a hyper-thermophilic archaebacterium, Pyrococcus horikoshii. DNA Res. 5, 55 –76.
Kawashima, T., Amano, N., Koike, H., Makino, S., Higuchi, S., Kawashima-Ohya, Y., Watanabe, K.,
Yamazaki, M., Kanehori, K., Kawamoto, T., Nunoshiba, T., Yamamoto, Y., Aramaki, H.,
Makino, K., and Suzuki, M. (2000). Archaeal adaptation to higher temperatures revealed by
genomic sequence of Thermoplasma volcanium. Proc. Natl. Acad. Sci. USA 97, 14257–14262.
Kerr, M. K., and Churchill, G. A. (2001a). Statistical design and the analysis of gene expression
microarray data. Genet. Res. 77, 123–128.
Kerr, M. K., and Churchill, G. A. (2001b). Experimental design for gene expression microarrays.
Biostatistics 2, 183– 201.
Khodursky, A. B., Peter, B. J., Cozzarelli, N. R., Botstein, D., Brown, P. O., and Yanofsky, C. (2000).
DNA microarray analysis of gene expression in response to physiological and genetic changes
that affect tryptophan metabolism in Escherichia coli. Proc. Natl. Acad. Sci. USA 97,
12170– 12175.
Khrapko, K. R., Lysov, Y. P., Khorlyn, A. A., Ivanov, I. B., Yershov, G. M., Vasilenko, S. K.,
Vlorentiev, V. L., and Mirzabekov, A. (1989). An oligonucleotide hybridization approach to
DNA sequencing. FEBS Lett. 256, 118 –122.
Kingsley, M. T., Straub, T. M., Call, D. R., Daly, D. S., Wunschel, S. C., and Chandler, D. P. (2002).
Fingerprinting closely related Xanthomonas pathovars with random nanamer oligonucleotide
microarrays. Appl. Environ. Microbiol. 68, 6361–6370.
Kleman-Leyer, K., Armbruster, D. W., and Daniels, C. J. (1997). Properties of H. volcanii tRNA
intron endonuclease reveal a relationship between the archaeal and eucaryal tRNA intron
processing systems. Cell 89, 839–847.
Klenk, H.-P., Clayton, R. A., Tomb, J.-F., White, O., Nelson, K. E., Ketchum, K. A., Dodson, R. J.,
Gwinn, M., Hickey, E. K., Peterson, J. D., Richardson, D. L., Kerlavage, A. R., Graham, D. E.,
Kyrpides, N. C., Fleischmann, R. D., Quackenbush, J., Lee, N. H., Sutton, G. G., Gill, S.,
Kirkness, E. F., Dougherty, B. A., McKenney, K., Adams, M. D., Loftus, B., Peterson, S., Reich,
C. I., McNeil, L. K., Badger, J. H., Glodek, A., Zhou, L. X., Overbeek, R., Gocayne, J. D.,
Weidman, J. F., McDonald, L., Utterback, T., Cotton, M. D., Spriggs, T., Artiach, P., Kaine, B. P.,
Sykes, S. M., Sadow, P. W., Dandrea, K. P., Bowman, C., Fujii, C., Garland, S. A., Mason, T. M.,
Olsen, G. J., Fraser, C. M., Smith, H. O., Woese, C. R., and Venter, J. C. (1997). The complete
genome sequence of the hyperthermophilic, sulphate-reducing archaeon Archaeoglobus fulgidus.
Nature 394, 364–370.
Knight, J. (2001). When the chips are down. Nature 410, 860–861.
Lange, C. C., Wackett, L. P., Minton, K. W., and Daly, M. J. (1998). Engineering a recombinant
Deinococcus radiodurans for organopollutant degradation in radioactive mixed waste
environments. Nat. Biotechnol. 16, 929–933.
Lamture, J. B., Beattie, K. L., Burke, B. E., Eggers, M. D., Ehrlich, D. J., Fowler, R., Hollis, M. A.,
Kosicki, B. B., Reich, R. K., and Smith, S. R. (1994). Direct detection of nucleic acid
hybridization on the surface of a charge coupled device. Nucl. Acids Res. 22, 2121–2125.
Lipshultz, R. J., Fodor, S. P. A., Gingeras, T. R., and Lockhart, D. J. (1999). High density synthetic
oligonucleotide arrays. Nat. Genet. 21, 20–24.
Liu, Y., Zhou, J., Omelchenko, M., Beliaev, A., Venkateswaran, A., Stair, J., Wu, L., Thompson,
D. K., Xu, D., Rogozin, I. B., Gaidamakova, E. K., Zhai, M., Makarova, K. S., Koonin, E. V., and
Daly, M. J. (2003). Transcriptome dynamics of Deinococcus radiodurans recovering from
ionizing radiation. Proc. Natl. Acad. Sci. USA 100, 4191–4196.
Lockhart, D. J., Dong, H., Byrne, M. C., Follettie, M. T., Gallo, M. V., Chee, M. S., Mittmann, M.,
Wang, C., Kobayashi, M., Horton, H., and Brown, E. L. (1996). Expression monitoring by
hybridization to high-density oligonucleotide arrays. Nat. Biotechnol. 14, 1675–1680.
Long, A. D., Mangalam, H. J., Chan, B. Y., Tolleri, L., Hatfield, G. W., and Baldi, P. (2001). Improved
statistical inference from DNA microarray data using analysis of variance and a Bayesian
statistical framework. Analysis of global gene expression in Escherichia coli K12. J. Biol. Chem.
276, 19937– 19944.
Lovley, D. (1991). Dissimilatory Fe(III) and Mn(IV) reduction. Microbiol. Rev. 55, 259–287.
Loy, A., Lahner, A., Lee, N., Adamczyk, J., Meier, H., Ernst, J., Schleifer, K. H., and Wagner, M.
(2002). Oligonucleotide microarray for 16S rRNA gene-based detection of all recognized
lineages of sulfate-reducing prokaryotes in the environment. Appl. Environ. Microbiol. 68,
5064–5081.
Mace, M. L., Montagu, J., Rose, S. D., and McGuinnes, G. (2000). Novel microarray printing and
detection technologies. In “Microarray Biochip Technology” (M. Schena, Ed.), pp. 39–64. Eaton
Makarova, K. S., Aravind, L., Wolf, Y. I., Tatusov, R. L., Minton, K. W., Koonin, E. V., and
Daly, M. J. (2001). Genome of the extremely radiation-resistant bacterium Deinococcus
radiodurans viewed from the perspective of comparative genomics. Microbiol. Mol. Biol.
Rev. 65, 44–79.
Martinsky, T., and Haje, P. (2000). Microarray tools, kits, reagents, and services. In “Microarray
Biochip Technology” (M . Schena, Ed.), pp. 201–220. Eaton Publishing, Natick, MA.
McGall, G. H., and Fidanza, J. A. (2001). Photolithographic synthesis of high-density oligonucleotide
arrays. In “DNA Arrays — Methods and Protocols” (J. B. Rampal, Ed.), pp. 71–101. Humana
Press, Totowa, NJ.
Minton, K. W. (1996). Repair of ionizing-radiation damage in the radiation resistant bacterium
Deinococcus radiodurans. Mutat. Res. 363, 1 –7.
Morrison, N., Rattray, M., Oliver, S. G., Hayes, A., Zhang, N., Brutsche, M., Penkett, C., Lockey, J.,
Rao, S., Hays, I., Jupp, R., and Brass, A. (1999). Robust normalization of microarray data over
multiple experiments. Nature Genetics Microarray Meeting: Technology, Application and
Analysis.
Moser, D., and Nealson, K. H. (1996). Growth of the facultative anaerobe Shewanella putrefaciens by
elemental sulfur reduction. Appl. Environ. Microbiol. 62, 2100–2105.
Murray, A. E., Lies, D., Li, G., Nealson, K., Zhou, J., and Tiedje, J. M. (2001). DNA/DNA
hybridization to microarrays reveals gene-specific differences between closely related microbial
genomes. Proc. Natl. Acad. Sci. USA 98, 9853–9858.
Nadon, R., Shi, P., Skandalis, A., Woody, E., Hubschle, H., Susko, E., Rghei, N., and Ramm, P.
(2001). Statistical inference methods for gene expression arrays. In “Proceedings of SPIE,
Microarrays: Optical Technologies and Informatics” (Bittner, Ed.), V4266, pp. 46 – 55.
International Society for Optical Engineering, Bellingham, WA.
Nadon, R., and Shoemaker, J. (2002). Statistical issues with microarrays: processing and analysis.
Trends Genet. 18, 265–271.
Nelson, K. E., Clayton, R. A., Gill, S. R., Gwinn, M. L., Dodson, R. J., Haft, D. H., Hickey, E. K.,
Peterson, L. D., Nelson, W. C., Ketchum, K. A., McDonald, L., Utterback, T. R., Malek, J. A.,
Linher, K. D., Garrett, M. M., Stewart, A. M., Cotton, M. D., Pratt, M. S., Phillips, C. A.,
Richardson, D., Heidelberg, J., Sutton, G. G., Fleischmann, R. D., Eisen, J. A., White, O.,
Salzberg, S. L., Smith, H. O., Venter, J. C., and Fraser, C. M. (1999). Evidence for lateral gene
transfer between Archaea and Bacteria from genome sequence of Thermotoga maritima. Nature
399, 323 –329.
Nealson, K. H., and Saffarini, D. A. (1994). Iron and manganese in anaerobic respiration:
environmental significance, physiology, and regulation. Ann. Rev. Microbiol. 48, 311–343.
Ng, W. V., Kennedy, S. P., Mahairas, G. G., Berquist, B., Pan, M., Shukla, H. D., Lasky, S. R., Baliga,
N. S., Thorsson, V., Sbrogna, J., Swartzell, S., Weir, D., Hall, J., Dahl, T. A., Welti, R., Goo,
Y. A., Leithauser, B., Keller, K., Cruz, R., Danson, M. J., Hough, D. W., Maddocks, D. G.,
Jablonski, P. E., Krebs, M. P., Angevine, C. M., Dale, H., Isenbarger, T. A., Peck, R. F.,
Pohlschroder, M., Spudich, J. L., Jung, K. H., Alam, M., Freitas, T., Hou, S. B., Daniels, C. J.,
Dennis, P. P., Omer, A. D., Ebhardt, H., Lowe, T. M., Liang, R., Riley, M., Hood, L., and
DasSarma, S. (2000). Genome sequence of Halobacterium species NRC-1. Proc. Natl. Acad. Sci.
USA 97, 12176–12181.
Nuwaysir, E. F., Huang, W., Albert, T. J., Singh, J., Nuwaysir, K., Pitas, A., Richmond, T., Gorski, T.,
Berg, J. P., Ballin, J., McCormick, M., Norton, J., Pollock, T., Sumwalt, T., Butcher, L., Porter,
D., Molla, M., Hall, C., Blattner, F., Sussman, M. R., Wallace, R. L., Cerrina, F., and Green, R. D.
(2002). Gene expression analysis using oligonucleotide arrays produced by maskless
photolithography. Genome Res. 12, 1749–1755.
Olsen, G. J., and Woese, C. R. (1997). Archaeal genomics: an overview. Cell 89, 991–994.
Pease, A. C., Solas, D., Sullivan, E. J., Cronin, M. T., Holmes, C. P., and Fodor, S. P. A. (1994).
Light-generated oligonucleotide arrays for rapid DNA sequence analysis. Proc. Natl. Acad. Sci.
USA 91, 5022–5026.
Petrov, A., Shah, S., Draghici, S., and Shams, S. (2002). Microarray image processing and quality
control. In “DNA Array Image Analysis — Nuts & Bolts” (S. Shah and G. Kamberova, Eds.),
pp. 99– 130. DNA Press, LLC, Eagleville, PA.
Pinkel, D., Segraves, R., Sudar, D., Clark, S., Poole, I., Kowbel, D., Collins, C., Kuo, W. L., Chen, C.,
Zhai, Y., Dairkee, S. H., Ljung, B. M., Gray, J. W., and Albertson, D. G. (1998). High resolution
analysis of DNA copy number variation using comparative genomic hybridization to
microarrays. Nat. Genet. 20, 207 –11.
Pollack, R. J., Perou, C., Alizadeh, A., Eisen, M., Pergamenschikov, A., Williams, C., Jeffrey, S.,
Botstein, D., and Brown, P. (1999). Genome-wide analysis of DNA copy-number changes using
cDNA microarrays. Nat. Am. 23, 41–46.
Ramsay, G. (1998). DNA chips: state-of-the art. Nat. Biotechnol. 1, 640– 644.
Rehman, F. N., Audeh, M., Abrams, E. S., Hammond, P. W., Kenney, M., and Boles, T. C. (1999).
Immobilization of acrylamide-modified oligonucleotides by co-polymerization. Nucl. Acids Res
27, 649 –655.
Richardson, D. J. (2000). Bacterial respiration: a flexible process for a changing environment.
Microbiol. 146, 551–571.
Robb, F. T., Maeder, D. L., Brown, J. R., DiRuggiero, J., Stump, M. D., Yeh, R. K., Weiss, R. B., and
Dunn, D. M. (2001). Genomic sequence of hyperthermophile Pyrococcus furiosus: implications
for physiology and enzymology. Meth. Enzymol. 330, 134– 157.
Rogers, Y., Jiang-Baucome, P., Huange, Z., Bogdanov, V., Anderson, S., and Boyce-Jacino, M. T.
(1999). Immobilization of oligonucleotides onto a glass support via disulfide bonds: A method for
preparation of DNA microarrays. Ann. Biochem. 266, 23 –30.
Rose, D. (2000). Microfluidic technologies and instrumentation for printing DNA microarrays. In
“Microarray Biochip Technology” (M. Schena, Ed.), pp. 19–38. Eaton Publishing, Natick, MA.
Rudi, K., Skulberg, O. M., Skulberg, R., and Jakobsen, K. S. (2000). Application of sequence-specific
labeled 16S rRNA gene oligonucleotide probes for genetic profiling of cyanbacterial abundance
and diversity by array hybridization. Appl. Environ. Microbiol. 66, 4004–4011.
Salama, N., Guillemin, K., McDaniel, T. K., Sherlock, G., Tompkins, L., and Falkow, S. (2000). A
whole-genome microarray reveals genetic diversity among Helicobacter pylori strains. Proc.
Natl. Acad. Sci. USA 97, 14668–14673.
Salunga, R. C., Guo, H., Luo, L., Bittner, A., Joy, K. D., Chambers, J. R., Wan, J. S., Jackson, M. R.,
and Erlander, M. G. (1999). Gene expression analysis via cDNA microarrays of laser
capture microdissected cells from fixed tissue. In “DNA Microarrays: A Practical Approach”
(M. Schena, Ed.), pp. 121 –136. Oxford University Press, New York, NY.
Schadt, E. E., Li, C., Su, C., and Wong, W. H. (2000). Analyzing high-density oligonucleotide gene
expression array data. J. Cell Biochem. 80, 192 –202.
Schena, M. (2003). “Microarray analysis.” John Wiley & Sons, New York, NY.
Schena, M., and Davis, R. W. (2000). Technology standards for microarray research. In “Microarray
Biochip Technology” (M. Schena, Ed.), pp. 1 –18. Eaton Publishing, Natick, MA.
Schena, M., Heller, R., Theriault, T., Konrad, K., Lachenmeier, E., and Davis, R. W. (1998).
Microarrays: biotechnology’s disovery platform for functional genomics. Trends Biotech. 16,
301 –306.
Schena, M., Shalon, D., Davis, R. W., and Brown, P. O. (1995). Quantitative monitoring of gene
expression patterns with a complementary DNA microarray. Science. 270, 467 –470.
Schena, M., Shalon, D., Heller, R., Chai, A., Brown, P., and Davis, R. W. (1996). Parallel human
genome analysis: microarray-based expression monitoring of 1000 genes. Proc. Natl. Acad. Sci.
USA 93, 10614– 10619.
Schermer, M. J. (1999). Confocal scanning microscopy in microarray detection. In “DNA
Microarrays: A Practical Approach” (M. Schena, Ed.), pp. 17–42. Oxford University Press,
New York, NY.
Schuchhardt, J., Beule, D., Malik, A., Wolski, E., Eickhoff, H., Lehrach, H., and Herzel, H. (2000).
Normalization strategies for cDNA microarrays. Nucl. Acids Res. 28, E47.
Schut, G. J., Zhou, J., and Adams, M. W. W. (2001). DNA microarray analysis of the
hyperthermophilic archaeon Pyrococcus furiosus: Evidence for a new type of sulfur-reducing
enzyme complex. J. Bacteriol. 183, 7027–7036.
Shalon, D., Smith, S. J., and Brown, P. O. (1996). A DNA microarray system for analyzing complex
DNA samples using two-color fluorescent probe hybridization. Genome Res. 6, 639–645.
Shchepinov, M. S., Case-Green, S. C., and Southern, E. M. (1997). Steric factors influencing
hybridisation of nucleic acids to oligonucleotide arrays. Nucl. Acids Res. 25, 1155–1161.
Shriver-Lake, L. C. (1998). Silane-modified surfaces for biomaterial immobilization. In “Immobilized
Biomolecules in Analysis” (T. Cass and F. S. Ligler, Eds.), pp. 1– 14. Oxford University Press,
New York, NY.
Slesarev, A. I., Mezhevaya, K. V., Makarova, K. S., Polushin, N. N., Shcherbinina, O. V., Shakhova,
V. V., Belova, G. I., Aravind, L., Natale, D. A., Rogozin, I. B., Tatusov, R. L., Wolf, Y. I., Stetter,
K. O., Malykh, A. G., Koonin, E. V., and Kozyavkin, S. A. (2002). The complete genome of
hyperthermophile Methanopyrus kandleri AV19 and monophyly of archaeal methanogens. Proc.
Natl. Acad. Sci. USA 99, 4644–4649.
Small, J., Call, D. R., Brockman, F. J., Straub, T. M., and Chandler, D. P. (2001). Direct detection of
16S rRNA in soil extracts by using oligonucleotide microarrays. Appl. Environ. Microbiol. 67,
4708–4716.
Smith, D. R., Doucette-Stamm, L. A., Deloughery, C., Lee, H., Dubois, J., Aldredge, T., Bashirzadeh,
R., Blakely, D., Cook, R., Gilbert, K., Harrison, D., Hoang, L., Keagle, P., Lumm, W., Pothier, B.,
Qiu, D., Spadafora, R., Vicaire, R., Wang, Y., Wierzbowski, J., Gibson, R., Jiwani, N., Caruso,
A., Bush, D., Reeve, J. N., et al., (1997). Complete genome sequence of Methanobacterium
thermoautotrophicum KH: Functional analysis and comparative genomics. J. Bacteriol. 179,
7135–7155.
Southern, E. M. (2001). DNA microarrays. History and overview. Meth. Mol. Biol. 170, 1–15.
Southern, E. M., Case-Green, S. C., Elder, J. K., Johnsone, M., Mir, K. U., Wang, L., and Williams,
J. C. (1994). Arrays of complementary oligonucleotides for analyzing the hybridizaiton behavior
of nucleic acids. Nucl. Acids Res. 22, 1368– 1373.
Southern, E. M., Maskos, U., and Elder, J. K. (1992). Analyzing and comparing nucleic acid
sequences by hybridization to arrays of oligonucleotides: evaluation using experimental models.
Genomics 13, 1108–1107.
Southern, E., Mir, K., and Shchepinov, M. (1999). Molecular interactions on microarrays. Nat. Genet.
21(Suppl. S), 5 –9.
Stears, R. L., Getts, R. C., and Gullans, S. R. (2000). A novel, sensitive detection system for high-
density microarrays using dendrimer technology. Physiol. Genomics 3, 93–99.
Stears, B. L., Martinsky, T., and Schena, M. (2003). Trends in microarray analysis. Nat. Med. 9,
140–145.
Steel, A. B., Levicky, R. L., Herne, T. M., and Tarlo, M. J. (2000). Immobilization of nucleic acids at
solid surfaces: effect of oligonucleotide length on layer assembly. Biophys. J79, 975–981.
Stillman, B. A., and Tonkinson, J. L. (2000). FAST slides: a novel surface for microarrays.
Biotechniques 29, 630–635.
Tamayo, P., Slonim, D., Mesirov, J., Zhu, Q., Kitareewan, S., Dmitrovsky, E., Lander, E. S., and
Golub, T. R. (1999). Interpreting patterns of gene expression with self-organizing maps:
methods and application to hematopoietic differentiation. Proc. Natl. Acad. Sci. USA 96,
2907–2912.
Taniguchi, M., Miura, K., Iwao, H., and Yamanaka, S. (2001). Quantitative assessment of DNA
microarrays — comparison with Northern blot analyses. Genomics 71, 34 –39.
Tao, H., Bausch, C., Richmond, C., Blattner, F. R., and Conway, T. (1999). Functional genomics:
expression analysis of Escherichia coli growing on minimal and rich media. J. Bacteriol. 181,
6425–6440.
Taroncher-Oldenburg, G., Griner, E. M., Francis, C. A., and Ward, B. B. (2003). Oligonucleotide
microarray for the study of functional gene diversity in the nitrogen cycle in the environment.
Appl. Environ. Microbiol. 69, 1159–1171.
Tavazoie, S., and Church, G. M. (1998). Quantitative whole-genome analysis of DNA-protein
interactions by in vivo methylase protection in E-coli. Nat. Biotech. 16, 566 –571.
Thompson, D. K., Beliaev, A. S., Giometti, C. S., Tollaksen, S. L., Khare, T., Lies, D. P., Nealson,
K. H., Lim, H., Yates, III, J., Brandt, C. C., Tiedje, J. M., and Zhou, J. (2002). Transcriptional and
proteomic analysis of a ferric uptake regulator (Fur) mutant of Shewanella oneidensis: Possible
involvement of Fur in energy metabolism, transcriptional regulation, and oxidative stress. Appl.
Toronen, P., Kolehmainen, M., Wong, G., and Castren, E. (1999). Analysis of gene expression data
using self-organizing maps. FEBS Lett. 451, 142 –146.
Tseng, G. C., Oh, M. K., Rohlin, L., Liao, J. C., and Wong, W. H. (2001). Issues in cDNA microarray
analysis: quality filtering, channel normalization, models of variations and assessment of gene
effects. Nucl. Acids Res. 29, 2549–2557.
Urakawa, H., Noble, P. A., El Fantroussi, S., Kelly, J. J., and Stahl, D. A. (2002). Single-base-pair
discrimination of terminal mismatches by using oligonucleotide microarrays and neural network
analyses. Appl. Environ. Microbiol. 68, 235–244.
Vasiliskov, A., Timofeev, E., Surzhikov, S., Drobyshev, A., Shick, V., and Mirzabekov, A. (1999).
Fabricaiton of microarray of gel-immobilized compounds on a chip by copolymerization.
Biotechniques 27, 592 –606.
Venkateswaran, A., McFarlan, S. C., Ghostal, D., Minton, K. W., Vasilenko, A., Makarova, K. S.,
Wackett, L. P., and Daly, M. J. (2000). Physiologic determinants of radiation resistance in
Deinococcus radiodurans. Appl. Environ. Microbiol. 66, 2620–2626.
Venkateswaran, K., Moser, D., Dollhopf, M., Lies, D. P., Saffarini, D. A., MacGregor, B. J.,
Ringelberg, D. B., White, D. C., Nishijima, M., and Sano, H. (1999). Polyphasic taxonomy of the
genus Shewanella and description of Shewanella oneidensis sp. nov. Int. J. Syst. Bacteriol. 49,
705 –724.
Verdnik, D., Handran, S., and Pickett, S. (2002). Key considerations for accurate microarray scanning
and image analysis. In “DNA Array Image Analysis — Nuts & Bolts” (S. Shah and G.
Kamberova, Eds.). DNA Press, LLC, Eagleville, PA.
Voordouw, G. (1998). Reverse sample genome probing of microbial community dynamics. ASM
News 64, 627–633.
Wang, D. G., Fan, J. B., Siao, C. J., Berno, A., Young, P., Sapolsky, R., Ghandour, G., Perkins,
N., Winchester, E., Spencer, J., Kruglyak, L., Stein, L., Hsie, L., Topaloglou, E., Hubbell,
E., Robinson, M., Mittmann, M. S., Morris, N., Shen, D., Kilburn, T., Rioux, J., Nusbaum,
C., Rozen, S., Hudson, T. J., Lander, E. S., and et al. (1998). Large-scale identification,
mapping, and genotyping of single-nucleotide polymorphisms in the human genome. Science
280, 1077– 1082.
Warrington, J. A., Dee, S., and Trulson, M. (2000). Large-scale genomic analysis using Affymetrix
GeneChipw probe arrays. In “Microarray Biochip Technology” (M. Schena, Ed.), pp. 119–148.
Eaton Publishing, Natick, MA.
White, O., Eisen, J. A., Heidelberg, J. F., Hickey, E. K., Peterson, J. D., Dodson, R. J., Haft, D. H.,
Gwinn, M. L., Nelson, W. C., Richardson, D. L., Moffat, K. S., Qin, H. Y., Jiang, L. X., Pamphile,
W., Crosby, M., Shen, M., Vamathevan, J. J., Lam, P., McDonald, L., Utterback, T., Zalewski, C.,
Makarova, K. S., Aravind, L., Daly, M. J., Minton, K. W., Fleischmann, R. D., Ketchum, K. A.,
Nelson, K. E., Salzberg, S., Smith, H. O., Venter, J. C., and Fraser, C. M. (1999). Genome
sequence of the radioresistant bacterium Deinococcus radiodurans R1. Science 286, 1571–1577.
Wilson, K. H., Wilson, W. J., Radosevich, J. L., DeSantis, T. Z., Viswanathan, V. S., Kuczmarski,
T. A., and Andersen, G. L. (2002). High-density microarray of small-subunit ribosomal DNA
probes. Appl. Environ. Microbiol. 68, 2535– 2541.
Wodicka, L., Dong, H., Mittmann, M., Ho, M.-H., and Lockhart, D. J. (1997). Genome-wide
expression monitoring in Saccharomyces cerevisiae. Nat. Biotechnol. 15, 1359–1367.
Woese, C. R. (1987). Bacterial evolution. Microbiol. Rev. 51, 221 –271.
Woese, C. R., Kandler, O., and Wheelis, M. L. (1990). Towards a natural system of organisms: proposal
for the domains Archaea, Bacteria, and Eucarya. Proc. Natl. Acad. Sci. USA 87, 4576–4579.
Worley, J., Bechtol, K., Penn, S., Roach, D., Hanzel, D., Trounstine, M., and Barker, D. (2000). A
systems approach to fabricating and analyzing DNA microarrays. In “Microarray Biochip
Technology” (M. Schena, Ed.), pp. 65–85. Eaton Publishing, Natick, MA.
Wu, L. Y., Thompson, D., Li, G.-S., Hurt, R., Huang, H., Tiedje, J. M., and Zhou, J.-Z. (2001).
Development and evaluation of functional gene arrays for detection of selected genes in the
environment. Appl. Environ. Microbiol. 67, 5780–5790.
Xu, D., Li, G., Wu, L., Zhou, J.-Z., and Xu, Y. (2002). PRIMEGENS: A computer program for robust
and efficient design of gene-specific targets on microarrays. Bioinformatics 18, 1432– 1437.
Yang, Y. H., and Speed, T. (2002). Design issues for cDNA microarray experiments. Nat. Rev. Genet.
3, 579–588.
Ye, R., Tao, W., Bedzyk, L., Young, T., Chen, M., and Li, L. (2000). Global gene expression profiles
of Bacillus subtilis grown under anaerobic conditions. J. Bacteriol. 182, 4458–4465.
Zammatteo, N., Jeanmart, L., Hamels, S., Courtois, S., Louette, P., Hevesi, L., and Remacle, J. (2000).
Comparison between different strategies of covalent attachment of DNA to glass surfaces to build
DNA microarrays. Anal. Biochem. 280, 143–150.
Zhao, X., Nampalli, S., Serino, A. J., and Kumar, S. (2001). Immobilization of oligodeoxyribonucleo-
tides with multiple anchors to microchips. Nucl. Acids Res. 29, 955–959.
Zhou, J.-Z. (2003). Microarrays for bacterial detection and microbial community analysis. Curr.
Opin. Microbiol. 6, 288– 294.
Zhou, Y.-X., Kalocsai, P., Chen, J.-Y., and Shams, S. (2000). Information processing issues and
solutions associated with microarray technology. In “Microarray Biochip Technology”
(M. Schena, Ed.), pp. 167–200. Eaton Publishing, Natick, MA.
Zhou, J., Palumbo, A. V., and Tiedje, J. M. (1997). Sensitive detection of a novel class of toluene-
degrading denitrifiers, Azoarcus tolulyticus, with small-subunit rRNA primers and probes. Appl.
Zhou, J., and Thompson, D. K. (2002). Challenges in applying microarrays to environmental studies.
Curr. Opin. Biotechnol. 13, 204 –207.
Zlatanova, J., and Mirzabekov, A. (2001). Gel-immobilized microarrays of nucleic acids and proteins:
production and application for macromolecular research. Meth. Mol. Biol. 170, 17–38.
THE AGRONOMY AND ECONOMY OF
BLACK PEPPER (PIPER NIGRUM L. ) —
THE “KING OF SPICES ”
K. P. Prabhakaran Nair
Distinguished Visiting Scientist, Indian Council of Agricultural Research,
New Delhi, India
I. Introduction
II. The Pepper Plant — Its Botany and Chemistry
A. Pepper Botany
B. Pepper Chemistry
III. Pepper Agronomy
A. The pepper Soils
B. Nutrition of Black Pepper
C. Evolution of Pepper Manuring
D. Response of Pepper to Mineral Nutrients
IV. The Role of “The Nutrient Buffer Power Concept” in Pepper
Nutrition
A. The Buffer Power and Effects on Nutrient Availability
B. Basic Concepts
C. Measuring the Nutrient Buffer Power and Its Importance in
Affecting Nutrient Concentration on Root Surfaces
D. Background Informaton on the Importance of Measuring Zn
Buffer Power
E. Quantifying Zn Buffer Power of Pepper Growing Soils
V. Establishing a Pepper Plantation
A. The Indian Experience
B. The Indonesian Experience
VI. Pepper Pests and Their Control
VII. The Processing of Black Pepper on Farm
A. Sun Drying of Pepper
B. Solar Drying of Pepper
C. Garbling, Cleaning, and Fractionation
D. Packaging and Storing
VIII. An Account of Indonesian Pepper Processing
IX. Industrial Processing of Black Pepper
A. White Pepper
B. Cryoground Pepper
C. Pepper Oil and Oleoresin
271
0065-2113/03 $35.00
272 K. P. P. NAIR
X. The Future of Global Pepper Economy

A. The Supply Side of Pepper Economy
B. The Demand Side of Pepper Economy
C. Prices and World Market
D. The Pepper Price Out Look by 2020
E. The Pepper Supply Out Look by 2020
F. The Pepper Demand Out Look by 2020
G. Country-wise Economic Growth Impacting Production
and Consumption
XI. Pepper Economy in India
A. Pepper Production Scenario in India
B. Bush Pepper
C. Economics of Pepper Production in the State of Kerala
D. Marketing
E. Pepper Futures Market
XII. Pepper Pharmacopoeia
A. Antiinflammatory and Central Nervous System (CNS)
Depressant Activity of Pepper
B. Effect on Hepatic Enzymes
C. Carcinogenic and Mutagenic Effects of Black Pepper
D. Pepper as An Antioxidant
E. Pepper as An Antimicrobial Agent
F. The Pharmacological Effect of Pepper on Human Health
G. Clinical Applications of Pepper
H. Toxicological Effects
I. The Insecticidal Activity of Pepper
XIII. Consumer Products of Black Pepper
XIV. Value Addition in Pepper
XV. Conclusions and a Peep Into Pepper’s Future
Acknowledgments
References
Black pepper, popularly known as the “King of Spices”, has a very

checkered history dating back to the times of Queen Sheeba and King
Solomon (BC 1015– BC 66) and has influenced the destiny of nations and
their people, spread across the world, both economically and culturally.
Today pepper commands the leading position among the different spices as
the spice of immense commercial importance in world trade and is finding its
way into the dietary habits of millions around the world, even among people
on the European and North American continents, hitherto unaccustomed to
its use. Pepper use ranges from a simple dietary constituent to that of
immense pharmacological benefits. Though beset with many problems, both
agronomic and economic, it is a safe bet that pepper will emerge as the
world’s most sought after spice, its predicted global demand escalating
colossally to about 280,000 metric tons by the year 2020 that will further
climb to 360,000 metric tons by the year 2050. This very exhaustive review
details the various constraints to enhancing productivity of pepper and charts
AGRONOMY AND ECONOMY OF BLACK PEPPER 273
contours of a new course of action. Among the primary constraints in pepper

production is the absence of an ideotype that combines many positive traits to
boost production potential, whilst at the same time resisting ravages of
nature such as the onslaught of the dreaded disease “Foot Rot” caused by
Phytophthora fungi, whch is the most devastating. Fertility management of
pepper soils is still rooted in classical “text book” knowledge. The review
also includes the relevance of a refreshing new concept, developed by the
author, which is now universally known as “the nutrient buffer power
concept”, in enhancing pepper productivity. q 2004 Academic Press.
I. INTRODUCTION
Known as the “King” of spices, black pepper (Piper nigrum), a perennial crop
of the tropics, is economically the most important and the most widely used spice
crop of the world. The history of spices is very much entwined with the history of
mankind. But, within the family of spices, black pepper predominates. In ancient
Egypt, when the mummified body of the Pharaoh was laid to rest in the Pyramids,
it was black pepper along with gold and silver that was kept adjacent to the body,
in the belief of the ancient Egyptians that even in the after life this very important
spice would be of use. The ancient scriptures, Bible, Koran, and the Vedas
mention the use of spices. According to the Bible, it was during the royal visit of
Queen Sheeba to King Solomon (BC 1015 –BC 66) that a caravan load of spices,
primarily pepper, was presented by the former to the latter. Nearly 3000 years
before the birth of Christ, both Babylonians and Assyrians were trading in spices,
primarily black pepper, with the people of the Malabar coast in the state of Kerala
on the Indian subcontinent. Also, the ancient Indian medical texts, such as
Ashtangahridaya and Samhitas, mention the use of pepper in rare and unique
medical formulations. That spices, and in particular, pepper, had such a lasting
impact on the economic prosperity of places is revealed by the fact that cities like
Alexandria, Genoa, and Venice can trace their economic prosperity back to the
vigorous trade in spices (Rosengarten, 1973). Parry (1969) observed that due to
the increased demand and consumption of pepper in England and Europe, a guild
known as “Pepperers” — the wealthiest of the merchants — was established in
London. The high price of pepper made it the exclusive commodity in use by the
rich for culinary purposes. But, it is interesting to note that with the arrival of
pepper, the western kitchen transformed when “dishes took on a fullness of flavor
previously unknown, beverages glowed with a redolent tang, and life experienced
a new sense of warmth and satisfaction” (Parry, 1969). King Solomon of Israel
and the Phoenician King, Hiram of Tyre, obtained their spices from the coast of
Malabar (Rosengarten, 1973). Though the Jews and Arabs were well into the
274 K. P. P. NAIR
spice trade by this time, it was the latter who had a domineering role in the spice
trade, which was later usurped by the Romans. However, the Arabs were privy to
the knowledge of the country of origin of pepper and even secretly kept the sea
route knowledge to themselves. But, the onslaught of the Roman empire changed
all that during the 1st century A.D. when Rome captured Egypt. It was around
A.D. 40, during the reign of the Roman Emperor Claudius that the enterprising
Greek merchant mariner Hippalus, who after discovering the full power and
velocity of wind movement of the Indian ocean, a secret guarded by the Arabs,
that it became known that a round trip to Egypt via India could be completed in
less than 1 year (Rosengarten, 1973). A consequence of the great economic scale
of this discovery was that when the sea route came to be known, the dependence
on the overland trade routes declined substantially and several wealthy cities on
the overland trade route went into economic penury (Ummer, 1989) and Hippalus
and many others reached the coast of Malabar and returned to Rome with loads of
pepper and other spices. The Roman dominance on the spice trade completely
eclipsed the Arabs. Interestingly, among all the spices, it was pepper that the
Romans fancied most despite its “obnoxious pungency” as noted by Pliny the
Elder in his Natural History compiled between A.D. 23 and A.D. 79. It is
interesting to note that just about the time Arabs were getting involved in the
spice trade with the Malabar coast, the Chinese also entered the fray. As early as
the 1st century A.D., a royal messenger reached the Malabar coast in search of
spices and the trade relationship between China and the Malabar coast began to
flourish as recorded by travelers, such as Sulaiman who reached the Malabar
coast in A.D. 851 (Ummer, 1989). It was the Chinese who played an important
part in spreading pepper to southeast Asia and far-east Asia and it was during
the voyages of Zheng He (1405 –1433) that China imported as much pepper as
the total quantity imported into Europe during the first half of the 10th century
(T’ien, 1981), and the Chinese trade was an imperial monopoly. It is interesting to
note that in the 15th century, soldiers in China were partly paid in pepper, and
even government officials, as observed by T’ien (1981). The fabulous Chinese
ships that carried pepper were greatly praised by the venerable explorer Marco
Polo (Mahindru, 1982).
The colonization of the Indian subcontinent had much to do with the pepper
trade. It was the lure of the spice trade that led Vasco de Gama, the great
Portuguese explorer to discover the sea route to India and he landed on the
Malabar coast at the Kappad beach near the presently named city of Kozhikode
(Calicut) on 20th May 1498. Vasco de Gama and his men returned to Portugal
immensely rich, but in the process also put an end to the Arab trade. The
Portuguese King, Dom Mannuel, was so much impressed with the pepper bounty
that Gama brought back home, that he sent a naval contingent with a fleet of
13 battle ships to India under the command of Pedro Alwarez Cabral in A.D.
1500 and went further to declare sovereignty over India, along with other
countries such as Ethiopia and Arabia (Rosengarten, 1973). The Portuguese
domination of Kerala through pepper production and trade was so complete that
this tiny state of India turned out to be the cradle of world pepper. The Portuguese
rulers were ruthless and their only aim was to make the most out of the pepper
trade. However, the scene changed in the first quarter of the 17th century with the
arrival of the Dutch. The Dutch were temporarily successful in elbowing out the
Portuguese until the British came on the scene in A.D. 1600. In hindsight, what is
most astonishing is how a trade war in pepper between the Dutch and the British
led to the establishment of the British empire on the Indian subcontinent. The
historians Collins and Lapiere in their monumental work “Freedom At Midnight”
(1976) make a remarkable observation as follows: “Sometimes history’s most
grandiose accomplishments have the most banal of origins. Great Britain was
set on the road to the great colonial adventure for five miserable shillings.
They represented the increase in price of a pound of pepper proclaimed by the
Dutch privateers who then controlled the spice trade. Incensed at what they
considered a wholly unwarranted gesture, 24 merchants of the city of London
gathered on the afternoon of 24th September 1599 A.D. in a decrepit building
on Leaden Hall Street. Their purpose was to find a modest trading firm with
an initial capital of £72,000 subscribed by 125 shareholders. Only the simplest of
concerns, i.e., profit, inspired their enterprise, which expanded and transformed,
would ultimately become the most noteworthy creation of the age of
imperialism — the British Raj.” What started as the British East India Company
on 31st December 1600, with the stamp of approval by Queen Elizabeth I, in
just 36 years, in A.D. 1636 from the day of 24th August 1600, when the 500 ton
ship Hector landed in the Surat port, north of Bombay, laid the long and
tortuous road to the subjugation of the vast millions of Indians through the
pepper trade. In more than one sense, pepper was the cause of India losing its
sovereignty.
It was only in the last quarter of the 18th century that the Americans entered
the pepper trade. The first sponsored trip to the East Indies was organized by
Capt. Jonathan Carnes in 1795. Though the American pepper trade flourished
until 1810, it later declined coinciding with the American Civil War in 1861.
Compared with the Portuguese, Dutch and British, the impact of Americans on
the pepper trade was only marginal. The Americans traveled to Sumatra to fetch
pepper. By 1933, pepper was introduced to Brazil, and in 1938, it reached the
Republic of Malagasy. By 1954, pepper was introduced to the African continent.
Within the Indian Republic, the tiny state of Kerala can pride itself as the home of
pepper, in particular, the coastal region of Malabar, in the state of Kerala, which
accounts for 95% of the country’s area and production (Anon, 1997). Besides
Kerala, two other states in the southern region, namely, Tamil Nadu and
Karnataka, put together, contribute the remainder. The first research station of
pepper in the world was established in Kerala in a small town named Panniyur
on the Malabar coast during 1952 –53. In addition, the first hybrid pepper —
Panniyur 1 — in the world was released by this station in 1966 (Fig. 1).
276 K. P. P. NAIR
Figure 1 Panniyur-1, the first Pepper hybrid developed in the world at the Panniyur Pepper
Research Station, in Kerala State, India.
Following research at Panniyur, pepper research began in Sarawak (Malaysia)

in 1955. Following the success of the “green revolution” in India, several
“Co-ordinated Research Projects” catering to the specific needs of individual
crops were established by the Indian Council of Agricultural Research,
headquartered in New Delhi. Thus, the All India Co-ordinated Research Project
on Spices, with an intense focus on pepper, was established in 1971. More than 30
years later the project has included other spice crops as well, yet pepper continues
to receive most attention against the background of the current liberalization and
globalization process and the World Trade Organization (WTO) mandated
changes. The future of pepper is most crucial to the economy of India vis-à-vis
the economy of the state of Kerala, where it is the economic mainstay.
Many other tropical countries have made concerted efforts to grow pepper in
view of its global economic importance, the most important being Indonesia
where the International Pepper Community (IPC) is headquartered in Jakarta.
Indonesia is second to India as a pepper growing country. It was either through
the Polynesian sea-farers or the Babylonian –Chinese sea route linking the
Malabar coast and southeast and far-east Asia that pepper reached Indonesia.
Indonesia, known as the Dutch East Indies during the pre-second world war
period, was the largest pepper producer. It was during the Japanese occupation
during the war that many plantations were abandoned and production declined
sharply. During the pre-war period, Indonesia had in all close to 30,000 ha of
pepper (Lawrence, 1981). Malaysia and Sri Lanka are the other two major
pepper producers. It was the European settlers who introduced pepper to
Malaysia, while in Sri Lanka, pepper was grown as a mixed crop with other
crops like cocoa and the foreign occupation of Sri Lanka helped pepper
cultivation expand in the country. In the southeast Asian countries such as,
Thailand, Vietnam, Cambodia, South Korea, and parts of south China, pepper
cultivation took hold in the post-war years. Within the southeast Asian
countries, Vietnam is beginning to emerge as a major pepper grower. Within the
South Pacific islands, Fiji is the most important pepper growing country. In
South America, Brazil is the leader followed by Mexico, Guatemala, Honduras,
Saint Lucia, Costa Rica, and Puerto Rico. Within the African continent,
Madagascar leads the pack followed by Malawi, Zimbabwe, Benin, Kenya,
Nigeria, Cameroon, Congo, Ethiopia, etc.
Within Asia, where pepper production is concentrated, Malaysia takes the
third place next to Indonesia, with an Agricultural Research Station in
Kutching, Sarawak. Indonesia, which takes the second place, has the Research
Institute for Spice and Medicinal Plants in Bogor. Though India tops the list
among the producers in acreage, with a total area of 191,426 ha and a
production of 56,200 metric tons, its productivity is the lowest in the world
with just 294 kg ha21, while Thailand with a total area of only 2808 ha tops
the list with a productivity of 3594 kg ha21 (Table I). The International Trade
Centre (ITC) in Geneva, estimates the current trade in spices at 400,000 –
450,000 metric tons with a total value of US $1.5 –2 billion annually. With an
annual growth rate of 3.6% in quantity and 8.4% in value in spices, pepper
contributes 34% of total trade in spices. Within the industrialized west,
Denmark tops the list in pepper consumption, followed closely by Germany
and Belgium. The USA is a sizable consumer, while Canada and Switzerland
tail the list.
Post-Doha negotiations, agriculture will increasingly play a crucial role in
world economy. Among the spices, pepper will play a major role. By 2010,
Table I
Acreage, Production and Productivity of Major Pepper Growing Countries in the Last Decade
of the 20th Century
Country Area (ha) Production (metric tons) Productivity (kg ha21)
Brazil 26,500 23,400 883

China 13,170 11,045 839
India 191,426 56,200 294
Indonesia 110,580 45,240 409
Madagascar 4228 2160 511
Malaysia 8960 16,920 1888
Mexico 1294 1112 859
Sri Lanka 12,080 5058 419
Thailand 2808 10,091 3594
Vietnam 15,700 17,266 1100
278 K. P. P. NAIR
projected world consumption will reach 230,000 metric tons, which scales up
to 280,000 metric tons by 2020, which means an annual increase of 5000
metric tons. Present production is close to 200,000 metric tons. For the next
two decades close to 100,000 metric tons will be needed to balance the
projected demand and consumption. World wide, especially in the industri-
alized countries, there is a growing demand for premium organically grown
pepper. The potential for the organic food market is close to US $8 billion
now in USA, which is followed by Germany and Japan, each with a market
share of close to US $2.5 billion. A substantial part of this market will be for
pepper. This would imply that future production strategies would need to
increasingly focus on clean pepper production, which has to withstand both
biotic and abiotic stresses without recourse to “high input chemical
technology” — the hall mark of the so-called “green revolution”. There are
areas of pepper production that simultaneously pose great challenges, while
opening up new avenues. One of the most daunting, in the former category, is
the evolution of a totally resistant pepper variety to the dreaded disease “Foot
Rot”, caused by the fungus Phytophthora, which has wiped out many a pepper
plantation. Also, pepper nutrition is still far from being thoroughly understood.
The fact that it is a perennial crop adds to the lack of thorough understanding.
Despite the complexity of soil science and the emergent soil management
practices, the basic concept of soil as a medium of plant growth can be
expected to persist for an indefinite length of time (Nair, 1996). But, it is
becoming clearer that the earlier views on soil as merely a “supportive
medium” for plant growth is giving place to new ones on “managerial
concepts” of this supportive medium. This is amply illustrated by the shift in
focus from the green revolution phase of the 1960s to mid-1970s where
application of increasing quantities of soil inputs, such as, fertilizers and
pesticides, was emphasized, to the “sustainable agriculture” phase from the
early 1980s to the present (probably to continue?); sustainable agriculture
places more reliance on biological processes by adopting genotypes to adverse
soil conditions, enhancing soil biological activity and optimizing nutrient
cycling to minimize external inputs, such as fertilizers, and maximize their
efficiency of use. In fact, the paradigm of the earlier phase has given way to
the emergent new paradigm (Sanchez, 1994), and this is clearly reflected in the
dialogue of the world leaders during the Earth Summit in 1991 in Rio de
Janeiro, Brazil, where Agenda 21 has incorporated six chapters on soil
management (Keating, 1993). This review on pepper, while discussing its
overall production profile in the world, will lay a special emphasis on the
second paradigm inasmuch as prescriptive soil management for pepper
production is concerned with regard to understanding the soil nutrient
bioavailability and its efficient management in the pepper production. On
account of the paucity of published literature on this aspect, the focus will
only be with regard to specific nutrients, such as zinc, which is becoming

increasingly important in pepper production.
II. THE PEPPER PLANT — ITS BOTANY

AND CHEMISTRY
A. PEPPER BOTANY
Black pepper (P. nigrum) is a perennial plant and derives its name Piper,
perhaps, from the Greek name for black pepper, Piperi (Rosengarten, 1973) and
most of the European names for black pepper were derived from the ancient
Indian language, Sanskrit, such as Pippali, the name for long pepper (P. longum).
It was the great botanist Linnaeus (1753) who established the genus Piper in his
Species Plantarum. In this monumental work, Linnaeus recognized 17 species in
the Piper family, all of which were included in the same genus. Ruiz and Pavon in
1794 introduced the second genus in the family, namely, Peperomia (Trelease
and Yuncker, 1950). The family name Piperaceae was used for the first time in
1815 by L.C. Rich in Humboldt, Bonpland, and Kunth’s Nova Genera et Species
Plantarum (Yuncker, 1958). All the species known in the family Piperaceae
during the early years of systematic classification were included in the classical
monographic study Systema Piperacearum published in 1843 by F.A.W. Miquel.
It was Rheede in the year 1678 who made the earliest record of the description
of Piper in the Indian subcontinent. In his Hortus Indicus Malabaricus, the
earliest printed document of plants on the Malabar coast, Rheede described five
types of wild pepper that included both black and long pepper. Linnaeus (1753)
included 17 species from India in his monumental work Species Plantarum. The
first major study of the Piper spp. from the Indian subcontinent was that of
Hooker (1886) in his book Flora of British India. However, the most
authoritative floristic study of the Western Ghats of southern India was that of
Gamble in 1925 in his book Flora of Presidency of Madras in which the
following 13 species with their taxonomic keys were described: P. argyrophyl-
lum, P. attenuatum, P. barberi, P. brachystachyum, P. galeatum, P. hapnium, P.
hookeri, P. hymenophyllum, P. longum, P. nigrum, P. schmidtii, P. trichostach-
yon, and P. wightii. Following the publication of Gamble, no new additions to the
list were made until 1981, when in this year a new species P. bababudani, from
the Bababudin hills of the state of Karnataka (the neighboring state of Kerala)
was added (Rahiman, 1981a); however, this was not published. Ravindran et al.
(1987) reported a new species, P. silentvalleyensis, the only bisexual wild species
from the world-renowned “Silent Valley” in the Western Ghats. Other new
species are P. pseudonigrum (Velayudhan and Amalraj, 1992), and P. sugandhi
(Nirmal Babu et al., 1993a).
280 K. P. P. NAIR
Black pepper cultivars could possibly have originated from wild ones through
the process of domestication and selection. More than a hundred cultivars are
known, but most have vanished due to the onset of devastating diseases like Foot
Rot and also replacement by hybrids. Human migration has contributed to the
spread of these cultivars. Pepper distribution is extensive in moist evergreen
forests and to a lesser extent in semi-evergreen and moist deciduous forests of the
Western Ghats of southern India, growing from almost sea level to a height of
1500 m above mean sea level. Population structure of any species is determined
mainly by the breeding system of the species, pollen, fruit, and seed dispersal
mechanism and the presence or absence of isolation mechanisms. In Piper, male,
female, and hermaphrodite forms exist. The cultivated P. nigrum is monoecious,
having hermaphrodite flowers, while the wild ones are mostly dioecious. Human
selection has played a major role in the directional evolution of hermaphroditism
in the cultivated pepper. Predominantly self-pollinated, pepper pollen dispersal is
aided by rain or dew drops and also by geitonogamy — the gravitational descent
of pollen. Though flowers are protogynous, in the absence of an active pollen
transfer mechanism, protogyny is ineffective in outbreeding. Active and efficient
pollen and seed dispersal mechanisms ensure gene flow within and between
population segments leading to the establishment of intergrading populations.
Absence of such a mechanism in Piper ensures effective isolation barriers among
individuals and population units. Segregation in the seedling progenies leads to
variations in such units, and also, mutations and chance crossing followed by
segregation. Any such variation stemming in population is fixed immediately
because of the vegetative mode of propagation and such a unit may over the
course of time, diverge from other similar units. Quite often different types of
P. nigrum or different species climb up a single tree, which enhances the chances
of outcrossing, which results in hybrid seedlings. Progenies from such chance
crosses grow, and later climb up the same or nearby trees and chance outcrossing
with the parental vine or its clonal or other seedling progenies, which result in
further back crossing or hybrid progenies, all of which lead to substantial
variation within the population. These forces acting together would have most
likely contributed to the evolution of many present day pepper cultivars
(Ravindran et al., 1990).
Though pepper has originated in the evergreen forests of the Western Ghats in
southern India, with extensive occurrence of wild pepper in the less disturbed
forest areas, no study has so far been carried out to find out the origin of P. nigrum.
The basic chromosome number of Piper is x ¼ 13 and that of P. nigrum 2n ¼ 52,
a tetraploid (Mathew, 1958, 1972; Jose and Sharma, 1984). On the basis of
morphological and biosystematic studies, Ravindran (1991) reported that three
species, namely, P. wightii, P. galeatum, and P. trichostachyon, could be the
putative parents of P. nigrum. These three are woody climbers with similar
texture and leaf morphology. Their spikes and fruits are more similar to that of
P. nigrum than those of other species. Of the three, P. wightii and P. galeatum are
the most probable ancestors of P. nigrum.
Pepper is a shade loving plant that requires a constant moisture supply
during dry spells as it has a high evapotranspiration coefficient (Raj, 1978).
However, even under favorable soil moisture conditions, when exposed to
direct sunlight the plant develops certain physiological disorders (Vijayakumar
et al., 1984). Pepper is highly sensitive to the growth-light regime (GLR) and
plant parts exposed to high GLR produce more fruits per unit surface area than
those exposed to low GLR (Montaya et al., 1961). Where grown in
permanently shaded situations productivity is poor. Under such conditions,
plants respond positively to radiation and a positive correlation has been
shown to exist between productivity and radiation. Except in the case of “bush
pepper”, normal pepper vines are grown on artificial support. These supports
are called “standards”. Ramadasan (1987) reports three types of pepper
canopies. Canopy shaping is largely based on the shading provided by live
support and that of the adjoining shading trees. When dead supports, such as
brick pillars or reinforced concrete poles are used, as in Thailand, the canopy
does not taper at the top. Trees that grow adjacent to dead supports lead to
tapering of the canopy at the base due to partial shading. Such tapering at the
base is not observed when there are no competing trees. In Sarawak, the wood
of Bornean ironwood (Eusideroxylon zwageri) is used for support and
generally, no shade tree is grown. In such dead supports, the top of the canopy
unfolds like an umbrella.
The pepper plant is endowed with two advantages that many others do not
have. First, it can be vegetatively propagated and second, it is viable for sexual
reproduction. Both offer much scope for the exploitation of hybrid vigor as well
as selection breeding. Published literature show that clonal selection, hybridiz-
ation, open pollinated progeny selection, mutation, and polyploidy have been
successfully experimented with to improve pepper yield and impart desirable
plant traits. In recent times, biotechnological methods are also being employed to
impart pathogen resistance.
Quality is the most important criterion in breeding as far as spices are
concerned and pepper is no exception. The quality of pepper is decided mainly by
the amounts of piperine, and oleoresin (essential oil). An alkaloid, piperine
contributes to the pungency of the berry, and oleoresin content enhances
the flavor. The unique pepper flavor is influenced by very many chemical
compounds present in the essential oil and a number of genes control this unique
character.
Conservation of pepper genetic resources is the key to sound pepper breeding.
Among the 18 biodiversity hot spots on this planet, the Western Ghats of
southern India is one of the most unique, with their different forest types, such as
tropical wet evergreen, tropical moist and dry deciduous, montane subtropical and
temperate ones. It is on the western side of the Western Ghats that the population
282 K. P. P. NAIR
and diversity of Piper is in abundance. It is important to realize that during the

past century the Western Ghats have been ravaged the most, leading to ecological
breakdown, due to clandestine timber felling and uncontrolled human
encroachment. The Western Ghats is home to spices that are endemic to the
region. Of the 17 species of Piper found here, 11 are endemic to the region
(Ravindran and Peter, 1995). More than a hundred cultivars were under
cultivation in the region, but most of them have vanished from the pepper
growers gardens, probably due to such dreaded diseases as Foot Rot
(Phytophthora) and slow decline and also, replacement of old and unproductive
cultivars by the few high yielding elite ones. It takes millennia for nature to
nurture plant diversity, but human activity, productive or otherwise, can change
all that in a very short span of time. As part of the conservation efforts, some
pepper growing countries, notably India, has made a concerted effort to collect
and catalogue pepper germplasm in a systematic manner. Mention should be
made of the efforts in this direction undertaken at the Indian Institute of Spices
Research at Kozhikode, Kerala State. The Institute has established a National
Conservatory for in situ conservation. Conservation is carried out in four stages
(Ravindran and Nirmal Babu, 1994a) as follows: conservation as a nursery gene
bank by trailing each accession in split bamboo pieces of about 2– 3 feet length
serially, which will be maintained and continuously multiplied. Conservation in
the clonal repository where 10 rooted cuttings from each accession are
maintained. Conservation in the field, as field gene bank, by planting accessions
for pilot yield performance and characterization and conservation in vitro and in
cryogen banks. With the establishment of the All India Co-ordinated Research
Project on Spices in 1971, collection and conservation of pepper germplasm on a
smaller scale has also been made by the Panniyur Pepper Research Station (the
first pepper research station established in the world) in the state of Kerala,
Pepper Research Station at Sirsi, in the state of Karnataka, Horticultural Research
Station, Chintapalli, in the state of Andhra Pradesh and Horticultural Research
Station, Yercaud, in the state of Tamil Nadu. Among all of these centers, the last
one has the largest collection, but it is the Indian Institute of Spices Research that
can pride itself as the one holding the largest collection in the world. On a much
smaller scale, collections and conservation have also been made, notably in
Indonesia and Malaysia, with the former holding 40 cultivars and seven Piper
species and Sarwak (Malaysia) holding 18 cultivars of P. nigrum, 18 identified
Piper spp. and 98 unidentified accessions (De Waard, 1984). Despite the fact that
they are not closely related to the cultivated black pepper, the vast number of
Piper spp. found in central and southern America, north eastern India, Indonesia,
and Malaysia have not been collected and systematically catalogued. The vast
germplasms in these regions could form a very important reservoir of gene pool
to develop varieties with in-built disease and pest resistance (Table II).
A new avenue for pepper improvement is through biotechnological research,
which of late has made great strides in many areas of biology. The
Table II
Some Improved Pepper Varieties from Major Pepper Growing Countries
Country of origin and Average yield

name of the variety (kg ha21) Remarks
India
Panniyur 1 1242 Well suited to most regions, performs rather poorly
at higher elevations and under shade
Panniyur 2 2750 Well suited to most regions and shade tolerant
Panniyur 3 1953 Well suited to most regions, rather late in maturity
Panniyur 4 1277 Consistently good yielder suited to most regions
Panniyur 5 1098 Most important characteristic is tolerance to nursery
diseases and shade
Subhakara 2352 Suited to most regions, berries of good quality
Sreekara 2677 Suited to most regions, berries of good quality
Panchami 2828 Suited to most regions, late in maturity
Pournami 2333 Good tolerance to root knot nematode
PLD-2 2475 Suited to most regions, berries of high quality
Indonesia
Natar1 – High yield and particularly tolerant to the dreaded
disease Foot Rot and nematodes
Natar 2 – High yield and particularly tolerant to the dreaded
disease Foot Rot and nematodes
Malaysia
Semongok Perak – High yield
Semongok Emas – Uniquely high yield, tolerant to Foot Rot, Black Berry
disease and pepper weevil
Madagascar
Sel. IV.1 – High yield
Sel. IV.2 – High yield
Sri Lanka
PW 14 – Claimed to be resistant to Radopholus similis
Note: Average yield unknown.
biotechnological approaches permit the researchers to manipulate plant tissues

and cells in vitro. A better understanding of genetics is provided by tissue culture
techniques and investigations at the molecular, cellular, and organismal levels.
The possibilities of cell culture and plant tissue culture for breeding and plant
propagation purposes has led to large-scale commercial exploitation. Technical
perfection by manipulating plant cells at the molecular level through recombinant
DNA technology has opened up enormous possibilities in the creation of
transgenic plants and this is an area that is assuming industrial proportions.
Inasmuch as pepper is concerned, one area where recombinant DNA technology
could be successfully exploited is the evolution of pepper varieties tolerant to the
284 K. P. P. NAIR
dreaded disease Foot Rot caused by Phytophthora capsici. None of the varieties
in cultivation is tolerant to this dreaded disease. Within the Piper spp., a distant
relative of P. nigrum, namely P. colubrinum from South America, is resistant to
Phytophthora and, hence, biotechnological approaches which will enable transfer
of resistance governing genes from P. colubrinum to P. nigrum will be the only
mode to contain the devastation from this dreaded disease.
Another area of interest is micropropagation, where in vitro culture methods
for pepper cloning is done. Micropropagation was first introduced by Broome and
Zimmerman (1978). For micropropagation, several plant parts, such as shoot tips,
nodal segments, and apical meristems, from both juvenile and mature plants have
been used (Mathews and Rao, 1984; Agarwal, 1988; Philip et al., 1992; Nazeem
et al., 1993; Nirmal Babu et al., 1993a,b; Joseph et al., 1996; Lisamma et al.,
1996). However, a serious limitation to micropropagation is the incidence of
bacterial contamination (Kelkar and Krishnamoorthy, 1996), though contami-
nation is minimal in seedling explants. The authors suggest the incorporation of
antibiotics in the culture media to arrest the endogenous bacterial contamination
in the in vitro pepper cultures. Field establishment of micropropagated pepper is
feasible, though endogenous bacterial contamination and phenolic exudates from
the cut surface could adversely affect field establishment (Raj Mohan, 1985;
Fitchet, 1988a,b). The authors suggest treating the explants with fungicides prior
to routine sterilization followed by frequent transfer to fresh medium to keep in
check the problem of contamination. Madhusudhanan and Rahiman (1966)
suggest the use of activated charcoal, at the rate of 200 mg l21, which could
reduce browning of the explants and culture medium. Micropropagation has been
in use for quite some time (Chu, 1981; Fitchet, 1988a,b). Culture media
commonly used in micropropagation are those of Murashige and Skoog (1962),
Schenk and Hildebrandt (1972) and McCown and Amos (1979). Indole acetic
acid (IAA), naphthalene acetic acid (NAA), 2,4,-dichlorophenoxy acetic acid
(2,4,-D) are some the commonly used growth regulators and additives used in the
culture formulation.
One other area of pepper improvement is the Agrobacterium mediated gene
transfer. Reference to the technique is made by Sasikumar and Veluthambi
(1994) and Sim et al. (1995). The technique can also be helpful in transferring
disease resistance genes from P. colubrinum to P. nigrum in the light of the
dreaded disease Foot Rot. Work on transgenic pepper to evolve delayed ripening
ensuring uniform maturity — which will ensure reduction in labor costs for
harvest — in Sarawak (Malaysia) has led to promising results.
Conservation of pepper germplasm assumes high practical importance in
improving the crop for enhanced productivity and acquisition of many
desirable traits, the most important being tolerance to dreaded diseases such
as Phytophthora caused Foot Rot. Normal conservation of crop germplasm is
through seed banks. Since pepper is propagated only vegetatively (seeds are
heterozygous and unpredictable in behavior) in vitro storage of germplasm is
a viable alternative. In vitro conservation of pepper germplasm and its related

species, such as P. barberi, P. colubrinum, P. beetle, and P. longum by
defining protocols have been made by Geetha et al. (1995) and Nirmal Babu
et al. (1996b) by maintaining cultures at reduced temperatures in the
presence of osmotic inhibitors at low nutrient levels and also by reducing to
the minimum evaporative losses using closed containers. When the conserved
materials were transferred to multiplication medium after storage, all the
species exhibited normal multiplication rates. When the normal-sized plantlets
were transplanted into field soil, an establishment rate exceeding 80% was
obtained. Normal pepper plants, morphologically similar to mother plants,
were observed to develop.
There is a steep decline in viability in pepper seeds associated with moisture
depletion. Seeds cryopreserved in liquid nitrogen at 2 1968C with an initial
moisture percentage of 12 have a survival rate of only 45%, while at 6% moisture
it reduced to just 10.5% (Chaudhary and Chandel, 1994) showing clearly how
seed moisture is so crucial to seed survival. There is yet another manner in which
pepper plants can be propagated and that is by way of producing synthetic seeds,
which consist of somatic embryos or shoot buds encapsulated in a protective
coating that is biodegradable. The system facilitates not only low-cost
propagation but also germplasm conservation and exchange. Encapsulation of
disease-free bud employing tissue culture techniques helps production of disease-
free plantlets. Pepper shoot buds 0.5 cm long were used for production of
synthetic seeds, which could be stored up to nine months in sterile water.
The method is described by Sajina et al. (1996).
B. PEPPER CHEMISTRY
It is the unique pungency and aroma of pepper that has both intrigued as well
as fascinated pepper chemists. As already mentioned in the review earlier, the
essential oil in the berry contributes to the aroma while the alkaloid piperine
imparts the unique pungency. Oleoresin, which has a very great commercial
value, is extracted from the dry powdered berries by solvents, and is the
product that imparts the unique aroma and pungency in pepper. Dramatically
put, it is the chemistry of pepper emanating from oleoresin that imparts that rare
and unique blend of both aroma and flavor, which is behind wars and
enslavement, as much as love and tragedy, as has been detailed in the
introductory part of this review. The earliest investigations on pepper chemistry
go back more than half a century (Guenther, 1950) and subsequent research
has enriched it (Wealth of India, 1969; Govindarajan, 1977; Lawrence, 1981;
Purseglove et al., 1981).
The various compounds occurring in Piper spp. are listed by Parmar et al.
(1977). It was in the early 19th century that the first report on essential oil in
286 K. P. P. NAIR
pepper was made by Dumas and later by Subeiran and Capitaine (Guenther,
1950). These researchers concluded that pepper oil is almost free of oxygenated
constituents. By treating a fraction of the oil boiling at 1768C with acid and
alcohol, Eberhardt obtained terpin hydrate and the presence of 1-phellendrene,
caryophyllene, and tentatively dipentene were reported by Schimmel and
Company and by Schreiner and Kremers (Guenther, 1950). The steam distillation
of essential oil obtained from dry powdered Malabar pepper berries showed the
presence of a-pinene, b-pinene, 1-a-phellendrene, DL -limonene, piperonal,
dihydrocarveol (a compound melting at 1618C), b-caryophyllene and a
piperidine complex (Hasselstrom et al., 1957). Additionally, the presence of
cryptone, epoxydihydrocaryophyllene, and possibly, citronellol and an azulene
were also reported. The presence of a- and b-pinenes, limonene, and
caryophyllene in the hydrocarbon part of black pepper oil was confirmed by
infrared spectroscopy (Jennings and Wrolstad, 1961).
It is important to mention that a renewed thrust was given to the study of
chemical compounds in pepper after the advent of gas chromatography. Modern
researchers used thin layer chromatography, gas chromatography, column
chromatography, and vacuum distillation, etc., to separate the constituents and
employed ultra violet, infra red, nuclear magnetic resonance, and mass
spectroscopy for identification.
As many as 135 compounds, consisting of monoterpenoids, sesquiterpenoids,
aliphatic, aromatic, and those of miscellaneous nature, have been reported (Ikeda
et al., 1962; Sharma et al., 1962; Nigam and Handa, 1964; Wrolstad and Jennings,
1965; Muller and Jennings, 1967; Muller et al., 1968; Richard and Jennings, 1971;
Russel and Else, 1973; Debrauwere and Verzele, 1975a,b, 1976; Lawrence, 1981;
Gopalakrishnan et al., 1993).
Different researchers have reported wide variations in the chemical
composition of essential oils in pepper and these variations originate on account
of several reasons, such as varietal differences, their geographic origin,
variations in the maturity of raw material, procedural differences in oil extraction
and non-resolution of constituents in early gas chromatographic analysis
employing packed columns. In general, composition of essential oils will depend
to a certain extent on the method of preparation, as for example, steam distillation
will give oils containing about 70– 80% monoterpene hydrocarbons, 20– 30%
sesquiterpene hydrocarbons and less than 4% oxygenated constituents. Vacuum
distilled oils will contain less monoterpene hydrocarbons and more sesquiterpene
hydrocarbons and oxygenated constituents. It is the result of incomplete
distillation and the poor recovery of the high boiling sesquiterpene hydrocarbons
and oxygenated constituents which lead to their low contents in steam distilled
oils compared with vacuum distilled oils. Seventeen pepper cultivars from the
state of Kerala were analyzed for their essential oil contents by Lewis et al. (1969)
and Richard et al. (1971) and comparable results were obtained. Gopalakrishnan
et al. (1993) analyzed the four Panniyur genotypes (Panniyur-1, Panniyur-2,
Panniyur-3, and Panniyur-4) developed at the Pepper Research Station in

Panniyur using a combination of gas chromatography and mass spectroscopy and
Kovats indices on a methyl silicone capillary column and found, on the whole, that
Panniyur-4 contained the most chemical constituents as compared with the other
three genotypes. Significant differences could also be found in 12 samples from
Lampong and 16 samples from Sarawak, both in Malaysia, using the procedure of
Richard et al. (1971) by Rusel and Else (1973).
An interesting aspect of research refers to sensory evaluation of essential oils
in pepper. There are but scanty references with regard to the relationship between
odor characteristics and oil composition. The characteristic odor of pepper oil has
been attributed to the meager amounts of oxygenated constituents (Hasselstrom
et al., 1957). Arctander (Purseglove et al., 1981) describes the odor of pepper oil
as fresh, dry –woody, warm – spicy, and identical to that of black pepper corn.
Pangburn et al. (1970) after a systematic and thorough sensory evaluation of
pepper oil obtained from pepper along the Malabar coast, using column
chromatographic fractions and mixtures of fractions of the oil, reported that the
early fraction was considered pepper-like and floral, the late fraction pepper-like,
fresh and woody and the middle fraction remaining in between. Direct sniffing at
the eluting port of the gas chromatographic columns helped distinguish the
distinct odor of black pepper and ascribe it to the three areas of the late fractions.
Distinctive odor analysis has been developed by Harper et al. and Sydow et al.
(Govindarajan, 1977). Using a descriptive odor profile based on a four-point
category scale and subjecting the oils to a ranking list, Gopalakrishnan et al.
(1993) have described the odor evaluation of the four Panniyur genotypes
described earlier in this review.
Besides aroma (odor), the other aspect that has been of interest to the
researchers is the pungency emanating from the alkaloid piperene. From the early
19th century onwards, the pungency of pepper has been investigated starting with
the first report of Oersted who isolated piperine in 1819 (Guenther, 1950).
Piperene is a yellow crystalline substance that was subsequently identified as the
trans form of piperoyl piperidine. Subsequent researchers showed that piperine
was not the only substance imparting pungency to pepper and Bucheim
(Govindarajan, 1977) obtained a dark oily resin, which he called “chavicine”,
subsequent to the removal of piperine from the oleoresin. Chavicine was
supposedly far more pungent than piperine leaving a much sharper bite on the
tongue than the crystalline piperine. This was shown to be incorrect by subsequent
researchers who demonstrated the extreme pungency of piperine in solution. The
debate, whether it was chavicine or other possible isomers, such as isopiperine or
isochavicine, which are more pungent than piperine, continued for almost a
century. However, recent investigations reveal that it is piperine which is the most
pungent and chavicine is a mixture of piperine and many other minor alkaloids.
The question why pepper blackens has interested many researchers. Lewis
et al. (1976) attributed blackening of pepper berries due to enzymatic oxidation
288 K. P. P. NAIR
of polyphenolic substrates present in the skin of green pepper. The blackening of

pepper is a chemical reaction akin to browning in fruits and vegetables and can
have an enzymatic or non-enzymatic origin. Formation of coloring pigments
because of enzymatic browning is triggered by the enzyme polyphenol oxidase.
It is due to the enzymatic oxidation of polyphenolic substrates in the skin of green
pepper that leads to blackening (Lewis et al., 1976). The authors developed a
process in which green pepper was blanched to arrest the enzyme reaction with
subsequent drying in a cross flow drier to obtain dehydrated green pepper. Pruthi
et al. (1976) reported that green pepper could also be preserved in brine
containing either acetic or citric acid.
In addition to the above chemical constituents, starch is a predominant
constituent of pepper and its content may vary between 35 and 40% in terms of
weight. In addition to starch, pepper also contains protein and fat. Reports on
both starch and protein are practically non-existent. As for fat, pepper contains
1.9 – 9% fat. Bedi et al. (1971) and Salzer (1975) have determined the fatty acid
composition of pepper and reported that of the different fatty acids, such as
palmitic, oleic, linoleic, and linolenic, it is the linoleic acid that is most abundant
(25 –35%) and the least is linolenic (8 –19%).
The survey of pepper chemistry shows that though some of the major
constituents have been researched during the last few decades, in the future, one
may still come across constituents hitherto unknown. The research so far has
mostly been confined to the varieties in cultivation in different regions, world
wide, and yet, there are far more cultivars that remain totally out of the chemists’
laboratory. It is important to research these, because one may come across unique
chemical compositions and flavor profiles in the different cultivars. As with
disease resistance, unique chemical constitution is also of immense commercial
importance. Only when pepper breeders, biotechnologists, and pepper chemists
join hands in an intense search for unique traits will the research on pepper
chemistry have any meaningful outcome, because it is only through these intense
efforts that unique pepper lines could be identified and isolated for future
breeding and crop improvement programs.
III. PEPPER AGRONOMY
Agriculture systems differ from natural systems in one fundamental aspect:

while there is a net outflow of nutrients by crop harvests from soils in the first, there
is no such thing in the second (Sanchez, 1994). This is because nutrient losses due
to physical effects of soil and water erosion are continually replenished by
weathering of primary minerals or atmospheric deposition. Hence, the crucial
element of sustainability is the nutrient factor. But, of all the factors, the nutrient
factor is the least resilient (Fresco and Kroonenberg, 1992). The thrust of high
input technology, the hallmark of the “green revolution”, in retrospect, or the

moderation by low input technology, the foundation stone of sustainable
agriculture, in prospect, both dwell on this least-resilient nutrient factor. If
the pool of nutrients in the soil, both native and added, could be considered as the
“capital”, efficient nutrient management might be analogous to raising the
“interest” accrued from this capital in such a way that there is no great danger of
erosion of this capital. Hence, sound prescriptive soil management should aim at
understanding the actual link between the capital and the interest so that
meaningful practices can be prescribed (Nair, 1996). This is all the more important
for a perennial crop like pepper whose life span can be upwards of 25 years. An
effort in this direction was initiated by Nair recently (Nair, 2002). But, before we
dwell on it, it is only fair that an objective review is made of what has been done so
far on the agronomy of black pepper.
A. THE PEPPER SOILS
Pepper easily grows on a variety of soils. De Waard (1969) reported that in

Malaysia pepper mostly grows on soils that are developed from slate or sandstone
and even on soils of alluvial origin low in fertility. Pepper is seen to grow in
Indonesia on all types of soils ranging from fertile and friable volcanic soils to
sticky clayey soils. In Sri Lanka, pepper growers prefer red clay loam or sandy
loam. In India, pepper is seen to grow on a variety of soils, but, in its native
Kerala state, it is grown mostly on red laterites. Pepper prefers a well-drained
soil having adequate water holding capacity, rich in humus and essential
plant nutrients. In India, the major pepper growing soils are Oxisols (6%),
Alfisols (70%), Mollisols (10%), and Entisols (4%). The state-wise distribution
is given below (Sadanandan, 1994) (Table III).
Table III
Type, Order, and Sub Order of Major Pepper Growing Soils and Their State-Wise Distribution
in India
Soil type Order Suborder Indian state
Forest loam Mollisols Udolls Kerala, Karnataka, and Tamil Nadu

Ustolls
Laterite Alfisols Ustalfs Kerala, Karnataka, and Tamil Nadu
Oxisols
Alluvium Entisols Ustorthcut Kerala and Karnataka
Red loam Alfisols Ustalfs Kerala and Karnataka
Ustults
290 K. P. P. NAIR
Forest loam is acidic in reaction, with a pH range of 5.0– 5.5 and is confined
mainly to the western Ghats Highlands. The soils are rich in organic carbon, well
drained, brown to black in color and are very well suited for pepper. Soils, in
general, are quite fertile, rich in N and K, medium in P. Laterites are, again,
acidic, with a pH range of 5.0 –6.2, low in fertility and invariably run into
problems with P fertilization because of high amounts of soluble Fe and Al,
which render applied P immobile. Excessive soluble Al leads to Al toxicity.
Of late micronutrient deficiencies, in particular, Zn, have been reported and
recently Nair (2002) has demonstrated the lacunae in Zn nutrition of pepper based
on classical textbook knowledge of Zn fertilization in these problematic soils.
Alluvium is confined to the river banks and their tributaries. The soils are only
moderately fertile and respond well to management. The soils are acidic in
nature, with the pH ranging from 5.0 to 6.5. Red loams are highly porous and
friable, low in fertility and like laterites contain high amounts of soluble Al and
Fe which lead to problems with P fertilization. Except in forest loam, where
pepper is grown as a monocrop, as the soil is highly suitable for the crop, in all the
others it is grown both singly or in association with other crops, such as coconut
(Cocus nucifera), arecanut (Areca catachu), etc. Unlike the large exclusive
pepper plantations, these soils support what is generally known as “homestead
farming” in Kerala — small, less than 5 ha, self-supporting family farming,
where even milch animals are reared in association.
B. NUTRITION OF BLACK PEPPER
The fundamental difference between agricultural systems and natural systems

is that while there is a net outflow of plant nutrients by crop harvests from soils in
the former, there is no such thing in the latter. This is because nutrient losses due to
physical effects of soil and water erosion are continually replenished by
weathering of primary minerals or atmospheric deposition in natural systems.
Hence, the crucial element of sustainability of crop production is the nutrient
factor. But, of all the factors, the nutrient factor is the least resilient to
management. The thrust of the “high input technology”, the hallmark of the “green
revolution”, in retrospect, or the moderation of the “low input technology”, the
foundation stone of “sustainable agriculture”, in prospect, both dwell on this least-
resilient nutrient factor. For production sustainability, the nutrient factor is the
most crucial. It is becoming increasingly clear that it is not the quantum per se of
a specific nutrient in the soil that is crucial as far as plant requirements are
concerned, but its bioavailability. In the final analysis, it is the plant and plant
alone, which will decide whether a nutrient inherently present in the soil or
externally added to it, will ultimately be “available” to the plant or not (Nair,
2002). Formidable effort has gone into the task of defining this “availability”. It is
the conviction of this author that if sound and accountable prescriptive fertilizer
recommendations have to be made to sustain crop production in the decades to

come, one must have a clear understanding of the dynamics of nutrient availabi-
lity from which accurate predictions can be made and sustainable and accountable
field practices are devised. These considerations particularly are important to
devise accurate fertilizer recommendations for pepper, because, unlike most
other annual or biennial crops, pepper is a perennial one and the utilization
pattern of applied or native nutrients over several years, often running into
decades, could be uniquely different. This part of the review will, at first, dwell on
the available pool of information relating to the general and prevalent mode of
pepper nutrition, and then will go on to the next stage, where it will encompass
the newer vision, which is now known, the world over, as “the nutrient buffer
power concept”.
C. EVOLUTION OF PEPPER MANURING
Pepper has been in cultivation for many decades now in India and elsewhere.
In earlier days, farmers used only meager quantities of organic manures, such as
leaf litter, animal manure (principally cow dung, which is a wide-spread practice
of homestead farming in Kerala even now) or slashed stems and leaves of live
support, such as Erythrina indica. Chemical manuring, through factory-
manufactured fertilizers, is a relatively recent phenomenon. It was the “pepper
boom” — the escalating prices of pepper in world market, especially in the post
second world war period — that enthused the pepper farmers to go in for
artificial manure. It was the spectacular initial yield increases in wheat and rice in
the North Indian regions — the green revolution effect owing to liberal doses of
chemical fertilizers — which prompted the affluent pepper farmers to go in for
factory-produced fertilizers. Though there is now a blacklash — in terms of
degraded soils, drying aquifers and vanishing biodiversity, all of which are
attributed to the indiscriminate use of chemical fertilizers — against the use for
chemical fertilizers, chemical fertilizer use in a systematic manner is still
confined only to affluent owners of pepper plantations. Organic manuring is still
the mode of pepper nutrition in homestead farming. Significantly, even pepper
planters of fairly large farms are gradually switching to organic farming, because,
“organic pepper” raised by the use of only organic manures fetches a much better
price and a steady and premium market is developing, especially in European
countries such as Germany, where the spices are much fancied lately.
D. RESPONSE OF PEPPER TO MINERAL NUTRIENTS
The major emphasis on pepper fertilization is still confined to nitrogen,

phosphorus, and potassium. There is an extensive body of literature on nitrogen
292 K. P. P. NAIR
fertilization in pepper globally (De Waard, 1964; Sim, 1971; Pillai and
Sasikumaran, 1976; Pillai et al., 1979, 1987; Mohanakumaran and Cheeran,
1981; Sadanandan, 1994). These routine studies have dealt with partitioning of N
in various plant parts (Adzemi et al., 1993), response functions ( Pillai et al., 1987;
Sadanandan, 1990,), greenhouse cultivation (Murni and Faodji, 1990) and foliar
N application (Anon, 1995). Adzemi et al. (1993) found the maximum N
concentration in leaves (2.30%) while the branches transported the most
(47.6 mg plant21 year21). In Kerala state, for Panniyur-1, the first hybrid
variety released in the world from the Pepper Research Station, Panniyur,
50 Kg N with 100 kg P2O5 and 150 kg K2O ha21 were found to be the
optimum rates for laterite soil (Pillai et al., 1987). In greenhouse conditions 1.1 g N
plant21 as urea and 2.0 g K2O plant21 produced the maximum of dry matter
(Murni and Faodji, 1990). Between Panniyur-1 and Karimunda, another
improved variety from the Panniyur Research Station, 292 kg ha21 year21 and
183 kg ha21 year21 of N, respectively, were found to be the optimum rates
(Sadanandan, 1990). In Sarawak (Malaysia), foliar application of 0.7% urea at
weekly intervals totaling nine sprays were found to increase pepper yield by 22%
(Anon, 1995) (Table IV).
Comparatively, research on P in pepper is less, viewed against that of
N. Unlike N, pepper stems contain the most P (3.96 g vine21) because of large
transport and a total of 22.8 kg ha21 of P is removed from the soil (Adzemi et al.,
1993). Availability of P in laterite soil is a universal problem, because of the
presence of excessive soluble Al and Fe, which render the applied P fertilizer
immobile and some researchers have focused on alternate sources of P as a
possible remedy. Phosphate rocks (PRs) are ideally suited for plantation crops
and where indigenously available could be profitably utilized. In northern India,
a popular PR is “Mussorie Rock Phosphate” (MRP). Sadanandan (1994) reported
that MRP is superior to ordinary super phosphate. Both cumulative yield increase
and relative agronomic effectiveness were superior in MRP as compared to
ordinary super phosphate.
Pepper is a prolific user of K and even with 2% content of K in the pepper
plant, K deficiency will manifest (De Waard, 1969). Field trials conducted using
Panniyur-1 as the test crop showed that as much as 200 g of K2O vine21 is
required (Pillai et al., 1987). Response functions to K fertilizer application
indicated that pepper needs a very high dose of K (270 kg K2O ha21) for
high yield (Sadanandan, 1990). Pepper plants remove as much as 203.2 kg ha21
of K and white pepper stores most of it (42.0 g vine21) as reported by Adzemi
et al. (1993). Most pepper growers adopt varying N:P:K ratios, while Pillai
et al. (1979) suggest an optimum ratio of 5:5:10 for N, P and K to obtain the
maximum yield.
As far as the secondary nutrients (Ca, Mg, and S) are concerned, the only
published work refers to Ca, but, indirectly through the effect of liming. Pillai et al.
(1979), however, did not observe any positive response to lime application in a
Table IV
Nutrient Deficiency Symptoms in Pepper
Nutrient Deficiency symptoms Cited by
Nitrogen General chlorosis and stunted growth. De Waard (1969), Nybe

Yellowing of older leaves followed and Nair (1986)
by younger ones. Bottom leaf tips
and margins become brown and
necrotic
Phosphorus Bronzing of older leaves accompanied De Waard (1969), Nybe
by necrosis of leaf tips and margins. and Nair (1986)
Stunted growth
Potassium Browning and necrosis of older leaf De Waard (1969), Nybe
tips and margins. Symptoms later and Nair (1986)
spread to younger leaves
Calcium Young leaves develop tiny, brown De Waard (1969), Nybe and
necrotic pin-head spots which Nair (1987a)
later spead to older leaves.
Interveinal chlorosis can also
be seen
Magnesium Pale yellow discoloration of leaf De Waard (1969), Nybe and
margins and tips, followed by Nair (1987a)
necrosis and defoliation. Major
veins remain green and laterals
turn yellow
Sulfur Late stage chlorosis in younger leaves, Chin et al. (1993), Nybe
turning to bright yellow color in and Nair (1987a)
interveinal areas. Premature
leaf fall and die back of
growing tip
Zinc Stunted growth, small leaves, Chin et al. (1993), Nybe and
interveinal chlorosis. Nair (1987c)
Leaf margins pucker
Manganese Interveinal chlorosis, with major Chin et al. (1993), Nybe and
veins remaining green. Chlorotic Nair (1987b)
leaves turn yellow or white later
and necrotic mature leaves.
Symptoms can often be
confused with that of
Mg deficiency
Iron Interveinal chlorosis in younger Chin et al. (1993), Nybe and
leaves, youngest leaves becoming Nair (1987b)
totally chlorotic
Copper Interveinal chlorosis in younger Chin et al. (1993), Nybe and
leaves, necrosis on leaf Nair (1987b)
tips and margins
Boron Stunted growth, necrosis and Chin et al. (1993), Nybe and
interveinal chlorosis. Necrotic Nair (1987c)
lesions are seen on main vein
294 K. P. P. NAIR
laterite soil growing Panniyur-1. Adzemi et al. (1993) report the highest
accumulation of both Ca and Mg in stem (11.5 and 5.98 g vine21, respectively)
with a total accumulation of 54.5 and 36.4 kg ha21, respectively.
Among the micronutrients, Zinc is the most important followed by Mo and B
(De Waard, 1969). Zn deficiency is beginning to be seen in many tropical
countries in a very significant manner adversely affecting crop yields. However,
to date, corrective measures still bank on routine soil testing with DTPA as the
most commonly used extractant. That such an approach does not lead to
satisfactory results uniformly is shown through the work of this author (Nair,
2002), which will be discussed subsequently in this section. A combination of Zn,
Mo, and B, along with N, P, and K was shown to result in the highest yield of two
locally developed varieties, namely, Subhakara and Sreekara at the Indian
Institute of Spices Research experimental farm (IISR, 1977) where Zn, B, and Mo
were used at a ratio of 5:2:1. Spraying a solution of Zn SO4 (0.5%) reduced spike
(inflorescence which eventually develops into pepper berries) shedding by 48.4%
(Geetha and Nair, 1990).
Organic manuring is a very commonly adopted practice in pepper production
in India and part of Asia. This can either be through the use of fresh vegetative
matter or through the use of “burnt earth” (Harden and White, 1934; Bergman,
1940). In the former category, freshly chopped materials, such as, leaves, stems,
etc., from a number of trees are used. The trees generally used are E. indica,
Garuga pinnata, Grevillea robusta, etc. (Sivakumar and Wahid, 1994). Organic
manures are widely used in pepper production in Sarawak (Malaysia), which
include soybean cake, guano, prawn, and fish refuse (Purseglove et al., 1981).
Holes are dug and about 85 g of manure are placed about 20 cm away from the
main stem of the vine. This is done once in two months. Some farmers dig
trenches all around the vines and put large quantities of manure once in two
months. Of late, sterilized animal meat and bone meal admixture, fortified with
potassium, are gaining popularity as an organic manure.
Another method of organic manuring is through the use of “burnt earth”. For
this, forests are first cleared and on top of the heaped vegetation soil from vacant
patches is spread and the heap is set fire to and the fire is on in a slow burning
fashion for two to three weeks.
At the end of this period, both the wood ash and burnt soil are thoroughly
mixed and applied at a rate of 18 kg vine21 year21. The application of burnt ash
has many beneficial effects in improving the physical, chemical, and biological
properties of soil. Compared to the acidic soil, having a pH of 4– 5, the burnt
earth has a pH of 7 –8 and acts as a good soil ameliorant in correcting pH. There is
extensive literature on this method (Harden and White, 1934; Bergman, 1940;
Huitema, 1941; De Waard, 1969). Owing to environmental hazards, especially
because of large-scale burning and the emission of smoke, the Government of
Sarawak banned the practice in 1940. Figures 2– 4 represent K, Mg, and Zn
deficiency, respectively.
Figure 2 K deficiency symptom.
Figure 3 Mg deficiency symptom.
Figure 4 Zinc deficiency symptom.

296 K. P. P. NAIR
IV. THE ROLE OF “THE NUTRIENT BUFFER POWER

CONCEPT” IN PEPPER NUTRITION
Historically, soil testing has been used to quantify availability of essential

plant nutrients to field grown crops. However, contemporary soil tests are based
on philosophies and procedures developed several decades ago without
significant changes in their general approach. For a soil test to be accurate, one
needs to clearly understand the physico-chemico-physiologic processes at the
soil –root interface, and an understanding of soils and plant root systems as
polycationic systems is as essential. It is this knowledge that leads to sound
prescriptive soil management practices in nutrient availability vis-à-vis fertilizer
application, because, of all the factors that govern sustainability of crop
production, the nutrient factor is the most important, and it is also the least
resilient to management. This section of the review will focus on the buffering of
plant nutrients, with specific reference to Zn, and discuss experimental results
that relate to pepper nutrition.
A. THE BUFFER POWER AND EFFECT ON NUTRIENT

AVAILABILITY
Before being able to understand the significance of the nutrient buffer power
concept on plant nutrient availability, certain basic concepts must be addressed
and the following review starts with this rationale.
B. BASIC CONCEPTS
In any nutrient management approach that is sound and reproducible, one

must start with a basic understanding of the chemical environment of plant roots.
When we consider this, the first term that we come across is the “soil solution”
because the plant root is bathed in it and is most affected by its chemical
properties. The Soil Science Society of America (1965) defines soil solution as
“the aqueous liquid phase of the soil and its solutes consisting of ions dissociated
from the surfaces of the soil particles and of other soluble materials”. Adams
(1971) has given a simple definition: “The soil solution is the aqueous component
of a soil at field moisture content”. Perhaps it is important to emphasize here that
much contemporary soil testing has considered a soil extract as synonymous with
the soil solution. Since soil extraction is supposed to simulate plant root
extraction, it is pertinent to consider the chemical environment of the root, though
briefly, from this angle. It is worth noting that the chemical environment of roots
in natural soil systems is so obviously complex that both soil scientists and plant
physiologists have been unable to provide a precise definition. If this complex
chemical system is to be accurately quantified, thermodynamic principles will

need to be used to evaluate experimental data. Even then, the limitations are
obvious, as in the case of K where the thermodynamic investigations are quite
often inapplicable under field conditions. This is because, although a quasi-
equilibrium in K exchange can be achieved in the laboratory, these conditions are
seldom, if ever, attained under field conditions (Sparks, 1987). Agricultural soils
are, for the most part in a state of disequilibrium owing to both fertilizer input and
nutrient uptake by plant root. It thus appears that a universal and accurate
definition of a root’s chemical environment awaits the proper application of
thermodynamics for the root’s ambient solution (Adams, 1974) or even kinetics,
as in the case of K (Sparks, 1987), where thermodynamics have been found to
be inadequate.
Soil extractions with different extractants provide a second approach in
defining the root’s chemical environment. This approach has been particularly
successful in understanding cases like P insolubility, soil acidity, and K fixation.
However, this approach also fails to define precisely the root’s chemical
environment. Though this approach also suffers deficiencies, such as the
extractants removing arbitrary and undetermined amounts of solid-phase
electrolytes and iron (or the extractants causing precipitation of salts or ions
from the soil solution) and the soil –plant interrelationship defined in terms of
solid-phase component of the soil, even though the solid phase is essentially inert
except as it maintains thermodynamic equilibria with the solution phase (Adams,
1974), the latter part could be researched more to understand how the solid
phase – solution phase equilibria can be interpreted to give a newer meaning to
quantifying nutrient availability. It is in this context that the role of the plant
nutrient “buffer power” assumes crucial importance.
The close, almost linear relationship in a low concentration range of
, 0.5 mM for NO2 þ þ 2 22
3 –N, NH4 – N, K , H2PO4 , and HPO4 , which has been
established by numerous solution culture experiments can be quantitively
described by the equation:
U ¼ 2p r aC r ; ð1Þ
where U is the uptake of a 1 m root segment, r is the root radius, Cr is the

concentration of the ion at the root surface, and a is the root absorbing power
(Mengel, 1985). The metabolic rate of the root determines its absorbing power.
A high root absorbing power would imply that a relatively high proportion of
nutrient ions coming in contact with the root surface is absorbed and vice versa.
The nutrient ion concentration at the root surface (Cr) depends on a since a high
root absorbing power tends to decrease Cr; it also depends on the rate of
movement from bulk soil towards the root (Mengel, 1985). Diffusion and/or mass
flow control this movement. But it is now established that nearly 95% of this
movement for nutrients such as P, K, and Zn (among heavy metals) and, possibly
298 K. P. P. NAIR
NHþ 4 , is by way of diffusion. When root uptake of an ion species is less than its
movement towards it, accumulation of the ion species on the root surface is bound
to occur, as has been shown to be the case with Ca2þ where mass flow contributes
to this accumulation (Barber, 1995). The diffusive path for ions such as P
and K, which plant roots take up at high rates but which are in low concentration
in the soil solution near the root, is the concentration gradient. In a sense, the
effective diffusion coefficient which quantifies the diffusive path and the buffer
power are analogous because the diffusive flux across the root surface is integrally
related to the nutrient buffer power. This has been shown to be true in the case of
P where a highly significant positive correlation between the two was found to
exist in 33 soil samples obtained from experimental sites in the USA and Canada
(Kovar and Barber, 1988). However, in a routine laboratory set up, it is far easier
to measure the buffer power than the effective diffusion coefficient and this
review will further focus on the question of how buffer power can be quantified
without recourse to cumbersome analytical techniques and how its integration into
routine soil test data will considerably improve predictability of nutrient uptake.
C. MEASURING THE NUTRIENT BUFFER POWER AND

ITS IMPORTANCE IN AFFECTING NUTRIENT CONCENTRATION
ON ROOT SURFACES
The ability to predict the mobility of dissolved chemicals, such as fertilizers,

in the soil is of considerable value in managing fertilizer applications. Soil testing
in its essence aims to achieve this. While modeling transport and retention of ions
from thermodynamic (Selim, 1992), kinetic (Sparks, 1989), and mechanistic
(Barber, 1984) angles could be very informative, the importance of translating
this information into practically feasible procedures in crop production calls for
an understanding not only of the basic concepts but of their intelligent application
as well. In a dynamic state of plant growth, the concentration of any nutrient on
the root surface is nearly impossible to measure since both the nutrient in the
plant tissue and the root absorbing power, which directly affects it, change
quickly due to root metabolic processes. The inability of even mechanical
mathematical models to accurately predict nutrient influx rate has recently been
highlighted (Lu and Miller, 1994). Hence, if an effective soil testing procedure is
to be devised for a nutrient, which is an alternative to defining the plant root’s
chemical environment, one must resolve the problems of quantifying the nutrient
concentration on the root surface indirectly, even if it is impossible to resolve it
directly, for the reasons mentioned above. Using Fick’s first law,
F ¼ 2DðdC=dxÞ ð2Þ
where F is the flux, dC/dx is the concentration gradient across a particular section,
and D is the diffusion coefficient, Nye (1979) has suggested that the formula
can be applied to both ions and molecules. The negative sign for D implies net
movement from a high to a low concentration. Although for molecules in simple
systems like dilute solutions D may be nearly constant over a range of
concentrations, for ions in complex systems like soils and clays, D will usually
depend on the concentration of the ion, and on that of other ions as well (Nye
1979). Nye (1979) has further suggested that though Fick’s first law may be
derived from thermodynamic principles in ideal systems, in a complex medium
such as the soil, the above equation may be regarded as giving an operational
definition of the diffusion coefficient. Thus, Nye (1979) defines the diffusion
coefficient as
D ¼ D1 u f1 ðdC1 =dCÞ þ DE ð3Þ
where D1 is the diffusion coefficient of the solute in free solution, u is the fraction
of the soil volume occupied by solution and gives the cross-section for diffusion,
f1 is an impedance factor, C1 is the concentration of solute in the soil solution, and
DE is an excess term which is zero when the ions or molecules on the solid have
no surface mobility, but represents their extra contribution to the diffusion
coefficient when they are mobile. DE can generally be neglected since only in rare
instances will it play any role in diffusion of plant nutrient ions in soil (Mengel,
1985). From the point of view of nutrient availability, dC1/dC, which represents
the concentration gradient, assumes crucial importance as we shall see below.
The term dC1/dC, where C1, is the concentration of the nutrient ion in the soil
solution and C is the concentration of the same ion species in the entire soil mass,
assumes considerable significance in lending a practical meaning to nutrient
availability. If we ascribe the term “capacity” or “quantity” to C and “intensity”
to C1, we have in this term an integral relationship between two parameters that
may crucially affect nutrient availability. Since the concentration gradient of the
depletion profile of the nutrient in the zone of nutrient uptake depends on the
concentration of the ion species in the entire soil mass (represented by “capacity”
or “quantity”) in relation to the rate at which this is lowered on the plant root
surface by uptake (represented by “intensity”), it could be argued that a
quantitative relationship between the two should represent the rate at which
nutrient depletion and/or replenishment in the rooting zone should occur (Nair,
1984a). This relationship has been functionally quantified by Nair and Mengel
(1984b) for P in eight widely differing Central European soils (Table V) and the
term dC1/dC has been referred to as the “nutrient buffer power”. Nair and Mengel
(1984b) used electroultrafiltration to quantify C1 while using an incubation and
extraction technique to quantify C. For P, C was found to closely approximate
isotopically exchangeable P (Keerthisinghe and Mengel, 1979), but in the
experiments conducted by Nair and Mengel (1984b) it was estimated by the
extraction of incubated soil with an extractant which was a mixture consisting of
0.1 M Ca lactateþ 0.1 M Ca acetateþ 0.3 M acetic acid at pH 4.1. The
extractant exchanges adsorbed phosphate and dissolves Ca phosphates except
300 K. P. P. NAIR
Table V
Comparison of P Buffer Power of Eight Widely Differing Central European Soils (Determined
by Two Different Techniques)
Regression r
Soil (1) (2) (1) (2)
Benzheimer Hof Y ¼ 18.8x þ 7.94 Y ¼ 0.23x þ 8.98 0.912 0.995

Hungen Y ¼ 38.2x 2 1.03 Y ¼ 0.25x þ 4.32 0.967 0.997
Oldenburg B6 Y ¼ 49.8x þ 0.52 Y ¼ 0.26x þ 0.72 0.994 0.999
Wolfersheim Y ¼ 70.3x þ 0.03 Y ¼ 0.27x þ 0.11 0.998 0.983
Obertshausen Y ¼ 70.5x þ 2.66 Y ¼ 0.30x þ 2.89 0.966 0.998
Oldenburg B3 Y ¼ 73.6x þ 2.07 Y ¼ 0.31x þ 0.61 0.994 0.997
Klein–Linden Y ¼ 75.0x þ 0.38 Y ¼ 0.32x þ 1.81 0.999 0.991
Gruningen Y ¼ 75.4x þ 0.89 Y ¼ 0.36x þ 3.62 1.000 0.996
Note: The b values in the regression functions represent the P buffer power of each soil. In regression
function (1) (after Nair and Mengel, 1984) Y ¼ CAL-P (Schüller’s method) and in regression function
(2) (after Nair, 1992) Y ¼ the author’s method. x in both refers to electroultrafiltrable P. Note the very
high r values in all the cases. The soils are arranged in their sequential increase in P buffer power.
apatites; the method known as the “CAL method”, developed by Schüller (1969),
is now widely used in Central Europe. In the case of Kþ and NHþ 4 – N, C denotes
the concentration of exchangeable, and to some extent non-exchangeable,
fractions (Mengel, 1985). Since very low concentrations in the range of 2.0 mM
may be attained on the root surface for both P and K (Claassen and Barber, 1976;
Claassen et al., 1981; Hendriks et al., 1981), Nair and Mengel (1984b) had to use
electroultrafiltration to quantify C1.
Thus, the nutrient depletion around the roots which is caused by the diffusive
flux of the nutrient towards the root surface is related to both the quantity and the
intensity parameters, and a quantifiable relationship between both represents the
buffer power specific to the nutrient and the soil. A growing root will at first
encounter a relatively high concentration of P, which is in the range of the
concentration of the bulk soil solution (Nair and Mengel, 1984b). As uptake
continues, depletion will occur at the root surface. This depletion profile gets
flatter with enhanced nutrient uptake (Lewis and Quirk, 1967; Claassen et al.,
1981; Hendriks et al., 1981). But it is the capacity of the soil to replenish this
depletion which ensures a supply of nutrient ions to the plant root without greatly
depressing its average concentration on the root surface. It is the nutrient’s buffer
power that decides these depletion and/or replenishment rates. A soil with a high
P buffer power implies that the P absorbed from the soil solution is rapidly
replenished. In such a case, P concentration at the root surface decreases only
slowly and mean P concentration at the root surface remains relatively high.
In soils with a low P buffer power, the reverse is true, and mean P concentration at
the root surface is rapidly diminished and remains relatively low. This has been
proved experimentally for P (Nair and Mengel, 1984b; Nair, 1992). This
phenomenon holds true for Zn2þ (Nair, 1984c) K (Nair et al., 1997) and NHþ 4 –N
as well (Mengel, 1985).
D. BACKGROUND INFORMATION ON THE IMPORTANCE

OF MEASURING Zn BUFFER POWER
There is a great paucity of published material on the effect of buffer power

on availability of heavy metals. Plants obtain most of their fertilizer Zn from
reaction products and not applied sources as such, implying that any source of
Zn added to soil has to necessarily conform to a chain reaction involving
adsorptive, desorptive, and resorptive processes that govern the maintenance of
an equilibrium between adequate Zn concentration in the soil solution nearest
to the zone of Zn depletion on the one hand and plant uptake on the other. The
Zn buffer power defines this. As Zn concentration in soil solution is normally
very low, the supply to plant roots by mass flow can only account for a very
small fraction of plant demand. For instance, with a transpiration coefficient of
300 l kg21 dry matter and a corresponding Zn concentration of 1027 M in the
soil solution, approximately 2 mg of Zn can be supplied by mass flow against
a demand of 10 – 30 mg Zn kg21 dry weight of plant tissue. In calcareous
soils, as the Zn concentration is of a much lower order of approximately
1028 M, the supply by mass flow could be very much lower (Marschner,
1994) indicating that mass flow can only contribute very negligibly to meet
plant needs of Zn. Hence, Zn movement to the plant root surface is principally
by diffusion and is essentially confined to a zone around the plant root which
hardly extends beyond the root hair cylinder (Marschner, 1994). In a review on
the mechanism of Zn uptake, Marschner (1994) indicated that flow culture
experiments with various species showed adequate ranges of Zn concentration
in the range of 6 £ 1028 –8 £ 1026 M, which are concentrations greater than
those that would be expected in the solution of most soils. He further pointed
out that although work using chelate-buffered solutions has indicated adequate
Zn concentrations between 10210 and 10211 M, extremely low adequate Zn
concentrations required a concomitant excess of about 100 mM Zn chelate as
buffer at the plasma membrane of the root cells. This implies a need for an
unlimited Zn pool for replenishment of Zn2þ at the plasma membrane. When
plants grow in soil, it is impossible to expect a Zn buffer of this size to exist,
and free Zn2þ and chelated-Zn concentrations will be at least threefold lower.
Hence, critical deficiency or sufficiency concentrations obtained through
research employing chelate-buffered solutions cannot be applied to soil-grown
plants.
302 K. P. P. NAIR
Most of the work on Zn availability to plants is based on chemical

extractions, among which DTPA extraction is the most frequently used. The
DTPA extraction quantifies a labile fraction of soil Zn comprising water
soluble, exchangeable, adsorbed, chelated and some occluded Zn. The critical
soil level of DTPA extractable Zn can vary from 0.3 to 1.4 mg kg21 soil which
equates to about 900– 4200 g ha21 of Zn in heavy soils and about 600– 2800 g
ha21 of Zn in light soils in the plough layer (0 – 20 cm). The crop requirements,
on the other hand, are quite small, in the range of 100 –300 g ha21 for a total
dry matter production of about 10 tons ha21 (Marschner, 1994). The
inadequacy of DTPA extraction to reflect plant Zn demand shows that other
important factors, such as replenishment of soil solution Zn (Nair, 1984c),
mobility, and transport to the root surface (Wilkinson et al., 1968; Nair et al.,
1984), and also the activity of the roots themselves (Wilkinson et al., 1968;
Marschner, 1994) are involved. Since the Zn buffer power is intricately
involved in all the three factors, the focus of this review is mainly on that
attribute.
As early as three decades ago it was suggested that colloidal Zn was
released by some specific process associated with root activity (Wilkinson et al.,
1968). Conditions in the rhizosphere and particularly root-induced changes
markedly affect Zn availability. A difference in rhizosphere pH of as much as
2, higher or lower compared to bulk soil, can be expected to occur as a result
of imbalance in ionic uptake. For instance, any acidifying fertilizer such
(NH4)2SO4 can result in a net excretion of Hþ ions and others, such as,
NH4NO3 can result in a net excretion of HCO2 3 or OH
2
ions. Additionally,
secretion of organic acids and enhanced CO2 production, as well, will affect
rhizosphere pH, and all of the above-mentioned changes will markedly affect
Zn availability. However, the scope of this review is confined to the kinetic/
dynamic aspects of the changes occurring in the rooting zone mirrored in the
Zn buffer power rather than changes in soil reaction in the rhizosphere per se
on Zn availability.
The distribution of Zn between the solid and solution phases can be described
by the buffer power. The availability of soil Zn to the plant depends on the initial
Zn concentration, Zn buffer power and effective diffusion coefficient (Barber,
1984). The Langmuir equation gives the relation between B and C1 as
C1 1 C
¼ þ 1
ðx=mÞ aB B
where C1 is the Zn concentration in the soil solution, x/m is the amount of Zn
adsorbed per unit of soil, B, the adsorption maximum and a is a constant related
to the soil’s bonding energy for Zn. A straight line is obtained when C1/(x/m) is
plotted against C1 with a slope of 1/B and intercept of 1/aB. The inverse of
C1/(x/m) ¼ b, the Zn buffer power, where C1 and x/m are both expressed in
volume units (Barber, 1984). Using this approach, Shuman (1975) estimated
the buffer power values varying from 5 to 100 for four soils representing
different major physiographic regions of Gerogia. Based on the diffusion model
of Drew et al. (1969), Nair (1984) has argued that the c in the equation
U ¼ 2paaC t (Drew et al., 1969), where, U is the quantity of Zn absorbed per
centimeter root length, a is the root radius in cm, a is the root absorbing power,
C 2 is the average Zn concentration on the root surface, and t is the duration of
the absorption period, in fact, represents an indirect measure of the Zn buffer
power. As we already know, the bulk of Zn uptake is by diffusion (Wilkinson
et al., 1968; Elgawahary et al., 1970; Barber, 1984). This diffusive process will
maintain a concentration gradient in the root zone. This concentration gradient
will directly affect Zn uptake because of its effect on the average Zn
concentration on the root surface. The Zn buffer power will affect this
concentration gradient, because the rate of Zn depletion and/or replenishment is
mirrored by it. In a sense, the effective diffusion coefficient and the buffer
power are analogous to each other for nutrients which are principally absorbed
by the plant root through diffusive process (Nair, 1989). Hence the crucial
question to examine would be the role of Zn buffer power in influencing Zn
availability for plant uptake.
E. QUANTIFYING Zn BUFFER POWER OF PEPPER

GROWING SOILS
Nair (1984c) has used electroultrafiltration to quantify Zn intensity in

measuring Zn buffer power of European soils. As the procedure is very highly
sophisticated and due to its non-availability in most of the developing countries,
a simple adsorption – desorption equlibrium technique was developed by the
author (Nair, 2002) to quantify the Zn buffer power of pepper growing soils. For
this, 200 g of representative soil samples from three locations in the state of
Kerala, where pepper is extensively grown, were incubated at 60% of the
maximum water holding capacity with graded rates of Zn over a fortnight, by
maintaining the water regime to constancy. At the end of the period the soil
samples were extracted with 0.01 M CaCl2 over a 24 h period and the extract
tested for Zn using atomic absorption spectrometry. This represented Zn intensity.
Separate extractions were made with DTPA, which represented Zn quantity.
Data in Table VI indicate that the Zn buffer power values of the experimental
soils varied widely.
Data in Table VI brings to light two very important facts. First, the very highly
significant correlation coefficients prove that the technique employed allows a
precise determination of the Zn buffer power of the soils. Second, the soils
varied widely in their Zn buffer power. It is this fact that has to be critically
viewed against the existing Zn fertilizer recommendation, emanating from
304 K. P. P. NAIR
Table VI
Zn Buffer Power of Pepper Growing Soils
Soil r-value b-value
Peruvannamuzhi 0.8337 ppp 0.7824

Thamarasseri 0.9304 ppp 1.5786
Ambalavayal 0.9604 ppp 3.0358
Note: ppp Significant at a confidence level of 0.1%; b values represent the Zn buffer power.
routine soil tests using the universally employed DTPA extraction to understand
the cruciality of Zn buffer power for precise formulations of Zn fertilizer
applications for pepper.
Data in Table VII clearly indicate that integration of the Zn buffer power has
improved the relationship between DTPA routine test versus both Zn
concentration and Zn uptake. More remarkably, the overall relationship between
DTPA test and dry matter production not only improved with Zn buffer power
integration in the computations, but turned from negative to positive. This
clearly shows that the commonly employed DTPA test could only be site-
specific, but on a larger scale it is the Zn buffer power that determines plant
performance. The experiment had only monitored dry matter production in the
pepper plant without taking it to berry formation as it takes about 3– 5 years for
berry formation. But, the data in Table VII clearly substantiate the cruciality of
the buffer power concept.
Data in Tables VI– VIII have to be viewed in conjunction to precisely
understand the cruciality of the buffer power concept. Pepper is the
economic mainstay of the state of Kerala and of late widespread Zn
deficiency has been observed in the state, and the onset of the dreaded disease
Foot Rot due to Phytophthora fungi is attributed, to a great extent, to this
deficiency in the soils. The scientists at the Indian Institute of Spices Research
Table VII
Correlation Coefficients (r) for the Inter-relationship between Routine DTPA Soil Test versus
Zn Concentration, Zn Uptake and Dry Matter Production without (1) and with (2) Zn Buffer
Power Integration in Pepper
Details 1 2
Zn concentration versus DTPA test 0.884 ppp 0.924 ppp

Zn uptake versus DTPA test 0.782 0.862 ppp
Dry matter production versus DTPA test 2 0.745 0.777 ppp
Note: ppp significant at a confidence level of 0.1%.

Table VIII
Pepper Yield from Farmers’ Fields (kg vine21) Weighted Against Zn Buffer Power
Yield
Region Targeted Actual Deviation (%)
Peruvannamuzhi 0.241 0.401 þ 66

Thamaraserri 0.490 0.487 þ 0.6
Note: Target weighting was done against the highest yield obtained from the Ambalavayal region.
Note the remarkable closeness between targeted and actual yields in Thamarasseri region.
Peruvannamuzhi soil is an atypical pepper soil.
have made a “blanket recommendation” of soil application of 23 kg ZnSO4 ha21

when the DTPA extractable Zn is less than 1.6 ppm. In terms of monetary input,
this would translate to an equivalent of US $25 ha21 in terms of farmer
investment. As the soils are variously Zn buffered (Table VI) a single “blanket
dose”, such as the one that is made, is totally unscientific, because, where the
farmer needs to apply only lesser quantities of Zn, he would end up applying
more, if he were to go by the routine recommendation based on the DTPA
extraction. To elaborate, the soils from the Ambalavayal region would only
require 25% of Zn needed in the soils of the Peruvannamuzhi region as the former
has a Zn buffer power nearly fourfold more compared to the latter. In the case of
the soils from the Thamarasseri region the quantity required would only be 50%.
In fact, the scientists at the Institute have made a gross underestimation of the Zn
supplying power of the soils in the state as their results are all based on the soils of
the Peruvannamuzhi region alone where the Institute has its experimental farm.
The soils here are obviously Zn impoverished, but that is not the case in the entire
state of Kerala. This also shows that, more often than not, recommendations from
experimental stations, when extrapolated on a large scale can run into problems,
especially when such recommendations are based on routine soil testing
procedures. These results (Nair, 2002) substantiate the earlier ones (Nair,
2003) on the importance of measuring Zn buffer; rather than the routine DTPA
extractable Zn alone, to make precise Zn fertilizer recommendations for wheat in
Central Asia (Turkey) where farmers have been advised to apply as much as
100 kg ZnSO4 ha21 as a “blanket recommendation”, without obtaining any
tangible and consistent response in wheat yield. Hence, the economic importance
of the buffer power concept, as against the routine soil testing procedures, needs
hardly any underscoring. Nair et al. (1997) have conclusively demonstrated the
importance of the buffer power concept in the case of K nutrition for another very
important perennial crop and spice, cardamom (Elettaria Cardamomum M.) in the
state of Kerala. Both pepper and cardamom are the staple spices of the state.
306 K. P. P. NAIR
V. ESTABLISHING A PEPPER PLANTATION
A. THE INDIAN EXPERIENCE
From a commercial view point, establishing a good pepper plantation is of

much significance. Unlike in countries such as Thailand, Malaysia, Brazil, etc.
where the life span of a pepper plantation is only 10 –15 years, at the end of
which replanting is done, in India a pepper plantation can last for more than a
quarter century. The prime reason for the shorter duration of the pepper
plantations in other countries, referred to above, unlike in India, is because, in
the former, pepper is trailed on non-living standards (support), while in India
pepper is always trailed on live support, such as E. indica. The starting point
to establish a good pepper plantation is the establishment of good planting
material. With the exception of India, most other countries use the orthotropic
(upward climbing shoot) shoot as the planting material. This is the best
planting material as they result in vigorous plants and bear fruiting laterals
right from the base itself and yield a lot earlier. In India, however, the non-
availability restricts the use of orthotropic shoots widely.
In earlier times, pepper planters used either runner shoots or climbing
orthotropic shoots to coincide with the onset of the southwest monsoon in India.
These days, pre-rooted cuttings from runner shoots are used instead. When
hanging shoots are used as planting material, they result in the formation weak
plants. Quite often, pruned material — pepper is pruned five to six months after
planting, and once again, after a year — is also used as planting material. These
are pruned stem cuttings.
Since pepper is only propagated through the type of planting material
described above, world-wide, there is a huge demand running into millions for
ready to use planting material. Production of pre-rooted cuttings in light
polyethylene bags (PVC bags) is the surest way of mass producing material once
the demand for planting material escalated in India. Their production is as
follows. Runner shoots from high yielding and healthy plants are kept coiled at
the base of wooden pegs fixed at the bottom of the plant so as to prevent the
growing shoots spreading on the soil and striking roots. These runner shoots can
be separated from the mother plant in winter (January – February) by snapping
them using scissors and are disinfected by dipping them in a fungicide solution,
such as Oxychloride or Bordeaux mixture for a minute, later surface dried and cut
into three nooded strips after clipping the leaves, planted either in raised soil beds
or PVC bags filled with a mixture of soil, sand and some farmyard manure, which
will provide enough nutrients in a growth substrate. Application of growth
regulators, such as IBA, by dipping the cuttings in the solution has been found to
enhance root proliferation (Pillai et al., 1982; Suparman and Zaubin, 1988). Care
must be taken to minimize fungal infection arising from the nursery bed.
In Sri Lanka growing planting material on split bamboo led to the production
of good planting material on a large scale. Popularly known as the “bamboo
technique” (Bavappa and Gurusinghae, 1978) the method consists of first digging
a trench (60 £ 40 cm2) of convenient length, filling it with a rooting medium —
preferably forest soil, sand and farmyard manure in a 1:1:1 ratio mixture and then
split bamboo poles (1.25 – 1.5 m long) or PVC pipes — both of about 10 cm in
diameter — are sunk into the trenches at 458 using a strong middle support.
Rooted cutting, one each per bamboo is planted in the trench and allowed to trail
onto the bamboo or PVC poles. The lower portion of the pole is filled with a
rooting medium (of weathered coir dust and farmyard manure in a 1:1 ratio) and
as the vine grows it is tied to the support. Vines are irrigated daily and to stimulate
rapid growth a nutrient solution consisting of N, P, K, and Mg (1 kg urea, 0.75 kg
super phosphate, 0.5 kg muriate of potash (MOP) and 0.25 kg magnesium sulfate
all mixed in 250 l of water) can be applied at the rate of 0.25 l per vine every
fortnight. It takes three to four months for the vine to reach the top of the pole, and
when it does, the terminal bud is nipped off and the stem is crushed about three
nodes above the base to activate the axillary buds and after about 10 days each
vine is cut at the crushed point and this is ready for transplanting into PVC bags,
filled with the rooting mixture as detailed above. Care should be taken, while
planting in PVC bags, not to injure the roots and not let the axil be below the
rooting medium. The PVC bags are placed in a cool place and covered with a
light PVC sheet (200 gage) to retain high internal humidity. Buds start to develop
in about three weeks time and then the PVC bags can be shifted to a place with
shade. Regenerated shoots from the stumps can be trained as before, which
permits the production of a continuous stream of rooted cuttings as planting
material. Four harvests are possible in 1 year, with a multiplication ratio of 1 –40.
A prolific root system develops and field establishment is very good. In Sarawak
(Malaysia), on average, 54 rooted cuttings could be obtained (Ghawas and
Miswan, 1984) and the method has been successfully field tested in the
experimental farm of the Indian Institute of Spices Research. Figure 5 shows the
“bamboo technique” developed in Sri Lanka for vegetative pepper propagation.
Pepper can be planted on flat surfaces and slopy land. If it is the latter, slopes
facing south should be avoided. On slopy land it is advisable to plant along the
contour lines to arrest top soil loss due to run off water during monsoon. An
important feature of pepper plantations in India is that the vines are trailed on
living standards, mainly E. indica or E. lithosperma (a thorny tree and the
trailing vines have strong footholds on the thorns) Glyricidia, G. robusta (Silky
Oak), G. pinnata, Ailanthus spp. etc. In homestead plantations, coconut and
arecanut trees act as the main crop in addition to being the standards. In the
southern state of Karnataka, pepper is grown along with coffee, where coffee is
the main crop, and the forest trees act as the standards. Glyricidia is the most
common standard in Sri Lanka and Malayasia. Pepper planting should precede
the onset of monsoon. Three noded 2 –3 rooted cuttings, prepared as explained
308 K. P. P. NAIR
Figure 5 Split-bamboo technique developed in Sri Lanka for vegetative pepper propagation.
earlier, are planted in shallow pits on the northern side of the standard. If the
cuttings are unrooted, 4 –5 could be planted. At least one node should be below
the soil for the establishment of roots. To accelerate growth, pits can be filled
with top soil, farmyard manure, rock phosphate, and neem cake mixture (Pillai,
1992). Plant density is around 1970 – 2000 ha21 (George, 1982) and can go up to
5000 ha21 when nutrient and carbon stress can be observed (Reddy et al.,
1992).
In order to support the climbing vine on the standard, it should be tied as
frequently as necessary, and for tying jute thread is used. Pruning of the terminal
part, though not extensively practiced in India, encourages growth and bearing
(Kurien and Nair, 1988). When a pepper plantation is established, starting from
the beginning of the year, specific operations are to be followed. In the first two
months, starting from January, when harvesting is completed, pruning the
hanging shoots, tying vines, and mulching the basin is done. In the following two
months, diseased vines are removed and the pits could be filled with top soil, 1 kg
lime and 5 kg farmyard manure mixture. Southwest monsoon establishes in
May –June and with this a mixture of 5 kg farmyard manure, 250 g rock
phosphate and 1 kg neem cake should be applied. In June – July, a Bordeaux
mixture (1%) and drenching the pits with copper oxychloride (0.2%) is done. The
most damaging “Pollu Beetle” (Longitarsus nigripennis), a deadly insect that
leaves the berries hollow (from which the name “Pollu” in the regional language
Malayalam of Kerala state is derived, meaning hollow) can be controlled by
spraying endosulfan (1.5 ml l21). Application of 50 g urea, 120 g MOP should
also be done. In September – October 50 g urea, 120 g of MOP and 50 g Mg SO4
are applied. Additionally, a Bordeaux mixture spray (1%) is also done. Tying of
vines should be continued. In November –December phytosanitary measures,
such as removal of infected plants, drenching with oxychloride (0.2%), mulching,

and shading the young vines must be done.
It is uncommon to irrigate a pepper crop, except in Thailand where it is grown
as an irrigated crop. Nevertheless, pepper plants can seriously suffer due to
moisture stress in the summer months, especially in India, when daytime
temperatures from April to May can soar to 358C plus. A field study conducted at
the Pepper Research Station in Panniyur, Kerala state, during 8 years (1988 – 96)
indicated that irrigation at IW/CPE ratio of 0.25 increased pepper yield, as much
as 90%, the positive effect being most manifest in Panniyur-1 and Karimunda
(Cultivar) both released from this station (Satheesan et al., 1997a).
An interesting aspect of pepper cultivation is the possibility of growing other
crops, in addition to pepper, in the same field or plot of land. This practice is more
in common when it is the case of homestead farming where income from the
accompanying crop adds to the profit from the main pepper crop. However, this is
not a general practice on large pepper plantations, because raising an additional
crop makes demands of its own, in terms of nutrient input, pest management, etc.
It is important to note that companion crops to pepper must best be raised in the
pre-bearing stage of pepper, which at best can extend from 3 to 5 years (at the
most). This is because, when fully grown, pepper canopy could completely shade
the accompanying crop and the latter would hardly yield. In an experiment at the
Indian Institute of Spices Research (Sadanandan, 1994) reported that an
accompanying crop of banana (cv. Mysore poovan) fetched an additional income
of approximately US $460 ha21. Pillai et al. (1987) have obtained comparable
results.
The recent glut in the pepper market and slump in prices have prompted many
farmers to look to companion crops to generate additional income. Simul-
taneously, the falling prices of coconut have prompted coconut farmers to use the
coconut trees as a live standard to raise pepper. In this connection, an interesting
experiment conducted at the Central Plantation Crops Research Institute
(CPCRI) at Kasargode, Kerala State, which has a mandate for research, primarily
on coconut, merits mention.
Pepper vines (Panniyur-1) were trailed on 60-year-old coconut palms and
when the vines reached a height of 4 –5 m, further growth was restricted to
enable farmers from climbing the palms to pluck the nuts. On average, 2 kg dry
pepper vine21 was obtained. Companion cropping with arecanut has also been
tried. An arecanut — pepper combination gave 3832 kg ha21 of dry dehusked
arecanut and 1418 kg ha21 dry pepper (Nair and Gopalasundaram, 1993), from
1000 vines ha21. Companion cropping is common in Brazil as well, where an
array of companion crops, such as rubber, cocoa, orange, lemon, clove, etc., are
extensively tried. Though growing mixed companion crops in pepper gardens is a
practice in many situations, it needs to be clearly understood that when inter
specific crops are grown in association, the biological implications could be
diverse. This diversity might originate not only in nutrient absorption pattern,
310 K. P. P. NAIR
Figure 6 A field view of an Indian Pepper Plantation.
moisture stress relationship, etc., but also can extend to other areas, such as
susceptibility/resistance to specific pests and diseases. These are the areas that
deserve thorough scientific scrutiny before value judgment can be made on the
suitability or otherwise of growing companion crops with the main crop of
pepper. Figure 6 shows a field view of an Indian Pepper Plantation.
B. THE INDONESIAN EXPERIENCE
Within Asia, next to India, Indonesia is the major pepper grower. Pepper was
the first spice Indonesia traded with Europe through Persia and Arabia and prior
to the second world war, 80% of world production was controlled by Indonesia. It
was the Japanese occupation during the war that left many pepper plantations
uncared for with the resulting decline in production. Compared to the pre-war
period when the country could boast of more than 20 million vines, after the war
it came down to just about a lakh of vines. The most interesting characteristic of
pepper cultivation in Indonesia is that most of the pepper gardens are owned by
small farmers unlike in India where large plantations could be found. Presently
the main producing areas are Lampung and Bangka with Lampung in the lead.
During the last quarter of the past century, a perceptible decline in area in
Lampung has taken place with shifts to Bangka, both east and west Kalimantan
and Sulawesi. In Bangka, pepper was first grown in Muntok and the Muntok
white pepper had a good world market.
In Indonesia, pepper is grown in various types of soil, such as andosol, grumsol,
latosol, podsol, regusol, etc. Well-drained alluvium rich in humus, pH above 5.8, is
ideal for pepper (Zaubin and Robbert, 1979). In general, pepper soils in Indonesia
are poor in fertility, texture, and structure and low in organic matter. The two main
pepper-growing areas, Bangka and west Kalimantan have poor, reddish brown,
and sandy soil. Soils in Lampung are reddish brown latosols and contain more
organic matter than the soils of Bangka. In Indonesia, the planting material is
cuttings from orthotropic shoots. Normally cuttings with five to seven internodes
are used, with two to four nodes buried in the soil. Cuttings are planted at an angle
of 35– 458 with reference to the standard. For the first three to four weeks, the
cuttings must be shaded from the sun. Cuttings are planted in pits of
0.6 £ 0.8 £ 0.6 m size. Rooted cuttings are also used. Precautions explained
earlier must also be adhered to in order to protect the cuttings. The procedures used
in the case of pre-rooted cuttings, except that they have a better ability to withstand
heat exposure from sunshine, are the same as above. Such planting materials are
selected from the nursery and it should be ensured that only well-rooted and well-
grown cuttings in PVC bags are transplanted in the main field.
Unlike in India, Sri Lanka, and the Philippians, in Indonesia the types of
standards used are both living and non-living. In the former category, we have
Glyricidia, Dadap, Kapok, etc., while in the latter category we have hard timber
poles, made of iron, wood, or concrete poles, etc. Experience shows that concrete
poles result in poor growth and productivity. In Lampung small farmers use
living standards while non-living ones are used in Bangka, east and west
Kalimantan. There is a gradual shift to living standards as non-living ones, of late
have become more expensive.
Pepper plantations are usually owned by small farmers, with holdings ranging
between 0.2 and 1.5 ha. On average, the size is 0.65 ha. In general, pepper
cultivation is intensive in Bangka and south Sumatra, while in Lampung it is
extensive, where pepper is planted under live standards and input rates, in terms
of fertilizer, pest management, etc., are meager.
The Indonesian pepper industry is mostly owned by small farmers and
planting pepper on a plantation scale did not last long in Java. The nature of
pepper cultivation and the degree of involvement of labor and capital have played
a very important role in determining the mode of the pepper farming system in
Indonesia. However, it is a tribute to the pepper industry in Indonesia that the IPC
office is headquartered in Jakarta. Figure 7 shows a field view of an Indonesian
Pepper Plantation.
VI. PEPPER PESTS AND THEIR CONTROL
The pepper plant is attacked by several diseases, caused by fungi, bacteria, virus
and mycoplasma, insects, and nematodes. Nutritional disorders aggravate the
impact of the pests, be they caused by any one of the above or soil and/or
environmentally triggered. However, it is important to make a distinction between,
for example, a deadly disease that is caused by a fungus, such as Phytophthora spp.,
312 K. P. P. NAIR
Figure 7 A field view of an Indonesian Pepper plantation.
from which specific symptoms in the plant will originate, and a nutrient deficiency
in the soil that may lead to an imbalance of the specific nutrient in the plant that will
aggravate the symptoms of the disease subsequently. It is important to recognize
that of all the diseases a pepper plant may suffer from, that caused by the
Phytophthora fungus is the most devastating. Hence, this review will primarily
focus on the “Foot Rot” disease caused by Phytophthora, which is the major
disease in India, Malaysia, and Indonesia causing very serious damage to pepper
production (Holliday and Mowat, 1963; Kueh and Sim, 1992b; Manohara et al.,
1992; Sarma et al., 1992c). In India, although wilt disease was reported as the
major cause of pepper crop loss, Phytophthora, as its causal agent, was first
reported only in 1966 by Samraj and Jose (1966). There are, in fact, 17 diseases that
cause loss in yield, but among these Phytophthora Foot Rot and “slow decline”,
which were originally referred to as “quick wilt” and “slow wilt”, respectively,
cause the most damage (Nambiar and Sarma, 1977; Nambiar, 1978; Das and
Cheeran, 1986). Phytophthora capsici is the causative fungus. The infection
occurs aerially as well as through soil and the severity of infection depends on the
site of infection (Mammootty 1978; Anadaraj and Sarma, 1995). When it is a foliar
infection one to many dark spots appear, which have characteristic fimbriate
margins, which advance and later coalesce leading to defoliation even before the
lesions spread to the entire lamina. The base of the vines can also be infected. On
tender shoots, the fungus profusely sporulates forming a white covering and when
the infection reaches the stem abrupt wilting of the entire plant takes place. Spikes,
when infected, lead to the formation of blackened berries. When the infection is
below the soil on roots, rotting and degeneration starts leading to yellowing,
defoliation, and drying up of the plants. Feeder root infection reaches the collar
through main roots, which causes the characteristic Foot Rot; hence the name
(Anandaraj et al., 1994).
Since Foot Rot results in sudden wilting followed by death of the plant the
disease was earlier referred to as “wilt” or “quick wilt” and the disease causing
fungus was identified as Phytophthora palmivora var piperis by Samraj and Jose
(1966) and as P. palmivora by Holliday and Movat (1963). In India, though
pepper is traditionally grown in the Western Ghats in the states of Kerala,
Karnataka, and Tamil Nadu, it has been introduced to non-traditional pepper
growing areas, such as the Andhra state and also the north-eastern states and the
incidence of Foot Rot has been reported from these places as well (Sarkar et al.,
1985). Up to 30% crop loss has been reported (Samraj and Jose,1966; Nambiar
and Sarma, 1977). The onset and spread of Foot Rot has been found to be closely
correlated to a wet and cloudy atmosphere prevalent during the southwest
monsoon period in the state of Kerala. A combination of factors, such as daily
rainfall of 1.6 –2.3 cm, ambient temperature ranging from 22.7 to 29.68C,
relative humidity of 81– 99% and daily sunshine duration of 2.8– 3.5 h favor the
spread of the disease (Ramachandran et al., 1988c, 1990).
Inasmuch as infection through soil substrate is concerned, the prime source of
infection is contaminated soil where new planting is done. The inoculum can
survive in the soil for upto 19 months in the absence of a host plant (Kueh and
Khew, 1982). The pathogen is normally concentrated to a depth of 30 cm
(Ramachandran et al., 1986) and the severity of concentration diminishes as one
moves deeper and away from the infected vines. Management of Foot Rot is,
perhaps, the most challenging aspect of pepper cultivation. There are various
aspects to the control measures, such as cultural, phytosanitary, biological, and
chemical. The most important among the cultural measures refer to maintenance
of good drainage. Stagnant water leads to anaerobic conditions triggering
germination of pathogen propagules. A decrease in the production of phenol
oxidase, phytoalexin and fixed nitrogen combined with suppressed mycorrhiza
leads to host plant susceptibility in anaerobic conditions (Drew and Lynch, 1980).
Good drainage blocks the buildup of P. capsici. Phytosanitation assumes an
important role in controlling the disease. Since the infected plants serve as the
foci of infection (Zadocks and Van den Bosch, 1994), their removal is most
crucial in controlling the disease. The initial occurrence and spread of the disease
is non-random and tends to cluster around the previously infected plants
(Anandaraj, 1997) and this is the reason why such infected plants have to be
removed forthwith. Further, as pepper grows on living standards, they develop a
canopy of their own generating a microclimate different from that of the ambient
situation, with high humidity and low temperature, which is ideal for P. capsici to
multiply fast and infect. Branches of the living standards should be regularly
lopped, especially during the rainy season, which facilitates the penetration of
direct sunlight that leads to the heating up of the soil and thereby reducing the
humid surroundings, thus acting as a check for the multiplication of the pathogen.
The loppings could be used as surface mulch, which while adding to the organic
matter of the substrate soil can also check weed growth. Since the prime source of
314 K. P. P. NAIR
the initial inoculum of P. capsici in pepper plantations is the contaminated soil,

infection of the foliage can occur through soil splashed onto tender shoots trailing
on the ground and from these shoots onto other parts of the vines through rain
splashes (Ramachandran et al., 1990). In order to arrest soil splashes, legume or
grass cover mulch is ideal (Ramachandran et al., 1991; Sarma et al., 1992c).
However, studies in Malaysia (Ahmed, 1993) indicate that pepper grew better in
clean-weeded plots unlike in those with a cover crop of Desmodium trifolium.
The Indian experience shows (Anandaraj, 1997) that after the cessation of the
southwest monsoon it would be ideal to clean up the pepper plots of the existing
weeds and rake up the soil which was seen to check the spread of the pathogen.
Chemical control of Phytophthora has revolved around the use of systemic
fungicides, such as metalaxyl formulations like Ridomil granules or Ridomil-
ziram (Ramachandran and Sarma, 1985; Ramachandran, 1990; Ramachandran
et al., 1988b, 1991; Kueh and Sim 1992a; Sarma et al., 1992c; Kueh et al., 1993)
and/or the classical Bordeaux mixture 1% spray and also pasting the collar
with Bordeaux mixture and drenching the basin/trench around the vine
with oxychloride (Malebennur et al., 1991; Nair and Sasikumaran, 1991;
Ramachandran et al., 1991; Lokesh and Gangadarappa, 1995). Owing to the
high cost of chemical control, it is not very popular with small farmers. Further,
Coffey (1991) has expressed doubts about the efficacy of chemical control of the
soil-borne diseases caused by Phytophthora.
In addition to the above measures of control, there is also the possibility of the
use of organic amendments and biological control. Inasmuch as the addition of
organic amendments to the soil are concerned to control the Phytophthora spp.,
the objective is to raise the growth of saprophytes (Kueh and Sim, 1992a) and it
has been observed that the enhanced activity of saprophytes checks the growth of
P. capsici dramatically (Anandaraj, 1997). Oil cakes, such as that obtained after
expelling oil from neem seed, groundnut, coconut, etc., can beneficially be used
to control the growth of the disease causing fungus (Sadanandan et al., 1992; Nair
et al., 1993). In addition, even chicken manure has been found to be useful.
Applications of organic amendments such as the ones cited above will have a
twofold beneficial effect: first it enhances the nutrition of the vines, and second, it
encourages the growth of saprophytes that check the population of P. capsici.
Both complement each other in the overall control of the fungal growth.
A recent development in the control of Phytophthora is the employment of
biological organisms, primarily from a fungal background. Mention must be
made of vesicular arbuscular mycorrhizae (VAM), Trichoderma and Gliocla-
dium (Manjunath and Bagyaraj, 1982; Ramesh, 1982; Datta, 1984; Anandaraj
and Sarma, 1994b; Anandaraj and Peter, 1996). Soil-borne pathogens are
amenable to biological control (Cook and Baker, 1983). Since P. capsici is a soil-
borne fungus and the prime source of contamination is infected soil, growth of
antagonistic fungi would check the build up of the pathogenic fungus. Interesting
work on isolation and mass multiplication of inexpensive carrier media for field
application of microorganisms having biological control properties has been

carried out (Anandaraj and Sarma, 1995; Sarma et al., 1996). An inexpensive
carrier medium for the culture of these beneficial organisms is water contained in
mature coconuts that is thrown away after shelling. Both Trichoderma and
Gliocladium have been found to grow well in this medium (Anadaraj and Sarma,
1997).
Interest in VAM has been mainly focused on its ability to solubilize native
phosphorus. Its use in the control of pepper diseases was started when it was first
observed that VAM suppressed the Foot Rot caused by Phytophthora in citrus
(Davis et al., 1978; Davis and Menge, 1981). Soil incorporation of VAM can also
be done along with Azotobacter and Azospirillum, which, besides controlling the
pathogen have beneficial growth effects (Govindan and Chandy, 1985; Bopaiah
and Khader, 1989) The positive effect of VAM is through the alteration of the
nature of root exudates and the rhizosphere microflora — which is generally
termed the “mycorrhizosphere effect” — and there is extensive research to
substantiate this view (Dehne, 1982; Graham, 1982, 1988; Linderman, 1988).
The positive effects of VAM have not only enhancing plant growth, but also, in
suppressing disease causing pathogens has been reported by Ewald (1991) and
Graham and Egel (1988). In view of the overall positive effect of VAM, it has
been recommended that VAM be incorporated into the soil right from the nursery
stage and onwards so that the beneficial effects of enhancing growth and
suppression of pathogenic effects is obtained right from the early stages (Sarma
et al., 1996). Field experience shows that strict phytosanitation in the nursery is a
very important pre-requisite in subsequently installing a disease-free pepper
plantation.
Next in importance to diseases caused by pathogenic fungi are the diseases
caused by nematodes. Crop productivity is seriously hampered by plant parasitic
nematodes and pepper is no exception. Of the 15,000 nematode species described
so far, 2200 are plant parasitic. Quite often nematode damage goes unnoticed.
Since nematode infestation starts at the root surface, pepper vines so infected will
subsequently suffer from a number of secondary complications and quite often
these symptoms can be mistaken for nutritional or other physiological disorders.
When susceptible crops are continuously grown, in the year after, they become a
good breeding ground for disease causing nematodes. A compilation of plant
parasitic nematodes associated with pepper in the major pepper growing
countries listed 48 species, that belong to 29 genera (Sundararaju et al., 1979),
while 54 species that belong to 30 genera have been listed by Ramana and
Eapen (1998). In India, 17 genera of nematodes associated with pepper were
listed from the two primary pepper growing states, Kerala and Karnataka
(Sundararaju et al., 1980), while in Indonesia 14 genera have been listed (Bridge,
1978; Mustika and Zainuddin, 1978). Among the nematode species, Meloidogyne
incognita and Radopholus similis are the most important that cause the most
severe damage. The former is commonly known as the “Root Knot Nematode”,
316 K. P. P. NAIR
while the latter as “The Burrowing Nematode”. On a global scale the former is
the most devastating.
The root knot nematodes have a specialized and complex relationship with the
host plant and they are sedentary obligate endoparasites. Their infestation of the
vine results in the formation of elongated swellings on the thick primary roots due
to multiple infections and typical knots or galls on either secondary or fibrous
roots because of hypertrophy and hyperplasia of the infected tissues. The name
root knot is derived because of these typical knots on the roots. In thick primary
roots a number of adult females with egg masses localize deep below the
epidermis and the entire length of the root turns into a gall and appears smooth
with infrequent swellings (Mohandas and Ramana, 1987). Since nematodes feed
on vascular tissues, a disruption in the arrangement and continuity of the vascular
bundles leads to impaired movement of both water and nutrients and as a
consequence, the plant is vitially affected. When infestation is severe, a large
amount of root mass is lost due to eventual decay of the galled roots, which in
turn very adversely affect the entire vine (Mohandas and Ramana, 1991; Siti
Hajijah, 1993). When root-knot infestation takes place, yellowing of the foliage
occurs resulting in stunted growth and eventual decline of the vine. Dense
yellowing of interveinal areas with the deep green veins prominently visible
(Ramana, 1992) can often be mistaken for nutritional deficiency. When vines are
infested with M. incognita, certain impaired physiological reactions, such as
lowered absorption and translocation of P, K, Zn, Mn, Cu, Ca, and Mg and their
accumulation in leaves have been observed (Ferraz et al., 1988) and also
reduction in total chlorophyll content of leaves (Ferraz et al., 1989). These
changes lead to stunted growth of the vines. When vines were inoculated with
M. incognita inoculum, a high concentration of total phenols without expression
of any resistance to the pest was observed (Ferraz et al., 1984). Additionally,
several changes in amino acids, organic acids and sugars were also observed in
vines infested with the nematode (Freire and Bridge, 1985b).
The burrowing nematode (R. similis) is an obligate and migratory
endoparasite, is extensively found in both tropical and subtropical regions of
the world, is a serious pest of pepper (Holdeman, 1986), and has a wild range of
hosts (about 370 plant species). The existence of R. similis was first reported in
the state of Kerala, India, on the banana host by Nair et al. (1966). The nematode
feeds on cortical tissues and produces elongated dark brown necrotic lesions on
the roots at the site of infection. The nematode pushes through the cell wall of
each cell after draining its contents and this burrowing phenomenon results in the
formation of tunnels in the root tissues. The nematode derives its name from this
trait of burrowing. When infestation is severe, many lesions coalesce and encircle
the root cortex and because of this damage to the cortical cells, the root
portion distal to these lesions gradually disintegrates. The vines tend to produce
new roots which, in turn, get infected leading to the formation of a bunch
of decayed root mass (Mohandas and Ramana, 1987). Yellowing of leaves,

defoliation and generally stunted growth result from infestation.
Another important disease, popularly known as the “Slow Decline Disease”,
caused by a combination of the two nematodes, M. incognita and R. similis,
and the fungus Fusarium spp., has been the cause for widespread destruction
of pepper in several countries. The disease is also known as “yellows”. It was
the cause of major pepper devastation in the Bangka Islands of Indonesia in
the fifties (Christie, 1959), and also in Guayana (Biessar, 1969), India
(Nambiar and Sarma, 1977), Malaysia (Kueh, 1979, 1990), Brazil (Sharma and
Loof, 1974; Ichinohe, 1975), and Thailand (Sher et al., 1969; Bridge, 1978).
The disease was first observed on the islands of Bangka (Van der Vecht, 1950)
in Indonesia and later spread to the islands of Belantung. The disease caused
losses of up to 30% annually (Sitepu and Kasim, 1991). The disease is widely
prevalent in Sarawak (Malaysia). The total life span of the vines can be
reduced to 8– 10 years, as against the normal 25 –30 years (Varughese and
Anusar, 1992). In Cambodia, where pepper is predominantly found in the
Kampot region, the disease was responsible for bringing down the vine
population from 2.5 million in 1942 to 0.5 million in 1953 (Hubert, 1957). In
the predominantly pepper growing state of Kerala in India, there is no precise
estimation of crop loss, but Menon (1949) reported a 10% loss of vines. It has
been observed that plants infested by root-knot nematodes are more
susceptible to the fungus Phytophthora (Winoto, 1972).
Fool-proof management of nematode infestation is an elusive target. It is
important to recognize that the primary source of infestation is nurseries.
Solarization and fumigation are effective in controlling nematode infestation in
nurseries. Additionally, incorporation of biocontrol agents, such as VAM, to
solarized soil has been found to be an encouraging method for the control of
nematode infestation (Ananadaraj and Sarma, 1994a,b; Sarma et al., 1996).
There are different angles to the control regime. For instance, in plantations it
has been observed that chemicals released during decomposition of organic
manure, for instance, Azadirachtin from neem cake and Ricin from castor
cake, are toxic to nematodes (Stirling, 1991). Mulching has been found to have
a positive effect on nematode control (Wahid, 1976; Ichinoe, 1980, 1985).
Hubert (1957) obtained very encouraging results in nematode control using
Eupatorium mulch. Of the various aspects of developing resistance/tolerance
to nematodes, the most reliable is the development of plant resistance. But, to
date, there are no varieties or cultivars that are totally resistant to nematode
attack. Obviously, the focus of pepper breeding has not been on
developing absolute nematode resistant varieties. In fact, polygenic-horizontal
resistance is more important than vertical-monogenic resistance (Fassuliotis
and Bhatt, 1982). Existence of physiological races or pathotypes in
nematodes, in particular Meloidogyne spp., is another factor that must be
keenly focused on in any breeding program in pepper for nematode resistance.
318 K. P. P. NAIR
Currently the efforts focus on developing tolerance to nematode infestation

rather than developing absolute resistance. Also, efforts are made to screen the
currently available cultivars for their reaction to nematodes. For instance, in
Sarawak (Malaysia), Kueh (1986) found cv. Uthirancotta to be the most
susceptible cultivar, while cultivars such as Balancotta, Belantung, Cheriaka-
niakkadan, Jambi, and Kalluvally are less susceptible to root-knot nematodes
under field conditions. Mustika (1990, 1991) found cv. Kuching to be tolerant
to M. incognita and R. similis, compared with Kalluvallay, Jambi, and Cunuk.
In India, none of the cultivars tested were found to be resistant to M. incognita
and R. similis compared with Kalluvally, Jambi, and Cunuk. In India, none of
the cultivars tested were found to be resistant to M. incognita (Jacob and
Kuriyan 1979a; Koshy and Sundararaju, 1979; Ramana and Mohandas, 1986)
or R. similis (Venkitesan and Setty, 1978; Ramana et al., 1987b). However, it
has been found that among the Indian cultivars, Pournami, which is a selection
from the germplasm collection at the Indian Institute of Spices Research, is
tolerant to M. incognita. This has been supported by field evidence that shows
that where this cultivar is cultivated, only a smaller population of nematodes
could thrive.
Another possibility in breeding resistance to nematode infestation is the
development of resistance from related wild species into the cultivated ones. The
related wild species, P. colubrinum and P. aduncum, are highly resistant to
M. incognita (Ramana and Mohandas, 1986; Paulus et al., 1993). Additionally,
P. hymenophyllum and P. attenuatum showed remarkable resistance to R. similis
(Venkitesan and Setty, 1978). Though P. colubrinum has been found to be
immune to R. similis (Ramana et al., 1994), crosses between P. nigrum and P.
colubrinum have been unsuccessful until now. The solution for cross transfer of
resistance from resistant to cultivated species might be found through the
biotechnological route, but, as of now, it still is a very challenging route. The
gene transfer mechanism in pepper is still very poorly understood.
In addition to the above, there is also the possibility of employing biological
organisms, as in the case of controlling diseases, to control nematode infestation.
Sewell (1965) defines biological control as “the induced or natural, direct or
indirect limitations of a harmful organism or its effects by another organism or
group of organisms”. Since a variety of microorganisms inhabit the soil, some of
which are either predatory or antagonistic to plant parasitic nematodes, such a
course of action could be taken. However, the question of nematode control
biologically is still a very open one.
Nevertheless, this approach could be a part of the overall integrated pest
management in pepper. As of now, the research efforts in the area of biocontrol is,
indeed, very sparse. A few examples, given below, illustrate the efficiency of
biocontrol.
In the area of biocontrol, fungal agents are the most important. Among them,
an opportunistic fungus, namely, Paecilomyces lilacinus, is an important one that
has great potential in the control of root knot and cyst nematodes. The fungus, on
contact with the egg masses of the nematodes, colonizes and grows rapidly. The
chitinolytic enzymes produced by the fungus help penetration of the eggs and
cause the suppression of nematodes. The efficacy of the fungus in controlling root
knot nematodes has been reported by Jatala (1986). Inoculation of the fungus into
rooted cuttings or seedlings of pepper significantly reduced the growth of the root
knot nematode, which in turn, reduced damage to the roots (Freire and Bridge,
1985d; Ramana, 1994; Sosamma and Koshy, 1995). However, it was less
efficient against the control of the borrowing nematode (Geetha, 1991; Ramana,
1994).
The next fungus that is effective against nematodes is Trichoderma spp. It has
been found effective in controlling root knot nematode (Eapen and Ramana,
1996), in addition to suppressing the P. capsici, as detailed earlier in this review.
Verticillium chlamydosporium is another fungus for control of root knot and cyst
nematodes (Kerry, 1990). The fungus colonizes the rhizosphere, infects adult
females and egg masses and reduces nematode multiplication by inhibiting the
hatching of eggs. Parasitization of root knot nematode eggs by the fungus in
Brazil has been reported (Freire and Bridge, 1985d). The occurrence of the
fungus in association with Trophotylenchulus piperis in pepper plantations was
first reported in India by Sreeja et al. (1996). In an in vitro test, the fungus was
found to colonize the egg masses of M. incognita and reduce hatching by almost
50% within five days of inoculation (Sreeja et al.,1996). In addition to these, the
VAM also has a suppressive effect on nematode population through the indirect
effect of enhancing plant vigor, which has been referred to earlier in this review.
Occurrence of VAM fungi on pepper roots was first reported in India by
Manjunath and Bagyaraj (1982). Bacteria come after fungus as nematicidal
agents. Within the bacteria is Pasteuria penetrans, which is an obligate parasite
of some nematodes. It infects the nematodes by direct penetration through the
cuticle by germinating spores sticking to the body surface of the nematodes. It has
been found to suppress both M. incognita and R. similis. In addition, Bacillus
pumilis, B. macerans and B. circulans also suppress M. incognita (Sheela et al.,
1993). Eapen et al. (1997) reported the inhibitory effect of fluorescent
pseudomonas, such as Pseudomonas fluorescens, on M. incognita. It is important
to realize that research in the area of biocontrol of nematodes is in its infant stage
and a number of studies are done in in vitro conditions and careful evaluation has
to be done before large-scale field recommendations can be made that are
applicable to pepper plantations. The experience, so far, has been promising,
though in a limited sense, and much more work needs to be done to probe the area
further.
Perhaps the most effective means of controlling nematode infestation is through
the use of chemicals. Nematicides — chemicals used to kill nematodes —
consist of two groups, namely, systemic nematicides and soil fumigants. Of the
two, the latter is more effective as nematodes generally inhabit the soil and once
320 K. P. P. NAIR
the nematicide is applied to the soil it is absorbed by the roots of the host plant
and translocated to shoots where they act as systemic inhibitors entering into
biochemical reactions within the host plant cell. Despite the effectiveness of
nematicides, their widespread use is not in vogue because of high costs and the
environmental adverse fallout from their use. Various nematicides, such as
Phenamiphos (Nambiar and Sarma, 1980), Aldicarbsulphone (Venkitesan and
Setty, 1978), Phorate and DBCP (Venkitesan and Charles, 1980) were found to
be effective. Among the fumigants, DD, methyl bromide and ethylene di bromide
have been found to be effective. With the advent of organic farming and greater
awareness among consumers of products of chemical-free agriculture, use of
nematicides is being increasingly restricted. Hopefully, with more research to
follow, better and safer nematicides will come into the market.
Next to the plant pathogenic diseases and plant parasitic nematodes as
pepper pests, comes the infestation by insects. Though a number of insect
pests attack the pepper plant, the most devastating is the “Pollu” beetle,
L. nigripennis Mots, though to a lesser extent others, such as, top shoot borer
(Cydia hemidoxa Meyr.), leaf gall thrips (Liothrips karnyi. Bagn.), and scale
insects (Lepidosaphes piperis Green and Aspidiotus destructor Sign.) also
could be considered as pests of major importance. However, this review will
primarily focus on the Pollu beetle. The insect pest derives the name Pollu
because of the hollow berry that results from its infestation, the word Pollu
in the regional Malayalam language of the state of Kerala meaning hollow.
The status and control of major insect pests in India have been reviewed
(Devasahayam et al., 1988; Premkumar et al., 1994).
The extent of attack by Pollu beetle varies and reports of damage varies from
6 to 21% (Thomas and Menon, 1939), while Rehiman and Nambiar (1967)
report damage as much as 30 –40% in the state of Kerala. Survey of one of the
major pepper growing districts of Kerala, Kannur district (where incidentally
the first research station in pepper in the world was established in the town
of Panniyur) showed that the single greatest cause for loss in pepper yield
was due to infestation by the Pollu beetle amounting to as much as 13%
(Prabhakaran, 1994).
The beetle measures about 2.5 mm £ 1.5 mm and feed on tender shoots,
leaves and spikes. When infestation is severe, leaves and spikes rot and drop
down. The grubs do the most damage. They bore into tender berries and empty
their contents, and the infested berries, in turn, become chlorotic first and then
turn black and crumble when pressed. Among the pepper cultivars, wide variation
has been observed inasmuch as susceptibility or resistance is concerned. Among
the four cultivars in which field observations were recorded, Kalluvally and
Karimunda were least susceptible to the pest in the northern and southern regions
of the state, respectively (Premkumar and Nair, 1988).
Inasmuch as control of the insect pest is concerned, both biocontrol and
chemical control assume importance in addition to plantation management.
Only few natural organisms exist which have a control trait on the Pollu beetle.
Among these, an unidentified entomophagous nematode (Mermithidae family),
a predatory spider (Araneae family) and Oecophylla smaragdina Fabr.
(Formicidae family) have been indentified. The extent of parasitization or
predation by these on the Pollu beetle is rather limited and their efficacy in
controlling the insect pest in the field conditions is open to question
(Devasahayam and Koya, 1994). Within the ambit of plantation management,
application of insecticides is the only viable proposition to effectively control the
Pollu beetle. A number of insecticides have been evaluated in the field for their
efficacy in controlling the pest. On balance, endosulfan spray (0.05%) has been
found to be most effective in the state of Kerala (Premkumar et al., 1986;
Nandakumar et al., 1987; Premkumar and Nair, 1987a). It must, however, be
noted that of late, there has emerged widespread opposition from environmental
activists in the state of Kerala to the use of endosulfan because of suspected
human health hazards, such as mongolian births, central nervous system (CNS)
disorders and even blood cancer. These maladies have been noted in the northern
part of the state of Kerala and the insecticide has been currently banned from use.
Application of insecticides, like that of fungicides or nematicides, leave
pesticide residues in the berries and with increasing health awareness within the
country and outside of it, consumers are opting for organically grown pepper.
Though biocontrol is a promising avenue, results obtained so far are of a very
preliminary nature and no hard and fast recommendations have emerged. A
number of byproducts have been used, both in vitro and in situ in the field.
Among them, mention must be made of neem oil, neem seed extract (Azadirachta
indica A. Juss.), nuxvomica (Strychnos nux-vomica L.) and also that of custard
apple (Anona squamosa L.) and the success rate has not always been very
consistent (Devasahayam and Leela, 1997), though they show potential for
further research (Devasahayam and Anandaraj, 1997).
As far as Malaysia and Indonesia are concerned, pepper weevil (Lophobaris
piperis) is the most serious insect pest. The grubs bore into the nodal region of
the climbing and flowering shoots, which result in infested plants. As the
infestation advances, the entire aerial portion of the vine wilts and collapses.
The infestation of the pest in the lower part of the vine results in the most
damage, amounting to a yield loss of around 50% at most and around 5% at
the least, in the upper part (Deciyanto and Suprapto, 1992). A number of
natural predators have been identified of which Spathius piperis has been
found to be the most effective in controlling the pest both in the laboratory and
also in the field. As much as 40% control has been reported (Deciyanto and
Suprapto, 1992). A number of insecticides have been recommended for the
control of pepper weevil. These include the spraying of endosulfan (0.2%),
parathion (0.2%), permithrin (0.2%), etc. Removal of affected branches and
stems would also help in reducing the level of pest infestation in the field
(Kueh, 1979; Deciyanto and Suprapto, 1992).
322 K. P. P. NAIR
VII. THE PROCESSING OF BLACK PEPPER ON-FARM
The processing of pepper starts from the harvesting operation, which is a long,
drawn out process. With the onset of the southwest monsoon in the months of May
and June the pepper plants start to flower. The harvest commences in December
and spills over into January of the following year when there is ample sunshine and
the rains completely cease. Varietal differences and the prevalent climatic pattern
decide the harvesting process. In the plains of the state of Kerala, the early
maturing varieties reach the harvest stage by November and the late maturing ones
by December – January. Harvesting of pepper is important in view of the end-
products, which have a very great industrial significance. Govindarajan (1979) has
made a detailed and systematic analysis of the harvesting dates against the back-
ground of the end-products as detailed in Tables IX and X.
Data in Table X give a clear idea how the chemical composition varies in
relation to maturity dates.
Table IX
Harvest Schedule in Relation to the End Products of Pepper
End products Time to harvest
Black pepper Fully mature and nearly ripe

White pepper Fully ripe
Canned pepper Four to five month old berries
Dehydrated green pepper Ten to fifteen days before full maturity
Oleoresin Fifteen to twenty days before full maturity
Pepper oil Fifteen to twenty days before full maturity
Pepper powder Fully mature berries when starch content peaks
Table X
Changes in Chemical Composition of Two Important Cultivars in Relation to Maturity Dates
Maturity date
Months after fruit setting
Cultivar Panniyur-1 Karimunda
End products 3.0 4.5 5.5 6.5 7.5 3.0 4.5 5.5 6.5 7.0
Volatile oil (%) 6.4 7.6 6.3 2.8 2.0 6.8 10.4 8.2 4.4 3.6
NVEE (%) 8.7 8.8 8.7 8.1 7.8 10.3 9.7 8.6 7.5 7.4
Piperine content (%) 1.9 2.6 2.7 3.1 3.5 1.9 2.4 2.4 3.1 3.1
Starch content (%) 2.5 3.7 5.1 10.2 16.8 2.6 4.9 6.2 15.3 15.3
Note: Non-Volatile Ether Extract (NVEE).

Harvesting commences mainly based on the visual appearance of the berries.

When a couple of berries in a spike turn orange or red in color harvesting
commences. Only the mature spikes are harvested. The immature spikes when
dry tend to produce shriveled berries and this in turn will reduce the quality of the
produce. However, in Sri Lanka, premature harvesting is done and the produce is
generally known as “light pepper” (Purseglove et al., 1981), which has a
premium in the market because it contains a higher amount of oleoresin and
volatile oils. Harvesting is done manually by climbing up the vine using a
bamboo pole and the produce is collected in a jute bag tied to the back of the
climber. Only in flat plantations can mechanical ladders attached to trolleys be
used for harvesting. In India, pepper plantations are usually on slopy terrain and
such harvesting can only be done by manual labor.
Once harvesting is completed and the produce brought to the farmyard, post-
harvesting operations have to begin. Proper post-harvesting handling of the
produce is very crucial to maintain the quality of the produce.
The first step in post-harvest handling is “decorning” or removal of the berries
from the stalk. It is synonymous to the threshing of wheat or rice grains. In most
of the situations, homestead farming and also large-scale plantations, decorning
is still done manually by spreading the produce on a clean floor in the farmyard
and laborers, mainly women, thresh employing their feet to trample the spikes to
separate the berries from the stalk. It is not a very hygienic way of doing it, yet, in
most Indian situations there seems to be no alternative to it. Mechanical threshers
are a rarity, while they are in use in Indonesia and Malaysia. When mechanical
threshers are used, the green spikes are slowly fed into the thresher in which a
rotating drum with aluminum blades removes the berries from the stalks. Speed
of the rotation is adjusted to be neither too slow nor too fast. The spikes are fed
between a moving drum with aluminum plates with suitable gaps and due to the
mild stretching action the berries are separated from the stalk. A thresher
powered by a 3 hp motor can handle 1.5 tons per hour. Before decorning it should
be ensured that the spikes are properly washed clean of all adhering dirt. Once the
berries are threshed, they should be graded using a mesh and normally three
grades are in vogue, namely, . 4.25 mm (large), 3.25 –4.25 mm (medium) and
, 3.25 mm (small). The first released hybrid Panniyur-1 belongs to the large size.
After grading, the berries are blanched and this is done in boiling water by
dipping the berries carried in bamboo baskets in boiling water contained in
troughs. This is done to clean the produce of all adhering dust particles and any
other extraneous particles, such as bird excreta. The produce is then dried to a
uniform moisture content (8 –10%) in open sunlight for three days consecutively.
While in India, blanching is done in boiling water for 1 min, in Papua New
Guinea, Indonesia and Micronesia it is done at 828C for 2 min (Pruthi, 1992). The
entire operation of washing, blanching and grading can be mechanized by
spraying boiling water onto pepper fruits as they move in a conveyor belt fitted
with mechanical brushes and the blanching time adjusted through the speed of
324 K. P. P. NAIR
the conveyor belt. At the end of the conveyor belt the blanched and washed
berries pass through sieves of various dimensions to separate “pin head” (good
and well sized berries) and light pepper. The graded pepper is later flushed with
dry air to remove surface moisture and then transported to the drying yard or to
artificial driers.
The phenolase enzyme which imparts the black color to the berries is activated
by blanching. It also ruptures the cells and thereby accelerates the escape of
moisture from the inner core and simultaneously enhances the black color with
the resinoids inside the berry. Therefore, blanched pepper will shine more and
dries at a faster rate. The black color which the berries acquire on drying is due to
the oxidation of colorless phenolic compounds present in the skin. Polyphenolase
(0-diphenol oxidase) present in the fruit wall converts colorless phenolic
substrates (3,4-dihydroxy phenyl ethanol glycoside) present in the cells to black
polymeric compounds (Variyar et al., 1988).
A. SUN DRYING OF PEPPER
Sun drying when done reduces the initial moisture content in the berries from
around 65% to within 10%. Most of the pepper growing countries practice sun
drying. To check the development of mould, the berry heap, which is spread on a
clean floor for drying, is periodically turned over. On average, sun drying leads to
a recovery of 29 – 38%. Non-uniformity and contamination by microorganisms
are the major disadvantages of sun drying. Pepper is dried on different surfaces,
such as bamboo mats, cement floors or black low density polyethylene sheets. It
has been observed that the black surface of the sheet, due to absorption and
retention of heat results in faster drying than when other surfaces are used
(Krishnamoorthy and Zachariah, 1992). In the Waynad district, in the state of
Kerala, a special type of drying is practiced where after two days of sun drying on
bamboo mats or a cement floor, the dried berries are collected in jute sacks and
stacked upright for two days in rooms. During this process of storing a
fermentation process sets in which imparts a uniform black color to the berries
leading to enhanced flavor. After two days of storage the berries are, once again,
spread on the cement floor or bamboo mats and dried in the conventional manner.
It is for this special type of drying that the pepper of this region is highly regarded
for its unique quality of color and flavor.
B. SOLAR DRYING OF PEPPER
Solar driers have been developed to dry agricultural products in India (Kachru
and Gupta, 1993) at the Central Institute for Agricultural Engineering in India
and in Germany at the Institute of Agricultural Engineering for the Tropics and
Table XI
Chemical Composition of Dried Pepper (in Percentage)
Chemical constituent Composition
Moisture 8.7– 14.0

Total nitrogen 1.5– 2.6
Volatile ether extract 0.3– 4.2
Non-volatile ether extract 3.9– 11.5
Alcohol extract 4.4– 12.0
Starch (acid hydrolysis) 28.0–49.0
Crude fiber 8.7– 18.0
Crude piperine 2.8– 9.0
Piperine 1.7– 7.4
Total ash 3.6– 5.7
Acid insoluble ash 0.03–0.55
After Pruthi (1993).
Subtropics, University of Hohenheim (Esper and Muhlbauer, 1996). In the solar

drier the change occurs mainly in the final drying phase, whereas sun drying
requires several days to reach the desired moisture level. Solar drying accelerates
the drying process leading to a considerable reduction in drying time. Additionally,
the maintenance of a constant temperature throughout the drying process ensures a
clear product free of microbial contamination or any other extraneous matter.
Drying of berries can also be done using mechanical driers. Different types are
in use, such as the “Copra Drier” which is used to dry coconut shell, husk etc., a
convection drier, which is a forced draft drier, a cascade type drier using indirect
heating by kerosene or gas, etc.
After drying, recovery normally does not exceed 70%. Dry recovery varies
from 29 to 38% depending on the variety. Table XI shows the composition of
dried pepper.
C. GARBLING, CLEANING, AND FRACTIONATION
Dried pepper is cleaned to get rid of dirt, grit, stone, leaves and any other
extraneous matter before packing. Pneumatic separators equipped with magnetic
separators are used to remove metallic contaminations, such as iron filings, nails,
etc. For destoning of spices a combination of air-classification and vibratory
conveying using inclined docks is very efficient (Ramanathan and Rao, 1974).
Some of the well established processing houses clean and grade pepper with the
help of multiple sieve-cum-air classifier type of machines whereby dust, stalks,
pinheads, hollows, immature pepper, red pepper and extra-bold pepper are
removed (Pruthi, 1992, 1993). Separated pepper is then washed and dried to make
it free of adhering fungus and other extraneous matter. Hence, to obtain good
326 K. P. P. NAIR
quality pepper of international standard the steps involved in processing are:

(1) drying, (2) separation of various fractions, (3) size segregation or grading,
(4) physical cleaning (washing and drying), and (5) packing. In India, farmers as
well as intermediaries bring the produce to the market usually in an ungraded
form and generally garbling and grading are done at the exporters’ premises.
Most of the exporters clean and grade using the garbling machines which remove
dust and chaff and grade the pepper according to densities. However, most of the
machines do not remove iron filings and heavy metals. These are generally hand
picked. Usually pepper collected by the large merchants (exporters) is first
cleaned manually by picking out the contaminants by hand, which is mostly done
by women. After this the pepper goes for garbling. A spiral separator has been
developed by Madasamy and Gothandapani (1995). This ideally removes all
extraneous materials. Traditionally, pepper is manually cleaned by winnowing
and then packed into jute bags. In some plantations cleaning is done through a
funnel-like arrangement in which dried pepper is fed and cleaned using a blower
at the outlet point when the pin heads and light berries are blown away.
The Directorate of Marketing and Inspection, Government of India, has
classified pepper into the following grades:
1. Malabar Garbled (MG Grades 1 and 2)

2. Malabar Ungarbled (MUG Grades 1 and 2)
3. Tellicherry Garbled Black Pepper Special Extra Bold (TGSEB)
4. Tellicherry Garbled Extra Bold (TGEB)
5. Tellicherry Garbled (TG)
6. Garbled Light Pepper (GL Special, GL Grades 1 and 2)
7. Ungarbled Light Pepper (UGL Special, UGL Grades 1 and 2)
8. Pin Heads (PH Grade Special and PH Grade 1)
9. Black Pepper (Non-specified) (NS Grade X)
Graded pepper fetches a higher price than ungraded, and in the international
market TGSEB fetches a premium. Good grabled pepper should have a bulk
density of 500 –600 g l21. Light berries should be less than 10% and pin heads
(unfertilized berries) less than 4%. When bulk density is low it indicates the
presence of more lightweight berries, with consequent less starch content, leading
to poor milling quality. Only when the end product has a good aroma and a biting
taste will it be considered good. The end product should contain at least 1.5%
volatile oil and 3% piperine (Lewis, 1984). Table XII contains data pertaining to
the physico-chemical composition of different grades of black pepper.
There is another grade of pepper known as “Half Pepper” which is placed in
between normal pepper and light pepper. The berries are slightly under mature
and therefore contain more of the active principle piperine. Due to immaturity the
berries may appear slightly wrinkled and this is ideal for extraction of oleoresin.
Table XII
Physico-chemical Composition of Different Grades of Black Pepper (in Percentage)
Grade Moisture Volatile oil Piperine NVEE Starch Crude fiber
MG 13.0 3.7 5.0 12.3 39.7 11.8

MUG 12.0 2.8 5.0 11.4 41.8 12.5
TGSEB 10.0 3.2 4.9 10.3 40.9 9.2
TGEB 13.0 2.2 4.4 9.1 39.7 11.8
Light pepper 13.0 2.9 4.1 13.5 14.6 27.8
Tellicherry ungarbled 12.0 4.0 6.3 13.5 39.3 11.0
Pin heads 13.0 0.6 0.8 7.1 11.5 27.4
High range ungarbled 12.0 2.6 4.0 11.1 41.8 10.5
Note: Non-volatile ether extract (NVEE).
The concentration of piperine in oleoresin is more important than the aroma,

which is determined by the oil content.
Using appropriate sieves the pepper is graded into different sizes. Very
small and underdeveloped berries are classified as “Pin Heads”. The bulk of the
pepper belongs to the average size which is known as “Malabar Garbled”.
The larger-sized ones are classified as “Tellicherry Bold”, “Tellicherry Extra
Bold”, “Tellicherry Special Extra Bold” and “Giant” (Kachru et al., 1990;
Mathew, 1992).
In normal pepper starch accounts for 34%, while it is 56.5% in white pepper
and 63.2% in decorticated white pepper. Of the 12% water-soluble nitrogen
present in the berries non-protein nitrogen constitutes about 82% and of this more
than 50% is made up of simple amino acids which act as a nutrient for humans
when pepper is consumed as a culinary material.
One of the most worrisome aspects of pepper processing is the adulteration
practiced by some of the unscrupulous merchants involved in the pepper trade.
When such adulterated pepper is exported and the adulteration subsequently
detected at the port of entry, it takes not only a heavy toll on the credibility of the
country exporting, but, with the recent provisions in the World Trade
Organisation (WTO) it can lead to international disputes. One of the most
common practices of adultering pepper is mixing it with papaya seeds, which
look exactly like pepper (Hartman et al., 1973; Sen and Roy, 1974).
D. PACKAGING AND STORING
One of the most important precautions to be taken in packaging pepper is

that, since it is hygroscopic in nature, it can easily absorb ambient moisture and
when such moisture-laden pepper is stored without care it will lead to the
formation of mould and insect infestation will follow as pepper contains a good
328 K. P. P. NAIR
amount of starch. Mould and insect infestation can lead to loss of aroma and
caking will result along with hydrolytic rancidity. Whole pepper is generally
packed and transported in jute bags (burlap bags) or polyethylene lined double
burlap bags. Dried pepper having 10 –11% moisture can be stored without the
mould infestation in jute bag with polyethylene lining or in laminated high-
density polyethylene bags (Balasubramanyam et al., 1978). Storing of pepper
should be done with great care and moisture should be in the range of 10 –11%
before storage. Store houses should be damp proof and should be free of rodent
attack, with controlled ventilation and devices for both humidity and temperature
control. It is important to fumigate the room before storage. The room walls
should be whitewashed with slaked lime and should be exclusively used for
storing pepper.
Another aspect which will impact the processing of pepper is the
microbiological contamination. Investigations conducted in Malaysia have
indicated that pepper berries collected from various farms had a total viable
count (TVC) between 105 and 1072g sample (Apun et al., 1993). The TVC dropped
on drying. TVC was found to be less when the produce was mechanically threshed
as compared to when threshing with feet. Washing the spikes before threshing and
blanching in boiling water minimizes the microbial load.
White pepper is one of the major pepper products consumers all over the world
prefer. Retting ripe berries in running water for about 7 –10 days is done to make
white pepper. The skin of the berries undergoes bacterial degradation and the
softened skin is removed by either decortication or trampling with feet as
explained earlier (Pruthi, 1993). An investigation done in Sarawak indicated that
pepper berries which have a specific gravity . 1.12 are the best for conversion to
white pepper and are not as good for making ordinary pepper (Anon, 1995).
VIII. AN ACCOUNT OF INDONESIAN

PEPPER PROCESSING
With Indonesia being a market leader in pepper, it will be informative to

review, though briefly, the processing of pepper in that country, more so because
the IPC Office is located in Jakarta, the capital of the country. In Indonesia most
of the pepper is grown by subsistent farmers. To improve pepper processing
technology, the Research Institute for Spice and Medicinal Plants has carried out
research on equipment design meant for pepper processing which has relevance
to on-farm processing. Immediate post-harvest treatment of pepper in
Indonesia is quite similar to that practiced in India, but the final cleaning,
grading and packing of the dried pepper is carried out by exporters. This is mainly
because the small-scale farmers do not have the infrastructure to carry out the
various aspects of processing. Most of the exporters have a complete set of
winnowing and shaking machine (for removal of extraneous materials), washing

machine, hot air dryer and screw type separation machine for size-separation.
The entire unit can process around 5 tons of black pepper per hour. Grading is
normally done according to requirements of overseas buyer and double-lined jute
bags are used for packing. Packed bags are stored in “go downs” prior to
shipment.
In the traditional method of making white pepper, the hygienic and quality
aspects are the major problems. The berries are usually soaked in rivers or ponds
with poor quality water. Microbial contamination is a major problem, including
contamination with waste and the product acquires a swampy odor that is difficult
to eliminate later. In addition, soaking in water for a long time results in the loss
of volatile oil and deterioration in aroma and flavor. The quality of white pepper
in the whole and ground forms is imparted by appearance, aroma and pungency.
Blanching or steaming fresh green berries for 10 –15 min for decortication and
then removing the rind in a fruit pulping machine has been tried (Purseglove et al.,
1981). Decortication by soaking in boiling water for 15 –25 min has been tried
and this led to the white pepper having a better aroma than when prepared by the
traditional method (Risfaheri and Hidayat, 1996). Prolonged soaking in boiling
water will lead to loss of volatile oil and adversely affects the aroma of white
pepper as prolonged soaking in boiling water is known to damage the flavor
profile of pepper (Pruthi, 1992). Soaking in boiling water softens the pepper fruit
wall and makes decortication easy.
Pepper oil extraction is an important aspect of pepper processing. It is
produced by steam distillation of pepper berries. Pepper oil does not have
piperine, which imparts pungency, and so is valued for its aroma in the fragrance
industry as well as the flavor industry. Pepper oil is used in high-grade perfumes
and the toiletry industry. In the flavor industry, it is primarily used in foodstuffs
requiring a high pepper aroma. Volatile oil derived from steam distillation is
almost colorless to slightly greenish. Mild in taste with no pungency at all,
Indonesia has put much emphasis on pepper oil production and its export.
Oleoresin, another by-product of pepper having odor, flavor and pungency, is
obtained by extraction of the berries using organic solvents. The organoleptic
properties of oleoresin are determined by its volatile oil and piperine contents and
their abundance primarily depends on the raw material used for extraction.
Ground pepper can be extracted with pure organic solvents, such as acetone,
ethanol, dichloroethane, etc. When freshly made, pepper oleoresin is a dark
green, viscous heavy liquid with a strong aroma, but, on standing, crystals of
piperine appear and the oleoresin requires mixing before use to ensure uniformity
and consistency. The components and the quality of the extract depend on the
raw material used and research in Indonesia indicates that light black berries are
better than pin heads for oleoresin, non-volatile ether extract (NVEE) and
piperine (Mapiliandri, 1989). Oleoresin offered for sale by some of the
principal producers contains 15 – 20% volatile oil and 35 – 55% piperine.
330 K. P. P. NAIR
Purseglove et al. (1981) have reported that as much as 1 kg of oleoresin can be

obtained from 8 kg of black pepper. From a review of research on
extraction technology it can be concluded that further research in this area is
needed to minimize oil loss and fractionate oil for further blending to produce oil
and oleoresin with specific quality characteristics.
IX. INDUSTRIAL PROCESSING OF BLACK PEPPER
Industrial processing of pepper has to be followed subsequent to on-farm

processing before the final product is ready for human consumption. In the pepper
processing plants, the pepper that has earlier gone through one round of on-farm
processing, where drying and grading to different standards has been done, goes
through further processing. The dried pepper is passed through mechanical sifters
for removal of pin heads, vegetable seeds, fine dust, sand etc., all of which are
extraneous materials most likely to be present, before winnowing and destoning
for removal of stalks, dust, light foreign matter and stones. Multiple sieve-cum-
air classifier type of machines and gravity separators are used for this purpose.
The pepper that is separated is then washed in mechanical washers fitted with
brushes to remove mould and dust and this gives a luster to the pepper berries.
Then the product is centrifuged to remove adhering water and then dried in
electric or diesel-fired indirect dryers. The dried pepper is then sent through
spirals for final cleaning followed by sterilization either by steam or gamma
radiation before being packed. When a pneumatic conveyor-cum-dryer-cum-
grader is used, the entire operation of drying and grading can be done in
sequential steps, consequently saving a lot of time.
There are different grades of pepper. Grading is done on the basis of size and
shape of the berries. As per the International Standards Organisation (ISO) the
following three grades of pepper exist:
1. Non-processed (NP)
2. Semi-processed (SP)
3. Processed (P)
There are permissible limits, set by ISO, for contaminants or light berries, pin
heads, broken berries, etc., in the final product and its bulk density according to
ISO specification (ISO/DIS 958-1, 1996). However, individual countries have
their own grades and these have been detailed earlier. Sterilization is important to
ensure good quality and freedom from microbial contaminants. Several methods
of sterilization are available, such as hot air/steam sterilization, extrusion,
hydrostatic/pressure sterilization, ozone sterilization, compressed carbon diodix-
ide treatment, irradiation, microwave heating, alcohol treatment, etc. In hot
air/steam sterilization the product is preheated to 50– 558C and subsequently

sterilized by a combination of indirect heating and direct steam injection. The
process substantially reduced microbial contaminants while not adversely
affecting the flavor, aroma or oil content (Schneider, 1993). Yet another process
of steam sterilization involves exposure to high pressures and temperatures for a
pre-determined time period in a series of chambers. The material to be sterilized
is placed in the first chamber and high pressure exerted and later transferred to
subsequent chambers following a similar procedure for specific time periods and
finally the sterilized material is depressurized in a chamber (Dudek, 1996). There
is a modified steam sterilization process that consists of a steam-jacketed
pressurized vessel with three temperature envelopes, namely, 50, 100 or 2008C
for different products. Since the system is pressurized, there will be no loss of oil
or moisture and the appearance of the product remains unchanged. Steam,
compressed air and nitrogen gas may be used depending on the end product
desired (Darrington, 1991). A continuous steam sterilization process developed
involves subjecting the pepper to a rapid flow of superheated steam during a pre-
determined time period followed by drying, re-humidification and packaging.
Both enzyme and microbial activities are significantly reduced to low levels and
no loss of flavor or aroma or oil was noted (Uijil, 1992). Leife (1992) describes a
system of steam sterilization where pepper is subjected to rapid pulses of steam,
and the steam condenses on the surface of the berries and contaminants are
removed. Steam sterilization is the ideal method in situations where, due to
national safety restrictions, both chemical and irradiation methods are prohibited.
Yet another method of sterilization is gamma irradiation. This method is
particularly useful against infestation by insects. As the process is a cold
treatment, loss of volatile oil, flavor and aroma are practically nil (Nair, 1993).
A low dose of 1 kGy (kiloGray) is sufficiently effective against insect infestation
whereas a 10-fold increase is needed to control microbial infestation. The process
does not lead to accumulation of any harmful residues unlike fumigation.
Depending on the source geometry and the conveyor system used, the size and
shape of the containers will be decided. Codex General Standards for irradiated
food and recommendations of the International Code of Practices for the
operation of irradiation facilities gives the guidelines for irradiation procedures.
It is important that Codex Standards are followed to facilitate international trade.
Pepper subjected to a radiation level of up to 10 kGy does not lose its aroma and
flavor, and there are no significant changes in the oil or piperine contents. When
the produce is heavily contaminated, irradiation up to 30 kGy is needed. Countries
such as USA, France, Austria, Germany, etc., have approved this irradiation
procedure for sterilization and a commercial application of the technology is
followed. More than 95% of the gamma irradiation facilities operating in the world
use Cobalt60. Gamma irradiation facilities are capital intensive and the unit cost of
processing is highly sensitive to the scale of operation. To minimize operational
costs, facilities need to operate round the clock at the maximum rated throughput
332 K. P. P. NAIR
that the facility is designed for. A gamma irradiator is simple to operate and
maintain. Chemiluminescence, thermoluminescence, and free radial dosimetry are
employed as routine methods for the detection of irradiated pepper. Other methods
in use are colored indicators on the containers which are to be irradiated, viscosity
measurements of gelatinized pepper and intensity measurements of reflected
signals of the electron spin resonance (ESR) spectrum.
Two other methods of sterilization are chemical and through the use of
microwave. Fumigation with ethylene dibromide to disinfect insect contami-
nation followed by fumigation with ethylene oxide or propylene oxide for
eliminating microbial contamination is the most commonly used chemical
sterilization process. Moisture content in the berries, temperature, time of
contact, and concentration of the fumigating gas decide the effectiveness of the
sterilization process. Fumigation results in the reduction of both volatile and non-
volatile oil contents. The main problem with the use of these fumigants is their
explosive, toxic, and irritant nature, and as such, unsafe when used without
adequate safety measures for the workers who are involved. Additionally,
harmful residues are left behind following fumigation. On account of these
problems many countries have banned the use of these fumigants.
The main advantage with microwave sterilization is that the microbial count is
drastically reduced. Microwave heating is done at 2450 MHz and high frequency
heating at 27.12 MHz. With an increase in moisture in the samples, microbial
efficiency is increased, but there are losses of both volatile and non-volatile oils.
Both chemical and microwave fumigation are comparable in their efficiency.
A number of value added pepper products of industrial origin exist. The first is
dehydrated green pepper. This product is prepared from immature green pepper
fruits of appropriate varieties by industrial processing. The fruit should be
uniformly sized with pungency, flavor, and color of green pepper. The fruits are
blanched in boiling water for a few minutes, the water drained, then cooled and
soaked in SO2 solution to fix the green color followed by drying in a cabinet drier
at 508C. Upon rehydration, the product will reconstitute to a good quality product
possessing the characteristic pungent spicy taste, color, and flavor of green
pepper, when one part by mass of dehydrated green pepper is cooked for 20 min
in the presence of 10 parts by mass of NaCl solution of 1% concentration. To
conform to international standards (ISO/DIS 10621, 1996), the product should
have a moisture content of less than 8%. The use of SO2 is a health hazard and
efforts are on way at different research institutions to find an alternative
(Sankarikutty et al., 1994). The next in line among the value added products is
canned green pepper. Of late, there has been a surging demand for canned green
pepper. The pepper fruits after removal from spikes are washed in running water
and kept soaked in water with a Cl concentration of 20 ppm for an hour. They are
then covered with 2% hot brine of 0.2% citric acid warmed to 808C, sealed air-
tight and kept in boiling water for 20 min. After this, they are immediately cooled
in a stream of running cold water. A better color was obtained using acetic acid
instead of citric acid, but, as per international standards, only citric acid is
permitted not exceeding 0.6% by mass of the packing medium and the covering
brine must be 1 –2% by mass of edible common salt. For making canned pepper,
berries harvested a month prior to the actual date of maturity are the best. Bottled
green pepper, dry packed green pepper, and freeze dried pepper are among the
other directly edible pepper products. Bottled green pepper is made by first
despiking fresh green pepper fruits of uniform size and maturity, immediately
after harvest, followed by cleaning, washing, and steeping in 20% brine solution
containing citric acid. This is then allowed to cure for 3–4 weeks. The liquid is
then drained off and fresh brine of 16% concentration together with 100 ppm SO2
and 0.2% citric acid is added. The resulting product is stored in containers kept
away from direct sunlight. As per international standards (ISO/DIS 11162, 1996),
the product will have the characteristic odor and flavor of fresh green pepper, the
color varying from pale green to green. Dry packed green pepper is produced
in the same way as bottled green pepper, with the only exception being that the
liquid at the final stage is drained off and packing is done in flexible pouches.
Dry packed green pepper, which has similar qualities to that of canned or bottled
green pepper, can replace the latter. Freeze-dried pepper retains the original
color and shape of green pepper and fetches a premium price in the market.
Its processing is a manufacturer’s secret.
A. WHITE PEPPER
White pepper is produced from fully ripened fruit after removal of the outer
pericarp either after or before drying. White pepper is preferred for use in food
products, such as colored sauces, salad dressing, soups, and mayonnaise, where
dark colored (black pepper) is undesirable. In some of the European countries it is
white pepper that is in common use. It is made by any of the following
methods: (1) steeping and retting in water, (2) steaming, (3) boiling, (4) chemical
treatment, (5) the simple decorticating process of ripened fresh or dry berries.
The water steeping and retting process is started when one or two fruits in the
spike start yellowing when the crop is harvested and ready for the treatment. The
produce after harvest is threshed and heaped in tanks through which water is
allowed to run for 7–10 days. Light-weight pin heads and light berries accumulate
on the surface, are removed and the remaining mass is rolled over at least three
times a day during the retting stage. On the eleventh day, the outer skin of the
berries is removed by gentle rubbing and the deskinned fruits (the pepper seeds)
are transferred to another tank which contains a bleaching solution. The produce
is left to stand in the bleaching solution for 2 days after which the solution is
drained, the seed washed and sun dried. The boiling or steaming processes
involve the steaming or boiling of the mature green fruits for 10 –15 min when
the outer skin of the fruits gets softened during the steaming process which is then
334 K. P. P. NAIR
removed by passing through a pulping machine. The deskinned fruit (pepper

seeds) are washed and treated with SO2 or bleaching powder solution,
subsequently washed in water and dried in the sun. The skin of the berries
recovered in this process can later be used for the recovery of pepper oil by steam
distillation though it may not be economical. The process has been developed by
the Central Food Technology Research Institute in Mysore, India.
There are a number of chemical processes to prepare white pepper by treating
the berries with acids or alkalis. As of now, no commercially viable production
unit exists. Manilal and Gopinathan (1995) have described a method by which
skin of both dried and fresh pepper is removed by microbial decortication.
By using decorticating machines, whole dried black pepper fruits can be
processed to white pepper. Since loss due to breakage will be high, production of
white pepper by this method will be expensive. The characteristic flavor and
aroma also could be missing in the final product and, as such, this method of
making white pepper is not preferred.
Among all the processes described above, the traditional retting process is the
most popular and the white pepper obtained by this method is preferred by
consumers. According to international standards (ISO/DIS 959-2, 1996) the
product should have a maximum of only 1% permissible extraneous matter, 4%
of broken berries and 15% of black berries. The minimum bulk density should be
600 g2l. On a comparative basis, while 100 kg of mature green pepper will yield
33 kg dry pepper, in the case of white pepper it will only be 25 kg. From this it
can be inferred that the cost of white pepper should be at least 35% above that of
black pepper. The total world demand for white pepper is around 38,000 metric
tons per annum, which is a quarter of the world-wide black pepper production.
There is yet another value added pepper product known as ground pepper. It
comprises both white and black pepper. Ground pepper is obtained by grinding
the normal pepper without adding any external material. Pepper has to be ground
to a specific size and for grinding, hammer mills with copper tipped hammers are
preferred to silica tipped plate mills as ground pepper obtained from the latter can
contain silica in excess of the permissible limit. The ground materials after
sieving, is packed and any above the specified size is sent back to the mill for
further grinding and size reduction.
When pepper is processed in modern spice grinding units, it is at first cleaned
of any extraneous matter and then passed through a magnetic separator for the
removal of metal particles. After this, it is passed through a vibrating screen for
further removal of extraneous matter and then sent to the hammer mill. The
ground product from the hammer mill is fed into a cyclone separator to recover
pepper powder. It is further sieved using appropriate sieves and the overflow is
recycled. Bag filters are used after the cyclone separators to prevent escape of fine
particles outside. The ground pepper is packed into air-tight containers prior to
shipping and the packages are made to specifications provided by the customer on
size, pungency, moisture content, ash content, crude fiber, volatile oil, piperine,
starch content, etc. White pepper is processed in the same manner as that
described above except that, instead of black pepper, the starting material is white
pepper. The ground pepper has a characteristic flavor, very aromatic, slightly
sharp and must be free of any extraneous odors and flavors including mouldy and
rancid odors and the international standards conform to ISO/DIS 959-2 (1996).
White pepper can also be made from black pepper using selective grinding
followed by sieving. The skin of the berries is a by-product and contains essential
oil and oleoresin, both of which can subsequently be extracted.
B. CRYOGROUND PEPPER
To cut down oxidative losses of essential oils and to assist in the fine grinding
of pepper by making the raw product brittle at low temperatures a new technique
known as cryogrinding at temperatures very much lower than the normal 1008C is
followed. The product disperses more uniformly in spice formulations. The
process involves injecting liquid nitrogen into the grinding zone. A temperature
controller maintains the desired product temperature by suitably adjusting the
liquid nitrogen flow rate. The exhausted gas is re-circulated for the pre-cooling
process.
C. PEPPER OIL AND OLEORESIN
Pepper oil and oleoresin are the two most important ingredients of pepper and
have a great economic value. The first imparts the aroma to the pepper while the
second gives the pungency and flavor. The essential oil can be recovered by
steam or water distillation. The essential oil contains mainly a mixture of terpenic
hydrocarbons and their oxygenated compounds having a boiling point in the
range of 80 – 2008C. Depending on the pepper cultivar, the agro climatic
situations in which the crop is grown, the management aspects and the grades, the
content of the oil varies. The pin heads contain the most sesquiterpenes. In order
to meet different flavor requirements it is possible to suitably blend oils from
different grades of pepper.
In industrial processing to recover essential oils, the product is first flaked
using roller mills and then steam distilled in stainless steel extractors. The
product to be distilled is first heaped into the distillation unit and then compacted
near the walls to prevent any steam escape. Through the bottom of the still, dry
steam is passed. The still also gets heated through the jacket provided for this
purpose. When the steam comes into contact with the pepper the temperature is
raised and the oil present in the oil cells of the berries vaporize, rising along with
the steam through the still. It is then condensed and since the oil is lighter than
water it floats on the surface of the condensing water. Using an oil or water
336 K. P. P. NAIR
separator the oil is then recovered. The oil recovered by steam distillation is
allowed to remain on anhydrous Na2SO4 to exclude traces of moisture. When
exposed to direct sunlight and air flow the hydrocarbons and the sesquiterpenes
present in the oil undergo oxidative changes especially during long storage.
Hence, it is always desirable to store the material in airtight containers, especially
since the end use occurs after a long time because of the process of export by
shipping to different countries.
Oleoresin is usually obtained by extraction of ground pepper with acetone,
ethanol, ethylene dichloride, ethyl acetate, etc. The main advantage of oleoresin
is that it is uniform in composition and strength. Contaminants such as mould and
fungus are absent in oleoresin and can be directly added to food stuff after
adjusting the level of concentration. Oleoresin comes in oil soluble, water
dispersible and dry forms. Currently, oleoresin is recovered either by a single-
stage or two-stage process. The process involves size reduction of the ground
pepper prior to solvent extraction done in stainless steel extractors. In the single-
stage process the oil is recovered along with the resins by solvent extraction
whereas in the two-stage process the ground pepper is first subjected to steam
distillation for the recovery of the essential oil. The composition of the oil and the
oleoresin content obtained in both the processes differ slightly from each other.
Due to the moisture present in the wet cake obtained after steam distillation, there
is the likelihood of oleoresin yield obtained in the two-stage process to be less
compared with that obtained in the single-stage process. Drying prior to solvent
extraction can prevent such loss in oleoresin yield. In the case of the single-stage
process, pepper is flaked to 1.0 –1.5 mm thickness and packed in stainless steel
extractors for extraction with organic solvent. Normally a solid to solvent ratio of
1:3 is maintained at an extraction temperature of 55 – 608C. The solvent is
continuously kept re-circulated to ensure efficient solid to solvent contact and
after 3 h of extraction the miscella is filtered and evaporated. The solids are
further contacted with lean solvents from other extractors and the entire
extraction process is completed after six or seven stages of extraction. The
miscella sent from the extractors are evaporated in a shell and tube evaporator at
temperatures not exceeding 808C in a vacuum of 250 mmHg. As the temperature
starts to rise and no further solvent is recovered from the concentrated miscella, it
is pumped into a high vacuum stripper. Final desolventization is done at a
vacuum of less than 20 mmHg and at no stage of the operation is the temperature
allowed to rise above 1008C. Initially, during vacuum distillation, the condensate
recovered will mostly be the organic solvent and towards the end of the
distillation the essential oil will also start coming out along with the organic
solvent. At this stage, some entrainer, such as alcohol or the monoterpene fraction
of the essential oil is added to the desolventiser. This is done to remove the final
traces of the organic solvent as well as to supplement the monoterpene fraction
of the pepper oil which might have been lost during the final solventization.
The product is pumped into storage tanks after confirming that the residual
solvent levels are within limits. Suitable blending is used to meet the customer
requirements in terms of both oil and piperine contents. Some customers specify
the homogeneity of oleoresin through a homogenizer which may be a colloid mill
or sand mill. The oleoresin, which is dark in color, is bleached using activated
carbon to obtain decolorized oleoresin. All over the world, oleoresin extraction is
done mostly in batch extractors. Investigations conducted at the Regional
Research Laboratory, Trivandrum, in the state of Kerala, India, showed that
extraction can be done in ambient conditions and at low residence time
(Sreekumar et al., 1993). Present day extraction, which takes more than 18 h to
process one batch, can be dispensed with employing continuous counter-current
extraction. Simultaneously, the quantity of solvent needed for extraction and its
loss in the extraction process can be substantially reduced.
Research in the laboratory has shown that extraction by enzymatic breakdown
of spice cell walls of the dried pepper can be accomplished. In this process, the
product is mixed with water, adjusting the pH by the addition of citric acid,
treated with enzymes, either individually or in combination, which is followed by
centrifugation. The enzymes used are those which are commercially available,
such as cellulase pectinase, hemicellulase, and liquefaction enzyme preparations.
Extracts having good sensory and compositional properties were obtained with
some of the enzyme combinations. Optimum results were obtained using a
combination of cellulase and pectinase preparation. The addition of hemi-
cellulase did not result in any improvement of flavor. This solvent-free extraction
procedure is an alternate route for the recovery of flavor compounds which is not
commercially possible at the moment. Another solvent-free extraction procedure
which has shown much promise is the super-critical carbon dioxide extraction.
Low energy consumption, high purity of resultant extracts, environmentally
acceptable, and possible fractional separation of the components are some of the
major advantages of this process.
The current demand for piperine is on the increase. Piperine can be produced
from the oleoresin in the concentrated form by centrifuging the oleoresin in a
basket centrifuge. Part of the oil along with some resin can be collected after
centrifugation and the centrifuged cake, which contains as much as 60% piperine,
is obtained. By washing the centrifuged cake with pepper oil and further
centrifuging, the piperine concentration can be further enhanced.
A number of secondary products have been developed from oleoresin to
improve solubility in food substances which are marked under various trade
names by the manufacturers with their flavor strength indicated on the respective
label. Standardized seasonings, which are able to withstand almost all processing
conditions, are some of the secondary products. Emulsions, solubilized spices,
dry soluble spices, encapsulated spices, heat resistant spices, fat based spices,
etc., are some of the additional secondary products. Emulsions are liquid
seasonings that are prepared by emulsifying blended pepper oil or oleoresins with
gum acacia or other permitted emulsifying agents. A stabilizer is added to check
338 K. P. P. NAIR
creaming and the products have only limited shelf life. Solubilized pepper is
blended pepper oil and/or oleoresin which is mixed with one of the polysorbate
esters in such a concentration to give a clear solution when mixed with fresh
water. Dry soluble pepper is prepared by dispersing the standardized oleoresin
onto an edible carrier-like salt, such as dextrose, rusk, etc., in order to give a
product which has a flavor strength equal to that of good or average quality
ground pepper.
In addition to the above-described, industrially produced value added by-
products of pepper, there are a few others, such as heat resistant pepper, fat based
pepper, extruded spices, and micro encapsulated flavors. Of these, micro
encapsulation is the most important. It is a process by which the flavor material is
encapsulated in a solid matrix and is ready for release as and when required.
Encapsulation can be achieved by a number of techniques, such as spray drying,
coacervation, polymerization, etc. Of these, spray drying is the most popular. The
process involves homogenization of the oil/water mixture in the presence of the
wall material and later removal of water under controlled conditions in a spray
drier. The advantage of the spray drying process over the others is that the
product, though in contact with existing gas at a higher temperature, will never
reach this high temperature within the short residence time. The oil/water
emulsion is atomized through an atomizer in the spray drier during spray drying.
Commonly used wall materials for encapsulation are selected from among
vegetable gums, starches, dextrins, proteins, sugars, and cellulose esters. The
wall material is selected so as to meet, as closely as possible, the properties, such
as low viscosity at a high solid state, ability to emulsify or disperse the active
material, non-reactivity with the material to be encapsulated both during
processing and on prolonged storage, uniform film forming property,
ready availability, etc. Investigations indicate that the addition of surfactants
during the process of encapsulation prior to spray drying will reduce the oil loss
(Anon, 1989). Another process for microencapsulation is the CR-100 process
(Anon, 1995; Findlay-Wilson, 1995; Mos, 1995) and this process overcomes
the limitations of the spray drying process such as reduction in flavor quality
and yield.
The heat resistant pepper is the double encapsulated product in which the
capsules are rendered water insoluble by a suitable coating and the contained
flavor is released only at high temperatures such as in the case of baking. Fat
based pepper is a blend of pepper oil and/or oleoresin in a liquid edible oil or
hydrogenate fat base formulated for use in products such as mayonnaise.
Extruded spices are spices sterilized, ground and encapsulated in a single step.
The process involves a combination of pressure changes, temperature shock, and
shear. The process, accomplished in a twin screw extruder, helps in retaining the
color and flavor in the original form because of the very short processing time.
When fresh spice is fed into the twin screw extruder the starchy materials get
gelatinized and form an encapsulated product. This process substantially reduces
contamination by bacteria, mould, and yeast. The product emerges from the
extruder as a “spice rope” which is then cut into pellets. Lucas Ingredient, Britain,
markets the products under the brand name “Master Spice” (Scott, 1992).
X. THE FUTURE OF THE GLOBAL PEPPER ECONOMY
The global pepper economy has different facets. The first is the supply side, the
second the demand side, the third the price projections, both supply and demand
included simultaneously, and the fourth referring to the prices and equilibrium in
the market. Bade and Smit (1992) have resorted to precise models to study the
above facets. This review does not provide an elaborate scope to study these
models, but the essential conclusions stemming from these models will be
discussed in this section. The various models are all mathematical in nature.
A. THE SUPPLY SIDE OF THE PEPPER ECONOMY
The supply side of pepper is approached on a national basis, examining the

area, production and exports from each country. The normal production is
obtained by multiplying the number of pepper vines by the average yield of a vine
in a country. Although measurement by vines must be considered much better, a
similar line of reasoning can be applied with area and average yield per hectare.
The major drawback of this method in comparison with using vines is that it adds
to a source of variation. Not only the per vine yield fluctuates, but also the number
of vines per hectare. In India this is a very significant variation, but in other
countries, for example in Thailand, the number of vines per hectare is almost the
same all over the country. If the intensity of cultivation is relatively constant and
there is no significant change in cultivation patterns, it is possible to estimate a
nation-wide normal average yield per hectare. In case the average number of
vines per hectare cannot be expected to be constant an assumption is needed
about the change. In case the intensity of cultivation increases or a superior
variety is introduced, it can then be assumed that there will be an upward trend in
the normal yield per hectare some 2 years after the start of the intensification
when the planted vines become productive. If normal yield can be used as a basis
for forecasting, the actual yield will deviate from it because of the influences of
weather, amount of fertilizer applied and the level of maintenance. Unfortunately,
these effects are not confined to only 1 year. When a year is very wet, the crop
will be small, but, most importantly, there will be also the incidence of dreaded
diseases such as Foot Rot. Invariably, these diseases will determine the yield in
340 K. P. P. NAIR
the succeeding year. Also, the effects of neglect or exceptionally good

maintenance will spread over more than 1 year.
All of the above illustrates that even when data on productive vines and
average yield are available, and quite reliable, there would be sufficient scope for
simulation and expert interpretation. However, global information is very
expensive to gather and gathering information on agriculture can even be a lot
more expensive and time consuming task. This is not a priority of the developing
countries. Planning depends largely on information and a model cannot
compensate for the lack of quality data. Inasmuch as quality of information is
concerned, it can only interpret and detect inconsistencies. Thus, in conclusion, it
can be said that the modeling of production and supply presented is only a
preliminary step on the road to a more sophisticated model analysis based on
superior data. In the following paragraphs, a brief review of the pepper economy
of important pepper producing countries, such as India, Brazil, Thailand,
Malaysia, and Sri Lanka is given.
1. India
It is a very important observation that the area under pepper in India would be
enough to meet the entire global demand if only average yields were a third of
what is obtained in Sarawak (Malaysia). This points to very important lacunae in
data collection, their precision and reporting. Except in large pepper plantations
in the state of Karnataka and Northern Kerala, pepper is still grown in homestead
gardens as an intercrop between coffee, cardamom and quite often trailed using
arecanut and coconut trees as standards. Up to 1986 a survey was conducted by
field extension agents in Kerala in randomly chosen parts of Kerala in such a way
that within 5 years every part was visited once. The total area under pepper
from the population was then multiplied by the inverse of the sample area
divided by state area ratio. The question asked was to estimate area on the basis
of 560 vines ha21. The method was applied, asking the same people, to get
production estimates. From 1987 onwards, the Department of Economics and
Statistics of the state is attempting to introduce a more sophisticated system
especially to estimate production. Though the method used is rather elementary,
other countries probably do not use better systems. A similar equation as that
used in India is used in Brazil as well.
On the market side, it is seen that an increase in price is an incentive
for enhanced acreage as more farmers resort to use of fertilizers and adopt better
maintenance practices followed by more picking rounds. The previous year’s
price is the trigger for these added incentives. Farmers also keep pepper in stock,
hoping to sell for a better price in the following year if a price drop is encountered
in the current season. Alternatively, when prices escalate and farmers release
their stocks it will seem as though production has gone up, which is not the case.
Indian exports depend largely on the current season’s crop. The other factor is the
influence of price changes on stocks of traders. When there is a sudden increase in
price, the total amount of exports may even exceed production. The year 1985
was exceptional. Again, 1991 and 1992 were also exceptional on account of the
restructuring of the Indian economy when globalization and liberalization
processes set in.
2. Brazil
In Brazil there are no time series on area by cultivation system, although

there are different cultivation systems in the country. To be sure, it is still totally
unclear how the Brazilian Pepper Exporters Association gathers the data which
are presented at the IPC meetings. Apparently, data on area and production do
indicate that pepper yields have gone up over the years in Brazil. But this is not
the case. If production is divided by area it would appear that there is a
decreasing trend over time. In fact, between 1978 and 1979 yields plummeted
despite area expansion suggesting that in those years data on area did not
include area with immature vines. After 1979 the correlation between area and
production was strong, although 1984 was an extraordinary year which had a
much higher yield than could be explained. This must be due to positive price
correlation.
Pepper consumption in Brazil is negligible compared with production and
presumably kept out of production statistics and as there are some other small
pepper producing regions outside of Para state, it can be expected that almost
all pepper produced will be exported. However, there is a rather constant amount
of about 1.5 thousand tons that is apparently not exported. Price is the
most important variable in the current productive area under pepper.
3. Indonesia
Indonesian data on aggregate area under pepper is rather scanty. The official
records claim that the total area did not change from 1983 until 1987. However,
records for Lampung and Bangka show substantial changes over these years. The
time series on area of Lampung and Bangka are still too short. Further, there is
hardly any information on area in Kalimantan. For future modeling research,
regional disaggregation of supply of Indonesia is important, especially because of
the special status of Bangka, which produces only white pepper. Prices have
played only a minor role in area expansion. Clearing of new land for pepper
production will play a predominant role in pepper production especially in
342 K. P. P. NAIR
Kalimantan and Sulawesi. For Indonesia the influences of price on stocks is far
less important than on maintenance, a similar conclusion to that for India.
The year 1993 was a bad year because of a serious set back in production and thus
on exports reflecting a serious situation for white pepper in Bangka.
4. Thailand
Thailand is a relative newcomer to pepper production. Until 1986, area was

quite steady, thereafter it expanded, but once again dropped between 1991 and
1993 when low prices ruled the market. At this time the government encouraged
farmers to diversify from pepper. Changes in prices were influenced by prices
prevailing in the previous year, while area was held steady. Productivity per hectare
is very high and the production was influenced by the previous year’s prices. Over
recent years, production has been on the decline. The country has a high domestic
consumption, which is around 5000 tons per year. Exports take a fixed share of
production, which is influenced by the previous year’s shortage or surplus.
5. Malaysia
Malaysian data on area and production of pepper, of which nearly 98% refers
to that of Sarawak, are rather unreliable. Foot Rot strongly influences yield. Data
on production is based on exports and as pepper farmers from Sarawak speculate
on the pepper market, export is not a reliable index to show production trends as
these farmers who speculate in the pepper trade are rich. Hence, large differences
between production and exports are bound to arise. From 1989 onwards there is a
decline in area arising from shortage of labor and its high cost, which did not
allow for profitable production to the extent as was the case in the past. The price
of the previous 2 years greatly influenced production per hectare, though this
influence is through the planting of new vines.
6. Sri Lanka
A traditional pepper producer, there is only a slight increase in area over the
years. Productivity is low and despite an increase in area, production did not
correspondingly increase, which is rather surprising. The production pattern in
1993 is just the opposite to that of Indonesia, a year that showed an incredible
increase in production and exports. Export in 1990 was exceptionally high,
exceeding production. Export can be explained on the basis of production, the
change in production and domestic prices.
7. Countries Other Than Those Named Above
In addition to the countries named above, countries such as Madagascar and

those on the African continent, such as Benin, Cameroon, Congo, Gabon,
Ethiopia, Nigeria, Zambia, Zimbabwe, etc., also produce pepper, of which
Madagascar with a production of about 2500 tons leads the pack. As long time
series and information on area and production are needed, as well as information
on internal markets, no tangible conclusions can be drawn from the existing
information.
B. THE DEMAND SIDE OF THE PEPPER ECONOMY
To make a clear analysis of the demand side of the pepper economy, precise
data on end-use of pepper is needed. To date, precise data on differentiation of
pepper use in food industries, institutional catering and household consumption
is unavailable and, as such, only rough estimates can be made. Further, a
distinction between black and white pepper is needed. The following country-
wise groupings will provide an insight into the demand side of the pepper
economy.
1. European Union
As a shift in the food consumption pattern takes place, for example

consumption of more pre-fabricated food, fast food, etc., the demand scene
will fluctuate. Western Europe is price conscious as far as stock formation is
concerned. Also, the level of stocks at the end of the previous year has a
reasonably significant influence.
2. Rest of Western Europe
Starting from 1998, Switzerland has import data on pure pepper, whereas until
then these figures also included pimento and capsicum. Price changes influence
stock formation.
3. Eastern Europe and CIS
The economic problems that overtook Eastern Europe and the CIS have had a
very negative effect on import since 1990. Price changes influence stock
formation.
344 K. P. P. NAIR
4. North America (USA and Canada)
There is a negative relationship between price levels and starting stocks. When
prices are low, traders will keep more stock expecting prices to increase again
sometime in the near future and if stocks are high, traders tend to sell.
5. Japan
Demand for pepper in Japan is pegged to the Gross Domestic Product (GDP)
in more than one way. A shift towards outdoor, ready to eat food, in particular
Western food, goes along with a rising GDP. Changes in stocks paralleled trends
in North America.
6. Australia and New Zealand
Stock levels are related to consumption and no price influence is found on the
demand side.
7. Latin America
Per capita import of pepper did not show any increase. Between 1976 –80 and
1985 – 88 significantly lower imports took place and stock formation depended on
prices and lagged stocks.
8. Asia and Pacific Countries, Singapore, Australia and

New Zealand (Excluding China)
Stock formation depended on prices and lagged stocks.
9. China
China was a small importer of pepper and the import spurted in the latter part
of the 1970s. Of late, imports have dwindled.
10. Middle East and North Africa
These regions have a rather irregular import pattern. Stock formation was
influenced by price changes.
11. Rest of Africa
Per capita import of pepper did not show any increase as was the case of Latin
America. Since 1990 imports surged. Stock formation was influenced by price
changes.
C. PRICES AND WORLD MARKET
As far as the world market is concerned Singapore has a special place for
pepper import and its commercial consequences. Of the total pepper exported
from Malaysia, 69% moves via Singapore. The function of Singapore as an
entreport is accounted for by the variable total exports of producing countries less
estimated world consumption, which is merely an estimate of the change of
stocks outside producing countries. Singapore is expected to import part of these
stocks and keep the major part of these stocks as carryover stocks. Some part of
pepper imports is consumed, but no statistics are available on pepper con-
sumption in Singapore. It is estimated that consumption is approximately 1% of
imports. The price pattern indicates that pepper traders are presumably more
interested in trade if prices are high, which does not seem unrealistic as margins
will definitely be correlated with the price peaks.
D. THE PEPPER PRICE OUTLOOK BY 2020
Perhaps the most important aspect of the price structure is the impact of the
size of the Indian supply responses to price fluctuation. This aspect has not been
sufficiently captured. For all countries the investment side is not yet represented
adequately for long-term analyses. Stock formation at various levels needs
further work as well. Price cycles of around 7 years will continue in future
showing that the 10 year cycles in income have very little impact. There will not
be an increase in average real prices. Obviously, with inflation around 4% per
year, nominal prices do increase on average.
346 K. P. P. NAIR
E. THE PEPPER SUPPLY OUTLOOK BY 2020
Projections indicate that production and exports are expected to decline in

Brazil, Malaysia, and Thailand because of the increase in labor costs. Brazil
shows a modest decline on average, while India and Indonesia show strong
growth. The same conclusion as for Brazil can be drawn for Malaysia and
Thailand. These countries are becoming too rich to grow labor intensive
agricultural crops such as pepper. Sri Lanka is expected to level off on account of
limited area. Exports from Vietnam and China, especially the former, is expected
to grow rapidly reaching an aggregate of over 60,000 tons. A large quantity of
pepper is imported into India from Vietnam, often clandestinely, in the garb of
value addition. Since Indian pepper fetches a premium in the world market,
unscrupulous traders mix the low quality and cheap Vietnamese pepper with
Malabar pepper and re-export as Malabar pepper to the US and Europe. On
detection by the consumers, this adversely affects Indian pepper in the world
market, price-wise and also future imports.
The gap left behind the relatively rich countries such as Brazil, Malaysia and
Thailand in production and exports is filled up by Vietnam and China and not by
the traditional major producers such as India and Indonesia. Exports from
Madagascar and other African countries are expected to be steady. This
inevitably leads to the important conclusion that almost all of pepper export will
originate in Asia.
F. THE PEPPER DEMAND OUTLOOK BY 2020
On the consumption side, North America, Japan, and the European Union will
show a steady growth. Developments in Eastern Europe will impact the
consumption pattern, so will the increased consumption in the Pacific countries
and other up coming Asian regions, such as Philippines, Korea, Cambodia, etc.
Another major consuming area is the Middle East and North Africa.
G. COUNTRY-WISE ECONOMIC GROWTH IMPACTING

PRODUCTION AND CONSUMPTION
It is informative to analyze the GDP/growth rates that impact pepper

production and consumption. The details of the estimates are presented by Burger
and Smit (1997). Historical developments have impacted consumption. For
instance, the late eighties have been years of political turmoil for Eastern Europe
and the CIS and this has adversely affected growth rates impacting consumption
negatively. On the other hand, countries in eastern, far eastern and south eastern
Asia have enjoyed very buoyant growth rates which positively impacted
consumption. However, such very high growth rates cannot be maintained for
decades, as has been clearly shown by other countries with high growth rates in
the past.
In the long run, the USA will average a growth rate of just below 2% per
annum, while Japan, despite its recovery, may not reach the growth rate of the
USA, after the period of high growth up to 1990. In Western Europe, growth rate
will hover around 1– 3%. Eastern Europe, CIS and former Yugoslavia are
expected to recover, but, on average, at rather moderate levels of economic
growth. Growth rates in Latin America will be somewhat higher than in Europe.
In south Asia, India is maintaining a growth rate around 5%, though expectations
were higher. Pakistan has a somewhat higher rate. In southeast Asia, Philippines
will show a moderate growth, while in the other countries, though showing higher
rates, a declining trend would be observed. The same applies to China as well.
Vietnam will grow around 4%. With the exception of Nigeria, Africa on the
whole is showing only very moderate growth rates.
The above growth scenario indicates that potential for pepper export from
producing countries to the USA and Europe remain bright in the long run. This is
primarily because of the fact that the processed and fast food cultures will unfold
further in the years to come. Southeast Asian countries are also potential markets.
There is not much scope for export into Africa. Latin America, which will have a
higher growth rate, is a potential market, but countries such as Brazil are already
important pepper producing countries and, as such, export from other pepper
producing countries has only limited scope for export to Latin America.
XI. PEPPER ECONOMY IN INDIA
As India is the most important producer of black pepper in the world, a

discussion on pepper economy and marketing patterns in the country is relevant.
Currently, India accounts for 49.5% of the total acreage in the world, but in 1951
it was 70%. Though the country holds the largest pepper acreage in the world and
contributes up to 29.8% of world production, its productivity is the lowest in the
world, an abysmally low figure of 294 kg ha21. In fact, the total contribution of
India to world production dropped from 66% in 1951 to the current 29.8%. This
has been the case for exports as well. From 1951 to 1991, the share of India in the
global pepper market fell from 56 to 23%. It is interesting to note that while
India’s position of pre-eminence in pepper production dwindled, other countries’
performance was on the rise. The last decade of the past century witnessed a
global downtrend in pepper production.
In India, pepper is primarily cultivated for export. From around 16,000 tons in
1950– 51 export has scaled up to 32,000 tons, a 100-fold increase, by the turn of
348 K. P. P. NAIR
the last century. The nineties witnessed a fluctuating market because of price
fluctuations. The latter half of the nineties witnessed a sharp increase in unit price
peaking at about US $50 kg21 in 1997 and though export per se witnessed a drop,
it was more than compensated by the escalating unit price. In the latter half of the
nineties global pepper production was around 189,000 tons of which more than
118,000 tons were exported as per the data provided by the IPC headquartered in
Jakarta. In these estimates India tops the list of producing countries with 65,000
tons followed by Indonesia. India exports spices to more than 120 countries
around the world and of these countries, export to 23 countries account for about
82% of the hard currency earned (in US $) by the country. In the case of pepper,
India is the major supplier to 83 countries and among the importers USA tops the
list with an annual growth rate of 5% in quantity and 3% in value. Of the total
spices imported into the USA, pepper accounts for about 15% of which 50%
orginates in India. USA is also the major importer of spice oils, primarily
oleoresin, and India supplies close to 86%. However, the recent trend shows a
decline in import into the USA. In the beginning of the last decade pepper import
from India to the USA was 56%, which declined to 30% by the middle of the
decade. Though USA tops the list of pepper importers from India, it is interesting
to note that per unit value realization is lowest from Russia, Canada, Italy, and
Poland, with Russia topping the list. It is also interesting to note that there is a
visible shift in the food habits of Americans, where the preference for hot spices
like pepper in the last two decades accounts for nearly 75% consumption. The US
trade association projected spice consumption to peak to 1000 million pounds by
the turn of the century (Cheriankunju, 1996a). Maintenance of quality assumes
the greatest importance in this regard. The United States Food and Drug
Administration (USFDA) detained 1164 consignments of Indian spices due to
contamination with seeds of noxious weeds in 1995. In fact, there has been a
steep escalation of detection of contaminated consignments from 140 in 1991 to
757 by 1994 (Sivadasan, 1996), which, indeed, is a very poor reflection of Indian
trade. With the WTO tightening its noose on sub-standard quality of imported
food articles from the developing countries, India would do well to clean its own
stables lest a very adverse fallout on the pepper market follows.
It is, however, encouraging to note that the Spices Board of India, the apex
body attached to the Ministry of Commerce, Government of India, has initiated a
number of measures to ensure that quality of pepper is guaranteed both at the
producer as well as the exporter levels.
The former USSR was the largest pepper importer from India, with a total
tonnage of 19,400 in 1989 –90, which came down dramatically to 6060 tons
by 1994– 95 primarily due to political turmoil in the region (Cheriankunju,
1996a). In spite of the economic instability, Russia continues to be a major
consumer of pepper from India. Russia with an import of about 3200 tons in
1994 – 95 accounted for US $6.05 million followed by Italy at US $3.62
million and Canada at US $2.54 million. Traditional export of pepper from
India to such countries as Hungary, North Korea, Maldives, Trinidad, Yemen,

United Arab Emirates, and Yugoslavia has considerably decreased and in
their place new markets, such as Chile, Kazhakistan, Tajikistan, and
Venezuela have emerged. Of these, Kazhakistan imports the most. Apart
from these, Japan is fast emerging as a potential market for Indian pepper.
This is mainly due to the fast changing food habits of the country where
ready-to-eat “fast foods” are emerging along the lines of western food habits,
where pepper forms an important culinary component. In 1994– 95 Japan
imported 41,911 metric tons of spices of which 17.8% was pepper. Most of
the pepper goes into industrial uses followed by domestic consumption and
other services sector. The annual rate of increase in spice import by Japan is
estimated to be 5% (Cheriankunju, 1996b) and share of Japan’s import from
India is stagnant at 5.5% while that from Malaysia, the bulk exporter to
Japan, stays at 68%. Indonesia contributes 31%. Indian pepper has a premium
in the world market and if strategic advantages are exploited following the
establishment of WTO, India can scale up its export to Japan. Additionally,
Israel is a potential market for Indian pepper. In 1994– 95 Israel imported
6902 tons valued at US $7.42 million (Anon, 1996). Among the potential
markets for Indian pepper, USA, Canada, Russia, UK and Japan are on the
top rungs as these countries have a lower level of instability in pepper
imports (Jeromi and Ramanathan, 1983). According to the International Trade
Center, Geneva, world trade in spices during the last decade averaged
450,000 metric tons valued at US $1500 million of which pepper is the major
component, which accounts for as much as 30– 35% in quantity and 20 – 25%
in value (Peter, 1996b).
A. PEPPER PRODUCTION SCENARIO IN INDIA
In 1950 –51, nearly 80,000 ha of cultivated area in India was under pepper of
which 98% was in Kerala State. An analysis of the trends in growth with regard to
area, production, and productivity is given in Table XIII. It can be surmised from
the table that during the last five decades, overall acreage, production, and
productivity have declined all over India and, in particular, in Kerala state. The
steepest fall was in productivity. On an all India basis, while acreage slipped by
only 22% from the fifties to the nineties, production and productivity declined by
54 and 89%, respectively. For Kerala, the respective figures for decline in area
and productivity are 25 and 96% while production increased by 70%. In fact, both
on an all India basis and for Kerala state as well, productivity showed very
dramatic negative trends.
The reason for the steep fall in productivity, especially in Kerala, is due to the
prevalence of diseases, of which the Phytophthora Foot Rot is the most serious
and there is, as yet, no permanent remedy for this dreaded disease. As the
350 K. P. P. NAIR
Table XIII
Area, Production and Productivity Growth Rates for Black Pepper in India and
the State of Kerala
Details 1950s 1960s 1970s 1980s 1990s Overall
Area India 0.87 1.12 1.05 1.81 1.11 0.68

Kerala 0.94 0.19 2 0.19 1.85 1.00 0.70
Production India 1.36 2 0.50 0.34 2.43 2 0.32 0.63
Kerala 0.40 2 0.88 2 0.40 2.26 0.70 0.68
Productivity India 0.44 2 1.40 1.27 0.63 2 1.40 2 0.65
Kerala 0.46 2 1.07 2 0.22 0.41 2 0.01 2 0.02
Note: Date given as log Y ¼ a þ bt in percent per annum.

Source: Economic Preview, Government of the State of Kerala.
question of area expansion has only very limited possibilities, if there is to be a

breakthrough in pepper production in Kerala, it will have to be through breeding
superior disease resistant varieties and good agronomy to match. Soil manage-
ment is still based on bookish knowledge and there is a very clear dearth of fresh
ideas.
Pepper cultivation in Kerala is largely concentrated in certain districts.
Kannur, Kozhikode, Kollam, and Thiruvananthapuram districts, in sequence,
contribute more than 80% of the total pepper area in the state and 75% production
of the state. By the early nineties, the share of these four districts declined to 33
and 29%, respectively, in acreage and production and others took over. Currently,
the districts of Idukki, Wayanad, and Kannur have emerged as the major pepper
producing districts of the state contributing 60% of area and 67% of the
production. Among the three, Idukki district leads with 21% area and 27%
production. It is followed by the Wayanad district, which during the past decade
has contributed 25% of pepper produced in the state. A very interesting
development of late is the emergence of the Idukki district as the front-runner in
the production of “organic pepper”, much valued in the USA and Europe. Based
on the cool climate of the region and relatively lower impact by the chemical
agriculture, the district holds out much promise to emerge as a global leader in the
production of organic pepper.
Pepper as a mono crop is confined to the district of Idukki and Wayanad which
constitute only 3% of the total area (George et al., 1989). There is ample scope to
cultivate pepper as an inter crop in coffee, tea and cardamom plantations. The
high range districts of Kerala, such as Idukki and Wayanad, where currently most
of these plantations are located, are potential areas where pepper can be grown as
an inter crop. Silver oak, used as a shade tree in these plantations can be used as
standards (support for the pepper vine to trail). There is a strong case to initiate
systematic field trials to develop a clear-cut package of practices in growing
pepper as an inter crop in these plantations. Isolated studies have been carried out.
However, no efforts on a large scale have been initiated. This is a very promising
area for enhancing pepper production in India, until now not tackled in a very
systematic and scientific manner.
Pepper is an important component of “homestead farming” in Kerala and
the vines are trailed on coconut trees, arecanut trees, mango trees, jackfruit
trees, and tamarind trees, which are generally grown in small farms owned
by small scale and marginal farmers. The level of inputs, namely fertilizers,
pesticides, etc., in these situations is low. Transformation of these homestead
gardens due to various socio-economic reasons largely reduced this practice.
Coupled with this, the incidence of the dreaded disease Foot Rot swept away
a substantial number of pepper plants in homestead farms. There is a strong
case for developing a package of practices at low input levels in these
homestead farms. There is a need to re-introduce pepper in homestead
farming, both in urban and rural areas. For this purpose, bush pepper offers
much promise.
B. BUSH PEPPER
Bush pepper holds out much promise in homestead farming. The

plagiotropic lateral fruiting branches of pepper exhibit sympodial growth
and these branches when rooted and planted grow into a bush. More fruiting
branches are produced by these bushes and flowering commences in the first
year of planting itself. A common method developed at the Indian Institute of
Spices Research (IISR) involves the following steps: (1) selecting healthy
lateral shoots of the previous year’s growth with two or three leaves,
(2) dipping the cut end in a commercially available rooting hormone, and
(3) planting them in moist rooting medium, which consists of weathered coir
dust, in a 200 gauge PVC bag of about 40 cm £ 25 cm size, and later the
bags are kept under shade. Prior to closing the bag, air is blown into it. In
about a month, the cuttings start producing roots. After about two months the
bags are opened and allowed to remain in that state for a few days before
transplanting into clay pots of convenient size. Almost 80% success can be
ensured provided the steps described above are taken meticulously. Bush
pepper can be planted at a minimum density of 2500 plants ha21 spaced
2 m £ 2 m and the population can be doubled if spacing is reduced. Record
yields of 1960 kg ha21 have been obtained using Panniyur-1 variety (Annual
Report of IISR, 1997) and yields in the range 1600 –1700 kg ha21 are quite
common. Bush pepper is commonly grown in Indonesia, Malaysia and the
Philippines. It is still very far from catching on in India, and the home state
of pepper, Kerala. Figure 8 shows Bush Pepper in India.
352 K. P. P. NAIR
Figure 8 Bush Pepper in India.
C. ECONOMICS OF PEPPER PRODUCTION IN THE

STATE OF KERALA
In India, Kerala is the most important pepper producing state and the
following discussion pertains to the state. Vinod (1984) and Santhosh (1985)
have attempted to make a dependable estimation of the cost structure of pepper
production in the Idukki and Kannur districts, respectively, which are at the
forefront of pepper production in the state of Kerala. Pepper production in the
state of Kerala is highly labor intensive and labor costs account for more than
50% of the total cost of production. An important reason for the decline in
pepper production in the state could be due to the non-availability of timely labor
and its very high wage structure on an all India basis. Among all the states of
India, Kerala’s labor wages are the highest. Though Idukki district leads in
cultivation, profitability is higher in Kannur district. The cost-benefit ratio
was 1.09 for Idukki district and 1.16 for Kannur district as per the findings of the
above mentioned authors. The pay back period of pepper cultivation was
estimated as 10 years in Idukki district and 11 years in Kannur district. The
corresponding net present worth is approximately US $93 for Idukki district and
US $148 for Kannur district and the internal rate of return 13.48 and 17.22%,
respectively.
D. MARKETING
Village traders dominate the local market inasmuch as small pepper

farmers are concerned in Kerala. The trade is based, in many ways, on faith
and long-term personal bonds, which is absent in the case of whole-sale

businessmen who deal in large quantities. Additionally, transportation to the
city or town market is quite often out of the reach of small farmers and this
also leads to the small farmer going to the village trader. By tradition, village
traders collect the product directly from the farmers’ yard or if not possible
from the neighboring village markets. Wholesale businessmen, though, offer
higher prices compared to the village trader. Village and internal wholesale
merchants are important market intermediaries (Vinod, 1984). Small and
marginal farmers generally affect village sales in the nature of a pre-harvest
contract whereby the farmer is obliged to sell his product at a pre-determined
price prior to harvest and if there arises any fluctuation in price subsequently,
both parties are bound to go by the earlier contract of sale agreed upon. The
up country wholesale merchants transport the produce to Kochi, which is the
port city of Kerala and the prime business center of the state, in the central
part of the state. Once the produce reaches Kochi, the commission agents
negotiate with the exporters or wholesale merchants or brokers. The brokers
act on behalf of the internal wholesale merchants or exporters and negotiate
with the commission agents. At the exporters’ premises the produce is
processed and graded. Prior to the official clearance of the produce for export,
it is subjected to check sampling and all the tests for grading. The
consignment thus cleared for export is then handed over to the clearing or
forwarding agency after completion of the paper work. Subsequent work is
carried out by the forwarding agency. Four major marketing channels were
identified in Idukki District, while there were five in Kannur district. The share
of the producer on the free on board price (FOB price) of pepper was
estimated to be 86.06%. Compared to this, the producer is the beneficiary of a
higher share, which is 88.80%, when the produce is internally consumed
moving through the domestic consumption channel. In either case, it is clear
that middlemen corner a good percentage of the value chain.
E. PEPPER FUTURES MARKET
The origin of futures trade is almost a century old in India. Futures trade is an
insurance against price fluctuations of commodities handled by traders. It was the
Indian Pepper and Spice Trade Association (IPSTA) who heralded the futures
trade in pepper in India in Kochi (Sethuram, 1995). The move had a very positive
effect on the transfer of risk and recovery of price. From hedgers the risk is
transferred to the speculators, who in turn provide the liquidity to the market.
Future markets are standardized with regard to trading specifications and the
terms of delivery. The international pepper price shows a trend which is
unfavorable to most of the pepper producing countries. This is largely attributable
to supply situations rather than production scenarios. The IPC in an effort to
354 K. P. P. NAIR
ensure an efficient price management system conducted a study, sponsored by the

United Nations Commission for Trade and Development (UNCTAD), on the
viability of pepper futures contract. Based on this a working group of
representatives from the IPC member countries was set up to examine the
requirements of a global pepper futures contract. Consequently, the IPSTA at
Kochi initiated action for the establishment of an International Pepper Exchange
(IPE) at Kochi. The expectations are that it will provide an incentive in pepper
trading, both in the domestic and international markets, ensure better prices to
producers and also extend a hedging facility to exporters of all the member
countries. Further, it is also proposed to link the Kochi pepper exchange with the
Kuala Lumpur Commodity Exchange at a future time, when a second trading
floor would be started. The globalization process, in general, and the increasing
volatility of pepper prices necessitated the establishment of a global pepper
market to trade pepper futures contract. This also facilitates a free trade regime in
pepper, as the product can be freely moved among the countries without tariff
barriers and the restrictions of quotas.
On account of the additional premium for Indian pepper, its prices in the
international market are relatively higher than that of pepper produced by other
countries. Low yield combined with high unit cost of production combined with
better prices in the domestic market leads to a higher international price. Also, the
improved stock holding capacity of the farmers influence price, because, in times
of price decline, farmers can hold on to their produce without unloading it to the
market. Both corporate procurement and speculation play a significant role in
pushing up the price. Futures trade ensures a constant supply and steady price for
the produce. One of the main criticisms against the exchange is how it can best
translate the benefits to the actual producers. Pepper production in Kerala is
mainly scattered among small farms and the market awareness of these small and
marginal farmers is indeed scanty. As much as 72% of the farmers belong to this
ignorant category (George et al., 1989). These small and individual farmers
neither have sufficient quantity to trade nor the technical expertise to benefit from
the pepper exchange. This has resulted in the futures trading in Kochi being
handled mainly by representatives of exporters. The history of futures trading in
pepper has clearly shown that producers have almost no say in its operations,
including the large ones. The Primary Agricultural Credit Societies (PACS) have
also not attempted to enter the futures market. The State Marketing Agencies,
such as the State Trading Corporation and the State Marketing Federation, both
governmental agencies, are wary of their association with the exchange set up as
they have met with large financial losses. This is only natural, because
governmental agencies in Kerala rarely have true accountability, which is a
legacy of the earlier socialistic pattern of governance in the state. If fair trading
has to be ensured, the possibilities of manipulative tactics, which create an
artificial market to the detriment of the interests of the producers, by the powerful
members of the exchange, have to be fully scotched. But this invariably does not
happen and the speculator calls the shots rather than the toiling producer,
however big or small he may be. Despite these shortcomings, the IPSTA is
hopeful of bridging the disparity, often quite wide in certain instances, between
farm gate price and market price, by eliminating needless middlemen involved in
transactions, which would ensure better prices for the producers. Liquidity of
traders is expected to increase as banks are now willing to advance as much as
80– 90% of the value of the produce when it is hedged against adverse price
fluctuations through the medium of futures contract. Reckless speculations will
be prevented by the imposition of limits on open position, daily price band
fluctuations bound both up and down and collection of special margin deposits to
curb speculative activity through financial restraints.
An important component of the economics of pepper growing in Kerala is the
availability of reliable market information to the farmers. As of now, most of the
pepper farmers have either no dependable access to reliable market information
or are illiterate so cannot make use of it. Hence, the need to educate farmers on
the importance of acquiring reliable market information and marketing processes
in pepper is of paramount importance. One of the prime reasons for pepper
farmers’ inability to realize a good price, especially in the case of small and
marginal farmers, is the lack of adequate storage facilities. Holding capacity of
the farmers can be enhanced through the provision of credit support. There is
good scope to attempt a co-operative marketing system as in the case of other
plantation crops in the states, such as rubber, coconut and arecanut. In the case of
these crops, there are Co-operative Marketing Societies spread across the state
and a similar attempt for pepper could be initiated also.
XII. PEPPER PHARMACOPOEIA
The use of black pepper as a drug, apart from its wide spectrum use as a food
additive, both in the Indian and Chinese systems of medicine is well documented
(Atal et al., 1975; Nadkarni, 1976; Kurup et al., 1979). Pepper is elaborately
described for its medicinal values in the ancient Indian system of medicine
Ayurveda. It is described as katu (pungent), tikta (bitter), and ushnaveerya
(potency). Further mention of pepper is also made in the control of the
principal sources of diseases, as described in Ayurveda, namely, vata and kapha,
in the ancient Indian language Sanskrit, the biological processes in the human
system controlled by the CNS and the later formation and regulation of all fluids
that have a preservative quality, such as mucus, synovia, etc. Pepper is described
as a drug which enhances digestive power in the body, improves appetite, cures
colds, coughs, dyspnoea, diseases of the throat, intermittent fever, colic,
dysentery, worm infestation, and also hemorrhoids. Pepper is also prescribed
to relieve toothache, muscular pain, inflammation, leucoderma, and even
356 K. P. P. NAIR
epileptic fits (Ayier and Kolammal, 1966; Kirtikar and Basu, 1975). Black pepper
is called maricha or marica in Sanskrit, which implies its ability to dispel poison.
In the Chinese system of medicine, pepper is used as an antidote against snake
and scorpion bites. The above descriptions explain the diverse actions of pepper
being used in the Indian system of medicine, either as such, or through its active
ingredients in many formulations. In sharp contrast to the use of pepper in the
Indian system of medicine, pepper finds no use in allopathy or homeopathy. The
beneficial effect of pepper in medical formulations can be attributed to piperine
and other phenolic amides and essential oils. But in Ayurveda, the active
ingredient based specific activities are not taken into consideration. Hence, the
pharmacological and toxicological aspects of pepper and its constituent
secondary metabolites were not studied. It must be, without doubt, mentioned
that many of the clinical uses of pepper are time-tested, over several generations,
and there is implicit faith among the Indian masses on the unquestionable
medicinal value of pepper. However, to scientifically establish the value of
pepper as a unique medicinal product, it is crucial to investigate, based on well
defined experimental protocols currently used in modern drug research and
development, the pharmacological aspects of pepper and its active ingredients.
Some of the preliminary investigations carried out over the last two decades in
India on the above lines provide valuable information. Antipyretic, analgesic,
antiinflammatory, antimicrobial and antineoplastic activities were reported after
in vitro and in vivo investigations. Of late, there is a surge of interest in
environmental friendly use of products to control crop pests and pepper has been
found to possess insecticidal as well as insect repellent properties. Developments
of new pepper based products as insecticides could be a boon to human use as
they would be free from the toxic nature of most of the insecticides currently in
use. Piperine is the major alkaloidal constituent of pepper. Systematic
pharmacological studies on piperine have revealed its analgesic (pain relieving),
antipyretic (fever relieving) and antiinflammatory properties, in addition to
rejuvenation of the CNS. The antimicrobial activities are provided mainly by the
essential oils.
In addition to the ancient Indian system of medicine, Ayurveda, there are other
native systems of medicines in India, such as Siddha and Yunani. Pepper is used
in these as well. In the Chinese system of medicine, pepper and chenghan (Dichora
febrifuga L.) are used in the treatment of malaria (Das et al., 1992). The active
ingredients of chenghan are febrifugin and iso-febrifugin, both of which show a
50% higher antimalarial property than the conventional quinine. As an antipyretic,
pepper is also used in the treatment of malaria (Nadkarni, 1976). The curative
effects claimed in the cases mentioned above in the case of pepper can be attributed
to the antipyretic and analgesic actions of the active ingredients. A strong
antipyretic effect on typhoid vaccinated rabbits at a dose of 30 mg kg21 body
weight after oral administration has been reported by Lee et al. (1984). In these
studies, acetaminophen was used as the reference drug. The antipyretic action of
piperine was found to be stronger than that of acetaminophen. To determine

the analgesic action, acetaminophen and aminopyrine were used as reference
compounds. Piperine showed a strong activity with an LD50 value of 3.7 mg kg21
dose using the writhing method and 104.7 mg kg21 using the tail-clip method.
A. ANTIINFLAMMATORY AND CENTRAL NERVOUS SYSTEM

(CNS) DEPRESSANT ACTIVITY OF PEPPER
Several researchers have reported the antiinflammatory property of pepper.

Piperene showed a significant inhibition of the increase in edema volule in a
carragenin induced test with an oral intake of 50 mg kg21 body weight (Lee et al.,
1984). The effect of different acute and chronic experimental models were
studied by Mujumdar et al. (1990a). These researchers evaluated the mechanism
of antiinflammatory activity by biochemical processes and concluded that
piperine acted positively in the case of early acute changes in inflammatory
processes. These studies have been further corroborated by those of Kapoor et al.
(1993).
In the case of epileptic patients, pepper is used in Ayurveda to induce sleep
(Kirtikar and Basu, 1975). In the pharmacological investigations pepper and
piperine were found to have a CNS depressant effect. Both pentetrazole seizure
and maximal electroshock seizure were inhibited by pepper and piperine (Won
et al., 1979). The LD50 value of 287.1 mg kg21 i.p. was found for piperine and a
corresponding oral value was 1638.8 mg kg21. In another investigation Shin et al.
(1980) showed that piperine at one-tenth its LD50 value had strong potentiating
effect on hexobarbitol induced hypnosis in mice. Decreased passivity, ptotic
symptoms and decrease in body temperature were also observed. Protection
against electroshock seizure and a muscle relaxant effect were observed at a
relatively low intraperitonial dosage range of an LD50 of 15.1 mg kg21 (Lee et al.,
1984). This dosage appeared almost equipotent to the reference compound
phenytoin. Petrol extract of pepper leaves was found to potentiate pheno
barbitone induced hypnosis in mice (Sridharan et al., 1978). Majumdar et al.
(1990b) have also reported that a high dosage of piperine potentiates the
phenobarbitone sleeping time by inhibiting the liver microsomal enzymes. It also
acted partially through stimulation of the pituitary adrenal axis. The above
investigations clearly show that piperine has the capacity to be an effective
ingredient of pepper in the treatment of “petit mal” (minor ailments).
Antiepilepsirine (AE) isolated from white pepper is a compound of very great
pharmacological interest. AE is 1-(3-benzodioxol-5yl)-1 (oxo-2-propenyl)-
piperidide. Perhaps this is the only compound derived from pepper which is
clinically used in the treatment of human epilepsy. This is an alternative to
dilantin therapy which is used in Chinese hospitals for the treatment of epilepsy
(Ebenhoech and Spadaro, 1992). It is probable that AE may release 5-hydroxy
358 K. P. P. NAIR
tryptamine (5-HT) from nerve endings which can intensify the anticonvulsive
state (Liu et al., 1984). Studies using rats as test animals show that AE increases
5-HT concentration in their brains which intensify the anticonvulsive state. AE
also raises the tryptophan level in the brain which causes elevation of serontin
and monoamine levels which leads to control of seizures.
B. EFFECT ON HEPATIC ENZYMES
While pepper or piperine positively affects liver function, as evidenced by the

research of many workers, they do in no way cause hepatic toxicity. Piperine
functions mainly as a chemopreventive substance through modulating enzyme
function. Hepatic toxicity of piperine in rats was studied by estimating the hepatic
mixed function oxidases and serum enzymes as specific markers of hepatotoxi-
city (Dalvi and Dalvi, 1991). An intragastric dose of 100 mg kg21 body weight
resulted in an increase in hepatic microsomonal enzymes 24 h after treatment. In
the experiment, cytochrome p-450, cytochrome b5, NADPH-cytochrome C
reductase, benzphetamine N-demethylase, aminopyrine N-demethylase and
aniline hydroxylase were estimated. An intraperitonial dose of 100 mg kg21
body weight did not produce any effect on the activity of the drug metabolizing
enzymes. However, when the dose was enhanced an eightfold to 800 mg kg21
body weight as intragastric and 100 mg kg21 body weight as intra peritonial, a
significant decrease in the levels of the enzymes was noted. But these treatments
did not affect those serum enzymes which are specific markers of toxic liver
conditions. Piperine has been found to provide significant protection against
chemically induced hepatotoxicity. Both in vitro and in vivo lipid peroxidation
and prevention of depletion of GSH and total thiols were observed (Kaul and
Kapil, 1993). Liquid peroxidation causes the production of free radicals which, in
turn, causes tissue damage. GSH conjugates xenobiotics which are excreted by
subsequent glucuronidation. In the above cited study the hepato protective action
of piperine was compared with a reference compound, silymarin, a known hepato
protective drug, and it was found that piperine has slightly lower activity. In a
feeding experiment using Swiss albino mice fed with a diet containing 1, 2, and
3% black pepper, w/w, for 10 and 20 days, a dose dependent increase in the
levels of the hepatic biotransformation enzymes, namely, glutathione-s-
transferase, cytochrome p-450, cytochrome b-5 and acid soluble sulfhydril-SH
were observed (Singh and Rao, 1993). Reen et al. (1993) in an in vitro study
found lowered levels of glucuronidation due to the inhibition of the enzyme
UDP-glucose dehydrogenase. Tripathi et al. (1979), while investigating the
hypoglycaemic action of several plants, found that pepper fruits are devoid of any
significant hypoglycaemic action in rabbits. The aqueous extract of pepper leaves
at a dose of 10 – 20 mg kg21 body weight led to a moderate increase in the blood
pressure of dogs (Sridharan et al., 1978). Both piperine and AE have detoxifying
qualities which may lead to an increase in the bioavailability of other drugs, as a

result of which they alter the pharmokinetic parameter of the epileptic (Bano
et al., 1969).
C. CARCINOGENIC AND MUTAGENIC EFFECTS OF

BLACK PEPPER
Positive effects of pepper as an anticarcinogenic and non-mutagenic source

have been found by investigators. The Ames test shows its non-mutagenic
quality. Chemical carcinogenesis has been found to be prevented by pepper
because of its stimulatory effect on xenobiotic biotransformation enzymes. The
antioxidant properties of piperine and associated unsaturated amides play a
preventive role in carcinogenesis. Dietary intake of natural antioxidants can be a
crucial aspect of building up the human body’s natural defense mechanisms
against the onslaught of diseases, which can be caused by degradative changes
brought about by mutagens. Additionally, the essential oil constituents in pepper
inhibit DNA adduct formation by xenobiotics. This observation shows that
pepper has anticarcinogenic potential. However, pepper extracts showed
enhanced incidence of tumor formation in mice and an elevated level of DNA
damage caused by piperine in cell culture investigations. Hexane, water and
alcohol extracts of pepper were tested for mutagenecity on Salmonella
typhimurium strains TA 98 and TA 100 by Ames assay and the results indicated
that the extracts were non-mutagenic. This investigation provides evidence that
water extract exerts an antimutagenic action on carcinogen induced mutagenesis
(Higashimoto et al., 1993). The chemoprotective role of pepper as well as its
constituents has been shown because of their positive effect on the activity of
biotransformation enzymes in the liver in a dose-dependent manner (Singh and
Rao, 1993). Nakatani et al. (1986) report a very significant antioxidant activity
due to the phenolic amides present in pepper. It is clearly established that
antioxidants exert a preventive role in carcinogenesis. The modulating effects of
the essential oil constituents of pepper were investigated by Hashim et al. (1994)
who showed that they have an inhibitory effect on carcinogenesis. The volatile oil
and its constituents suppress the formation of DNA adducts with aflatoxin B1.
The microsomal enzymes modulate this action. Feeding powdered pepper to
mice at 1.66 w/w dose showed no positive impact on carcinogenesis, while
feeding and painting of the mice body with solvent extract of pepper, 2 mg for 3
days per week for three months, showed an increased incidence of tumor
formation (Shwaireb et al., 1990). The activity of methylcholanthrene, a potent
carcinogenic compound, was found to be reduced by a pepper terpenoid
d-limonene (Wuba et al., 1992). Two minor constituents of pepper, safrole and
tannic acid, have minor carcinogenic activity. In a tissue culture study using V-79
lung fibroblast cell lines it was found that piperine treated cell lines showed
360 K. P. P. NAIR
increased DNA damage compared with untreated ones (Chu et al., 1994).
Piperine treatment lowered the activity of the enzymes glutathione-s-transferase
and uridine diphosphate glucuronyl transferase indicating cytotoxic potential.
The in vivo formation of n-nitroso compounds from naturally occurring amines
and amides contribute to the carcinogenic potential of certain foods and food
additives. Piperine and other phenolic amides present in pepper are also known
for their conversion to n-nitroso compounds in acidic conditions and, hence, are
treated as carcinogenic (Lin, 1986). However, it can be inferred that the presence
of a conjugated unsaturated system in the phenolic amide prevents the oxidation
of the amide nitrogen to n-nitroso compounds to a great extent. Additionally, the
essential oil constituents in pepper would also ensure DNA stability because of
their anticarcinogenic potential. All of the above cited research results point to an
ambivalent nature of pepper, both as an anticarcinogenic agent, and as a pro-
carcinogenic agent. But, the overwhelming evidence indicates that it is more of
the former than the latter. It is because of this belief that pepper has been a crucial
ingredient of the Indian systems of medicine, primarily Ayurveda and to a lesser
extent Siddha and Yunani, for centuries, and continues to be so even in the
modern day.
D. PEPPER AS AN ANTIOXIDANT
Antioxidants are one of the most crucial biochemical compounds in the human
system ensuring good health. They scavenge free radicals, which trigger many
untoward biochemical reactions in the human system, and control lipid
peroxidation in mammals. Of late, there is considerable interest in antioxidants.
Lipid peroxidation is a chain reaction that is a constant source of free radicals,
which initiate further peroxidation, which trigger deterioration of food, but also
damage tissue, both of animal and plant origin, which in human beings cause
many inflammatory diseases, ageing, atherosclerosis and also cancer. Many
investigations reveal that pepper and its phenolic constituents, such as amides,
possess good antioxidant properties. Tocopherol and vitamin-C are two important
natural antioxidants. Chiapault et al. (1955) have investigated the antioxidant
property of spices in a two-phase aqueous fat system. The investigations indicated
that pepper has an antioxidant index of 6.1 while turmeric and clove show values
of 29.6 and 103.0, respectively. The antioxidant property of pepper has been
attributed to its tocopherol content (Saito and Asari, 1976). Revankar and Sen
(1974), however, attribute the antioxidant property of pepper to its polyphenolic
content. These authors investigated the effect of pepper oleoresin on fish oil and
arrived at this conclusion. Abdel-Fattah and E1-Zeany (1979) observed that the
autoxidation of unsaturated fatty acids and proteins was delayed by the addition
of pepper and substantial protection against oxidative degradation was obtained.
Nakatani et al. (1986), while investigating the family Piperaceae, established the
fact that all the five phenolic amides present in P. nigrum possess very good
antioxidant properties, which were found to be superior to the synthetic
antioxidants such as butylated hydroxy toluene and butylated hydroxy anisole.
E. PEPPER AS AN ANTIMICROBIAL AGENT
Volatile oils, which are an active constituent of most spices, of which pepper is
the most important, impart antiseptic, antibacterial and antifungal properties. The
positive effect of volatile oils can be attributed to their terpenoid constituents.
Pepper possesses both bactericidal and bacteriostatic properties. These properties
aid in enhancing the shelf life of foods to which pepper has been added. Even the
leaf extract of pepper possesses the antibacterial activity (Subramanyam et al.,
1957). However, the extract was not inhibitory to the growth of E. coli,
Aerobacter aerugenosa, Lactobacillus casei, Staphylococcus faecalis, S. aureus,
and S. sonnei (Subramanyam et al.,1957). The essential oil obtained from pepper
was found to be inhibitory to the growth of a penicillin-C resistant strain of
S. aureus. Jain and Kar (1971) documented the inhibitory action of pepper oil
on Vibrio cholerae, S. albus, Clostridium dipthereae, Shigella dysenteriae,
Streptomyces faecalis, S. pyogenes, B. pumilis, B. subtlis, Micrococcus sp.,
Pseudomonas pyogenes, P. solanacearum and Salmonella typhimurium. The
antibacterial action was determined by the agar well diffusion technique using
cephazolin as standard (Pever and Anesini, 1994). The mycelial growth and
aflatoxin synthesis by Aspergillus parasiticus were inhibited by pepper oil at a
concentration of 0.2 –1% (Tantaoui and Beraoud, 1994). Antifungal activity
against Candida albicans (Jain and Jain, 1972) was exhibited by pepper leaf oil.
Rao and Nigam (1976) reported a similar effect of pepper leaf oil on A. flavus.
The antibacterial effects of pepper extract, essential oil and isolated piperine in
vitro against sausage micro flora, L. plantrum, Micrococcus specialis and
Streptococcus faecalis have been reported by Salzer et al. (1977) in which the
authors noted that only isolated piperine displayed microbial growth inhibiting
effects at a normal dose. Pepper powder and extract were active only at high
concentrations. An alcoholic extract of pepper was found active against the
deadly food borne bacterium C. botulinum (Huhtanen, 1988). Another potential
activity was revealed by a study carried out using tincture of pepper by Houghton
et al. (1994). These authors investigated the antibacterial activity against nine
strains of Mycobacterium tuberculosis and found that growth of all the nine
strains was inhibited. The mode in which pepper acts positively is through its
effect on enhancing bioavailability of administered medicaments, when they
contain pepper or its active ingredients and uptake of proteins and amino acids
from ingested food. In an investigation using Trikatu, which is an Ayurvedic
preparation containing P. nigrum, P. longum and turmeric (Zinjiberus officinale,
another spice with great therapeutic value), it was observed that the pepper
362 K. P. P. NAIR
containing preparation enhances the bioavailability of other medicaments (Johri

and Zutshi, 1992). At a dose of 25– 250 m mol21, piperine enhanced the uptake of
1-leucine, 1-iso leucine and 1-valine and increased lipid peroxidation in an
in vitro study using intestinal epithelial cells of rats (Johri et al., 1992). Probably,
piperine interacts internally to enhance the permeability of intestinal cells.
Protein uptake from pulses was enhanced by up to 1.5% with the addition of
pepper (Pradeep and Geervani, 1994). The study involved mainly two pulses in
the Indian vegetarian diet, namely, winged bean and horsegram. The results
clearly suggest that the Indian vegetarian diet, where spice is a common
ingredient, has a decided advantage in human health.
F. THE PHARMACOLOGICAL EFFECT OF PEPPER ON

HUMAN HEALTH
The precise role of pepper on human health can only be understood through
well-structured studies, which are only very limited in number to date. However,
from the above it can be concluded that where extrapolations and comparisons
can be made, pepper, as a whole product, or its active ingredients have been
found to have a very positive role on human health. A brief description of the
limited number of studies is given here.
When used in high doses, gastric mucosal injury caused by pepper is
comparable with that of aspirin. This observation was made in a double blind
study of intragastric administration of pepper to human volunteers (Mayore et al.,
1987). In this investigation, healthy human volunteers were given meals
containing 1.5 g of pepper; 655 g of aspirin and distilled water were used as
positive and negative controls. This is a short term study and it must be pointed
out that long term effects of daily pepper intake, through food or medicaments,
are unknown. Vezyuez et al. (1992) investigated the effect of intestinal peristalsis
by measuring the orocaecal transit time (OCTT) utilizing a lactulose hydrogen
breath test on healthy human subjects. They were given 1.5 g of pepper in
gelatine capsules and the OCTT was measured on several days and it was found
that OCTT increased significantly after administration of pepper. This finding has
great clinical importance in the management of various gastrointestinal tract
disorders. An equally important investigation reveals the effectiveness of pepper
extract volatiles in the treatment of cessation of smoking. Results of the
investigation on human subjects reveal that cigarette substitutes, which deliver
pepper volatile compounds, alleviated smoking withdrawal symptoms. The
results of both the investigations detailed above have very great potential for
further experimentation and, perhaps, pharmacological exploitation for further
therapeutic benefits for human health.
G. CLINICAL APPLICATIONS OF PEPPER
It is only in the case of the Indian system of medicine, primarily Ayurveda, and
to a lesser extent Siddha and Yunani, that pepper has been used in many clinical
applications. No such use is made of pepper in allopathy. When clinically used,
none of the active pepper ingredients are used as such. Dried black pepper with
other medicinal plants, also in the dried form, are used in the preparation of
specific formulations. The most widely used formulation in Ayurveda is
Dasamulakatutrayadi Kashayam, a formulation that is used in the treatment of
asthma. Then comes Ashtacurnam, which is used to get over dyspepsia,
flatulence, etc. The formulation also has stomachic and carminative effects.
Amritharishtam is used as an antipyretic and also in the case of women with
excessive menstrual bleeding. Muricadi Thailam is used as an antiinflammatory
agent and also to relieve rheumatic pain. To cure coughs and bronchitis,
Dasamularasayanam is used and Gulgulutiktaka-Ghrtam has both analgesic and
antiinflammatory properties. Though pepper as such is widely used in the
preparation of Ayurvedic medicines, its antipyretic, analgesic and antiinflamma-
tory properties merit further research for use in allopathy, since, to date, pepper is
not used in the preparation of any of the allopathic medicines. Its antioxidant and
antimicrobial activities are also worth investigating further for possible use in the
manufacture of allopathic medicines. Additionally, pepper has a good dietary
value. It has a high fiber content (15 – 33%), iron (54 –62 mg g21), calcium
(1 – 1.5%) and also appreciable amounts of essential amino acids (Uma Pradeep
et al., 1993). As a good digestive, pepper enhances the secretion and flow of
salivary enzyme amylase.
H. TOXICOLOGICAL EFFECTS
Published reports do not detail any toxic effects of pepper. This may be due to
the relatively small amounts used in many medicinal formulations. Since, in most
cases, the total contents of piperine and associated phenolic amides used would
add up to just about 7 – 9%, w/w, and that of the volatile oils 2 –4%, which are
negligible amounts, no untoward health hazard has been encountered so far. At
this level, the actual doses of the different constituents available from the quantity
of pepper powder, oleoresin or extractive used, will be very unlikely to trigger
any toxic side effects in human body. In fact, the very pungency of piperine and
strong flavor of the volatile oils, act as deterrents against excessive human
consumption. The FAO and WHO Experts Committee on Food Additives do not
prescribe any limit to acceptable intake of piperine and the volatile oils. The
major untoward consequence of pepper use is the gastric mucosal injury at a dose
of 1.5 g kg21 of food. It enhances the DNA adduct formation. The extract of
pepper produces enhanced cancer incidence in mice. When mice were fed with
364 K. P. P. NAIR
extract of black pepper it resulted in the formation of heptocellular carcinoma,

lymphosarcoma and fibrosarcoma (El Mofty et al., 1991). However, a large
number of investigations show that pepper has anticarcinogenic attributes. Before
conclusive evidence can be arrived at, whether or not pepper and its active
ingredients trigger anticarcinogenic or procarcinogenic reactions in human
beings, much more detailed scientific scrutiny needs to be carried out. As of now,
such studies are conspicuously lacking. The observation that pepper enhances the
liver microsomal enzyme activity must also be investigated further. This will
help decide whether or not constant use of pepper in the human diet leads to
production of non-specific enzyme induction. Pepper is a universal dietary
component on the Indian subcontinent and the quantity used is much more than
used anywhere else in the world. Such information has very vital implications of
health for a population that currently stands at more than a billion.
I. THE INSECTICIDAL ACTIVITY OF PEPPER
The primary advantage of using plant derived insecticides in agriculture is that

while the product is toxic to the pest in question, it is non-toxic to human beings,
including those who handle it in the field, unlike most other chemically produced
commercial insecticides, which have come under attack from the environmen-
talists. Plant derived products have been used since time immemorial both in
agriculture and for domestic purposes. The extract of pepper, volatile oil
components, and the different phenolic amides present in pepper have shown
insecticidal, insect repelling, and ovicidal activities to various plant insects and
pests harmful to human beings. The major alkaloid of pepper, namely piperine, is
more toxic than pyrethrine, a standard insecticide for house flies (Harvill et al.,
1943). In a number of instances the insecticidal or insect repellent activity was
obtained at low concentrations. Hence, this aspect needs to be further probed for
positive field application. The volatile oil of pepper was found to cause mortality
of the cigarette beetle (Lasioderma serricorne) (Samuel et al., 1984). Several
groups of researchers investigated the effectiveness of the alcohol extract of
pepper and found it effective against cotton boll weevil, rice weevil and
Drosophila (Su, 1977; Scott and McKibben, 1978; Barakat et al., 1985). The
LD50 value for the tropical application of the crude and purified extracts, obtained
after 24 h showed a mortality rate of 3.4 and 4.8 mg, respectively, per adult insect
of Sitophilus oryzae and 4.5 and 7.2 mg, respectively, per adult insect of
Callosobrunchus maculatus. The oleoresin, piperine and other amides in pepper
were found to be lethal to culex mosquito (Culex pipiens) and also to pulse weevil
(Su, 1977). Pepper amides, such as piperonal, piperine, piperoline, pellitorine,
pepercide and dihydropipercide, also were investigated by Miyakado et al.
(1979), Masakazu et al. (1980) and Deshmukh et al. (1982) for their effectiveness
against insects and were found to be effective against pulse beetle, rice weevil,
and Drosophila, where larval growth inhibition was found. Desai et al. (1997)
showed that 100% mortality within 24 h at a dose of 80 ppm was obtained using
acetone extract of pepper in the case of Anopheles subpietus larvae.
XIII. CONSUMER PRODUCTS OUT OF BLACK PEPPER
Of all the spices used in the world, black pepper, is, perhaps, the most widely
used one in the kitchen, perfumery, medicine, and industry. Most of these
consumer products that have their origin in black pepper are the “value added”
ones. For instance, when pepper is used in the food processing industry, it is not
pepper as such, but the oleoresin extracted through solvent extraction and its
pungency and flavor are the ones that add to the value of the end product made by
the food processing industry.
XIV. VALUE ADDITION IN PEPPER
There has been a dramatic advancement in the field of value addition of black
pepper and diversification of processed pepper products. The value added pepper
products are classified as follows:
1. Green pepper based products

2. Black and white pepper based products
3. Pepper by-products
The above may be further classified as follows (Pruthi, 1997).
1. Green Pepper Based Products
a. Canned green pepper in brine,

b. Bottled green pepper in brine,
c. Bulk packed green pepper in brine,
d. Cured green pepper (without any covering tissue),
e. Frozen green pepper,
f. Freeze-dried green pepper,
g. Sun dried or dehydrated green pepper,
h. Green pepper pickles in oil, vinegar or brine,
i. Green pepper mixed pickle in oil, vinegar or brine,
366 K. P. P. NAIR
j. Green pepper flavored products, and

k. Green pepper paste.
2. Black and White Pepper Based Products
a. Black pepper powder,

b. White pepper powder,
c. White pepper whole,
d. Pepper oleoresin,
e. Pepper oil,
f. Other pepper products,
g. By products from pepper waste,
h. Miscellaneous forms of utilization, such as in medicine, culinary, and
industry.
The Pepper Marketing Board of Malaysia (PMBM), as well as many privately

owned industrial houses have brought to the market a number of innovative end
products of pepper and have facilitated planning and product development of
black pepper. The following structure has been proposed by PMBM (Abdulla,
1997). The pepper products are whole black pepper in retail packs for table use,
ground pepper in retail packs and dispensers. Pepper based products are flavored
ground pepper, such as lemon pepper, garlic pepper, sauces, and marinades,
which have pepper as the primary component and also pepper paste. Spice
mixtures and blends are curry powders and spice blends for specific cuisines,
such as, five-spice powder, soup blends, etc.
In addition to the above, there are pepper flavored products, which are pepper
mayonnaise, pepper tofu, pepper cookies, pepper keropok (which are prawn or fish
crackers). Products that use pepper extracts are pepper candy, pepper perfume,
etc. Pepper is also put to other auxiliary uses, such as in the case of pepper stalk,
which is used as a substrate for growing mushrooms.
There are a number of common black pepper products of which the most
important is ground pepper. This is the exclusive form in which black pepper is
used on the table. It is also the only spice served along with food on board and
also used in restaurants specializing in fast foods. Pepper is very commonly used
for seasoning food at different stages of preparation. The following factors have
to be kept in mind in the case of ground pepper use, which is the easiest to
manufacture and market. Freedom from bacteria and mould is the most
important, as otherwise it is a potential health hazard. During the grinding
process, volatile oil content can be adversely affected. It should be ensured that it
is not so. Since high moisture content in the berries will adversely affect shelf life,
moisture during storage should be kept to a minimum. Optimum particle-size
should be ensured for free flow during the period of storage and active use.
Another very important aspect is packaging. Without good packaging, the

product will not only not have a good market value but also can deteriorate during
storage. Above all, marketable pepper should be free from all extraneous matter.
Flavored pepper is an important product with good market for culinary
purposes. Lemon and garlic pepper are examples. The former contains lime
powder, pepper powder and salt and the latter is a blend of dehydrated garlic
powder and dried pepper powder. Both are table condiments, especially in non-
vegetarian dishes, containing chicken, fish, etc. The PMBM has developed these
products. Pepper sauce is another product of PMBM which can be used as a
marinade especially in meat-based dishes, such as steak, lamb, etc., and is also
used in stir-frying. The preparation of this product involves the use of black
pepper, soybean extract, garlic and other ingredients to obtain a relatively mild
dark sauce. Pepper is also used extensively in the preparation of a variety of
sauces including Worcestershire sauce. Spice mixtures and blends are
extensively used in continental kitchens as well as in regional kitchens. Pepper
based spice mixes, available in departmental stores and supermarkets, are widely
used in different meat based preparations. There are also soup mixes which
contain pepper. Private industrial houses and PMBM have developed a variety of
pepper flavored products, such as pepper flavored mayonnaise, egg tofu, savory
pepper cookies, traditional biscuits with pepper flavor, pepper flavored prawn and
fish crackers, etc. Many traditional dishes and snacks, when dressed with pepper
will result in enhanced taste and flavor. Pepper flavored vinegar and pepper salts
are excellent taste enhancers. Pepper is widely used in industrially manufactured
food items. Almost all fish and meat based products use pepper for seasoning,
which enhances both flavor and pungency. Industrially manufactured foods, such
as soups, pickles, etc., use pepper extracts or powdered pepper. Additionally,
there are pepper flavored beverages, such as, pepper tea, pepper coffee, pepper
flavored milk, etc. Some brands of brandy contain piperine and pepper oil, which
impart a pungent taste and an “exotic” flavor. In the perfumery industry, volatile
oil from pepper is used to impart an exclusive spicy “oriental” touch. Some of the
more popular brands of perfume that contain pepper volatile oil are Revlon’s
“Charlie” and Christian Dior’s “Poison” and Malaysian perfumes, such as
“Sensai” and “Amila” (Ng, 1993).
3. Green Pepper Based Products
Pepper fruits that are not fully mature are used in the manufacture of green
pepper based products, such as pepper in brine, pepper in oil, pepper in vinegar,
desiccated green pepper, freeze-dried green pepper, pepper paste, etc. Almost all
of the green pepper products are used by the catering sector to be served with meat
dishes, such as steak and pork. The food industry also uses green pepper in the
production of certain specialized cheeses. The other green pepper based products
368 K. P. P. NAIR
are pickled green pepper and pepper spike. Tender pepper spikes and fruits alone or
in combination with tender cardamom, pickled in vinegar and salt or sugar make
very delicious dishes which can be served with dinner. Green pepper paste is
another product. The green pepper paste, in polypacks or bottles, are now a
common sight in supermarkets. This can be used as a substitute to pepper powder
and it gives a refreshing taste with flavor and a unique “bite”, which enhances the
palatability of fish and meat dishes, and is a common choice of well trained chefs.
4. Other Ancillary Pepper Based By-products
An industrial area worth greater research and development is the ancillary

pepper-based by-products, which are of uncommon use. For instance, Malaysia
has succeeded in producing paper and board from pepper stalks. These materials,
especially paper made of pepper stalk make good invitation cards, having a unique
texture and color pattern (Ng, 1993). Pepper stalk makes a good substrate for oyster
cultivation when grown in a 1:1 mixture of pepper stalk and shredded paper (Siti
Hajijah and Bong, 1993). In Malaysia, the remnants of the pepper processing
industry, which consists mainly of stalk and pericarp, are powdered and used as an
organic manure. This is called pepper dust. It can enrich soil fertility when used in
combination with other organic agricultural wastes (Ng, 1993).
In addition to the above, there are a number of other pepper-based commercial
formulations. Though a number of commercial formulations, such as canned
green pepper, green pepper in brine, pepper oil, oleoresin, etc., are readily
available in the market, they still are to find widespread acceptability by the
consumer. The most recent use of the spice processing technology is the
encapsulation of the flavoring components, namely, spice oils and oleoresin.
These can be fully exploited commercially on a large scale, provided capital and
entrepreneurship are readily available. This is yet to happen.
There is now a changing consumer preference for white pepper from black
pepper in some western countries. White pepper is used both as table pepper and
also for enriching the flavor and taste of crab soups, sea food salads, casseroles,
chicken, fish, and egg preparations. Sauces like mayonnaise also contain white
pepper as an ingredient. One of the primary reasons for a shift in preference to
white pepper from black pepper in western countries is because the black specks
from the skin of black pepper are unseemly to the eye on a dining table.
The pepper product that is becoming increasingly popular in European
countries is the green pepper. As the piperine and pepper oil levels are high in
slightly immature corns, they are preferred in the preparation of green pepper.
This gives a characteristic flavor and is soft with a “bite” — a unique feel on the
palate. These qualities make it ideal for garnishing meat dishes. The major
products are green pepper in brine, canned green pepper, dehydrated green
pepper, freeze-dried green pepper, green pepper paste, etc. While an increasing
number of consumers prefer green pepper in brine and green pepper paste, and
their demand is on the increase, demand for canned pepper is not much as its cost
of production is quite high. India, Brazil, and Malaysia are the major producers of
these value added pepper products. The catering industry is the main source of
demand, followed by the food manufacturing industry for these products.
Among the flavoring agents, the most important are oleoresin, pepper oil, and
encapsulated pepper. The most extensive use of oleoresin is in the flavoring of
meat. The other end uses of oleoresin are for preparing pickles, sauces, gravies,
dressing “chutneys” (unique Indian spicy eatables), soups, and snacks. In most
meat preparations pepper is used to impart flavor. The same task is accomplished
by oleoresin as well. One of the most recent advances made by the Regional
Research Laboratory in Trivandrum, Kerala State, India, is in developing the
spray drying encapsulation technique for oleoresin. The encapsulated flavor
powder is used in a variety of products, such as cake mixes, dry beverage mixes,
desserts, soup mixes, dusting on potato chips, and nuts. These have an emerging
market all over the world.
XV. CONCLUSIONS AND A PEEP INTO

PEPPER’S FUTURE
Pepper has certainly had a checkered history. Within the past, present and
future, pepper can indeed be termed the “King of Spices”. Though beset by many
problems, both economic and agronomic, it is a safe bet that pepper will emerge
as the world’s most sought after spice. It has been estimated that by the year 2020
global demand for pepper would be about 28,0000 metric tons, which is projected
to escalate to 360,000 metric tons by the year 2050. This would entail almost
doubling the present production in the first half century of the current
millennium. Where will the additional output come from? Increase of pepper
area could only be marginal on a global basis. India, the major producer, will,
perhaps, experience only marginal area increases. On the Asian subcontinent, it is
only in China and Vietnam, the latter already emerging as a key producer, where
marginal area increases can be expected. Like India, only a marginal increase in
area can be expected in Indonesia. Malaysia and Thailand, the other two pepper
producers of Asia, are already experiencing a decrease in area. This is also true
for Brazil, a key producer on the Latin American continent. Labor is becoming
increasingly difficult to obtain and very expensive when found and, hence,
countries such as Malaysia and Thailand on the Asian continent and Brazil on the
Latin American continent are moving away from the labor intensive pepper
production. The impact of globalization is clearly seen in these countries, where
market forces are shying away from such labor intensive crops such as pepper in
370 K. P. P. NAIR
preference to more mechanically managed crops. By the very nature of the

pepper plant and the pepper canopy, mechanization has only very limited
possibilities.
From the foregoing review, it can clearly be seen that one of the most pressing
needs of pepper production is to have a pepper ideotype that combines many
positive attributes in terms of boosting production potential, at the same time,
resisting many ravages of nature, especially those of diseases, among which the
most predominant is Foot Rot, caused by the Phytophthora fungi. There is yet no
single pepper variety that is universally resistant to this dreaded disease, though
some lines have been identified which have shown regional promise.
Conventional breeding, which will take years to accomplish, since pepper is a
long duration perennial, will not be the answer to resolve this major problem.
Biotechnology is a promising area, but there are vast gaps in knowledge.
Behavior of the pepper plant at the molecular level is far from clearly understood.
Though tissue culture is a promising area in terms of cutting down the long gap in
producing a desirable ideotype, this will not give final answers because there are
far too many gray area variables in the technique. Gene sequencing, currently a
very highly sophisticated technology with very high levels of investments, opens
up a door in the case of pepper as well. At the current level of global scientific
knowledge and available technical expertise in the case of pepper, it is a safe bet
to say that, for whom pepper is a great fascination among crops, it is a very long
road to traverse. Sophisticated instrumentation is only part of the answer, with the
availability of superlative technical skills being of greater importance, which the
pepper producing countries, India included, clearly lack. What one might expect
on the pepper front, at least for the few decades in the current century to come, is
the following.
As science advances and global populations become more and more health
conscious, the demand for organically grown food would enlarge. One of the
most promising areas for future pepper development will be in the area of
producing “organically grown” or what one might term “clean pepper”. The
environmental scare generated by the so-called “green revolution”, the hall
mark of which has been extensive and indiscriminate use of chemical
fertilizers and pesticides, is driving more and more people to organically
grown food and the demand for organically grown pepper is no exception.
The potential market for organically grown food products, including pepper,
is most promising in the USA followed by Germany. The rigors of growing
organic foods are encompassed in the Codex Alimentaris stipulations
developed by France and the USA has its own stipulations. In a country
like India where pepper production on the farm and its further processing is
not subject to the rigors of quality control, export of organically grown
pepper can only be put on sound footing if production and processing
methods are well streamlined and rigorously controlled for quality, sanitation
being the most important. The same would hold good inasmuch as the prevalent
situations in other pepper producing countries like Indonesia, Vietnam,

Malaysia, Thailand, etc., are concerned.
Another area which has much promise is the development of pepper derived
food additives. The potential of this is gradually unfolding on an industrial
scale. Heavy investments are necessary to make the various products price
competitive. Some regional success stories do exist, though. For instance, the
pepper based products, which have a ready regional market. Yet, these products
do not have a global reach. The potential market for these products would be
Europe and the USA, where a gradual change in food habits is taking place.
Both post-harvest technology and product development would undergo
remarkable changes in the future. Imaginative product diversification would
lead to the development of an array of novel products for the consumer,
currently unheard of. Pepper with high flavor, but low pungency, will begin to
be increasingly used in the manufacture of liquors, fruit juices, bakery products
and even choice perfumes. A possible challenge to the development of pepper
varieties with high flavor will be the creation of the in vitro pepper flavor from
bioprocessors. If this comes about, the only way natural pepper can be salvaged
is a preference of the consumer for the clean, natural spice, rather than the
“laboratory created spice”.
Another important area that needs to be thoroughly researched is the
management of soil fertility of pepper soils. Almost the entire approach to
fertility management in pepper soils is based on classical textbook knowledge,
where empirical recommendations generated from micro plots are extrapolated
to large-scale field conditions. These experimental micro plots are nothing but
artefacts and many of the recommendations which emanate from such studies,
when applied to large-scale plantations, turn out to be quite off the mark in
reproducing accountable results. This often shatters farmers’ confidence. A
very significant departure is being made after the concerted efforts of the
author, who has developed an entirely new approach to soil testing and
fertilizer management, based on nutrient buffering. The concept is now
known, universally, as “the nutrient buffer power concept”. A detailed
discussion on the concept and its relevance in pepper nutrition is given in
Chapter IV. In India, in the state of Kerala, where pepper is grown to the
largest extent, it has been observed that fertilizer input can be significantly
reduced by taking into consideration the buffer power of the nutrient
concerned. Experimental and field results (Nair, 2002) show that, as against
the routine soil testing and fertilizer recommendations, addition of Zn
fertilizer, a very crucial input in pepper production, can be reduced by almost
75% in some soils.
There is an urgent need to extend the concept to other crucial nutrients, such
as phosphorus and potassium, and possibly nitrogen as well. Of the first two,
phosphorus is more important in view of the fact that soils of Kerala State,
the home of pepper, are lateritic and much of the applied phosphate fertilizer,
372 K. P. P. NAIR
based on routine soil testing is not available to the crops. It is in this context
that this new concept holds out much hope for pepper farming. However, it
must be emphasized that the success of a new approach, to a great extent, rests
with the ingenuity of those applying it to suit the demands of a new situation.
This principle is no exception to making the nutrient buffer power concept
succeed in the case of pepper production, as has been the case with other
crops, such as, maize, rye, white clover, and even a perennial crop like
cardamom, tested by the author (Nair et al., 1997). The fact that pepper is a
perennial crop makes it all the more important because, unlike an annual one,
where a mid course correction can be effected in the following season, the
fertilizer regime has to be correct from the very beginning, because pepper
grows for upwards of 25 years. Unlike routine soil testing, the new approach
calls for an accurate determination of the buffer power of the nutrient in
question at the very start. Once this is accomplished, the buffer power factor
can be integrated into the computations with the routine soil test data and
accurate fertilizer recommendations can be made on the basis of this new
information. This implies that, in addition to obtaining routine soil test data,
one also needs to know the buffer power. The author has obtained very
encouraging results with the new concept in pepper production in Kerala with
regard to Zn fertilization, which signals a very promising change in the
fertilizer practices. Hopefully, the new concept could very successfully be
extended to other important plant nutrients.
Agriculture is the engine for development, nationally and globally, as is being
increasingly proved against the unfolding scenario of globalization. A nation
with a strong agricultural base can be expected to be in the forefront of economic
development. Pepper is the crop of the tropics and, in that sense, is very much a
part of the development of the third world economy. A lot has been achieved in
pepper production, but, before it moves to center stage, as a commanding crop of
third world economy, much more needs to be done. This calls for the concerted
efforts of not just agronomists, soil scientists or plant breeders dealing with
pepper, but a whole range of visionary thinkers and planners, who can really
make the crop, not just the king of spices, but the monarch of third world
agricultural economy.
ACKNOWLEDGMENT
With love I dedicate this chapter to my wife, Pankajam, my all, who sustains me
in this very difficult journey, that life is.
REFERENCES
Abdel-Fattah, I. E., and EI-Zeany, B. A. (1979). Effect of spices on the autooxidation of fatty foods.
Rev. Ital. Soztanze Grasse. 56, 441 –443.
Abdulla, A. (1997). Product development and promotion of pepper products. In “Proceedings of the
22nd Pepper Tech. Meet”, pp. 1–6, Cochin, Kerala State, India.
Adams, F. (1971). Ionic concentrations and activities in soil solutions. Soil Sci. Soc. Am. Proc. 35,
420 –426.
Adzemi, M. A., Hamdan, J., and Sjahril, J. S. (1993). Micro-nutrient removal studies on black pepper
(Piper nigrum L.) in Sarawak. In “The Pepper Industry: Problems and Prospects” (M. Y. Ibrahim,
C. F. J. Bong, and I. P. Ipor, Eds.), pp. 104– 109. Universiti Pertanian Malaysia, Sarawak.
Agarwal, S. (1988). Shoot tip culture of pepper and its micropropagation. Current Sci. 57,
1347–1348.
Ahmed, K. (1993). The effect of Nitrogen on the growth of Piper nigrum L. planted with Desmodium
trifolium cover. In “The Pepper Industry: Problems and Prospects” (M. Y. Ibrahim, C. F. J. Bong,
and I. P. Ipor, Eds.), pp. 97–105. Universiti Pertanian Malaysia, Sarawak.
Anandaraj, M. (1997). Ecology of Phytophthora capsici (Leonian 1922, emend A. Alizadeh and
P.H. Tsao) In “Causal Organism of Foot Rot Disease of Black Pepper (Piper nigrum L.)”, p. 154.
Ph. D. Thesis, University of Calicut, Kerala State, India.
Anandaraj, M., and Peter, K. V. (1996). In “Biological Control in Spices”, p. 52. Indian Institute of
Spices Research, Calicut, Kerala State, India.
Anandaraj, M., and Sarma, Y. R. (1994a). Effect of Vesicular Arbuscular Mycorrhiza on rooting of
black pepper (Piper nigrum L.). J. Spices Aromatic Crops. 3, 39– 42.
Anandaraj, M., and Sarma, Y. R. (1994b). Biological control of black pepper diseases. Indian Cocoa,
Aracanut Spices J. 18, 22–23.
Anandaraj, M., and Sarma, Y. R. (1995). Diseases of black pepper (Piper nigrum L.) and their
management. J. Spices Aromatic Crops 4, 17–23.
Anandaraj, M., and Sarma, Y. R. (1997). Mature coconut water for mass culture of biocontrol agents.
J. Plantation Crops. 25, 112 –114.
Anandaraj, M., Sarma, Y. R., and Ramachandran, N. (1994). Phytophthora root rot of black pepper in
relation to age of the host and its culmination in foot rot. Indian Phytopath. 47, 203–206.
Annual Report. (1997). “Indian Institute of Spices Research-Vision 2020”. IISR Perspective Plan,
Calicut, Kerala State, India.
Anon. (1989). “Technology transfer document of microencapsulated spice flovours Regional
Research Laboratory”, pp. 9–10. Council of Scientific and Industrial Research, New Delhi, India,
Trivandrum, India.
Anon. (1995). Research Branch, Department of Agriculture, Sarawak, Malaysia, Annual Report.
Anon. (1996). Israel — A promising market for Indian spices. Spice India 9, 2–3.
Anon. (1997). Spice export grows to an all time record. Spice India 10, 2–4.
Apun, K., Siti Hajijah, A. S., and Yusof, I. M. (1993). Microbiological properties of pepper during
processing. In “The Pepper Industry: Problems and Prospects” (M. Y. Ibrahim, C. F. J. Bong, and
I. P. Ipor, Eds.), pp. 337 –341. Universiti Pertanian Malaysia, Sarawak.
Atal, C. K., Dhar, K. L., and Singh, J. (1975). The Chemistry of Indian Piper. Lloydia 38, 250–264.
Ayier, K. N., and Kolammal, M. (1966). “Pharmacognosy of Ayurvedic Plants”, 9, p. 63. Kerala
University, Trivandrum, Kerala State, India.
Bade, J., and Smit, H. P. (1992). An annual model of the world pepper economy, p. 63. Seminar on
Commodity Analysis, Jakarta, Indonesia, ESI-VU, Amsterdam.
Balasubramanyam, N., Mahadevan, B., and Anandaswamy, B. (1978). Packaging and storage studies
on ground block pepper (Piper nigrum L.) in flexible consumer packages. Indian Spices 15(4),
6 –11.
374 K. P. P. NAIR
Bano, G., Alma, V., Raina, U., and Zutshi, V. (1969). The effect of piperine on the pharmokinetics of
phenytoin in healthy volunteers. Planta Medica 12, 566–569.
Barakat, A. A., Family, H. S. M., Kandil, M. A., and Ebrahim, N. M. M. (1985). Toxicity of the
extract of black pepper, cumin, fennel, chamomile and Lupin against Drosophila ceratetis and
Spodoptera. Indian J. Agric. Sci. 55, 116 –120.
Barber, S. A. (1984). “Soil nutrient bioavailability: a mechnistic approach”. Wiley, New York.
Barber, S. A. (1995). “Soil nutrient bioavailability: a mechnistic approach”. Wiley, New York.
Bavappa, K. V. A., and Gurusinghae (1978). Rapid multiplication of black pepper for commercial
planting. J. Plantation Crops 6, 92– 95.
Bergman, A. (1940). De pepper culture en-handel op Bangka. Landbouw 16, 139–256.
Biessar, S. (1969). Plant parasitic nematodes of crops in Guyana. PANS 15, 74–75.
Bopaiah, B. M., and Khader, K. B. A. (1989). Effect of bio-fertilizers on growth of black pepper (Piper
nigrum L.). Indian J. Agric. Sci. 59, 682–683.
Bridge, J. (1978). Plant Nematodes Associated with Cloves and Black Pepper in Sumatra and Bangka,
Indonesia. In “O.D.M. Technical Report”, p. 19. U.K. Ministry of Overseas Development, U.K.
Broome, O. C., and Zimmerman, R. N. (1978). In vitro propagation of black pepper. Hortic. Sci. 43,
151–153.
Burger, W. C. (1972). The evolutionary trends in the Central American species of Piper. Brittonia. 24,
356–362.
Burger, K., and Smit, H. P. (1997). The Natural Rubber Market, Woodhead Publishing Limited,
Cambridge, p. 354.
Chaudhary, R., and Chandel, K. P. S. (1994). Germination studies and cryo preservation of seeds of
black pepper (Piper nigrum L.), a recalicitrant species. Cryo-Lett. 15, 145–150.
Cheriankunju, N. E. (1996a). Know your market for spices — Japan. Spice India 9, 5–7.
Cheriankunju, N. E. (1996b). Know your market for spices — USA. Spice India 9, 2–5.
Chiapault, J. R., Mizuna, G. R., Hawkins, J. M., and Lundberg, W. O. (1955). The antioxident
properties of spices in oil-in-water emulsions. Food Res. 20, 443–451.
Christie, J. R. (1959). “Plant Nematodes — Their Bionomics and Control”, p. 256. University of
Florida, Gainesville, Florida, USA.
Chu, C. V., Chang, J. P., and Wang, C. J. (1994). Modulatory effect of piperine on benzopyrene
cytotoxicity and DNA adduct formation to V-79 lung fibroblast cells. Food Chem. Toxicol. 39,
373–377.
Chua, B. K. (1981). Studies on in vitro propagation of black pepper (Piper nigrum L.). MARDI Res.
Bull. 8, 155 –162.
Claassen, N., and Barber, S. A. (1976). Simulation model for nutrient uptake from soil by growing
plant root system. Agron. J. 68, 961–964.
Claassen, N., Hendriks, K., and Jungk, A. (1981). Rubidium — Verarmung des wurzelnahen Bodens
durch Maispflanzen. Z. Pflanzenern. Bodenk. 144, 533 –545.
Coffey, M. D. (1991). Strategies for the integrated control of soil borne Phytophthora species.
In “Phytophthora” (J. A. Lucas, R. C. Shattock, D. S. Shaw, and L. R. Cooke, Eds.), pp. 411– 432.
Cambridge University Press.
Collins, L., and Lapieree, D. (1976). Freedom at Midnight. Vikas Publication House, New Delhi,
India.
Cook, R. J., and Baker, R. (1983). “The Nature and Practice of Biological Control of Plant Pathogens”,
p. 53. American Phytopathological Soc., St. Paul, Minnesota, U.S.A.
Dalvi, P. R., and Dalvi, P. S. (1991). Comparison of the effect of piperine administered intragastrically
and intraperitonially on the liver and liver mixed function oxidases in rats. Drug Metabol. Drug
Interact. 9, 23–30.
Darrington, H. (1991). Herbs and spices. Food Manuf. 66(11), 21– 23.
Das, T. P. M., and Cheeran, A. (1986). Infectivity of Phytophthora species on cash crops in Kerala.
Agric. Res. J. Kerala 24, 7– 13.
Das, P. C., Das, A., Mandal, S., and Chatterjee, A. (1992). On the validity of the ethnic use of
P. nigrum and M. indica L. In “Proceedings of the International Seminar on Traditional
Medicine”, p. 84. Calcutta, India.
Datta, P. K. (1984). “Studies on two Phytophthora diseases (Koleroga of arecanut and black pepper
wilt) in Shimoga District, Karnataka State, India”, p. 121. Ph.D. University of Agricultural
Sciences, College of Agriculture, Dharwad, Karnataka State, India.
Davis, R. M., and Menge, J. A. (1981). Phytophthora parasitica inoculation and intensity of
vesicular–arbuscular mycorrhizae in citrus. New Phytol. 87, 705 –715.
Davis, R. M., Menge, J. A., and Zentmyer, G. A. (1978). Influence of vesicular arbuscular mycorrhiza
on Phytophthora root rot of three crop plants. Phytopathology 68, 1614–1617.
De Waard, P. W. F. (1964). Pepper cultivation in Sarawak. World Crops 16, 24–30.
De Waard, P. W. F. (1969). “Foliar diagnosis, nutrition and yield stability of black pepper (Piper
nigrum L) in Sarawak”, p. 150. Comm. No. 58, Royal Tropical Institute, Amsterdam, The
Netherlands.
De Waard, P. W. F. (1979). Yellow disease complex in black pepper on the island of Bangka,
Indonesia. J. Plantation Crops 7, 42 –49.
De Waard, P. W. F. (1984). Piper nigrum L. In “Plant Resources of South East Asia” (E. West Phal
and P. C. M Jansen, Eds.), pp. 225–230. Rijksherbarium, The Netherlands.
Debrauwere, J., and Verzele, M. (1975a). Constituents of pepper VI. Oxygenated fraction of pepper
essential oil. Bull. Soc. Chem. Belg. 84, 167 –177.
Debrauwere, J., and Verzele, M. (1975b). New Constituents of the oxygenated fraction of pepper
essential oil. J. Sci. Food Agric. 26, 1887–1894.
Debrauwere, J., and Verzele, M. (1976). The hydrocarbons of pepper essential oil. J. Chrom. Sci. 14,
296 –298.
Deciyanto, S., and Suprapto (1992). Research progress on important insect pests of black pepper in
Indonesia. In “Proceedings of the International Workshop on Black Pepper Diseases 3–5
December, 1991, Bandar Lampung, Indonesia” (P. Wahid, D. Sitepu, D. Deciyanto, and
U. Suparman, Eds.), pp. 237 –244. Research Institute for Spice and Medicinal Crops, Bogor,
Indonesia.
Dehne, H. W. (1982). Interaction between vesicular–arbuscular mycorrhizal fungi and plant
pathogens. Phytopathology 72, 1115–1119.
Desai, A. E., Ladhe, R. U., and Deorag, B. M. (1997). Larvicidal activity of acetone extract of Piper
nigrum seeds to a major vector, Anopheles subpictus from Nashik region. Geobios 24, 13– 16.
Deshmukh, P. B., Chavan, S. R., and Renapurkar, D. M. (1982). A study of insecticidal activity of
twenty indigenous plants. Pesticides 16, 7.
Devasahayam, S., and Anandaraj, M. (1997). Black pepper (Piper nigrum L.) research. In “Neem in
Sustainable Agriculture” (S. S. Narwal, P. Tauro, and S. S. Bisla, Eds.), pp. 117 –122. Scientific
Publishers, Jodhpur.
Devasahayam, S., and Koya, K. M. A. (1994). Natural enemies of major insect pests of black pepper
(Piper nigrum L.) in India. J. Spices Aromatic Crops 3, 50–55.
Devasahayam, S., and Leela, N. K. (1997). Evaluation of plant products for antifeedant activity
against pollu beetle (Longitarsus nigripennis Motschulsky), a major pest of black pepper.
In “Abst. First Natn. Sym. Pest Management in Horticultural Crops”, p. 85. Environmental
Implications and Thrusts, Bangalore.
Devasahayam, S., Premkumar, T., and Koya, K. M. A. (1988). Insect pests of black pepper Piper
nigrum L. in India — a review. J. Plantation Crops 16, 1 –11.
Drew, M. C., and Lynch, J. M. (1980). Soil anaerobiosis, microorganisms and root function. Ann. Rev.
Phytopath. 18, 37– 67.
Drew, M. C., Nye, P. H., and Vaidyanathan, L. V. (1969). The supply of nutrient ions by diffusion to
plant roots in soil. I. Absorption of potassium by cylindrical roots of onion and leek. Plant Soil 30,
252 –270.
376 K. P. P. NAIR
Dudek, D. H. (1996). “Sterilization method and apparatus for spices and herbs”, US Patent US
5523053 [US 260068 (940615)].
Eapen, S. J., and Ramana, K. V. (1996). Biological control of plant parasitic nematodes of spices. In
“Biological Control in Spices” (M. Anandaraj and K. V. Peter, Eds.), pp. 20 –32. Indian Institute
of Spices Research, Calicut, Kerala State, India.
Eapen, S. J., Ramana, K. V., and Sarma, Y. R. (1997). Evaluation of Psuedomonas fluorescens isolates
for control of Meloidogyne incognita in black pepper (Piper nigrum L.). In “Biotechnology of
Spices, Medicinal and Aromatic Plants” (S. Edison, K. V. Ramana, B. Sasikumar, K. Nirmal
Babu, and S. J. Eapen, Eds.). Indian Society for Spices, Calicut, Kerala State, India.
Ebenhoech, A., and Spadaro, O. (1992). Antiepilepsirine: a new Chinese anticonvulsant herb drug.
J. Eco. Tax. Bot. 16, 99–102.
El Mofty, M. M., Khudoley, V. V., and Shwaireb, M. H. (1991). Carcinogenic effect of force-feeding
an extract of black pepper in Egyptian toads. Oncology 48, 347–350.
Elgawhary, S. M., Lindsay, W. L., and Kemper, W. D. (1970). Effects of EDTA on the self-diffusion
of zinc in aqueous solution and in soil. Soil Sci. Soc. Am. Proc. 34, 66–70.
Esper, A., and Muhlbauer, W. (1996). Solar tunnel dryer for fruits. Plant Res. Develop. 44, 61–79.
Ewald, S. (1991). Vesicular – Arbuscular Mycorrhiza Management in Tropical Agrosystems.
Technical Cooperation, Federal Republic of Germany, 371.
Fassuliotis, G., and Bhatt, D. P. (1982). Potential of tissue culture for breeding root knot nematode
resistance in vegetables. J. Nematol. 14, 10–14.
Ferraz, E. C. A., Orchard, J. E., and Lopez, A. S. (1984). Reactions of black pepper to Meloidogyne
incognita in relation to total phenol. Revista Theobroma. 14, 217–227.
Ferraz, E. C. A., Lordello, L. G. E., and de Santana, C. J. L. (1988). Nutrient absorption of black
pepper vine (Piper nigrum L.) infested with Meloidogyne incognita Kofoid & White (1919),
Chitwood (1949). In “Boletin Tecnico Centro de Pesquisas do Cacau, Brazil, No. 160”, p. 34.
Ferraz, E. C. A., Lordello, L. G. E., and Gonzaga, E. (1989). Influence of Meloidogyne incognita
(Kofoid and White 1919) Chitwood 1949 on chlorophyll content of black pepper (Piper nigrum L).
Agrotropica 1, 57–62.
Findlay-Wilson, I. Z. F. L. (1995). Microencapsulated flavours for a great taste. Internationale
Zeitschrift für Lebensmittel — Technik, Marketing, Verpackung und Analytik 46(10), 76 –78.
Fitchet, M. (1988a). Establishment of black pepper in tissue culture “Information Bulletin No. 189.”
Citrus and Subtropical Fruits Research Institute, South Africa.
Fitchet, M. (1988b). Progress with in vitro experiments of black pepper. In “Information Bulletin No.
196”. Citrus and Subtropical Fruits Research Institute, South Africa.
Freire, F. C. O., and Bridge, J. (1985b). Biochemical changes induced in roots and xylem sap of black
pepper by Meloidogyne incognita. Fitopatologia Brasileira 10, 483 –497.
Freire, F. C. O, and Bridge, J. (1985d). Parasitism of eggs, females of juveniles of Meloidogyne
incognita by Paecilomyces lilacinus and Verticillium chlamydosporium. Fitopatologia Brasileira
10, 577 –596.
Fresco, L. O., and Kroonenberg, S. B. (1992). Time and spatial scales in ecological sustainability.
Land Use Policy. July, 1992, 155–167.
Gamble, J. S. (1925). Flora of the Presidency of Madras. In “Bot. Sur. India”, Vol. II. Calcutta (Repr.).
Geetha, S. M. (1991). In “Studies on the Biology, Pathogenicity and Biocontrol of Different
Populations of Radopholus similis”, p. 196. Ph. D. Thesis, Kerala University, Trivandrum, India.
Geetha, C. K., and Nair, P. C. S. (1990). Effect of plant growth regulators and zinc on spike shedding
and quality of pepper. Agric. Res. J. Kerala 14, 10–12.
Geetha, S. P., Manjula, C., and Sajina, A. (1995). In vitro conservation of genetic resources of spices.
In “Proceedings of the Seventh Kerala Science Congress, January 27 –29”, pp. 12 –16, Science
Technology and Environment Committee, Kerala Government.
George, C. K. (1982). Pepper cultivation in Malaysia. Indian Cocoa, Arecanut and Spices J. 5,
75–76.
George, P. S., Nair, K. N., and Pushpangadan, K. (1989). “The Pepper Economy of India”. Oxford and
IBH Publishing Co. Pvt. Ltd., New Delhi, India.
Ghawas, M., and Miswan, J. (1984). Satu Kaedah memprecepatkan Pembiakn lada hitam (Piper
nigrum L.). Teknol. Pertanian. MARDI Res. J. 5, 50–54.
Gopalakrishnan, M., Menon, N., Padmakumari, K. P., Jayalakshmi, A., and Narayanan, C. S. (1993).
GC analysis and odor profiles of four new Indian genotypes of Piper nigrum L. J. Essential Oil
Res. 5, 247–253.
Govindan, M., and Chandy, K. C. (1985). Utilization of the diazotroph, Azospirillum for inducing
rooting in pepper cuttings (Piper nigrum). Curr. Sci. 54, 1186–1188.
Govindarajan, V. S. (1977). Pepper — Chemistry, Technology and Quality Evaluation. Crit. Rev.
Food. Sci. Nut. 9, 115– 225.
Govindarajan, V. S. (1979). Pepper — chemistry, technology and quality evalutation. CRC Crit. Rev.
Food. Sci. Nut. 9, 1–115.
Graham, J. H. (1982). Effect of citrus root exudates on germination of chlamydospores of the
vesicular–arbuscular mycorrhizal fungus Glomus epigaeum. Mycologia 74, 831 –835.
Graham, J. H. (1988). Interactions of mycorrhizal fungi with soil borne plant pathogens and other
organisms, an introduction. Phytopathology 78, 365–366.
Graham, J. H., and Egel, D. S. (1988). Phytophthora root rot development on mycorrhizal and
phosphorus-fertilized non-mycorrhizal sweet orange seedlings. Plant Dis. 72, 614–615.
Guenther, E. (1950). Essential oils of the plant family Piperaceae. In “The Essential Oils”, vol 5,
pp. 135–161. Robert Krieger Pub. Co., New York.
Harden, H. J., and White, J. T. H. (1934). Quoted by De Waard (1969).
Hartman, C. P., Divakar, N. G., and Nagaraja Rao, U. N. (1973). A study of identification of papaya
seeds in pepper. J. Food. Sci. Technol. 10, 43– 45.
Harvill, E. K., Hartzell, A., and Arthur, J. M. (1943). Toxicity of piperine solution to house flies
control. Boyce Thompson Inst. 13, 87–92 (Merck Index, 1976, 1266).
Hashim, S., Aboobaker, V. S., Madhubala, R., Bhattacharya, R. K., and Rao, A. R. (1994). Modulatory
effects of essential oils from spices on the formation of DNA adducts by aflatoxin B in vitro. Nutr.
Cancer 21, 169 –175.
Hasselstrom, T. F., Hewitt, E. J., Konigsbacher, K. S., and Ritter, J. J. (1957). Composition of oil of
black pepper. J. Agric. Food Chem. 5, 53 –55.
Hendriks, L., Classen, N., and Jungk, A. (1981). Phosphatverarmung des wurzelnahen Bodens und
Phosphataufnahme von Mais und Raps. Z. Pflanzenern. Bodenk. 144, 486 –499.
Higashimoto, M., Purinatrapiban, J., Katoka, K., Kinouchi, T., Vinitkekumnuen, U., Akimoto, M., and
Dinishi, M. (1993). Mutagenicity and antimutagenicity of extracts of three medicinal plants.
Mutat. Res. 303, 135– 142.
Holdeman, Q. L. (1986). “The Burrowing Nematode, Radopholus similis sensu lato”, p. 52. Division
of Plant Industry, Department of Food and Agriculture and the Agriculture Commissioners of
California, California, U. S. A.
Holliday, P., and Mowat, W. P. (1963). Foot rot of Piper nigrum L. (Phytophthora palmivora).
In “Phytopathological paper no. 5”, p. 62. Commonwealth Mycological Institute, Kew, Surrey.
Hooker, J. D. (1886). “The flora of British India”, vol. 5, pp. 79 –95. L. Reeva and Co., London.
Houghton, P. J., Astaniou, A., Grange, J. M., and Yates, M. (1994). Antibacterial effects of extracts
and constituents of Piper nigrum. L. Pharm. Pharmacol. 46(Suppl 2), 1042.
Hubert, F. P. (1957). Diseases of some export crops in Indonesia. Plant Dis. Reptr. 41, 55–63.
Huhtanen, C. N. (1988). Inhibition of Clostridium botulinum by spice extracts and aliphatic alcohols.
J. Food Prot. 43, 195 –196.
Huitema, W. K. (1941). Quoted from De Waard (1969).
Humboldt, Alexander von., Bonpland., and Kunth. (1815). Cited by Burger (1972).
Ichinohe, M. (1975). Infestation of black pepper vines by the root knot nematode, Meloidogyne
incognita at Tome-Acu, Para, Brazil. Jap. J. Nematol 5, 36–40.
378 K. P. P. NAIR
Ichinohe, M. (1980). Studies on the root knot nematode of black pepper plantation in Amazon.
In “Annual Report of the Society of Plant Protection of North Japan, No. 31”, pp. 1–8.
Ichinohe, M. (1985). Integrated control of root knot nematode, Meloidogyne incognita on black
pepper plantations in the Amazonian region. Agric. Ecosystems Environ. 12, 271–283.
IISR. (1997). “Annual Report of Indian Institute of Spices Research 1996–97”, pp. 50– 51. Calicut,
Kerala State, India.
Ikeda, R. N., Stanley, W. L., Vannier, S. H., and Spitler, F. M. (1962). The monoterpene composition
of some essential oils. J. Food Sci. 27, 455– 458.
Jacob, J. A., and Kuriyan, K. J. (1979a). Screening of pepper varieties for resistance against root knot
nematode (Meloidogyne incognita). Agric. Res J. Kerala 17, 90.
Jain, S. R., and Jain, M. R. (1972). Antifungal studies on some indigenous volatile oils and their
combinations. Planta Medica 22, 136.
Jain, S. R., and Kar, A. (1971). Antibacterial activity of some essential oils and their combinations.
Planta Medica 22, 118.
Jatala, P. (1986). Biological control of plant parasitic nematodes. Ann. Rev. Phytopath. 24, 453– 489.
Jennings, W. G., and Wrolstad, R. E. (1961). Volatile constituents of black pepper. J. Food Sci. 26,
499.
Jeromi, P. D., and Ramanathan, A. (1983). World pepper market and India: An anlaysis of growth and
instability. Ind. J. Agric. Econ. 48, 89 –97.
Johri, R. K., and Zutshi, U. (1992). An Ayurvedic formulation of “Trikatu” and its constituents.
J. Ethnopharmacol. 37, 85–91.
Johri, R. K., Thusu, N., Khajuria, A., and Zutshi, U. (1992). Piperine mediated changes in permeability
of rat intestinal epithelial cells. Biochem.Pharmacol. 43, 1401–1407.
Jose, J., and Sharma, A. K. (1984). Chromosome studies in the genus Piper L. J. Indian Bot. Soc. 63,
313–319.
Joseph, B., Joseph, D., and Philip, V. J. (1996). Plant regeneration from somatic embryos in black
pepper. Plant Cell, Tissue and Organ Culture 47, 87–90.
Kachru, K. P., Patil, R. T., and Srivastava, P. K. (1990). Post harvest technology of pepper. Spice India
3(8), 12–16.
Kachru, K. P., and Gupta, R. K. (1993). Drying of spices-status and challenges. In “Proceedings of the
National Seminar of Post Harvest Technology of Spices”, pp. 15 –27, Trivandrum, India.
Kapoor, V. K., Chawla, A. S., Manojkumar, and Pradeepkumar (1993). Search for antiinflammatory
agents. India Drugs 30, 481 –493.
Kaul, I. B., and Kapil, A. (1993). Evaluation of liver protective potential of piperine-an active
principle of black pepper. Planta Medica 59, 413– 417.
Keating, M. (1993). The Earth Summit’s Agenda for Change. In “A Plain Language Version of
Agenda 21 and the Other Rio Agreements”. Centre for our Common Future, Geneva,
Switzerland.
Keerthisinghe, G., und Mengel, K. (1979). Phosphatpufferung verschiedener Böden und ihre
Veränderung infolge Phosphatalterung. Mitt. Dtsch. Bödenk, Ges. 29, 217–230.
Kelker, S. M., and Krishnamoorthy, K. V. (1996). Control of bacterial contamination in in vitro
cultures of pepper (Piper sp.) In “Proc National Seminar on Biotechnology of Spices and
Aromatic Crops”, 24–25 April 1996. Calicut, Kerala State, India.
Kerry, B. R. (1990). An assessment of progress towards microbial control of plant parasitic
nematodes. J. Nematol. 22 (Suppl.), 621– 631.
Kirtikar, K. R., and Basu, B. D. (1975). Indian Medicinal Plants III, 2133–2134 Bishan Singh
Mahendrapal Singh. Dehra Dun, India.
Koshy, P. K., and Sunderaraju, P. (1979). Response of seven black pepper cultivars to Meloidogyne
incognita. Nematol. Medit. 7, 123–125.
Kovar, J. L., and Barber, S. A. (1988). Phosphorus supply characteristics of 33 soils as influenced by
seven rates of phosphorus addition. Soil Sci. Soc. Am. J. 52, 160 –165.
Krishnamoorthy, B., and John Zachariah, T. (1992). Drying black pepper on polyethylene materials.
Indian Cocoa, Arecanut and Spices J. 15, 75.
Kueh, T. K. (1979). “Pests, Diseases and Disorders of Black Pepper in Sarawak Semongok
Agricultural Research Centre”, p. 68. Department of Agriculture, Sarawak, Malaysia.
Kueh, T. K. (1986). Pests and diseases of black pepper-a review. In “Pepper in Malaysia” (C. F. J.
Bong and M. S. Saad, Eds.), pp. 115 –133. Universiti Pertanian Malaysia Cawangan, Sarawak,
Kuching, Malaysia.
Kueh, T. K. (1990). Major diseases of black pepper and their management. The Planter. 66, 59–69.
Kueh, T. K., and Khew, K. L. (1982). Survival of Phytophthora palmvivora in soil and after passing
through alimentary canals of snails. Plant Dis. 66, 897–899.
Kueh, T. K., and Sim, S. L. (1992a). Etiology and control of Phytophthora foot rot of black pepper in
Sarawak, Malaysia. In “Proceedings of the International Workshop on Black Pepper Diseases.
Bander Lampung, Indonesia” (P. Wahid, D. Sitepu, S. Deciyanto, and U. Suparman, Eds.),
pp. 155–162. Institute for Spice and Medicinal Crops, Bogor, Indonesia.
Kueh, T. K., and Sim, S. L. (1992b). Important pepper diseases in Sarawak, Malaysia. In “Proceedings
of the International Workshop on Black Pepper Diseases Bander Lampung, Indonesia” (P. Wahid,
D. Sitepu, S. Deciyanto, and U. Suparman, Eds.), pp. 101 –117.
Kueh, T. K., Fatimah, O., and Lim, J. L. (1993). Management of black berry disease of black pepper.
In “The pepper Industry: Problems and Prospects” (M. Y. Ibrahim, C. F. J. Bong, and I. B. Ipor,
Eds.), pp. 162– 168. Universiti Pertanian Malaysia, Sarawak, Malaysia.
Kurup, P. N. V., Ramdas, V. N. K., and Joshi, P. (1979). “Handbook of Medicinal Plants”. p. 143,
New Delhi.
Kurien, S., and Nair, P. C. S. (1988). Effect of pruning on yield in pepper (Piper nigrum L.). Agric.
Res. J. Kerala 26, 137– 139.
Lawrence, B. M. (1981). “Major tropical spices-pepper (Piper nigrum L.) Essential oils 1979–80”,
pp. 141–228. Allured Publications Corporation, USA.
Lee, E. B., Shin, K. H., and Woo, W. S. (1984). Pharmacological Study of Piperine. Arch. Pharm. Res.
7, 127–132.
Leife, A. (1992). Sterilization of spices with pulsed steam. Livsmedelsteknik 34(1/2), 24 –27.
Lewis, Y. S. (1984). “Spices and Herbs for Food Industry”, p. 69–77. Food Trade Press, England.
Lewis, D. G., and Quirk, J. P. (1967). Phosphate diffusion in soil and uptake by plants III 31P-
movement and uptake by plants as indicated by 32P-autoradiography. Plant Soil 26, 445 –453.
Lewis, Y. S., Nambudiri, E. S., and Krishnamurthy, N. (1969). Composition of pepper oil. Perfum.
Essent. Oil Rec. 60, 259 –262.
Lewis, Y. S., Krishnamurthy, N., Nambudiri, E. S., Sankarikutty Amma, B., Shivshankar, S., and
Mathew, A. G. (1976). The need for growing pepper cultivars to suit pepper products.
In “Proceedings of the International Seminar on Pepper”, pp. 4–9, Spices Export Promotion
Council, Cochin, Kerala State, India.
Lin, J. K. (1986). Food-borne amines and amides as potential precursors of endogenous carcinogens.
In “Proceedings of the National Science Council of Republic of China”. (B), 10, pp. 20–34.
Linderman, R. G. (1988). Mycorrhizal interactions with the rhizosphere microflora: The micorrhizo-
sphere effect. Phytopathology 78, 366–370.
Linnaeus, C. (1753). “Species Plantarum”. London.
Lissamma, J., Nazeem, P. A., Mini, S. T., Shaji, P., and Mini, B. (1996). In vitro techniques for mass
multiplication of black pepper (Piper Nigrum L.) and ex vitro performance of the plantlets.
J. Plantation Crops 24 (Suppl.), 511–516.
Liu, G., et al., (1984). Quoted from Ebenhoech and Spadaro (1992).
Lokesh, M. S., and Gangadarappa, P. M. (1995). Management of Phytophthora foot rot and nematode
diseases in black pepper (Piper nigrum L.). J. Spices and Aromatic Crops 4, 61 –63.
Lu, S., and Miller, M. H. (1994). Prediction of phosphorus uptake by field-grown maize with Barber-
Cushman model. Soil Sci. Soc. Am. J. 58, 852 –857.
380 K. P. P. NAIR
Madasamy, M., and Gothandapani, L. (1995). Spiral separator for cleaning. Spice India 8(11),
7–8.
Madhusudhanan, K., and Rahiman, B. A. (1996). In vitro response of Piper species on activated
charcoal supplemented media. In “Biotechnology of Spices, Medicinal and Aromatic Plants”
(S. Edison, K. V Ramana, B. Sasikumar, K. Nirmal Babu, and S. J. Eapen, Eds.), pp. 16 –19.
Indian Soc. for Spices, Calicut, Kerala State, India.
Mahindru, S. N. (1982). “Spices in Indian Life”. Sultan Chand & Sons, New Delhi, India.
Majumdar, A. M., Dhuley, J. N., Deshmukh, V. K., Raman, P. H., Thorat, S. L., and Naik, S. R.
(1990b). Effect of piperine on pentobarbitone induced hypnosis in rats. Indian J. Exp. Biol. 28,
486–487.
Malebennur, N. S., Ganadarappa, P. M., and Hegde, H. G. (1991). Chemical control of foot rot of
black pepper caused by Phytophthora capsici (Phytophthora palmivora MF 4). Indian Cocoa,
Arecanut and Spices J. 14, 148–149.
Mamootty, K. P. (1978). “Quick wilt disease of pepper (Piper nigrum Linn.)-1 Symtomatological
studies on the quick wilt disease of pepper”, p. 87, M. Sc(Ag.) Thesis, Kerala Agricultural
University, Thrissur, Kerala State, India.
Manilal, V. B., and Gopinathan, K. M. (1995). “Annual Report, Regional Research Laboratory,
Council of Scientific and Industrial Research”, pp. 9–10. Trivandrum, Kerala State, India.
Manjunath, A., and Bagyaraj, D. J. (1982). Vesicular arbuscular mycorrhizas in three plantation crops
and cultivars of field bean. Curr. Sci. 51, 707 –709.
Manohara, D., Kasim, R., and Sitepu, D. (1992). Current research status of foot rot disease in
Indonesia. In “Proceedings of the International Workshop on Black Pepper Diseases Bander,
Lampung, Indonesia” (P. Wahid, D. Sitepu, S. Deciyanto, and U. Suparman, Eds.), pp. 144 –154.
Institute for Spice and Medicinal Crops, Bogor, Indonesia.
Mapiliandri, I. (1989). The extraction of black pepper oleoresin. In “Thesis from Diplome. IV.
Ministry of Industry, Indonesia, Ministry of Agriculture (1996)”. Statistik Perkebunan, Jakarta.
Marschner, H. (1994). Micronutr. News Information. 14, 3–5.
Masakazu, M., Isamu, N., and Hirosuki, Y. (1980). The Piperaceae amides Part III. Insecticidal joint
action of pipercide and co-occuring compounds isolated from Piper nigrum. Agri. Biol. Chem. 44,
1071–1073.
Mathew, P. M. (1958). Studies on Piperaceae. J. Indian Bot. Soc. 37, 155–171.
Mathew, P. M. (1972). Karyomorphological studies on Piper nigrum. In “Proceedings of the National
Symposium on Plantation Crops.” J. Plantation Crops, 1, pp. 15–18, (Suppl).
Mathew, A. G. (1992). Chemical constituents of pepper. International Pepper News Bull. 16(2),
18–21.
Mathews, V. H., and Rao, P. S. (1984). In vitro responses in black pepper (Piper nigrum). Curr. Sci.
53, 183 –186.
Mayore, M., Smith, J. L., and Garhaur, D. Y. (1987). Effect of red pepper and black pepper on the
stomach. Am J. Gastroenterol. 82, 211–214.
McCown, H. B., and Amos, R. (1979). Initial trials of micropropagation with birch. Proc. Inter. Plant
Prop. Soc. 29, 387–393.
Mengel, K. (1985). Dynamics and availability of major nutrients in soils. Adv. Soil Sci. 2, 65– 131.
Menon, K. K. (1949). The survey of pollu and root diseases of pepper. Indian J. Agric. Sci. 19,
89–136.
Miquel, F. A. W. (1843). “Systema Piperacearum”. Rotterdam, The Netherlands.
Miyakado, M., Nakayama, I., Yoshoka, H., and Nakatani, N. N. (1979). The Piperaceae amides:
structure of pipercide. Agri. Biol. Chem. 43, 1609–1611.
Mohanakumaran, B., and Abi, Cheeran. (1981). Nutritional requirement of pepper vine trailed on
living and dead standards. Agric. Res. J. Kerala. 19, 3– 5.
Mohandas, C., and Ramana, K. V. (1987). Slow wilt disease of back pepper and its control. Indian
Cocoa, Arecanut and Spices J. 11, 10 –11.
Mohandas, C., and Ramana, K. V. (1991). Pathogenicity of Meloidogyne incognita and Radopholus
similis on black pepper (Piper nigrum L.) J. Plantation Crops. 19, 41 –43.
Montaya, L. A., Sylvain, P. G., and Umana, R. (1961). Effect of light intensity and nitrogen
fertilization upon growth differentiation balance in Coffea arabica L. Coffee (Turrialba) 3,
97 –115.
Mos, A. M. (1995). Microencapsulation enhances flavourings and ingredients. Voedingsmiddelen-
technologie 28(23), 64–65.
Mujumdar, A. M., Dhuley, J. N., Deshmukh, V. K., Raman, P. H., and Naik, S. R. (1990a).
Antiinflammatory activity of piperine. Japn. J. Med. Sci. Biol. 43, 95–100.
Muller, C. J., and Jennings, W. G. (1967). Constituents of black pepper. Some sesquiterpene
hydrocarbons. J. Agric. Food Chem. 15, 762 –766.
Muller, C. J., Creveling, R. K., and Jennings, W. G. (1968). Some minor sesquiterpene hydrocarbons
of black pepper. J. Agric. Food Chem. 16, 113–117.
Murashige, T., and Skoog, F. (1962). A revised medium for rapid growth and bioassays with tobacco
tissue culture. Physiol. Plant 15, 473–497.
Murni, A. M., and Faodji, R. (1990). The combination effect of KCl with two nitrogen sources on the
growth of black pepper. Bull. Penelitian Tanaman Rempah dan Obat. 2(2), 79–84.
Mustika, I. (1990). Studies on the interactions of Meloidogyne incognita, Radopholus similis and
Fusarium solani on black pepper (Piper nigrum L.). p. 127, Ph. D. Thesis, Wageningen
Agricultural University, Wageningen, The Netherlands.
Mustika, I. (1991). Response of four black pepper cultivars to infection by Radopholus similis,
Meloidogyne incognita and Fusarium solani. Industrial Crops Res. J. 4(1), 17–22.
Mustika, I., and Zainuddin, N. (1978). Efficacy tests of some nematicides for the control of nematodes
on black pepper. Pemberitaan L.P.T.I. 30, 1–10.
Nadkarni, K. M. (1976). “Indian Materia Medica”, pp. 971–972. Popular Prakashan, Bombay,
India.
Nair, K. P. P. (1984a). Towards a better approach to soil testing based on the buffer power concept.
In “Proc 6th. Int. Colloquium for the Optimization of Plant Nutrition”, vol 4, 2 –8, September
(Prevel Pierre-Martin, Ed.), pp. 1221–1228. Montpellier, France.
Nair, K. P. P. (1984c). Zinc buffer power as an important criterion for a dependable assessment of
plant uptake. Plant Soil 81, 209–215.
Nair, K. P. P. (1989). Comments on Phosphorus supply characteristics of 33 soils as influenced by
seven rates of phosphorus addition. Soil Sci. Soc. Am. J. 53, 984.
Nair, K. P. P. (1992). Measuring P buffer power to improve routine soil testing for phosphate. Eur.
J. Agron. 1(2), 79–84.
Nair, P. M. (1993). Spice irradiation-Present and future scenario. In “Post Harvest Technology of
Spices, Indian Soc. Spices” (C. S Narayanan, B. Sankarikutty, A. Nirmala Menon, P. N.
Ravindran, and B. Sasikumar, Eds.), pp. 28 –33. Indian Institute of Spices Research, Calicut,
Kerala State, India.
Nair, K. P. P. (1996). The Buffering Power of Plant Nutrients and Effects on Availability. Adv. Agron.
57, 237–287.
Nair, K. P. P. (2002). Sustaining crop production in the developing world through The Nutrient Buffer
Power Concept. In “Proceedings of the 17th World Soil Science Congress”, 14–21 August, vol. 2,
p. 652, Bangkok, Thailand.
Nair, K. P. P. (2003). Measuring Zn buffer power to predict Zn availability. J. Plant Nutr.
(in Press).
Nair, M. K., and Gopalasundaram, P. (1993). Nutrient management in tree based cropping system.
In “Potassium for Plantation Crops” (Mahatim Singh, and M. K. Mishra, Eds.), pp. 200–211.
Potash Research Institute of India, Gurgaon, India.
Nair, K. P. P., and Mengel, K. (1984b). Importance of phosphate buffer power for phosphate uptake by
rye. Soil Sci. Soc. Am. J. 48, 92–95.
382 K. P. P. NAIR
Nair, P. K. U., and Sasikumaran, S. (1991). Effect of some fungicides on quick wilt (foot rot) disease
of black pepper. Indian Cocoa, Arecanut and Spices J. 14, 95 –96.
Nair, M. R. G. K., Das, N. M., and Menon, M. R. (1966). On the occurrence of the burrowing
nematode, Radopholus similis (Cobb 1893) Thorne 1949, on banana in Kerala. Indian J. Ent. 28,
553–554.
Nair, K. P. P., Nand, Ram., and Sharma, P. K. (1984d). Quantitative relationship between zinc
transport and plant uptake. Plant Soil 81, 217–220.
Nair, P. K. U., Mammootty, K. P., Sasikumaran, S., and Pillay, V. S. (1993). Phytophthora foot rot of
black pepper (Piper nigrum L.). A management study with organic amendments. Indian Cocoa,
Arecanut and Spices J. 17, 1–2.
Nair, K. P. P., Sadanandan, A. K., Hamza, S., and Jose, Abraham. (1997). The Importance of
Potassium Buffer Power in the Growth and Yield of Cardamom. J. Plant Nutr. 20(7 & 8),
987–997.
Nakatani, N., Inatani, R., Ohta, H., and Nishioka, A. (1986). Chemical constituents of pepper and
application to food preservation. Naturally occurring antioxidant Environ. Health Perspect. 67,
135–147.
Nambiar, K. K. N. (1978). Diseases of pepper in India. In “Proceedings of the National Seminar on
Pepper” (M. K Nair and M. Haridasan, Eds.), pp. 11–14. Central Plantation Crops Research
Institute, Calicut, Kerala State, India.
Nambiar, K. K. N., and Sarma, Y. R. (1977). Wilt diseases of black pepper. J. Plantation Crops 5,
92–103.
Nambiar, K. K. N., and Sarma, Y. R. (1980). Factors associated with slow wilt of pepper.
In “Proceedings of the PLACROSYM-II 1979.” pp. 348 –358, Indian Society for Plantation
Crops, Kasaragod, Kerala State, India.
Nandakumar, C., Ragunath, P., Visalakshi, A., and Mohandas, N. (1987). Management of pollu beetle
on pepper by adjusting the spray schedules “Abst. Natn. Symp. Integrated Pest Control-Progress
and Perspectives.” p. 50, Trivandrum, Kerala State, India.
Nazeem, P. A., Joseph, L., Thampi, M. S., Sujatha, R., and Nair, G. S. (1993). In vitro culture system
for indirect organogenesis for black pepper (Piper nigrum L.) In “Golden Jubilee Symposium on
Horticultural Research-Changing Scenario,” 24–28 May, p. 250, Bangalore, India.
Ng, S. C. (1993). Uses of pepper. In “The Pepper Industry: Problems and Prospects” (M. Y. Ibrahim,
C. F. J. Bong, and I. B. Ipor, Eds.), pp. 347–353. Universiti Pertanian Malaysia, Sarawak,
Malaysia.
Nigam, M. D., and Handa, K. L. (1964). Constituents of essential oils. Indian Perfumer. 8, 15–17.
Nirmal Babu, K., Naik, V. G., and Ravidran, P. N. (1993a). Two new taxa of Piper (Piperaceae) from
Kerala, India with a note on their origin and inter-relationships. J. Spices and Aromatic Crops 2,
26–33.
Nirmal Babu, K., Rema, J., Lukose, R., Ravindran, P. N., Johnson George, K., Sasikumar, B., and
Peter, K. V. (1993b). “Spices Biotechnology at National Research Centre for Spices”, Technical
Report II, Calicut, Kerala State, India.
Nirmal Babu, K., Geetha, S. P., Manjula, C., Sajina, A., Minoo, D., Samsudeen, K., Ravindran, P. N.,
and Peter, K. V. (1996b). Biotechnology-its role in conservation of genetic resources of
spices. In “Biotechnology for Development.” (M. R. Das and Satish Mundayoor, Eds.),
pp. 198 –212. State Committee on Science, Technology and Environment, Trivandrum, Kerala
State, India.
Nybe, E. V., and Nair, P. C. S. (1986). Nutrient deficiency in black pepper (Piper nigrum L.)
I. Nitrogen, Phosphorus and Potassium. Agric. Res. J. Kerala 24, 132–150.
Nye, P. H. (1979). Diffusion of ions and uncharged solutes in soils and clays. Adv. Agron. 31,
225–272.
Pangburn, R. M., Jennings, W. G., and Noelting, C. F. (1970). Preliminary examination of odour
quality of black pepper oil. Flavour Ind. 1, 763.
Parmar, V. S., Jain, S. C., Bisht, K. S., Jain, R., Taneja, P., Jha, A., Tyagi, O. D., Prasad, A. K.,
Wengel, J., Olsen, C. E., and Boll, P. M. (1977). Phytochemistry of the genus Piper.
Phytochemistry 46, 597–673.
Parry, J. W. (1969). “Spices Vol 1. Chemical Publishing Co. Inc.” New York, USA.
Paulus, A. D., Eng, L., Teo, C. H., and Sim, S. L. (1993). Screening black pepper genotypes and Piper
spp. for resistance to root knot nematode. In “The Black Pepper Industry-Problems and
Prospects” (M. Y. Ibrahim, C. F. J. Bong, and I. P. Ipor, Eds.), pp. 132–139. Universiti Pertanian
Malaysia, Sarawak, Malaysia.
Perez, C., and Anesini (1994). Antibacterial activity of elementary plants against Staphylococcus
aureus growth. American J. Chinese Medicine 22, 169–174.
Peter, K. V. (1996a). Emerging potential for global spices trade. The Hindu Businessline, 21st
December 1996.
Peter, K. V. (1996b). Spices research and development-an updated overview. Employment News 21, 1.
Philip, V. J., Joseph, D., Trigs, G. S., and Dickinson, N. M. (1992). Micropropagation of black pepper
(Piper nigrum L.) through shoot tip cultures. Plant Cell Rep 12, 41– 44.
Pillai, V. S. (1992). Management of improved cultivars of black pepper. In “Black Pepper and
Cardamon: Problems and Prospects” (Y. R. Sharma, S. Devasahayam, and M. Ananadaraj, Eds.),
pp. 17–19. Indian Soc. Spices, Calicut, Kerala State, India.
Pillai, V. S., and Sasikumaran, S. (1976). A note on the chemical composition of pepper plant (Piper
nigrum L.). Arecanut and Spices Bull. 8, 13– 17.
Pillai, V. S., Chandy, K. C., Sasikumaran, S., and Nambiar, P. K. V. (1979). Response of Panniyur-1
variety to nitrogen and lime application. Indian Cocoa, Arecanut and Spices J. 3(2), 35 –38.
Pillai, V. S., Ali, A. B., and Chandy, K. C. (1982). Effect of 3-IBA on root initiation and
development in stem cuttings of pepper (Piper nigrum L.). Indian Cocoa, Arecanut and Spices J.
6(1), 7–9.
Pillai, V. S., Sasikumaran, S., and Nambiar, P. K. V. (1987). N, P and K requirement of black pepper.
Agric. Res. J. Kerala 25(1), 74 –80.
Prabhakaran, P. V. (1994). Assessment of losses due to pests, diseases and drought in black pepper.
J. Indian Soc. Agric. Stat. 47, 59–60.
Pradeep, K. U., and Geervani, P. (1994). Influence of spices on protein utilisation of winged bean and
horsegram. Plant Foods Hum. Nut. 46, 187–193.
Premkumar, T., and Nair, M. R. G. K. (1987a). Reduction in damage by Longitarsus nigripennis to
berries of black pepper by insecticides. Madras Agric. J. 74, 326 –327.
Premkumar, T., and Nair, M. R. G. K. (1988). Effect of some planting conditions on infestation
of black pepper by Longitarsus nigripennis Mots. Indian Cocoa, Arecanut and Spices J. 10,
83 –84.
Premkumar, T., Banerjee, S. K., Devasahayam, S., and Koya, K. M. A. (1986). Effect of different
insecticides on the control of pollu beetle Longitarsus nigripennis Mots., a major pest of black
pepper Piper nigrum L. Entomon. 11, 219– 221.
Premkumar, T., Devasahayam, S., and Koya, K. M. A. (1994). Pests on spice crops. In “Advances in
Horticulture-Plantation and Spice Crops Part 2” (K. L. Chadha and P. Rethinam, Eds.), vol. 10,
pp. 787–823. Malhotra Publishing House, New Delhi, India.
Pruthi, J. S. (1992). Advances in sun/solar drying and dehydration of pepper (Piper nigrum L.).
International Pepper News Bull. 16(2), 6–17.
Pruthi, J. S. (1993). “Major Spices of India: Crop Management and Post-Harvest Technology”,
pp. 44–105. Indian Council of Agricultural Research, New Dehhi, India.
Pruthi, J. S. (1997). Diversification in pepper utilization. Part I & II. Green pepper products.
International Pepper News Bull. 14(4), 5–9; 14(5), 6–9.
Pruthi, J. S., Bhat, A. V., Kutty, S. K., Varkey, A. G., and Gopalakrishnan, M. (1976). Preservation of
fresh green pepper by canning, bottling and other methods. In “International Seminar on Pepper”,
pp. 64–72. Spices Export Promotion Council of India, Cochin, Kerala State, India.
384 K. P. P. NAIR
Purseglove, J. W., Brown, E. G., Green, C. L., and Robbins, S. R. J. (1981). “Spices”, vol. 1, pp. 1 –99.
Longman, London.
Rahiman, B. A. (1981a). Biosystematic studies in varieties and spices of Piper occurring in Karnataka
region. Ph. D. Thesis, University of Mysore, Karnataka State, India.
Raj, H. G. (1978). “A comparison of the systems of cultivation of Black pepper-Piper nigrum L. in
Malaysia and Indonesia”, pp. 65–75. The Directorate of Extesion Education, Kerala Agricultural
University, Trichur, Kerala State, India.
Raj Mohan, K. (1985). “Standardization of tissue culture techniques in important horticultural crops”.
Ph. D. Thesis, Kerala Agricultural University, Thrissur, Kerala State, India.
Ramachandran, N. (1990). “Bioefficacy of systemic fungicides against Phytophthora infections in black
pepper (Piper nigrum L.)”, p. 138, Ph.D Thesis, University of Calicut, Calicut, Kerala State, India.
Ramachandran, N., and Sarma, Y. R. (1985). Efficacy of three systemic fungicides in controlling
Phytophthora infections in black pepper. Indian Phytopath. 38, 160–162.
Ramachandran, N., Sarma, Y. R., and Nambiar, K. K. N. (1986). Spatial distribution of Phytophthora
palmivora MF4 in the root zones of Piper nigrum. Indian Phytopath. 39, 414 –417.
Ramachandran, N., Sarma, Y. R., and Anandaraj, M. (1988b). Sensitivity of Phytophthora species
affecting different plantation crops in Kerala to Metalaxyl. Indian Phytopath. 41, 438 –442.
Ramachandran, N., Sarma, Y. R., Anandaraj, M., and Abraham, J. (1988c). Effect of climatic factors
on Phytophthora leaf infection in black pepper grown in arecanut–black pepper mixed cropping
system. J. Plantation Crops 16, 110–118.
Ramachandran, N., Sarma, Y. R., and Anandaraj, M. (1990). Vertical progression and spread of
Phytophthora leaf infection in black pepper in arecanut–black pepper mixed cropping system.
Indian Phytopath. 43, 414 –419.
Ramachandran, N., Sarma, Y. R., and Anandaraj, M. (1991). Management of Phytophthora infections
in black pepper. In “Diseases of Black Pepper” (Y. R. Sarma and T. Premkumar, Eds.),
pp. 158 –174. National Research Centre for Spices, Calicut, Kerala State, India.
Ramadasan, A. (1987). Canopy development and yield of adult pepper vines in relation to light
interception. Indian Cocoa, Arecanut and Spices J. 11, 43–44.
Ramana, K. V. (1992). “Final Report of the project: Role of Nematodes in the incidence of Slow
Decline (Slow Wilt Disease) of Black Pepper and Screening Pepper Germplasm against
Nematodes”, p. 149. National Research Centre for Spices, Calicut, Kerala State, India.
Ramana, K. V. (1994). Efficacy of Paecilomyces lilacinus (Thom). Samson in suppressing nematode
infestations in black pepper (Piper nigrum L.). J. Spices and Aromatic Crops 3, 130 –134.
Ramana, K. V., and Eapen, S. J. (1998). Plant parasitic nematodes associated with spices and
condiments. In “Nematode Diseases in Plants” (P. C. Trivedi, Ed.), pp. 217 –251. C. B. S.
Publishers and Distributors, New Delhi, India.
Ramana, K. V., and Mohandas, C. (1986). Reaction of black pepper germplasm to root knot nematode
Meloidogyne incognita. Indian J. Nematol. 16, 138–139.
Ramana, K. V., Mohandas, C., and Ravindran, P. N. (1987b). Reaction of black pepper germplasm to
the burrowing nematode (Radopholus similis). J. Plantation Crops. 15, 65–66.
Ramana, K. V., Mohandas, C., and Eapen, S. J. (1994). “Plant Parasitic Nematodes and Slow Decline
Disease of Black Pepper”, p. 14. Tech. Bull. National Research Centre for Spices, Calicut, Kerala
State, India.
Ramanathan, P. K., and Rao, P. N. S. (1974). Equipment needs of spice industry. In “Proceedings of
the Symposium on Development and Prospects of Spice Industry in India”, pp. 82–84. Central
Food Technology Research Institute, Mysore, Karnataka State, India.
Ramesh, C. R. (1982). Root infection and population density of VA mycorrhizal fungi in a coconut
based multi-storied cropping. In “Proceedings of the Plantation Crops Symposium (PLACRO-
SYM V)” (K. V. A. Bavappa, R. D. Iyer, P. K. Das, M. R. Sethuraj, V. Ranganathan, K. K. N.
Nambiar, and M. J. Chacko, Eds.), pp. 548 –554. Central Plantation Crops Research Institute,
Kasaragod, Kerala State, India.
Rao, C. S. S., and Nigam, S. S. (1976). Antimicrobial acitivity of some Indian essential oils. Indian
Drugs 14, 62.
Ravindran, P. N. (1991). “Studies on Black Pepper and Some of its Wild Relatives”. Ph. D. Thesis
University of Calicut, Calicut, Kerala State, India.
Ravindran, P. N., and Nirmal Babu, K. (1994a). Genetic Resources of Black Pepper. In “Advances in
Horticulture” (K. L. Chadha and P. Rethinam, Eds.), pp. 99–120. Malhotra Publications Co.,
New Delhi, India.
Ravindran, P. N., and Peter, K. V. (1995). Biodiversity of major spices and their conservation in India.
In “Proceedings of the South Asia Coordinators Meeting on Plant Genetic Resources. IPGRI”
(R. K. Arora and V. R. Rao, Eds.), pp. 123–134. New Delhi, India.
Ravindran, P. N., Nair, M. K., and Nair, R. A. (1987). Two taxa of Piper (Piperaceae) from Silent
Valley Forests, Kerala. J. F. Eco. Tax. Bot. 10, 167 –169.
Ravindran, P. N., Nair, M. K., Nair, R. A., Nirmal Babu, K., and Chandran, K. (1990). Ecological and
taxonomical notes on Piper spp. from Silent Valley Forests, Kerala. J. Bombay Nat. His. Soc. 87,
421 –426.
Reddy, B. N., Sadanandan, A. K., Sivaraman, K., and Abraham, J. (1992). Effects of planting
density on black pepper on yield and nutrient availability in soil. J. Plantation Crops 20, 10–13
(Suppl.).
Reen, R. K., Jamwal, D. S., Taneja, S. C., Koul, J. L., Dubey, P. K., Wiebel, F. J., and Singh, J. (1993).
Impairment of UDP-GDH and glucuronidation activities in liver and small intestine of rat and
guinea pig in vitro by piperine. Biochem. Pharmacy 46, 229–238.
Rehiman, P. V. A., and Nambiar, P. K. V. (1967). Insecticidal control of pollu beetle (Longitarsus
nigripennis) on pepper. Madras Agric. J. 54, 39 –40.
Revankar, G. D., and Sen, D. P. (1974). Antioxidant effects of a spice mixture on sardine oil. J. Food
Sci. Technol (India) 11, 31– 32.
Rheede, H. Van (1678). “Hortus Indicus Malabaricus”, vol. 7, pp. 23–31. Amstelodami, pl. 12–16.
Richard, H. M., and Jennings, W. G. (1971). Volatile composition of black pepper. J. Food Sci. 36,
584 –589.
Richard, H. M., Russel, G. F., and Jennings, W. G. (1971). The volatile components of black pepper
varieties. J. Chromatogr. Sci. 9, 560 –566.
Risfaheri Hidayat, T. (1996). Study on decorticating of pepper berries by soaking in boiling water
method. International Pepper News Bull 20(1), 17–20.
Rosengarten, F. Jr. (1973). The Book of Spices. Pyramid Books, New York, U.S. A.
Ruiz, Pavon. (1794). Cited by Trelease and Yuncker (1950).
Russel, G. F., and Else, J. (1973). Volatile compositional differences between cultivars of black pepper
(Piper nigrum). J. Assoc. off. Anal. Chem. 56, 344– 351.
Sadanandan, A. K. (1990). “Response of pepper to applied potassium (Abs.)”, p. 149. National
Seminar on Potassium on Plantation Crops, Bangalore, India.
Sadanandan, A. K. (1994). Nutrition of black pepper. In “Advances in Horticulture. Vol. 9-Plantation
and Spice Crops (Part I).” (K. L. Chadha and P. Rethinam, Eds.), pp. 423 –456.
Sadanandan, A. K., and Hamza, S. (1996). Studies on nutritional requirement of bush pepper for yield
and quality (Abs.). In “Proceedings of the PLACROSYM-XII, RRII”, p. 48, Kottayam, Kerala
State, India.
Sadanandan, A. K., Abraham, J., and Anandaraj, M. (1992). Impact of high production technology on
productivity and disease incidence of black pepper. In “Proceedings of the PLACROSYM-IX”
(K. K. N. Nambiar, R. D. Iyer, and E. V. V. B. Rao, Eds.), 20, pp. 384–387. Indian Soc.
Plantation Crops, Kasaragod, Kerala State, India.
Saito, Y., and Asari, T. (1976). Studies on the antioxidant properties of spices. I. Total tocopherol
content in spices. J. Japanese. Soc. Food. Nutr. 29, 289–292.
Sajina, A., Minoo, D., Geetha, S. P., Samsudeen, K., Rema, J., Nirmal Babu, K., Ravindran, P. N., and
Peter, K. V. (1996). Production of synthetic seeds in spices. In “Biotechnology of Spices,
386 K. P. P. NAIR
Medicinal and Aromatic Plants” (S. Edison, K. V. Ramana, B. Sasikumar, K. Nirmal Babu, and
S. J. Eapen, Eds.), pp. 65–69. Indian Soc. for Spices, Calicut, Kerala State, India.
Salzer, U. J. (1975). Uber die Fettsaurezussamensetzung de lipoide einiger gewurze. Fette Stefen
Anstrichm. 77, 446 –450.
Salzer, U. J., Broker, U., Klie, H. P., and Liepe, H. U. (1977). Effect of pepper and pepper constituents
on the microflora of sausage products. Fleischwirtschafe 57, 2011–2121.
Samraj, J., and Jose, P. C. (1966). A Phytophthora wilt of Pepper (Piper nigrum). Sci Cult. 32, 90 –92.
Samuel, R., Prabhu, V. K. K., and Narayanan, C. S. (1984). Influence of spice essential oils on the life
history of Lasioderma seriorne. Entomon. 9, 209.
Sanchez, P. A. (July 10–16, 1994). Tropical soil fertility research: Towards the second paradigm.
In “Proceedings of the XV ISSS Symposium”, vol. 1, pp. 89–104. Acapulco, Mexico.
Sankarikutty, B., Padmakumari, K. P., Howa, Umma., and Narayanan, C. S. (1994). pp. 7–8. Annual
Report, Regional Research Laboratory, Council of Scientific and Industrial Research,
Trivandrum, Kerala State, India.
Santosh, P. (1985). “Cost of cultivation and marketing of pepper in Cannanore District”. M. Sc. (Ag)
Thesis, Kerala Agricultural University, Thrissur, Kerala State, India.
Sarkar, B. B., Choudhary, D. M., Nath, S. C., Bhattacharjee, N. R., and Barna, D. T. C. (1985). Wilt
disease of pepper and its control in Tripura. Indian Cocoa, Arecanut and Spices J. 8, 63–66.
Sarma, Y. R., Anandaraj, M., and Ramana, K. V. (1992c). Present status of black pepper diseases in
India and their management. In “Proceedings of the International Workshop on Black Pepper
Diseases Bander Lampung, Indonessia” (P. Wahid, D. Sitepu, S. Deciyanto, and U. Suparman,
Eds.), pp. 67–68. Institute for Spice and Medicinal Crops, Bogor, Indonesia.
Sarma, Y. R., Anandaraj, M., and Venugopal, M. N. (1996). Biological control of diseases in spices.
In “Biological Control in Spices” (M. Anandaraj and K. V. Peter, Eds.), pp. 1– 19. Indian Institute
of Spices Research, Calicut, Kerala State, India.
Sasikumar, B., and Veluthambi, K. (1994). Kanamycin sensitivity of cultured tissues of Piper
nigrum L. J. Spices and Aromatic Crops 3, 158–160.
Satheesan, K. N., Rajagopalan, A., Sukumarapillai, V., Jagadeesh Kumar, T. N., Mammooty, K. P.,
and Unnikrishnan Nair, P. K. (1997a). Influence of irrigation levels on yield, quality and
incidence of pests and diseases in black pepper (Piper nigrum L.) (Abs.). In “National Seminar on
Water Management for Sustainable Production and Quality of Spices”, p. 21, Madikheri,
Karnataka State, India.
Schenk, R. U., and Hildebrandt, A. C. (1972). Medium and techniques for induction and growth of
monocotyledonous and dicotyledonous plant cell cultures. Can. J. Bot. 50, 199– 204.
Schneider, B. (1993). Steam sterilization of spices. Fleischwirtschaft 73(6), 646–649.
Schüller, H. (1969). Die CAL-Methode, eine neue Methode zur Bestimmung des pflanzen verfügbaren
Phosphates in Böden. Z. Pflanzenernähr Bodenkd. 123, 48 –63.
Scott, R. (1992). Master spice: Extruded low count spices. In “Proceedings of the New Technologies
Symposium Bristol”, pp. 6– 8, U. K.
Scott, W. P., and McKibben, G. H. (1978). Toxicity of black pepper extract to ball weevil. J.Eco. Ent.
71, 343 –344.
Selim, H. M. (1992). Modeling the transport and retention of inorganics in soils. Adv. Agron. 47,
331–384.
Sen, A. R., and Roy, B. R. (1974). Adulteration in spices. In “Proceedings of the National Symposium
on Development and Prospects of Spice Industry in India”. Central Food Technology Research
Institute, Mysore, Karnataka State, India.
Sethuram, P. (1995). Cochin Pepper market-a journey through history. Ind. Spices. 32, 3.
Sewell, G. W. F. (1965). The effect of altered physical conditions of soil on biological control.
In “Ecology and Soil Borne Plant Pathogens” (K. F. Baker and W. C. Snyder, Eds.), pp. 479– 494.
University of California Press, Berkeley, U. S. A.
Sharma, R. D., and Loof, P. A. A. (1974). Nematodes of cocoa region of Bahia, Brazil IV. Nematodes
in the rhizosphere of pepper (Piper nigrum L.) and clove (Eugenia caryophylla Thumb). Revista
Theobroma. 4, 26 –32.
Sharma, N. L., Nigam, N. C., and Handa, K. L. (1962). Pepper husk in oil. Perfuem. Kosmet. 43,
505 –506.
Sheela, M. S., Venkitesan, T. S., and Mohandas, N. (1993). Status of Bacillus spp. as biocontrol agents
of root knot nematode (Meloidogyne incognita) infesting black pepper (Piper nigrum L.).
J. Plantation Crops 21, 218– 222 (Suppl.).
Sher, S. A., Chunram, C., and Pholcharoen, S. (1969). Pepper yellows disease and nematodes in
Thailand. FAO Plant Prot. Bull. 17, 33.
Shin, K. H., Yun, H. S., Won, W. S., and Lee, C. K. (1980). Pharmacological activity of Piper
retrofractrum. Soul. Tacha. Sean. Young Opi. 18, 87 –89. (Chem. Abstr. 93, 215484s).
Shuman, L. M. (1975). The effect of soil properties on zinc adsorption by soils. Soil Sci. Soc. Am.
Proc. 39, 454– 458.
Shwaireb, M. H., Waba, H., El Mofty, M. M., and Dutter, A. (1990). Carcinogenesis induced by black
pepper and modulated by vitamin-A. Exp. Pathol. 40, 233–238.
Sim, E. S. (1971). Dry matter production and major nutrient contents of black pepper (Piper nigrum
L.). MalayAgric. J. 48, 73.
Sim, S. L., Jafar, R., Grierson, D., Power, J. B., and Davery, M. R. (1995). Applications of molecular
biology and genetic manipulation to pepper (Piper nigrum L.) breeding: present state and
prospects In “Abstracts FAO/IAEA International Symposium on the Use of Induced Mutations
and Molecular Techniques for Crop Improvement”. Vienna, Austria.
Singh, A., and Rao, A. R. (1993). Evaluation of the modulatory influence of black pepper on the
hepatic detoxification system. Cancer Lett. 72, 5– 9.
Sitepu, D., and Kasim, R. (1991). Black pepper diseases in Indonesia and their control strategy.
In “Diseases of Black Pepper” (Y. R. Sarma and T. Premkumar, Eds.), pp. 13–28. National
Research Centre for Spices, Calicut, Kerala State, India.
Siti Hajijah, A. S. (1993). Observations of root knot infestation on pepper (Piper nigrum L.) in
Sarawak. In “The Pepper Industry-Problems and Prospects” (M. Y. Ibrahim, C. F. J. Bong, and
I. B. Ipor, Eds.), pp. 140–147. Universiti Pertanian Malaysia, Sarawak, Malaysia.
Siti Hajijah, A. S., and Bong, C. F. J. (1993). Pepper fruit stalk as substrate for mushroom cultivation.
In “The Pepper Industry: Problems and Prospects” (M. Y. Ibrahim, C. F. J. Bong, and I. B. Ipor,
Eds.), pp. 342– 346. Universiti, Pertanian Malaysia, Sarawak, Malaysia.
Sivadasan, C. R. (1996). Import regulations in USA. Spice India 9, 2–10.
Sivakumar, C., and Wahid, P. A. (1994). Effect of application of organic materials on growth and
foliar nutrient contents of black pepper (Piper nigrum L.). J. Spices and Aromatic Crops 3,
135 –141.
Soil Science Society of America, (1965). Glossary of soil science terms. Soil Sci. Soc. Am. Proc. 29,
330 –351 (Committee Report).
Sosamma, V. K., and Koshy, P. K. (1995). Effect of Pasteuria penetrans and Paecilomyces lilacinus
on population build up of root knot nematode, Meloidogyne incognita on black pepper. Indian
J. Nematol. 25, 16–17 (Abst.).
Sparks, D. L. (1987). Potassium dynamics in soils. Adv. Soil Sci. 6, 2–63.
Sparks, D. L. (1989). “Kinetics of Soil Chemical Processes”. Academic Press, San Diego, CA.
Sreeja, T. P., Eapen, S. J., and Ramana, K. V. (1996). Occurrence of Verticillium chlamydosporium
Goddard in a black pepper (Piper nigrum L.) garden in Kerala, India. J. Spices and Aromatic
Crops 5, 143–147.
Sreekumar, M. M., Sankarikutty, B., Nirmala Menon, A., Padmakumari, K. P., and Narayanan, C. S.
(1993). pp. 5–6. Annual Report. Regional Research Laboratory, Council of Scientific and
Industrial Research, Trivandrum, Kerala State, India.
388 K. P. P. NAIR
Sridharan, K., Kalla, A. K., and Singh, J. (1978). Chemical and pharmacological screening of Piper
nigrum L. leaves. J. Res. Indian Med. Yoga and Homeop. 13, 107.
Stirling, G. R. (1991). In “Biological Control of Plant Parasitic Nematodes”, p. 282. C. A. B.
International, Wallingford, U. K.
Su, H. C. F. (1977). Insecticidal properties of black pepper on cotton weevil and cow-pea weevil.
J. Eco. Ent. 70, 18 –21.
Subramanyam, V., Sreenivasamoorthy, N., Krishnamoorthy, K., and Swaminathan, M. (1957).
Studies on the antibacterial effects of spices. J. Scien. Ind. Res. 16, 240.
Sundararaju, P., Koshy, P. K., and Sosamma, V. K. (1979). Plant parasitic nematodes associated with
spices. J. Plantation Crops 7, 15–26.
Sundararaju, P., Koshy, P. K., and Sosamma, V. K. (1980). Survey of nematodes associated with
spices in Kerala and Karnataka. In “Proceedings of the PLACROSYM-II, 1979”, pp. 39 –44.
Indian Society for Plantation Crops, Kasaragod, Kerala State, India.
Suparman, U., and Zaubin, R. (1988). Effect of defoliation, IBA and Saccharose on root growth of
black pepper cuttings. Industrial Crop Res. J. 1, 54 –58.
Tantaoui, E. A., and Beraoud, L. (1994). Inhibition of growth and aflatoxin production in Aspergillus
parasiticus by essential oils. J. Environ. Toxicol. Pathol. Oncol. 13, 67 –72.
Thomas, K. M., and Menon, K. K. (1939). The present position of pollu disease of pepper in Malabar.
Madras Agric. J. 17, 347 –356.
Tien, Ju-K’ang (1981). China and the pepper trade. Hemisphere, pp. 220 –222.
Trelease, W., and Yuncker, T. G. (1950). “The Piperaceae of Northern South America”. University of
Illinois, U. S. A.
Tripathi, S. N., Tiwari, C. N., Upadhyaya, B. N., and Singh, K. R. S. (1979). Screening of
hypoglycaemic action in certain indigenous drugs. J. Res. Ind. Med. Yoga and Homeop. 14, 159.
Uijil, C., and den., H. (1992). New continuous steaming method for herbs and spices. Voedings-
middelentechnologie 25(6), 40–43.
Uma Pradeep, K., Geervani, P., and Eggum, B. O. (1993). Common Indian Spices: nutrient
composition, consumption and contribution to dietary values. Plant Foods Hum. Nutr. 44,
137–148.
Ummer, C. (1989). Indian Spices-from the leaves of history. In “Spice Fair Commemorative
Volume”, pp. 27–40.
Van der Vecht, J. (1950). Plant parasitic nematodes. In “Diseases of Cultivated Plants in Indonesian
Colonies” (L. G. E. Karshoven and J. van der Vecht, Eds.), pp. 16–45. I. S’ gravenhage, W. van
Woeve.
Variyar, P. S., Pendharkar, M. B., Banerjee, A., and Bandopadhyaya, C. (1988). Blackening in green
pepper berries. Phytochemistry 27, 715–717.
Varughese, J., and Anuar, M. A. (1992). Etiology and control of slow wilt disease in Johore, Malaysia.
In “Proceedings of the International Workshop on Black Pepper Diseases” (P. Wahid, D. Sitepu,
S. Deciyanto, and U. Suparman, Eds.), pp. 188–197. Research Institute for Spice and Medicinal
Crops, Bogor, Indonesia.
Velayudhan, K. C., and Amalraj, V. A. (1992). Piper pseudonigrum — a new species from Western
Ghats. J. Eco. Tax. Bot. 16, 247 –250.
Venkitesan, T. S., and Charles, J. S. (1980). A note on the chemical control of nematodes infesting
pepper vines in Kerala. In “Proceedings of the PLACROSYM-II 1979”, pp. 27–30. Indian
Society for Plantation Crops, Kasaragod, Kerala State, India.
Venkitesan, T. S., and Setty, K. G. H. (1978). Reaction of 27 black pepper cultivars and wild forms to
the burrowing nematode Radopholus similis(Cobb) Thorne. J. Plantation Crops 6, 81 –84.
Venkitesan, T. S., and Setty, K. G. H. (1979). Control of the burrowing nematode, Radopholus similis
on black pepper. Pesticides 13, 40–42.
Vezyuez Olivincia, W., Shah, P., and Pitchumoni, C. S. (1992). The effect of red and black pepper on
orocecal transit time. J. M. College Nutr. 11, 228–231.
Vijayakumar, K. R., Unni, P. N., and Vamadevan, V. K. (1984). Prevention of photo-induced

chlorophyll loss by the use of lime reflectant on the leaves of black pepper (Piper nigrum L.).
Agric. For. Meteorol. 34, 17–20.
Vinod, G. (1984). “Cost of cultivation and marketing of pepper in Idukki District”. M. Sc.(Ag) Thesis.
Kerala Agricultural University, Thrissur, Kerala State, India.
Wahid, P. (1976). Studies on yellows disease in black pepper on the island of Bangka. Pembr. LPTI
21, 64–79.
Wealth of India, (1969). Piper nigrum Linn (Piperaceae). In “Wealth of India-Raw Materials”, vol. 8,
pp. 99 –115. Publications and Information Directorate, Council of Scientific and Industrial
Research, New Delhi.
Wilkinson, H. F., Loneragan, J. F., and Quirk, J. P. (1968). The movement of zinc to plant roots. Soil
Sci. Soc. Am. Proc. 32, 831–833.
Winoto, S. R. (1972). Effect of Meloidogyne species on the growth of Piper nigrum L. Malaysian
Agric. Res. 1, 86–90. Seoul Teahakkyo Saengyak Yonguso Opidukdip.
Won, W. S., Lee, E. B., and Shin, K. H. (1979). CNS depressant activity of piperine. 18, 66–70.
(Chem. Ab. 93, 215482q).
Wrolstad, R. E., and Jennings, W. G. (1965). Volatile constituents of black pepper III. The
monoterpene hydrocarbon fraction. J. Food Sci. 30, 274–279.
Wuba, H., El-Mofty, M. M., Schweureb., M. H., and Dutter, A. (1992). Carcinogenicity testing of
some constituents of black pepper. Exp. Pathol. 44, 61– 65.
Yuncker, T. G. (1958). The Piperaceae: A family profile. Brittonia 10, 1 –17.
Zadocks, J. C., and Van den Bosch, F. (1994). On the spread of plant disease: A theory of foci. Ann.
Rev. Phytopath. 32, 503 –521.
Zaubin, R., and Robbert (1979). Pengaruh keasaman tanah terhadap pertumbuhan tanaman lada. Pebr.
Littri. 33, 1987.
This page is intentionally left blank
SOIL MINERAL – ORGANIC
MATTER – MICROORGANISM
INTERACTIONS: FUNDAMENTALS
AND IMPACTS
P. M. Huang
Department of Soil Science, University of Saskatchewan, 51 Campus Drive,
Saskatoon, SK, Canada S7N 5A8
I. Introduction
II. Mechanisms of Binding of Nonhumic Organics and Humic
Substances by Soil Mineral Colloids
A. Nature of Mineral Colloid Surfaces
B. Binding of Nonhumic Organics
C. Binding of Humic Substances
III. Influence of Organic Substances on the Formation, Transform-
ation, and Surface Properties of Metal Oxides
A. Aluminum Oxides
B. Iron Oxides
IV. Role of Soil Minerals in Abiotic Catalysis of the Formation of
Humic Substances
A. Polyphenol Pathway
B. Maillard Reaction
V. Interactions of Soil Mineral Colloids with Microorganisms
A. Surface Interactions
B. Influence on Microbial Activity and Survival
VI. Interactions of Soil Colloids with Enzymes
A. Mineral Colloid – and Humic– enzyme Complexes
B. Effects on Enzymatic Activity
VII. Microbial Mediation of Soil Mineral Weathering and
Transformation
A. Mineral Weathering
B. Fine-grained Mineral Development
VIII. Interactions of Soil Mineral Colloids with Organic Substances
and Microorganisms in Relation to Soil Structure Stability
A. Organo-mineral Complexes in Relation to Soil Structure
B. Dynamics of Aggregate Turnover
IX. Influence of Mineral Colloids on Biogeochemical Cycling of C,
N, P, and S in Soil
A. Role of Mineral Colloids in Soil Organic Matter Storage
and Turnover
B. Decomposition and Stabilization of C, N, P, and S in
Relation to Primary and Secondary Particles
391
0065-2113/03 $35.00
392 P. M. HUANG
X. Effects of Interactions Between Microorganisms and Soil

Colloids on the Transformation of Organic Pollutants
A. Catalytic Transformations of Organic Pollutants
B. Binding of Organic Pollutants, Enzymes, and Microorgan-
isms on Mineral and Humic Surfaces and the Effects on
Pollutant Bioavailability
XI. Impact of Interactions of Physicochemical, Biochemical, and
Biological Processes on Metal Transformation
A. Redox Reactions of Metals
B. Complexation Reactions of Metals
C. Adsorption – desorption Reactions of Metals
D. Precipitation – dissolution Reactions of Metals
E. Metal Sorption and Uptake by Microorganisms and
Biomineralization
F. Metal Transformation by Microbial Excretions and
Mycorrhizal Infection
XII. Foreseeable Impacts of Soil Mineral –organic Matter – micro-
organism Interactions on the Ecosystem
A. Global Ion Cycling
B. Global Climate Change
C. Biodiversity
D. Biological Productivity and Human Nutrition
E. Geomedicine
F. Ecotoxicology and Human Health
G. Biotechnology
H. Risk Assessment
I. Risk Management, Remediation, and Restoration
XIII. Summary and Conclusions
Acknowledgment
References
Minerals, organic matter, and microorganisms are integral parts of the

pedosphere and related environments. These three components are not
separate entities but rather a united system constantly in association and
interactions with each other in the terrestrial environment. Interactions of
these components mediated by soil solution and atmosphere govern
mechanisms of mineral weathering reactions, formation of short-range
ordered (SRO) metal oxides, abiotic catalysis of the formation of humic
substances, formation of organo-mineral complexes, microbial ecology,
enzymatic activity, soil structure stability, dynamics of aggregate turnover,
biogeochemical cycling of C, N, P, and S, and transformations and dynamics
of metals and organic pollutants in the terrestrial environment. Foreseeable
impacts of these interactions include global ion cycling and climate change,
biodiversity, biological productivity and human nutrition, geomedicine,
ecotoxicology and human health, biotechnology in relation to food
production and security, risk assessment, and ecosystem restoration.
Therefore, soil mineral – organic matter –microorganism interactions play a
key role in influencing agricultural sustainability and ecosystem health.
Fundamental understanding of these interactions at the molecular level is
SOIL MINERAL – ORGANIC MATTER– MICROORGANISM 393
essential for developing innovative management strategies for land and water
resources. q 2004 Academic Press.
I. INTRODUCTION
Soil is an integral compartment of the environment and the central organizer

of the terrestrial ecosystem. Minerals, organic components, and microorganisms
are three major solid components of the soil. They profoundly affect the physical,
chemical, and biological properties and processes of terrestrial systems. To date,
scientific accomplishments in individual disciplines of the chemistry of soil
minerals, soil organic matter, and soil microbiology are commendable. How-
ever, minerals, organic matter, and microorganisms should not be considered as
separate entities but rather a united system constantly in close association and
interactions with each other in soil environments. Interactions of these com-
ponents have enormous impact on terrestrial processes critical to environmental
quality and ecosystem health.
In view of the significance of soil mineral – organic matter –microorganism
interactions, the International Society of Soil Science (ISSS) established the
Working Group MO “Interactions of Soil Minerals with Organic Components
and Microorganisms” in 1990. The objective of this Working Group is to promote
research and education on the interactions of these major solid components of
soil and the impact on the production of foodstuffs and fibers, the sustainability of
the environment, and ecosystem health including human health on the global
scale. Since its establishment, the Working Group has contributed substantially to
advance the knowledge on physical/chemical/biological interfacial interactions
in soil environments.
The study of the interactions of these soil components has to be considered
from the molecular level to field/landscape systems and is, indeed, essential in
stimulating further research to uncover the dynamics and mechanisms of
soil processes. Major research thrusts of these interactive soil processes include:
(1) mineral and biological catalysis and enzyme – mineral interactions leading to
humus and organo-mineral complex formation, (2) surface reactions of micro-
and macrobiota and biomolecules with soil particles, (3) the effect of soil
interactive processes on the structure, dynamics, and activities of microbial
communities, and (4) the impact of soil interactive processes on porosity
formation by structure or organization development and on biogeochemical
transformation and transport of chemical and biological components at different
spatial scales, and on ecosystem integrity.
394 P. M. HUANG
This paper addresses a variety of issues on fundamentals of soil mineral –

organic component – microorganism interactions at the molecular and micro-
scopic levels and the impact on the terrestrial ecosystem.
II. MECHANISMS OF BINDING OF NONHUMIC

ORGANICS AND HUMIC SUBSTANCES
BY SOIL MINERAL COLLOIDS
A. NATURE OF MINERAL COLLOID SURFACES
The surface chemistry of soils is dependent on weathering processes of the

parent materials that are initiated by geological, hydrological, and biological
agents. Therefore, soil formation processes have tremendous impact on the
formation and modification of surface functional groups of soils.
The siloxane surface of 2:1 expansible layer silicates is negatively charged.
The ditrigonal cavities of the surface can thus adsorb positively charged ions and
polymers to form hydroxy-interlayered minerals in soil weathering processes, as
illustrated in Fig. 1. These hydroxy interlayered minerals occur in soils
throughout the world and are present in soils of several orders of soil taxonomy
systems (Rich, 1968; Barnhisel and Bertsch, 1989). In addition to the formation
of hydroxy-interlayered minerals, positively charged hydroxy polymers can be
adsorbed on the edges and external surfaces of phyllosilicates (Huang and Kozak
1970) as illustrated in Fig. 1.
The surfaces of soil mineral colloids can also complex with organics
(Mortland, 1986; Schnitzer, 1986). It is generally accepted that organics can form
inner-sphere and/or outer-sphere complexes with surface functional groups of
minerals. Stereochemistry has a much more important role in surface
complexation with organics than in the formation of surface complexes with
inorganic polymers (Sposito, 1984). The surface complexation reactions may
result in the development of thick envelopes of colloidal organics on metal oxides
and aluminosilicates.
The complexity of the surface properties of soil mineral colloids is further
manifested by the formation of iron oxyhydroxide or calcium carbonate coatings
on the external surfaces of layer silicates, the growth of two-dimensional
solid solutions of trace metal oxides on the periphery of Mn oxide nodules
(Sposito, 1984), the coating of calcium carbonate on Mn oxide (Oscarson et al.,
1983), and the coating of smectite by hydroxy-Al-organic complexes (Goh and
Huang, 1986). Moreover, similar effects can occur on the surfaces of other
minerals in soils (Johnston and Tombacz, 2002). Therefore, the surface chemistry
of soil minerals is mainly that of the functional groups exposed on the coatings
rather than that of the groups in the underlying matrix. These surface functional
Figure 1 Structural models depicting the hydroxy-Al coatings on the interlamellar and external
planar surfaces and edges of mica-vermiculite. From Huang (1980).
groups have a very significant influence on the transformation of organics in soil

environments.
B. BINDING OF NONHUMIC ORGANICS
Nonhumic organic molecules interact with the surfaces of mineral colloids by

a variety of mechanisms as discussed below. The reaction mechanisms mainly
depend on the intrinsic nature and properties of the organic species, the kind of
exchangeable cation on the surface of the mineral, the water content of the
system, and the properties of the mineral colloids (Mortland, 1986; Dec et al.,
2002).
1. Ion Exchange
Organic cations such as protonated amines and quarternary ammonium cations

will enter into exchange reactions with metal cations normally occupying ex-
change sites on the surfaces of mineral colloids. Organic cations are often preferred
over metal cations because of size considerations. However, small organic cations
396 P. M. HUANG
can be replaced by metal cations. Organic cations will also exchange with other
species of the organic cations on the exchange sites of mineral colloids.
2. Protonation
Some organic molecules may become cationic after adsorption at the surfaces
of mineral colloids by protonation. The relative basicity of the organic species
and the Brønsted acidity of the mineral surface determine whether or not
protonation happens. Organic molecules which possess basic amino groups have
the possibility of accepting protons from the mineral surface to render the species
cationic. The degree of protonation of the organic molecule is determined by its
intrinsic basicity and the ability of the surfaces of the mineral colloids to donate
protons. Although protons can occupy some exchange sites, many of the protons
available may arise from hydrolysis of water associated with exchangeable metal
cations. Therefore, the nature of exchangeable cation, the water content, and the
nature of mineral colloids play an important role in influencing the degree of
protonation of organic molecules.
3. Coordination and Ion-dipole
Coordination or ion-dipole interaction is another kind of mineral colloid–

organic interaction. Where exchangeable cations are of the transition metal type,
unfilled d orbitals will allow the formation of coordinate covalent bonds wherein
an electron donor such as N or O on functional groups of organic molecules are
available as ligands. Where exchangeable cations are not able to form classical
coordination complexes, they will interact with polar molecules through ion-
dipole processes.
4. Hydrogen Bonding
Hydrogen bonding is another type of interaction possible between organic

molecules and mineral surfaces. Where a polar organic molecule cannot displace
a water molecule solvating an exchangeable metal cation, it settles for a
secondary role by H bonding with the directly coordinated water. This type of
arrangement is referred to as water bridge where water acts as the linking entity
between an exchangeable metal cation and a polar organic molecule. Another
H-bonding interaction may take place between an organic cation on an exchange
site and another organic molecule. For example, a protonated amine on an
exchange site is H bonding with a carbonyl group.
5. Hydrophobic Bonding on Clay –Organic Complexes
Natural clay – organic complexes may have adsorption properties quite

different from pure clays. Many organic molecules including aromatics and
particularly the hologenated types such as chlorinated and brominated phenyls
and biphenyls are adsorbed a little if any on clay surfaces. In natural environment,
they are more likely to be adsorbed on organic components of clay –organic
complexes. The surfaces of clays can be made hydrophobic after reaction with
organic molecules that have some hydrophobic properties. When this happens,
the surfaces of clay –organic complexes become hydrophobic and in turn
organophilic. Therefore, hydrophobic portions of the adsorbate can react with the
hydrophobic organic portions of the clay – organic complexes. Such adsorption
interactions may take place since natural organic residues may have regions of
hydrophobicity wherein hydrophobic molecules might be bound.
6. Anion Adsorption
Organic anions may be adsorbed on the surfaces of mineral colloids. Although

pure clay minerals are negatively charged, hydroxy-interlayered clay minerals
(Fig. 1) have positively charged sites. Organic anions can interact with polyvalent
cations such as Al and Fe on the surface coatings of clay minerals. Such an
interaction has also been suggested for humic (HAs) and fulvic (FAs) acids in soils.
At low water content, the dissolved organic species approach the surfaces of mineral
colloids more closely than in a suspension environment, and, therefore, have a better
opportunity for interaction with polyvalent metal cations on the surface coatings.
C. BINDING OF HUMIC SUBSTANCES
The mechanisms of binding of nonhumic substances by mineral colloids

discussed above are fundamental in understanding the formation of mineral
colloid –humic complexes. Comprehensive reviews of clay – humic interactions
have been published (Theng, 1979; Schnitzer, 1995, 2000). The binding sites of
mineral colloids for humic substances include external surfaces and interlayers of
mineral colloids as discussed below.
1. Adsorption on External Mineral Surfaces
The extent of adsorption of humic substances on external mineral surfaces

depends on the physical and chemical characteristics of the surface, the pH of the
system, and its water contents. The formation of a wide range of mineral –humic
398 P. M. HUANG
associations involve chemical bonding with widely differing strengths. One can
visualize that complexation reactions of HA and FA with structural cations of
edges and hydroxy Al (or Fe) coatings on mineral colloids are important binding
mechanisms. Hydrogen bonding is also of considerable importance as clearly
indicated in IR spectra of HA- and FA-mineral complexes (Schnitzer and Khan,
1972). These reactions apparently involve H and O of COOH and OH groups in
HA and FA and O and H of external planar surfaces and edges. Cations with
high hydration energies on mineral surfaces react via water bridges with HA- and
FA- functional groups. Fulvic acid can also be adsorbed by sepiolite, apparently
through displacement of bound and/or zeolitic water in the structural channels by
undissociated FA (Kodama and Schnitzer, 1974).
2. Adsorption in Clay Interlayers
Interlayer adsorption of FA by expansible layer silicates is pH dependent,

being greatest at low pH, and no longer occurring at pH 5.0 (Fig. 2). An inflection
occurs in the adsorption – pH curve near the pH corresponding to the pK of FA.
In this type of reaction, the FA apparently penetrate the coordination shell of
the dominant cation in the clay and displace water coordinated to the cation in the
clay interlayer. The ease with which water can be displaced depends on the
affinity for water of the dominant interlayer cation and also on the degree of
dissociation of the FA. Since the latter is very low at low pH, undissociated FA
Figure 2 Effect of pH on basal spacing (d001) and adsorption of FA. From Schnitzer and
Kodama (1966).
can penetrate the clay interlayer. Therefore, interlayer adsorption of FA is

greatest at low pHs.
Wang and Huang (1986) reported that one of the well-defined precursors for
the formation of humic substances, hydroquinone, can be transformed in aqueous
solution even at near-neutral pH (6.5) to humic macromolecules and adsorbed in
the interlayers of nontronite (Fe(III)-bearing smectite) saturated with Ca (Fig. 3),
which is the most common and most abundant exchangeable cation in soils and
sediments. The decrease in the organic carbon contents of the complex on heating
to between 150 and 5008C is accompanied by a decrease in basal d values. The
heat treatment at 5008C destroys the humic macromolecule interlayers, as
indicated by the disappearance of the slight skewness to low angles of the X-ray
diffraction pattern. Most of the interlayer humic macromolecules are highly
resistant to alkali extraction and are, thus, humin-type materials. Therefore, the
formation of humic substance interlayers in 2:1 expansible layer silicates,
Figure 3 X-ray diffractograms of Ca-nontronite, influenced by deposition of humic

macromolecules in the interlayers through polymerization of hydroquinone and the associated
reactions. Diffractograms were recorded on a Philips X-ray diffractometer using Fe-Ka radiation
with an Mn filter. (a) Ca-nontronite; (b, c) humic macromolecule-Ca-nontronite complex before
(b) and after (c) NaOH extraction. From Wang and Huang (1986).
400 P. M. HUANG
through polymerization of phenol monomers and associated reactions in soils and

sediments, merits close attention.
III. INFLUENCE OF ORGANIC SUBSTANCES

ON THE FORMATION, TRANSFORMATION,
AND SURFACE PROPERTIES OF METAL OXIDES
Metal oxides are ubiquitous in soil. They play a very significant role in
influencing soil behavior and, thus, have great impact on the ecosystem. They
may exist as crystalline minerals, as SRO minerals, or noncrystalline precipitates,
which are partly present as coatings on clay minerals and humic substances. The
SRO Al and Fe oxides are undoubtedly among the most reactive inorganic
components of acidic and neutral soils (Kämpf et al., 2000; Huang et al., 2002;
Bigham et al., 2002). Organic compounds exert an important influence on the
hydrolytic reactions of Al and Fe and on the formation, transformation, and
surface properties of these metal oxides.
A. ALUMINUM OXIDES
Numerous conceptual models have been proposed to explain reaction

pathways of Al hydroxide formation from hydrolyzed Al species. The most
recently proposed model (modified from Bertsch and Parker, 1996) includes
concepts from the Al(OH)3 fragments of Hsu (1988) and more recent data
on Al13 polynuclear species ðAlO4 Al12 ðOHÞ24 ðH2 OÞ7þ 12 Þ (Fig. 4). This model
incorporates multiple reaction paths, many of which might occur simultaneously
but at different reaction rates. Reaction Path I is based on the Al(OH)3
fragments model and is believed to be common when solutions are brought to
supersaturation through simple dilution, or when systems are neutralized very
rapidly, within seconds to hours. Under these conditions, little or no Al13 forms;
the formation of Al(OH)3 nuclei is rapid, and gibbsite appears within days or
weeks of aging. Reaction paths II and III both involve the formation of Al13.
Path II is applicable when initial conditions favor formation of soluble Al13. The
individual Al13 ions can remain in solution for months to years, but eventually
they transfer by one of three pathways. If no Al13 nuclei are present, the soluble
Al13 ions can slowly dissociate into Al3þ ions and deposit onto Al(OH)3 nuclei
(Path IIa). These soluble Al13 may also aggregate and form Al13 nuclei (Path
IIb). If some Al13 nuclei are already present, the remaining Al13 ions can be
deposited onto them within weeks and months (Path IIc). Path III is applicable
when Al13 forms under conditions that promote the rapid formation of Al13
aggregates via an anion bridging mechanism, i.e., outer-sphere associations.
Figure 4 A proposed reaction scheme for Al hydroxide mineral formation from hydrolyzed
aluminum solutions. From Huang et al. (2002).
These aggregates rapidly rearrange themselves in Al13 nuclei. The Al13 nuclei in
Paths IIb, IIc, and III then transform into a poorly ordered phase, i.e.,
microcrystalline boehmite (pseudobehmite). This phase then transforms into
gibbsite or sometimes bayerite. The Al(OH)3 fragments model and the Al13
polymer model are complementary. The best model to describe Al
transformations in soils is still an open question particularly since the existence
of Al13 in nature remains to be established.
The relative importance of each pathway (Fig. 4) and thus the rate at which
gibbsite forms in a given system depend greatly on reaction conditions. The
hydrolysis and polymerization of Al and the subsequent transformation into more
crystalline phases via the various pathways described above are strongly influenced
by the nature and concentration of soluble inorganic and organic ions and solid-
state ions such as clay minerals and humus (Huang and Violante, 1986; Huang,
1988; Bertsch and Parker, 1996; Krishnamurti et al., 1999; Huang et al., 2002).
The influence of organic acids has been studied extensively, with most of the
focus on the particular solid phases that form (Huang et al. 2002). The influence
of a particular organic acid is generally related to the stability constant of the
complex that the acid forms with Al (Table I). Therefore, p-hydroxybenzoic acid,
which forms the least stable complex with Al, does not inhibit the crystallization
of Al hydroxides, whereas aspartic, tannic, malic, and citric acids increasingly
402 P. M. HUANG
Table I
Stability Constants of the Complexes Formed Between Al and Five Organic Acids at 258C
Stability constants of the complexes
Organic acids log K1 log K2
p-hydroxybenzoic acid 1.66 –

Aspartic acid 2.60 –
Tannic acid 3.78 –
Malic acid 5.14 8.52
Citric acid 7.37 13.90
From Kwong and Huang (1979a).
retard crystallization (Fig. 5). Besides the stability constant of the complex, the
concentration of the organic acid is important. At certain low concentrations,
the presence of some organic acids actually promotes the crystallization of
particular Al(OH)3 polymorphs, but above a certain critical concentration, it
disrupts crystallization (Huang and Violante, 1986). This is because organic acids
replace water molecules that would otherwise coordinate with the Al3þ ion and
the extent to which this occurs depends upon the chemical affinity of the organic
acid for the Al, that is its stability constant, and its concentration relative to Al.
Figure 5 The X-ray diffraction patterns of hydrolytic precipitation products of Al, showing how
four different organic acids influence the transformation to more crystalline phases. The initial Al
concentration was 1.1 £ 1023 M at an OH/Al molar ratio of 3 and the solution was aged for 40 d at
room temperature in the presence of 1024 M organic acid. From Kwong and Huang (1979b).
Humic substances such as FA and HA also influence the transformation of Al

by promoting the formation of microcrystalline boehmite and hampering the
formation of more crystalline phases (Kodama and Schnitzer, 1980; Singer and
Huang, 1990). Fulvic acids and HA resemble aliphatic acids, such as citric and
malic acids, in that they contain COOH and aliphatic OH groups. They also
resemble quercetin and tannic acid, because they contain phenolic hydroxyl
and ketonic C ¼ O groups. Through these functional groups, FA and HA form
stable complexes with Al and inhibit the crystallization of Al hydroxides.
The influence of FA and HA also resembles that of salts in that they favor the
formation of Al –O – Al (oxo) over that of Al – (OH) –Al (01) linkages.
Organics can have an enormous impact on the surface properties of the Al
transformation products. For instance, the presence of organic acids during aging
of Al hydroxide gels for 40 d increases the specific surface of the products up to
30-fold over that of the control, and higher acid concentrations result in higher
specific surface (Table II). The surface charge characteristics of the products also
dramatically changed (Kwong and Huang, 1978, 1979c, 1981).
The intermediate transformation products of Al, which include soluble
mononuclear and polynuclear Al species and colloidal SRO Al hydroxides, are
the most reactive Al species in influencing physical, chemical, and biological
processes of soils and associated environments. These Al species influence
environmental quality and ecosystem health by impacting acidification of the
terrestrial environment, formation of humic substances, development of soil
structure, dynamics of organic C cycling and global warming, transformations of
nutrients and pollutants, biological productivity and food chain contamination,
microbial and enzymatic activity, aquatic ecosystem integrity, and human and
animal health (Huang et al., 2002).
Table II
Specific Surface Area of Al Hydroxide Precipitation Products Formed in the Presence of Tannic
Acid and Selected Low Molecular Weight Organic Acids in Systems at Initial Al Concentrations
of 1.1 £ 1023 M and OH/Al Molar Ratios of 3.0 and Aged for 40 d at Room Temperature
Organic acid concentration (mol L21)
Organic acid 1.0 £ 1026 1.0 £ 1024
(m2 g21)
None 20 20
p-hydroxybenzoic 22 28
Aspartic 27 587
Tannic 95 195
Malic 36 635
Citric 117 295
From Kwong and Huang (1981).

404 P. M. HUANG
B. IRON OXIDES
The formation pathways of different Fe oxides in soils may be elucidated on

the basis of observation of the pedogenic environment under which a certain
phase or association of phases occur in a particular soil or soil sequence. Based on
field observations and laboratory synthesis under near-earth surface conditions,
some general pathways of formation and transformation of Fe oxides in soil
environments have been proposed (Schwertmann and Taylor, 1989; Cornell and
Schwertmann, 1996). Soil environmental factors, which influence the modes of
formation of Fe oxides, include the initial valence of the Fe source, its
concentration in soil solution, the rate of oxidation of Fe (II), and parameters such
as pH, Eh, temperature, soil moisture, soil organic matter, activity of Al, and
other inorganic and organic ionic factors.
Soil organic matter plays a vital role in the formation of pedogenic Fe oxides
(Fig. 6). In soil environments where the amount of organic matter is low, for
example in subsoils, the Fe supplied will form goethite and hematite depending
on environmental factors (Schwertmann, 1985). As the organic matter content
increases, more of the Fe will be complexed with organics leading to decreases
in Fe activity. The activity of Fe(III) ions is so low that only the solubility
product of goethite (10241 –10242), but not the solubility product of ferrihydrite
(10237 – 10239) is exceeded. Consequently, goethite but not the ferrihydrite may
form. Therefore, no hematite will form in an environment where the organic
matter content is high, since ferrihydrite is deemed a necessary precursor for
Figure 6 Tentative schematic representation of the effect of organic matter content and rate of
soluble Fe supply on the formation of various Fe forms in soils. From Schwertmann et al. (1986).
hematite. This trend is generally observed in the soils in the temperate and cool
regions as well as in wet depressions and surface soils of the subtropical and
tropical regions. At a higher content of organic matter, the rate of Fe supply is
high, ferrihydrite will form and may survive for pedogenic times. If the content of
organic matter is even higher such as in O horizons or in peaty environments, all
of the Fe may be in the form of Fe –organic complexes. The interaction of organic
matter with surface Fe may be the main mechanism for inhibiting the formation
of crystalline Fe oxides in organic C-rich horizons. Laboratory studies verify this
postulation and elucidate the mechanism of the inhibition (Cornell and
Schwertmann, 1996).
In an aqueous weathering environment, Fe oxides generally form via solution
transformation. The oxidation products of Fe(II) solutions are, thus, important
because it is in this valence state that Fe is commonly mobilized during
weathering under the EH – pH region of natural soil environments. The kinetics of
Fe(II) oxidation and the nature of Fe oxides formed as influenced by a series of
organic ligands have been reported (Krishnamurti and Huang, 1990, 1991, 1993).
The relative stability of the Fe(II) – ligand complex influences the rate of Fe(II)
oxidation. The rate constant of Fe(II) oxidation at the same ligand/Fe molar ratio
generally decreases with the increase in the stability constant of Fe(II) –ligand
complex (Krishnamurti and Huang, 1990). In the absence of any complexing
ligands, the precipitation product formed at pH 6.0 is dominantly goethite with
small amounts of poorly crystalline lepidocrocite. In the presence of oxalate
(log K ¼ 2.52, where K is the stability constant of the Fe(II) – ligand complex),
the slower rate of formation of Fe(III) and its hydrolysis apparently modifies the
oxygen coordination in the edge-sharing Fe(O, OH)6 octahedron, resulting in the
formation of lepidocrocite at the expense of goethite. The strong complexation
of Fe(II) with tartrate (Log K ¼ 4.85) influences the kinetics of Fe(II) oxidation
and Fe(III) hydrolysis and perturbs the crystallization of hydrolytic products of
Fe(III), resulting in the formation of X-ray noncrystalline Fe oxides.
The rate constant of Fe(II) oxidation also decreases with the increase in the
perturbing ligand/Fe(II) molar ratio due to the increasing formation of the Fe(II) –
ligand complex (Table III). The retardation of Fe(II) oxidation inhibits the
nucleation of goethite and promotes the crystal growth of lepidocrocite especially
at a citrate/Fe(II) molar ratio of 0.001. Further increase in the amount of citrate
distorts the structural order, resulting in the increasing inhibition of crystal
growth of lepidocrocite and the formation of noncrystalline Fe oxides.
The surface of Fe oxides is the region of their interactions with the soil solution,
organic and inorganic particles, plant root, microorganisms, and other soil biota.
Therefore, surface properties of Fe oxides should have profound impacts on
microaggregate formation, water flux, nutrient and pollutant flux, and the ability of
soils to promote plant growth, to maintain reasonable soil biotic habitat, and to
respond to management, and resist degradation. However, most studies on the
surface chemistry of Fe oxides have been conducted on the two common Fe oxide
406 P. M. HUANG
Table III
Rate Constants of Fe(II) Oxidation and Nature of Fe Oxides Formed in Absence and Presence of
Citrate in Ferrous Perchlorate-NaOH System at pH 6.00 and 23.58C
XRDa 020 peak of lepidocrocite
Initial citrate/Fe Rate constant Dominant WHH ^ 0.02 Area

molar ratio (min21) £ 104 Fe-minerals (82u) (mm2)
0 41.3 ^ 2.0 G,L 1.40 40 ^ 5

0.0003 28.6 ^ 3.5 L,G 1.00 97 ^ 5
0.0005 26.9 ^ 1.8 L 1.00 124 ^ 12
0.0010 24.4 ^ 3.0 L 0.80 263 ^ 3
0.0050 20.2 ^ 1.8 L 1.10 82 ^ 4
0.0060 19.1 ^ 1.5 L 1.20 25 ^ 3
0.0080 16.8 ^ 0.6 L 1.20 10 ^ 3
0.0100 15.2 ^ 0.7 PC n.a.b n.a.
0.1000 7.6 ^ 0.7 no ppt n.a. n.a.
From Krishnamurti and Huang (1991).

a
X-ray powder diffraction data. WHH ¼ width at half height; L ¼ lepidocrocite; G ¼ goethite;
PC ¼ poorly crystalline.
b
n.a. ¼ not applicable.
minerals, goethite and hematite. This is based on the assumption that all the Fe
oxides could have similar surface properties. Xue and Huang (1995) reported that
the presence of citric acid, one of the most common organic ligands in soils, during
the formation of Fe oxides substantially modifies the characteristics of surface
properties of the oxide of Fe. Further, the fine scale morphology, mean surface
roughness, fractal dimension, specific surface, and microporosity of the Fe oxides
depend on the citrate concentration in the solution in which the Fe oxides are
formed (Liu and Huang, 1999). The surface properties of Fe oxides formed under
the influence of organic ligands merit close attention in advancing our
understanding of their surface chemistry pertaining to dynamics and transforma-
tions of nutrients and pollutants in terrestrial and aquatic environments.
IV. ROLE OF SOIL MINERALS IN ABIOTIC CATALYSIS

OF THE FORMATION OF HUMIC SUBSTANCES
A. POLYPHENOL PATHWAY
The mechanisms of the formation of humic substances are very complicated

processes in which a variety of organic components, such as phenolic
compounds, carbohydrates, and nitrogenous substances, can participate as
starting materials. Humic substances can be synthesized through biotic and

abiotic processes. Soil minerals play an important role in catalyzing the abiotic
polymerization of phenolic compounds and the polycondensation of
phenolic compounds and amino acids and the subsequent formation of humic
substances.
Kumada and Kato (1970) are among the pioneers in the study of browning
of pyrogallol catalyzed by clay-size layer silicates. In a series of studies, Wang
and his co-workers reported that layer silicates catalyze the abiotic formation of
humic substances through oxidative polymerization of phenolic compounds
common in soil environments (Wang and Li, 1977; Wang et al., 1986).
Intracrystal surfaces of transition metal saturated layer silicates can catalyze the
formation of aromatic cation (Pinnavaia et al., 1974). Aromatic molecules may
donate electrons to metal ions, such as Cu((II) and Fe(III), on the cation-exchange
complex of smectite (Mortland and Halloran, 1976). The ability of muscovite to
catalyze the oxidation of some phenolic compounds to form HA has also been
demonstrated (Filip et al., 1977). Since the early 1980s, Huang and co-workers
have studied the sequence of catalytic power of layer silicates and their reaction
sites in the polymerization of phenolic compounds and the subsequent formation
of humic substances. The promoting effect of 2:1 layer silicates is higher than 1:1
layer silicates (Shindo and Huang, 1985a). This is attributed to the higher specific
surface and higher lattice imperfection of the former than the latter. The edges of
kaolinite are virtually the only catalytic sites for the formation of polyphenol-
derived humic substances. The edges of nontronite have a very important role as
catalytic sites in the formation of hydroquinone-derived humic macromolecules
(Wang and Huang, 1988). The Fe(III) in the octahedral sheet of nontronite also
serves as a Lewis acid site to catalyze the oxidative polymerization of
hydroquinone (Wang and Huang, 1986). Further, nontronite has the ability to
cleave the ring of pyrogallol, catechol and hydroquinone; the extent of ring
cleavage of the polyphenols to release CO2 varies with their structures (Wang and
Huang, 1994).
Iron oxides and oxyhydroxides can catalyze polymerization of hydroquinone
(Scheffer et al., 1959). The importance of hydrous oxides of Fe in catalyzing
the oxidative polymerization of phenolic compounds is influenced by the
structure and functionality of phenolic compounds. Hydrous oxides of Al also
catalyze the oxidative polymerization of phenolic compounds (Wang et al.,
1983a; Shindo and Huang, 1984; Wang and Huang, 2000a). Aluminum and Fe
oxides and oxyhydroxides may form surface coatings on layer silicates and humic
substances. The ability of various Al and Fe oxides and oxyhydroxides and their
surface coatings on soil particles to promote oxidative polymerization of phenolic
compounds remains to be uncovered. Among metal oxides, hydroxides, and
oxyhydroxides, Mn oxides are the most powerful catalysts in the transformation
of phenolic compounds (Shindo and Huang, 1982, 1984). Manganese oxides
(birnessite, cryptomelane, and pyrolusite), which are common in soils, act as
408 P. M. HUANG
Figure 7 Solid-state 13C NMR spectra of synthesized HA: (a) pyrogallol and (b) birnessite–
pyrogallol systems. The dry, powdered HA formed in the reaction systems was analyzed with
cross-polarization and magic-angle spinning (CPMAS), proton decoupling; contact time 2.00 ms,
recycle time 1.00 s, spectral width 20,000.0 Hz, spinning speed 3700 Hz, number of scans 60,000
and spectrometer frequency 37.7 MHz. The band positions and assignments for the spectra are
as follows: 196.3 ppm (carbons in aldehydes and ketones), 180.9– 170.7 ppm (carboxyl carbons
Lewis acids that accept electrons from phenolic compounds, leading to their
formation of semiquinone and their oxidative polymerization and formation of
humic substances. The solid state cross-polarization magic-angle spinning
(CPMAS) 13C NMR spectrum of HA (MW . 1000) formed in the Mn (IV)
pyrogallol system (Fig. 7b) is very similar to those of HAs extracted from
natural soils (Hatcher et al., 1981; Schnitzer and Preston, 1983). The spectrum
(Fig. 7b) and the spectra of natural humic substances in the cited references
show clearly the presence of carbons in aldehydes and ketones (196.7 ppm),
carbonyl carbons in carboxyl and ester groups (173.3 ppm), aromatic carbons
(141.4, 132.5, 120.3 ppm), CO carbon– alcohols, esters, ethers, carbohydrates
(79.4 ppm), and alkyl carbons (41.3, 34.4, and 19.6 ppm). However, the
spectrum of synthesized HA (MW . 1000) formed in the pyrogallol system in
the absence of Mn(IV) oxide (Fig. 7a) is very different from those of natural
humic substances. Wang and Huang (2000a, b) reported that the catalytic power
of metal oxides in the abiotic ring cleavage of pyrogallol and the
polycondensation of the resultant fragments varies greatly with the nature of
the oxides. The abiotic ring cleavage of polyphenols catalyzed by SRO oxides
of Al and especially Fe and Mn (Wang and Huang, 2000a, b) may account, in
part, for the carboxylic group contents and the origin of the aliphaticity of
humic substances in soils (Schnitzer, 1977; Preston et al., 1982). Therefore, the
catalytic power of these oxides in affecting the C turnover and humic substance
formation via abiotic processes in soils and related environments deserve
attention.
Poorly crystalline aluminosilicates are common in soils (Wada, 1989). Allo-
phane has the ability to catalyze the polymerization of polyphenols (Kyuma and
Kawaguchi, 1964). Humified polyphenols resemble natural humic substances in
their infrared spectra (Wang et al., 1983b). There are a series of poorly
crystalline aluminosilicates in soil environments (Huang, 1991). Their ability to
catalyze the transformation of phenolic compounds to humic substances remains
obscure.
Primary minerals differ in their ability to catalyze the abiotic polymeri-
zation of hydroquinone. The sequence of the catalytic power of primary minerals is
tephroite . actinolite . hornblende . fayalite . augite . biotite . musco-
vite < albite < orthoclase < microcline < quartz (Shindo and Huang, 1985b).
Except for the tephroite, which is a Mn-bearing olivine, the hydroquinone-derived
products are largely low-molecular weight humic polymers.
in carboxyl and ester groups), 152.5–111.4 ppm (aromatic carbons and olefinic carbons),
82.9 ppm (CO carbons—alcohols, esters, ethers, and carbohydrates), and 43.0 ppm (alkyl
carbons). The formation of alkyl carbons, CO carbons, and carboxyl carbons in (b) is due to
the ring cleavage and fragmentation of pyrogallol catalyzed by birnessite. From Wang and
Huang (1992).
410 P. M. HUANG
B. MAILLARD REACTION
The Maillard reaction (Maillard, 1913) is thought to be a major pathway in

humification because of significant similarities between humic substances and
melanoidins formed through this pathway involving sugar-amino acid conden-
sations (Ikan et al., 1996). Evershed et al. (1997) detected the presence of
characteristic products of the Maillard reaction (alkyl pyrazines) in archae-
ological plant remains up to 1500 years from Egypt. Despite the importance of
the Maillard reaction, the mechanisms and rates of polycondensations of sugars
and amino compounds in nature remain obscure (Ikan et al., 1996).
Jokic et al. (2001a) reported that birnessite (d-MnO2) significantly increases the
extent of humification of the glucose –glycine system over the pH range of 6– 8
(Fig. 8). They verified the formation of humic substances in the glucose – glycine –
birnessite system by measuring the CPMAS 13C NMR spectrum of the isolated FA.
The chemical shifts of FA formed in the Maillard reaction system (Jokic et al.,
2001a) resemble those of natural humic substances such as soil and stream FAs
(Malcolm, 1989; Schnitzer, 2000). Their data clearly show that birnessite greatly
accelerates the Maillard reaction between glucose and glycine forming poly-
condensation products resembling humic substances under ambient conditions.
Furthermore, light intensity of 168 m E S21m22 exerts an increase in the browning
of the glucose –glycine –birnessite system compared with the same system,
which is kept in the dark (Jokic et al., 2001b). Equally important is that even in
complete darkness, birnessite catalyzes the Maillard reaction between glucose and
glycine. Therefore, birnessite catalysis can occur in soil or sediment environments
at any depth, but the presence of sunlight should strongly accelerate the reaction.
V. INTERACTIONS OF SOIL MINERAL COLLOIDS

WITH MICROORGANISMS
A. SURFACE INTERACTIONS
Soil is a habitat for myriads of microorganisms. Although the microbial

biomass constitutes only a very small proportion (, 3%) of the total organic
carbon in soil, it is the most active and dynamic fraction of the living organic
pool. Among the inorganic components of soil, mineral colloids are the most
reactive because of their large specific surface and reactive charge characteristics.
Being enriched in ions, water, and organic matter relative to the bulk soil, the
surface of mineral colloids serves as a preferred habitat for soil microorganisms.
That complex formation between mineral colloids and microorganisms can and
does occur, when the components are brought together in the laboratory, has been
amply documented (Marshall, 1980).
Figure 8 Absorbance versus wavelength plots in the Maillard reaction between glucose
and glycine as influenced by birnessite catalysis. (a), (b) and (c): 30 d reaction period. (d), (e) and
(f): 15 d reaction period. From Jokic et al. (2001a).
Although the surface of bacterial cells and crystalline clay minerals are both
negatively charged, bacteria have a propensity for producing extracellular
polysaccharides (EPS) which bind simultaneously to cell and clay surfaces
through cation bridging involving polyvalent cations (Fig. 9). Cation bridging
may be direct [Scheme (1)] or effected by H-bonding to water molecules in the
primary hydration shell of polyvalent cation [Scheme (2)]:
½Clay2 · · · Mnþ · · · 2 ½EPS – B ð1Þ
or
½Clay21 · · · Mnþ · · · H 2
OH · · · ½EPS – B ð2Þ
412 P. M. HUANG
Figure 9 Diagram illustrating the interaction of bacteria and fungi with mineral particles in a soil
aggregate. Bacterial cells with a coat of extracellular polysaccharides (EPS) are enveloped by clay
particles. The pore space where clays and bacteria interact, bounded by silt- and sand-size particles, is
relatively enriched in organic matter including EPS residues. Fungal hyphae are attached to the
outside surface of an aggregate. Inset shows an enlarged view of a bacterial cell with its complement
of EPS. At normal soil pH conditions, the cell has a net negative surface charge. Most clay particles
adhere to the cell surface by bridging through polyvalent cations, represented by Mnþ (a) although
some may be attached directly by electrostatic interactions, either in face-to-face (b), or edge-to-face
(c) association. From Theng and Orchard (1995).
where Mnþ denotes a polyvalent cation of valency n, and B is a bacterial cell

(Theng and Orchard, 1995). The predominant mechanism would depend on the
nature of the polyvalent cation but, more so, on the hydration status of the soil.
Dehydrating conditions would clearly favor mechanism (1). In addition, short-
range interactions (van der Waals, H-bonding) may come into play because the
EPS chains are induced to make close contact with the surface, increasing the net
energy of interactions.
Mineral colloids in soil are generally coated with hydroxy Al (or Fe) polymers
in soil (Fig. 1). These coated minerals thus behave as a positively charged species
or display amphoteric characteristics. Therefore, mineral colloids can strongly

interact with negatively charged microbial cells in soil environments. This type
of bonding, which is much stronger than cation bridging, is also expected to
occur with Al and Fe oxides over the pH range of soils. The attachment of
microorganisms to SRO minerals and the crystal edges of layer silicates
through electrostatic interactions would also be predicted to occur when the
soil pH falls below 6 because all of these surfaces would then be positively
charged.
Furthermore, in the majority of cases, minerals in topsoils are partially
covered with organic materials, especially humic substances (HS), which are
microbially resistant and the most prevalent. The most common mode of mineral
colloid-organic material-microorganism interactions may be represented by
Scheme (3).
½Mineral colloid – HS2 · · · Mnþ · · · 2 ½EPS – B ð3Þ
In humic-rich calcareous Mollisols, Ca would be the predominant bridging

cation. In Andisols, Oxisols, Ultisols, and the B horizon of Spodosals, HS
largely occur as complexes with Al and Fe, or their respective poorly
crystalline oxides (Theng et al., 1989; Oades et al., 1989). In soils with little
organic matter and in subsoils, mineral colloid– microorganism interactions
are largely influenced by the mineralogical composition and pH of the
system.
To date, the research data on the interactions of microorganisms with
crystalline layer silicates with different structural configurations and surface
properties are limited and even less information is available with respect to SRO
minerals and Al and Fe oxides. Much of the evidence on the mechanisms for
direct in situ-as opposed to in vitro-surface interactions of mineral colloids with
microorganisms is circumstantial and equivocal (Stotzky, 1986; Theng and
Orchard, 1995).
B. INFLUENCE ON MICROBIAL ACTIVITY

AND SURVIVAL
Mineral colloids, in general, can directly or indirectly influence the activity of

microorganisms in their immediate vicinity. The effect of mineral colloids may
be positive, negative, or sometimes so small as to escape detection.
Mineral colloids have a stimulating effect on the activity of adhering
bacteria by keeping the pH of microhabitats within the optimum
physiological range for growth. The content and type of mineral colloids
are influential in determining the balance between different microbial
414 P. M. HUANG
populations in soil. A well-known example is the failure of some fungi to

thrive and spread in certain soils (Stotzky, 1986). This is largely attributed to
the presence of montmorillonite in the suppressive soils. Montmorillonite can
serve as a proton sink and is thus able to promote the growth of acid-
sensitive bacteria in these soils. This gives bacteria a selective advantage
over fungi in competing for available nutrients. Therefore, fungal growth and
proliferation are effectively suppressed.
Mineral colloids can also stimulate microbial activity by sorbing
metabolites that would otherwise have an adverse effect on microbial growth
(Filip et al., 1972; Filip and Hattori, 1984). Montmorillonite is more effective
than kaolinite and finely ground quartz. Furthermore, surfaces of mineral
colloids can adsorb other toxic substances to microorganisms such as
antibiotics and pesticides (Theng and Orchard, 1995).
Adhering microorganisms may benefit from being close to nutrients
adsorbed on mineral colloids. However, the potential substrates may not be
readily available or physically accessible (Fletcher, 1991). Furthermore,
beyond a certain concentration, the addition of mineral colloids to the system
may result in a reduction in microbial activity (Marshall, 1971). This is
attributable to restricted diffusion of oxygen and nutrients to microbial cell as
it is progressively enveloped by mineral colloids.
However, by forming an envelope around bacteria cells, mineral colloids
may provide protection from extreme fluctuations in physicochemical
environments, and thus enhance bacterial survival (Stotzky, 1986; Theng
and Orchard, 1995). The protective effect of mineral colloids, especially
montmorillonite, is manifested in the ability of soil microorganisms to
withstand exposure to hypertonic osmotic pressure, dessication, and
ultraviolet radiation. Besides serving as a substratum for bacteria to adhere
to and as a protective envelope of bacterial cells, mineral colloids can act as
a cementing agent of soil particles. Therefore, the addition of clay to soils,
especially those of light texture, would modify the spatial arrangement of
particles and especially the pore –size distribution within soil aggregates.
Such a modification of aggregate structure often benefits the bacterial
population by increasing the proportion of pores of a certain size range
(, 6 mm) from which bacteria could freely enter to colonize pores but
bacterial predators notably protozoa are effectively barred from entering such
pores due to steric hindrance. Another survival mechanism is adhesion of
bacteria to mineral colloids, since these surfaces are relatively enriched in
substrates and water. Even if they lack available nutrients, solid surfaces may
trigger cells to diminish in size, resulting in an increase in the surface-to-
volume ratio and packing density (Kjelleberg et al., 1982). Such bacterial
cells can respond rapidly to conditions of high nutrient and water supply and
regain their normal size and growth pattern.
VI. INTERACTIONS OF SOIL COLLOIDS

WITH ENZYMES
A. MINERAL COLLOID – AND HUMIC – ENZYME COMPLEXES
Extracellular enzymes are rapidly sorbed at mineral colloids and humic

substances in soil environments. Mineral colloids have a high affinity for
enzymes although this is not always synonymous with the retention of catalytic
ability. Enzymes associated with soil humic substances retain their activity for
long periods.
Mineral colloid –enzyme interactions have been documented (Theng, 1979;
Burns, 1986; Stotzky, 1986; Naidja et al., 2000; Violante and Gianfieda, 2000).
The mechanisms whereby enzymes are retained by surfaces of mineral colloids
have usually been assumed to include cation – exchange reactions but not all
observations have supported this view. If this were true, one would expect
the adsorption to increase as pH values decrease below the isoelectric point of the
enzyme. Under these conditions, the protonation of amino groups should give rise
to an increase in positive charges, resulting in more adsorption to mineral
colloids. However, the adsorption of polysaccharases is highest at their
isoelectric points when the enzymes have no net charge. Therefore, besides
cation-exchange reactions, adsorption of enzymes by mineral colloids may
proceed through ionic, covalent, hydrophobic, and hydrogen bonding, and van
der Waals forces. Although individual van der Waal’s forces are regarded as
weak, retention may be cumulative if the enzyme is in close contact with the
adsorbing surface.
Naidja et al. (2002) reported that the tyrosinase protein have a very high
affinity to the surfaces of hydrous manganese oxide birnessite. Analysis by
atomic force microscopy (AFM) confirms the globular structure and spheroid
shape of the protein molecules and clearly shows the unfolding in an ellipsoidal
shape of the protein molecules after adsorption and immobilization on the
mineral surfaces, especially the frayed edges of the phyllomanganate layers
(Fig. 10). AFM pictures along with X-ray diffraction and infrared absorption data
of the protein– birnessite complex show the coating of the mineral surfaces by
the protein molecules as well as the changes in the protein conformation,
resulting in the unfolding and flattening of the protein molecules after binding to
mineral surfaces (Naidja et al., 2002). The information obtained is of
fundamental significance for understanding the nature of enzyme –mineral
complexes.
Relatively few studies have been made of enzyme adsorption to humic
substances, probably because of the biochemical heterogeneity, dynamic nature,
and physicochemical complexity of soil organic matter. Nevertheless, it is well
416 P. M. HUANG
Figure 10 Atomic force micrographs of wet samples of free protein and protein–birnessite
complexes after immobilication of the protein at an initial protein concentration of 3 g/g, pH 6.0,
and room temperature (238C) without air-drying: (a) Free tyrosinase sample, (b) protein –
birnessite complex formed from the tyrosinase sample, (c) free bovine serum albumin (BSA), and
(d) protein–birnessite complex formed from the BSA. The same scale was used from (a) to (d). From
Naidja et al. (2002).
known that enzymes are stabilized by association with soil organic matter
(Conrad, 1942; Burns, 1986; Nannipieri and Gianfreda, 1999). Many mechan-
isms have been proposed to account for the stability of enzymes which are
complexed with soil humic substances. These include ion exchange, H-bonding,
entrapment within three-dimensional micelles, lipophylic reactions, and covalent
bonding during organic matter genesis. Although it is not inconceivable that
some or even all of these reactions and processes are involved, the intrinsic
mechanisms of interactions of enzymes with humic substances are still obscure.
Functional groups of enzymes implicated in covalent bonding to humic polymers
include terminal and basic amino, carboxyl, sulfhydryl, phenolic, and imidazole
groups. All these may be involved in the stabilization process provided they do
not form part of the active sites of the enzyme and not crucial to the retention of
its tertiary structure. Enzyme– humic complexes can also be attached to mineral
colloids, and in some instances, this further enhances enzyme stability. Recently,
Violante and Gianfreda (2000) reported the stability of enzymes bound to organo-
mineral complexes. Their findings are basic to understanding enzymatic activities
in soils and related environments.
B. EFFECTS ON ENZYMATIC ACTIVITY
Enzymes adsorbed or immobilized on mineral colloids operate in a

structurally restricted system. The behavior of such enzymes toward their
respective substrate is, thus, different from that occurring in a homogeneous,
aqueous solution. When enzymes are adsorbed on mineral colloids, changes in
the tertiary structure (i.e., the folding of the helix or coil into a compact, globular
molecule stabilized by interfold hydrogen bonding, van der Waals and
hydrophobic interactions) of the enzymes and their active sites decreases activity
or eliminate it altogether (Burns, 1986). However, there are notable exceptions to
the adsorption-decline in activity rule. Gainfreda and Bollag (1994) investigated
the behavior of a laccase and a peroxidase in the presence of a montmorillonite, a
kaolinite, and a silt loam soil. The various supports show different enzyme
immobilization capabilities (Table IV). There is considerable variation in the
retained activities of the two enzymes on the four supports, as well as the same
enzyme on the four supports. The residual specific activities (calculated as
percentages of the specific activity of the free enzyme) of laccase and peroxidase
Table IV
Immobilization of a Laccase (from Trametes versicolor) and a Peroxidase (from Horseradish) on
Different Supports
Enzymatic activity
Enzyme and Protein Adsorbeda Units Specific Residual specific

support (mg/%) adsorbedb activityc activityd(%)
Laccase
Glass beads 0.452/56 28.8 63.7 236
Montmorillonite 0.622/71 19.8 31.8 118
Kaolinite 0.566/64 13.1 23.1 85.5
Soil 0.644/73 15.7 24.4 90.4
Peroxidase
Glass beads 0.092/17 8.4 91.6 93.8
Montmorillonite 0.224/43 23 102.8 105.2
Kaolinite 0.120/23 9.5 78.9 80.7
Soil 0.162/31 15 92.6 94.8
From Gianfreda and Bollag (1994).

a
Difference between proteins initially added to 200 mg of support (0.88 mg laccase and 0.52 mg of
peroxidase) and those recovered in the supernatant and washings.
b
Expressed as mmol O2 consumed min21 for laccase and mmol guaiacol transformed min21 for
peroxidase.
c
Units adsorbed/protein adsorbed.
d
Calculated as percentage of the specific activity (sa) of the free enzyme (laccase, sa ¼ 27 mmol
min21 mg21; peroxidase, sa ¼ 97.7 mmol min21 mg21).
418 P. M. HUANG
immobilized on all supports are high. Further, laccase immobilized on

montmorillonite shows specific activities higher than that of the free enzymes.
This may be attributable to steric modification of the immobilized enzyme. The
data clearly indicate that the performance of enzymes in soil environments is
significantly affected by soil mineral colloids. The study of interactions of
enzymes with soil inorganic constituents was mainly concerned with crystalline
clay minerals. The role of SRO oxides, hydroxides, and oxyhydroxides of Al, Fe,
and Mn in influencing enzyme activity involved in the transformations of natural
organics and xenobiotics (Huang, 1990) has now been recognized (Naidja et al.,
1997, 2000; Huang et al., 1999; Violante and Gianfreda, 2000) and merits further
attention.
A significant portion of immobilized enzymes in soils is associated with the
humic fraction, not by adsorption or occlusion, but by the formation of enzyme –
phenolic copolymers during the genesis of humic substances (Burns, 1986; Huang,
1990). These immobilized enzymes are active and extremely stable. They can be
adsorbed on mineral colloids. The catalytic role of soil minerals with different
structural configuration and surface properties in the formation of enzyme–
phenolic co-polymers and the subsequent impact on the activity and stability of
enzymes in the environment deserve close attention.
VII. MICROBIAL MEDIATION OF SOIL MINERAL

WEATHERING AND TRANSFORMATION
A. MINERAL WEATHERING
Chemical weathering of minerals, as a part of the soil formation process, can

be enhanced by microbial activity by a factor as high as 106 (Kurek, 2002).
Microorganisms can dissolve minerals by direct and indirect action under aerobic
and anaerobic conditions (Robert and Berthelin, 1986; Ehrlich, 2002; Kurek,
2002). The modes of attack of minerals by microorganisms include: (1) direct
enzymatic oxidation or reduction of a reduced or oxidized mineral component,
(2) indirect attack with a metabolically produced redox agent or inorganic and
organic acid, (3) indirect attack by metabolically produced alkali, usually in the
form of ammonia, (4) indirect attack with a metabolically produced ligand that
form a highly soluble product with a mineral component, and (5) indirect attack
by biopolymer. The mode of microbial attack of a mineral may involve a
combination of some of these mechanisms. In some cases of attack, the
microorganisms may be dispersed in the soil solution; in others, they may grow in
biofilms on the surface of the susceptible minerals.
When oxidized metal compounds such as Fe(III), Mn(IV), and As(V) act
as electron acceptors, anaerobic respiration becomes an example of
direct dissolving action under anaerobic conditions (Ahmann et al., 1994;

Ehrlich, 2002). Oxidation of sulfur entities of metal sulfides to obtain
energy is an example of direct dissolving action under aerobic conditions
(Kurek, 2002). Volatization of metals and metalloids or biomethylated metal
and metalloid compounds from the soil into the atmosphere can be
a mechanism of detoxification for toxic elements such as Hg, As, and Se
(Gadd, 1993).
Uptake and concentration of elements by absorption and/or adsorption were
observed with bacteria and fungi for P, Ca, Fe, and K. Weed et al. (1969) have
shown that fungi can adsorb K from solution and thus shift K equilibrium in
suspensions of trioctahedral and dioctahedral micas and transfer them to
vermiculites. Such process can also occur for many major and trace elements
(Robert and Berthelin, 1986).
Some microorganisms can promote the transformation of one mineral into
another by a process called diagenesis. Mineral diagenesis can be an indirect
effect of aerobic and anaerobic microbial metabolism (Ehrlich, 2002; Kurek,
2002). The formation of a new mineral can be resulted from a chemical reaction
between a product of microbial dissolution of a mineral and appropriate cations
present in the environment.
B. FINE-GRAINED MINERAL DEVELOPMENT
Geochemical modeling of metal speciation and transport is only beginning to

include bacteria as geochemically active surfaces. The physical and chemical
characteristics of bacteria, such as their large surface area-to-volume ratio, serve
to increase the metal-binding capacity of their charged surfaces leading to
precipitation and formation of mineral phases on their cell walls or other surface
polymers (McLean et al., 2002).
The mechanisms by which bacteria initiate the formation of minerals in bulk
solution vary widely between species. There may be a combination of
biochemical and surface-mediated reactions during the process. Bacteria
surface layers may passively adsorb and indirectly serve as a nucleation
template. Bacteria can also more directly initiate mineral precipitation by
producing reactive compounds (e.g., enzymes, metallothioneins, siderophores),
which bind metals or catalyze metal transformations. Further,
bacteria can instigate the spontaneous precipitation of metals by altering the
chemistry of their microenvironments (Beveridge et al., 1997; Douglas and
Beveridge, 1998).
Metal complexing ligands such as metallothioneins and siderophores serve to
sequester metals from the environment for incorporation into cellular
components (McLean et al., 2002). Reactive inorganic ligands such as sulfide
420 P. M. HUANG
and phosphate may also be produced as cellular metabolic by-products. Sulfide

reacts with metal ions to form metal sulfides, a common reaction in anoxic
environments with sulfate-reducing bacteria. Phosphate pumped out of cells
can also react with metal ions to form cell-associated and extracellular
precipitates.
Another example of microbially mediated fine-grained mineral development
is the formation of Mn oxides. Microbial oxidation of Mn(II) is a major process
that can produce Mn oxide coatings on soil particles 105 times faster than abiotic
oxidation (Tebo et al., 1997). This microbially mediated formation is illustrated
in Fig. 11. Manganese oxides are highly reactive minerals and help restrict the
mobility of metals in soils and sediments through adsorption on their surfaces.
Biogenic Mn oxides formed by Leptothrix discophora SS-1 have
significantly higher Pb adsorption capacity and larger specific surface than
abiotically precipitated Mn oxides (Nelson et al., 1999). Therefore, bioformation
of minerals should have a significant bearing on remediation of metal
contamination.
Figure 11 Thin section of Leptothrix sp., which is precipitating manganese oxide on its
outermost structure called a sheath. The arrows point to the manganese mineral phase identified by
EDS. Scale bar ¼ 150 nm. From McLean et al. (2002).
VIII. INTERACTIONS OF SOIL MINERAL COLLOIDS

WITH ORGANIC SUBSTANCES AND MICROORGANISMS
IN RELATION TO SOIL STRUCTURE STABILITY
A. ORGANO-MINERAL COMPLEXES IN RELATION

TO SOIL STRUCTURE
The processes by which soil organic matter strengthens the bonds between soil
mineral particles are complicated. Root exudation and microbial action produce
organic compounds with a range of composition and molecular weights; these
compounds interact with the mineral particles, which also vary in size, shape,
crystallinity, and electric charge (Emerson et al., 1986). Interactions between soil
mineral particles, organic substances, and microorganisms can occur at many
different size scale, as these materials have a large size range in soils (Fig. 12).
Therefore, it is important to indicate the size scale being considered when
discussing soil structure and the mechanisms of its stabilization, because the
potential mechanisms for stabilization vary with aggregate size.
The adsorption of organic molecules such as microbially derived poly-
saccharides and other unaltered and altered biomolecules onto mineral surfaces
can enhance the stability of individual clay microstructures. Adsorption is also
important in binding together the clay microstructures and silt particles into small
microaggregates with 2 – 50 mm diameters and densities . 2.0 Mg m23
(Baldock, 2002). The high stability of these small microaggregates is
demonstrated by their resistance to ultrasonification. Many microaggregates
exist as pieces of fungal hyphae, bacteria or bacterial colonies coated with EPS
and clay minerals (Oades and Waters, 1991). The polysaccharides are present
throughout the matrix but concentrated in pores between clay microstructures.
Particulate organic matter (POM) are, important stabilizing agents at larger
size scales: large microaggregates and small macroaggregates (Jastrow and
Miller, 1997). Soil structure can be stabilized by POM through two mechanisms
related to its physical properties and its susceptibility to biological decomposition
(Baldock, 2002). POM can bridge the failure zones that exist between adjacent
stable aggregates. The bridging can result from a combination of binding to
aggregate surfaces, penetration through aggregates, and the formation of a
network capable of holding groups of aggregates together. The ability of POM to
stabilize soil structure by this bridging mechanism is controlled by its physical
size and morphology. POM can also enhance the stability of soil structure by
providing a substrate for microorganisms to enhance the production of fungal
hyphae and microbial metabolites such as polysaccharides. The two mechanisms
of structural stabilization by POM can operate independently or synergistically.
POM is often found at the core of microaggregates , 250 mm (Oades and
Waters, 1991). The size of the microaggregate formed depends on the size of
422 P. M. HUANG
Figure 12 Size scales associated with soil mineral particles, organic components, pores and
aggregations of mineral and organic components. The definitions of pore size have used those
developed by IUPAC (micropores , 2 nm, mesopores 2 – 50 nm and macropores . 50 nm).
Alternatively, the pore sizes corresponding to the lower (Cm ¼ 2 1500 kPa) and upper
(Cm ¼ 2100 kPa) limits of water availability to plants may be used to define the boundaries
between the different classes of pore size. Cm is soil water metric potential. From Baldock (2002).
Figure 13 A conceptual model of aggregate hierarchy in soils where organic materials play an
important role in the stabilization of aggregates. From Jastrow and Miller (1997).
the POM and the nature and amount of binding materials secreted by
microorganisms as decomposition proceeds. While POM continues to provide
a substrate for microorganisms, the production of biochemical aggregating agents
continues, and structural stability is maintained.
Mechanisms of stabilization of soil structure can operate over larger distances
to bind microaggregates together to form macroaggregates as illustrated in
Fig. 12. In view of the distances involved, the stabilization of macroaggregates is
related to the presence of nonliving POM capable of spanning distances
. 100 mm or the existence of a network of fungal hyphae and plant roots that
physically enmeshes microaggregates (Fig. 13). The death of roots and hyphae
growing within and through macroaggregates results in the formation of bio-
chemical binding agents capable of stabilizing the structure of macroaggregates.
B. DYNAMICS OF AGGREGATE TURNOVER
Except for the biologically unavailable char/charcoal fraction, all components

of soil organic matter contribute to the stabilization of soil structure in mineral
soils. The diverse range of living and nonliving components varying in size,
chemical composition, and extent of decomposition provides soil organic matter
with the capacity to stabilize aggregations of soil particles at size scales varying
424 P. M. HUANG
over nine orders of magnitude (Fig. 12). Therefore, it is essential to maintain the
proper balance of components of soil organic matter to ensure the stability of the
entire soil matrix.
Since organic matter responsible for the stabilization of soil structure is not
inert and thus subject to decomposition, aggregation is a dynamic process in soils
(Baldbok, 2002). Biological activity has the potential not only to stabilize soil
structure through the production of organic substances capable of binding soil
particles, but also to destabilize soil structure by decomposing organic binding
agents. The balance between these two processes dictates the level of soil
structural stability. A continual addition of organic matter is essential to ensure
that the contents of each type of aggregating agent are adequate to maintain
structural stability of soils. Therefore, management that provides an ongoing
addition of organic materials has the greatest potential to maintain or enhance soil
structural stability. The model proposed by Golchin et al. (1997) can be used to
relate the physical and chemical properties of soil organic materials and their
distribution and cycling to the stabilization of soil structural form (Fig. 14). The
major organic materials contributing to the structural stability of soils are
perceived to be free and occluded POM fractions and the biomolecules
synthesized and exuded by roots, mycorrhizal fungi, and other microorganisms
in soils. This model depicts the dynamics of soil aggregation and is most
applicable to the aggregation of soils with appreciable contents of clays that
protect aggregating agents from rapid decomposition (Baldock and Skjemstad,
2000). In sandy soils, the protection mechanisms that stabilize organic binding
agents from microbial attack are less effective. Therefore, aggregation in sandy
soils is a much more dynamic process. The large size of the mineral particles and
pores in sandy soils reduces the effectiveness of polysaccharides and other
unaltered and altered biomolecules to stabilize soil structure. These biomolecules
are important to the adherence of soil particles to plant roots and fungal hyphae
and soil microorganisms to mineral particles. However, in sandy soils, they do
not have the capacity to span the distances between the large primary particles
and stabilize soil structure.
In the model depicting the dynamics of soil aggregation (Fig. 14), freshly
deposited POM is typically present as the free POM fraction of soil organic
matter. The binding of soil particles to free POM is limited. Therefore, POM has
little direct influence on soil structural stability (Stage 1, Fig. 14). Once free POM
is colonized by soil microorganisms and decomposition is initiated, mineral
particles adhere to the extracellular mucilage released by decomposer organisms
(Stage 2, Fig. 14). The remaining free POM becomes occluded within a new
macroaggregate. Further decomposition of occluded POM results in a decrease in
the size of POM through preferential utilization of exposed POM that has not
been encapsulated with mineral particles. The ability of pieces of POM to
maintain the stability of the original macroaggregate decreases and the
macroaggregate deteriorates into smaller microaggregates on exposure to
SOIL MINERAL – ORGANIC MATTER– MICROORGANISM
Figure 14 A model depicting the dynamics of soil aggregation and the roles of particulate organic matter (POM) and microbial metabolites in the stabilization
425
of soil aggregates. From Baldock (2002).
426 P. M. HUANG
disruptive forces (Stage 3, Fig. 14). Once carbohydrates are removed from the
microaggregate POM, the more recalcitron cores are no longer capable of
maintaining the microaggregate stability. The microaggregate then deteriorates
into aggregations of soil particles bound together with remnant microbial
mucilage and other unaltered and altered biomolecules (Stages 4 and 5, Fig. 14).
Although mechanisms exist that can contribute to some protection against
biological attack, all components of soil organic matter involved in aggregation
decompose with time. A continuous input of organic materials, mainly from plant
production is, thus, essential to maintain or enhance the stability of soil structure.
IX. INFLUENCE OF MINERAL COLLOIDS ON

BIOGEOCHEMICAL CYCLING OF C, N, P, AND S IN SOIL
A. ROLE OF MINERAL COLLOIDS IN SOIL ORGANIC MATTER

STORAGE AND TURNOVER
Most of the organic C in soils is degraded to inorganic forms slowly, on

timescales from centuries to millennia. Soil minerals are known to play a
stabilizing role in soil organic matter. The interaction of Al and Fe with humus is
of primary importance in the determination of the content of organic matter in
tropical and temperate soils (Wada, 1995). The Al and Fe that complex and
stabilize humus against microbial decomposition are released from soil minerals
during soil formation and present as polymeric hydroxy cations. Their supply
rates apparently control the content of soil organic matter to a great extent, as
demonstrated by the relationship between pyrophosphate-extractable C and
pyrophosphate-extractable Al plus Fe (Fig. 15).
Torn et al. (1997) used radiocarbon analyses to explore interactions between
soil mineralogy and soil organic C along two natural gradients — of soil-age
and of climate — in volcanic soil environments. The total stock of organic C in
soil increases with substrate age up to 150 kyr, to 60 kg m22, and then
decreases to 31 kg m22 at the oldest site (Fig. 16a). Carbon inventory in the
surface horizons (O and A horizons) reaches its maximum at 20 kyr, and
varies less with age than did deep soil C. Most of the decrease in soil organic
C stored in older substrates is attributed to faster turnover of soil C, rather than
to a decrease in plant productivity. The loss of net primary productivity is too
small to explain the loss of soil C. The K 14C (0/00), which is the turnover time
of soil organic matter (the reciprocal of the decomposition rate), shows that the
surface horizons are dominated by fast-cycling organic matter (Fig. 16b). The
K 14C (0/00) values also show the turnover of soil organic matter decreases
with substrate time up to 150 kyr and then increases at the oldest site. During
the first 150 kyr of soil development, the volcanic parent material weathers to
Figure 15 Relationship between pyrophosphate-extractable C and pyrophosphate-extractable

(Al þ Fe) in Japanese Andisols of different ages. From Wada and Higashi. (1976).
metastable noncrystalline minerals. Thereafter, the amount of noncrystalline

minerals increases up to 150 kyr and then declines with greater age (Fig. 16c).
In contrast, the amount of more stable crystalline minerals remains low until
150 kyr, then increases steeply (Fig. 16d). Soil organic carbon content follows
a similar trend, accumulating to a maximum after 150 kgr, and then decreasing
by 50% over the next four million years (Fig. 16a). The abundance of
noncrystalline minerals accounts for . 40% of the variation in organic C
content across all the mineral horizons, substrate age, and soil orders
(excluding the O and A horizons dominated by fast-cycling plant litter).
Noncrystalline minerals also strongly influence turnover of soil organic matter.
Organic matter D 14C is highly and negatively correlated with abundance of
noncrystalline minerals (R 2 ¼ 0.62) except in the oldest site, which has , 10%
noncrystalline minerals. In contrast, there is no discernible correlation between
the abundance of crystalline minerals and C content or turnover time across
sites. The amount of C stabilized per gram is much greater for noncrystalline
than for crystalline minerals. A positive relationship between noncrystalline
minerals and organic C also exists in soils through the climate gradient. This
indicates that the accumulation and subsequent loss of organic matter are
largely driven by changes in the millennial scale cycling of mineral-stabilized
C, rather than by changes in the amount of fast-cycling organic matter or in net
primary productivity.
Kennedy et al. (2002) showed that 85% of variation in total organic C could be
explained by mineral surface area in a black shale deposit from two locations in
428 P. M. HUANG
Figure 16 Soil inventory of carbon in soil organic matter (SOM) (a), K14C of SOM (b),
noncrystalline minerals (c), and crystalline minerals (d) versus age of soil substrate. Filled circles,
total profile; filled triangles, surface (O and A) horizons. From Torn et al. (1997).
the late Cretaceous Western Interior Seaway, United States. This relation
suggests that adsorption of organic C compounds on mineral colloids plays a
fundamental role in the preservation of organic C. Their data also provided
evidence for organic matter within smectite interlayer, which is in accord with the
finding of Wang and Huang (1986).
Therefore, soil mineral –organic interactions are important in determining the

quantity of organic matter stored in soil, its turnover time, and atmosphere –
ecosystem carbon fluxes during long-term soil development. An understanding of
how minerals influence the dynamics of soil organic matter should yield
significant improvements to our understanding of the role of soils in the global
ion cycling.
B. DECOMPOSITION AND STABILIZATION OF C, N, P,

AND S IN RELATION TO PRIMARY AND
SECONDARY PARTICLES
Generally, more than 95% of the N and S and between 20 and 90% of the P
in surface soils are present in soil organic matter (Guggenberger and Haider,
2002). The close relationship between the organic forms of C, N, P, and S is
well established. The turnover of organic C is closely associated with the
dynamics of N, P, and S in soils. Soil mineral colloids exert a profound
influence on the stabilization and degradation of soil organic matter and its
associated nutrients. Chemical and physical interactions of minerals with soil
organic matter result predominantly in the stabilization of the organic
substances. However, minerals can also contribute to a more rapid cycling of
associated nutrients of soil organic matter. Some soil minerals such as birnessite
may enhance the cycling of the organically bound nutrients by catalytic
degradation of the organic substances (Wang and Huang, 1987, 1992, 2000b;
Huang, 2000a).
The release of the stored organic nutrients depends on the mean residence time
of soil organic matter associated with the mineral phase. The mean residence time
of soil organic matter varies widely with the type of the organomineral
associations and the spatial location within the aggregate structure of soil
(Table V). Some of soil organic matter of rapid turnover is probably related to
microbial residues, which are often loosely bound to layer silicates. In contrast,
soil organic matter of slow turnover is tightly bound to Al and Fe oxides,
hydroxides, and/or oxyhydroxides by formation of inner-sphere complexes. In
soil environments, primary particles (layer silicate clay minerals and Al and Fe
oxides, hydroxides, and/or oxyhydroxides) are usually clustered to larger
entities — the secondary soil particles or aggregates. The arrangement of primary
and secondary particles results in the formation of a series of pores, which range
in size from nanometers to centimeters and is defined as soil structure.
Aggregation and soil structure determine the pore-size distribution and, hence,
the accessibility of substrates to exoenzymes, microorganisms, and faunal
grazers. Pools I and II are generally related to plant fragments, divided into easily
available cell constituents and lignocelluloses at various degrees of degradation,
respectively. As these fragments are more degraded and reduced in size,
430 P. M. HUANG
Table V
Comparison of Estimated Mean Residence Times of Soil Organic Matter in Soil Physical
Fractions
Mean residence time (years)
Jenkinson and
Pool Rayner (1977) Parton et al. (1987) Buyanovsky et al. (1994) Carter (1996)
I Decomposable Metabolic Vegetative fragments Litter, 1– 3

plant material, plant residues, 2– 0.2 mm, 0.5–1
0.24 0.1– 1
II Resistant plant Structural plant Vegetative fragments Free POM
material, 3.33 residues, 1– 5 0.05–0.025 mm, 1–3 (light fraction), 1–15
III Soil biomass, Active SOM OM in aggregates Microbial biomass,
2.44 pool, 1–5 2– 1 mm, 1–4 0.1 –0.4
IV Physically Slow SOM OM in aggregates Intermicroaggregate
stabilized pool, 25–50 1– 0.1 mm, 2–10 OMa, 5–50
OM, 72
V Chemically Passive SOM OM in fine silt, ~400 Intramicroaggregate
stabilized OM, pool, 1000– 1500 OMb physically
2857 sequestered, 50– 1000
OM in fine clay, ~1000 Chemically sequestered,
1000–3000
From Guggenberger and Haider (2002).

a
Organic matter stored within macroaggregates but external to microaggregates; includes coarse
occluded POM and microbial organic matter.
b
Organic matter stored within microaggregates; includes fine occluded POM and microbial derived
organic matter.
the microbially available components are exhausted and the residues are more
resistant to degradation. Pool III consists of soil biomass and readily available
organic matter within large aggregates. These materials have a relatively
short mean residence time. On average, the Pools I–III account for about
20 – 30% of the total C in soil organic matter. These pools must be renewed
continuously by fresh plant residues to maintain a relatively constant nutrient
level and release by mineralization. The balance between decay and renewal
processes, which is sensitive to management, controls the availability of N, P, and
S. Pool IV is physically protected soil organic matter and is also affected by
cultivation, because physical disturbance such as ploughing destroys macro-
aggregates and large microaggregates. Pool V is chemically stabilized soil
organic matter, which has the longest mean residence time. This pool represents
50 – 70% of the total C in soil organic matter and is seldom affected by
management practices.
Organic substances may weather soil minerals (Robert and Berthelin,
1986; Tan, 1986) and impede crystallization of secondary minerals
(Schwertmann, 1966; Kwong and Huang, 1975; Inoue and Huang, 1984; Huang
and Violante, 1986; Schwertmann et al., 1986; Huang et al., 2002). Noncrystal-
line minerals have the ability to stabilize soil organic matter and, thus, reduce the
turnover of C, N, P, and S (Torn et al., 1997; Guggenberger and Haider, 2002).
Organic substances also act as binding agents to promote aggregation, which in
turn, reduce the turnover of these nutrient elements through occlusion by
minerals. Hence, in mutual interactions, noncrystalline minerals can retard the
organic matter from being biodegraded, which in turn, inhibits the transformation
of noncrystalline mineral phases into more stable crystalline minerals. Therefore,
there are distinct interactive mechanisms between soil minerals and organic
matter, which have direct effects on the cycling of C, N, P, and S in the
environment.
X. EFFECTS OF INTERACTIONS BETWEEN

MICROORGANISMS AND SOIL COLLOIDS ON THE
TRANSFORMATION OF ORGANIC POLLUTANTS
A. CATALYTIC TRANSFORMATIONS OF ORGANIC POLLUTANTS
Both enzymes and mineral colloids are involved in catalytic transformations

of organic pollutants in soil environments. Enzymes are biotic catalysts and
mineral colloids are abiotic catalysts. Frequently it is difficult to determine
whether an organic pollutant is transformed abiotically or biotically (Huang,
1990). In many cases, significant abiotic and biotic catalytic reactions take place
simultaneously (Huang and Bollag, 1999).
1. Biotic Catalysis
Microorganisms are the major source of extracellular cellular enzymes in

soils. However, considerable amounts of soil enzymes can also be derived
from plant debris, plant root exudates, and soil fauna. The degradation of
organic pollutants by extracellular enzymes is well documented (Dec et al.,
2002).
Malathion, for instance, is hydrolyzed to a monocarboxylic acid (via loss of an
ethyl group) by a carboxyl esterase isolated from soil (Getzin and Rosefield,
1971). The enzyme catalysis exhibits typical Michaelis-Menton kinetics.
Enzymes commonly occurring in soil, such as esterases, amidases, phosphatases,
and proteases catalyze the hydrolysis of the respective chemical bonds in
xenobiotic molecules. (Dec et al., 2002).
432 P. M. HUANG
Extracellular phenoloxidases have the ability to catalyze the transformation

of phenolic or anilinic compounds to their polymerized products (Bollag,
1992). For example, 2,4-dichlorophenol, an intermediate of 2,4-D (dichloro-
phenoxy acetic acid), is coupled to itself by a laccase of Rhizoctonia praticola
to yield dimeric to pentameric products. In the presence of this laccase,
2,4-dichlorophenol reacts with various humic constituents (e.g., syringic acid,
oricinol, or vanillin) to form a variety of hybrid oligomers. Various phenols
and anilines are also enzymatically incorporated to natural HAs (Bollag et al.,
1998). These oxidative coupling reactions were carried out only in vitro, there
are indications they may also occur in soils. Phenoloxidases are believed to
participate in humus formation by mediating the polymerization of the
monomer products of microbial metabolism. They are also implicated in
binding of xenobiotics to soil by catalysis of the oxidation of pollutants to free
radicals, followed by chemical coupling of oxidation products to humic
substances. Further, there are indications that participation of enzymes in
humification and binding may be triggered by the presence of toxic substrates
that induce the production of extracellular phenoloxidases (Dec and Bollag,
2000).
2. Abiotic Catalytic Transformation
Most of organic chemicals, including xenobiotics, exhibit a strong affinity to

humic substances. However, transformation of xenobiotics in terrestrial systems
is greatly influenced by mineral components of soil (Huang, 1995). Mineral
colloids, their high concentration in soil, large specific surface, and relatively
high charge density, contribute to the overall xenobiotic transformation at least as
much as does the organic matter.
Organic matter can induce surface-catalyzed reactions of adsorbed
pesticides, but theoretically it could also hinder the degradation of some
pesticides by decreasing both their availability to microbial attack and their
concentration in the soil solution (Huang and Bollag, 1999). The significance
of soil mineral-catalyzed abiotic transformation of xenobiotics in the
environments has become widely recognized only recently (Cheng, 1990;
Huang, 1990, 2000a). The processes of adsorption and abiotic degradation of
xenobiotics through the action of the surfaces of soil minerals vary with the
structural and surface properties of the minerals, saturating cations and
hydration status, molecular structures of xenobiotics, and associated environ-
mental factors.
Besides clay minerals with well-ordered crystalline structures, soils contain
mineral colloids such as metal oxides, aluminosilicates, and carbonates that
exist in poorly ordered forms (Huang, 1995). These SRO mineral colloids
may exist as discrete phases in soils. They may also form coatings on
crystalline minerals and organic matter and, thus, alter the surface chemistry
of soil components. These SRO mineral colloids appear to play a dual role in
humification and binding interactions between xenobiotics and soil. First of
all, they may be directly involved in the catalytic oxidation of organic
chemicals with the formation of organic residues (Chorover et al., 1999;
Karthikeyan et al., 1999). Secondary, they may influence the efficiency of
enzymes that contribute to the same reactions (Claus and Filip, 1990;
Gianfreda and Bollag, 1994; Huang et al., 1999). The contribution of these
SRO mineral colloids to the transformation of organic chemicals merits
increasing attention.
(a) Oxidative transformation of organic compounds by soil minerals. Soil
minerals play an important role in catalyzing the abiotic transformation of
phenolic compounds (Huang, 1990, 2000a). Abiotic catalysts include primary
minerals, layer silicates, metal oxides, hydroxides, oxyhydroxides, and poorly
crystalline aluminosilicates. They promote the transformation of phenolic
compounds through oxidative polymerization, ring cleavage, decarboxylation,
and/or dealkylation. The ability of soil inorganic constituents to catalyze the
transformation substantially varies with their structural configuration and
surface chemistry and the structure and functionality of the organics involved.
Manganese(IV) oxides are most powerful soil minerals in catalyzing abiotic
oxidation of phenolic compounds (Shindo and Huang, 1982, 1984; Huang,
2000a). Organic compounds such as 2,4-D and ethyl ether can be degraded by
the catalysis of birnessite (d-MnO2) (Chenney et al., 1996). These organic
compounds can be adsorbed on birnessite and rapidly oxidized, both
producing CO2 as a major product (Fig. 17), but by somewhat different
mechanisms. In the case of 2,4-D only, methanol-extractable Mn2þ is
detected. Therefore, the solid degradation of organochlorine herbicides can
occur by catalysis of birnessite, which is a common component in the
environment.
(b) Brønsted and Lewis acidity of mineral surfaces and hydrolysis of
organic compounds. Mineral surfaces may detoxify adsorbed organic
pollutants by catalysis through their ability to behave as Brønsted acids
and donate protons or to act as Lewis acids and accept electron pairs.
Brønsted acidity derives essentially from the dissociation of water molecules
coordinated to surface-bound cations. Therefore, this acidity is strongly
influenced by the hydration status and polarizing power of surface-bound and
structural cations on mineral colloids (Mortland, 1986). Certain organic
pollutants such as organophosphate and s-triazine pesticides can be degraded
by catalysis of mineral colloids through their surface acidity (Brown and White,
1969; Mingelgrin et al., 1977). This type of reaction deserves attention in a
wide range of other organic compounds in soil environments (McBride, 1994;
Huang, 2000a). Besides the Brønsted acidity, the Lewis acidity of metals
434 P. M. HUANG
Figure 17 Gas chromatographic measurements of CO2 and oxygen in the headspace of reaction
vials and 2,4-D at 1, 2, 3, and 4 h, respectively. The values indicate the mole percent of CO2 and
oxygen in the headspace gas. †, CO2 from air control; B, CO2 from ethyl ether control; O, CO2 from
0.3 mmol 2,4-D plus ethyl ether sample; W, oxygen from air control; A, oxygen from the ethyl ether
control; K, oxygen from 0.3 mmol 2,4-D plus ethyl ether. The error bars indicate the standard error for
two measurements. From Cheney et al. (1996).
such as Al and Fe exposed at the edges of minerals is important in mineral-

catalyzed hydrolysis reactions. Aluminum and Fe oxides in water and
especially in the dry state have the ability to catalyze organic hydrolysis
reactions, at least for those that are hydroxyl ion-catalyzed (Hoffmann, 1990).
Dissolved metals and metal-containing surfaces play a significant role in the
catalysis of the hydrolysis of organic pollutants. The ability of a metal ion to
catalyze the hydrolysis varies with its ability to complex with reactant
molecules and shift electron density and conformation in ways favorable to
the reaction (Hoffmann, 1980). Metals vary in their complex formation
constants, which reflect differences in metal – ligand bond strengths and
solvation forces (Stone and Torrents, 1995). The reactivity of hydrolyzable
organic pollutants arises from the presence of electrophilic (electron deficient)
sites within the molecules. The SN2 mechanism (neucleophilic substitution —
bimolecular) involves attack of electrophilic sites by OH2 or H2O, generation
of a higher number intermediate, subsequent elimination of the leaving group,
and the formation of a hydrolysis product (X and Y in this mechanism are
elements or functional groups attached to C (Stone and Torrent, 1995):
In the case of the SN1 mechanism (nucleophilic substitution — monomole-

cular), the reaction proceeds with the loss of the leaving groups to generate a
lower coordination number intermediate and then the generation of the
hydrolysis product by neucleophilic addition (X, Y, and Z in this mechanism
are elements and functional groups attached to P (Stone and Torrent, 1995):
Metal ions can catalyze hydrolysis in a way similar to acid catalysis. Metal ions
and protons coordinate to organic pollutants so that the electron density is shifted
away from the site of nucleophilic attack to facilitate the reaction. Protons have an
extremely high charge density and great polarizing power; metal catalysis is, thus,
insignificant in acidic conditions. By contrast, metal ions can readily coordinate
two or more ligand donor sites on a molecule and can greatly outnumber protons in
neutral and alkaline conditions. Many organic compounds are susceptible to
catalysis by metals exposed on the surfaces of metal oxides/hydroxides and
aluminosilicates. These include carboxylic acid ester, amides, anilides, phosphate-
containing esters, and other hydrolyzable compounds (Stone and Torrent, 1995;
Huang, 2000a). Surface catalysis by metals on minerals is observed when all
participating reactants are adsorbed to a significant extent and when rate constants
for the reactions at the mineral – water interface are comparable with or exceed rate
constants for the reactions in homogeneous solution.
(c) Catalytic effects of humic substances on the transformation of organic
pollutants. Soil organic matter and especially humic substances exert catalytic or
inhibiting effects on the abiotic hydrolysis of organic pollutants (Senesi and
Miano, 1995). In aqueous systems, HAs and FAs have the ability to enhance the
acid hydrolysis of phenoxy acetic acids and esters (Struif et al., 1975) and chloro-
s-triazines (Li and Felbeck, 1972; Khan, 1978) and to retard the alkaline hydrolysis
of the n-octyl ester of 2,4-D (2,4 DOE) (Perdue and Wolfe, 1982). The
mechanisms proposed to explain the catalytic effects of HAs and FAs on the
dechlorohydroxylation of the chloro-s-triazines, simazine, atrazine, and propazine
consists of the interaction through H-bonding between the surface carboxyl groups
of HAs and FAs and the heterocyclic nitrogen atoms of the triazine ring (Fig. 18).
This interaction would strengthen the electron-withdrawing effect of the electron-
deficient carbon bearing the chlorine atom and, thus, reduce the activation energy
barrier for the hydrolytic cleavage of the C –Cl bond and facilitate the replacement
of chlorine by the weakly nucleophilic water (Fig. 18). This mechanism would
also explain the correlation between the amount of chloro-s-triazines adsorbed
by soil organic matter and their chemical hydrolysis by soil catalysis.
436 P. M. HUANG
Figure 18 The mechanism of the catalytic effects of humic substances on the dechloro-
hydroxylation of the chloro-s-triazines. From Senesi and Miano (1995).
B. BINDING OF ORGANIC POLLUTANTS, ENZYMES,

AND MICROORGANISMS ON MINERAL AND HUMIC SURFACES
AND THE EFFECTS ON POLLUTANT BIOAVAILABILITY
1. Biodegradation of Bound Xenobiotics
The degradation of organic pollutants may be considerably reduced when they

are retained by soil colloids. The major reason for reduced biodegradation rates is
the diminished bioavailability of chemicals involved in binding processes
(Alexander, 1995). The availability of sorbed xenobiotics to microorganisms
varies with the chemical properties of the pollutant, the nature of the sorbent, the
mechanism of sorption, and the properties of the degradative organisms (Guerin
and Boyd, 1992).
For instance, phenol (a nonionic compound at neutral pH) adsorbed to stream
sediments is available to microorganisms, whereas cationic surfactants are
unavailable (Shimp and Young, 1988). The herbicide diquat sorbed to the
external exchange sites of nonexpansible clay mineral, kaolinite, is readily
accessible to microorganisms; however, diquat molecules entrapped in the
interlayer space of the expansible clay mineral, montmorillonite, escape
biodegradation (Weber and Coble, 1968).
The mechanism of sorption may change with residence time of xenobiotics
present in soil, leading to changes in their bioavailability (Alexander, 1995). For
example, a strain of pseudomonas mineralized as much as 45% of phenanthrene
during a 5-day incubation with the compound that was in soil for less than 1 day
(Hatzinger and Alexander, 1995). When phenanthrene was aged for 315 days,
only 5% of the compound was subject to microbial mineralization. The initial fast
adsorption of phenanthrene molecules on soil colloids apparently did not

constitute as big a barrier to biodegradation as the slow diffusion of the chemical
across organic matter and its sequestration in inaccessible microsites within the
soil matrix.
Contaminants sorbed by soil particles may be available to certain
microorganisms and unavailable to others. For instance, soil-sorbed naphthalene
was degraded by pseudomonas putida strain 17484, but it did not undergo
degradation by a gram-negative soil isolate, designated NP-Alk (Guerin and
Boyd, 1992; Crocker et al., 1995). Apparently, by directly mineralizing surface-
localized, labile sorbed naphthalene, the pseudomonas strain established
concentration gradients that promoted desorptive diffusion and mineralization
of nonlabile naphthalene portioned into soil organic matter. On the other hand,
naphthalene degradation of NP-Alk appeared to rely mainly on the passive
desorption of the surface-sorbed compound (Guerin and Boyd, 1992).
Some investigations have provided evidence contradictory to the general
principle that binding interactions should reduce the rates of the transformation of
xenobiotics (Dec et al., 2002). The enhanced microbial degradation of
xenobiotics in the presence of soil colloids is attributed to their increased
concentration in the vicinity of the microbial cells attached to organic and/or
mineral surfaces.
2. Adsorption of Microorganisms and Biodegradation of Xenobiotics
The attachment of microorganisms to solid surfaces is a common phenomenon

in soil environments. Microorganisms may use many different mechanisms to
achieve and maintain the attachment (Dec et al., 2002). Adsorption is critical to
the initial adhesion of microbial cells to solid surfaces. The next step may involve
special attachment structures such as pili or holdfasts. In most cases,
microorganisms manage to strengthen the attachment by means of extracellular
polymeric adhesives such as polysaccharides and proteins, which are frequently
their own metabolic products. Both electrostatic and hydrophobic interactions are
involved in microbial attachment (Fletcher, 1996).
Microorganisms are closely associated with mineral colloids in surface soils
(Huang and Bollag, 1999). Sorption of microorganisms may be extremely
extensive, especially in soils with high contents of clay and organic matter. Some
sorbed microorganisms may be released back to the solution, but some may be
sorbed irreversibly (Dec et al., 2002). The effect of adsorption of microorganisms
on their ability to degrade organic compounds is difficult to predict. The
microbial degradation may be enhanced or decreased, or may remain unchanged
as the cells undergo the adsorption to solid surfaces. This appears to depend on
the nature and properties of solid surfaces and substrates and the kinds of
438 P. M. HUANG
microorganisms. However, the utilization of sorbed organic compounds by

microorganisms that adhere to the same surfaces is a common phenomenon
(Alexander, 1999). For instance, naphthalene is metabolized in soil mainly by
sorbed degraders whose population is twice as great as that of the free degraders
(Di Grazia et al., 1990).
Knowledge on the effects of mineral colloids and other soil components on
microbial activity under field conditions is relatively limited (Haider, 1995).
These effects may vary with the type of soil components and microorganisms.
Microbial activity (growth, metabolism, and uptake of substrates) may be
stimulated by clays to a greater extent in bacteria than in fungi (Stotzky, 1986).
The impact of structural configuration and surface properties of soil mineral and
organic components on the activity of microorganisms and their ability to
degrade organic pollutants with different structure and functionality deserves
increasing attention for years to come.
3. Immobilization of Extracellular Enzymes and Xenobiotic

Transformation
Soil is a living system in which enzymes are present either free in solution or
bound onto mineral colloids, humic substances, and mineral colloid-humic
complexes (Section VI). This immobilization is the major factor that determines
their performance in soil environments. Therefore, immobilization of enzymes
may have a considerable impact on the rate of xenobiotic degradation.
After immobilization, enzymes show generally increased stability toward
physical, chemical, and biological denaturation (Ruggiero et al., 1996; Naidja
et al., 2000). Immobilization may have a considerable effect on the activity
and kinetic behavior of enzymes in xenobiotic transformation due to steric and
diffusional restrictions, direct involvement of the active sites in binding to the
support, and modified conformation of the immobilized enzymes (Matinek and
Mozhaev, 1985; Quiquampoix, 1987; Naidja et al., 2000, 2002). Conformational
alteration of enzyme molecules may result in the decrease in accessibility of the
active sites to the substrate, causing a setback in substrate transformation
(Goldstein, 1976). In general, the intimate association of enzymes with soil
colloids should have a negative effect on xenobiotic degradation (Ruggiero et al.,
1996). However, there are notable exceptions to this trend (Burns, 1986; Claus
and Filip, 1990; Wada et al., 1992; Gianfreda and Bollag, 1994; Naidja et al.,
2000).
Transformation and degradation of anthropogenic organic compounds,
including chlorinated phenols, detergents, pesticides, and other organic
pollutants can be mediated by various enzymes (Munnecke et al., 1982; Bollag,
1992; Naidja et al., 2000). Enzymes have great potential for in situ treatment
of environmental pollutants in soils and waters. The lifetime and stability of
enzymes can be increased by immobilization on soil colloids (Burns, 1986; Dec

et al., 2002). Enzymes immobilized by mineral colloids can be used more
efficiently, with a possible longer lifetime, higher stability, and possible
reusability (Ruggiero et al., 1989; Naidja et al., 2000). Further, many mineral
colloids are efficient abiotic catalysts in the transformation of organic pollutants
(Huang, 1990; Naidja et al., 2000; Dec et al., 2002). It has even been suggested
that bioremediation efforts should include the possibility of site-specific abiotic
reactions (Morra, 1996). However, information on the involvement of
immobilized enzymes in the transformation of organic pollutants in soil is
sparse (Dec et al., 2002).
To cope with the problems caused by increasing amounts of agrochemical and
industrial effluents disposed of in soil and water environments, a fundamental
understanding of enzyme-soil colloid interactions and their role in remediation of
anthropogenic organic compounds in the environment is essential for the
development of effective, durable, and practical treatment methods. A thorough
understanding of the use of immobilized enzymes in remediation of contaminated
soil environments is needed. However, the mechanisms of immobilization of
enzymes on mineral and organic colloids and the resultant changes of protein
conformation of enzymes remain obscure. A wide range of soil minerals have yet
to be studied that may have a high capability to immobilize enzymes, adsorb
substrates and reaction products and be catalytically active in the transformation
and degradation of organic pollutants. Further, the role of abiotic catalysis of the
mineral supports (Naidja and Huang, 1996), the hidden half in the enzyme-
mineral colloid complexes, is still at an early stage and deserves an intensive
investigation. Although coating of the clay surfaces with Al hydroxides may
improve their ability to immobilize enzymes and in some cases enhance their
activity (Naidja et al., 1997), the effect of the nature of coatings of Al and Fe
hydro(xides) and oxyhydroxides on the activity of immobilized enzymes is still
unknown.
In bioremediation processes, incorporation of xenobiotics into humus through
oxidative coupling reactions mediated by enzymes, mineral colloids, or mineral
colloid-immobilized enzymes decreases the mobility and toxicity of phenolic
pollutants (Naidja et al., 2000). Researchers should continue to develop new
methods for maximizing the binding process of xenobiotics to humus, for
instance, through the use of immobilized enzymes or abiotic catalysis or the
addition of copolymerizing agents (Bollag et al., 1992). The effect of xenobiotics
on the activity of mineral colloid-immobilized enzymes also warrants in-depth
studies.
To date, the knowledge on detailed surface structures of the enzyme –mineral
colloid and enzyme – humus complexes at the molecular level still needs to be
advanced. Atomic force and scanning tunneling microscopies and synchrotron-
based spectroscopic methods such as extended X-ray absorption fine structure
(EXAFS) and X-ray absorption near edge fine structure (XANES) can be used to
440 P. M. HUANG
better elucidate the mechanisms of the formation of enzyme– mineral colloid and
enzyme – humus complexes and their catalysis in the transformation of
anthropogenic organic compounds. More research is needed to uncover new
enzymes and new mineral composite supports to be used for catalytic degradation
of a wide range of industrial and agricultural pollutants in soil and water
environments.
XI. IMPACT OF INTERACTIONS OF

PHYSICOCHEMICAL, BIOCHEMICAL, AND BIOLOGICAL
PROCESSES ON METAL TRANSFORMATION
The transformation of metals in the environment is influenced by interactions

of physicochemical, biochemical, and biological processes. The impacts of these
interactive processes on metal transformation are especially important in the soil
rhizosphere where the kinds and concentrations of substrates are different from
those of the bulk soil because of root exudation (McLaughlin et al., 1998; Huang
and Germida, 2002). This leads to colonization by different populations of
bacteria, fungi, protozoa, and nematodes. Plant – microbe interactions result in
intense biological processes in the rhizosphere. These interactions, in turn, affect
physicochemical reactions in the rhizosphere. Physicochemical properties that
can be different in the rhizosphere include acidity, concentration of complexing
biomolecules, redox potential, ionic strength, moisture, and nutrient status. The
total rhizosphere environment is governed by an interacting trinity of the soil, the
plant, and the organisms associated with the root (Lynch, 1990a,b). Therefore,
reactions and processes in the soil rhizosphere, which is the bottleneck of metal
contamination of the terrestrial food chain, can only be interpreted satisfactorily
with interdisciplinary approaches. Much of the research on physicochemical
reactions of metal transformation in soil has used well-defined model systems
that simulate bulk soil characteristics. The impacts of interactions of
physicochemical, biochemical, and biological processes in soils on metal
transformation and bioavailability, food chain contamination, and ecosystem
health merit increasing attention.
A. REDOX REACTIONS OF METALS
Redox reactions are important in controlling the chemical speciation of a

number of contaminant metals, notably As, Se, Cr, Pu, Co, Pb, Ni, and Cu
(Oscarson et al., 1981; Bartlett and James 1993; Alloway, 1995; Myneni et al.,
1997; Huang, 2000a; Sparks, 2002). Redox reactions also are important in
controlling the transformation and reactivity of Mn and Fe oxides in soils, which
have enormous capacities to adsorb metal pollutants and are major sinks of these
pollutants (Huang and Germida, 2002). Further, reduction of sulfate to sulfide in
anaerobic environments also affects the transformation, solubility, and
availability of metal pollutants through the formation of highly insoluble metal
sulfides (Alloway, 1995).
The transformation, mobility, bioavailability, and toxicity of metals, which
exist in more than one oxidation state, are significantly influenced by redox
reactions. Massechelyn and Patrick (1994) have summarized the critical redox
potentials for the transformation of some metal contaminants in soils. There has
been little study of how changes in soil redox potential in the rhizosphere could
affect the chemistry of metal contamination. However, the creation of an oxidized
zone adjacent to the plant root in wetland soils has been identified as one process
affecting the chemistry of Zn, Cu, and As in soils. In wetland soils, it has been
well established that steep gradients in redox potentials develop around plant
roots as a result of CO2 release from the roots. This process is reflected in
precipitation of FeOOH (iron plaque) on the roots (Otte et al., 1989; Kirk and
Bajita, 1995). Compared with the surrounding soil, these Fe-rich plaques on the
roots of the salt-marsh plant, Aster tripolium, are enriched in Zn and Cu (Otte
et al., 1989). Zinc also accumulates in the rhizosphere of rice (Oryza sativa L.),
which is the result of the formation of a zone of oxidation of Fe2þ to Fe3þ
adjacent to the roots (Kirk and Bajita, 1995). Zinc concentration in red roots (with
iron plaque) is higher than in white roots; a positive effect of the Fe concentration
on the root surface, up to a certain level, on Zn uptake into the xylem fluid has
been demonstrated (Otte et al., 1989). Above this level of Fe coating, Zn uptake
by the plant is reduced, which is attributed to complete coating of the root surface
by FeOOH and blocking of absorption sites.
In reduced conditions, As is mobilized as a result of reduction of Fe and Mn
oxides and reduction of As(V) to As(III). However, in the rhizosphere in
wetlands, As is immobilized because of the oxidation to As(V) and adsorption to
FeOOH (Otte et al., 1991). Therefore, As has been found to accumulate in the
rhizosphere of many plants, but most of the As is likely to be retained on the root
surface (Otte et al., 1991, 1995).
B. COMPLEXATION REACTIONS OF METALS
A series of complexation reactions in the soil solution affect metal

transformation in the rhizosphere. Complexation reactions of metals with ligands
in the soil solution are significant in determining the chemical behavior,
availability, and toxicity of metals in the rhizosphere. In view of the occurrence
of organic ligands in the rhizosphere due to root exudates and microbial
metabolites (McLaughlin et al., 1998; Marschner, 1998) and the stability
constants of the complexes of metals with these ligands (NIST, 1997), a large
442 P. M. HUANG
fraction of the soluble metal ions in the soil solution may actually be complexed
with a series of organic ligands commonly present in the rhizosphere.
The study of metal speciation in the soil solution has been encouraged by the
free metal ion hypothesis in environmental toxicology (Lund, 1990). This
hypothesis states that the toxicity or bioavailability of a metal is related to the
activity of the free aquo ion. Although this hypothesis is gaining popularity in
studies of soil – plant relations (Parker et al., 1995), some evidence is now
emerging that free metal ion hypothesis may not be valid in all situations (Tessier
and Turner, 1995). At the same free metal activity, plant uptake of metals varies
with the kinds of chelators present in solution. Further, given the same chelate,
total metal concentration in solution affects plant uptake of metal. Either kinetic
limitations to dissociation of the chelate or uptake of the intact chelate could
explain these observations (Laurie et al., 1991).
Seasonal changes of the concentrations of such metals as Cu, Mn, Zn, and Co
in the rhizosphere are related to the presence of complexing agents of biological
origin (Nielsen, 1976; Linehan et al., 1989). Krishnamurti et al. (1996) reported
variation in pH and the cadmium availability index (CAI) of the bulk and
rhizosphere soils after 2 weeks of crop growth. The pH of the rhizosphere soil
was lower than that of the corresponding bulk soil, and the CAI values were
higher, indicating that more Cd was complexed with the low-molecular-weight
organic acids (LMWOAs) at the soil-root interface. The plant and prolific
microbial activity is expected to result in increased amounts of LMWOAs at the
soil –root interface. Therefore, a larger fraction of the metal contaminant will be
in a complexed and usually soluble form in the rhizosphere than in the bulk soil.
C. ADSORPTION – DESORPTION REACTIONS

OF METALS
The transformation of metals is significantly influenced by adsorption –

desorption reactions in soil environments. The reactions are affected by
biochemical and biological processes, which should be especially important in
the rhizosphere. Few studies have investigated the adsorption –desorption
reactions in the rhizosphere. Most studies examined these reactions in bulk
soils and under modifying conditions to simulate rhizosphere conditions.
Root exudates and microbial metabolites include a series of LMWOAs, which
are capable of forming complexes with metal ions (Robert and Berthelin, 1986;
Stevenson, 1994). There is increasing evidence to show that LMWOAs may
modify the transformation and dynamics of metals in soils, especially in the
rhizosphere (McLaughlin et al., 1998). Krishnamurti et al. (1997) reported that
the increase in Cd release in the presence of LMWOAs can be explained by the
surface complexation of the particulate-bound Cd in soil with LMWOAs, which
is reflected in the increase in the release of Cd from the soils with the increase in
the stability constant of Cd –LMWOA complexes (Fig. 19). Further, the rate
coefficients of the Cd release from the soils, calculated from the parabolic
diffusion equation are substantially influenced by LMWOAs (Table VI). The rate
coefficients of Cd release within each ligand vary from soil to soil. The
complexibility of soil Cd should vary with the nature of the particulate-bound Cd
of the soil. Therefore, the rate of Cd release by each ligand should vary with the
nature of the particulate-bound Cd of the soils.
The activity of Cd species in the soil solution of the soil – root interface
governs the amount of labile soil Cd. The importance of the metal –organic
complex-bound particulate Cd species in determining the bioavailability of soil
Cd has been shown by Krishnamurti et al. (1995a). The rate coefficients of Cd
release from the soils by LMWOAs (Table VI), which is a measure of the rate of
the release of soil Cd to soil solution through complexation of soil Cd with
LMWOAs, follow the same order as that of the CAI values of the soils
(Krishnamurti et al., 1995b). Furthermore, the amounts of the Cd released from
the soils by renewal of LMWOAs (Krishnamurti et al., 1997), which is an
indication of Cd sustaining power of the soils, also follow the same trend
Figure 19 Relationship between Cd released from the soils by selected low molecular weight
organic acids (LMWOAs) during the reaction period of 0.25 h and the logarithm of the stability
constant of Cd-LMWOA complexes. From Krishnamurti et al. (1997).
444 P. M. HUANG
Table VI
Overall Diffusion Coefficients of Cd Release from the Soils by 1022 M Low Molecular Weight
Organic Acids (LMWOAs) During the 0.25 to 1 h Reaction Period
Overall diffusion coefficient
Acetic acid Citric acid Fumaric acid Oxalic acid Succinic acid
Soil site (mmol (mmol (mmol (mmol (mmol
kg21 h20.5) kg21 h20.5) kg21 h20.5) kg21 h20.5) kg21 h20.5)
Luseland 0.112 ^ 0.010 0.200 ^ 0.015 0.199 ^ 0.012 0.079 ^ 0.006 0.090 ^ 0.005
Waitville 0.046 ^ 0.004 0.049 ^ 0.003 0.050 ^ 0.005 0.036 ^ 0.004 0.019 ^ 0.003
Jedbergh 0.036 ^ 0.005 0.196 ^ 0.009 0.041 ^ 0.003 0.026 ^ 0.004 0.009 ^ 0.003
From Krishnamurti et al. (1997).
as the CAI values of the soils. The data indicate the importance of the kinetics of
Cd release from soils by LMWOAs found in the rhizosphere in understanding Cd
availability.
More research is needed to understand the dynamics of the adsorption –
desorption of metals as influenced by biochemical and biological processes. Such
information is fundamental to understanding the pathways of the contamination
of metals to the terrestrial food chain and to developing innovative management
strategies to protect ecosystem health.
D. PRECIPITATION – DISSOLUTION REACTIONS

OF METALS
For many of the most abundant elements such as Al, Fe, and Mn, precipitation
of mineral forms is common and may control their solubility. For most of the
trace metals, direct precipitation from solution through homogeneous nucleation
appears to be less likely than adsorption –desorption by virtue of the low
concentrations of these metals in soil solutions in well aerated dryland soils
(McBride, 1989; Tiller, 1996; Christensen and Huang, 1999). In reduced
environments where the sulfide concentration is sufficiently high, precipitation of
trace metals as sulfides may have a significant role in metal transformation
(Robert and Berthelin, 1986).
In aerobic soils, although precipitation of trace metals through homogeneous
nucleation is not likely, heterogeneous nucleation may play a significant role in
metal transformation because of the presence of mineral, organic, and microbial
surfaces that catalyze the nucleation set of crystallization (Huang and Germida,
2002). The energy barrier to nucleation is reduced or removed by surfaces. This
is especially true in cases where there are crystallographic similarities between
the surface and the precipitating phase. This catalytic process reduces the extent
of supersaturation necessary for precipitation to occur. However, precipitation
reactions are often slower than adsorption – desorption reactions in soil
environments. Besides physicochemical reactions, metals have easy access to
bacterial surfaces through diffusion. Metal sorption and precipitation on
bacterial surfaces are interfacial effect. Surface metal concentrations frequently
exceed the stoichiometry expected per reactive chemical sites within the cell
walls (Beveridge, 1989; McLean et al., 2002). The sorption of metals can be so
great that precipitates can be formed and distinct metallic minerals are
eventually formed through biomineralization as discussed in section XI E in this
article.
Activities of free metal ions in the rhizosphere may be decreased through the
uptake by plants and microorganisms. Because concentrations of complexing
organic ligands in the rhizosphere are higher than in bulk soils, metal
contaminants are substantially complexed with organic ligands. And activities
of free metal ions should, thus, be decreased further. Therefore, compared with
bulk soils, the activities of trace metal ions in the soil solution of the rhizosphere
in aerobic dryland soils appear to be even less controlled by precipitation through
homogeneous nucleation. However, in the rhizosphere, precipitation of metals
through heterogeneous nucleation on microbial surfaces on one hand and metal
mobilization by biomolecules on the other hand as a result of intense biological
activity warrant in-depth research.
E. METAL SORPTION AND UPTAKE BY MICROORGANISMS

AND BIOMINERALIZATION
All microorganisms contain biopolymers, such as proteins, nucleic acids,

and polysaccharides, which provide sites where metal ions can bind (Hughes
and Poole, 1989). These binding sites include negatively charged groups such
as carboxylate, thiolate, phosphate, and groups such as amines, which
coordinate to the metal center through lone pairs of electrons. Because of the
ability of these biopolymers to bind metals, large concentrations of metals are
frequently associated, not only with living microbial biomass, but also with
dead cells (Berthelin et al., 1995). Many metals bind with various degrees of
tenacity to the largely anionic outer surface layers of microbial cells. Binding
of metals by microbial cells alters cell wall composition and induces
morphological, ultrastructural, and surface charge changes (Venkateswerlu
et al., 1989; Collins and Stotzky, 1996). Some metals are bound by cell walls
to a greater extent than by clay minerals (Table VII), indicating that bacterial
cell walls and membranes may act as foci for accumulation of metals in
soils.
446 P. M. HUANG
Table VII
Metals Bound by Native B. subtilis Walls, E. coli Envelopes, Kaolinite, and Smectite
Amount of metal bound (mmol g21) (oven-day weight)a
Metal Walls Envelopes Kaolinite Smectite
Ag 423 ^ 15 176 ^ 3 0.46 ^ 0.02 43 ^ 0.3

Cu 530 ^ 13 172 ^ 9 5 ^ 0.03 197 ^ 4
Ni 654 ^ 25 190 ^ 3 4 ^ 0.2 173 ^ 10
Cd 683 ^ 19 221 ^ 6 6 ^ 0.2 1 ^ 0.02
Pb 543 ^ 11 254 ^ 5 3 ^ 0.2 118 ^ 6
Zn 973 ^ 13 529 ^ 32 37 ^ 1 65 ^ 2
Cr 435 ^ 37 102 ^ 2 8 ^ 0.5 39 ^ 5
From Walker et al. (1989).

a
The data represent the average of three to five determinations for each sample from duplicate
experiments and the standard error.
Microbial biomineralization (the formation of minerals by microorganisms)

is another important activity of microorganisms that is now being defined
(Beveridge, 1989). Its scope is much larger than initially thought, as it
involves metal transformation and the development of fine-grain minerals of
tremendous range and kind (cf. Section VIIB). Research on biomineralization
indicates that specific molecular interactions at inorganic– organic interfaces
can result in the controlled nucleation and growth of inorganic crystals
(Mann et al., 1993). A central tenet of biomineralization is that the
nucleation, growth, morphology, and aggregation (assembly) of the inorganic
crystals are regulated by organized assemblies of organic macromolecules,
‘the organic matrix’. Control over the crystallochemical properties of the
biominerals is achieved by specific processes involving molecular recognition
at inorganic-organic interfaces. Electrostatic binding or association, geometric
matching (epitaxis), and stereochemical correspondence are important in these
recognitions. The subtle differences in the kinetics of these recognition
processes on different crystal faces lead to specific changes in crystal
morphology. Biomineralization has global consequences in dynamics,
toxicity, and fate of metal pollutants.
Compared with any other life form, bacteria may have a greater capacity to
sorb and precipitate metals from solution resulting in mineral formation as they
have the highest surface area to volume ratio (Beveridge, 1989). Although in
most environments, soluble metal ions are present in low concentrations,
bacterial cells have a remarkable ability to concentrate metal ions from solu-
tions. The research on the pH-dependent binding of metal ions to bacterial cell
walls and the effect on metal bioavailability is currently emerging (Huang and
Bollag, 1999).
F. METAL TRANSFORMATION BY MICROBIAL EXCRETIONS

AND MYCORRHIZAL INFECTION
Microorganisms have a range of metal transport systems, which are often

highly specific for certain metals and capable of accumulating metals against
large concentration gradients (Kurek, 2002; cf. Section VIIA). Certain
microorganisms synthesize compounds that bind specific metals with high
affinity (Lynch, 1990a,b). For instance, some microorganisms make Fe-binding
siderophores, which are biomolecules usually a phenolate or hydroxamate ligand
(Neilands, 1981). Other microorganisms produce compounds called metallo-
thioneins that are small cystine rich protein that strongly bind Cd, Cu, and Zn.
Ligands of this type and related biomolecules are of significance, because they
influence the transformation, transport, bioavailability, and toxicity of metals in
soils especially in the rhizosphere.
Most plants in natural habitats form association with mycorrhizae.
Arbuscular mycorrhizal fungi (AMF) are obligate symbionts, and infection
of plant roots exerts a metabolic load on the host plant (Reid, 1990). During
infection and colonization of host plant roots, mycorrhizal fungi produce
mycelium inside root cortical cells. They may form storage structures termed
vesicles, and they also may form other structures which are referred to
arbuscules and serve as the site of ion exchange between the host plant and
the mycorrhizal fungus. The fungi also form extra cellular hyphae that
penetrate out of the root and explore the soil in search of nutrients including
metals. Therefore, mycorrhizal infection may expand the volume of soil the
root can penetrate. Uptake of metals may, thus, be facilitated by mycorrhizae
because these elements are diffusion limited in soils.
Therefore, the effect of mycorrhizal infection on metal uptake depends on
the ability of the fungal symbiont to absorb metals and to transfer them to the
symbiotic roots through extensive vegetative mycelium. Further, mycorrhizal
fungi release LMWOAs (e.g., oxalic acid) into soils, which enhance the
solubilization of particulate-bound metals and might facilitate uptake of metals.
However, there are considerable gaps in knowledge on the mechanisms fungi
influence the transformation and uptake of metals.
XII. FORESEEABLE IMPACTS OF SOIL

MINERAL– ORGANIC MATTER– MICROORGANISM
INTERACTIONS ON THE ECOSYSTEM
Soil is a focal point of the ecosystem (Odum, 1989; Coleman, et al., 1998).
Physical, chemical, and biological processes are interacting in soil environments,
and are governed by soil mineral – organic matter – microorganism interactions.
448 P. M. HUANG
Therefore, interactions of these three major solid components of soils have

enormous impacts on reactions and processes critical to ecosystem health
(Huang, 2002).
A. GLOBAL ION CYCLING
Major biogeochemical transformations of elements include the cycling of

C, N, P, S, and metals (Butcher et al., 1992), which should be very much
influenced by interactions of soil minerals with organic matter and
microorganisms. Transformation of soil organic matter are closely related to
the mobilization of C, N, P, and S into the soil solution or their immobilization
from the solution and to the release or fixation of trace gases such as CH4,
CO2, OCS (carboxide sulfide), H2 NO2, and NO. These microbially mediated
processes are the basis for the interrelationship of the C, N, P, and S cycling
within the soil-plant system (McGill and Cole, 1981). Interactions of soil
minerals with microorganisms and organic components have an important role
in influencing the stability and degradation of soil organic matter and its
associated nutrients (Guggenburger and Haider, 2002) directly affecting the
global cycling of C, N, P, and S.
Metals are part of natural biogeochemical cycles. One of the character-
istics of the cycle of metal mobilization and deposition is that the form of the
metal is changed. This change in speciation of a metal has a profound effect
on its fate and impact on ecosystem health (Benjamin and Honeyman, 1992;
Hayes and Traina, 1998). Metals are found in the environment in solid,
solution, and gaseous phases, associated with thousands of different
compounds. These associations often reflect the affinity of metal ions for
other atoms with free electron pairs, particularly O, N, and S. The critical
processes controlling global metal cycling are adsorption – desorption,
precipitation –dissolution, complexation, and volatilization (Benjamin and
Honeyman, 1992; Hayes and Traina, 1998; Kurek, 2002; McLean et al.,
2002). Transport in solution or aqueous suspension is a major mechanism for
metal movement in the ecosystem. This transport process is greatly influenced
by adsorption – desorption on surfaces of minerals and organic matter,
precipitation –dissolution especially in reduced environments where sulfide
concentration is sufficiently high, and a series of inorganic and organic
complexation reactions in dissolved and particulate phases. Through their
effect on the chemical environment in soils and sediments, microorganisms
can help dissolve, complex, or precipitate metals and can also directly
mediate reactions involving metals, such as initiation of fine-grain mineral
formation.
Transport of particles suspended in the air is an important process for
transporting many metals to regions far from their sources. A few metals, most
notably Hg, can exist as gases at ambient temperatures. In the case of Hg,
reduction of Hg2þ to Hg8 and alkylation to form methyl- or dimethylmercury can
result in the loss of Hg from the aqueous phase. Microorganisms can also convert
the methylated forms to Hg8, which is more volatile and less toxic. Several other
metals such as As and Se also form organometallic compounds, which can be
mediated by microorganisms (Gadd, 1993). These volatile organometallic
compounds can dominate transport of the metal in local environments. However,
mediation of alkylation of metals such as Hg by bacteria is substantially
influenced by Hg speciation on surfaces of mineral colloids (Table VIII). Soil
minerals, organic matter, and microorganisms have their respective roles in
influencing metal speciation, toxicity, and cycling.
B. GLOBAL CLIMATE CHANGE
Many biogeochemical and physical processes are involved in determining the

climate of the Earth (Charlson et al., 1992). The biogeochemical cycles of C, N,
and S are central to the radiative properties of the atmosphere. Carbon and N form
radiatively important gases. Sulfur is a crucial component of clouds and most
aerosols. The cycles of C and N contribute to the long-wave radiative properties,
Table VIII
Biomethylation of Hg(II) Adsorbed on Mineral Colloids Common in Freshwater Sediments, by
P. fluorescens Isolate BPL85-48 during a 25 h Incubation Period
Optical density
Sample ID Hg(II)sourcea Absorbance at 530 nm RGIb CH3Hgþ (ng L21)c
Blankd – 0.551 1.28 a –

Controle Hg(NO3)2 0.430 1.00 d 32.86 ^ 0.67 a
KGa-1 Kaolinite 0.423 0.98 d 30.53 ^ 1.64 ab
STx-1 Montmorillonite 0.451 1.05 c 25.96 ^ 4.17 b
MnO2 Birnessite 0.478 1.11 b ND f c
From Farrell et al. (1998).

a
Total concentration of Hg(II), added as Hg(NO3)2 or in adsorbed form, was 6 mmol/100 ml.
b
Relative growth index ¼ optical density of colloid-amended medium/optical density of the control.
Values followed by the same letter are not significantly different (P # 0.05; least significant difference
test, LSD ¼ 0.04).
c
Values followed by the same letter are not significantly different (P # 0.05; least significant
difference test, LSD ¼ 5.30 ng CH3Hgþ L21).
d
Isolate grown in the M-IIY medium in the absence of Hg(II).
e
Isolate grown in M-IIY medium supplemented with Hg(NO3)2; total Hg(II) concentration ¼ 60 mM.
f
ND, not detectable.
450 P. M. HUANG
whereas the S cycle influences the short-wave radiative properties; all of the
cycles are severely perturbed by human activity (Table IX).
Transformation of C, N, and S in soils as influenced by land management
and the impact on their ion cycling and global climate change should not be
overlooked (Lal, 1998; Guggenburger and Haider, 2002). Jenkinson et al.
(1991) estimated the additional degradative effects on soil organic matter if
the global annual mean temperature rises during the next 60 years by 38C.
According to their estimate, about 100 Gt C(1 Gt ¼ 109 t) should be
additionally evolved as CO2 from soil organic matter (1600 Gt C). This will
increase the present CO2 concentration in the atmosphere by 14%, whereas
the combustion of fossil fuels (5.4 Gt C yr21) should add during this 60 year
period 330 Gt C to the atmosphere. Microbial by-products and resistant plant
residues adsorbed on soil particles have turnover times in terms of years.
Fulvic acids have turnover times in terms of hundreds of years, whereas HAs
and humins usually approach a few thousand years in their turnover time
(Paul and Van Vee, 1978). Although the HAs and humins constitute by far
the majority of the organic C in a system, they contribute only a small
proportion to the annual cycling of C within the soil because of their very
slow turnover rate. The undecomposed litter also includes the soil biomass
and microbial metabolites. These, together with the plant residues, constitute
the active fraction of organic matter that has a prominent role in the cycling
of elements such as C, N, and S annually. The influence of crystalline and
noncrystalline mineral colloids, which differ in structural configuration and
surface properties, on the biodegradation, turnover, and stabilization of
organic components, the cycling of C, N, and S, and the impact on global
climate changes merits close attention (Huang and Schnitzer, 1986; Berthelin
et al., 1999; Violante et al., 2002a, b).
Table IX
Radiatively Important Trace Species in the Atmosphere: Percent Change in Flux Measured
Relative to the Pre-industrial Age
Cycle change Species % Change
Long-wave absorbers
Water H2O (vapor) Not known
Carbon CO2 þ 50%
CH4 . þ 65%
Nitrogen N2 O þ 25%
Halogens Chlorofluorocarbons þ1
Short-wave reflectors
Sulfur SO22
4 þ 230%
From Charlson et al. (1992).

C. BIODIVERSITY
The functioning and stability of terrestrial ecosystems are determined by plant

biodiversity and species composition (Tilman et al., 1996; Hooper and Vitousek,
1997). However, the ecological mechanisms by which they are regulated and
maintained are not well understood. These mechanisms need to be identified to
ensure successful management for conservation and restoration of diverse natural
ecosystems.
Van der Heijden et al. (1998) recently reported that below-ground diversity of
AMF is a major factor contributing to the maintenance of plant biodiversity and
to ecosystem functioning. These results emphasize the need to protect AMF and
to consider these fungi in future management practices in order to maintain
diverse ecosystems. Research provides support for the view that flouristically rich
systems are more productive (Tilman et al., 1996), show greater stability under
stress (Tilman and Downing, 1994), and are more likely to provide alleviation of
global problems posed by atmospheric CO2 enrichment (Naeem et al., 1994).
A recognition of these properties, coupled with an increasing awareness that the
diversity of terrestrial vegetation systems is everywhere under stress, has
encouraged the research on mechanisms that determine and affect species
composition in plant communities.
The scientific evidence indicates that below-ground microbial diversity
substantially influences plant biodiversity and ecosystem variability. Further,
microbial events are significantly affected by surface-reactive particles (Stotzky,
1986, 2002; Burns, 2002). However, our knowledge on the effect of mineral –
organic component – microorganism interactions on below-ground microbial
diversity and the impact on above-ground biodiversity remains to be advanced.
D. BIOLOGICAL PRODUCTIVITY AND HUMAN NUTRITION
Soil is the life-sustaining material, which is the structurally porous and

biologically active medium that has developed on the continental land surface on
our planet. This material is created and continues to evolve through weathering
processes driven by biological, climatic, geological, topographic, and chrono-
logical influences. Soil is the cradle of agriculture including crop and animal
production and a fountain for sustaining human nutrition. However, it is the
fragile epidermis of the planet Earth that can sustain human nutrition or cause
starvation for humans depending on our management of soil resources.
Interactions of soil minerals with organic components and microorganisms
exert enormous influences on the transformation and dynamics of soil organic
matter (cf. Sections II, IV, IX), nutrient cycling and bioavailability (cf. Section IX;
Barber, 1995), efficacy and toxicity of pesticides (cf. Section X; Cheng, 1990),
microbial metabolic processes, growth, adhesion, and ecology (cf. Section V),
452 P. M. HUANG
enzymatic activity (cf. Section VI), and soil physical properties (cf. Section VIII).
Therefore, interactions of soil minerals with organic components and microor-
ganisms should have great impacts on plant nutrition and biological productivity of
soils.
In the rhizosphere, the narrow zone of soil surrounding a living plant that is
subject to influence by the root and its exudates, more intense microbial activity
and larger microbial populations occur than in the bulk soil (Waisel et al., 1996).
Up to 18% of the C assimilated through photosynthesis can be released from
roots. Since the rhizosphere is rich in root exudates, microbial population can be
10 – 100 times larger than the population in bulk soil (Sposito and Reginato,
1992). The rhizosphere typically extends away from the root for up to 2 mm, but
some organisms (e.g., fungi) may be stimulated up to 5 mm away. The
rhizosphere is bathed in root exudates and microbial metabolites. Both the
amounts and proportion of organic compounds of root exudates vary
substantially with plant species and cultivars. Further, the same plant cultivar
grown in different soils varies in the kind and amount of LMWOAs present in the
rhizosphere (Szmigielska et al., 1997). The chemistry and biology at the soil –
root interface, thus, differs significantly from soil to soil. The soil rhizosphere is
the bottleneck of the nutrient-feeding zone in soils. Therefore, the dynamics,
transformation, and bioavailability of nutrients are bound to be influenced greatly
by the chemistry and biology at the soil-root interface. The intense soil mineral –
organic component-microorganism interactions in the rhizosphere, thus, deserve
close attention in the development of innovative management strategies for land
resources to increase biological productivity and sustain human nutrition.
E. GEOMEDICINE
Geomedicine may be defined as the science dealing with the environmental

factors, which influence the geographical distribution of pathological and
nutritional problems relating to human and animal health (Låg, 1980).
Geomedicine is a young science with very old roots. Knowledge on soil science
is needed for solving many geomedical problems (Låg, 1994). Hunger and
malnutrition are serious issues for large groups of populations, especially in
developing countries. In addition to prevention of starvation, promotion of better
nourishment quality of food and feed is important. Pollution of the environment
and related health problems have increased rapidly in many industrialized
countries. Effects of many soil chemical, physical and biological factors and
processes on geomedical problems should be studied in depth as they impact on
the quality of vegetation and the food and feed produced.
Biogeochemistry of trace elements is greatly influenced by soil – plant –
microbe interactions (Chang et al., 2002). Therefore, soil mineral – organic
component – microorganism interactions deserve close attention in the

transformation, dynamics, and bioavailability of trace elements, many of which
are of concern to animal nutrition and health and well-being of humankind
(National Academy of Sciences, 1974, 1977). They include Se, Fe, I, Zn, Cu, Mn,
Mo, Cr, F, Co, Si, V, Ni, As, and Mg. One trace element may serve in one, several
or dozens of different metalloenzymes or tissue constituents.
All the trace elements appear to have vital metabolic roles in critical steps in
health risks built into human genetics. To facilitate fundamental understanding of
the linkage of trace elements in the soil-plant-environment-animal-human
systems, and to provide practical solutions to their deficiency and toxicity
problems, we need to advance the knowledge on soil mineral-organic
component-microorganism interactions affecting the transformation, dynamics,
and bioavailability of trace elements in soils and associated environments.
F. ECOTOXICOLOGY AND HUMAN HEALTH
Soil plays the central role as the organizer of terrestrial ecosystem (Coleman
et al., 1998). Furthermore, it may be perceived as the center of the ecosystem,
which evolves because of interactions of the lithosphere, hydrosphere,
atmosphere, and biosphere. A factor of central importance of soil to ecological
studies is that soils on a global scale have a range of characteristics, which enable
an enormous array of microorganisms, plants, animals, and humans to co-exist
and thrive. Humans have exploited the ability of soils to provide massive
amounts of food. About . 40% of the net primary production of the world is
exploited by humans (Vitrousek et al., 1986). The exploitation is increasing with
the addition of 87.6 million people to the global population every year, with the
rate addition steadily increasing (Brown and Flavin, 1996). The impact of
population and the accompanying intensification of agriculture and industrial-
ization on ecotoxicology and human health is, thus, of increasing concern.
Among the environmental compartments, about 90% of environmental
pollutants are bound with soil particles and 9% of the pollutants are bound
with aquatic sediments (Table X). These soil- and sediment-bound pollutants are
in dynamic equilibrium with the hydrosphere, atmosphere, and biosphere. Soil
mineral – organic component – microorganism interactions may enhance the
release of environmental pollutants from soils and sediments and, thus, pose a
threat to ecosystem health, including human health.
Land resource management to sustain the capacity of a soil to function within
ecosystem boundaries to improve biological productivity, maintain environmen-
tal quality, and promote plant and animal growth and human health is the central
role of soil scientists in contributing to human welfare. Interactions of soil
minerals with organic components and microorganisms have vital roles in
governing transformation, speciation, transport, bioavailability, and toxicity of
454 P. M. HUANG
Table X
Theoretical Pollutant Distribution in the Environment at Equilibrium
Compartment Concentrationa (mol m23) % Distribution in compartmenta
Air 4 £ 10210 0.35

Water 1027 0.01
Sediment 1022 9.10
Soil 1022 90.50
Aquatic biota 1027 0.01
From Crosby (1982).

a
Assumes approximately 100 kg of pollutant (MW 100) introduced into 10 km3 of the environment.
organics, metals, and other inorganics of agricultural and ecological concerns

(Fig. 20). These interactions are fundamental to understanding and regulating the
ecosystem at the molecular level.
G. BIOTECHNOLOGY
The probability and frequency of the transfer of genetic information, both

intra-and interspecifically, between microorganisms, especially bacteria, in soil
and other natural habitats have become important concerns (Stotzky, 1986,
2002). Although genetic transfer has been extensively studied in pure culture
systems, few studies have been conducted in natural microhabitats. More studies
are necessary to clarify the role of mineral colloids, organic matter, and organo-
mineral complexes in not only the survival of genetically engineered
microorganisms but also the transfer of genetic information by conjugation in
soil. Further, because of the potential hazards associated with the transfer of
genes — whether introduced accidentally or deliberately — in soil and contiguous
habitats (e.g., groundwater, surface water, atmosphere, plants, animals, human
beings), the importance of mineral – organic matter –microbe interactions in
influencing the fate of genetically engineered microbes in soil and related
environments should receive increasing attention.
Surface-active particles, primarily mineral colloids and humic substances, in
soil and other natural habitats are important in the persistence of biomolecules
which in the absence of such particles are rapidly degraded or inactivated by the
indigenous microbiota. Many biomolecules are important in the ecology,
biodiversity, evolution of microorganisms, environmental protection, and
biological control of pests (Stotzky, 2002). Insecticidal proteins produced by
Bacillus thuringiensis and by transgenic plants containing toxin genes, enzymes,
bacterial transforming DNA, and viruses are examples of the persistence to
biodegradation and the retention of biological activity of biomolecules when
Figure 20 Impact of environmental soil chemistry on agricultural sustainability and the

ecosystem health. Environmental soil chemistry is governed by mineral – organic matter –
microorganism interactions. From Huang (2000b).
bound on such particles. Therefore, these surface-active particles affect the

transfer of genetic information among bacteria by conjugation, transformation,
and transduction in soil. All these aspects affect ecosystem health.
The gains in food production provided by the Green Revolution have reached
their ceiling while world population continues to rise. The application of
456 P. M. HUANG
advances in plant biotechnology is going to be essential especially in developing

countries if farmers’ yields and yield ceilings are to be raised, excessive pesticide
use reduced, the nutrient value of basic foods increased, and farmers on less
favored lands provided with varieties better able to tolerate drought, salinity and
lack of soil nutrients (Conway and Toenniessen, 1999). However, when
biotechnology is applied at the field scale, the impact of soil mineral – organic
matter –microorganism interactions on the effectiveness of growing genetically
modified crops in different soil environments so as to achieve the objectives of
being more productive, yet less damaging to the ecosystem and consequently
more beneficial to human health remains to be uncovered. We need to assure that
biotechnology is applied to agriculture only when this can be done safely and
effectively in helping to achieve future food security and protect ecosystem
health for the world.
H. RISK ASSESSMENT
One of the great ecological challenges today is to determine if soil

contamination will lead to adverse effects on beneficial organisms. The mere
presence of a chemical in air, water, and soil, biota, or food can raise public fears
and a call for regulation of that chemical or its removal from the environment in
which it was found. Environmental monitoring may be able to show that the
chemical is a contaminant because concentrations are above some established
background. Contaminants have background levels in soils and biological tissues
as a result of their persistence and global dispersion. What is not known is if the
measured concentration of a contaminant can have an adverse effect on an
organism. To answer this question, we must use the tool of risk assessment, which
combines exposure assessment (how is the organism exposed to the chemical and
how much exposure is there?) and effect assessment (how does the chemical harm
the organism? how much exposure to the chemical is required to cause harm?).
Assessing exposure to contaminants in soil environments includes the
determination of the pathways to exposure (inhalation, drinking water, food or
direct soil ingestion) and the extent of contaminant transfer from soil to the
various pathway components (Logan, 1998). Possible soil pathways of human
contaminant exposure are presented in Table XI. In environmental exposure
assessment, a transfer coefficient (K) must be determined for each component of
the pathway, namely, the fraction of the component that is transferred from one
component to another (e.g., from soil to plant, soil to air, soil to surface water).
Such an approach is used in Mackay’s general fugacity model (Mackay and
Paterson, 1991). The transfer coefficient can be viewed as an availability index,
although other terms (extractability, solubility, accessibility, volatility) can also
be used. This will depend on the mechanism of transfer, which should be greatly
influenced by interactions of soil minerals with organic components and
Table XI
Pathways for Human Exposure to Soil-Borne Contaminants. The Most Exposed Individual
(MEI) of the Population is also Identified. Based on U.S. Environmental Protection Agency
Exposure Assessment for Land Applied Contaminants in Sewage Sludge (U.S. Environmental
Protection Agency, 1995)
Pathway MEI
Soil ! plant ! human Human lifetime plant ingestion; general population

Soil ! plant ! human Human lifetime plant ingestion; home gardener
Soil ! human Child
Soil ! plant ! animal ! human Human lifetime ingestion of animal products;
animals raised on forage
Soil ! animal ! human Human lifetime ingestion of animal products;
animals ingest soil
Soil ! dust ! human Human lifetime dust inhalation
Soil ! surface water ! human Human lifetime ingestion of surface water
and fish
Soil ! ground water ! human Human lifetime ingestion of groundwater
Soil ! air ! human Human lifetime inhalation of volatilized contaminants
From Logan (1998).
microorganisms. These interactions should profoundly influence the transform-

ation, dynamics, bioavailability, and toxicity of contaminants in soil and related
environments (Huang, 2002). Because of these interactions, some pollutants that
would be otherwise at an acceptable level may be released and enter the pathway
affecting health. Therefore, soil mineral –organic component –microorganism
interactions merit attention in soil pathways of human contaminant exposure and
risk assessment.
I. RISK MANAGEMENT, REMEDIATION,

AND RESTORATION
Land use, past and present, critically influences the extent and intensity of soil
contamination. Contamination of soils from anthropogenic chemicals and their
subsequent degradation has become a major concern because of the critical role
of soil resources in sustaining agricultural production, economic development,
and ecosystem health. Both inorganic and organic anthropogenic compounds in
soils may not only adversely affect their production potential but also deteriorate
the quality of the food chain, the air, the surface water, and the underlying
groundwater. Therefore, it is essential to manage the risk through use of
remediation technology to restore ecosystem health
Selected ameliorants have been used to remediate metal contaminated soils
(Table XII). These ameliorants are rather inexpensive and readily available on a
458 P. M. HUANG
Table XII
Selected Ameliorants that are Adapted to Metal Contaminated Soils
Target Soil processes

Technique contaminants involved Constraints
Limestone Metals, Precipitation, Ineffective for oxyanions;

radionuclides sorption certain crops (lettuce,
spinach, tuber, and others);
short term
Zeolite Metals, Ion exchange, Insufficient data;
radionuclides sorption, fixation short term
Apatite Metals Sorption, Selective; insufficient data
precipitation,
complexation
Clay Metals, Ion exchange, Type of clay; short term
mineral radionuclides sorption, fixation
From Adriano et al. (1998).
global basis. However, they may provide only interim solutions in stabilizing
contaminants. Soil processes involved in remediation by these ameliorants
include precipitation, sorption, ion exchange, fixation, and complexation. These
processes may be substantially influenced by soil mineral – organic component –
microorganism interactions especially in the rhizosphere soils where these
interactions are intense.
Plants have been described in engineering terms as solar driven pumps and
filtering systems that extract and concentrate elements from the environment
(Cunningham and Berti, 1993; Salts et al., 1995). Some of the metals that may
accumulate in plants are those that have nutritional value. Certain plants also
have the ability to accumulate metals that have no known biological function.
The value of metal-accumulating plants has recently been realized (Cunningham
and Berti, 1993; Pierzynski et al., 2000). Further, plant-assisted bioremediation
are used to remediate soil contaminated with organic compounds (Walton and
Anderson, 1992; Anderson et al., 1993). In phytoremediation, a wide range of
processes are involved in the soil rhizosphere where rhizospheric microorgan-
isms interact intensely with soil minerals and root exudates. These interactions
should have a profound influence on transformations of metals and anthropogenic
organic compounds and the efficacy of phytoremediation.
Xenobiotics can also be degraded by oxidative coupling reactions that are
catalyzed either biologically by polyphenol oxidases and peroxides (Bollag,
1992) or abiotically by metal oxides and clay minerals (Huang, 2000a). These
processes may result in the detoxification of xenobiotics and have been suggested
as a means of decontamination. Abiotic and biotic catalysts co-exist in soil
environments. Abiotic catalysts can influence microbial formation of enzymes
and enzymatic activity (cf. Section VI). Many soil abiotic catalysts can also
immobilize enzymes and alter their performance. Further, many soil abiotic
catalysts influence the activity of the desorbed enzymes. Therefore, the influence
of interactions of abiotic and biotic catalysts on the transformation and toxicity of
xenobiotics and the impact on risk management, remediation, and restoration of
ecosystem health is an issue that needs to be addressed in the years to come.
XIII. SUMMARY AND CONCLUSIONS
Soils are among the major compartments of the ecosystem. Their colloidal
or particulate constituents, be minerals, organic matter, or microorganisms,
play key roles in affecting physical, chemical, and biological processes in the
pedosphere. Minerals, organic matter, and microorganisms are constantly in
close association and in interactions with each other in soil environments.
Interactions of these three solid components mediated by soil solution and
atmosphere govern mineral weathering transformations, formation of organo-
mineral complexes, microbial and enzymatic activities, soil structural
stability, dynamics of aggregate turnover, biogeochemical cycling of C, N,
P, and S, and transformation and dynamics of metals and organic pollutants
in the terrestrial system.
Therefore, foreseeable impacts of interactions of minerals with organic matter
and microorganisms in soil environments include: (1) global ion cycling and
climate change, (2) biodiversity, (3) biological productivity and human nutrition,
(4) geomedicine, (5) industrialization, waste disposal, ecotoxicology, and human
health with respect to the bioavailability and toxicity of metals, metalloids, other
inorganics, xenobiotics, biomolecules, and pathogens, (6) biotechnology
development in relation to agricultural sustainability and ecosystem integrity,
(7) ecosystem risk assessment, and (8) ecosystem risk management and
restoration.
Fundamental understanding of soil mineral– organic matter –microorganism
interactions at the molecular level is essential to understanding and regulating the
behavior of the terrestrial ecosystem on the global scale. Future research on this
extremely important and exciting area of science should be stimulated to uncover
the dynamics and mechanisms of environmental processes in nature and the
impact on human welfare.
ACKNOWLEDGMENT
This study was supported by discovery grant GP 2383 — Huang of the Natural
Sciences and Engineering Research Council of Canada.
460 P. M. HUANG
REFERENCES
Adriano, D. C., Chlopecka, A., and Kaplan, D. I. (1998). Role of soil chemistry in soil remediation and
ecosystem conservation. In “Soil Chemistry and Ecosystem Health” (P. M. Huang, D. C.
Adriano, T. J. Logan, and R. T. Checkai, Eds.), pp. 361 –386. Soil Sci. Soc. Am., Madison, WI,
SSSA Spec. Publ. 52.
Ahmann, D., Robert, A. L., Krumholz, L. R., and Morel, F. M. M. (1994). Microbe grows by reducing
arsenic. Nature (London) 371, 750.
Alexander, M. (1995). How toxic are toxic chemicals in soil. Environ. Sci. Technol. 29, 2713– 2717.
Alexander, M. (1999). “Biodegradation and bioremediation”, 2nd ed. Academic Press, San Diego, CA.
Alloway, B. J. (1995). Soil processes and the behavior of metals. In “Heavy Metals in Soils” (B. J.
Alloway, Ed.), pp. 11–37. Blackie Academic & Professional, London.
Anderson, T. A., Guthrie, E. A., and Walton, B. T. (1993). Bioremediation. Environ. Sci. Technol. 27,
2630–2636.
Baldock, J. A. (2002). Interactions of organic materials and microorganisms with minerals in the
stabilization of structure. In “Interactions Between Soil Particles and Microorganisms. Impact
on the Terrestrial Ecosystem”. (P. M. Huang, J.-M. Bollag, and N. Senesi, Eds.), Vol. 8,
pp. 85–131. IUPAC Series on Analytical and Physical Chemistry of Environmental Systems.
Wiley, Chichester, UK.
Baldock, J. A., and Skjemstad, J. O. (2000). The role of soil mineral matrix in protecting natural
organic materials against biological attack. Org. Geochem. 31, 697– 710.
Barber, S. A. (1995). “Soil Nutrient Bioavailability: A Mechanistic Approach”, 2nd ed. Wiley,
New York.
Barnhisel, R. I., and Bertsch, P. M. (1989). Chlorites and hydroxy-interlayered vermiculite and
smectite. In “Minerals in Soil Environments” (J. B. Dixon and S. B. Weed, Eds.), pp. 729– 788.
Soil Sci. Soc. Am., Madison, WI.
Bartlett, R. J., and James, B. R. (1993). Redox chemistry of soils. Adv. Agron. 50, 151– 208.
Benjamin, M. M., and Honeyman, B. D. (1992). Trace metals. In “Global Biogeochemical Cycles”
(S. S. Butcher, R. J. Charlson, G. H. Orians, and G. V. Wolfe, Eds.), pp. 317 –352. Academic
Press, London.
Berthelin, J., Huang, P. M., Bollag, J.-M., and Andreu, F. (1999). “Effect of Mineral–Organic–
Microorganism Interactions on Soil and Freshwater Environments”. Plenum, New York.
Berthelin, J., Munier-Lamy, C., and Leyval, C. (1995). Effect of microorganisms on mobility of
heavy metals in soils. In “Environmental Impact of Soil Component Interactions. Volume II
Metals, Other Inorganics and Microbial Activities” (P. M. Huang, J. Berthelin, J.-M. Bollag,
W. B. McGill, and A. L. Page, Eds.), pp. 3– 17. CRC Press/Lewis Publishers, Boca Raton, FL.
Bertsch, P. M., and Parker, D. R. (1996). Aqueous polynuclear aluminum species. In “The
Environmental Chemistry of Aluminum” (G. Sposito, Ed.), pp. 117–168. CRC Press/Lewis
Publishers, Boca Raton, FL.
Beverdge, T. J. (1989). Role of cellular design in bacterial metal accumulation and mineralization.
Ann. Rev. Microbiol. 43, 147– 171.
Beveridge, T. J., Makin, S. A., Kadurugamuwa, J. L., and Li, Z. (1997). Interactions between biofilms
and the environment. FEMS Microbial Rev. 20, 291–303.
Bigham, J. M., Fritzpatrick, R. W., and Schulze, D. G. (2002). Iron oxides. In “Soil Mineralogy with
Environmental Applications” (J. B. Dixon and D. G. Schulze, Eds.), pp. 323–366. Soil Sci. Soc.
Am., Madison, WI, SSSA Book Ser. 7.
Bollag, J.-M. (1992). Decontaminating soil with enzymes: an in situ method using phenolic and
anilinic compounds. Environ. Sci. Technol. 26, 1876–1881.
Bollag, J.-M., Dec, J., and Huang, P. M. (1998). Formation mechanisms of complex organic structures
in soil habitats. Adv. Agron. 63, 237 –266.
Brown, L. R., and Flavin, C. (1996). “State of the World”. W.W. Norton Co., New York.
Brown, C. B., and White, J. L. (1969). Reactions of 12 s-triazines with soil clays. Soil Sci. Soc. Am.
Proc. 33, 863– 867.
Burns, R. G. (1986). Interactions of enzymes with soil mineral and organic colloids. In “Interactions of
Soil Minerals with Natural Organics and Microbes. SSSA Spec. Publ. 17” (P. M. Huang and
M. Schnitzer, Eds.), pp. 429– 451. Soil Sci. Soc. Am., Madison, WI.
Burns, R. G., and Stach, J. E. M. (2002). Microbial ecology of soil biofilms: Substrate bioavailability,
bioremediation, and complexity. In “Soil Mineral–Organic Matter–Microorganism Interactions
and Ecosystem Health. Vol. 28B. Ecological Significance of the Interactions Among Clay
Minerals, Organic Matter and Soil Biota” (A. Violante, P. M. Huang, J.-M. Bollag, and
L. Gianfreda, Eds.), pp. 17–42. Elsevier, Amsterdam.
Butcher, S. S., Charlson, R. J., Orions, G. H., and Wolfe, G. V. (1992). “Global Biogeochemical
Cycles”. Academic Press, London.
Buyanovsky, G. A., Aslam, M., and Wagner, G. H. (1994). Carbon turnover in soil physical fractions.
Soil Sci. Soc. Am. J. 58, 1167–1173.
Carter, M. R. (1996). Analysis of soil organic matter storage in agroecosystems. In “Structure
and Organic Matter Storage in Agricultural Soils” (M. R. Carter and B. S. Stewart, Eds.),
pp. 3–11. Adv. Soil Sci., CRC Press, Boca Raton, FL.
Chang, A. C., Page, A. L., and Koo, B.-J. (2002). Biogeochemistry of phosphorus, iron, and trace
elements in soils as influenced by soil — plant–microbial interactions. In “Soil Mineral–
Organic Matter –Microorganism Interactions and Ecosystem Health. Vol. 28B. Ecological
Significance of the Interactions Among Clay Minerals, Organic Matter and Soil Biota”
(A. Violante, P. M. Huang, J.-M. Bollag, and L. Gianfreda, Eds.), pp. 43–57. Elsevier,
Amsterdam.
Charlson, R. J., Orians, G. H., Wolfe, G. V., and Butcher, S. S. (1992). Human modification of global
biogeochemical cycles. In “Global Biogeochemical Cycles” (S. S. Butcher, R. J. Charlson, G. H.
Orians, and G. V. Wolfe, Eds.), pp. 353 –361. Academic Press, London.
Cheney, M. A., Sposito, G., McGrath, A. E., and Criddle, R. S. (1996.). Abiotic degradation of 2,4-D
(dichlorophenoxyacetic acid) on synthetic birnessite: calorespirometric method. Colloids Surf.
A. Phys. Eng. Asp 107, 131–140.
Cheng, H. H. (1990). “Pesticides in the Soil Environment: Processes, Impacts, and Modeling”.
Soil Sci. Soc. Am., Madison, WI.
Chorover, J., Amistadi, M. K., Burgos, W. D., and Hatcher, P. G. (1999). Quinoline sorption on
kaolinite–humic acid complexes. Soil Sci. Soc. Am. J. 63, 850–857.
Christensen, T. H., and Huang, P. M. (1999). Solid phase cadmium and the reactions of aqueous
cadmium with soil surfaces. In “Cadmium in Soils and Plants” (M. J. McLaughlin and B. R.
Singh, Eds.), pp. 65 –96. Kluwer Academic Publishers, Dordrecht, The Netherlands.
Claus, H., and Filip, Z. K. (1990). Effects of clays and other solids on the activity of phenoloxidases
produced by some fungi and actinomycetos. Soil Biol. Biochem. 22, 483–488.
Coleman, D. C., Hendrix, P. F., and Odum, E. P. (1998). Ecosystem health: an overview. In “Soil
Chemistry and Ecosystem Health” (P. M. Huang, D. C. Adriano, T. J. Logan, and R. T. Checkai,
Eds.), pp. 1–20. Soil Sci. Soc. Am., Madison WI, SSSA Spec. Publ. 52.
Collins, Y. E., and Stotzky, G. (1996). Changes in the surface charge of bacteria caused by heavy
metals does not affect survival. Can. J. Microbiol. 42, 621– 627.
Conrad, J. P. (1942). The occurrence and origin of urease-like activities in soils. Soil Sci. 54, 367–380.
Conway, G., and Toenniessen, G. (1999). Feeding the twenty-first century. Nature (London) 402,
C55–C58.
Cornell, R. M., and Schwertmann, U. (1996). “The Iron Oxides”. VCH Publ., Weinheim, Germany.
Crocker, F. H., Guerin, W. F., and Boyd, S. A. (1995). Bioavailability of naphthalene sorbed to
cationic surfactant-modified smectite clay. Environ. Sci. Technol. 29, 2953–2958.
Crosby, D. G. (1982). Environmental chemistry: an overview. Environ. Toxicol. Chem. 1, 1–8.
462 P. M. HUANG
Cunningham, S. D., and Berti, W. R. (1993). Remediation of contaminated soils with green plants: an
overview. In Vitro. Cell. Dev. Bio. 29P, 207 –212.
Dec, J., and Bollag, J.-M. (2000). Phenoloxidase-mediated interactions of phenols and anilines with
humic materials. J. Environ. Qual. 29, 665–676.
Dec, J., Bollag, J.-M., Huang, P. M., and Senesi, N. (2002). Impact of interactions between
microorganisms and soil colloids on transformation of organic pollutants. In “Interactions
Between Soil Particles and Microorganisms. Impact on Terrestrial Ecosystem”. (P. M. Huang, J.-
M. Bollag, and N. Senesi, Eds.), Vol. 8, pp. 323–378. IUPAC Series of Analytical and Physical
Chemistry of Environmental Systems. Wiley, Chichester, England.
Di Grazia, P. M., Blackburn, J. W., Bienkowski, P. R., Hilton, B., Reed, G. D., King, J. M. H., and
Sayler, G. S. (1990). Development of a system analysis approach for resolving the structure of
biodegrading systems. Appl. Biochem. Biotechnol. 24/25, 237–252.
Douglas, S., and Beveridge, T. J. (1998). Mineral formation by bacteria in natural microbial
communities. FEMS Microbiol. Ecol. 26, 79–88.
Ehrlich, H. L. (2002). Interactions between microorganisms and minerals under anaerobic conditions.
In “Interactions Between Soil Particles and Microorganisms. Impact on the Terrestrial
Ecosystem”. (P. M. Huang, J.-M. Bollag, and N. Senesi, Eds.), Vol. 8, pp. 459 –494. IUPAC
Series on Analytical and Physical Chemistry of Environmental Systems. Wiley, Chichester, UK.
Emerson, W. W., Foster, R. C., and Oades, J. M. (1986). Organo-mineral complexes in relation to soil
aggregation and structure. In “Interactions of Soil Minerals with Natural Organics and Microbes,
SSSA Spec. Publ. 17” (P. M. Huang and M. Schnitzer, Eds.), pp. 521–548. Soil Sci. Soc. Am.,
Madison, WI.
Evershed, R. P., Bland, H. A., van Bergen, P. F., Carter, J. F., Horton, M. C., and Rowley-Conwy, P. A.
(1997). Volatile compounds in archaeological plant remains and the Maillard reaction during
decay of organic matter. Science 278, 432– 433.
Farrell, R. E., Huang, P. M., and Germida, J. J. (1998). Biomethylation of mercury (II) adsorbed on
mineral colloids common in freshwater sediments. Appl. Organomet. Chem. 12, 613–620.
Filip, Z., and Hattori, T. (1984). Utilization of substrates and transformation of solid substrata.
In “Microbial Adhesion and Aggregation” (K. C. Marshall, Ed.), pp. 251– 282. Dahlem
Konferenzen, Springer-Verlag, Berlin.
Filip, Z., Haider, K., and Martin, J. P. (1972). Influence of clay minerals on growth and metabolic
activity of Epicoccum nigrum Stachybotrys chartarum. Soil Biol. Biochem. 4, 135–145.
Filip, Z., Flaig, W., and Rietz, E. (1977). Oxidation of some phenolic substances as influenced by clay
minerals. In “Isot. Radiat. Soil Org. Matter Stud., II”, pp. 91– 96. IAEA, Bull., Vienna.
Fletcher, M. (1991). The physiological activity of bacteria attached to solid surfaces. Adv. Microb.
Phyiol. 32, 53 –85.
Fletcher, M. (1996). “Bacterial Adhesion: Molecular and Ecological Diversity”. Wiley-Liss,
New York.
Gadd, G. M. (1993). Microbial formation and transformation of organometallic and organometalloid
compounds. FEMS Microbiol. Rev. 11, 297– 310.
Getzin, L. W., and Rosefield, I. (1971). Partial purification and properties of a soil enzyme that
degrades the insecticide malathion. Biochim. Biophys. Acta 235, 442–453.
Gianfreda, L., and Bollag, J.-M. (1994). Effect of soils on the behavior of immobilized enzymes. Soil
Sci. Soc. Am. J. 58, 1672–1681.
Goh, T. B., and Huang, P. M. (1986). Influence of citric and tannic acids on hydroxy-Al interlayering
in montmorillonite. Clays Clay Miner. 34, 37–44.
Golchin, A., Baldock, J. A., and Oades, J. M. (1997). A model linking organic matter decomposition,
chemistry and aggregate dynamics. In “Soil Processes and the Carbon Cycle” (R. Lal, et al., Ed.)
pp. 245 –266. CRC Press, Boca Raton, FL.
Goldstein, L. (1976). Kinetic behavior of immobilized enzyme systems. In “Methods in Enzymology,
Volume XLIV” (K. Mosbach, Ed.), pp. 397–443. Academic Press, Orlando, FL.
Guerin, W. F., and Boyd, S. A. (1992). Differential bioavailability of soil-sorbed naphthalene to two
bacterial species. Appl. Environ. Microbiol. 58, 1142–1152.
Guggenberger, G., and Haider, K. M. (2002). Effect of mineral colloids on biogeochemical cycling of
C, N, P and S in soil. In “Interactions Between Soil Particles and Microorganisms. Impact on the
Terrestrial Ecosystem”. (P. M. Huang, J.-M. Bollag, and N. Senesi, Eds.), Vol. 8, pp. 267–322.
IUPAC Series on Analytical and Physical Chemistry of Environmental Systems. Wiley,
Chichester, UK.
Haider, K. (1995). Sorption phenomena between inorganic and organic compounds in soils: Impacts
on transformation processes. In “Environmental Impact of Soil Component Interactions. Volume
I Natural and Anthropogenic Organics” (P. M. Huang, J. Berthelin, J.-M. Bollag, W. B. McGill,
and A. L. Page, Eds.), pp. 21–45. CRC Press/Lewis Publishers, Boca Raton, FL.
Hatcher, P. G., Schnitzer, M., Dennis, L. W., and Maciel, G. E. (1981). Aromaticity of humic
substances in soils Soil Sci. Soc. Am. J. 45, 1089–1094.
Hatzinger, P. B., and Alexander, M. (1995). Effect of aging of chemicals in soil on their
biodegradability and extractability. Environ. Sci. Technol. 29, 537– 545.
Hayes, K. F., and Traina, S. J. (1998). Metal ion speciation and its significance in ecosystem health.
In “Soil Chemistry and Ecosystem Health” (P. M. Huang, D. C. Adriano, T. J. Logan, and R. T.
Chechai, Eds.), pp. 45–84. Soil Sci. Soc. Am., Madison, WI, SSSA Spec. Publ. 52.
Hoffmann, M. R. (1980). Trace metal catalysis in aquatic environments. Environ. Sci. Technol. 14,
1061–1066.
Hoffmann, M. R. (1990). Catalysis in aquatic environments. In “Aquatic Chemical Kinetics”
(W. Stumm, Ed.), p. 71–111. Wiley, New York.
Hooper, D. U., and Vitousek, P. M. (1997). The effect of plant composition and diversity on ecosystem
processes. Science 277, 1302–1305.
Hsu, P. H. (1988). Mechanisms of gibbsite crystallization from partially neutralized aluminum
chloride solutions. Clays Clay Miner. 36, 25–30.
Huang, P. M. (1980). Adsorption processes in soil. In “The Handbook of Environmental Chemistry,
Vol. 2/Part A. Reactions and Processes” (O. Hutzinger, Ed.), pp. 47 –59. Springer-Verlag, Berlin.
Huang, P. M. (1988). Ionic factors affecting aluminum transformation and the impact on soil and
environmental sciences. Adv. Soil Sci. 8, 1–78.
Huang, P. M. (1990). Role of soil minerals in transformations of natural organics and xenobiotics.
In “Soil Biochemistry Volume 6” (J.-M. Bollag and G. Stotzky, Eds.), pp. 29 –115. Marcel
Dekker, New York.
Huang, P. M. (1991). Ionic factors affecting the formation of short-range ordered aluminosilicates.
Soil Sci. Soc. Am. J. 55, 1172–1180.
Huang, P. M. (1995). The role of short-range ordered mineral colloids in abiotic transformations of
organic compounds in the environment. In “Environmental Impact of Soil Component Inter-
actions. Volume I, Natural and Anthropogenic Organics” (P. M. Huang, J. Berthelin, J.-M. Bollag,
W. B. McGill, and A. L. Page, Eds.), pp. 135–167. CRC Press/Lewis Publishers, Boca Raton, FL.
Huang, P. M. (2000a). Abiotic catalysis. In “Handbook of Soil Science” (M. E. Sumner, Ed.),
pp. B303–B332. CRC Press, Boca Raton, FL.
Huang, P. M. (2000b). “Environmental Soil Chemistry and Its Impact on Agriculture and the
Ecosystem”. Università degli Studi di Napoli Federico II, Portici, Italy.
Huang, P. M. (2002). Foreseeable impacts of soil mineral–organic component–microorganism
interactions on society: ecosystem health. In “Soil Mineral–Organic Matter– Microorganism
Interactions and Ecosystem Health”. (A. Violante, P. M. Huang, J.-M. Bollag, and L. Gianfreda,
Eds.), Vol. 28A, pp. 1–36. Development in Soil Science. Elsevier, Amsterdam.
Huang, P. M., and Bollag, J.-M. (1999). Minerals — Organics–Microorganisms Interactions in the
Soil Environment. In “Structure and Surface Reactions of Soil Particles”. (P. M. Huang,
N. Senesi, and J. Buffle, Eds.), Vol. 4, pp. 3 –39. IUPAC Series on Analytical and Physical
Chemistry of Environmental Systems. Wiley, Chichester, UK.
464 P. M. HUANG
Huang, P. M., and Germida, J. J. (2002). Chemical and biological processes in the rhizosphere:
metal pollutants. In “Interactions Between Soil Particles and Microorganisms. Impact on the
Terrestrial Ecosystem”. (P. M. Huang, J.-M. Bollag, and N. Senesi, Eds.), Vol. 8, pp. 381– 438.
IUPAC Series on Analytical and Physical Chemistry of Environmental Systems. Wiley,
Chichester, UK.
Huang, P. M., and Kozak, L. M. (1970). Adsorption of hydroxy-aluminum polymers by muscovite and
biotite. Nature (London) 228, 1084–1085.
Huang, P. M., and Schnitzer, M. (1986). “Interactions of Soil Minerals with Natural Organics and
Microbes”. Soil Sci. Soc. Am., Madison, WI, SSSA Spec. Publ. 17.
Huang, P. M., and Violante, A. (1986). Influence of organic acids on crystallization and surface
properties of precipitation products of aluminum. In “Interactions of Soil Minerals with Natural
Organics and Microbes” (P. M. Huang and M. Schnitzer, Eds.), pp. 159 –221. Soil Sci. Soc. Am.,
Madison, WI, SSSA Spec. Publ. 17.
Huang, P. M., Wang, M. C., and Wang, M. K. (1999). Catalytic transformation of phenolic
compounds in the soils. In “Principles and Practices in Plant Ecology. Allelochemicals
interactions” (Inderjit, K. M. M. Dakshini and C. L. Foy, Eds.), pp. 287 –306. CRC Press, Boca
Raton, FL.
Huang, P. M., Wang, M. K., Kämpf, N., and Schulze, D. G. (2002). Aluminum hydroxides. In “Soil
Mineralogy with Environmental Applications” (J. B. Dixon and D. G. Schulze, Eds.),
pp. 261 –289. Soil Sci. Soc. Am., Madison, WI, SSSA Book Ser. 7.
Hughes, M. N., and Poole, R. K. (1989). “Metals and microorganisms”. Chapman and Hall,
New York.
Ikan, R., Rubinsztain, Y., Nissenbaum, A., and Kaplan, I. R. (1996). Geochemical aspects of the
Maillard reaction. In “The Maillard Reaction: Consequences for the Chemical and Life Sciences”
(R. Ikan, Ed.), pp. 1–25. Wiley, Chichester, UK.
Inoue, K., and Huang, P. M. (1984). Influence of citric acid on the natural formation of imogolite.
Nature (London) 308, 58–60.
Jastrow, J. D., and Miller, R. M. (1997). Soil aggregate stabilization and carbon sequestration:
feedbacks through organomineral associations. In “Soil Processes and the Carbon Cycle” (R. Lal,
et al., Eds.), pp. 207 –223. CRC Press, Boca Raton, FL.
Jenkinson, D. S., and Rayner, J. H. (1977). The turnover of soil organic matter in some of the
Rothamsted classical experiments. Soil Sci. 123, 298–305.
Jenkinson, D. S., Adams, D. E., and Wild, A. (1991). Model estimates of CO2 emissions from soil in
response to global warming. Nature (London) 351, 304– 306.
Johnston, C. T., and Tombácz, E. (2002). Surface chemistry of soil minerals. In “Soil Mineralogy with
Environmental Applications” (J. B. Dixon and D. G. Schulze, Eds.), pp. 37–67.
Jokic, A., Frenkel, A. I., Vairavamurthy, M. A., and Huang, P. M. (2001a). Birnessite catalysis of the
Maillard reaction: its significance in natural humification. Geophys. Res. Lett. 28, 3899–3902.
Jokic, A., Frenkel, A. I., and Huang, P. M. (2001b). Effect of light on birnessite catalysis of the
Maillard reaction and its implication in humification. Can. J. Soil Sci. 81, 277–283.
Kämpf, N., Scheinost, A. C., and Schulze, D. G. (2000). Oxide minerals. In “Handbook of Soil
Science” (M. E. Sumner, Ed.), pp. F 125 –168. CRC Press, Boca Raton, FL.
Karthikeyan, K. G., Chorover, J., Bortiatynski, J. M., and Hatcher, P. G. (1999). Interaction of 1-
naphthol and its oxidation products with aluminum hydroxides. Environ. Sci. Technol. 33,
4009–4015.
Kennedy, M. J., Pevear, D. R., and Hill, R. J. (2002). Mineral surface control of organic carbon in
black shale. Science 295, 657–660.
Khan, S. U. (1978). Kinetics of hydrolysis of atrazine in aqueous fulvic acid solution. Pestic. Sci. 9,
39–43.
Kirk, G. J. D., and Bajita, J. B. (1995). Root-induced iron oxidation, pH changes and zinc
solubilization in the rhizosphere of lowland rice. New Phytol. 131, 129–137.
Kjelleberg, S., Humphrey, B. A., and Marshall, K. C. (1982). Effects of interfaces on small, starved
marine bacteria. Appl. Environ. Microbial. 43, 1166–1172.
Kodama, H., and Schnitzer, M. (1974). Adsorption of fulvic acid by non-expanding clay minerals.
Trans. Int. Congr. Soil Sci., 10th 11, 51–56.
Kodama, H., and Schnitzer, M. (1980). Effect of fulvic acid on the crystallization of aluminum
hydroxides. Geoderma 24, 195– 205.
Krishnamurti, G. S. R., and Huang, P. M. (1990). Kinetics of Fe(II) oxygenation and the nature of
hydrolytic products as influenced by ligands. In “Proc. 9th Int. Clay Conf., Strasbourg” (V. C.
Farmer and Y. Tardy, Eds.), pp. 195– 204. Publié avec le concours de l’Université Louis Pasteur
et du Center National de la Recherche Scientifique, Strasbourg, France.
Krishnamurti, G. S. R., and Huang, P. M. (1991). Influence of citrate on the kinetics of Fe(II) oxidation
and the formation of iron oxyhydroxides. Clays Clay Miner. 39, 28– 34.
Krishnamurti, G. S. R., and Huang, P. M. (1993). Formation of lepidocrocite from Fe(II) solutions:
stabilization by citrate. Soil Sci. Soc. Am. J. 57, 861 –867.
Krishnamurti, G. S. R., Huang, P. M., Van Rees, K. C. J., Kozak, L. M., and Rostad, H. P. W. (1995a).
Speciation of particulate-bound Cd in soils and its bioavailability. Analyst 120, 659 –665.
Krishnamurti, G. S. R., Huang, P. M., Van Rees, K. C. J., Kozak, L. M., and Rostad, H. P. W. (1995b).
A new soil test method for the determination of plant-available cadmium in soils. Commun. Soil
Sci. Plant Anal. 26, 2857–2867.
Krishnamurti, G. S. R., Huang, P. M., and Van Rees, K. C. J. (1996). Studies on soil rhizosphere:
speciation and availability of Cd. Chem. Speciat Bioavailab 8, 23–28.
Krishnamurti, G. S. R., Cieslinski, G., Huang, P. M., and Van Rees, K. C. J. (1997). Kinetics of
cadmium release from soils as influenced by organic acids: implication in cadmium availability.
J. Environ. Qual. 26, 271 –277.
Krishnamurti, G. S. R., Wang, M. K., and Huang, P. M. (1999). Role of tartaric acid in the inhibition of
the formation of Al13 tridecamer using sulfate precipitation. Clays Clay Miner. 47, 658–663.
Kumada, K., and Kato, H. (1970). Browning of pyrogallol as affected by clay minerals. Soil Sci. Plant
Nutr. 16, 195–200.
Kurek, E. (2002). Microbial mobilization of metals from soil minerals under aerobic conditions.
In “Interactions Between Soil Particles and Microorganisms. Impact on the Terrestrial
Ecosystem”. (P. M. Huang, J.-M. Bollag, and N. Senesi, Eds.), Vol. 8, pp. 189–225. IUPAC
Series on Analytical and Physical Chemistry of Environmental Systems. Wiley, Chichester, UK.
Kwong, K. F. Ng Kee, and Huang, P. M. (1975). Influence of citric acid on the crystallization of
aluminum hydroxides. Clays Clay Miner. 23, 164–165.
Kwong, K. F. Ng Kee, and Huang, P. M. (1978). Sorption of phosphate by hydrolytic products of
aluminum. Nature (London) 271, 336–338.
Kwong, K. F. Ng Kee, and Huang, P. M. (1979a). The relative influence of low-molecular-weight
complexing organic acids on the hydrolysis and precipitation of aluminum. Soil Sci. 128,
337 –342.
Kwong, K. F. Ng Kee, and Huang, P. M. (1979b). Nature of hydrolytic precipitation products of
aluminum as influenced by low molecular weight complexing organic aicds. In “Int. Clay Conf.
Proc., 6th, Oxford” (V. C. Farmer and M. M. Mortland, Eds.), pp. 527 –536. Elsevier, Oxford,
UK.
Kwong, K. F. Ng Kee, and Huang, P. M. (1979c). Surface reactivity of aluminum hydroxides
precipitated in the presence of low-molecular-weight organic acids. Soil Sci. Soc. Am. J. 43,
1107–1113.
Kwong, K. F. Ng Kee, and Huang, P. M. (1981). Comparison of the influence of tannic acid and
selected low-molecular-weight organic acids on precipitation products of aluminum. Geoderma
26, 179–193.
Kyuma, K., and Kawaguchi, K. (1964). Oxidative changes of polyphenols as influenced by allophane.
Soil Sci. Soc. Am. Proc. 28, 371– 374.
466 P. M. HUANG
Låg, J. (1980). Survey of Geomedical Problems: Including Some Examples from Investigations
Carried Out in Norway “Geomedical Aspects in Present and Future Research”. Universitatsfor-
laget, Olso.
Låg, J. (1994). Geomedicine, an expanded application of soil science. In “Interactions of Soil
Components, Agricultural Ecosystems, and Health”. Vol. 3a, pp. 557 –567, Transactions 15th
World Congress of Soil Science. Commission II Symposium, Acapulco, Mexico.
Lal, R. (1998). “Soil Processes and the Carbon Cycle”. CRC Press, Boca Raton, FL.
Laurie, S. H., Tancock, N., McGrath, S. P., and Sanders, J. R. (1991). Influence of complexation on the
uptake by plants of iron, manganese, copper and zinc: Effect of DTPA in a multi-metal and
computer simulation study. J. Exp. Bot. 42, 509–513.
Li, G. C., and Felbeck, G. T. Jr. (1972). A study of the mechanism of atrazine adsorption by humic
acid from muck soil. Soil Sci. 113, 140 –148.
Linehan, D. J., Sinclair, A. H., and Mitchell, M. C. (1989). Seasonal changes in Cu, Mn, Zn and
Co concentrations in soil in the root zone of barley (Hordeum vulgare L). J. Soil Sci. 40,
103–115.
Liu, C., and Huang, P. M. (1999). Atomic force microscopy and surface characteristics of iron oxides
formed in citrate solutions. Soil Sci. Soc. Am. J. 63, 65–72.
Logan, T. J. (1998). Impact of soil chemical reactions on food chain contamination and environmental
quality. In “Future Prospects for Soil Chemistry” (P. M. Huang, D. L. Sparks, and S. A. Boyd,
Eds.), pp. 191–204, SSSA Spec. Publ. 55, Soil Sci. Soc. Am., Madison, WI.
Lund, W. (1990). Speciation analysis — why and how? Fresenius J. Anal. Chem. 337, 557–564.
Lynch, J. M. (1990a). Introduction: some consequences of microbial rhizosphere competence for plant
and soil. In “The Rhizosphere” (J. M. Lynch, Ed.), p. 1–10. Wiley, Chichester, UK.
Lynch, J. M. (1990b). Microbial metabolites. In “The Rhizosphere” (J. M. Lynch, Ed.), p. 177– 206.
Wiley, New York.
Mackay, D., and Paterson, S. (1991). Evaluating the multimedia fate of organic chemicals: a level III
fugacity model. Environ. Sci. Technol. 25, 427 –436.
Maillard, L. C. (1913). Formation de matières humiques par action de polypeptides sur sucres. C.R.
Acad. Sci. 156, 148 –149.
Malcolm, R. L. (1989). Application of solid-state 13C NMR spectroscopy to geochemical studies
of humic substances. In “Humic Substances II. In Search of Structure” (M. H. B. Hayes,
P. MacCarthy, R. L. Malcolm, and R. S. Swift, Eds.), pp. 339 –372. Wiley, Chichester, UK.
Mann, S., Archibald, D. A., Didymus, J. M., Douglas, T., Heywood, R. R., Meldrum, F. C., and
Reeves, N. J. (1993). Crystallization at inorganic – organic interfaces: biominerals and
biomimetic synthesis. Science 261, 1286–1292.
Marschner, H. (1998). Soil– root interface: biological and biochemical processes. In “Soil Chemistry
and Ecosystem Health” (P. M. Huang, D. C. Adriano, T. J. Logan, and R. T. Checkai, Eds.),
pp. 191 –231, SSSA Spec. Publ. 52, Soil Sci. Soc. Am., Madison WI.
Marshall, K. C. (1971). Sorptive interactions between soil particles and microorganisms. In “Soil
Biochemistry” (A. D. McLaren and J. Skujins, Eds.), Vol. 2, pp. 409–442. Marcel Dekker,
New York.
Marshall, K. C. (1980). Adsorption of microorganisms to soils and sediments. In “Adsorption
of Microorganisms to Surfaces” (G. Britton and K. C. Marshall, Eds.), pp. 317– 330. Wiley,
New York.
Martinek, K., and Mozhaev, V. V. (1985). Immobilization of enzymes: an approach to fundamental
studies in biochemistry. Adv. Enzymol. 57, 179–249.
Masschelyn, P. H., and Patrick, W. H. Jr. (1994). Selecium, arsenic and chromium redox chemistry
in wetland soils and sediments. In “Biochemistry of Trace Elements” (D. C. Adriano, Ed.),
pp. 615 –625. Science and Technology Letters, Northwood.
McBride, M. B. (1989). Reactions controlling heavy metal solubility in soil. Adv. Soil Sci. 10, 1 –56.
McBridge, M. B. (1994). “Environmental Chemistry of Soils”. Oxford University Press, Oxford.
McGill, W. B., and Cole, C. V. (1981). Comparative aspects of cycling of organic C, N, S and P
through soil organic matter. Geoderma 26, 267 –286.
McLaughlin, M. J., Smolders, E., and Merckx, R. (1998). Soil– root interface: physicochemical
processes. In “Soil Chemistry and Ecosystem Health.” (P. M. Huang, D. C. Adriano,, T. J.
Logan,, and R. T. Checkai,, Eds.), pp. 233 –277, SSSA Spec. Publ. 52, Soil Sci. Soc. Am.,
Madison, WI.
McLean, J. S., Lee, J.-U., and Beveridge, T. L. (2002). Interactions of bacteria and environmental
metals, fine-grained mineral development and bioremediation strategies. In “Interactions
Between Soil Particles and Microorganisms. Impact on the Terrestrial Ecosystem”. (P. M.
Huang, M. J.-Bollag, and N. Senesi, Eds.), Vol. 8, pp. 227 –261. IUPAC Series on Analytical and
Physical Chemistry of Environmental Systems. Wiley, Chichester, UK.
Mingelgrin, U., Saltzman, S., and Yaron, B. (1977). A possible model for the surface-induced
hydrolysis of organophosphorus pesticides on kaolinite clays. Soil Sci. Soc. Am. J. 41, 519–523.
Morra, M. J. (1996). Bioremediation in soil: influence of soil properties on organic contaminants and
bacteria. In “Bioremediation: Principles and Applications” (R. L. Crawford and D. L. Crawford,
Eds.), pp. 35–60. Cambridge University Press, Cambridge, UK.
Mortland, M. M. (1986). Mechanisms of adsorption of nonhumic organic species by clays.
In “Interactions of Soil Minerals with Natural Organics and Microbes” (P. M. Huang and
M. Schnitzer, Eds.), pp. 59–76, SSSA Special Publication No. 17. Soil Sci. Soc. Am., Madison, WI.
Mortland, M. M., and Halloran, L. J. (1976). Polymerization of aromatic molecules on smectites. Soil
Sci. Soc. Am. J. 40, 367–370.
Munnecke, D. M., Johnson, L. M., Tablot, H. W., and Barik, S. (1982). Microbial metabolism and
enzymology of selected pesticides. In “Biodegradation and Detoxification of Environmental
Pollutants” (A. M. Chakrabarty, Ed.), pp. 1 –32. CRC Press, Boca Raton, FL.
Myneni, S. C. B., Tokunaga, T. K., and Brown, G. E. Jr. (1997). Abiotic selenium redox
transformations in the presence of Fe(II, III) oxdies. Science 278, 1108–1109.
Naeem, S., Thompson, L. J., Lawler, S. P., Lawton, J. H., and Woodfin, R. M. (1994). Declining
biodiversity can alter the performance of ecosystems. Nature (London) 368, 734–737.
Naidja, A., Huang, P. M., and Bollag, J.-M. (2000). Enzyme–Clay Interactions and Their Impact on
Transformations of Natural and Anthropogenic Organic compounds in Soils. J. Environ. Qual.
29, 677–691.
Naidja, A., and Huang, P. M. (1996). Deamination of aspartic acid by aspartase — Ca –
montmorillonite complex. J. Mol. Catal. 106, 255–265.
Naidja, A., Huang, P. M., and Bollag, J.-M. (1997). Activity of tyrosinase immobilized on
hydroxyaluminum –montmorillonite complexes. J. Mol. Catal. A: Chem. 115, 305 –316.
Naidja, A., Liu, C., and Huang, P. M. (2002). Formation of protein– birnessite complex: XRD, FTIR,
and AFM analysis. J. Colloid Interface Sci. 251, 46 –56.
Nannipieri, P., and Gianfreda, L. (1999). Kinetics of enzyme reactions in soil environments.
In “Structure and Surface Reactions of Soil Particles.” (P. M. Huang, N. Senesi, and J. Buffle,
Eds.), Vol. 4, pp. 449 – 479. IUPAC Series on Analytical and Physical Chemistry of
Environmental Systems. Wiley, Chichester, UK.
National Academy of Sciences (U.S.), (1974). Geochemistry and the Environment. “The Relation of
Selected Trace Elements to Health and Disease”, Vol. I. Washington, DC.
National Academy of Sciences (U.S.), (1997). Geochemistry and the Environment. “The Relation of
Other Selected Trace Elements to Health and Disease”, Vol. II. Washington, DC.
Nelson, Y. M., Lion, L. W., Ghiorse, W. C., and Shuler, M. L. (1999). Production of biogenic Mn
oxides by Leptothrix discophora SS-1 in a chemically defined growth medium and evaluation of
their Pb adsorption characteristics. Appl. Environ. Microbial. 65, 175–180.
Nielands, J. B. (1981). Microbial iron compounds. Ann. Rev. Biochem. 50, 715 –731.
Nielsen, N. E. (1976). The effect of plants on the copper concentration in the soil solution. Plant Soil
45, 679–687.
468 P. M. HUANG
NIST, (1997). Standard Reference Database 46. “Critically Selected Stability Constants of Metal
Complexes, Version 5.0”. National Institute of Standards and Technology, US Department of
Commerce, Gaithersburg, MD.
Oades, J. M., and Waters, A. G. (1991). Aggregate hierarchy in soils. Aust. J. Soil Res. 29, 815 –828.
Oades, J. M., Gillman, G. P., and Uehara, G. (1989). Interactions of soil organic matter and variable-
charge clays. In “Dynamics of Soil Organic Matter in Tropical Ecosystems”. (D. C. Coleman,
J. M. Oades, and G. Uehara, Eds.), pp. 69–95. Nif TAL Project. University of Hawaii, Honolulu.
Odum, E. P. (1989). “Ecology and our endangered life-support systems”. Sinauer Assoc., Sunderland,
MA.
Oscarson, D. W., Huang, P. M., Defosse, C., and Herbillon, A. (1981). Oxidative power of Mn(IV) and
Fe(III) oxides with respect to As(III) in terrestrial and aquatic environments. Nature (London)
291, 50 –51.
Oscarson, D. W., Huang, P. M., Hammer, U. T., and Liaw, W. K. (1983). Oxidation and sorption of
arsenite by manganese dioxide as influenced by surface coatings of iron and aluminum oxides and
calcium carbonate. Water Air Soil Poll. 20, 233–244.
Otte, M. L., Dekkers, M. J., Rozema, J., and Broekman, R. A. (1991). Uptake of arsenic by Aster
tripolium in relation to rhizosphere oxidation. Can. J. Bot. 69, 2670–2677.
Otte, M. L., Kearns, C. C., and Doyle, M. O. (1995). Accumulation of arsenic and zinc in the
rhizosphere of wetland plants. Bull. Environ. Contam. Toxicol. 55, 154 –161.
Otte, M. L., Rozem, J., Koster, L., Haarsma, M. S., and Broekman, R. A. (1989). Iron plaque on roots
of Aster tripolium L.: interaction with Zn uptake. New Phytol. 111, 309–317.
Parker, D. R., Chaney, R. L., and Norvell, W. A. (1995). Chemical equilibrium models: applications to
plant nutrition. In “Soil Chemical Equilibrium and Reaction Models”. “SSSA Spec. Publ.” (R. H.
Loeppert, et al., Eds.), Vol. 42, pp. 163–200. American Society of Agronomy and Soil Science
Society of America, Madison, WI.
Parton, W. J., Schimel, D. S., Cole, C. V., and Ojima, D. S. (1987). Analysis of factors controlling soil
organic matter levels in Great Plains grasslands. Soil Sci. Soc. Am. J. 51, 1173– 1179.
Paul, E. A., and Van Veen, J. A. (1978). The use of tracers to determine the dynamics of organic
matter. Trans. Int. Soc. Soil Sci. (Edmonton) 3, 61–102.
Perdue, E. M., and Wolfe, N. L. (1982). Modification of pollutant hydrolysis kinetics in the presence
of humic substances. Environ. Sci. Technol. 16, 847–852.
Pierzynski, G. M., Sim, J. T., and Vance, G. F. (2000). “Soil and Environmental Quality”. CRC Press,
Boca Raton, FL.
Pinnavaia, T. J., Hall, P. L., Cady, S. S., and Mortland, M. M. (1974). Aromatic radical cation
formation on intracrystal surfaces of transition metal layer silicates. J. Phys. Chem. 78, 994– 999.
Preston, C. M., Rauthan, B. S., Rodger, C., and Ripmeester, J. A. (1982). A hydrogen-1, carbon-13,
and nitrogen-15 nuclear magnetic resonance study of p-benzoquinone polymers incorporating
amino nitrogen compounds ("synthetic humic acids"). Soil Sci. 134, 277 –293.
Quiquampoix, H. (1987). A stepwise approach to the understanding of extracellular enzyme activity in
soil. I. Effect of electrostatic interactions on the conformation of a b-D-glucosidase adsorbed on
different mineral surfaces. Biochimie 69, 753– 763.
Reid, C. P. P. (1990). Mycorrhizas. In “The rhizosphere” (J. M. Lynch, Ed.), pp. 281–315. Wiley,
New York.
Rich, C. I. (1968). Hydroxy interlayers in expansible layer silicates. Clays Clay Miner. 16, 15–30.
Robert, M., and Berthelin, J. (1986). Role of biological and biochemical factors in soil mineral
weathering. In “Interactions of Soil Minerals with Natural Organics and Microbes. SSSA Spec.
Publ. 17” (P. M. Huang and M. Schnitzer, Eds.), pp. 453–495. Soil Sci. Soc. Am., Madison, WI.
Ruggiero, P., Sakar, J. M., and Bollag, J.-M. (1989). Detoxification of 2,4-dichlorophenol by a laccase
immobilized on soil and clay. Soil Sci. 147, 361–370.
Ruggiero, P., Dec, J., and Bollag, J.-M. (1996). Soil as a catalytic system. In “Soil Biochemistry
Volume 9” (G. Stotzky and J.-M. Bollag, Eds.), pp. 79. Marcel Dekker, New York.
Salts, D. E., Blaylock, M., Kumar, N., Dushenkov, V., Ensley, B. D., Chet, I., and Raskin, I. (1995).
Phytoremediation: a novel strategy for the removal of toxic metals from the environment using
plants. Biotechnology, 13, 468 –474.
Scheffer, F., Meyer, B., and Niederbudde, E. A. (1959). Huminstaffbildung unter katalyischer
Einwirkung natürlich vorkommender Eisenverbindungen im Modellversuch. Z. Pflanzenernaehr.
Bodenkd. 87, 26–44.
Schnitzer, M. (1977). Recent findings on the characterization of humic substances extracted from soils
from widely differing climatic zones. In “Soil Organic Matter Studies II”, pp. 77–101. IAEA
Bull., Vienna.
Schnitzer, M. (1995). Organic–inorganic interactions in soils and their effects on soil quality.
In “Environmental Impact of Soil Component Interactions, Vol. I Natural and Anthropogenic
Organics” (P. M. Huang, J. Berthelin, J.-M. Bollag, W. B. McGill, and A. L. Page, Eds.),
pp. 3–45. CRC/Lewis Publishers, Boca Raton, FL.
Schnitzer, M. (1986). Binding in humic substances by soil mineral colloids. In “Interactions of
Soil Minerals with Natural Organics and Microbes” (P. M. Huang and M. Schnitzer, Eds.),
pp. 77–101. Soil Sci. Soc. Am., Madison, WI, SSSA Special Publication No. 17.
Schnitzer, M. (2000). A lifetime perspective on the chemistry of soil organic matter. Adv. Agron. 68,
1 –58.
Schnitzer, M., and Khan, S. U. (1972). “Humic Substances in the Environment”. Marcel Dekker,
New York.
Schnitzer, M., and Kodama, H. (1966). Montmorillonite: effect of pH on its adsorption of a soil humic
compound. Science, 153, 70–71.
Schnitzer, M., and Preston, C. M. (1983). Effects of acid hydrolysis on the 13C NMR spectra of humic
substances. Plant Soil, 75, 201–211.
Schwertmann, U. (1966). Inhibitory effect of soil organic matter on the crystallization of amorphous
ferric hydroxide. Nature (London) 212, 645 –646.
Schwertmann, U. (1985). The effect of pedogenetic environments on iron oxide minerals. Adv. Soil
Sci. 1, 171 –200.
Schwertmann, U., and Taylor, R. M. (1989). Iron oxides. In “Minerals in Soil Environments” (J. B.
Dixon and S. B. Weed, Eds.), pp. 379–438. Soil Sci. Soc. Am., Madison WI.
Schwertmann, U., Kodama, H., and Fisher, W. R. (1986). Mutual interactions between organics and
iron oxides. In “Interactions of Soil Minerals with Natural Organics and Microbes” (P. M. Huang
and M. Schnitzer, Eds.), pp. 223 –250. Soil Sci. Soc. Am., Madison, WI, SSSA Spec. Publ. 17.
Senesi, N., and Miano, T. M. (1995). The role of abiotic interactions with humic substances on the
environmental impact of organic pollutants. In “Environmental Impact of Soil Component
Interactions. Volume I Natural and Anthropogenic Organics” (P. M. Huang, J. Berthelin, J.-M.
Bollag, W. B. McGill, and A. L. Page, Eds.), pp. 311– 335. CRC/Lewis Publishers, Boca Raton,
FL.
Shimp, R. J., and Young, R. L. (1988). Availability of organic chemicals for biodegradation in settled
bottom sediments. Ecotoxicol. Environ. Safety 15, 31–45.
Shindo, H., and Huang, P. M. (1982). Role of Mn(IV) oxide in abiotic formation of humic substances
in the environment. Nature (London) 298, 363–365.
Shindo, H., and Huang, P. M. (1984). Catalytic effects of manganese (IV), iron (III), aluminum, and
silicon oxides on the formation of phenolic polymers. Soil Sci. Soc. Am. J. 48, 927–934.
Shindo, H., and Huang, P. M. (1985a). The catalytic power of inorganic components in the abiotic
synthesis of hydroquinone-derived humic polymers. Appl. Clay Sci. 1, 71–81.
Shindo, H., and Huang, P. M. (1985b). Catalytic polymerization of hydroquinone by primary
minerals. Soil Sci. 139, 505–511.
Singer, A., and Huang, P. M. (1990). Effect of humic acid on the crystallization of aluminum
hydroxides. Clays Clay Miner. 38, 47–52.
Sparks, D. L. (2002). “Environmental Soil Chemistry”. Academic Press, San Diego, CA.
470 P. M. HUANG
Sposito, G. (1984). “The Surface Chemistry of Soils”. Oxford University Press, Oxford.
Sposito, G., and Reginato, R. J. (1992). “Opportunities in Basic Soil Science Research”. Soil Sci. Soc.
Am., Madison, WI.
Stevenson, F. J. (1994). Humus Chemistry “Genesis, Composition, Reactions.”, 2nd ed. Wiley,
New York.
Stone, A. T., and Torrents, A. (1995). The role of dissolved metals and metal-containing surfaces in
catalyzing the hydrolysis of organic pollutants. In “Environmental Impact of Soil Component
Interactions.” (P. M. Huang, J. Berthelin, J.-M. Bollag, W. B. McGill, and A. L. Page, Eds.),
Vol. I., pp. 275–298. Natural and Anthropogenic Organics. CRC/Lewis Publishers, Boca Raton,
FL.
Stotzky, G. (1986). Influence of soil mineral colloids on metabolic processes, growth, adhesion, and
ecology of microbes and viruses. In “Interactions of Soil Minerals with Natural Organics and
Microbes” (P. M. Huang and M. Schnitzer, Eds.), pp. 305–428. Soil Sci. Soc. Am., Madison, WI,
SSSA Spec. Publ. 17.
Stotzky, G. (2002). Clays and humic acids affect persistence and biological activity of insecticidal
proteins from Bacillus Thuringiensis in soil. In “Soil Mineral–Organic Matter–Microorganism
Interactions and Ecosystem Health”. Ecological Significance of the Interactions Among Clay
Minerals, Organic Matter and Soil Biota. (A. Violante, P. M. Huang, J.-M. bollag, and
L. Gianfreda, Eds.), Vol. 28B, pp. 1–16. Elsevier, Amsterdam.
Struif, B., Weil, L., and Quentin, K. E. (1975). The behavior of herbicides, phenoxy acetic acids and
their esters in water. Vom. Wasser 45, 53 –73.
Szmigielska, A. M., Cieslinski, G., Van Rees, K. C. J., and Huang, P. M. (1997). Low molecular
weight dicarboxylic acids in rhizosphere soil of durum wheat. Commun. Soil Sci. Plant Anal. 28,
99–111.
Tan, K. H. (1986). Degradation of soil minerals by organic acids. In “Interactions of Soil Minerals
with Natural Organics and Microbes” (P. M. Huang and M. Schnitzer, Eds.), pp. 1 –27. Soil Sci.
Soc., Am., Madison, WI, SSSA Spec. Publ. 17.
Tebo, B. M., Ghiorse, W. C, van Waasbergen, L. G., Siering, P. L., and Caspi, R. (1997). Bacterially
mediated mineral formation: insights into manganese (II) oxidation from molecular genetic and
biochemical studies. In “Geomicrobiology: Interactions Between Microbes and Minerals,
Reviews in Mineralogy” (J. F. Banfield and K. H. Nelson, Eds.), Vol. 35, pp. 225– 266.
Mineralogical Society of America, Washington, DC.
Tessier, A., and Turner, D. R. (1995). “Metal Speciation and Bioavailability in Aquatic Systems”.
IUPAC Series on Analytical and Physical Chemistry of Environmental Systems, Vol. 3. Wiley,
Chichester, UK.
Theng, B. K. G. (1979). “Formation and Properties of Clay–Polymer Complexes”. Elsevier,
New York.
Theng, B. K. G., and Orchard, V. A. (1995). Interactions of clays with microorganisms and bacterial
survival in soil: a physicochemical perspective. In “Environmental Impact of Soil Component
Interactions”. “Metals, Other Inorganics and Microbial Activities” (P. M. Huang, J. Berthelin,
J.-M. Bollag, W. G. McGill, and A. L. Page, Eds.), Vol. II, pp. 123 –143. CRC/Lewis Publishers,
Boca Raton, FL.
Theng, B. K. G., Tate, K. R., and Sollins, P. (1989). Constituents of organic matter in temperate
and tropical soils. In “Dynamics of Soil Organic Matter in Tropical Ecosystems”
(D. C. Coleman, J. M. Oades, and G. Uehara, Eds.), pp. 5–31, Nif TAL Project, University
of Hawaii, Honolulu.
Tiller, K. G. (1996). Soil contamination issues: past, present and future, a personal perspective
In “Contaminants and the Soil Environment in the Australasia-Pacific Region” (R. Naidu, et al.,
Eds.), pp. 1–27. Kluwer Academic Publ., Dordrecht, The Netherlands.
Tilman, D., and Downing, J. A. (1994). Biodiversity and stability in grassland. Nature (London) 367,
363–365.
Tilman, D., Wedlin, D., and Knops, J. (1996). Productivity and sustainability influenced by
biodiversity in grassland ecosystems. Nature (London) 379, 718–720.
Torn, M. S., Trumbore, S. E., Chadwick, O. A., Vitousek, P. M., and Hendricks, D. M. (1997). Mineral
control of soil organic carbon storage and turnover. Nature (London) 389, 170–173.
U.S. Environmental Protection Agency., (1995). “A Guide to the Biosolid Risk Assessments for the
USEPA Part 503 Rule” Office of Wastewater Management. USEPA, Washington, DC, USEPA,
832-B-93-005.
van der Heijden, M. G. A., Klironomos, J. N., Ursic, M., Moutoglis, P., Streitwolf-Engel, R., Boller,
T., Wiemken, A., and Sanders, I. R. (1998). Mycorrhizal fungal diversity determines plant
biodiversity, ecosystem variability and productivity. Nature (London) 396, 69–72.
Venkateswerlu, G., Yorder, M. J., and Stotzky, G. (1989). Morphological ultra-structural, and
chemical changes induced in Cunninghamella blakesleeana by copper and cobalt. Appl.
Microbiol. Biotechnol. 31, 204–210.
Violante, A., and Gianfreda, L. (2000). Role of biomolecules in the formation and reactivity towards
nutrients and organics of variable charge minerals and organo-mineral complexes in soil
environment. In “Soil Biochemistry” (J.-M. Bollag and G. Stotzky, Eds.), Vol. 10, pp. 207–270.
Marcel Dekker, New York.
Violante, A., Huang, P. M., Bollag, J.-M., and Gianfreda, L. (2002a). Soil Mineral –Organic Matter–
Microorganism Interactions and Ecosystem Health. In “Dynamics, Mobility and Transformation
of Pollutants and Nutrients.” Vol. 28A. Elsevier, Amsterdam.
Violante, A., Huang, P. M., Bollag, J.-M., and Gianfreda, L. (2002b). Soil Mineral– Organic Matter–
Microorganism Interactions and Ecosystem Health. In “Ecological Significance of the Interactions
Among Clay Minerals, Organic matter and Soil Biota.”, Vol. 28B. Elsevier, Amsterdam.
Vitrousek, P. M., Ehrlich, P. R., Ehrlich, A. H., and Mattson, P. A. (1986). Human appropriation of the
products of photosynthesis. Bio. Sci. 36, 368–373.
Wada, K. (1989). Allophane and imogolite. In “Minerals in Soil Environments”, 2nd ed. (J. B. Dixon
and S. B. Weed, Eds.), pp. 1051–1087. Soil Sci. Soc. Am., Madison, WI, SSSA Book Ser. 1.
Wada, K. (1995). Role of aluminum and iron in the accumulation of organic matter in soils with
variable charge. In “Environmental Impact of Soil Component Interactions”. “Natural and
Anthropogenic Organics.” (P. M. Huang, J. Berthelin, J.-M. Bollag, W. B. McGill, and A. L.
Page, Eds.), Vol. 1, pp. 47– 58. CRC/Lewis Publishers, Boca Raton, FL.
Wada, K., and Higashi, T. (1976). The categories of aluminum– and iron –humus complexes in Ando
soils determined by selective dissolution. J. Soil Sci. 27, 357– 368.
Wada, S., Ichikawa, H., and Tatsumi, K. (1992). Removal of phenols with tyrosinase immobilized on
magnetite. Water Sci. Technol. 26, 2057– 2059.
Waisel, Y., Eshel, A., and Kafkafi, U. (1996). “Plant Roots. The Hidden Half.”, 2nd Ed. Marcel
Dekker, New York.
Walker, S. G, Flemming, C. A., Ferris, F. G., Beveridge, T. J., and Bailey, G. W. (1989).
Physicochemical interaction of Escherichia coli cell envelopes and Bacillus subtilis cell walls
with two clays and ability of the composite to immobilize heavy metals from solution. Appl.
Environ. Microbiol. 55, 2976– 2984.
Walton, B. T., and Anderson, T. A. (1992). Plant-microbe treatment systems for toxic waste. Current
Opinion Biotech. 3, 267–270.
Wang, M. C., and Huang, P. M. (1986). Humic macromolecular interlayering in nontronite through
interaction with phenol monomers. Nature (London) 323, 529 –531.
Wang, M. C., and Huang, P. M. (1987). Polycondensation of pyrogallol and glycine and the associated
reactions as catalyzed by birnessite. Sci. Tot. Environ. 62, 435 –442.
Wang, M. C., and Huang, P. M. (1988). Catalytic power of nontronite, kaolinite, and quartz and their
reaction sites in the formation of hydroquinone-derived polymers. Appl. Clay Sci. 4, 43– 57.
Wang, M. C., and Huang, P. M. (1992). Significance of Mn(IV) oxide in the abiotic ring cleavage of
polyphenol in natural environments. Sci. Tot. Environ. 113, 147 –157.
472 P. M. HUANG
Wang, M. C., and Huang, P. M. (1994). Structural role of polyphenols in influencing the ring cleavage
and related chemical reactions as catalyzed by nontronite. In “Humic Substances in the Global
Environment and Implications in Human Health.” (N. Senesi and T. M. Miano, Eds.),
pp. 173 –180. Elsevier, Amsterdam.
Wang, M. C., and Huang, P. M. (2000a). Characteristics of pyrogallol-derived polymers formed by
catalysis of oxides. Soil Sci. 165, 737–747.
Wang, M. C., and Huang, P. M. (2000b). Ring cleavage and oxidative transformation of pyrogallol
catalyzed by Mn, Fe, Al, and Si oxides. Soil Sci. 165, 934–942.
Wang, T. S. C., and Li, S. W. (1977). Clay minerals as heterogeneous catalysts in preparation of model
humic substances. Z. Pflanzenernaehr. Bodenkd. 140, 669 –676.
Wang, T. S. C., Huang, P. M., Chou, C. H., and Chen, J. H. (1986). The role of soil minerals in the
abiotic polymerization of phenolic compounds and formation of humic substances. In
“Interactions of Soil Minerals with Natural Organics and Microbes” (P. M. Huang and M.
Schnitzer, Eds.), pp. 251 –281. Soil Sci. Soc. Am., Madison WI, SSSA Spec. Publ. 17.
Wang, T. S. C., Wang, M. C., and Huang, P. M. (1983a). Catalytic synthesis of humic substances by
using alumina as catalysts. Soil Sci. 136, 226 –230.
Wang, T. S. C., Wang, M. C., Ferng, Y. L., and Huang, P. M. (1983b). Catalytic synthesis of humic
substances by natural clays, silts, and soils. Soil Sci. 135, 350–360.
Weber, J. B., and Coble, H. D. (1968). Microbial decomposition of diquat adsorbed on
montmorillonite and kaolinite clays. J. Agric. Food Chem. 16, 475–478.
Weed, S. B., Davey, C. B., and Cook, M. G. (1969). Weathering of micas by fungi. Soil Sci. Soc. Am.
Proc. 33, 702–706.
Xue, J., and Huang, P. M. (1995). Zn adsorption–desorption on short-range ordered iron oxides as
influenced by citric acid during its formation. Geoderma 64, 343–356.
LOW EXTERNAL INPUT
TECHNOLOGIES FOR LIVELIHOOD
IMPROVEMENT IN SUBSISTENCE
AGRICULTURE
Anil Graves,1 Robin Matthews1 and Kevin Waldie2
1
Institute of Water and Environment, Cranfield University, Silsoe, Bedfordshire
MK45 4DT, United Kingdom
2
International and Rural Development Department, University of Reading, Reading
RG6 6AH, United Kingdom
I. Introduction
II. The Technologies
A. Intercropping
B. Alley Cropping
C. Cover Crops and Green Manures
D. Biomass Transfer Techniques
E. Compost
F. Animal Manure
G. Improved Fallows
III. Generic Issues
A. Soil Fertility Management
B. Socio-economic Issues
IV. Discussion
A. Integrated Nutrient Management
B. A Systems Perspective
C. Modelling
V. Concluding Remarks
Acknowledgements
References
I. INTRODUCTION
The global population is currently predicted to reach 9.3 billion people by the
year 2050, a 50% increase over the current level, after which it will level off due
to falling fertility rates and family sizes (UNPD, 2001). Of these, 84% will live
473
Copyright q 2004 by Academic Press. All rights of reproduction in any form reserved
0065-2113/03 $35.00
474 A. GRAVES, R. MATTHEWS AND K. WALDIE
in countries that are currently classified as being in the developing world

(Pretty, 1999). This rise in population, together with a desire for a wider variety
in diet brought about by greater purchasing power through a steady
improvement in incomes, is predicted to increase food demand over the period
1990 – 2050 by 2.4 times in Asia, 1.9 times in Latin America and the Caribbean,
and 5-fold in Africa (FAO, 1996). Moreover, the increase in production
required to meet this demand will need to be achieved with less water,
less labour, and less land, and without adversely affecting the environment
(Dowling et al., 1998).
How this is to be achieved is the topic of some debate (e.g., Pretty, 1999;
Crosson and Anderson, 2002). Over the 40 years since 1960, the global
population has doubled; despite this, food production has more than kept pace,
resulting in a 24% increase in per capita world food production and a 40%
reduction in food prices in real terms (although these figures do mask some
striking imbalances — per capita food production has fallen 20% in Africa, for
example). The total number of undernourished people in the world has also
fallen significantly over the same period. This has been largely achieved by the
use of “Green Revolution” technologies, i.e., high-yielding cereal varieties,
together with high levels of inputs such as water from irrigation systems,
fertiliser to provide the nutrients needed by the varieties, and pesticides to
control any associated weeds, pests and diseases. These technologies generally
need a relatively high capital investment, either by, or on the part of farmers,
and also need a well-functioning economic and physical infrastructure for
effective implementation. However, an estimated 30 –35% of the world’s
population (i.e., 1.9 – 2.1 billion people) do not have access to such
infrastructures, are remote from markets, practice subsistence agriculture on
marginal soils, and lack access to knowledge on how to improve their situation
(Pretty, 1999).
One school of thought is that a similar high external input agriculture (HEIA)
approach as used in the last 40 years can also be used to address the demand
for food in the next 50 years by improving the productivity of this group of
subsistence farmers, perhaps using new emerging technologies such as genetic
modification (e.g., Crosson and Anderson, 2002). A second school of thought is
that such an approach is not sustainable, and moreover, is damaging to the
environment as the inputs of fertilisers and chemicals accumulate in neighbouring
ecosystems. Thus, technologies using low levels of external inputs readily
available either on-farm or from nearby off-farm sources are seen by some
experts as more appropriate and sustainable (Pretty, 1995). This approach, often
referred to as low external input agriculture (LEIA), emphasises the use of
techniques that integrate natural processes such as nutrient cycling, biological
nitrogen fixation (BNF), soil regeneration and natural enemies of pests, into food
production processes (Pieri, 1995; Snapp et al., 1998). Efforts are also made to
minimise losses from the system, such as by leaching or removal of crop residues.
LOW EXTERNAL INPUT TECHNOLOGIES 475
The use of non-renewable inputs such as pesticides and fertilisers that can
damage the environment or harm the health of farmers and consumers is also
minimised, and more emphasis is placed on the use of such techniques as, for
example, intercropping, agroforestry, cover crops, or animal manure. Usually,
but not always, such technologies are more labour-intensive than the HEIA
approach (Deugd et al., 1998). In many cases, LEIA technologies are not new,
but are variations of those practised by farmers for generations, who have sought
to make use of resources such as vegetation or animal manure that have always
been ready to hand. Wolf (1986) has estimated that about 1.4 billion people
(25% of the world’s population), depend on this type of agriculture for their
livelihood.
Thus, the heart of the debate is not about whether either approach “works,” as
clearly both do, and have done, under the appropriate conditions and according to
their own criteria. Rather, the central question concerns which approach can best
address the future demand for food production while protecting the environment
as much as possible. Within this general debate, more specific questions relate
to whether LEIA technologies really have the capability to maintain or increase
productivity per unit area above current levels (e.g., Crosson and Anderson,
2002). Certainly, there is evidence to suggest that the relative rate of increase in
crop yields through the use of Green Revolution technologies is slowing (Mann,
1997), although Crosson and Anderson (2002) argue that this is more likely due
to the practice of quoting annual percentage increases of a constantly increasing
baseline rather than absolute annual growth. Proponents of LEIA technologies
often claim that the reliance on local sources of inputs is more sustainable, but
the analysis of De Jager et al. (2001) suggests there is little difference between
the two approaches in this respect, with both mining similar quantities of soil
nutrients to generate farm income. However, despite the continuing debate on the
relative performance of the two approaches, there are few studies that compare
yields and production under the same soil and climatic conditions and over wide
areas. With LEIA technologies in particular, there is little in the literature on the
issues that need to be faced in scaling up production from plot level to supplying
inputs and meeting food demand on a larger scale.
The purpose of this review is to examine research that has been conducted in
recent years on a number of LEIA technologies in the context of subsistence
agriculture — firstly, to determine our current state of knowledge, and secondly,
to assess their potential to improve livelihoods of subsistence farmers and
contribute to long-term sustainability of the resource base. The review was
prompted by concerns within the United Kingdom Department for International
Development (DFID) that, despite considerable research on, and dissemination
of, a wide range of LEIA technologies, there had been little clear evidence of
widespread uptake by farmers. Although this could partly be ascribed to
inadequate attention to promotion pathways and dissemination by research
centres, there was some concern that more fundamental reasons may be
responsible. In this review, therefore, we have attempted to step back and take a
wider view of some of these technologies, examining both their biophysical and
socio-economic aspects to try understand the reasons for their variable uptake by
farmers. We have focused on those technologies designed to address problems of
soil fertility and weed control; these include intercropping, alley cropping, cover
cropping and green manuring, biomass transfer, compost, animal manure, and
improved and enriched fallows.
II. THE TECHNOLOGIES
A. INTERCROPPING
Intercropping is the growing of two or more crops on the same piece of land
within the same year. Various forms of intercropping have been a central feature
of many tropical agricultural systems for centuries. Vandermeer (1989) has
proposed that intercropping can be divided into three general categories — full,
relay and sequential intercropping — depending on the extent of physical
association between the crops. Full intercropping involves complete association
between crops planted at the same time, while relay cropping involves only
partial association, in which a second crop is planted into an already standing
crop before it is harvested. Sequential intercropping, where there is no physical
association, is the extreme case where two crops are grown on the same land in
the same year but not at the same time. Cover crops are a special case of
intercropping and are discussed in more detail in Section II.C; in this section, we
discuss cases where the intercrop components are both food crops.
The main advantages of intercropping are in reducing the risk of total crop
failure, and in product diversification — food crops are often mixed with cash
crops to help ensure both subsistence and disposable income (Vandermeer, 1989;
Singh and Jodha, 1990). There is some evidence to suggest that in intercropping
systems, the microclimate surrounding the lower crop is more conducive to plant
growth than in a sole crop (Matthews et al., 1991), and that an intercrop is more
efficient at using resources such as light and water (Azam-Ali et al., 1990).
Cereals and legumes are often mixed, but probably more for dietary reasons than
for any beneficial effect from the nitrogen-fixing ability of the legumes. In
Zimbabwe, for example, farmers intercrop sorghum with cowpeas, pumpkins,
cucumbers and watermelon to provide nutritional and livelihood benefits
(Chivasa et al., 2000). Thus, any soil fertility benefits that can be obtained by
intercropping leguminous grain crops with other food crops should probably be
seen as a useful spin-off rather than the main purpose of the practice. Moreover,
main crop yields can even be reduced by intercropping techniques, both as a
result of loss of land to the legume, and also to competition for resources
(Vandermeer, 1989; Snapp et al., 1998). In the long-term, cereal/legume

intercrops are still likely to require fertilisers for the provision of phosphorus (P),
potassium (K) and micro-nutrients in order to maintain satisfactory yields
(Coultas et al., 1996; Kumwenda et al., 1997a).
Nevertheless, yield advantages through intercropping are well documented.
For example, Rao (2000) found that, compared to sole maize, maize intercropped
with pigeonpea (Cajanus cajan) in Kenya yielded 24% more and was 49% more
profitable, even though the pigeonpea was affected by pests and diseases. In an
interesting variation, sequential intercropping of rose (Rosa damascena) with
potatoes, maize and cowpea greatly increased the land equivalent ratio (LER) and
provided large economic gains (Yaseen et al., 2001). Similarly, positive effects
on soil fertility improvement have been observed. Kumara Rao et al. (1981)
estimated that leaf abscission during the growth of a pigeonpea intercrop was
equivalent to the addition of between 10 and 40 kg N ha21. The root system of
pigeonpea may also recycle N from deeper layers, and in some areas, the build-up
of sub-surface nitrates at about 1 –3 m has been observed (Farrell et al., 1996;
Hartemink et al., 1996). Morris et al. (1990) observed N transfer from arrow-leaf
clover to rye grass, and suggested that in mixed stands of legumes and
non-legumes, direct transfer of N during the growing season was possible,
although this was likely to be 10% or less of the total N fixed. It is also
possible that growing non-legumes with legumes encourages legumes to
respond by fixing more N than they might do in a pure stand, so long as the
legumes dominate the mixture (Marschner, 1995). Snapp et al. (1998) considered
soybean (Glycine max), pigeon pea (C. cajan), groundnuts (Arachis hypogaea),
dolichos bean (Dolichos lablab), and cowpea to be among the most promising
grain legumes in southern Africa for both food provision and fertility
enhancement.
Although grain legume intercrops can often help to increase the resource use
efficiency and stabilise yields of the main crop under optimal plant growing
conditions, this is not always the case. In India, Indonesia, and the Philippines,
Ali (1999) found that although intercropping could help to increase the yield of
rice, it also increased the variability of yield. Many green manure crops have been
selected on the assumption that maximum production is desirable. However,
yield stability under adverse conditions may be more important to many farmers
than high productivity under good conditions. Ironically, it is usually under
adverse conditions that intercrop competition is most intense, and it is during
these conditions that the farmer can least afford the technology to fail.
Intercropping is most likely to be practised on small farms, in areas where land
is scarce, forcing the simultaneous production of different crops on the same area
of land. For example, Ali (1999), in a survey of data from India, Nepal and the
Philippines, found that the attractiveness of intercropping increased as land and
labour costs grew. Lower rainfall and/or a unimodal distribution of rain may
encourage intercropping as farmers try to maximise their use of water, although,
in the extreme, this can result in competition for a scarce resource. Mixed-
cropping techniques are also more likely to be used by farmers relying on hand-
held implements for tillage (Ruthenberg, 1980). The need for simultaneous
production of different food crops and/or cash crops can also encourage
intercropping. Relatively better-off farmers with large farms are less reliant on
intercropping, being able to fallow and/or control production with other inputs
such as water and inorganic fertilisers.
The Machobane farming system in Lesotho (IIRR, 1998) is an example of a
system incorporating intercropping. Developed in the 1950s by an agronomist,
James Machobane, and based on experiments on his own farm, the approach is a
complex, integrated farming system designed to improve the productivity of
small-scale mountain farms in Lesotho. Based on 0.4 ha (1 acre), an application
of 7.5 t FW ha21 of a mixture of animal manure and wood ash (from household
ash) is made each year, with the proportion of each depending on soil type.
Enough for the 0.4 ha can generally be met from household waste (1 – 2 t year21
of animal manure, 2 t year21 of ash, (Pantanali, 1996)). Wheat, peas, and possibly
potatoes as a cash crop, are planted as intercrops in April –May for harvesting the
following January – March, and summer crops such as maize, beans, sorghum,
and possibly pumpkins and water melons, are planted in August –October for
harvesting in November –December (Fig. 1). Crop residues are left in the field to
allow humus to build up, and the field is ploughed only once every five years. The
intercropping/relay-cropping pattern allows food crops to be produced almost
all the year round, and there is always some crop cover throughout the year so
that erosion, a major problem in Lesotho, is minimised. Despite the crop cover,
weeding is essential, and represents a major labour input into the system. Some of
the crops (e.g., pumpkins) help to reduce pests, and the use of chemical pesticides
is discouraged. Overall, labour inputs are high, and perhaps reflecting this,
annual productivity is three times higher than the traditional system, allowing a
household of five people to be self-sufficient on 0.4 ha of land (Pretty, 1999).
Moreover, the potato crop is a source of cash, and income fluctuations over the
year are lower through less reliance on a single crop. Farmers also claim a better
resistance of crops to drought.
Promotion of the approach emphasises self-reliance, appreciation of the
resource base, readiness for hard work, learning by doing, and a duty to help
neighbours, on the part of farmers, and lack of success of the system is
sometimes blamed on non-compliance with some of those conditions (Pantanali,
1996). Due to the high labour inputs (peak labour demand estimated at 14 days
(0.4 ha)21 month21), critics of the system regard it as little more than “garden-
ing,” with limited relevance to Lesotho’s major agrarian problems. Unfortu-
nately, little data appear to have yet been collected to measure the impact
of the system in terms of production, gross income and net returns to
labour, either compared to traditional cropping methods, or to recommended
Figure 1 Crop calendar for the Machobane farming system in Lesotho (from IIRR, 1998).
“modern” improved technology practices based on monocropping, mechanical

power, chemical fertilisers, and pesticides (Pantanali, 1996).
Long-term sustainability of the system, however, depends on the production
of animal manure, and hence the availability of pasture, which fortunately is avail-
able in the mountains. As with many LEIA systems, therefore, there is a reliance on
nutrients collected and concentrated from a much wider area. More manure could
be made available if less was used for fuel, but in this case, fuelwood would have
to be grown, which again could be done on the land saved through intensification
of production. Thus, through careful integration of crops, livestock and trees, the
long-term sustainability of the system seems possible (Pantanali, 1996).
In relation to its applicability to other areas, the system’s economic sustain-
ability rests on being able to grow crops all year round, which will not be feasible in
areas with a pronounced dry season unless irrigation is available. Even in Lesotho,
the system cannot be practised in all areas due to severe winter conditions with
snowfall preventing the growth of many of the crops (Pantanali, 1996).
B. ALLEY CROPPING
Alley cropping is an agroforestry practice developed in the 1970s at the

International Institute for Tropical Agriculture in Nigeria (Kang et al., 1981), in
which hedgerows of trees and shrubs are established and annual crops are
cultivated in the alleys between the hedgerows. The hedgerows are pruned before
planting the crop and periodically while it is growing to prevent shading, with the
prunings being applied to the soil as green manure and/or mulch. Between
cropping cycles, hedgerows are usually allowed to grow without pruning. It was
originally hoped that by incorporating fast-growing nitrogen-fixing woody
perennials with crops, their ability to cycle nutrients, suppress weeds, and reduce
erosion would create soil conditions similar to those in the fallow phase of
shifting cultivation. In this way, the cropping and fallow phases could take place
simultaneously on the same land, allowing the land to be cropped for an extended
period when long fallow periods are not feasible under the particular socio-
economic conditions. Researchers saw the technology as the combination of
farmers’ accumulated traditional wisdom with the efficiency of modern science
(Kang, 1993).
Initial results from on-station experiments were promising. In Nigeria, for
example, prunings from Leucaena leucocephala increased maize grain yields
from 1.9 to 3.5 t ha21 (Kang et al., 1981), while a Gliricidia sepium alley system
on a degraded soil increased maize yields from 1.74 to 2.42 t ha21 (Atta-Krah
and Sumberg, 1988). Increases in yields of banana were obtained when alley-
cropped with Enterolobium cyclocarpum, and of cowpea when alley-cropped
with Enterolobium cyclocarpum and Dialium guianense (Oko et al., 2000). In the
fourth and fifth years of an alley-crop trial in Burundi, Calliandra calothyrsus
increased maize yields by 29 – 63%; Leucaena diversifolia by 27 –43%, and
Senna spectabilis by 24 –38% (Akyeampong, 1999).
However, these results were obtained in humid regions on soils of high base
status. Results from semi-arid regions were less positive. Yields of sorghum,
castor and cowpea were found to be lower when alley-cropped with Leucaena
than when grown alone (Singh et al., 1989), with the magnitude of the yield
depression correlating with distance from the hedgerows. This was attributed to
competition between trees and crops for water. Similar depressions of yield under
alley cropping were found in Peru (Szott, 1987) and Zambia (Matthews et al.,
1992a,b), although, in the latter work, Leuceana was found to have a positive
effect on maize yields if lime was applied first. In smallholder farms in southern
Africa, little benefit was derived from grain legume intercrops, particularly in
adverse weather conditions (Mukurumbira, cited in Snapp et al., 1998).
Hedgerow intercropping did not increase maize yields in below average rainfall
years, indicating that competition by trees predominated over benefits to soil
fertility (Snapp et al., 1998). Similarly, Vanlauwe et al. (2001) found reduced
crop yields under alley cropping in the absence of mineral fertilizers.
Competition, especially for light, is particularly acute with perennial legumes,
which are larger than the main crop (Ong, 1994). In general, it appears that where
resources are scarce, competition for resources such as water and nutrients is not
offset by benefits to fertility (Manu et al., 1994; Rao et al., 1997). Root pruning
has been found to reduce competition for water, but this adds to the labour
requirement, and may affect the ability of the tree to access water and nutrients
below the crop zone (Ong et al., 2002).
In some cases, other negative effects of trees on the crop have been found. For
example, Adeorike (2001) found that Inga edulis, Anthonatha macrophylla, and
Dactyladenia barterri had allelotrophic effects on maize. In Kenya, decreasing
the width between Leucaena hedgerows was found to increase the incidence of
angular leaf spot (Phaeoisariopsis griseola) and anthracnose (Colletotrichum
lindemuthianum) on beans (Phaseolus vulgaris), with the bean rows closest to
the hedgerows the most affected (Koech and Whitbread, 2000). Microclimate
changes caused by the hedgerows appeared to best explain the distribution of the
diseases on the beans. In some cases, more bird damage has been reported in alley
crops than in monocrops (e.g., Hoang Fagerström et al., 2001).
Some control of weeds was achieved in alley crops if the hedgerow canopy
was maintained during the fallow period. For example, in Nigeria, uncut
hedgerows of Gliricidia and Leucaena decreased the shoot biomass of the weed
Imperata cylindrica by about 80% (Anoka et al., 1991). Similarly, Yamoah et al.
(1986b) found that unpruned hedgerows of Flemingia macrophylla, Gliricidia
sepium, and Cassia siamea were able to reduce weed yields. Shifts in weed
composition were also observed — Siaw et al. (1991), for example, reported a
significant change towards more broadleaf weeds after alley cropping with
Leucaena and Dactyladenia barter. In most alley-cropping systems, the weed
suppression effect of the hedgerows probably has not been fully exploited, and
further studies of the effect of different hedgerow species, fallowing and
manipulation of cutting regimes may improve the effectiveness of the system in
reducing weed infestation.
Alley cropping does seem to have favourable effects on soil physical and
chemical properties through the addition of large amounts of organic matter from
the prunings. Levels of organic C, total N, extractable P, Mg and K, and pH have
been shown to increase under alley cropping under a range of conditions (e.g.,
Kang et al., 1985; Lal, 1989b; Dalland et al., 1993). Similarly, lower bulk density
and penetration resistance, and higher infiltration rate and pore volume fraction,
were found under Leucaena alley crops in Zambia, which was ascribed to
increased levels of soil organic matter (SOM) (Dalland et al., 1993). The
magnitude of these effects, however, varied with hedgerow species and
management — Leucaena prunings increased crop yields more than Flemingia
congesta due to the faster release of nutrients as a result of a lower C/N ratio.
There is also some evidence that nutrient recycling is enhanced by alley cropping
so that the downward displacement of nutrients is reduced. Hauser (1990), for
example, attributed higher concentrations of N, K, Ca and Mg in the surface soil
than in the subsoil under Leuceana hedgerows to leaf litter fall and nutrient
uptake by the trees from the subsoil. Between the rows, there were lower nutrient
levels in the surface soil due to crop uptake and higher levels in the subsoil
due to leaching.
Alley cropping also seems to have been successful in reducing runoff and soil
erosion on sloping land. For example, in Nigeria, Lal (1989a) found a reduction
of 73 and 83% in soil erosion under alley crops of Gliricidia and Leucaena,
respectively. In the Philippines, Paningbatan (1990) recorded soil erosion over
a three-month period of 41 t ha21 with Desmanthus hedgerows and contour
cultivation, and only 0.2 t ha21 with hedgerows, application of prunings as a
mulch, and zero tillage, compared to 127 t ha21 in the control treatment. In
northern Vietnam, contour planting of hedgerows on sloping lands reduced soil
loss from 18 to 7.4 t ha21 year21 and also produced 2.5 –12 t DM ha21 year21
for green manure (Nguyen The et al., 2001). Farmers were aware that soil loss
resulted from cultivating annual crops on sloping lands without adequate
protection, but were constrained by lack of labour and capital (Brodd and
Osanius, 2002).
Initial uptake of alley cropping by farmers in Nigeria and Benin was good, but
nearly half of those who adopted the technology subsequently abandoned it. In
Nigeria, no new adoption occurred after 1990 (Douthwaite et al., 2002). Follow-
up studies concluded that the initial enthusiasm shown by farmers was probably
more related to the incentives offered by researchers (free establishment of the
alley fields, weeding, provision of animals, animal vaccination, fertilisers and
seed of improved crop varieties), and to the prestige of contact with international
researchers, than to the characteristics of the technology itself (Whittome, 1994).
The Nigerian farmers gave the high labour demand for establishment and
management of the hedgerows, and incorporation of the biomass into the soil, as
the main constraints. Other studies have shown similar results (e.g., Reynolds
et al., 1991; David, 1995; Craswell et al., 1998), with many adopters specifically
citing the labour required for pruning as being the most difficult aspect of alley
cropping. Interestingly though, Hoang Fagerström et al. (2001) note no difference
in labour requirements between a monocrop and Tephrosia alley-crop in
Vietnam — the extra labour required for hedgerow management was balanced
by a reduction in labour associated with crop husbandry, such as weeding.
An abundance of available land has also been found to be a factor constraining
uptake of alley cropping – Whittome (1994), in his study of farmer experiences in
Nigeria and Benin, found that in most cases, land was still sufficiently abundant
for them not to consider soil fertility decline as a problem, and therefore not to
find alley cropping attractive. In Kenya, Swinkels and Frankel (1997) found that
alley farming was most attractive in areas where the population density was high,
farms were small, and labour was plentiful. Access to capital is another key
factor. For example, Cenas et al. (1996) have shown that adoption may be higher
where farmers have off-farm sources of income, relatively large farms, and were
interested in cash cropping. This was not only because of the high labour costs
involved in establishment of hedgerows (Nelson et al., 1996), but also because
alley cropping alone could not maintain total soil fertility requirements and
money had therefore to be spent on fertiliser for full benefits (Wendt et al., 1994).
Moreover, benefits from alley cropping take some time to accrue (Carter, 1995),
and resource-poor farmers may feel that these are not realised rapidly enough to
meet their current needs and that the long-term benefits do not outweigh the more
immediate costs of establishment (Nelson and Cramb, 1998). Indeed, traditional
cropping practices often create greater net revenue than alley cropping over the
first 4– 5 years (Nelson et al., 1996), and this may discourage the use of alley
cropping despite greater long-term benefits (Nelson et al., 1998).
Security of tenure and long-term access to land are important issues affecting
uptake in some countries. Tenant farmers, for example, are unlikely to want
to bear the full cost of the technique while the benefit is shared with the
landlord (Nelson et al., 1998). Similarly, systems based on revolving cultivation
of land amongst family members, short-term tenancy, and share cropping
tenancy arrangements may have the same effect. Where farmers have long-term
security of tenure over discrete areas of land, alley cropping may be more
relevant (Carter, 1995).
It is important to note that there are cases of farmers adapting the basic alley
cropping practice to fit their own needs. For example, in the Philippines, farmers
often increased alley spacing, planted single rather than double hedgerows to
reduce planting density, and reduced trimming frequencies and mulch application
(Garcia et al., 2002). Some farmers even used alternative tree species so that the
hedgerow could be used for other purposes. These modifications may have
reduced the value of alley cropping as a soil fertility-enhancing technique, but
have allowed it to fit within the constraints of the farmer and to answer a wider set
of needs. In some cases, there has even been an evolution from alley cropping
into intercropping two crop species — in eastern Indonesia, for example,
Harsono (1996) describes the replacement of hedgerows with strips of grain
legumes such as soybean which were shown to increase net profits. In other cases,
alley cropping has had some success for reasons other than soil fertility
enhancement or erosion control. For example, farmers have planted hedgerows
for the provision of poles, medicines, plants, fibre, fruit, and fuel (Cenas et al.,
1996), while in the Amarasi district of Indonesia, Field et al. (1992) noted that
Leucaena alley crops could provide fodder to allow farmers to develop intensive
livestock systems. The return from the sale of cattle could be used for the
purchase of food, and as farmers become less reliant on annual crops for
subsistence, more appropriate perennial plant systems could be established on
steeper land prone to erosion.
These modifications by farmers illustrate an important point regarding
adoption of alley cropping, or any technique for that matter — if it is to be
successful, it must address a range of farmers’ requirements, some of which
may not necessarily be related to the researchers’ original intentions, in this case,
those of soil fertility enhancement or erosion control (Douthwaite et al., 2002).
Alley cropping (and agroforestry technologies in general) must become more

“versatile,” capable of meeting a range of needs in response to changes in the
socio-economic circumstances of the farmer (Vosti et al., 1998). Gender issues
are also important — adoption is more likely if it is able to meet the needs of both
men (e.g., for poles or fodder) and women (e.g., for fibre, fuel-wood and mulch)
(Rocheleau and Rocheleau, 1990).
C. COVER CROPS AND GREEN MANURES
A cover crop is a crop grown to provide soil cover to prevent erosion by wind
and water, regardless of whether it is later incorporated. Green manuring involves
the incorporation of a crop while it is still mainly green into the soil for the
purpose of soil improvement. Cover crops and green manures are generally
annual, biennial, or perennial herbaceous plants grown in a pure or mixed stand
during all or part of the year, and as such can be seen as a special case of
intercropping. In addition to providing ground cover and, in the case of a legume,
producing N, they may also help suppress weeds and reduce insect pests and
diseases. Catch crops are cover crops that have been planted specifically to
reduce losses of nutrients by leaching following a main crop.
To compete with weeds, which by definition are aggressive plants, cover crops
need to have an appropriate canopy architecture. A spreading cover crop is more
likely to suppress weeds than a cover crop with an erect habit. For example, in
trials in rice systems, in the Ichilo Sara area of Bolivia, Pound et al. (1999) noted
that the performance of Arachis pintoi as a cover crop was highly variable and
rejected by farmers due to its inability to suppress weeds (especially Imperata
contracta), largely as a result of poor growth and lack of full cover. Similarly, in
Ghana, Jackson et al. (1999) found that C. cajan (pigeonpea) was slow-growing,
low yielding, and incapable of suppressing weeds due to its poor ground
coverage. However, even the use of aggressive cover crops that spread and shade
well may fail to suppress the growth of certain shade-tolerant weed species.
Pound et al. (1999) found that Mucuna pruriens and Calopogonium mucunoides
could not suppress the growth of weed species such as Axonopus compressus,
Cyperaceas, Panicum spp., Leersia spp., and shade-tolerant species such as
Drymaria, Commelina and Talinum.
The duration of the cover crop may also determine its effectiveness in
controlling weed growth. If its duration is less than the period between harvest of
the main crop and the planting of the next main crop, weeds may proliferate
during the time gap, and nutrients may even be released for uptake by weeds
rather than by the subsequent crop. For example, in experiments on winter cover
crops grown in rotation with rice, Pound et al. (1999) found that certain cover
crops with relatively short durations could actually increase the number of weeds
in comparison with the traditional practice of leaving a winter fallow. In these
cases, the weeds had time to multiply before the planting date of the next rice
crop, possibly taking advantage of N that was fixed by the leguminous cover
crops. Cover crops such as alfalfa, cowpea, carioca beans or Mucuna deeringiana
had very short durations, and tended to produce comparatively little biomass as a
result. Weed counts in these plots were substantially greater than for the
traditional fallow. Species with a longer duration, such as Mucuna pruriens,
C. cajan, and Canavalia ensiformis, continued growing up to the first planting of
the rice, produced more biomass, and continued to suppress the growth of weeds
over this time, although not significantly more than the traditional fallow
treatment as far as grass and broadleaf weeds were concerned. Rice yields also
showed no significant differences between different cover-crop treatments,
despite the large variation in grass weed density (Pound et al., 1999).
A cover crop with too long a duration may also cause problems. If the use of
cover crops in the system is to be sustained, particularly in isolated areas, farmers
need to be able to collect seed for the next season. If the crop fails to flower and
seed before the next main crop planting, then seed must be obtained externally.
Most legume genotypes appear to be adapted to quite narrow biophysical
conditions, and, therefore, must be tailored to specific environments if they are to
be successful. Keatinge et al. (1998) showed that Vicia faba, Vicia villosa ssp.
dasycarpa, and Lupinus mutabilis would be suitable as autumn-sown cover crops
across most of the mid-hills of Nepal if early sowing was possible. Vicia sativa
and Trifolium resupinatum, on the other hand, were only likely to mature
soon enough at lower elevations. Similar exercises were conducted for hillside
regions in Bolivia (Wheeler et al., 1999) in which potential cover crops,
not grown locally, were recommended for further trials, and also in Uganda
(Keatinge et al., 1999).
In some cases, the growth of cover crops grown in association with another
crop has been found to be too aggressive, resulting in undue competition with the
main crop for nutrients, space, water, and light. If left unchecked this can lead to
complete domination of the main crop by the cover crop, i.e., the latter is
beginning to behave like a weed itself. For example, Pound et al. (1999) found
that Mucuna pruriens was rejected by some farmers in the Ichilo-Sara area of
Bolivia because it dominated Bactris gasipaes, a local palm, and banana, limiting
their growth and development. Similarly, in on-farm trials of a Calopogonium/
rice intercropping system, farmers found that Calopogonium tended to climb
over the rice and cause it too lodge. This occurred especially when rice and
Calopogonium were sown simultaneously or when long-duration rice varieties
were used (Pound et al., 1999).
Nevertheless, cover crops have had some success in addressing problems of
soil fertility and weed control. It has been shown that short-term fallows of
herbaceous crops such as Mucuna pruriens (velvet bean) and Stylosanthes
hamata (Stylo) can help increase main crop yields compared with continuous
cropping, and that weed densities can be reduced (Tarawali et al., 1999).
Farmers seem to be well aware of these benefits (Buckles and Triomphe, 1999;
Franzel, 1999), and because of this, the adoption of cover crops by farmers has
been relatively widespread (Sanchez, 1999).
In Benin, for example, the ability of Mucuna grown as a cover crop to suppress
the weed Imperata cylindrica was discovered almost by accident by researchers
with the Recherche Appliqueé Milieu Reél (RAMR) Project, and was
subsequently promoted amongst farmers by various formal and informal sector
organisations. It was estimated that about 10,000 farmers tested Mucuna between
1988 – 1996 (Tarawali et al., 1999), although Elbasha et al. (1999) report that this
was on only 1000 ha. Other estimates put the adoption figure as high as 100,000
farmers (Versteeg et al., 1998). An adoption study by RAMR showed that about
25% of farmers had adopted the technique (defined as having used it at least
twice), whilst about 35% had rejected it despite still having an Imperata problem
on their fields (Versteeg et al., 1998). Adopters cited the need to control Imperata
infestations as the primary motivation for their using Mucuna, rather than soil
fertility enhancement, although apparently some reported benefits from higher
maize yields (3 – 4 t ha21 without application of nitrogen fertilizer and with less
labour input for weeding, compared to 1.3 t ha21 without Mucuna (Pretty,
1999)). Non-adopters said that leaving the field unproductive during the minor
season was a major disincentive, as well as the lack of a use for the grain
produced by Mucuna, which is toxic (due to the presence of l-Dopa) unless
treated properly (Versteeg et al., 1998). However, in reality, lack of a market was
not always a problem, as demand for Mucuna seed grew as use of the technique
spread. Interestingly, adoption was lower in areas where land was less scarce,
although some farmers in these areas discovered that Mucuna made good
fodder for livestock, and could also be used to suppress the parasitic weed
Striga hermonthica.
The benefit/cost analysis over a period of eight years indicated a ratio of 1.24
when Mucuna was included in the system, and 0.62 for the system without
Mucuna, with the ratio as high as 3.56 if Mucuna seeds were sold (UNEP, 1999).
Positive returns were achieved in the second year of establishment at both the
farm and regional levels. It has been estimated that Mucuna grown as a cover
crop can provide more than 100 kg N ha21 to a following maize crop, and, as
such, its adoption throughout the Mono Province of Benin would result in savings
of about 6.5 million kg of N, or about US$1.85 million year21. However, yearly
analysis of the benefit/cost ratio indicated a declining trend over time, suggesting
that addition of external inputs (probably P and K fertilizer) are required in order
to achieve full sustainability (Pretty, 1999).
Widespread adoption of Mucuna as a cover crop has also been evident in
Honduras, where, without any extension support, it spread from farmer to farmer
since the 1970s, when, due to population pressure on the plains, hillside areas
were first used for agricultural production (Buckles and Triomphe, 1999).
Traditional practice was to crop maize twice a year in the wet and dry seasons for
two years, then return the field to fallow for four years, a system which was labour
intensive (land preparation, planting, weed control), land extensive (due to the
need to have additional plots to meet food requirements during the fallow), and
capital intensive (due to outlay on herbicides and labour for weeding). In the new
system, farmers plant Mucuna as a relay crop into the maize in the dry season so
that it grows alone over the wet season, suppressing other weed species at the
same time. At the beginning of the next dry season, the farmer slashes the
Mucuna and sows the following maize crop directly into the decaying mulch, and
so the cycle continues. The system gives benefits in terms of reduced labour
(15 – 20% less) and increased yields after the second year — while the traditional
system provided four harvests over six years from a single plot, the maize –
Mucuna system produces six harvests with yields 50 –100% higher.
On average, those who have adopted the Mucuna technique planted twice as
much maize as those who did not. Despite this, the total amount of land occupied
by their cropping system was less, as they no longer needed large areas to fallow,
although, interestingly, overall deforestation rates continued to increase due to an
influx of migrants into the area (Humphries, 1996). Experimental evidence
indicated that the system was capable of maintaining soil N and OC, Ca, pH and
P levels. This was largely achieved through a large biomass production of about
10– 12 t ha21, and large amounts of N (about 300 N kg ha21) being contributed
through this biomass, although, of course, only a proportion of this represents a
net addition to the system through nitrogen fixation. By 1992, 65% of farmers
were using the maize – Mucuna system, with a further 19% having used it in the
past (Buckles and Triomphe, 1999).
However, more recently there has been a widespread decline in the number of
farmers using the system; such that with abandonment running at 10% per year,
only 39% of farmers were still using it by 1997. Neill and Lee (2001) provide
interesting insights into why this has occurred. Their surveys showed that there
was no single overriding factor involved. Firstly, a minimum farm size of 2– 3 ha
is required to meet household food requirements during the wet season while
some fields are under Mucuna, and farmers must have security of land tenure to
adopt the system. However, over the last 30 years, there has been a steady decline
in farm size and insecurity of tenure has increased. Secondly, the increase in
extensive cattle production in the region has decreased the availability of land for
rent and off-farm work, both necessary requirements for the smallest farmers if
they are to adopt the maize –Mucuna system. Thirdly, improved road access in
the region has made other alternatives, such as fruit-trees and off-farm work,
more attractive than maize growing. Fourthly, the incursion of the noxious weed
Rottboellia cochinchinensis into the area in the last few years has drastically
reduced maize yields (by 50 –72%) by establishing in gaps within the Mucuna
caused by farmers relying on it to regenerate itself rather than deliberately
reseeding it at the beginning of each new season. The presence of Rottboellia has
negated the two advantages of the maize – Mucuna system — that of lower
labour requirements and higher productivity. Thus, it is clear that the

sustainability of a technique is a function both of the agronomic performance
of a system, and of the socio-economic context in which it is located, including
the knowledge and understanding of the farmers. Neill and Lee (2001) make the
point that farmers’ inadequate grasp of agronomic aspects such as reseeding,
nutrient cycles, nitrogen fixation and herbicide effects, brought about by too rapid
adoption of the system through hearing about it from neighbours, may also have
contributed to the decline of the system.
In both the West African and Central American cases, adoption of Mucuna
cover cropping was due to a specific set of biophysical and socio-economic
circumstances, which may not always be the same in other countries. In Benin,
the critical conditions associated with its adoption were a decline in soil fertility,
the lack of fertilizer, land scarcity, and weed encroachment, all of which
combined to induce farmers to adopt a new technique which might not otherwise
have been accepted (UNEP, 1999). In semi-arid Africa, however, establishment
and management of cover crops has proven difficult where rainfall is less than
800 mm or on very clayey soils, so the practice is not likely to have much success
there (Ganry et al., 2001). In Hondurus, adoption was widespread as farmers
appreciated the benefits of the technique, which included higher maize yields,
improved soil fertility with ease of land preparation, and moisture conservation,
with weed control and erosion control of lesser importance (Buckles and
Triomphe, 1999). The seasonality of maize prices also encouraged the uptake of
the technique, as maize planted during the dry season commanded a higher than
average value. These factors all helped to improve productivity both to land and
to labour. The benefits were such that farmers who rented land were even willing
to pay a premium on land that had been under the Mucuna technique. This also
encouraged landlords who owned more land than they were able to cultivate
under maize to invest in the technique. Thus, while rental of land generally
discourages investment in techniques by tenants, the availability of land for rent
and the value placed on Mucuna-treated fields encouraged the initial spread of the
technique in this particular case.
In general, therefore, intermediate intensities of land use are likely to be the
best context for cover crops, as the opportunity cost of land will be high in land
extensive systems, while the opportunity cost of labour and capital will be high in
land extensive systems (Tarawali et al., 1999). Security of land tenure is also
important as several years are required to reduce weeds and enhance soil physical
and chemical properties to a level that might sustain another arable rotation
(Tarawali et al., 1999). Some financial and labour investment in herbicides and
fertiliser may also be necessary as evidence suggests that cover crops are not
capable of indefinitely sustaining arable crop production, even in rotations
(Tarawali et al., 1999). Purchase of seed can also be costly and this has been
identified as a major constraint in some areas. Those farmers with intermediate
levels of wealth and/or off-farm incomes may prefer to use cover crops in
improved fallows (Tarawali et al., 1999), as the opportunity cost of their

labour is relatively high (Franzel, 1999). Poorer farmers are likely to make use of
natural fallow unless they are provided with credit facilities and/or other
incentives (Tarawali et al., 1999) or have a large labour pool. However, the
opportunity cost of capital, land, and labour may be relatively high. Wealthier
farmers may decide to use inorganic fertilisers during the cropping cycle,
although as they often have fairly large farms, part of the farms may be under
fallow. Labour demand and the timeliness of that demand may also be proble-
matic and cover crops will probably have the best chance of being adopted by
households with sufficient surplus labour. This can sometimes be difficult as
labour availability in rural areas often declines as farmers attempt to broaden
their livelihood strategies with off-farm work.
In other areas, cover crops have been found in niche uses with high-value tree
crops, even though little extension effort has been made for this. For example,
Thomas et al. (1991) discuss the use of cover crops as green manures as a source
of nutrients for coconut. Similarly, Lehmann (2000) found that Pueraria
phaseoloides cover crops in tropical fruit-tree systems enhanced nutrient cycling
and reduced leaching, although competition for other nutrients (e.g., P) was a
problem. In the tropical rainforest of Nigeria, farmers found that cover cropping
was the most effective way of controlling soil erosion (Akinyemi and Kalejaiye,
1998). Another possible use is for feed meal production (Stur et al., 2002).
It should be noted that cover crops have not always controlled weeds or
increased crop yields, primarily due to competition with the main crop. For
example, maize intercropped with the cover crops Pueraria phaseoloides, Vigna
unguiculata (cowpea), and Mucuna pruriens in SW Nigeria yielded 2.15 t ha21
(þ 15% increase), 1.92 t ha21 (þ 3%), and 1.71 t ha21 (2 9%), respectively,
compared to the 1.87 t ha21 for maize grown with no cover crop, the differences
reflecting the relative competitive ability of the legume crops (Kirchhof and
Salako, 2000). In some cases, the decomposition of green manure (particularly
legumes) can release organic acids and/or other compounds which may affect
germination, seedling growth, and yield of the following crop (Boddey et al.,
1997). In Costa Rica, intercropping maize with Mucuna deeringiana and
Canavalia ensiformis helped to control itchgrass (Rottboellia cochinchinensis)
infestation to some extent, which usually resulted in improved maize yields,
although on occasions yields were lower due to competition from the cover
crop, especially Mucuna (Valverde et al., 1999). Similarly, in Benin, although
the main reason for the widespread adoption of Mucuna cover cropping was its
ability to control weeds, it was found that Imperata, in particular, was
suppressed but not eliminated, and that it regained its original strength after
one or two years of maize cropping (UNEP, 1999). This was because Mucuna has
a limited effect on the rhizomes of Imperata (Akobundu et al., 2000), which
suggests that eradication of Imperata may require a more integrated approach
using cover crops in combination with other management techniques such as

tillage or herbicides.
The use of cover crops (traditional or new) may also sometimes encourage the
introduction of pests, which may hamper the growth of the main crop or cause
harm to the farmers themselves. In Bolivia, Pound et al. (1999) noted the
increased presence of rats, snakes, and red spider mites with the use of Pueraria
phaseoloides and Arachis pintoi as cover crops, while in West Africa, one of
the constraints to adoption of mucuna cover crops was the fear possessed by
farmers for the presence of snakes within the canopy (Galiba et al., 1998).
Another concern with the use of green manures in flooded systems such as rice
agriculture, which is receiving increased attention, is its influence on methane
emissions into the atmosphere and subsequent contribution to global warming
(e.g., Matthews et al., 2000).
D. BIOMASS TRANSFER TECHNIQUES
In an effort to relocate nutrients from forests to agricultural land, tropical

farmers have traditionally used a variety of biomass transfer techniques (e.g.,
Young, 1987; Nyathi and Campbell, 1993). In most cases, this has involved the
use of naturally occurring biomass (i.e., tree or grass material), and rarely
biomass that has been specifically planted for that purpose. Recently, however,
the attention of researchers has focused on transfer of biomass from deliber-
ately planted “biomass banks” of species such as Tithonia diversifolia (ICRAF,
1997; Gachengo et al., 1999; Jama et al., 2000), Gliricidia sepium (Rao and
Mathuva, 2000), Calliandra calothyrsus, and L. leucocephala (Mugendi et al.,
1999) as a means of providing nutrients for crop growth, and organic material
for physical improvement of the soil. The use of so-called cut-and-carry
grasses is another technique where biomass is harvested and transported, in
this case specifically to provide fodder for animals (e.g., Tanner et al., 1993;
Stur et al., 2002).
While similar in principle to alley cropping in that plant biomass is cut and
incorporated into the soil to release nutrients for crops and to help improve SOM
levels, one of the potential advantages of biomass banks is that direct competition
between the main crop and that used to supply the biomass is minimised, if not
eliminated altogether. The evidence so far suggests that biomass transfer
techniques can help to increase soil fertility and sustain or increase crop yields
(Mugendi et al., 1999; Rao and Mathuva, 2000). However, for the technique to be
successful, the quality of this biomass needs to be high (Snapp et al., 1998), and
very large amounts of biomass are required to supply “ideal” quantities of
nutrients to crops (Gachengo et al., 1999). Labour for the collection,
transportation, and incorporation of the OM into the soil must also be plentiful
(Snapp et al., 1998; Gachengo et al., 1999; Jama et al., 2000). For example, in the
subhumid highlands of Kenya, biomass transfer of Caliandra calothyrsus and

Leucaena leucocephela prunings was found to increase maize yields compared to
both unfertilised and fertilised sole maize (Mugendi et al., 1999). Similarly, Rao
(2000) found that transfer of biomass from a Gliricidia stand increased maize
yields on an equivalent area by 27%, but the high labour cost for harvesting,
transfer, and application of prunings, made the technique uneconomical. These
issues are discussed in more detail later in this review. In Vietnam, Hoang
Fagerström et al. (2001) found that Tephrosia grown in a “biomass bank” and
used as mulch on rice fields increased production over a four-year period by 55%
compared to the monocropped control, if the extra land used for growing the
Tephrosia was not included. This increase dropped to 29% if the total land area
was taken into account.
The use of the shrub Tithonia diversifolia (Mexican sunflower) as a source of
P has also received some interest recently (e.g., Jama et al., 2000). Tithonia
appears to have the ability to extract relatively high quantities of P from the soil
giving it a high leaf P content (. 0.25%), and although it is a non-leguminous
species, it also has a relatively high biomass N content of about 3.4% (Gachengo
et al., 1998), above the level required to prevent net immobilisation of N (Palm
et al., 1997b). In addition, its leaf lignin and polyphenol contents are less than 15
and 4%, respectively, so that it decays relatively easily. As it grows abundantly in
the wild in many areas, and indeed, is often cultivated as a farm boundary hedge
in Africa and Asia, it has been suggested for use as a biomass transfer technique
(ICRAF, 1997; Jama et al., 2000). Experimental evidence suggests that addition
of N and P through the application of Tithonia biomass may increase yields more
than the use of equivalent quantities of mineral N and P (Jama et al., 2000). This
has been attributed to the presence of K, Ca, and Mg in the biomass which might
ameliorate deficiencies of these nutrients in the soil, and possibly also because of
an improvement to soil physical characteristics. In addition to providing nutrients,
Tithonia has been shown to reduce P sorption and increase soil microbial biomass
(Jama et al., 2000).
Various techniques have been found to increase Tithonia biomass production
and therefore the quantity of nutrients extracted from the soil and made available
for transfer elsewhere. The use of woody cuttings for propagation instead of soft
stem cuttings, for example, was found to increase its biomass production, as was
the application of mineral P fertiliser (Jama et al., 2000). However, its ability to
extract P appears to be less if it is grown as a stand rather than as a hedge (George
et al., 2001), suggesting that it is taking up P from a wider area than just that
occupied by the hedge. It may, therefore, not be so effective if grown as a biomass
bank. The P content of the soil also influences the concentration in the leaves, and
continual P-mining through removal of Tithonia prunings is likely to eventually
reduce the quantity and quality of the biomass to the level below which
immobilisation of P is likely to occur (, 2.4 g P kg21). Moreover, it needs to be
remembered that only the importation of Tithonia biomass will result in a net
increase of P on the farm — that taken from boundary hedges will merely
redistribute it within the farm.
Most literature on established biomass banks appears to be research oriented
in nature, and little evidence exists of their deliberate and successful introduction
onto farm land, probably for the reasons mentioned above. As with many other
LEIA technologies, if biomass transfer techniques are to be used, considerable
areas of land will be needed to grow sufficient quantities of plant material, which
is clearly a limitation if land is scarce and farms are small. Consequently, it is
unlikely that biomass banks for the sole purpose of soil fertility enhancement are
likely to be widely adopted by farmers. There is also the disadvantage that there is
a relatively long time-lag for benefits on soil fertility of such techniques to accrue
(Snapp et al., 1998). Farmers may be interested in biomass banks if they cannot
effectively use all their land for cultivation. However, in such cases, it is more
likely that they may opt to fallow this land to regenerate soil fertility. Biomass
banks could be established on strategically located common land, which would
be especially valuable for poor farmers, but for this to be workable, it would be
necessary for a system of access agreements to be developed. The question of
who was responsible for the establishment and maintenance of such stands would
also need consideration, particularly as there are costs involved in purchase of
seeds and seedlings, planting, and the care required during the initial establish-
ment (i.e., weeding, etc.). Similarly, the sustainability of the system, with
constant removal of nutrients in the biomass, would need to be addressed. Snapp
et al. (1998) have noted that even large naturally occurring woodlands, such as
the miombo woodlands in southern Africa, cannot be indefinitely mined for
nutrients.
In most cases, it is unlikely that biomass transfer techniques will be capable of
supplying the full fertility needs of a farm, and as with other LEIA fertility
enhancing techniques, it may be best to see them as a component of an integrated
nutrient management (INM) system involving external supplies of inorganic
nutrients. It may be that there are specific niche roles which will make them
useful on small areas within a single farm (e.g., for home gardens), or on
degraded common land, although it is more likely in this case that they will fulfil
other important needs, such as the provision of fuel-wood and/or fodder. Here
they would move closer to the role played by natural forests, in which case, they
may help relieve some of the pressure on the latter.
However, there is a greater likelihood of adoption of biomass transfer
techniques where there is some immediate benefit to be obtained by the farmer.
The use of cut-and-carry grasses as animal fodder, for example, has the advantage
that some animal products, including milk, meat and draught power, are available
almost immediately. Improvements to soil fertility through the application of
manure and urine may be a secondary result. Farmers are usually well aware of
the beneficial effects of manure on soil fertility — there is some evidence that
they feed their livestock much more than required for optimal live-weight gain, in
order to provide manure for arable crop production (Tanner et al., 1993).
Nevertheless, the essential difference of “processing” plant biomass through
animals first to gain immediate benefits (e.g., meat, milk, and draft power), rather
than using the biomass directly to improve soil fertility, is likely to be a
determining factor of whether biomass banks are adopted by farmers or not. Of
course, the intensification of agriculture with cut-and-carry grasses and/or fodder
banks is most likely to occur where animals are already a major component of the
agricultural system and where satisfactory and alternative feeding strategies
do not already exist (Reynolds et al., 1991; Rachmat et al., 1992). In some areas,
the importance of cut-and-carry and zero-grazing techniques may increase as
population increases and less and less land is available for free-grazing on
communal land (Murwira et al., 1995), or where access to naturally occurring
vegetation is limited, either because it does not exist or because access to it is
blocked. However, in either case, sufficient land still needs to be found
somewhere to grow the biomass (Gashaw et al., 1991).
As with other biomass transfer technologies, the evidence indicates that the
amount of labour required for cut-and-carry techniques for fodder provision is
often a disincentive to adoption (Mogaka, 1993; Wandera et al., 1993; Sanchez
and Rosales Mendez, 1999). Cut-and-carry grasses may also not supply the
full fodder requirements of livestock, necessitating supplementary feeding.
Capital will be needed to pay for this, and for the inevitable veterinary fees
associated with keeping livestock healthy (Mogaka, 1993). In addition, the
decline of productivity of the fodder banks, as nutrients are removed may require
investment in fertilisers to maintain productivity (Wandera et al., 1993). Finding
suitable grasses, as well as issues related to land tenure are other important
considerations. On the whole, the deliberate establishment and maintenance of
fodder banks and cut-and-carry systems is probably most likely where farmers
have some surplus capital and labour, but where land scarcity has resulted in
limited access to natural vegetation. The ownership of cattle and the establish-
ment of biomass stands involve costs that very poor farmers are unlikely to be
able to meet.
E. COMPOST
Compost is the aerobic, thermophilic decomposition of organic wastes to a

relatively stable humus. Although it makes use of the same decomposition
processes occurring naturally, the aim in compost-making is to control the
conditions to a level that allows faster decomposition. The biophysical conditions
that are required for effective composting are generally those that are required by
the micro-organisms at various stages of the composting process, i.e., good
moisture levels, moderate temperatures, mixed quality organic matter, and a
fairly neutral pH range.
Composting is not a new technique for the improvement of soil fertility and
structure, and tropical farmers have been aware for centuries of its impact on crop
yields, soil structure and fertility, crop growth and vigour (Diop, 1999; Onduru
et al., 1999). For example, Ouédraogo et al. (2001) observed that yields of
sorghum in Burkino Faso could be tripled by the application of 10 t ha21 of
compost. Another benefit noted is the reduced need for capital inputs (Onduru
et al., 1999), although some capital may be necessary for farmers to adopt the
technology (Girish and Chandrashekar, 2000; Slingerland and Stork, 2000).
Compared to techniques such as alley cropping, composting in relation to
subsistence agriculture seems to have received little attention by researchers.
Much of the literature available tends to be written from a purely technical
standpoint, and is often from the perspective of agriculture in developed
countries. Relatively few studies have considered the on-farm issues of using
compost, although the few that do give a good indication of the constraints
involved.
The major problem associated with the use of compost is the high labour
requirements (Onduru et al., 1999). In particular, female-headed households can
have considerable difficulty in undertaking some of the heavier tasks involved in
composting, such as preparing compost pits (Diop, 1999). In Uganda, Briggs and
Twomlow (2002) found that the poorest households did not make compost at all
because of the labour and time requirements. Transportation of biomass and
compost is also problematic (Apiradee, 1988; Adeoye et al., 1996). Also, like the
other LEIA techniques already discussed, large quantities of biomass are
required, and questions arise as to where farmers can obtain this (Onduru et al.,
1999; Ouédraogo et al., 2001). This is particularly relevant where there are
competing demands for such resources, for example as mulch, fuel or fodder
(Drechsel and Reck, 1998), and where land to produce the biomass is scarce.
Resource-poor farmers may have problems providing land for processing of
“ideal” quantities of biomass, although this is not generally cited as a limitation,
probably because fairly small quantities of compost are usually produced. On the
other hand, some systems seem capable of producing quite large quantities;
Briggs and Twomlow (2002) found that smallholder households in Uganda
produced 40 kg of fresh organic waste per day, or about 9.2 t DM year21, 25% of
which was used to make compost, with the rest either fed directly to livestock or
applied directly to household plots. Composting may sometimes be constrained
by lack of water (Apiradee, 1988; Diop, 1999), which is needed to aid
decomposition, and by lack of biostarter (Apiradee, 1988), although it appears
that animal manure and inorganic fertiliser may be used. Lack of tools for
composting may also be a problem in some areas (Ouédraogo et al., 2001).
An example of the use of compost in subsistence agriculture is provided by
the work of the Rodale Institute, which has been working since 1987 with farmers
in the Peanut Basin region of Senegal to help offset the rising cost of fertiliser due
to removal of subsidies (Diop, 1999). Composting is not a new technique in
Senegal, so research has focused on attempting to improve existing techniques

involved in its production and use. Farmers are encouraged to collect crop
residues for compost making rather than burning them, and to incorporate the
resulting compost in the soil rather than leaving it lying on the surface. Pit
composts are also being developed — these are about 1 m deep, lined with
cement and bricks, and covered to prevent contamination from wind-blown sand.
In the simplest and most efficient approach, which does not require turning,
rainfall in the wet season is used as a water source, with compost being produced
within 18 months. In a more labour-demanding approach during the dry season,
water is added manually when the compost humidity drops below a certain level.
The compost must be turned every 15 days, but is ready within 45 days. For
larger-scale compost making with this method, three pits are used with compost
being transferred from one pit to the next every 15 days. Yields of both groundnut
and millet, the traditional crops, have been approximately tripled through the
application of 2 t ha21 of compost (Diop, 1999), although the area of crop
involved in this case is not stated.
Another example is the Zai system, traditionally used by farmers in Burkina
Faso in times of drought (Roose et al., 1999). The system involves digging
holes (20,000– 25,000 ha21), typically 30 cm wide and 20 cm deep, filling them
with compost, and planting seeds of sorghum, millet, and cowpea into them.
The compost is made from farmyard manure, plant residues, garbage, and
rock phosphate — a natural product from mines in Burkina Faso. Crop yields
can be more than 10-fold higher using the system than otherwise (from 150 to
1700 kg ha21), and the holes can be reused for 3 years. Trees may also sprout
spontaneously in the holes.
As composting is labour intensive, it is probably most appropriately used close
to the homestead, on specific crops (e.g., Briggs and Twomlow, 2002). Evidence
also suggests that progress may be made by improving the technical knowledge
of farmers, so that composting practice is improved (e.g., Sutihar, 1984; Adeoye
et al., 1996; Wakle et al., 1999). For example, the quality of compost can be
greatly enhanced by mixing it with a combination of inorganic chemicals (van
den Berghe et al., 1994), or by combining it with manure (Onduru et al., 1999).
During processing, protecting it from heat and direct light may reduce
volatilisation of the nutrients, while protecting it from rain may prevent similar
losses by leaching (Diop, 1999). It seems unlikely that composting would be a
new technique in many areas, unless a rapid transition from an abundance of
land to scarcity is occurring. Progress is more likely to be made by determining
whether composting practices can be improved within the socio-economic
constraints of such farmers, such as with the Rodale Institute work. Recommend-
ing the use of animal manure in compost, for example, would not be appropriate
if animal manure is the only source of fuel.
Thus, as with other OM management techniques, composting is likely to
provide only a partial solution to the problem of decline in soil fertility and soil
structure in subsistence agriculture, but should be an important component in

the basket of options that a farmer could use. Since composting is very labour
intensive, it will probably continue to be most used close to the homestead, or on
small areas of high-value crops. Extending the use of composting to large or distant
fields may require unrealistic levels of labour for most farmers.
F. ANIMAL MANURE
As with several of the other LEIA technologies, the beneficial impacts of

applying animal manure to land are well known to many tropical farmers and
are also well documented in the scientific literature (e.g., Prudencio, 1993). As
already mentioned in relation to the Machobane System, the use of animal
manure as a source of crop nutrients is often a nutrient-harvesting technique in
which nutrients are gathered through grazing of a relatively large area and
concentrated on a smaller area where crops are grown. Even where animals are
stall-fed, the nutrients they consume must be brought to them from elsewhere,
either as collected fodder or as purchased concentrates.
The rumen provides an ideal environment for the decomposition of organic
matter, and is a way of improving the rate of decomposition in suboptimal
environments, such as those with low temperatures or dry conditions. Livestock
in developing countries generally feed on low-quality grasses and/or crop
residues, which have significant fractions of indigestible materials such as lignin
and cellulose – lignin complexes, as well as low N contents (Leng, 1990).
Resource-poor farmers often cannot afford to purchase supplements, and instead
use tree and forage legumes to improve poor-quality diets (Coppock and Reed,
1992). The quality of a diet can influence feed intake, feed digestibility, and the
partitioning of N between the urine and faeces (e.g., Somda et al., 1995). This
latter characteristic is important, as a large proportion of the N excreted in the
form of urine can be lost from the system through volatilisation and leaching,
whereas that in faeces can be collected, stored and used to benefit crop growth
(Delve et al., 2001). The manure of animals which are fed highly digestible diets
is more susceptible to N losses than that of animals fed a greater amount of
roughage (Powell and Williams, 1993), while feeding on browse species has been
found to shift the balance of excreted N away from the urine more towards the
faeces (e.g., Reed, 1986). For example, Coppock and Reed (1992) found that
animals fed Acacia tortilis pods excreted only half as much N in urine as those
fed cowpea hay. Similarly, inclusion in the diet of Flemingia congesta, which has
a high tannin content, was found to reduce N excreted in the urine by up to 35%
(Fassler and Lascano, 1995).
Hypothetically, a farmer with access to organic material is faced with a choice
of whether to apply it directly to a crop as an organic fertiliser, or to use it
to feed livestock first with the manure produced then being applied to the crop.
The optimal choice depends on the quality of the biomass available. It is

generally accepted that material with a C/N ratio of more than 20 kg C (kg N)21
results in immediate immobilisation of N by microbial biomass when
incorporated into the soil (Senesi, 1989) which is likely to reduce crop yields
by reducing the amount of mineral N available. Passing this type of material (e.g.,
cereal straws) through an animal rumen first can result in a large reduction in its
C/N ratio, thereby reducing the amount of immobilisation. For example, Delve
et al. (2001) found that cattle faeces from a diet of barley straw had a C/N ratio of
27 kg C (kg N)21 compared to that of 86 kg C (kg N)21 of the plant material
only; consequently on incorporation, the faeces had a faster mineralisation rate
and showed less net N immobilisation than did the fresh barley straw alone. Crop
uptake of N was also greater from incorporated faeces than from the fresh straw.
On the other hand, incorporation of high-quality plant material usually results in a
higher recovery of N than of faeces derived from the same plant materials (e.g.,
Catchpoole and Blair, 1990). Delve et al. (2001) found that leaves of Calliandra
calothyrsus with a C/N ratio of 13 kg C (kg N)21 resulted in steady net N
mineralisation, but that manure from a diet containing 30% Calliandra
calothyrsus resulted in net immobilisation between 3 and 16 weeks. The results
from these studies indicate that the effect of passing material with C/N ratios
ranging from 13 to 86 kg C (kg N)21 through an animal rumen is to make the C/N
ratio converge to a value between 20 and 27 kg C (kg N)21. Thus, from a nitrogen
supply point of view, it would seem better to apply high-quality organic matter
(C/N ratio , 20 kg C (kg N)21) directly to the soil, but to pass low-quality
biomass through an animal first, and then apply the manure to the soil. Of course,
this does not take into account other benefits such as a better supply of milk and
meat that might come from feeding animals high-quality fodder.
The beneficial effects of animal manures on crop yields, when applied in
sufficient quantities, are well documented. Not only can they enhance immediate
crop yields (Selvarajan and Krishnamoorthy, 1990; Ali, 1996; Drechsel and
Reck, 1998), they can also provide residual benefits for following years (Singh
and Desai, 1991). However, Giller et al. (1997) note that, although animal
manures are of major importance in nutrient cycling, they are generally of poor
quality as a supply of plant nutrients. In Rwanda, for example, Roose et al. (1997)
found that the application of manure at 10 t ha21 had no effect on maize yields.
Even the effect of applying manure at the rate of 20 t ha21 was limited to a single
season, as the yield of the second crop (sorghum), showed no significant differ-
ence compared to the other treatments, which was ascribed to the low amount
of P supplied in the manure. The benefits to the soil from applying manure,
therefore, may be more related to improvements in physical characteristics rather
than the provision of nutrients, especially in the quantities that farmers can supply
(Singh and Desai, 1991).
It is important to remember that the use of manure is related to its role as
part of a larger system. In many subsistence-farming systems, there is a close
interdependence between crop production, animal production, and possibly

neighbouring forests and/or rangelands. In Nepal, for example, animals are
grazed in the forest, crop residues and forage collected from the forest are used to
feed housed animals, animal manure is applied to cropped areas, crop residues
and forest litter are used for animal bedding, and animals are used to provide
tillage and transport (Pilbeam et al., 1999b). In certain areas of western Africa,
some arable farmers make arrangements with itinerant herdsmen to corral
livestock on their land (Waldie, 1990; Enyong et al., 1999). Farmers may also
move their homesteads from place to place so that crops can be grown on that
land to benefit from the manure left over by livestock (Ruthenberg, 1980). In
upland Java, livestock may be fed with far greater amounts of biomass than is
needed for optimal live-weight gain, the rationale being the production of
manure-compost that is collected for intensive upland agriculture (Tanner et al.,
1993). In smallholder farms in Uganda, Briggs and Twomlow (2002) calculated
that an average of 4.6 t of manure was produced annually from a household with
2.2 ha of land — this was mainly from goats grazing on communal land during
the day, but tethered near the household at night — a figure that could be
increased if better manure management was practised.
In many cases, there may be a high opportunity cost of using manure as a
fertiliser, and farmers may often value it more for uses other than soil fertility
maintenance. Benefits obtained from manure include the provision of material for
plastering and building, and fuel for heating and cooking (Jeffery et al., 1989;
Murwira et al., 1995). Provision of milk for human consumption from keeping
livestock has already been mentioned. Many of these activities (e.g., the
production of fuel cakes or milk) have direct economic value in themselves
(Jeffery et al., 1989). As discussed for the Machobane system, competing uses of
manure inevitably reduce the amount available for soil chemical and physical
improvement.
If manure is to be extensively used to enhance soil fertility, it will need to be
culturally acceptable to farmers, which is most likely to occur where livestock are
an integral part of the farming system already. This is more likely to be in areas
where population pressure is higher, labour availability is higher, and land is
scarce, such as in Nepal (Murwira et al., 1995). Where land is not scarce, farmers
may find it more practical to improve soil fertility and structure through natural or
improved fallow, although they may still use manure on plots or kitchen gardens
close to the homestead if they own poultry, smallstock, or cattle.
Both the production and use of manure are labour intensive (Enyong et al.,
1999). Households without adequate labour, or the means to procure it (e.g., in
communal work groups or through purchase), may only be able to use small
amounts of manure. In general, investment in manure to improve soil increases
where an enabling environment for agricultural production is provided. For
example, in areas where there are market outlets and effective extension networks
providing technical guidelines on good manure practice, farmers may be
encouraged to increase productivity, stimulating the use of manure as one of the

various options available for increasing crop production (Enyong et al., 1999).
However, the fragmentation of fields occurring in many developing countries may
also make it more difficult to transport manure, reducing farmers’ willingness to
apply it to fields at a distance from the homestead (Enyong et al., 1999).
In general, as with other LEIA techniques, the use of manure should be seen as
a component in an INM system. Indeed, Sherchan et al. (1999) found that the
highest yields of maize and wheat in Nepal were obtained from a combination of
manure and inorganic fertiliser. Alternatively, it could be used in certain niches,
say around the homestead with specialised high-value crops, where the distance
to transport it is not a major constraint.
G. IMPROVED FALLOWS
Shifting cultivators have traditionally alternated periods of crop production

with periods of fallow in order to restore soil fertility and suppress weeds. In some
cases, the cropping period only lasts 4– 5 years while the fallow period may be as
long as 30 years (e.g., Matthews et al., 1992a), during which time the land is
usually unproductive in terms of generating a livelihood. In recent years,
researchers have focused on ways to shorten this period, and/or to make some use
of the land while it is fallow. Thus, an “accelerated fallow” is where specific fast-
growing leguminous or non-leguminous trees, shrubs, legumes, and other plants
are used to improve soil fertility faster than would occur otherwise, while an
“enriched fallow” is where trees or shrubs of economic value are planted into the
fallow so that the farmer can derive some income from them while the land is
regenerating (Garrity and Lai, 2000).
Szott et al. (1999) have reviewed the processes involved in fallowing from an
ecological perspective. Under a natural fallow, the time needed to restore N to
original levels can be less than two years, but can be up to 20 years for other
nutrients such as Ca. The rapid restoration of N is likely due to N fixation (up to
300 kg N ha – 1 year21 on high base status soils, (Giller, 2001)) and retention in
the growing biomass. Most evaluations of the performance of improved fallows
have compared crop yields after the fallow period with those after a natural
fallow of the same duration. In general, it has been found that short-term fallows
(i.e., , 3 years) growing leguminous trees or shrubs can increase crop yields
compared to the natural fallow control (Szott et al., 1999). Nitrogen fixation is the
main reason for this, but the retrieval of subsoil N by deep-rooted species from
below the crop root depth can also be important, and may be of the same
magnitude as N-fixation, particularly if nitrate has accumulated in the subsoil
during the cropping phase before the fallow (e.g., Mekonnen et al., 1997).
Improvement in soil structure may also be important — Salako et al. (2001),
for example, found that soil physical properties improved under fallows of
Pueraria phaseoloides, Acacia auriculiformis, Leucaena leucocephela, Senna

siamea, and Acacia leptocarpa, as well as under a natural fallow. Grass fallows
do not usually increase crop yields to the same degree, if at all, probably due to
immobilisation of N (Barrios et al., 1997) although in some cases, they do
improve soil structure (Padwick, 1983). It is noteworthy that crop yields were
significantly lower following fallows of Ageratum conizoides, Tithonia
diversifolia, Sida rhombifolia, or Rottboellia cochichinensis than following a
natural fallow, despite having more N and P in the fallow biomass (Szott et al.,
1999). On soils that are severely deficient in P, short-duration improved fallows
have been shown to need additions of fertiliser P to maintain crop yields (Jama
et al., 1998). Benefits of short-duration fallows to crop yields are sometimes
related to the amount of biomass accumulated during the fallow. Maroko et al.
(1998), for example, reported no increase in yields of the second maize crop after
a Sesbania fallow, which had accumulated 14 t ha21, but a significant increase in
crop yield with a 23 t ha21 Sesbania fallow. In most cases, there is little residual
effect of the fallow biomass after two crops (Drechsel et al., 1996). Interestingly,
the use of short-term cover crops such as Mucuna in fallows may be detrimental
to long-term accumulation of nutrients and biomass by impeding the establish-
ment of trees and shrubs (e.g., Szott and Palm, 1996). Kang (1997) has
recommended that short fallows be used where the duration of the cropping
period is at least half of the total cropping/fallow cycle.
There may also be other benefits derived from improved fallows. For example,
Brodd and Osanius (2002) report reduced soil erosion under an improved fallow
of Tephrosia candida in northern Vietnam compared with shifting cultivation.
Similarly, Gallagher et al. (1999), in a review of literature, have shown that
improved fallows of woody perennials and herbaceous cover crops could
suppress weeds, particularly over a number of years, and might be an important
component of integrated weed management (IWM) strategies.
Perhaps the main advantage of accelerated and enriched fallow systems is that
that they are modifications of an existing system, requiring only minor changes to
existing farmer practice. Accelerated fallows can be seen as a natural progression
from shifting cultivation and other long-fallow techniques, and may therefore be
an appropriate choice where these are no longer sufficient to maintain
productivity due to population pressure (Sanchez, 1999). From the biophysical
point of view, due to the deeper-rooting characteristics of the woody species
usually used in accelerated fallows, nutrients from below the rooting depth of
arable crops can be made available again, and the use of appropriate leguminous
species can result in improved rates of addition of N to the system. Also, there is
no direct competition for resources with main crops as with some of the other
LEIA techniques such as intercropping or alley cropping (Sanchez, 1999). The
downside, compared with continuous cropping systems at least, is that production
is lost from the land set aside for fallow. Hoang Fagerström et al. (2001) found
that the overall rice production from both a natural fallow and a Tephrosia fallow
with 2-year crop/2-year fallow cycle were 18 and 6% less, respectively, than
production from continuous cropping over the four-year period, despite annual
yields in the latter being considerably lower than in the fallow treatments. Similar
results were obtained by Swinkels et al. (1997) in western Kenya.
Accelerated fallows are more likely to be used by farmers in areas where
increases in the population density is starting to make the long periods required
by natural fallow impracticable. At higher population densities, however, scarcity
of land means that there is a high opportunity cost in putting land to fallow, and
intensive continuous cultivation systems may dominate (Drechsel et al., 1996).
Accelerated fallows are, therefore, most relevant in the intermediate stage
between extensive and intensive land use (Franzel, 1999). However, if they are to
be adopted more widely, farmers need to be aware that there is a problem to be
addressed. This may be declining yields (Franzel, 1999) or fertility (Degrande
and Duguma, 2000), or controlling weeds (Tarawali et al., 1999). Security of land
tenure is also an important consideration, as farmers are unlikely to be willing to
invest time and effort in establishing accelerated fallows if they are not the ones
to receive the benefits (Seif El Din and Raintree, 1987; Long and Nair, 1999;
Tarawali et al., 1999). Institutional support, in the form of seed programmes and
training of extension agents and farmers, has been found to be important in
improving the adoption of accelerated fallow techniques (Franzel, 1999). Other
important requirements may also be to provide adequate and/or improved
germplasm (Place and Dewees, 1999).
A major disadvantage of both natural and accelerated fallow systems is the
length of time it takes for any financial benefits to accumulate (Grist et al., 1999),
particularly with the latter, since natural fallows may provide resources that
accelerated fallows do not. Kaya et al. (2000) concluded that improved fallows in
Mali were not attractive to farmers if their sole purpose was soil fertility
improvement. Enriched fallows address this problem to some extent, in that
species that are able to provide some economic benefit, such as fruit or nuts, are
planted in preference to species that only improve soil fertility (Cairns and
Garrity, 1999; Franzel, 1999; Sanchez, 1999). Other practical benefits to farmers
may include production of fodder (Kaya et al., 2000), honey (Franzel, 1999),
firewood, or bean poles (Drechsel et al., 1996), or light timber for construction
(Franzel, 1999). From an ecological viewpoint, Styger et al. (1999) have
suggested that the judicial identification, selection, and domestication of
preferred forest fruit trees could be used as a means of preserving biodiversity
in forest margin areas that are under pressure.
Where such multipurpose tree (MPT) techniques are successful, farmers may
be encouraged enough to develop them into permanent agroforestry systems
(Cairns and Garrity, 1999), and, indeed, in several places many have done this
already. Farmers in Benin, for example, plant oil-palm trees in a fallow of about
12– 15 years (Versteeg et al., 1998). This restores soil fertility, but also provides
subsistence and cash income, even upon clearance of the trees when “palm wine”
is produced. However, removal of harvestable products (and the nutrients they

contain) lengthens the time taken for regeneration of soil fertility, and in extreme
situations, may eventually mine the soil of nutrients (Franzel, 1999). These
deficits will have to be made up with other sources of nutrients.
Pradeepmani (1988) has discussed some of the issues affecting farmers’
decisions to plant MPT species. These include having adequate land, time,
labour, knowledge, and inputs, being able to protect trees properly, and success
with tree survival. Security of tenure and access to markets are also important
(Mahamoudou and Meritan, 1998; Hellin et al., 1999). A certain level of access
to capital also appears to encourage adoption. Where these factors are not in
place, farmers tend to increase the rate at which they discount future benefits,
making such techniques socio-economically unviable, and reducing both the
“action time horizon” and the “planning time horizon” (Vosti and Witcover,
1996). Efforts to encourage planting of MPTs species through training visits,
extension of effective methods for protecting trees (which is often expensive),
and government land tax incentives, were also noted as important factors
(Pradeepmani, 1988). Although the multipurpose nature of many trees may serve
as “pull” factors, strong “push” factors can also operate at the localised scale. For
example, shortages of agricultural labour, high cost of agricultural inputs, and
shortages of power and water have all been reported to encourage farmers to plant
MPTs (Dasthagir et al., 1996). In most cases, economic factors are more
important than ecological factors in influencing farmers’ decisions to plant MPTs
(Mahamoudou and Meritan, 1998). Certainly, in a study of the reasons for the
adoption of MPTs on homestead land by farmers in Bangladesh, direct economic
concerns were reported as being uppermost in the minds of farmers (Salam et al.,
2000). Other factors in order of importance were the provision of fruit, firewood,
and building materials for subsistence, emergency cash needs, maintenance of
ecological balance, and protection from strong winds. Further analysis showed
that tree planting increased with increases in the amount of land owned, the level
of non-agricultural income, the market costs of fuel-wood, the male membership
of the household, and the knowledge of extension activities.
Compared to natural fallow systems, more labour is required for both
accelerated and enriched fallows, primarily for planting of the fallow species and
weeding to ensure their establishment (Drechsel et al., 1996; Grist et al., 1999).
Hoang Fagerström et al. (2001) give values of 386 and 600 labour days ha21 over
a 4 year period for a natural fallow and a Tephrosia fallow, respectively. There is
a clear danger that this additional labour demand could clash with those for
planting and management of other crops (Franzel, 1999). On the other hand, the
labour demand of improved fallows is relatively flexible, certainly compared to
alternative systems such as alley cropping where the timeliness of pruning is very
important. Some capital input is needed for improved fallows, mainly for the
purchase of seeds or seedlings of the species to be planted. If farmers have
insufficient capital, they tend to use natural fallows for soil fertility improvement,
whereas better-off farmers will tend to purchase inorganic fertilisers (Franzel,

1999). Improved fallows may therefore, be most appropriate for farmers at an
intermediate level with some disposable income.
Adoption of improved fallow techniques by farmers has been significant in
some regions. Citing various sources, Sanchez (1999) claims that there is large-
scale adoption of improved short-term fallows (i.e., , 5 years in duration)
occurring in Central America, Brazil, Southeast Asia, East Africa, and southern
Africa, with perhaps hundreds of thousands of farmers using the technique. The
majority of species used are Sesbania, Leucaena, Mucuna, Centrosema,
Pueraria, Crotalaria, Cajanus, Indigofera, and Mimosa.
An example is the “Qezungual System,” which has been indigenously
developed in western Honduras (Hellin et al., 1999). The technique can be
described as a triple-level agroforestry system, combining crops such as maize,
sorghum and beans, numerous pollarded trees and shrubs (about 1.5 m high)
and high-value trees, particularly fruit trees and timber trees. The system
has generally developed on land that has been under secondary vegetation, or
less commonly, on land that is under primary forest. A main characteristic of the
technique is the reduction of labour requirements for the establishment of the
valuable fruit and timber trees species, as these are simply selected when the land
is cleared for agriculture. Other less valuable trees and shrubs are pollarded and
the land is prepared for cultivation. This also reduces the time required for
benefits to accrue to the farmer, and may reduce the need for inputs requiring
capital (for example, seedlings) and labour (maintenance of vulnerable
seedlings). Competition between perennial plants and food crops is greatly
reduced by the pollarding, and can be manipulated by gradual clearing of
perennial plants, if necessary. Natural regeneration is managed by selecting
specific trees for production, whilst others are pollarded. Within the pollarded
areas, crops may be rotated and areas left fallow to control pests. From
discussions with farmers, there appear to be several benefits from the use of this
technique. Pollarded plots have higher agricultural production and can also be
cultivated for a longer time than unpollarded plots. Soil moisture is conserved
because of reduced soil evaporation due to the mulch from the pollards, and
perhaps because the pollards improve soil physical structure and allow increases
in WHC. The system provides multiple benefits for subsistence and cash income
(fruit, food crops, timber and firewood). Disadvantages are that moisture levels
can become too high in times of high rainfall, leading to fungal attacks on crops.
Birds may also be attracted by the trees and pollards, causing reduction of food
crop yields. Animal or mechanised traction cannot be used to cultivate the land
due to the random presence of the trees and pollards.
Various factors have encouraged farmers to adopt the Qezungual system.
Land scarcity is important (most farms are about 2.5 ha) as, where land is
abundant, farmers generally continue to use natural fallows. Absence of fire as a
management tool is also important, otherwise the trees are destroyed. Lack of
animal or mechanised traction is another factor, as the pattern of trees and

pollards is fairly random, although with selective thinning, it might also be
possible to develop pathways for animal and mechanical traction tillage. Possibly
the most important factor is that it addresses a problem that the farmers find
important — soil moisture. In dissemination of the use of the system, it
is promoted as a soil moisture technology, rather than an erosion control
technology, although erosion is definitely a problem in some areas.
III. GENERIC ISSUES
A. SOIL FERTILITY MANAGEMENT
1. Biomass Quantity
Relatively little quantitative information exists on the ideal level of SOM.

Brady (1990) suggests that it should be around 5%, corresponding to a soil organic
carbon fraction (SOC) of about 3%. In sandy southern African soils, 1– 1.5% SOC
has been recommended as the long-term agroecologically viable minimum
(Araki, 1993). Similarly, research from western African countries suggests that
when SOC levels fall below 1%, severe physical soil degradation can be expected
to take place (Pieri, 1995). The threshold level of SOC required to prevent severe
physical degradation of a soil is also related to the soil’s texture — for soils with a
low-sand content, an SOC content of 0.9% may be adequate, but for sandy soils,
this may have to be as high as 1.5% (Araki, 1993). Thus, although there appears to
be some variation as to an “ideal” level, somewhere between 1 and 3% SOC is
what many researchers would consider necessary.
Young (1989) has estimated indicative quantities of plant biomass required
for maintenance of SOC in soils in various agroecological zones, based on typical
topsoil organic carbon levels and approximate oxidation and erosion losses.
Above-ground plant biomass required inputs were estimated to be about
8.4 t DM ha21 year21 for humid regions, 4.2 t DM ha21 year21 for subhumid
regions, and about 2.1 t DM ha21 year21 for semi-arid areas, with below-ground
biomass inputs about 70% of these figures. These are broad estimates and other
evidence suggests different quantities are required. For example, in southern
Africa, Snapp (1998) estimated that annual applications of about 10 t DM ha21
year21 of high-quality plant biomass (see below for discussion on biomass
quality), or 7 t DM ha21 year21 of low-quality residue, were necessary to main-
tain a minimum level of 1% SOC in sandy loam soils in the subhumid tropics,
assuming a decomposition rate of 0.05 year21 (Janssen, 1993). Thus, assuming a
linear relationship between SOC levels and biomass inputs, the “ideal” level of
3% SOC mentioned previously may require as much as 30 t DM ha21 year21 in
similar conditions. Even if the relationship is not linear, the data suggest that on the
whole, large amounts of biomass are required to maintain the physical condition
of the soil at a level that can support continuous and sustained crop production.
The supply of N and P to meet crop nutrient requirements also requires large
quantities of biomass, not only because of the relatively low concentration of
nutrients in biomass, but also because the fraction actually recovered by crops is
generally low for organic inputs due to losses to the atmosphere, surface waters,
or ground water (Gregory et al., 2002), although this can vary substantially
depending on biophysical conditions. Giller and Cadisch (1995) have suggested a
value of 20% for most organic inputs. Gachengo et al. (1999) found that the
recovery of N in Tithonia prunings was about 25% by a first maize crop in
western Kenya. Other evidence suggests that N recovery by the first crop after
OM incorporation is generally between 9 and 28% of the N supplied in the OM,
but may be as low as 2 –10% in a second crop (Snapp et al., 1998). However, the
rate of N recovery can be improved by the addition of limiting nutrients to ensure
that the growth of the main crop is not limited. For example, Snapp et al. (1998)
found that the recovery fraction of organic N increased from 25 to 46% in the first
year when 25 kg P ha21 was applied to correct the P deficiency at the site.
To supply adequate N for a typical 5 t ha21 maize crop removing about
100 kg N ha21, therefore, more than 14 t DM ha21 of biomass would be required
(assuming a 20% recovery rate and 3.5% leaf biomass N content (Jama et al.,
2000)). For animal manure with 1.5% N content (Lekasi et al., 1998), 33 t DM ha21
would be required. These figures would equate to 60 and 67 t ha21 of fresh
material, respectively, assuming a 20% dry matter content for fresh leaf biomass
and a 40% dry matter content for fresh manure. For P, Jama et al. (2000)
calculated that 5 t DM ha21 would be required to supply the 18 kg P ha21 needed
to overcome moderate P deficiencies. With a 20% recovery rate, this is equivalent
to an application of 25 t DM ha21 or 125 t ha21 fresh weight. In severely
P-deficient soils, even more would be required. The quantities required for the
effective management of other soil nutrients is also large. For example, Jackson
et al. (1999) found that to correct zinc deficiencies of soils near Wenchi in Ghana,
about 20 t ha21 (dry or fresh weight not specified) of poultry manure was
required. The amount of sheep or cattle manure required was estimated to be
between 40 and 60 t ha21.
If these levels of biomass are required to have an appreciable effect on SOC,
the question arises as to how easily these quantities can be produced by
smallholders. In the studies where alley cropping has been shown to benefit crop
yields, tree biomass production has been in the order of 6 –8 t ha21 year21, using
L. leucocephala (Kang et al., 1985). For Flemingia congesta, Budelman (1988)
recorded an annual dry matter production of 12.4 t ha21 year21 in the Ivory
Coast, while Yamoah et al. (1986a) measured 16.9 t DM ha21 year21 for
Flemingia congesta in Nigeria. However, in many cases, biomass production is
not this high. In an alley cropping system in Costa Rica (humid), the net primary
production of Calliandra calothyrus was about 4.4 t DM ha21 year21. Of this,

about 2.8 t DM ha21 year21 was estimated to be leaf production (Baggio and
Heuveldorp, 1988) and of possible use as a green manure. In Nepal, C. cajan was
able to produce about 3– 4 t DM ha21 year21, Centrosema pubescens about 5 t
DM ha21 year21, and Calopogonium mucunoides about 4 t DM ha21 year21
(Pande, 1997). In Malawi, Saka et al. (1995) showed that the leaf biomass
production of three hedgerow species (Gliricidia sepium, Leuceana leucoce-
phala, and Senna spectabilis) varied between 0.5 and 2 t DM ha21 year21, and
did not affect the SOC level over a 1 year period. Also in Malawi, Kanyama-Phiri
et al. (1997) found that Sesbania sesban produced about 2– 3 t DM ha21 year21
of high-quality leaf biomass. This was in addition to the fuel-wood produced
during the 10 months of growth between January and October. In Zambia,
although Flemingia congesta produced a maximum of 3 t DM ha21 year21 in one
trial (Table I), the mean production of all the species was only 1.3 t DM ha21
year21 (Matthews et al., 1992a).
In many cases, therefore, biomass production does not seem high enough to
have any appreciable effect on SOC levels. Cover crops are likely to produce
even less biomass annually due to their shorter duration of growth compared to
woody perennials. Increasing SOC to “ideal” levels, therefore, will in most cases,
necessitate the importation of additional amounts of organic material.
The question is, therefore, where is this biomass to come from? If the farmer is
to grow it, can it be produced in sufficient quantities to have an appreciable effect
Table I
Mean Annual Biomass Production (t ha21 y21) of Different Tree Species in Agroforestry Trials
at Kasama, Northern Province, Zambia. (Developed from Matthews et al., 1992a,b)
Trial no. Species 1987 1988 1989 1990
D11 Flemingia congesta 1.40 2.45 2.91

Tephrosia vogelii 2.19 0.33
Cassia spectabilis 0.56
Calliandra calothyrsus 0.64 1.51
D31 Leucaena leucocephala 2.46 2.23 2.47 1.48
Albizzia falcataria 0.67
Flemingia congesta 0.44 1.09 0.73
Gliricidia sepium 0.94 0.79 1.08 0.47
D32 Flemingia congesta 0.83 0.74 0.51 0.60
Cassia spectabilis 0.54 1.31 0.84 1.12
Sesbania sesban 0.72 0.44 0.07
D33 Flemingia congesta 1.05 1.13 1.36 0.82
Cassia spectabilis 0.56 2.14 1.40 0.97
on SOC levels? Assuming an annual biomass production of 2 t DM ha21 year21,

3 ha of land would be required to supply 1 ha of cropped land area with enough
biomass to maintain the SOC level at just 1%. Significantly more would be
required to raise it to the 3% level indicated earlier. In the initial years of
hedgerow intercrop systems, when farmers are most likely to reject or accept a
new technique, tree biomass production is likely to be well below 2 t DM ha21
(e.g., Matthews et al., 1992a). This is clearly insufficient to maintain SOC at even
the 1% level. From the farmer’s perspective, if this biomass is to be grown
on-farm, the area available for food crop production must necessarily be less. If it
is to be supplied from outside the farm, the transport of such large quantities
of organic material would require considerable labour and physical effort (Rao
and Mathuva, 2000; Hoang Fagerström et al., 2001). Furthermore, labour will be
required for incorporation of the biomass into the soil. In many situations,
therefore, it may simply not be possible to use plant biomass to increase SOC
within farmer constraints.
Supplying adequate amounts of organic matter through animal manure is also
difficult. The dry matter quantities required for soil physical improvement are
similar to the amounts required when using plant biomass (Euroconsult, 1989).
Pilbeam et al. (1999b), in deriving an N balance for a hypothetical household
with 1 ha of agricultural land in the mid-hills of Nepal, estimated that the total
feed requirement for buffalo, assuming a live-weight of 450 kg, was about 2.6 t
DM year21. For cattle, assuming a live-weight of 250 kg, feed requirements were
estimated to be 1.8 t DM year21. Thus, to balance the SOM losses given by
Young (1989), about 3 buffaloes or 5 cows would need to be kept for every 1 ha
in humid regions and one buffalo or 1.5 cows per hectare in semi-arid regions,
assuming that nearly 100% of the consumed biomass passes through the animals.
For comparison, Nandwa and Bekunda (1998) calculated that between 2 and 8
cattle would be needed to supply enough nutrients for a 2—3 t ha21 maize crop.
For a typical farm in the mid-hills of Nepal, the analysis by Pilbeam et al. (1999b)
suggests that about 2.5 t of animal feed comes from dry and green crop residues,
presumably from on-farm sources. However, to maintain SOC levels of one
hectare of land through the use of animal manure alone, this still leaves a
requirement of about 6 t DM of fodder from off-farm sources.
These biomass requirements are rough guides, but serve to show that the
quantity of animal manure needed to maintain soil physical characteristics and
nutrient levels effectively are substantial. In most cases, it is unlikely that
resource-poor farmers would have access to on-farm sources of manure in
sufficient quantities to supply the total OM requirements of their cropped land,
particularly as there are competing demands for its use, such as for building
material or fuel (e.g., Nandwa and Bekunda, 1998). In such cases, reliable access
to off-farm land for fodder collection will be a major requirement for the use of
animal manure. In the Embu District of the Kenyan highlands, application
rates of 5– 8 t FM ha21 (~2 –3 t DM ha21) are recommended to farmers
(Lekasi et al., 1998), but average rates applied by farmers are often much
higher at 11 t FM ha21 (~4.4 t DM ha21), and could even exceed 17 t FM ha21
(~7 t DM ha21). Despite these relatively high rates of application, many farmers
felt that they would use even more if it were available.
In general, the smaller the animal, the higher is the nutrient concentration
in its manure. For example, poultry manure has much higher levels of N and
P (up to 4.8 and 1.8%, respectively) than cattle manure (1.5% N and 0.14%
P) (Reddy et al., 2000). The amount of poultry manure needed to supply the
requisite amount of N and P to a crop would, therefore, be correspondingly
lower, as would the labour required for transporting it. For example, Swift
et al. (1994) estimated that 1 –2 tonnes of poultry manure would be required
to fertilise a 2 tonne maize crop, compared to 7 tonnes of low-quality animal
manure or 10 tonnes of straw. On the other hand, smaller animals produce
lower quantities of manure per animal, and finding poultry manure in
sufficient quantities could prove difficult, as it is unlikely that the numbers of
poultry typically found on resource-poor farms would supply more than a
few kilograms of manure annually.
Overall, therefore, the quantity of biomass available to small farmers, either
from plants or from animal manure, is likely to constrain the degree to which soil
fertility can be maintained or even improved. Production of this biomass on farm
must be weighed against the loss of any other use the land may be put to,
particularly the growing of crops. Off-farm sources may be an option, as in Nepal,
but this will not always be the case.
However, an important effect of the addition of OM to the soil, and one that
may be immediately appreciated by farmers, is an improvement in the workability
of the soil, so that farming operations, particularly ploughing or hand hoeing, are
eased. This is likely to be particularly important where continuous cultivation of
land is already practised and SOC has decreased, and bulk density has increased
as a result. For example, in on-farm trials in Ghana and Bolivia, farmers reported
the benefits in terms of the “softness,” “looseness” or “coolness” of soils after
using OM from cover crops and animal manure (Kiff et al., 1999; Pound et al.,
1999). The extent, to which these benefits were a result of the incorporation of the
OM in the soil, or the growth of cover crop, was not reported. However, what is
important in that soil workability was a key factor by which farmers appreciated
the effect of OM techniques in the short-term rather than in the long term.
2. Biomass Quality
The “quality” of biomass is a function of its nutrient content and its resistance
to breakdown. Biomass quality has two opposing effects, in that lower quality OM
is likely to have a larger impact on SOM levels than high-quality material, which
mineralises more rapidly. On the other hand, higher quality organic material
contributes more to the nutrient status of the soil (N, P, K, and micronutrients), and
is important for maintaining soil microbial activity and the soil buffering capacity.
Successful OM management depends on finding the correct balance between these
two effects. This applies to plant OM, whether green manure or crop residues, as
well as to animal manure. Crop residues and other low-quality organic material,
particularly if added in large quantities, may temporarily induce N or P
deficiencies in the soil due to microbial immobilisation, thereby reducing crop
yields. Palm et al. (1997b), for example, have shown that addition of OM
containing less than about 0.25% P to the soil is likely to cause net immobilisation
of P. Such deficiencies may have to be overcome through the use of inorganic
fertilisers (Muriwara and Kirchmann, 1993).
The ratio of carbon to nitrogen (i.e., the C/N ratio) in organic material is often
used as a measure of its quality. More recently, the concentrations of lignin
and polyphenols have also been found to be important, particularly the lignin/N,
polyphenol/N, and (lignin þ polyphenol)/N ratios (Snapp et al., 1998). High-
quality organic inputs are low in lignin and polyphenol and high in N (Palm et al.,
1996), and as such decompose more quickly. It has been estimated that they
release between 70 and 95% of their N within a season under tropical conditions
(Giller and Cadisch, 1995). The quality of plant biomass is not constant, but
varies with age and whether a plant is leguminous or non-leguminous. In general,
young plant material has low C/N ratios ensuring that its nutrients will be
released quickly when it is incorporated into the soil, while material from older
plants (or plant organs) of the same species generally has a higher C/N ratio.
Data on C, N, P, and K contents for a range of organic materials used in low-
external input agriculture are available in the Organic Resources Database
developed as part of the Tropical Soils Biology and Fertility Programme (TSBF)
(Palm et al., 2001).
The nutrient content of animal manure depends on the species (Table II), the
diet of the animal, and on how the manure is collected, stored, and applied. Diet is
particularly important in relation to the partitioning of N between the faeces and
the urine (Snapp et al., 1998). High-quality diets (low in lignin and polyphenols)
result in more N being excreted in the urine than in the faeces (Somda et al.,
1995). N that is excreted in the urine is much more quickly volatilised, and urine
is also more difficult to collect. Animals fed with a tannin-rich diet tend to excrete
more of the N in their faeces. However, recent results suggest that this kind of N
is very resistant to mineralisation (Mafongoya et al., 1997a). The beneficial
effects of N released from manure as NHþ 2
4 and NO3 appear to be directly after
application. However, poor-quality manure has been found to result in prolonged
periods of N immobilisation, and under these conditions, N availability has been
increased with inorganic sources (Muriwara and Kirchmann, 1993). Table II
shows that for manures from most domesticated animals, P content is above the
threshold at which P immobilisation is likely to occur, but that N content is often
below the critical limit required to prevent N immobilisation.
Table II
Mean Content of C, N, P and K in Manure from Various Domestic Animals in Muranga and
Kiambu Districts, Kenya
C N P K C/N ratio
Manure type (%) (%) (%) (%) (kg C (kg N)21)
Cattle 35 1.4 0.60 0.59 26

Cattle and compost 25 1.3 0.44 0.36 21
Goat 32 1.5 0.40 0.53 22
Pig 40 2.0 1.19 0.49 21
Poultry (broilers) 41 2.4 1.60 0.41 17
Poultry (local) 22 1.2 0.91 0.26 19
Rabbit 33 1.6 0.40 0.50 20
Sheep 33 1.5 0.33 0.44 22
(From Lekasi et al., 1998).
In the absence of inherently fertile soils and inorganic fertilisers, improved

crop yields are usually achieved only with high-quality OM. Low- or even
medium-quality residues have generally been unsatisfactory (Snapp et al.,
1998). In Kenya, for example, Nandwa (1995) found that incorporation of
maize stover (low-quality OM) reduced maize grain yield by between 3 and
30%. Maize stover has also been shown to reduce crop yields in Zimbabwe
(Rodell et al., 1980; Muriwara and Kirchmann, 1993). In India, Goyal et al.
(1992) found that a combination of wheat straw (low-quality OM) and urea
reduced yields, while a combination of Sesbania green manure (high-quality
OM) and urea increased yields compared with the application of urea alone. The
reduction in yields with low-quality OM is generally attributed to the
immobilisation of N by microbial growth. Added organic material with a
high C/N ratio (e.g., . 20 kg C (kg N)21) provides adequate C substrate, but as
the C/N ratio of microbial biomass is lower (3—14 kg C (kg N)21), N is taken
up from the mineral N pool in the soil to meet the shortfall, thereby reducing
that available for crop uptake. This can be a major problem if it coincides with
critical growth stages of the crop. Mafongoya et al. (1997b) have shown that N
immobilisation also occurs when the lignin and polyphenol content of the
residues incorporated into the soil were over 15 and 3%, respectively.
Interestingly, N immobilisation resulting from high polyphenol levels seems
to last much longer than that resulting from low C/N ratios (Palm et al., 1996).
As the C content of dry biomass is usually around 0.4 kg C (kg DM)21,
the critical C/N ratio of 20 kg C (kg N)21 corresponds to a biomass N content
of between 2.0 and 2.5% (Palm et al., 1997b). Table III shows the N contents
for a range of organic materials, from which it can be seen that some cover
crop species, the leguminous tree species, and Tithonia diversifolia all contain
N levels above 2.5%. The animal manures, on the other hand, particularly
Table III
Mean, Maximum and Minimum N Percentages for Various Organic Materials
Organic matter source Mean Max Min
Non-legume Tithonia diversifolia 3.38 4.59 1.10

Legumes (cover crops) Crotalaria juncea 3.47 6.30 0.80
Canavalia ensiformis 2.89 4.74 0.23
Mucuna pruriens 3.23 6.05 0.83
Legumes (trees) Leuceana leucocephala 3.68 6.32 1.04
Gliricidia sepium 3.38 5.33 1.33
Animal manure Cattle 1.04 4.15 0.30
Cattle (Lekasi et al., 1998) 1.40 2.00 0.50
Pigs 3.79 4.25 3.08
Pigs (Lekasi et al., 1998) 2.00 2.20 1.50
Poultry 4.02 6.73 1.85
Poultry (Lekasi et al., 1998) 2.40 2.60 2.30
Crop residues Maize 1.01 3.07 0.25
Sorghum 0.63 0.63 0.63
Compiled from the Organic Resources Database (Palm et al., 2001) and Lekasi et al. (1998).
from cattle, tend to have mean N contents that are likely to lead to net
immobilisation of N in the soil. Data from Lekasi et al. (1998) also suggest
the pig and poultry manures are marginal in terms of N levels, although the
data from the Organic Resources Database show more favourable levels of
mean N for these animals. The low N contents in most cereal crop residues
also suggest that there may be negative yield effects on the crops to which
they are added, due to net N immobilisation.
These issues are important when adding OM to the soil, whether it is to
improve the physical or chemical characteristics of the soil. Where long-term
improvement to soil physical characteristics is important, the requirement will be
for moderate- to low-quality OM to be applied. However, this may result in N and
P immobilisation, particularly without supplementary use of inorganic fertilisers,
and crop production may therefore decline. Where an improvement to the soil
nutrient status in the short-term is needed, high-quality OM with low C/N ratios
should be applied.
3. Nutrient Mining
In view of the large quantities of OM required and also because of the

resulting uptake of nutrients during the production of this biomass, there is
the question of whether such production is sustainable in the context of
subsistence agriculture in the tropics. For example, biomass transfer systems,
such as cut-and-carry grasses or fodder banks, may function effectively for a

certain period of time and relieve pressure on other sources of biomass, in
particular, forest biomass and on-farm sources of biomass. However, it is
unlikely that such systems can exist long without additional inputs, as nutrients
are “mined” from the soil in which the biomass is being produced. In Kenya, it
was found that the introduction of fodder banks was effective for a short while,
but it soon became apparent that their productivity was declining due to the
continuous removal of biomass and hence of nutrients from the soil (Wandera
et al., 1993).
Similar considerations apply to the use of Tithonia diversifolia as a supply of
P to crops. Tithonia is able to scavenge relatively large quantities of P from the
soil and to provide biomass with relatively high concentrations of P for
incorporation as OM. However, this does not add to the net amount of P in the
soil, but rather is a way of redistributing it spatially. Clearly, Tithonia will
eventually mine the soil of P and other nutrients, which is unsustainable at the
level of the whole farm. Unless resource-poor farmers have access to large areas
of off-farm land growing Tithonia, it is unlikely that the system can be sustained
over long periods of time. One solution might be to fertilise the on-farm biomass
banks or hedges, but in this case farmers may as well fertilise the crops directly.
The sustainability of the System for Rice Intensification (SRI) technology
developed in Madagascar in the 1980s by Fr. Henri de Laulaníe (Stoop et al.,
2002) and promoted as a successful example of low-external input agriculture
(e.g., Pretty and Koohafkan, 2002) can also be questioned from this point of view.
The approach focuses on early transplanting, planting of single plants rather than
hills, a square arrangement of plants rather than in rows, frequent weeding, and
non-flooding of the soil during the vegetative period, resulting in less seed and
water being required. Compost can be added as a source of nutrients, although
this is not an essential part of the package. Huge increases in rice yields up to
21 t ha21 without the use of purchased inputs of fertiliser and pesticide have been
claimed (Uphoff, 1999). Some of the individual components (e.g., adequate
spacing and weeding) are commonly recommended practices, but the remarkable
yield increases are claimed to arise from the “synergistic” effect of all of these
used in combination (Stoop et al., 2002). Part of the success of the approach
appears to be in maintaining the soil in an aerobic state rather than completely
flooded as in traditional irrigated rice cultivation, resulting in higher rates of
organic matter mineralisation (Stoop et al., 2002), so that more nutrients can be
supplied from the soil rather than from fertiliser applications. Certainly,
anaerobic conditions are known to depress mineralisation rates, and there are
also reports of increases in rice yields in well-drained soils compared to flooded
soils (e.g., Ramasamy et al., 1997). While the yields of 21 t ha21 are undoubtedly
an extreme (and indeed questionable physiologically), average yields of
8.8 t ha21 in farmers fields have been reported (Stoop et al., 2002). Such yields
will remove about 130 kg N ha21 along with other nutrients at each harvest.
Even if such low-fertility soils as reported (Uphoff, 1999) are able to supply such
quantities of N, it is difficult to see how such yields could be sustained for long
without further addition of nutrients. If compost is to be the main source of
nutrients, then quantities in the order of 33 t DM ha21 (at 2% N content and
20% recovery rate) would be required to replace the N removed by the rice crop.
The 21 t ha21 rice yields were reportedly achieved with applications of 40 t FW
ha21 (~16 t DM ha21) compost made from leaves of Tephrosia, Crotalaria, and
banana along with rice straw (Uphoff, 1999). It is not easy to see how such
quantities of compost could be produced and transported by a single household
on a sustainable basis, except for very small areas of crop. Estimates of annual
household waste production (potentially available for making compost) vary
widely from ~100 kg DM year21 in Uganda (Wortmann and Kaizzi, 1998) to as
high as 9.8 t DM year21 for households of 8 –14 people, also in Uganda (Briggs
and Twomlow, 2002), although at between 700 and 1225 kg DM person21
year21, this last value appears somewhat high. Pantanali (1996) gives an inter-
mediate value of 2 t DM year21 for the typical annual production of household
waste in Lesotho, which, assuming a family of eight persons, represents about
250 kg DM person21 year21. Estimates for low-income households in South
Africa are as low as 40 kg person21 year21 (Durban Metropolitan, 1999). For
comparison, average household waste produced per person in England is about
500 kg person21 year21. Whatever figure is used, it would seem that household
waste alone is unlikely to be able to meet the nutrient requirements of such high-
yielding crops, and that nutrients, either in inorganic or organic form, would have
to be collected from a wider area for the enhancement of the cropped area.
4. Biological Nitrogen Fixation
Biological nitrogen fixation (BNF) by legumes is a key process in LEIA

technologies as it potentially results in a net addition of N to the system.
However, the quantity of N fixed by legumes is difficult to quantify and varies
according to the species involved and location. Webster (1998) noted that
estimates of the amount of N fixed by groundnuts and grain legumes range from
about 25 to 200 kg N ha21 during growing seasons of 60 –120 days. Bouldin et al.
(1979) found that some legumes seemed to fix N at relatively high rates — for
example, values up to 535 kg N ha21 have been recorded for Crotalaria spp. and
400 kg N ha21 for other pure legume green manure crops. Moore (1962) reported
that Centrosema pubescens (star grass) fixed N at the rate of 280 kg ha21 year21.
Reviewing the contribution of BNF in alley cropping, Sanginga et al. (1995)
quote measured values of BNF of between 100 and 300 kg N ha21 year21
for some tree species such as L. leucocephala, Gliricidia sepium, and
Acacia mangium, but values as low as 20 kg N ha21 year21 for others such as
Faidherbia albida and Acacia senegal. Values compiled by Brady (1990)

indicate that rates of N fixation by a range of legumes vary between 5 and 300 kg
N ha21 year21, with an average of about 100 kg N ha21 year21.
The amount of N fixed biologically that is actually taken up by the main crop
is difficult to determine with accuracy. In their review, Sanginga et al. (1995)
estimate the contributions of N from trees in alley crops to an associated food
crop to range from 25 –102 kg N ha21 season21 from root turnover and nodule
decay, with a further 40 –70 kg N ha21 season21 being supplied through
prunings. This represents recoveries of N of about 30%, although values as low as
5 –10% were reported. Thus, between 65 and 172 kg N ha21 season21 could
potentially be supplied to the main crop over a season, representing enough N to
support maize crops yielding between 3.5 and 8.5 t ha21.
These figures, however, are based on BNF rates towards the high end of the
range. Taking the average value of 100 kg N ha21 year21, and assuming that the
recovery by the crop was the same 20% suggested by Giller and Cadisch (1995),
with the rest lost through immobilisation, adsorption, volatilisation, leaching, and
denitrification, BNF would provide around 20 kg N ha21, or roughly the N
requirements of a 1 t ha21 maize crop. Moreover, this figure assumes a pure stand
of the legume — if it is a component of mixed species system, then the addition
of N to the system would be correspondingly less (Snapp et al., 1998). Kumara
Rao et al. (1981) showed that both relay and sequential intercrop systems in
southern Malawi produced insufficient quantities of biologically fixed N to
maintain a maize biomass yield of 4 t ha21 on infertile land. They estimated the
N contribution of Sesbania sesban when relay intercropped in low-fertility areas
to be 28 kg N ha21, and that from a pigeonpea/groundnut intercrop to be about
20 kg N ha21. In comparison, the quantity of N required to sustain maize yields
was estimated to be about 100 kg N ha21. Similar calculations were made by
Matthews et al. (1992a) for alley crops in northern Zambia — a net amount of
14 kg N ha21 year21 was estimated to be added to the system through BNF, only
13% of the local recommended fertiliser N application rate for maize. Despite
this, maize crop yields of 2 – 2.5 t ha21 were achieved for 4 years, which, with an
annual removal of 30 – 35 kg N ha21 year21 in the harvest, suggests that some
nutrient mining may have been occurring.
The growth rate of the legume is a strong determinant of the rate of BNF. Any
factor that reduces this growth, therefore, will also reduce the introduction of N
into the system through BNF. In particular, competitive main crops such as maize
may reduce the growth, and hence BNF rates, of legumes growing in an intercrop
(Shumba et al., 1990; Muza, 1995; Kumwenda et al., 1997b). For example, Patra
and Poi (1998) noted that intercropping maize with various legumes caused the
number of nitrogen-fixing nodules on the legume crops to decrease due to
shading, while Fujita et al. (1993) found that artificial shading reduced BNF in
several pasture legume species. In agroforestry systems, competition for light can
be managed with timely pruning of the perennial species. It has been suggested
that competition for water and nutrients can also be managed with pruning,
although the effectiveness of this is still debated (Snapp et al., 1998). Nair (1993)
has pointed out that controlling the below-ground competition of roots is far more
difficult than controlling the above-ground competition of canopies for light.
Reduced light intensity as a result of interception from a dominant main crop has
also been found to reduce nodulation (Webster and Wilson, 1998).
Another factor influencing the net contribution of N to the system is the
amount that is removed if the legume crop is harvested for grain or forage. Where
the legume produces a useable product, farmers are likely to harvest and use the
product. In the case of edible grain legume crops grown either as intercrops or in
rotation, harvesting the grain can result in lower than expected residual N effects
on the following crop, particularly if the grain legume has a high harvest index. In
Brazil, residual N benefits from using soybean as a leguminous green manure
were greatly reduced by harvesting (Boddey et al., 1997). Farmers are likely to
leave grain for incorporation as a green manure only where they do not market or
consume it themselves. Boddey et al. (1997) have suggested that the solution may
be to develop or use varieties of soybean that produce extra biomass while
producing the same grain yield, thereby reducing the harvest index and allowing
more N to be incorporated into the soil with the extra plant biomass. Grain
legumes with a naturally low harvest index, such as pigeonpea or groundnut, or
plants with no commercial or subsistence value could also be used. However,
such solutions are only likely where improved N management is more important
than short-term subsistence or cash benefits.
Legumes may also have other important functions within the smallholder’s
strategy. For example, in the semi-arid areas of northern Namibia, McDonagh
and Hillyer (2000) found that for cowpea, bambara and groundnut intercrops to
be able to make any contribution of N to the system, there should be no grazing or
burning of legume residues; but as cattle were an integral part of the system, this
was unlikely to be a popular option with the farmers. Increasing the legume plant
density just to the point where it began to affect the growth of the pearl millet
contributed only about 4 kg N ha21 to the system, which increased millet yields
by 80 kg ha21, a somewhat insignificant amount. They concluded that grain
legumes alone are unlikely to be able to improve soil fertility in the area
substantially, and that external fertiliser inputs would be necessary, although the
uncertain rainfall makes investment in soil fertility unattractive for farmers there.
It would appear, therefore, that in many cases BNF approaches are unlikely to
provide more than a small fraction of the N requirements of main crops, unless
low yields are accepted, or unless the farmer has access to other sources of
leguminous biomass. Certainly, the use of these techniques to supply the full N
requirements of the crop will require more land under the legume than under the
main crop, which deprives the farmer of land that could otherwise be used for
subsistence or cash crops. The use of mixed species intercropping is likely to
decrease main crop yields through competition, particularly in difficult
biophysical conditions, where legume growth is limited, for example, by lack of

P, or where lack of rainfall exacerbates competition for water. Nevertheless,
Snapp et al. (1998) suggest that farmers may contemplate the use of such systems
as part of the solution to N management, providing that the area available from
growing the main crop is not reduced too much.
B. SOCIO-ECONOMIC ISSUES
As the technologies we have discussed are ultimately used by people, social

and economic factors are likely to be as important, if not more so, than the
biophysical characteristics of the technologies. Amede (2001), for example,
noted that in the Ethiopian highlands, adoption of various techniques such as
cover crops and crop residue incorporation related as much to farm size and
availability of labour as to the conditions of the soil. In this section, we discuss
various socio-economic factors that can influence the successful uptake of
LEIA technologies.
1. Land
The amount of land required to produce enough biomass to maintain or

improve the SOC level of cropped land or to supply sufficient nutrients to meet
crop requirements has already been discussed. The calculations described
previously indicate that in general, a minimum of 3 ha of land is required to
produce enough plant biomass to maintain SOC at 1% or to meet the nutrient
requirements of 1 ha of cropped land. Essentially, in such a system, nutrients are
being harvested from a wider area to enhance productivity within a smaller area,
as in some slash-and-burn cultivation systems (e.g., chitemene and fundikila in
Zambia, (Matthews et al., 1992a)). In many cases, depending on the type of
biomass, climate and soil conditions, this ratio may be considerably more than
3:1. In the case of animal manures, where animals are used to gather nutrients
from a wider area, the ratio is likely to be higher because of the losses of nutrients
during manure storage and N losses in the urine not incurred in direct
incorporation of biomass in the soil. Palm et al. (1997a) estimated that between
14 and 42 ha of miombo woodland would need to be grazed to provide enough N
in manure for a 2 t ha21 maize crop, while ratios of up to 45:1 have been
estimated for other extensive systems (Turner, 1995). In many situations, it is
very unlikely that resource-poor farmers will have access to sufficient land, over
and above that needed for crop production, to produce the appropriate quantities
of biomass. Thus, if organic materials are to be used solely for soil fertility
maintenance or improvement, in most cases, it would seem that they must be
obtained from off-farm sources such as forests or, possibly, as waste from
urban areas.
Even with intercropping or alley cropping, where a leguminous species is
grown on the same piece of land as the main crop, there are costs; firstly, in that
the area taken up by the legume cannot be used by the main crop, and secondly,
due to potential competition for resources such as light, water, and nutrients. Few
studies have demonstrated the possibility of achieving increases in main crop
yield per unit of total system area while maintaining soil fertility. Relay cover
cropping with annual legumes, or perennial legumes like C. cajan for a single
season, may address the issue of direct competition, but is often not feasible due
to lack of rainfall outside the main cropping season. Thus, legume growth and
hence BNF rates do not compensate for the N that is removed in the harvest of the
main crop. On the other hand, if rainfall does continue, farmers may be more
likely to grow a second food crop rather than an unproductive legume.
Improved fallows may be an option for resource-poor farmers where fallows
are already an accepted part of the cropping system. However, where land is
limited and continuous cropping exists, improved fallows would have to replace a
main crop, which is not feasible for most resource-poor farmers. Furthermore,
resource poor-farmers are unlikely to be able to meet the financial and labour
costs required for the establishment of improved fallows. A natural fallow,
particularly one of several years, may provide resources that an improved fallow
cannot. Plants, animals, or residual germination and growth of cultivated crops
provide important products for farmers and soil regeneration under a natural
fallow may not be very different compared to that under an improved fallow,
especially if the natural fallow contains leguminous plants.
Suggestions have been made that field boundaries could be used to source
nutrients, particularly P, from biomass transfer species such as Tithonia
diversifolia (e.g., Briggs and Twomlow, 2002). While this may make some
contribution, a quick calculation suffices to show how much P, for example,
might be supplied in this way. If we assume that a landholding is 1 ha in area
consisting of 3 –4 fields, (typical for many small-holdings in Nepal and Ghana),
there would be about 600 m of field boundaries. Tithonia growing on these
boundaries producing 1 kg DW m21 from biannual pruning (Jama et al., 2000),
would therefore supply only 0.6 t DM ha21, or about 2.2 kg P ha21, equivalent to
about 15% of the 15 kg P ha21 removed in the harvest of a typical 5 t ha21 maize
crop. Alternatively, this amount of P would only support a 0.7 t ha21 maize crop
in P-deficient soils, and probably far less if due consideration is given to the usual
crop recovery rates of nutrients from organic material. It must also be borne in
mind that the Tithonia hedge has probably extracted some of this P from the
adjoining cropped fields in the first place.
In some situations, there may be ways of making better use of land that is
unproductive at certain times of the year. For example, in the Barind Tract of
Bangladesh, land is often left fallow during the dry season following the main
rice crop, as it is often difficult to establish a second crop during this time due to
drying of the soil surface layers (Musa et al., 2001). This is despite the fact that
there is usually sufficient residual water further down in the soil profile left from
the irrigated rice crop. In recent years, drought-tolerant crops such as chickpea
have been introduced, but establishment is not certain, and complete crop failure
may result. Recent research has shown that “priming” chick-pea seed, by soaking
it overnight in water, can result in a marked improvement in crop establishment,
making the difference between a healthy crop and no crop at all (Musa et al.,
2001). Attractions of the technique are that it is simple for farmers to implement,
requires no expensive inputs, and is aimed at resource-poor farmers rather than at
those with mechanised systems (Harris et al., 1999). Farmers welcome the ability
to gain an extra crop in the sequence, particularly of chickpea, which currently
commands good prices in the market, at little extra cost in land, labour, or capital.
Of course, there are questions of whether such systems of increased
intensification are sustainable, or whether soil fertility decline and possibly
weed encroachment is enhanced, although the experience in Bangladesh would
suggest that this is not the case. The area in question was converted from forest
about 150 years ago, and although current SOM levels are very low (0.5 –0.8%),
reasonable main crop yields seem to be obtained year after year with appropriate
inputs of inorganic fertilisers. Whether a further crop in the sequence will reduce
SOM levels even lower remains to be seen.
Land tenure is also an important issue influencing the use of LEIA
technologies. It is generally thought that land users who do not own their land
have less incentive to invest in technologies that take some time for soil fertility
benefits to accrue. Share-cropping is one such example, where a farmer exchanges
a proportion of farm output in exchange for the right to crop an area of land (Ellis,
1988). The system has generally been considered as being less economically
efficient than if the land is owned, as share-croppers are thought to input labour
only at a level that maximises their own perceived share of farm production, which
is less than what they would be prepared to give if they receive the total production
from the farm (Todaro, 1999). Nelson et al. (1998) analysed the economics of
upland agriculture in the Philippines, concluding that share-cropping would
reduce the economic attractiveness of alley cropping techniques compared with
alternative techniques (Fig. 2). This was because it was assumed that, under the
particular share-cropping arrangement, landlords would not contribute to the
establishment costs of the hedgerows, while a portion of the main crop would be
given to them as part of the tenancy agreement. However, other analysis suggests
that share-cropped farms are not necessarily inherently less efficient than owner-
managed farms (Reid, 1976). Also, share-cropping may at least give very poor
farmers the opportunity to farm, and some evidence suggests that it is the poorest
and most unskilled who stand to benefit from share cropping (Reid, 1976),
especially where the landlord also wishes to maximise the productivity of the
farm. Similarly, Ayuk (2001) in reviewing a number of studies, concluded that
LOW EXTERNAL INPUT TECHNOLOGIES
Figure 2 (a) The impact of sharecropping on the net present value of open field, fallow and hedgerow intercropping in upland Philippine agriculture at a
discount rate of 25%. (b) The net present value without the impact of sharecropping is also shown for all three systems at a discount rate of 25%. (Source: Nelson
et al., 1998).
519
the relationship between land ownership and willingness to invest in technologies

to improve SOM was not a simple one. While in some cases, there was a strong
correlation between long-term use rights and soil fertility improvement in the
form of manuring and mulching (e.g., Blarel, 1994), in other cases, there was no
correlation at all (e.g., Golan, 1994). Interestingly, a number of studies indicate
that traditional tenure systems do not inhibit land users from investing in
improving their land, and that a greater danger may be the granting of titles to
urban dwellers who have no interest in agriculture (Ayuk, 2001).
Fragmentation of land is another factor that may influence the adoption of
certain LEIA technologies. Fragmented landholdings may result in a single
farmer having to transport inputs to several isolated plots of land in several
different locations. This difficulty is particularly great with the use of biomass
transfer techniques, where several tonnes of biomass per hectare may be required.
In the mountainous terrain found in Nepal or Bolivia, for example, transporting
heavy loads to small isolated plots of land is extremely arduous. This is an
extremely labour-intensive process if the farmer has to do it alone, or expensive if
he/she hires labourers. Moreover, land fragmentation may result in decreasing
field sizes, which makes the implementation of certain techniques impractical.
Trees used for green manure or fodder, for example, may shade out the crop if
planted on the borders of small fields.
2. Labour
The labour required to make use of LEIA technologies may also constrain
their adoption. Often household labour may need to be supplemented with that
purchased from off the farm. While such contributions are seldom accounted for
in analyses of technology costs and benefits, especially when female labour is
involved, it is important to recognise that the use of domestic labour represents a
real opportunity cost. The need for external labour, which will generally involve a
cash transaction and therefore directly affect household finances, is more readily
acknowledged. Both aspects are important.
Ali (1999), in a study of farmers in Asia, found the cost of labour to be partly
responsible for making nutrient supply through organic matter less cost-effective
than through mineral fertilisers, a situation which is likely to get worse due to
rapidly rising wages. In India, the requirement for a pair of bullocks and a
ploughman was 10.5 days ha21 in a rice/green manure/rice rotation, while in
Nepal, the number of days needed in a wheat/green manure/wheat system was
11 days ha21. In both cases, the cost of this was about US$40 ha21 which largely
accounted for the differences in economic performance between green manure
and mineral fertilisers. This occurred despite the reduced need for weeding
and the increased yields obtained. Ali (1999) also analysed the economic
performance of using grain legumes rather than green manures, and found that
these were more profitable, despite increased labour requirements, because the
legume produced grain that could be sold. However, acceptance was not as high
as expected because the incorporation of the legume biomass delayed the
planting of the monsoon crop and increased the burden of labour at a time when it
was already high. The grain legume was also susceptible to pests and diseases.
Similarly, the cutting and carrying of biomass, such as from Tithonia, is
extremely labour intensive, particularly if it is to supply the full crop P
requirements in a P-deficient soil (Buresh and Niang, 1997). For a typical crop
requirement of 15 kg P ha21, the application of about 20 t DM ha21 of Tithonia
biomass would be needed (assuming a 20% recovery rate of the P by the crop),
equivalent to about 100 t ha21 of fresh biomass. Harvesting at the rate of 100 kg
FW man-day21 (ICRAF, 1997), 1000 man-days of labour would be needed just to
harvest the biomass! These are not the only costs. Tithonia also needs to be
propagated and prepared for incorporation in to the soil (ICRAF, 1997), and,
although it does not have thorns, it is difficult to handle because it is sticky and
exudes a pungent smell (Jiri and Waddington, 1998). In addition, because of its
ability to regenerate, it may invade farmland (Jama et al., 2000), thereby
increasing the labour required by a farmer to control it. The implication is that
either labour must be plentiful and cheap, or that the crops fertilised with Tithonia
should be high-value crops. As an example, Jama et al. (2000) cite data from
ICRAF showing that under farmer-managed conditions in western Kenya,
investing in Tithonia fertilisation was viable for high-value kale (Brassica
olecacea), but uneconomical when used with a low-value crop such as maize.
Supplying sufficient P through animal manure also requires substantial labour.
For example, the supply of 15 kg P ha21 through cattle manure (using a mean
concentration of P of 0.138% and a 20% recovery rate of P by the crop) would
require 55 t DM ha21 of manure, equivalent to 275 t FW ha21. Assuming that the
farmer had to transport the manure manually, that 20 kg of manure per load could
be carried at 5 km hr21, and that this load had to be transported 100 m from the
source, transporting this 15 kg P would require about 550 man-hours of labour.
In comparison, transporting the same amount of P in poultry manure, the same
distance would take only about 43 man-hours. Loads may often have to be carried
much further than this, and where large amounts of manure have to be transported
long distances, farmers may struggle to provide the labour required. Where the
manure has to be transported in mountainous terrain, such as in Nepal, the
amount of time required to transport manure will be even greater.
For many farmers, weeding is one of the most labour-demanding activities
undertaken. Gill (1982) noted that hand-hoe weeding in India required between
200 and 400 man-hours ha21 and that two weedings were needed during the
growth and development of many field crops. Van Tienhoven et al. (1982)
found that between 13 and 37 man-days ha21 of labour were required to
weed a maize – bean production system in the Jinotega region of Nicaragua.
This accounted for between 21 and 35% of family labour. Ruthenberg (1980) has
compiled data of labour requirements for weeding from various sources — in
Ghana, for example, weeding in a maize system required 31% of the total labour
(about 186 man-hours ha21), while in Columbia, weeding a cassava crop required
about 55% of total farm labour (about 408 man-hours ha21). Other traditional
agricultural systems cited were less intensive, although they all required at least
20% of total labour requirements for weeding.
In the Ichilo-Sara area of Bolivia, Pound et al. (1999) found that weeding could
require from 35 man-days ha21 of labour in the first year of a cropping cycle to 53
man-days ha21 in the third year as weeds started to dominate the system. The
increase in weed cover was associated with large decreases in rice yield, with
yields in the third year only about 30% of those in the first year. This was probably
due to the combined effect of weed growth and declining soil fertility. The labour
requirement for weeding was reduced when Calopogonium was sown as an
intercrop 25 days after the rice planting, but not if the Calopogonium was sown 45
days after the rice planting or as a cover crop after the rice was harvested. Pound
et al. (1999) make the point that these reductions in labour were probably not great
enough to be of practical significance to subsistence farmers, and clearly, would
not make a substantial difference to their livelihoods.
The system for rice intensification (SRI) discussed above (Stoop et al., 2002),
provides an interesting example of how labour requirements can limit the uptake
of an improved practice. Although labour requirements are high (38 – 54% more
than traditional methods), returns to labour are also high ($3.87 day21 compared
to $2.61 day21), a characteristic which has been seen as a major advantage of the
technology compared to traditional approaches. Despite these apparent
advantages, farmer adoption of SRI in the areas where it was promoted has
been low, abandonment of the method by those farmers who originally adopted it
has been high, and those who continue to practice the method rarely do so on
more than half of their land (Moser and Barrett, 2002). Participatory surveys
showed that this was because the recommended technological package required
significant additional labour inputs due to the extra weeding and water
management involved (the latter on a daily basis), during the time of year
when poorer farmers need to seek employment with other farmers to earn cash
to meet immediate consumption needs. Those who did try SRI were less likely to
rely on agricultural day labour as a source of income, and were also more likely to
have larger farms due to the economies of scale in offsetting the initial costs
involved in levelling their fields and in redesigning their irrigation systems to
allow more precise water control. Interestingly, adopters were also less likely to
have a relatively high salaried income as the opportunity cost of foregoing their
salary to supervise hired labour for the SRI was greater than the returns gained.
Thus, poorer farmers for whom the technology was designed to help were
less able to take advantage of it. This example highlights the need to evaluate
new technologies in the context of the whole agricultural system — even
considering returns to labour in this case was not sufficient to predict the
uptake of SRI, as the timing of labour and income requirements throughout
the year was the most important factor. Assertions by the proponents of SRI
that “farmer perceptions and practices are not necessarily wise or optimal”
(Uphoff, 1999) may, therefore, be somewhat premature when the wider
picture is taken into account. Moser and Barrett (2002) make the point that
opportunity cost is often overlooked in evaluating LEIA technologies, i.e.,
even though the financial cost of inputs may be low, this does not mean that
the technology is without other costs. They offer an interesting comparison
with the technique of off-season cropping (OSC), which was introduced in
the same areas in Madagascar as SRI, in which crops such as potatoes are
grown in the winter season after the rice harvest. Adoption of the OSC
technique has been high at 84% of households (across a range of wealth
classes), and with no disadopters to date (2002). The key to its success
appears to be the ease with which it fits into the existing agricultural system.
Despite relatively high input costs for the purchase of seed and fertiliser,
these occur at a time when farmers have just completed the rice harvest, and
have more time and money. Moreover, the OSC harvest provides them with
food and working capital at the beginning of the rice season, freeing the
household from needing to do off-farm work to earn wages to purchase food.
Farmers also perceive a carry-over benefit of the fertiliser applications on soil
fertility in the following rice crop. A large proportion of the farmers adopting
the technique had learned of it from other farmers, suggesting that it was
easy to learn.
The importance of the labour profiles of new technologies fitting in with
existing labour patterns has also been noted by Hoang Fagerström et al.
(2001) — biomass banks of Tephrosia were not popular with farmers in Vietnam
as the labour required for cutting and transporting the biomass coincided with
busy periods for other farm activities. On the other hand, a two-year crop/two-
year fallow cycle for upland rice fitted in well with the existing split of work
between upland and lowland cultivation.
3. Economics
In addition to its labour requirements, the adoption of a particular technique

will also depend on its economic benefits in relation to other options, regardless
of its biophysical attributes. For example, in a study of the economic viability of
combined fertiliser, green manure and grain legume techniques, Ali (1999) found
that the short-term benefits of green manure were “negative or trivial” compared
with inorganic fertiliser, despite on-farm experiments showing that yields were
higher with combined fertiliser and green manure treatments. On the other hand,
farmers using grain legumes obtained short-term benefits from the sale of the
grain as well as long-term benefits, compared to inorganic fertiliser systems. Ali
(1999) concluded that recommended techniques need to have short-term
economic benefits, or risk being rejected by farmers despite any long-term
advantages they might have.
The relative costs of inorganic sources of nutrients is a major determinant of
the use of LEIA techniques. In Nepal, for example, Pilbeam et al. (1999a) showed
that margins were generally negative when manure was applied either alone or in
combination with fertilizer, but positive with applications of fertilizer. Similarly,
in Zambia, subsidies on N fertiliser in the 1980s stimulated many subsistence
farmers to adopt high-input maize production based on mineral fertiliser inputs
(Matthews et al., 1992a). An economic analysis showed that under these
conditions, the best net returns were obtained from applying as much mineral
fertiliser N as possible (up to 120 kg N ha21 in the study) rather than using alley
cropping (Matthews et al., 1992b). However, when the fertiliser subsidies were
removed in 1990 through donor pressure, farmers reverted to their traditional
low-input chitemene and fundikila shifting-cultivation systems. Under these
conditions, alley cropping with Leucaena always gave a small positive return,
and indeed, the highest net return was obtained from Leucaena and 60 kg N ha21
fertiliser. This analysis did not, however, include the cost of labour; if it had, the
positive net returns from alley cropping would probably have disappeared.
Sometimes, however, the yield increase under alley cropping may be
economically viable. Chianu (2002) used partial budget analysis to show that
compared to a bush fallow, alley cropping L. leucocephala becomes
advantageous during longer fallow periods due to the production of fuelwood.
However, it was noted that yield variability, labour scarcity, and risk aversion
could influence the technology choice of the farmer. Similarly, novel alley
cropping systems may be economically viable to farmers. In India, introducing
geranium (Pelargonium spp.) into alleys of Eucalyptus citriodora did not affect
the essential oil yield of the latter, but resulted in higher monetary benefit over
sole Eucalyptus plots (Singh et al., 1998). Intercropping with Java citronella and
lemongrass also resulted in higher net benefits than from Eucalyptus alone,
although lemongrass did reduce Eucalyptus yields.
In India, Pakistan, the Philippines, and Indonesia the prices of both land and
labour have increased relative to inorganic fertiliser since the 1970s;
consequently, the use of the latter has increased at the expense of labour-
intensive, land-extensive, organic matter approaches to maintaining soil fertility
(Ali, 1999). In Taiwan, for example, green manure crops have decreased from an
area of 153,000 ha in 1954 to 11,000 ha in 1991, while in India, Nepal, and
Pakistan, green manure is no longer widely used. Ali (1999) calculated the
economics of using Sesbania, Azolla, and rice straw as green manures in India,
Indonesia, and the Philippines, assuming that a Sesbania green manure crop
would provide 70 kg N ha21, Azolla 30 kg N ha21, and rice straw 18 kg N ha21.
Ignoring the opportunity cost of growing green manure on productive land, he

found that the benefit/cost ratio of all three techniques was less than 1.0. This
indicated that at these labour/fertiliser price ratios, green manure was not a cost-
efficient option in any of the three countries compared to inorganic fertiliser. Ali
(1999) calculated that for green manure to be competitive with fertiliser and with a
zero opportunity cost of land, the ratio of wages (US$ day21) to fertiliser price
(US$ kg21 N) should not exceed 3.0 for Sesbania and Azolla, and 2.0 for rice
straw. As the ratio was higher than this in India, Indonesia, and the Philippines,
this explained the low use of green manure techniques there, while a more favou-
rable ratio accounted for the greater use of green manure in Myanmar and China.
Interestingly, Swinkels et al. (1997) found that farmers in the high population
density areas of western Kenya were more likely to practice fallowing due to the
availability of off-farm income, contradicting the view that continuous cropping
on depleted soils is the final stage in the land intensification process (e.g.,
Ruthenberg, 1980). Cropping depleted land gave farmers poorer returns to
labour, and, therefore, it was more rational to take work to buy food and allow the
soil fertility to regenerate. Only a 21% increase in crop yield following a one-year
fallow was necessary to compensate for the loss of returns during the fallow,
mainly due to the savings in crop husbandry costs. Similar behaviour has been
observed in Zambia (Low, 1988), Kenya and India (Dewees and Saxena, 1995),
and Indonesia (Nibbering, 1991).
As land and labour prices rise, there is pressure to shift away from green
manure towards intercropping, grain legumes, compost, and animal manure.
In India, farmers growing Crotalaria juncea and Tephrosia purpurea as green
manure after winter rice moved to grain legumes such as C. cajan and Vigna spp.,
particularly if irrigation water was available. In the central terai of Nepal,
Crotalaria juncea and Sesbania rostrata have given way to grain legumes such as
mung-bean, particularly if water is available during the dry season. In many
cases, only dramatic increases in fertiliser prices due to scarcity of fossil fuels
may make green manure viable in these countries, because transport networks are
well developed (Ali, 1999). Grain legumes, on the other hand, have potential if
the grain has economic value, although the removal of nutrients in the grain will
largely undermine their utility as a source of N.
Economic analyses of improved fallow systems have given conflicting results.
In upland northern Vietnam, a Tephrosia fallow had a negative net present value
(NPV) and was judged not to be a rational choice where natural fallow was still
viable (Hoang Fagerström et al., 2001). In Mali, Kaya et al. (2000) concluded
that improved fallows were only financially attractive if fodder had a value and if
subsequent crop yields exceeded the regional average of 2500 kg ha21. This is in
contrast to the example in western Kenya, where improved fallows were seen to
be a rational choice because of the relatively small increase in crop yields
required to break even, provided labour demands of establishing the fallow were
low (Swinkels et al., 1997). A study in West Africa showed that cumulative net
Figure 3 Cumulative discounted net benefits for continuous maize with no fertiliser, continuous
maize with recommended, one-year Sesbania fallow fertiliser, two-year Sesbania fallow, and three-
year Sesbania fallow. (Drawn from data of Kwesiga et al., 1999).
benefits were $1664 ha21 under a Pueraria fallow, $1121 ha21 under
Chromlaena odorata natural fallow and cover crops, and $1113 ha21 under
continuous cropping at farmer input levels (Tian et al., 2001). In Cameroon,
Adesina and Coulibaly (1998) found that improved fallows of Tephrosia,
Sesbania, Mucuna, and Calliandra were all profitable whether used alone or in
conjunction with inorganic fertilisers. In Zambia, Kwesiga et al. (1999) showed
that one- and two-year Sesbania fallows gave higher net returns than continuous
unfertilised maize crops, while a three-year Sesbania fallow was the same
(Fig. 3). Fertilised continuous maize, however, gave more than twice the net
return of the best fallow treatment. These authors make the point that the timing
of cash flow from improved fallows is important for subsistence farmers — even
though the net return of the 2-year fallow was higher than the unfertilised maize
after 4 years, for the first 2 years the net returns were negative, and the farmer
would need to have other sources of food and income.
IV. DISCUSSION
A. INTEGRATED NUTRIENT MANAGEMENT
In general, LEIA technologies aim at reducing losses, while also introducing

inputs from organic sources, while HEIA technologies aim at ensuring that
agronomic inputs (i.e., fertilisers, pesticides, water, energy) into the production
system are maintained at a high level, with less attention being paid to reducing
losses from the system. Gregory et al. (2002) refer to these two approaches as
Type I and Type II intensification, respectively. It is clear from the analysis in this
review that the LEIA techniques we have discussed cannot alone meet crop
nutrient requirements if potential yields are to be obtained. On the other hand,
sole reliance on HEIA technologies is also not an option for many resource-poor
farmers due to cost and availability of the inputs required, and is also not
desirable because of the associated health and environmental pollution problems.
There is, therefore, a growing consensus amongst researchers that the debate
on whether HEIA or LEIA technologies are the most appropriate is largely
irrelevant, and that the best way forward is through an INM approach, in which a
combination of techniques from the two extremes is used (e.g., Sanchez et al.,
2001). INM has been defined as the “judicious manipulation of nutrient stocks
and flows, in order to achieve a satisfactory and sustainable level of agricultural
production” (e.g., Deugd et al., 1998). Certainly, there is ample evidence (e.g.,
Palm et al., 1997a; Jadhao et al., 1999; Prasad et al., 2002) that the highest yields
and returns can be obtained by a combination of maximising inputs into the
system, both from organic and inorganic sources, and at the same time, reducing
losses. Inorganic fertilisers have the advantage that nutrient concentrations are
much higher than in organic material, so that handling and incorporation into the
soil is greatly facilitated. On the other hand, organic matter in the soil acts as both
a “binder” for added inorganic nutrients so that they are less likely to be lost by
leaching and volatilisation and more likely to be taken up by a crop, and also as a
source of nutrients in its own right through decomposition. It is also critical in
improving the physical structure of the soil. The resulting greater efficiency of
nutrient use through the combined use of both inorganic and organic sources of
nutrients and reduced losses is referred to as Type III intensification by Gregory
et al. (2002).
The challenge, therefore, is to identify those techniques from the continuum
between the HEIA and LEIA extremes, which can complement each other to
achieve this goal of sufficient and sustainable production, regardless of whether
they are labelled “organic” or “inorganic.” In general terms, Sanchez et al. (2001)
have proposed a combination of (1) biological N fixation by short-term
leguminous fallows, (2) applications of mineral P fertilisers, (3) enhanced P
cycling using Tithonia, (4) use of trees to maximise nutrient cycling, (5) return of
crop residues, (6) soil erosion prevention, (7) improved crop management
practices such as the use of better varieties, and (8) improved availability and
timeliness of supply of inorganic fertilisers. For this approach to be feasible, ways
need to be found to extract the indigenous deposits of rock phosphate present in
many countries, particularly those in sub-Saharan Africa (Ayuk, 2001), and make
it available to farmers at reasonable cost. At the farm level, improved
management of organic resources, such as in the storage and application of
compost and manure, may help reduce losses by volatilisation (Pilbeam et al.,
1999b). Similarly, the development of alternative energy sources such as
woodlots and the introduction of more efficient stoves may help to free up crop
residues currently used as fuel sources, which could then become available for
SOM maintenance and improvement (Ayuk, 2001).
Many of the nutrient management approaches in subsistence systems involve
moving nutrients from one part of the landscape at various scales to another
where they are more useful. For instance, several of the techniques we have
reviewed involve collection of nutrients from a wider area through grazing or
biomass harvesting and concentrating these on a smaller cropped area. In Nepal,
for example, the fertility built up in forested areas over long periods of time is
transferred gradually to nearby farms through collection of fodder for animals
and biomass for enhancing on-farm soil fertility (Pilbeam et al., 1999b). In much
of the analysis of farming systems, however, this heterogeneity is often ignored
and the fertility of a farm is assumed to be constant across all fields or parts of the
farm. In reality, farmers are usually very adept (consciously or subconsciously) at
manipulating this heterogeneity to improve their livelihoods. Both Wortmann
and Kaizzi (1998) and Briggs and Twomlow (2002) describe flows of organic
material from distant maize fields to higher value banana plantations nearer the
household on smallholder farms in Uganda. In many countries, the soil fertility of
small areas used as home gardens is enhanced by incorporating household waste
containing nutrients gathered from a wider area, both from other parts of the farm
through consumption of crops harvested there, and from off-farm sources such as
food bought in a market. Higher-value crops such as vegetables or fruit, which
would not grow well elsewhere on the farm, are often grown in these high-
fertility areas.
Indeed, Sanchez et al. (2001) have suggested that the growing of high-value
crops may be the most direct way out of poverty. For example, they quote high-
value vegetables such as kale, tomatoes, and onions in Kenya as being able to
increase net profits from US$91 to US$1665 ha21 year – 1. Whether this is viable
on a large scale will depend on broader economic development and the
availability of markets, storage and processing facilities, and urban population
growth rates. High-value tree crops may also be promising. For example,
extractions from the bark of Prunus africana can be used to treat prostate gland-
related diseases, and has an annual market value of $220 million per year (Sanchez
et al., 2001). The demand has been so high that the species is now on the CITES
(Convention on International Trade in Endangered Species) list, but is now in the
process of being domesticated. Other examples include bush mango (Irvingia
gabonensis) in West Africa, and Sclerocarya birrea from the miombo woodlands
of southern Africa, which are used to make liqueurs. The danger is, however, that
without replenishment in some form or another, eventually the fertility of
those parts of the farm providing nutrients for these high-value crops will decline
to a point where it is no longer possible for them to act as a nutrient source.
Clearly, such systems are complex, and there are questions regarding overall
sustainability, but with the advent of new crops, new markets, and better
infrastructure in many regions, it may be worthwhile re-examining these
traditional systems of nutrient redistribution carefully to see if there are any
possibilities of adding value to the overall production system through innovative
practices that farmers have not yet discovered.
The NUTMON project is an example of an approach to develop INM
strategies at the farm level (de Jager et al., 1998). The prime objective of the
project was to investigate technologies that mitigate against nutrient depletion
and possibly add nutrients which are economically viable and socially
acceptable. This was done by monitoring nutrient inflows and outflows of a
farm, calculating the balances, and quantifying the impact of INM practices on
soil fertility, and hence agricultural production and sustainability. The approach
included a diagnosis phase during which both qualitative and quantitative
assessments of nutrient management and stocks were made from farmer
interviews and by taking and analysing soil samples. Then, the most appropriate
technologies for a specific farming system were determined from a combination
of the quantitative nutrient flows and balances, the economic performance
indicators, and farmers’ perceptions, which were subsequently tested through on-
farm trials. Existing indigenous or science-based technologies as well as any new
ideas or modifications of existing technologies were considered. The approach
allowed the study of a farm in a holistic way, taking into account the effects of
many different household activities on the nutrient stocks and flows and the
economic performance of the farm, providing a way to assess the constraints to
adoption of alternative INM technologies in relation to economic viability and
demand for labour. Briggs and Twomlow (2002) followed a similar approach in
determining flows of organic material within smallholder farms in southwest
Uganda. By helping farmers to conceptualise and draw diagrams of these flows
within their farms, several sources of organic material, such as hedgerows,
weeds, fallowed areas, and ash residues, were identified which farmers had not
previously recognised as potentially contributing to the household’s fertility
management practices.
At the national level, Cuba provides an interesting example of how INM
practices can work (Carney, 1993). With the collapse of the Soviet empire at the
end of the Cold war in 1989, the country was deprived of a source of imported
fertilisers, pesticides, and fossil fuel. To cope, it was forced to look for other ways
of sustaining its agricultural production, and focused on more efficient nutrient
recycling and reuse of organic urban waste, and biological pest control. Although
food production initially dropped by 30%, it has since risen steadily, and to date
(2003) there are no food shortages despite soils being severely degraded, a high
population, and a continuing economic blockade. Better integration between
rural and urban areas may be a key focus for other countries — in general in
developing countries, there is a net flow of nutrients from the former to the latter
in the form of produce for human and animal consumption, which is not often
returned due to the inability or reluctance to process human and other waste. In
Ghana, for example, poultry manure produced in urban areas is considered to
have little value and is often dumped on the roadside (Quansah et al., 2001).
Thus, ways in which these nutrients can be retained and returned to the rural
areas, rather than being lost to the groundwater or sea as usually occurs, need to
be developed, taking into account, of course, the costs involved in processing and
transport, and possible disease risks. Currently, the economics are not
favourable — Palm et al. (1997a) showed that in Kenya, N and P in urban
compost cost US$0.5 kg21 and US$1.2 kg21, respectively, compared to
US$0.42 kg21 and US$0.18 kg21 in purchased inorganic fertilisers.
It is unlikely that there are any techniques that will provide universal
solutions — it is much more likely that progress will be made by farmers taking
an idea and adapting it to their own particular “microniche.” These microniches
will be unique for every farmer — not only will the biophysical environment
vary, but each farmer will also have different perspectives based on their interests
and experiences and be influenced by his/her own particular worldview (Scoones
and Toulmin, 1998). It is perhaps more important that farmers have a good
understanding of the principles involved in nutrient management, pest manage-
ment, and crop interactions, rather than a detailed knowledge of a particular
technique, and know how to apply this understanding to their own situation.
Deugd et al. (1998) have emphasised the importance of any improvements to
nutrient management being through farmer participation and learning. Thus,
improvements should not be seen as optimal solutions to scientifically well-
defined problems, but more as stages in an adaptive learning process within a
complex and changing environment.
B. A SYSTEMS PERSPECTIVE
It is evident from many of the examples we have reviewed, that a major reason
for the lack of widespread uptake of particular technologies is the failure to see
them as part of a larger system. For example, the decline in numbers of farmers
using the maize –Mucuna system in Honduras was found to be due largely to
external socio-economic factors independent of the agronomic performance of
the system itself (Neill and Lee, 2001). Similarly, despite the large increases in
rice yields and high returns to labour claimed for the System of Rice
Intensification, it has not been adopted widely by farmers, and a sizeable
fraction of those who did adopt it are now in the process of abandoning it (Moser
and Barrett, 2002). This was ascribed to the large amount of labour required by
the technology at the time of the year when poor farmers, facing cash shortages,
need to work for other farmers to earn enough to buy food. It remains to be
seen whether provision of credit facilities can ease this financial pressure to seek
off-farm work so that technologies such as SRI can be adopted. It is also

questionable whether the large quantities of biomass required for making the
compost used in the system would be available for the sustainable uptake of the
technology on a large scale, both in terms of the numbers of farmers involved, and
in terms of the actual cropped area under the technology. This biomass must of
necessity, come from a wider area than that actually cropped. Widespread uptake
of the Machobane system in Lesotho (Pantanali, 1996), and the Zai system in
Burkino Faso (Roose et al., 1999), both of which also rely on large supplies of
compost for their success, is likely to face the same constraint. Thus, failure to
consider the whole livelihood system, including the influence of temporal flows
of labour and money on farmer decision making, can lead to false conclusions
and unrealistic expectations on the part of the scientists about the appropriateness
of a particular technology. Ironically, the SRI technology has been more widely
adopted by middle-income farmers so far than by the poor farmers it was
designed to help (Moser and Barrett, 2002).
It is important, therefore, for researchers to try and view the situation from a
farmer’s perspective (Douthwaite et al., 2002). In general, most researchers
conceptualise problems associated with LEIA systems in developing countries in
purely biophysical terms, particularly in relation to soil fertility decline and weed
encroachment; less attention has been paid to the socio-economic issues (Ayuk,
2001). While this approach is useful from the perspective of scientific research,
farmers seldom think in these terms. Rather, they are more concerned with how
particular practices relate to their broader livelihoods. In considering whether or
not to adopt a particular research product such as alley cropping, they are more
likely to be influenced by how their livelihood will benefit in terms of the extra
food, cash, or quality of life it is likely to provide, than they are from technically
framed arguments concerned with nitrogen fixation and the like.
This is not an argument about a “reductionist” versus an “holistic” approach.
The fallaciousness of drawing a dichotomy between the two approaches has been
pointed out by Kline (1995). Neither is superior to the other, and we would argue
that a “reductionist” approach is essential, provided it is contextualised within a
broader framework of analysis. An example of such a framework is the
sustainable livelihoods (SL) approach currently being promoted by a number of
aid agencies as a way of thinking more broadly about the objectives, scope and
priorities for development, in order to enhance progress on the elimination of
poverty (Ashley and Carney, 1999). The main feature of the SL approach is that it
places people at “centre stage” of these discussions, rather than natural resources
or commodities as has been the case in the past, and considers their assets
(natural, human, financial, physical, and social capital) and their external
environment (trends, shocks, and transforming structures and processes). A key
concept is that of “sustainability” — a livelihood is defined as sustainable where
it can cope with (and recover from) stresses and shocks, and maintain or enhance
its capabilities and assets both now and in the future, while not undermining
the natural resource base (Carney, 1998). An important point is that agriculture is
seen as a system, rather than as simplified issues arising from single discipline
perspectives. We would argue, therefore, that a more appropriate approach might
be to consider the farmers’ problems from a livelihoods perspective, identify
potential solutions to these problems, and match or develop relevant techniques in
order to achieve these solutions. An illustrative example is provided in Table IV.
The three broad problems we have defined in Table IV are included in the
livelihood outcomes of the SL framework, while the various techniques we have
been reviewing contribute to the various livelihood strategies that subsistence
farmers can adopt to achieve these outcomes. We do not claim that Table IV is
exhaustive, but have attempted to present a different way of looking at possible
interventions, one that perhaps corresponds more closely to problems perceived
by farmers. Whether or not a particular technique is adopted will depend on the
balance between the perceived benefits and the costs of obtaining these benefits,
particularly, but not exclusively, in terms of the land, labour, and capital that is
required. Although the abilities of a technique to meet researchers’ expectations,
such as soil fertility enhancement or better weed control, is important, these
abilities are not necessarily how the farmers value them. This approach also
allows the consideration of other options besides natural resource management
techniques. For example, a cash-generating activity might be for some of the
household members to seek work in a local town or abroad. In many cases, this
could be a better option than trying to grow a cash crop for this purpose, as returns
Table IV
Possible Approach to Addressing Relevant Problems of Subsistence Farmers and Matching of
LEIA Techniques and Practices to Problem Solutions
Problem being addressed Solution Technique
More food for the Increased yields Improved varieties, intercropping, animal
household manure, composting
Extra food source Multipurpose trees, cover crops,
intercropping
More cash generated for Increase sales of surplus Enriched fallow, cover crops,
the household produce animal manure, multipurpose trees,
composting, crop residues
Enhanced quality of life Reduce labour Improved fallows
for members of the
household
More varied diet Crop diversification, cover crops,
multipurpose trees
Ease of cultivation Animal manure, cover crops
Fuel source Animal manure, multipurpose trees,
crop residues
Aesthetic value Tithonia hedgerows
to labour may be greater. Thus, not only is there a much wider range of possible
solutions that can be evaluated, but also any improved solutions developed are
much more likely to be adopted by farmers.
Gender issues are also important. Women make up a sizeable proportion of
tropical farmers, and it is they who most often focus on subsistence crops,
generally using lower inputs of organic and inorganic fertilisers than men
(Gladwin et al., 1997). In a study on constraints faced by women using organic
agriculture, Gladwin et al. (1997) found that lack of capital prevented them from
investing in either organic or inorganic fertilisers, lack of land limited their use of
low-input organic techniques, and lack of labour limited their ability to undertake
the activities that were required to implement such techniques, particularly as
most women were also solely responsible for household duties and child care. In
Senegal, women have even less rights to long-term use of land than do men
(Golan, 1994), and therefore have no incentive to make long-term investments to
improve soil productivity. Another gender-related problem that may act against
the uptake of any improved technique is that additional incomes arising from
sales of produce may go directly to the men in households, who are less likely
than women to invest in children and the household as a whole (Pretty and
Koohafkan, 2002).
Cultural traditions may also restrict the uptake and use of particular
techniques, despite any other benefits they may have. For example, in relation
to the potential use of animal manure in Ghana, Kiff et al., (1997) found from
surveys in a number of villages that farmers knew that manure could be used, but
generally found its use unattractive due to its supply being unreliable, too much
effort involved in its collection, and because they felt that manuring techniques
were regressive and old-fashioned. In certain areas, there is also the cultural
problem of persuading farmers who have no tradition of cattle husbandry to
develop the knowledge and interest to keep them on their farms (Dickson and
Benneh, 1995). In Nepal on the other hand, there is a long tradition of keeping
livestock, and the integration of animal manure into farm nutrient management is
well developed. Animals are multifunctional in Nepalese agricultural systems,
and provide meat, milk, ghee, curd, and draught power, as well as manure. Many
families may own more than a single species of animal, with the most common
combination (60% of those owning animals) being cattle, goats, chickens and
buffaloes (Pilbeam et al., 1999b). Often there are competing demands for manure
produced by livestock, particularly for use as fuel or in construction, which may
make its availability as a nutrient source scarce.
The use of the livelihoods-oriented approach described above may help to
identify the real limitations of agricultural production systems more clearly. For
example, the destruction of primary forest at the forest margins in Bolivia and
Brazil is driven by other more powerful factors than soil fertility or weed
encroachment issues. There, the underlying causes are economic or political
in nature — aided by government subsidies, wealthy landowners buy out the
frontier colonists to obtain land for cattle ranching, so that the latter then move on
and clear new land. Converting primary forest to pasture is therefore an effective
“cash generating” option for resource poor farmers that fits neatly within wider
economic and political realities. Consequently, the introduction of LEIA soil
fertility and weed management techniques in an attempt to stabilise agricultural
production at the forest margin, whilst technically feasible, is rendered almost
irrelevant by these broader political and economic factors (Muchagata, 1997).
C. MODELLING
To integrate all the various enterprises on a farm and to understand the often
complex differences between farms and farmers, tools are needed. The SL
framework just described is a useful start in this direction, but is limited in that it
does not capture the dynamic and spatial nature of farming systems, particularly
in relation to nutrient flows (Scoones and Toulmin, 1998). Simulation modelling
is a tool that offers the potential to integrate knowledge from a range of different
disciplines and to explore the complex relationships between them in a dynamic
and spatial way. For this reason, we believe that it is important that effort is made
in developing simulation models of subsistence agricultural systems so that the
processes involved are made explicit and to identify gaps in our knowledge.
Because of the long-term nature of many of the underlying processes, such
modelling offers a cost-effective and relatively quick way of obtaining answers to
questions regarding potential interventions. Such models could also help to
explore some of the wider global issues such as climate change, deforestation,
and desertification from the livelihoods perspective discussed above.
The type of modelling we propose to be the most appropriate for this task is an
integration of the key biophysical and socio-economic processes at the level of a
household. A large number of household models incorporating these aspects
already exist (e.g., de Jager et al., 1998; Pagiola and Holden, 2001), but most of
these are static models providing only a snapshot of the state of a household at an
instant in time and do not capture the dynamic characteristics of household
activities (Scoones and Toulmin, 1998). In a recent attempt to make progress in
this area, Shepherd and Soule (1998) developed a farm simulation model to
assess the long-term impact of existing soil management strategies on the
productivity, profitability, and sustainability of farms in west Kenya. The model
ran in time-steps of one year and linked soil management practices, nutrient
availability, crop and livestock productivity, and farm economics. Crop types
included weeds, fodder, grass, shrubs, and two types of grain crops. Growth of the
plants was determined by N and P availability. A wide range of soil management
options were simulated, including crop residue and manure management, soil
erosion control measures, biomass transfer, improved fallow, green manuring,
crop rotations, and N and P fertiliser application.
The model was used to examine the sustainability of existing farming systems
in the Vihiga district in western Kenya. Three household types were simulated to
represent the range of resource endowments of households in the study area. Data
to initialise the model for the different household types were collected through
participatory research in the area, which indicated large differences in farm size,
quantity and quality of livestock, soil and plant management, food consumption
patterns, and sources of income. It was shown that the low and medium resource
endowment farms had declining SOM, negative C, N, and P budgets, and low
productivity and profitability. The high resource endowment farms, on the other
hand, had increasing SOM, low soil nutrient losses, and were productive and
profitable. This disaggregation into household types according to resource
endowments highlighted the dangers of relying on nutrient balances of an
“average” farm type — most previous studies in Africa have generally shown
negative nutrient balances using this approach. Nevertheless, in this particular
case, the low and medium resource households represented around 90% of the
total in the study area, suggesting that overall a negative nutrient budget was
likely. There was also the question of where the increasing nutrients of the high
resource households were coming from — their greater purchasing power may
have meant that there was nutrient flow from the poorer households to the richer
ones. Shepherd and Soule (1998) concluded that the ability of the high resource
households to manage their farms profitably and sustainably indicates that it is
possible, but that capital is required. Strategies they suggested to improve
livelihoods included: (1) an increase in the value of farm input, (2) an increase in
high quality nutrient inputs at low cash and labour costs to the farmer, and (3) an
increase in off-farm income.
In another example, as part of a project evaluating integrated agriculture –
aquaculture farming systems in the Philippines, Schaber (1997) developed a
whole-farm model called FARMSIM to quantify flows of nutrients between the
different farming enterprises. The ORYZA_0 rice simulation model was
combined with a fish-pond model (for Nile Tilapia Oreochromis niloticus (L.))
and models of pigs, chickens and buffaloes. After the rice was harvested, the
straw was assumed to be composted, the bran to be fed to the pigs and to the fish,
and broken rice to be used as chicken food. If this supply of food was less than the
demand, then more had to be purchased externally. Weeds from the field bunds
were fed to the buffaloes. Manure from the pigs and buffaloes was fed to the fish,
although buffalo manure could also be applied to the rice fields. The model was
then used to evaluate three different scenarios in terms of the efficiency of N use
(defined as the output N as a ratio of the input N) of the farm as a whole. In the
first scenario, a conventional farming system with a mono-cultural rice field, two
buffaloes and fifteen chickens was assumed. High levels (200 kg N ha21) of
commercial fertiliser were applied to the rice field, and the output of rice grain
was high. In the second scenario, diversification increased as pigs were
introduced. The third scenario represented a fully integrated, diversified farming
system, including a fish-pond, with all reusable products from each farm
enterprise being used as inputs to other enterprises where appropriate. The
predicted efficiencies of N use in each of the three scenarios were 13.3, 18.7, and
21.6% respectively. It was concluded that the greater the number of enterprises
there were on the farm, the greater the efficiency of N use of the farm as a whole.
A similar household model is currently being developed as a tool to help
evaluate the relevance of potential soil fertility-enhancing techniques to
livelihoods of farmers in the mid-hills of Nepal (Matthews, 2000). In addition
to the biophysical processes of crop and animal growth, and water and nitrogen
fluxes through the household, economic and labour flows have also been
incorporated, along with household resources such as food, money, manure,
fodder, and fertiliser. The model, therefore, incorporates elements of the natural,
human, and financial capitals in the SL framework described above. Various
types of household can be accommodated, ranging from resource poor to
resource rich. The model will be used in the first instance to evaluate potential
interventions in the existing system and the likelihood of uptake of these
interventions, using criteria such as their contribution to household finances, food
production, alleviation of risk, and labour demands in relation to other farm
enterprises.
So far, these models are not spatially explicit, but they do have the capability
to become so. This is important because, as we have already discussed, the spatial
relationship between different parts of the landscape is often central to the
functioning of a farming system, with some areas acting as nutrient sources and
others as nutrient sinks. The net movement of nutrients from fields far from the
household to those close by on smallholder farms in Uganda (Wortmann and
Kaizzi, 1998; Briggs and Twomlow, 2002) has already been mentioned, while in
Burkino Faso, Prudencio (1993) found that 85% of household manure was
applied to nearby plots and only 15% to distant plots. Similarly, farmers often
consciously manipulate erosion and run-off from dry toplands as a means of
concentrating moisture and nutrients on the wetter and more accessible
bottomlands where they can benefit from them (Scoones and Toulmin, 1998).
Thus, extrapolation of results from measurements at a single site can be highly
misleading. The household models described above could take this into account
by incorporating a number of associated land management units (e.g., fields) that
are linked spatially, both in the horizontal and vertical directions if necessary.
Heterogeneity at the village level could then be described by linking a number of
such individual household models together (Matthews and Stephens, 2002). Such
models would then provide insights into the movement of nutrients within the
landscape, and possibly suggest ways in which this could be optimised to benefit
people’s livelihoods.
At some point in the future, these models should also incorporate household
decision-making processes, based on a labour and economic analysis each year of
the various household enterprises (crops, livestock, off-farm work, etc.), also
taking into account subsistence needs and attitude to risk. However, considerable
thought needs to be given to the dynamic processes involved in household
decision-making, and how these are influenced by the biophysical environment.
Some progress has been made by Pagiola and Holden (2001) and Angelsen and
Kaimowitz (2001) in determining when forest clearing is likely to be a rational
decision for farmers. The emerging field of multi-agent systems simulation, in
which artificial intelligence approaches are incorporated and interactions
between individuals are central (e.g., Bousquet et al., 1999), may be one way
in which further progress can be made in this direction. This is one area where
multidisciplinary research involving biophysical scientists and social scientists is
likely to be fruitful.
It is important to emphasise that while such models cannot be used to predict
the behaviour of specific households precisely, they are useful as tools for
understanding and testing hypotheses regarding the processes involved in
interactions between the biophysical and socio-economic environments of
subsistence farmers, and how these processes relate to their livelihoods and
poverty. Exploration of viable pathways out of poverty is more important than the
prediction of final endpoints. In the context of LEIA technologies, some of the
types of questions that can be addressed with such models are as follows:
1. The potential of LEIA techniques: While LEIA techniques such as the ones we
have discussed in this review can make a useful contribution to maintenance
of soil fertility, they are unable to supply enough nutrients required to achieve
the genetic potential of high-yielding crops. However, it would be useful to
know what level of crop yields could be sustained by the sole use of such
techniques in different environments and contexts, and how farmer livelihoods
are affected by this.
2. Evaluation of fallow types: Natural fallows offer a way of regenerating soil
fertility, but land must be set aside for long periods of time. Where land is
relatively plentiful, natural fallowing is a rational strategy. However, where
population increases and the availability of land decreases, the opportunity
cost of setting aside land for long periods of time rises significantly. Improved
fallows may be able to speed up the regeneration process, but at what level of
land availability does it become worthwhile for a farmer to consider the
technique? Similar questions can be asked for enriched fallows, where the
regenerative process is accompanied by income generation, taking into
account the possibly slower regeneration rate due to removal of harvested
material. Experimental determination of these issues is time-consuming and
expensive.
3. Managing variation in natural resources: By concentrating resources in one
area at the expense of another, higher value crops may be grown, leading to
an improvement in cash income for the household, some of which could be
re-invested in the poorer areas of the farm, thereby improving the overall
fertility of the farm in the long term. Many of the LEIA systems we have
discussed (e.g., the Machobane system, SRI, the Zai system) essentially
concentrate nutrients from a wider area into a smaller area that can be cropped.
What is the optimum way to do this for a given biophysical and socio-
economic situation?
4. Effect of a change in farmer perceptions: Recent work in Ghana has shown
that in evaluating different practices, farmers do not always value the
opportunity cost of their own labour, though they more readily assess the
financial cost of hiring others to work on their farms (Galpin et al., 2000).
Participatory interaction, however, has brought some of them to consider that
their own time and labour should be factors taken account of in the evaluation.
It would be interesting to compare the likelihood of adoption of various
techniques (both traditional and researcher-generated) with and without
consideration of the labour involved. Would patterns of development be
different in each case? Do more sustainable practices result from taking labour
into account? Or is the concept of opportunity cost of labour meaningless
when there are so few options available in which it could be deployed,
anyway?
5. Effect of current socio-economic trends: In Nepal in recent years, a decline in
soil fertility has been ascribed by farmers themselves to a decline in manure
applications, due in turn to a decline in livestock numbers brought about by a
reduction in the household labour pool with more and more children going to
school (Ellis-Jones, pers. comm.). School leavers are not interested in
returning to work on the farm, preferring to find jobs in the towns and cities.
What effect is this likely to have on the fertility of the soil in the first instance,
and on the overall livelihood of the household, bearing in mind that urban jobs
represent a potential source of cash income for the household in the future? Is
it a good livelihood strategy to invest in the education of one’s children, and at
what cost is this to the biophysical environment? Should government policies
aim to encourage the educated to take up farming, or is it desirable that hill
agriculture continues to decline?
6. Trajectories out of poverty: The question of whether there are “natural”
processes (in the broadest sense, including both biophysical and socio-
economic processes) that can lead to the evolution of one agricultural system
into another needs to be explored. Given that it is perfectly rational for poor
people to adopt short-term strategies that attempt to maximise their livelihood
outcomes, can improved (or even new) strategies be developed or promoted to
hasten the change from shifting cultivation systems to more settled patterns of
agriculture? Can LEIA techniques improve livelihoods even if they are used
efficiently, or are external inputs essential? Could governments adopt
particular policies that would facilitate the transition process? How is the
distribution of wealth influenced by different processes of transition? What are
the long-term environmental consequences of such transitions?
V. CONCLUDING REMARKS
For any agricultural system to be sustainable, regardless of whether it is

classified as LEIA or HEIA, inputs must not be less than outputs — if they are,
then mining of existing stocks occurs, and the system must deteriorate in time.
Where there is constant removal of nutrients in crops harvested for human
consumption and these are not returned to the system, then no such system can be
sustainable without some further input from outside. No group of technologies,
whether based on organic or inorganic sources of nutrients, can contravene this
basic law of mass conservation.
In this review, we have attempted to analyse the biophysical and socio-
economic characteristics of a number of LEIA techniques that have been
evaluated as potential improvements to subsistence agricultural systems. These
techniques included intercropping, alley cropping, cover crops and green
manures, biomass transfer, compost, use of animal manure, and improved and
enriched fallows. It is important to remember that farmers have been practising
variations of these techniques for generations, using organic materials that have
been available within their immediate vicinity. While population density has
been low, and land has been abundant, these systems have been sustainable, but
with the population explosion in the 20th century, the parameters have now
changed, and there is no guarantee that these systems will continue to function in
a sustainable way. Indeed, many are beginning to break down as cropping periods
are extended and fallow periods are shortened (e.g., Chidumayo, 1987). There is
no doubt that there is an urgent need to find ways to make these systems more
productive and sustainable, since it is upon the outputs of such systems that the
lives and livelihoods of so many of the world’s poor ultimately depend.
While there may be some scope in making existing systems more efficient
by reducing losses of nutrients, identifying new sources of organic material, or
spatially manipulating nutrient concentrations within farms (Wortmann and
Kaizzi, 1998; Briggs and Twomlow, 2002), most of the LEIA techniques we have
reviewed, when used alone, appear to have limited potential to increase food
production dramatically. In the case of those techniques aimed at maintaining or
improving soil fertility, the nutrient content and the quantity of biomass that can be
produced within the resources available to such farmers is insufficient to meet the
requirements of most crops, certainly at a reasonable yield level, although where
existing yields are very low, substantial relative yield increases may be possible
(e.g., Pretty et al., 2003). The contribution that these large relative, but small
absolute, increases in yields make to overall global food production needs to be
determined by weighting each yield level with an appropriate land area. Similarly,
those LEIA techniques aimed at weed control in the studies reviewed, while often
being able to reduce weed populations, had little effect on crop yields. In both
cases, there is also the possibility that crop yields may actually be depressed,
either as a result of competition for resources, particularly in difficult conditions,

or as a result of N or P immobilisation if low-quality organic material is incorpor-
ated into the soil. While LEIA technologies are often promoted as “natural” (e.g.,
Pretty, 1999), as we have seen, in reality many of them have very high demands on
land and labour. Thus, the debate over the relative merits of LEIA and HEIA
technologies essentially hinges on whether the inputs should be capital intensive
or land and/or labour intensive. Ironically, all of these are assets that by definition,
resource-poor farmers have in short supply. Any hopes of promotion of their use to
bring about radical changes to existing cultivation systems to address the problem
of future global food demand, and to provide impetus to lift such farmers out of the
poverty spiral, is, therefore, likely to be more utopian than realistic. A similar
conclusion was also reached by Sanchez et al. (2001), who argued that sole use of
LEIA technologies is only likely to perpetuate food insecurity and poverty.
The future, therefore, must surely lie in the integration of the two approaches,
using LEIA technologies when organic sources of nutrients are available, but also
being prepared to supplement these with external supplies if necessary and when
it is economic to do so. Taking a systems perspective is essential — there is no
doubt that higher inputs of either inorganic or organic forms of nutrients can
increase crop yields of individual plots or fields — the real issue, however, is
whether the supplies of these inputs are sustainable at higher scales, and whether
there is sufficient labour and capital in the system for their transport and handling.
There is little point in increasing crop yields per se if these cannot be feasibly
scaled up or maintained at that level for long. This applies to both LEIA and
HEIA technologies — while promotion of the use of external supplies of
fertiliser has been rightly criticised for being out of reach for many subsistence
farmers, so might the large amounts of land required to provide sufficient
nutrients in biomass, or the large amounts of labour at the right time to harvest,
transport and incorporate this biomass.
There is a tendency by researchers and proponents of both LEIA and HEIA
approaches to quote only yield increases and to ignore the overall constraints of
the system required to produce these increases (e.g., Uphoff, 1999; Singh and
Sharma, 2001), which can be misleading. We would make a plea, therefore, that
as well as providing adequate biophysical information on soils and climate, all
future reports of yield increases in LEIA techniques state the area of land on
which the crop yield has been measured, accompanied by estimates of the amount
of land from which the particular technique has gathered its nutrients from, and
the amount of labour required to harvest and transport these nutrients to the
cropped area. Similar analyses should be made for existing farmer practice, and
for HEIA techniques in terms of capital costs of purchasing, transporting, and
applying the external inputs. The economic costs and benefits of the technique
should also be determined, preferably dynamically, so that it is possible to see
whether resource demand matches resource availability throughout the year. In
this way, a more objective appraisal of all techniques may be possible.
Future research to improve subsistence agriculture needs, therefore, to focus

more on developing interventions to meet farmer realities, such as increased food
security, improved cash generation, reduced risk, and enhanced quality of life.
This may necessitate consideration of a broader range of interventions than just
natural resource management options. In this process, both “holistic” and
“reductionist” approaches are necessary and valuable in an iterative process of
investigation and analysis. The starting point should be from a holistic viewpoint,
the analysis of problems in the system and development of solutions should be
reductionist, and any successful solutions to the problems should be evaluated
holistically again. Many of the processes, both biophysical and socio-economic
and their interactions, are poorly understood, and it is essential that future research
addresses this.
This lack of knowledge is compounded by the large degree of heterogeneity of
the production systems involved, both at the system level with different cultivation
systems in the different countries, and also at the individual farm level with
between-farm variability in terms of farmer aspirations and attitudes, and within-
farm variability in resources. However, it is important that this heterogeneity is
preserved as it contributes to the resilience, and hence the sustainability, of the
production system. This was highlighted in the example of the maize – Mucuna
system in Honduras in which over-reliance on one cover crop species was probably
a major factor allowing the incursion of Rottboellia as Mucuna as more effective
against broad-leaf weeds than grasses (Neill and Lee, 2001). Of all of the LEIA
and HEIA technologies on offer, there is no single one that is a panacea for the
problems faced by subsistence farmers. Each has particular strengths and
weaknesses, and the challenge is to identify these strengths and combine them
into integrated systems capable of adapting to changing circumstances when
necessary.
ACKNOWLEDGMENTS
We are grateful to the Natural Resources Systems Programme of the United

Kingdom’s Department for International Development (DFID) for funding this
work. The review is condensed and updated from the Final Report of DFID
Project R7560.
REFERENCES
Adeorike, V., Ogburia, M. N., and Anegbeh, P. (2001). Evaluation of the allelopathic influence of
selected multipurpose tree species on maize (Zea mays) under a simulated field condition.
Tropicultura 19(4), 191 –193.
Adeoye, G. O., Sridhar, M. K. C., and Oyenekan, O. T. (1996). A survey of compost utilization for
ornamental plants in Nigeria. Compost Sci Utilization 4(4), 81–83.
Adesina, A., and Coulibaly, N. (1998). Policy and competitiveness of agroforestry-based technologies
for maize production in Cameroon: an application of policy analysis matrix. Agric. Econ. 19,
1–13.
Akinyemi, J. O., and Kalejaiye, A. S. (1998). The strategies for implementing the soil conservation
practices among the rural farmers in the tropical rain forest zone of Nigeria. Int. J. Tropic Agric.
14(1–4), 35–48.
Akobundu, I. O., Udensi, U. E., and Chikoye, D. (2000). Velvetbean (Mucuna spp.) suppresses
speargrass (Imperata cylindrica (L.) Raeuschel) and increases maize yield. Int. J. Pest Manag.
46(2), 103–108.
Akyeampong, E. (1999). Suitability of six hedgerow species for alley cropping on an acid infertile soil
in Burundi. Tropic. Agric. 76(2), 82 –87.
Ali, M. (1996). Quantifying the socio-economic determinants of sustainable crop production: an
application to wheat cultivation in the Tarai of Nepal. Agric. Econ. 14(1), 45–60.
Ali, M. (1999). Evaluation of green manure technology in tropical lowland rice systems. Field Crops
Res. 61, 61 –78.
Amede, T., Belachew, T., and Geta, E. (2001). Reversing the degradation of arable land in the
Ethiopian Highlands. Managing Africa’s Soils 23, 23.
Angelsen, A., and Kaimowitz, D. (2001). When does technological change in agriculture promote
deforestation? In “Tradeoffs or Synergies? Agricultural Intensification, Economic Development,
and the Environment” (D. R. Lee and C. B. Barrett, Eds.), pp. 89 –114. CAB International,
Wallingford, UK.
Anoka, U. A., Akobundu, I. O., and Okonkwo, S. N. C. (1991). Effects of Gliricidia sepium and
Leucaena leucocephala on growth and development of Imperata cylindrica. Agrofor. Syst. 16,
1–12.
Apiradee, K. (1988). Factors affecting the adoption of farmers’ compost making and application in
Amphoe Khao Yoi, Changwat Phetchaburi, Thailand. Kaesart University, Bangkok.
Araki, S. (1993). The effect of burning, ash fertilisation, and soil organic matter on productivity of the
chitemene shifting cultivation in northern Zambia. In “Soil Organic Matter Dynamics and
Sustainability of Tropical Agriculture” (K. Mulongoy and R. Merckz, Eds.), pp. 367–375. John
Wiley & Sons, New York.
Ashley, C., and Carney, D. (1999). Sustainable Livelihoods: lessons from early experience, p. 55.
Department for International Development, London, UK.
Atta-Krah, A. N., and Sumberg, J. E. (1988). Studies with Gliricidia sepium for crop/livestock
production systems in West Africa. Agrofor. Syst. 6, 97 –118.
Ayuk, E. T. (2001). Social, economic and policy dimensions of soil organic matter management in
sub-Sahara Africa: challenges and opportunities. Nutrient Cycling in Agroecosystems 61,
183–195.
Azam-Ali, S. N., Matthews, R. B., Williams, J. H., and Peacock, J. M. (1990). Light use, water uptake
and performance of individual components of a sorghum/groundnut intercrop. Exp. Agric. 26,
413–427.
Baggio, A., and Heuveldorp, J. (1988). Initial performance of Calliandra calothyrus Messn. in live
fences for the production of biomass. Agrofor. Syst. 2, 19–29.
Barrios, E., Kwesiga, F., Buresh, R. J., and Sprent, J. I. (1997). Light fraction soil organic matter and
available nitrogen following trees and maize. Soil Sci. Soc. Am. J. 61, 826–831.
van den Berghe, C., Sota, P., and Mujawayezu, A. (1994). Study on mineral complementation in the
composting process in farmers conditions of Mugamba, Burundi. Tropicultura 12(2), 55– 60.
Blarel, B. (1994). Tenure and agricultural production under land scarcity: the case of Rwanda.
In “Searching for Land Tenure Security in Africa” (J. W. Bruce and S. N. Migot-Adholla, Eds.),
pp. 71– 96. Kendall/Hunt Publishing Company, Dubuque, Iowa, USA.
Boddey, R. M., de Sa, J. C. M., Alves, B. J. R., and Urquiaga, S. (1997). The contribution of biological
nitrogen fixation for sustainable agricultural systems in the tropics. Soil Biol. Biochem. 29(5/6),
287 –799.
Bouldin, D. R., Mughogho, S., Lathwell, D. J., and Scott, T. W. (1979). Nitrogen Fixation by Legumes
in the Tropics. (International Agricultural Mimeograph). Report No. 75. Department of
Agronomy, Cornell University, New York.
Bousquet, F., Barreteau, O., Le Page, C., and Weber, J. (1999). An environmental modelling
approach: the use of multi-agent simulations. In “Advances in Environmental and Ecological
Modelling” (F. Blasco and A. Weill, Eds.), pp. 113–122. Elsevier, Paris, France.
Brady, N. (1990). In “The Nature and Properties of Soils”. Macmillan Publishing Company, New York.
Briggs, L., and Twomlow, S. J. (2002). Organic material flows within a smallholder highland farming
system of south west Uganda. Agric. Ecosyst. Environ. 89, 191–212.
Brodd, P., and Osanius, M. (2002). Erosion and land use at plot and catchment levels — a case study
in northern Vietnam, p. 62. Minor Field Studies International Office, Swedish University of
Agricultural Sciences, Uppsala, Sweden.
Buckles, D., and Triomphe, B. (1999). Adoption of mucuna in the farming systems of northern
Honduras. Agrofor. Syst. 47, 67 –91.
Budelman, A. (1988). Leaf dry matter productivity of three selected perennial leguminous species in
humid tropical Ivory Coast. Agrofor. Syst. 7, 42–62.
Buresh, R. J., and Niang, A. I. (1997). Tithonia diversifolia as a green manure: awareness,
expectations and realities. Agrofor. Forum 8(3), 29–31.
Cairns, M., and Garrity, D. P. (1999). Improving shifting cultivation in Southeast Asia by building on
indigenous fallow management strategies. Agrofor. Syst. 47, 37–48.
Carney, J. (1993). Low input sustainable agriculture in Cuba. Agric. Human Values 10(3) (Special
Issue).
Carney, D. (1998). Implementing the sustainable rural livelihoods approach, Sustainable Rural
Livelihoods — what contribution can we make?, pp. 3–23. Department for International
Development, London, UK.
Carter, J. (1995). Alley farming: have resource-poor farmers benefited? Natural Resources
Perspectives No. 3. Overseas Development Institute, London.
Catchpoole, D. W., and Blair, G. J. (1990). Forage tree legumes. III. Release of nitrogen from leaf,
faeces and urine derived from Leucaena and Gliricidia leaf. Aust. J. Agric. Res. 41, 539–547.
Cenas, P. A. A., Pandey, S., and Pandey, S. (1996). Determinants of adoption and retention of contour
hedgerows in Claveria, Misamis Oriental. Philippine J. Crop Sci. 21(1–2), 34 –39.
Chianu, J. N., Akintola, J. O., and Kormawa, P. M. (2002). Profitability of cassava-maize production
under different fallow systems and land-use intensities in the derived savanna of southwest
Nigeria. Exp. Agric. 38(1), 51 –63.
Chidumayo, E. N. (1987). A shifting cultivation land use system under population pressure in Zambia.
Agrofor. Syst. 5, 15 –25.
Chivasa, W., Harris, D., Chiduza, C., Nyamudeza, P., and Mashingaidze, A. B. (2000). Biodiversity
on-farm in semi-arid agriculture: a case study from a smallholder farming system in Zimbabwe.
Zimbabwe Sci. News 34(1), 13–18.
Coppock, D. L., and Reed, J. D. (1992). Cultivated and native browse legumes as calf supplements in
Ethiopia. J. Range Manag. 45, 231 –238.
Coultas, C. L., Post, T. J., Jones, J. B. Jr, and Hsieh, Y. P. (1996). Use of velvet bean to improve soil
fertility and weed control in corn production in northern Belize. Commun. Soil Sci. Plant Anal.
27(9– 10), 2171–2196.
Craswell, E. T., Sajjapongse, A., Howlett, D. J. B., and Dowling, A. J. (1998). Agroforestry in the
management of sloping lands in Asia and the Pacific. Agrofor. Syst. 38, 121–137.
Crosson, P., and Anderson, J. R. (2002). Technologies for meeting future global demand for food.
Discussion Paper 02-02. Resources for the Future, 1616 P Street, NW, Washington D.C. 20036.
Dalland, A., Våje, P. I., Matthews, R. B., and Singh, B. R. (1993). The potential of alley cropping in
improvement of cultivation systems in the high rainfall areas of Zambia. III. Soil chemical and
physical properties. Agrofor. Syst. 21, 117 –132.
Dasthagir, M. G., Kohli, R. K., Arya, K. S., and Atul (1996). Review of research on tropical
agroforestry. In “IUFRO-DNAES International Meeting; Resource Inventory Techniques to
Support Agroforestry and Environment”. pp. 306 –309. DAV College, Chandigarh, India.
David, S. (1995). What do farmers think? Farmer evaluations of hedgerow intercropping under semi-
arid conditions. Agrofor. Syst. 32, 15–28.
Degrande, A., and Duguma, B. (2000). Adoption potential of rotational hedgerow intercropping in the
humid lowlands of Cameroon, p. 8. Agricultural Research and Extension Network, Network
Paper No. 103. Overseas Development Institute, London, UK.
Delve, R. J., Cadisch, G., Tanner, J. C., Thorpe, W., Thorne, P. J., and Giller, K. E. (2001).
Implications of livestock feeding management on soil fertility in the smallholder farming systems
of sub-Saharan Africa. Agric. Ecosyst. Environ. 84, 227–243.
Deugd, M., Röling, N., and Smaling, E. M. A. (1998). A new praxeology for integrated nutrient
management, facilitating innovation with and by farmers. Agric. Ecosyst. Environ. 71, 269– 283.
Dewees, P. A., and Saxena, N. C. (1995). Tree planting and household land and labour allocation: case
studies from Kenya and India. In “Tree Management in Farmer Strategies: Responses to
Agricultural Intensification” (J. E. M. Arnold and P. A. Dewees, Eds.), pp. 198– 241. Oxford
University Press, Oxford.
Dickson, K. B., and Benneh, G. (1995). In “A New Geography of Ghana”. Longman, Harlow.
Diop, A. M. (1999). Sustainable agriculture: new paradigms and old practices? Increased
production with management of organic inputs in Senegal. Environ. Develop. Sustainability 1,
285–296.
Douthwaite, B., Manyong, V. M., Keatinge, J. D. H., and Chianu, J. (2002). The adoption of alley
farming and Mucuna: lessons for research, development and extension. Agrofor. Syst. 56(3),
193–202.
Dowling, N. G., Greenfield, S. M., and Fischer, K. S. (1998). Sustainability of Rice in the Global Food
System, p. 404. Pacific Basin Study Centre, Davio, California, and International Rice Research
Institute, Manila.
Drechsel, P., and Reck, B. (1998). Composted shrub-prunings and other organic manures for
smallholder farming systems in southern Rwanda. Agrofor. Syst. 39, 1–12.
Drechsel, P., Steiner, K. G., and Hagedorn, F. (1996). A review on the potential of improved fallows
and green manure in Rwanda. Agrofor Syst. 33(2), 109 –136.
Durban Metropolitan, (1999). Durban Metro State of the Environment and Development: 1999
(Accessed 12-Dec-2002 online at http://www.ceroi.net/reports/durban/issues/waste/state.htm).
Elbasha, E., Thornton, P. K., and Tarawali, G. (1999). An ex post economic impact assessment of
planted forages in West Africa. ILRI Impact Assessment Series, No. 2. International Livestock
Research Institute (ILRI), Nairobi.
Ellis, F. (1988). In “Peasant Economics: Farm Households and Agrarian Change”. Cambridge
University Press, Cambridge.
Enyong, L. A., Debrah, S. K., and Bationo, A. (1999). Farmers perceptions and attitudes towards
introduced soil fertility enhancing technologies in western Africa. Nutrient Cycling in
Agroecosystems 53, 177 –187.
Euroconsult, (1989). “Agricultural Compendium for Rural Development in the Tropics and
Subtropics”. The Netherlands Ministry of Agriculture and Fisheries, Amsterdam.
FAO, (1996). “Food Requirements and Population Growth”. Food and Agriculture Organisation,
Rome, Italy.
Farrell, R. E., Sandercock, P. J., Pennock, D. J., and van Kessel, C. (1996). Landscape-scale variations
in leached nitrate: relationship to denitrification and natural N15 abundance. Soil Sci. Soc. Am. J.
60, 1410–1415.
Fassler, O. M., and Lascano, C. E. (1995). The effect of mixtures of sun-dried tropical shrub legumes
on intake and nitrogen balance by sheep. Trop. Grasslands 29, 92–96.
Field, S. P., Momuat, E. O., Ayre-Smith, R., and Bamualim, A. (1992). Developing land rehabilitation
technologies compatible with farmer needs and resources. J. Asian Farm. Syst. Assoc. 1(3),
385 –394.
Franzel, S. (1999). Socioeconomic factors affecting the adoption potential of improved tree fallows in
Africa. Agrofor. Syst. 47, 305–321.
Fujita, K., Matsumoto, K., Ofosu-Budu, G. K., and Ogata, S. (1993). Effect of shading on growth and
dinitrogen fixation of kudzu and tropical pasture legumes. Soil Sci. Plant Nutr. 39(1), 43–54.
Gachengo, C., Palm, C., Adams, E., Giller, K. E., Delve, R. J., and Cadisch, G. (1998). Organic
Resources Database. Tropical Soil Biology and Fertility Programme (TSBF), Nairobi and Wye
College, University of London.
Gachengo, C. N., Palm, C. A., Jama, B., and Othieno, C. (1999). Tithonia and Senna green manures
and inorganic fertilizers as phosphorus sources for maize in western Kenya. Agrofor. Syst. 44(1),
21 –36.
Galiba, M., Vissoh, P., Dagbenonbakin, G., and Fagbohoun, F. (1998). The reactions and fears of
farmers utilizing velvet bean (Mucuna pruriens). Cover crops in West Africa: contributing to
sustainable agriculture. In “Proceedings of a conference 1–3 October 1996, Cotonou, Benin”
(D. Buckles, A. Eteka, O. Osiname, and G. Galiano, Eds.), pp. 55–64. International Development
Research Centre (IDRC), Ottawa, Canada.
Gallagher, R. S., Fernandes, E. C. M., and McCallie, E. L. (1999). Weed management through short-
term improved fallows in tropical agroecosystems. Agrofor. Syst. 47, 197–221.
Galpin, M., Dorward, P., and Shepherd, D. (2000). Participatory farm management methods for
agricultural research and extension: a training manual. Agricultural Extension and Rural
Development Department, Reading University, Reading, UK.
Ganry, F., Feller, C., Harmand, J. M., Guibert, H., Martius, C., Tiessen, H., and Vlek, P. L. G. (2001).
Management of soil organic matter in semiarid Africa for annual cropping systems. Nutrient
Cycling in Agroecosystems 61(1– 2), 105–118.
Garcia, J. N. M., Gerrits, R. V., and Cramb, R. A. (2002). Adoption and maintenance of contour bunds
and hedgerows in a dynamic environment: experience in the Philippine uplands. Mountain Res.
Develop. 22(1), 10 –13.
Garrity, D. P., and Lai, C. K. (2000). Innovation in shifting cultivation in Asia: indigenous fallow
management. Ileia Newslett. 16(3), 5–6.
Gashaw, G., Varvikko, T., and Khalili, H. (1991). Smallholder Dairy Production in Northern Shoa
Province, Central Highlands of Ethiopia. Assessment of Feed Resources in Relation to the
Requirements on a Smallholder Farm (Final Report), p. 25. International Livestock Centre for
Africa (ILCA), Addis Ababa.
George, T. S., Gregory, P. J., Robinson, J. S., Buresh, R. J., and Jama, B. A. (2001). Tithonia
diversifolia: variations in leaf nutrient concentration and implications for biomass transfer.
Agrofor. Syst. 52(3), 199–205.
Gill, H. S. (1982). The role of hand and mechanical weeding in weed management in the advancing
countries. Improving Weed Management (Plant Production and Protection Paper 44). FAO,
Rome, Italy.
Giller, K. E. (2001). “Nitrogen Fixation in Tropical Cropping Systems”, p. 313. CAB International,
Wallingford, UK.
Giller, K. E., and Cadisch, G. (1995). Future benefits from biological nitrogen fixation: an ecological
approach to agriculture. Plant Soil 174, 255 –277.
Giller, E. K., Cadisch, G., Ehaliotis, C., Adams, E., Sakala, W. D., and Mafongoya, P. L. (1997).
Building soil nitrogen capital in Africa. In “Replenishing Soil Fertility in Africa”. SSSA Special
Publication No. 51 (R. J. Buresh, P. A. Sanchez, and F. Calhoun, Eds.), pp. 151 –192. Soil
Science Society of America, Madison, Wisconsin.
Girish, G. K., and Chandrashekar, C. P. (2000). Farmers practice of nutrient management for
sugarcane in GLBC area of Karnataka — a status paper. Cooperative Sugar 31(6), 477 –484.
Gladwin, H., Buhr, K. L., Goldman, A., Hiebsch, C., Hildebrand, P. E., Kidder, G., Langham, M., Lee,
D., Nkedi-Kizza, P., and Williams, D. (1997). Gender and Soil Fertility in Africa. In “SSSA
Special Publication No. 51: Replenishing Soil Fertility in Africa” (R. J. Buresh, P. A. Sanchez,
and F. Calhoun, Eds.), pp. 219–236. Soil Science Society of America, Wisconsin.
Golan, E. H. (1994). Land tenure reform in the Peanut Basin of Senegal. In “Searching for Land
Tenure Security in Africa” (J. W. Bruce and S. N. Migot-Adholla, Eds.), pp. 231– 249. Kendall/
Hunt Publishing Company, Dubuque, Iowa, USA.
Goyal, S., Mishra, M. M., Hooda, I. S., and Singh, R. (1992). Organic matter– microbial biomass
relationship in field experiments under tropical conditions: effects of inorganic fertilization and
organic ammendments. Soil Biol. Biochem. 24, 1081–1084.
Gregory, P. J., Ingram, J. S. I., Andersson, R., Betts, R. A., Brovkin, V., Chase, T. N., Grace, P. R.,
Gray, A. J., Hamilton, N., Hardy, T. B., Howden, S. M., Jenkins, A., Meybeck, M., Olsson, M.,
Ortiz-Monasterio, I., Palm, C. A., Payn, T. W., Rummukainen, M., Schulze, R. E., Thiem, M.,
Valentin, C., and Wilkinson, M. J. (2002). Environmental consequences of alternative practices
for intensifying crop production. Agric. Ecosyst. Environ. 88, 279–290.
Grist, P., Menz, K., and Nelson, R. (1999). Multipurpose trees as improved fallow: an economic
assessment. Int. Tree Crops J. 10(1), 19–36.
Harris, D., Joshi, A., Khan, P. A., Gothkar, P., and Sodhi, P. S. (1999). On-farm seed priming in semi-
arid agriculture: development and evaluation in maize, rice and chickpea in India using
participatory methods. Exp. Agric. 35, 15–29.
Harsono, A. (1996). Potency and opportunity of soybean farming system on upland in Sumbawa
(NTB) and Sikka (NTT) districts. In “Proceedings of a Seminar on Legumes and Tuber Crops
Research Results in 1995” (N. Marwoto, A. Saleh, Kasno, and Sunardi, Eds.). Malang, Indonesia.
Hartemink, A. E., Buresh, R. J., Jama, B., and Janssen, B. H. (1996). Soil nitrate and water dynamics
in Sesbania fallows, weed fallows and maize. Soil Sci. Soc. Am. J. 60, 568–574.
Hauser, S. (1990). Water and nutrient dynamics under alley cropping versus monocropping in the
humid–subhumid transition zone. Transactions of the 14th International Soil Science Society
Congress, Kyoto, Japan, 204– 209.
Hellin, J., Welchez, L. A., and Cherrett, I. (1999). The Quezungual system: an indigenous agroforestry
system from western Honduras. Agrofor. Syst. 46(3), 229 –237.
Hoang Fagerström, M. H., van Noordwijk, M., Thai, Phien, and Nguyen Cong, Vinh (2001).
Innovations within upland rice-based systems in northern Vietnam with Tephrosia candida as
fallow species, hedgerow, or mulch: net returns and farmers’ response. Agric. Ecosyst. Environ.
86(1), 23–37.
Humphries, S. (1996). In Milk cows, migrants, and land markets: unravelling the complexities of
forest to pasture conversion in northern Honduras. University of Guelph, Guelph, Ontario,
Canada.
ICRAF, (1997). Using the Wild Sunflower, Tithonia, in Kenya for Soil Fertility and Crop Yield
Improvement. International Centre for Research in Agroforestry, Nairobi, Kenya.
IIRR, (1998). The Machobane farming system in Lesotho, Sustainable Agricultural Extension Manual
for Eastern and Southern Africa. pp. 171 –176. International Institute of Rural Reconstruction,
Africa Regional Office, PO Box 66873, Nairobi, Kenya.
Jackson, D., Kiff, E., and Chan, M.-K. (1999). Integrated Food Crop Systems Project, Ghana:
Development and Promotion of Improved Techniques of Water and Soil Fertility Management
for the Sustainable Production of Crops on Land in the Humid Forest Belt. (Progress Against
Milestones Review). Report No. 4242. Natural Resources Institute, London.
Jadhao, S. L., Giri, D. G., Kharat, B. S., Deshmukh, J. P., and Kubde, K. J. (1999). Effect of integrated
nutrient management on productivity of sorghum-wheat cropping system. Ann. Plant Physiol.
13(2), 128–132.
de Jager, A., Nandwa, S. M., and Okoth, P. F. (1998). Monitoring nutrient flows and economic
performance in African farming systems (NUTMON) I. Concepts and methodologies. Agric.
Ecosyst. Environ. 71, 37– 48.
de Jager, A., Onduru, D., van Wijk, M. S., Vlaming, J., and Gachini, G. N. (2001). Assessing
sustainability of low-external-input farm management systems with the nutrient monitoring
approach: a case study in Kenya. Agric. Syst. 69, 99–118.
Jama, B., Buresh, R. J., and Place, F. M. (1998). Sesbania tree fallows on phosphorus-deficient sites:
maize yield and financial benefit. Agron. J. 90, 717–726.
Jama, B., Palm, C. A., Buresh, R. J., Niang, A., Gachengo, C., Nziguheba, G., and Amadalo, B.
(2000). Tithonia diversifolia as a green manure for soil fertility improvement in western Kenya: a
review. Agrofor. Syst. 49, 201 –221.
Janssen, B. H. (1993). Integrated nutrient management: the use of organic matter and mineral
fertilizers. In “The Role of Plant Nutrients for Sustainable Food Crop Production in sub-Saharan
Africa” (H. Van Reuler and W. Prins, Eds.), pp. 89 –106. VKP, Leidschendam.
Jeffery, R., Jeffery, P., and Lyon, A. (1989). Cattle-dung and dung-work: an aspect of the value of
women’s work in rural north India. Int. Asienforum 20(1–2), 21–42.
Jiri, O., and Waddington, S. R. (1998). Leaf prunings from two species of Tithonia raise maize grain
yields in Zimbabwe, but take a lot of labour! pp. 4–5. In “Newsletter of Soil Fertility Network”,
Harare, Zimbabwe.
Kang, B. (1993). Alley cropping: past achievements and future directions. Agrofor. Syst. 23,
141 –155.
Kang, B. T. (1997). Alley cropping — soil productivity and nutrient recycling. For. Ecol. Manag. 91,
75 –82.
Kang, B. T., Wilson, G. F., and Sipkens, L. (1981). Alley cropping maize (Zea mays L.) and Leucaena
(Leucaena leucocephala LAM) in Southern Nigeria. Plant Soil 63, 165–197.
Kang, B. T., Grimme, H., and Lawson, T. L. (1985). Alley cropping sequentially cropped maize and
cowpea with leucaena on a sandy soil in southern Nigeria. Plant Soil 85, 267–276.
Kanyama-Phiri, G., and Snapp, S. S. (1997). Developing improved organic matter technologies in
partnership with smallholder farmers in southern Malawi. In “Africa Crop Science Conference,
Pretoria, S. Africa”.
Kaya, B., Hildebrand, P. E., and Nair, P. K. R. (2000). Modelling changes in farming systems with the
adoption of improved fallows in southern Mali. Agric. Syst. 66, 51–68.
Keatinge, J. D. H., Qi, A., Wheeler, T. R., Ellis, R. H., and Summerfield, R. J. (1998). Effects of
temperature and photoperiod on phenology as a guide to the selection of annual legume cover and
green manure crops for hillside farming systems. Field Crops Res. 57, 139–152.
Keatinge, J. D. H., Qi, A., Wheeler, T. R., Musitwa, F., Franks, N. A. P., Kuribuza, D., Ellis,
R. H., and Summerfield, R. J. (1999). Potential annual-sown legumes for low-input systems
in the East African Highlands of southwestern Uganda. Mountain Res. Develop. 19(4),
345 –353.
Kiff, E., Chan, M.-K., and Jackson, D. (1997). Integrated Food Crop Systems Project, Ghana:
Development and Promotion of Improved Techniques Of Water and Soil Fertility Management
for the Sustainable Production of Crops on Land in the Humid Forest Belt. (Inception Report).
Report No. R6789 AO683. Natural Resources Institute, London.
Kiff, E., O’Connell, N., Musa, H., and Nelson, D. (1999). Integrated Food Crop Systems Project,
Ghana: Development and Promotion of Improved Techniques of Soil and Fertility Management
for the Sustainable Production of Crops on Land in the Humid Forest Belt. Report No. R6789
AO683. Natural Resources Institute, London.
Kirchhof, G., and Salako, F. K. (2000). Residual tillage and bush-fallow effects on soil properties and
maize intercropped with legumes on a tropical Alfisol. Soil Use Manag. 16(3), 183–188.
Kline, S. J. (1995). In “Conceptual Foundations for Multi-disciplinary Thinking”. Stanford University
Press, Stanford, California.
Koech, E. K., and Whitbread, R. (2000). Disease incidence and severity on beans in alleys between
Leucaena hedgerows in Kenya. Agrofor. Syst. 49(1), 85–101.
Kumara Rao, J. V. D. K., Dart, P. J., Matsumato, T., and Day, J. M. (1981). Nitrogen fixation by
pigeonpea. In “International Workshop on Pigeonpeas” (Y. I. Nene, Ed.), pp. 190– 199.
International Crops Research Institute for the Semi-Arid Tropics (ICRISAT), Patancheru, Andhra
Pradesh, India.
Kumwenda, J. D. T., Saka, A. R., Snapp, S. S., Ganunga, R. P., and Benson, T. (1997a.). Effects
of organic legume residues and inorganic fertilizer nitrogen on maize yield in Malawi.
In “Soil Fertility Research for Maize-Based Farming Systems in Malawi and Zimbabwe”
(S. R. Waddington, H. K. Murwira, J. D. T. Kumwenda, and D. Hikwa, Eds.), CIMMYT, Mutare.
Kumwenda, J. D. T., Waddington, S., Snapp, S. S., Jones, R. B., and Blackie, M. J. (1997b). Soil
fertility management in the smallholder maize-based cropping systems of eastern and southern
Africa. In “The Emerging Maize Revolution” (C. Eisher, Ed.), pp. 153– 172. Michigan State
University, Michigan.
Kwesiga, F. R., Franzel, S., Place, F., Phiri, D., and Simwanza, C. P. (1999). Sesbania sesban
improved fallows in eastern Zambia: their inception, development, and farmer enthusiasm.
Agrofor. Syst. 47, 49–66.
Lal, R. (1989a). Agroforestry systems and soil surface management of a tropical alfisol. II. Water
runoff, soil erosion and nutrient loss. Agrofor Syst. 8, 97–112.
Lal, R. (1989b). Agroforestry systems and soil surface management of a tropical alfisol. III. Changes
in soil chemical properties. Agrofor. Syst. 8, 113–132.
Lehmann, J., da Silva, J. P., Trujillo, L., Blanke, M., and Pohlan, J. (2000). Legume cover crops and
nutrient cycling in tropical fruit tree production. Acta Hortic. 531, 65–72.
Lekasi, J. K., Tanner, J. C., Kimani, S. K., and Harris, P. J. C. (1998). Manure Management in the
Kenya Highlands: Practice and Potential, p. 32. HDRA Publications, UK.
Leng, R. A. (1990). Factors affecting utilisation of poor quality forages by runinants under tropical
conditions. Nutr. Res. Rev. 3, 277 –303.
Long, A. J., and Nair, P. K. R. (1999). Trees outside forests: agro-, community, and urban forestry.
New Forests 17, 145–174.
Low, A. (1988). Farm household economics and the design and impact of biological research in
southern Africa. Agricultural Administration and Extension 29, 23–34.
Mafongoya, P. L., Barak, P., and Reed, J. D. (1997a). Carbon, nitrogen and phosphorus mineralization
of tree leaves and manure. Biol. Fert. Soils 30(4), 298 –305.
Mafongoya, P. L., Dzowela, B. H., and Nair, P. K. (1997b). Effect of multi-purpose tree, age of
cutting, and drying method on pruning quality. In “Driven by Nature; Plant Litter Quality and
Decomposition” (G. Cadisch and K. E. Giller, Eds.), pp. 167–174. CAB International,
Wallingford, UK.
Mahamoudou, S., and Meritan, R. (1998). Agricultural intensification and agroforestry in the Comoros
Islands. Cahiers Agric. 7(3), 207–212.
Mann, C. (1997). Reseeding the Green Revolution. Science 227(August), 1038–1043.
Manu, V. T., Halavatau, S., and Craswell, E. T. (1994). Agroforestry in the food production systems in
the South Pacific. In Soil Organic Matter Management for Sustainable Agriculture. “Proceedings
of a Conference held at Ubon, Thailand, 24 –26 August 1994” (R. D. B. Lefroy and G. J. Blair,
Eds.), pp. 63–68. ACIAR Proceedings No. 56.
Maroko, J., Buresh, R. J., and Smithson, P. C. (1998). Soil nitrogen availability as affected by fallow–
maize systems on two soils in Kenya. Biol. Fert. Soils 26, 229 –234.
Marschner, H. (1995). In “Mineral Nutrition of Higher Plants”. Academic Press Ltd, London.
Matthews, R. B. (2000). The role of quantitative methods in integrating biophysical and socio-
economic aspects of soil fertility management. In “Proceedings of a workshop on ‘Biophysical
and socio-economic tools for assessing soil fertility’ held at ARS-Lumle, Nepal, 17 –18 October
2000” (J. Ellis-Jones, Ed.). Silsoe Research Institute, UK, Silsoe.
Matthews, R. B., Azam-Ali, S. N., Saffell, R. A., Peacock, J. M., and Williams, J. H. (1991). Plant
growth and development in relation to the microclimate of a sorghum/groundnut intercrop. Agric.
For. Meteorol. 53, 285–301.
Matthews, R. B., Holden, S. T., Lungu, S., and Volk, J. (1992a). The potential of alley cropping in
improvement of cultivation systems in the high rainfall areas of Zambia. I. Chitemene and
fundikila. Agroforestry Systems 17, 219– 240.
Matthews R. B., and Stephens, W. (Eds.), (2002). In “Crop–Soil Simulation Models: applications in
Developing Countries.” p. 277. CAB International, Wallingford, UK.
Matthews, R. B., Volk, J., Lungu, S., Holden, S. T., and Solberg, K. (1992b). The potential of alley
cropping in improvement of cultivation systems in the high rainfall areas of Zambia. II. Maize
production. Agrofor. Syst. 17, 241–261.
Matthews, R. B., Wassmann, R., Knox, J. W., and Buendia, L. (2000). Using a crop/soil simulation
model and GIS techniques to assess methane emissions from rice fields in Asia. IV. Upscaling of
crop management scenarios to national levels. Nutrient Cycling in Agroecosystems 58, 201–217.
McDonagh, J., and Hillyer, A. (2000). Soil fertility, legumes and livelihoods in northern Namibia.
In DFID Plant Sciences Research Programme Annual Report. 1999/2000 (J. Witcombe and
D. Harris, Eds.), pp. 39–47. Centre for Arid Zone Studies, University of Bangor, Bangor, UK.
Mekonnen, K., Buresh, R. J., and Jama, B. (1997). Root and inorganic nitrogen distributions in
sesbania fallow, natural fallow, and maize fields. Plant Soil 188, 319 –327.
Mogaka, L. M. (1993). An economic evaluation of zero-grazing feeding system for high yielding cows
on smallholder farms in Kenya. In “Sustainable Feed Production and Utilisation for Smallholder
Livestock Enterprises in sub-Saharan Africa” (J. Ndikumana and W. d. Leeuw, Eds.), AFRNET,
Nairobi.
Moore, A. W. (1962). The influence of a legume on soil fertility under a grazed tropical pasture.
Empire J. Exp. Agric. 30, 239– 248.
Morris, D. R., Weaver, R. W., Smith, G. R., and Rouquette, F. M. (1990). Nitrogen transfer from
arrowleaf clover to ryegrass in field plantings. Plant and Soil 128, 293 –297.
Moser, C. M., and Barrett, C. B. (2002). The system of rice intensification in practice: explaining low
farmer adoption and high disadoption in Madagascar. Paper presented at the Water-wise Rice
Production Workshop, 8 –11 April 2002. International Rice Research Institute, Los Baños,
Philippines.
Muchagata, M. G. (1997). Forests and people: the role of forest production in frontier farming systems
in eastern Amazonia. (Development Studies Occasional Paper 36). School of Development
Studies, University of East Anglia, Norwich, UK. 79 pp.
Mugendi, D. N., Nair, P. K. R., Mugwe, J. N., MK, O. N., and Woomer, P. L. (1999). Alley cropping
of maize with Calliandra and Leucaena in the subhumid highlands of Kenya. Part 1. Soil-fertility
changes and maize yield. Agrofor. Syst. 46(1), 39–50.
Muriwara, H. K., and Kirchmann, H. (1993). Nitrogen dynamics and maize growth in a Zimbabwean
sandy soil under manure fertilization. Commun. Soil Sci. Plant Anal. 24, 2343–2359.
Murwira, K. H., Swift, M. J., and Frost, P. G. H. (1995). Manure as a key resource in sustainable
agriculture. In International Conference on Livestock and Sustainable Nutrient Cycling in Mixed
Farming Systems of Sub-Saharan Africa (J. M. Powell, S. Fernandez-Rivera, T. O. Williams, and
C. Renard, Eds.), pp. 131–148. ILCA, Addis Ababa.
Musa, A. M., Harris, D., Johansen, C., and Kumar, J. (2001). Short duration chickpea to replace fallow
after Aman rice: the role of on-farm seed priming in the high Barind Tract of Bangladesh. Exp.
Agric. 37(4), 509–521.
Muza, L. (1995). Green manuring through legume relay cropping in maize in communal areas of
Zimbabwe. In “Second South Eastern Africa Soil Fertility Network Workshop”. CIMMYT,
Kadoma, Zimbabwe.
Nair, P. (1993). An Introduction to Agroforestry. Kluwer Academic Publishers, Dordrecht,
Netherlands.
Nandwa, S. M. (1995). Synchronization of nitrogen mineralization with N uptake through maize

stover placements and N fertilization under continuous maize monocropping systems in Kenya,
p. 344. University of Exeter, Exeter, UK, Ph.D. Thesis.
Nandwa, S. M., and Bekunda, M. A. (1998). Research on nutrient flows and balances in East and
Southern Africa: state-of-the-art. Agric. Ecosyst. Environ. 71, 5–18.
Neill, S. P., and Lee, D. R. (2001). Explaining the adoption and disadoption of sustainable agriculture:
the case of cover crops in northern Honduras. Econ. Develop. Cult. Change 49(4), 793–820.
Nelson, R. A., and Cramb, R. A. (1998). Economic incentives for farmers in the Philippine uplands to
adopt hedgerow intercropping. J. Environ. Manag. 54(2), 83 –100.
Nelson, R. A., Grist, P. G., Menz, K. M., Paningbatan, E. P., and Mamicpic, M. A. (1996). Cost-benefit
analysis of hedgerow intercropping in the Philippine uplands using SCUAF (Soil Changes Under
Agroforestry). SEARCA/SEAMEO Regional Center for Graduate Study and Research in
Agriculture, College, Laguna.
Nelson, R. A., Cramb, R. A., and Mamicpic, M. A. (1998). Erosion/productivity modelling of maize
farming in the Philippine uplands. Part III: Economic analysis of alternative farming methods.
Agric. Syst. 58(2), 165– 183.
Nguyen The, D., Klinnert, C., Martius, C., Tiessen, H., and Vlek, P. L. G. (2001). Problems with and
local solutions for organic matter management in Vietnam. Nutrient Cycling in Agroecosystems
61(1–2), 89–97.
Nibbering, J. W. (1991). Crisis and resilience in upland land use in Java. In “Indonesia: Resources,
Ecology and Environment” (J. Hardjono, Ed.), pp. 104–132. Oxford University Press, New York.
Nyathi, P., and Campbell, B. M. (1993). The acquisition and use of miombo litter by small-scale
farmers in Masvingo, Zimbabwe. Agrofor. Syst. 22, 43–48.
Oko, B. F. D., Shiyam, J. O., Eneji, A. E., Honna, T., and Yamamoto, S. (2000). Yield of selected food
crops under alley-cropping with some hedgerow species in humid tropical southeastern Nigeria.
Jpn. J. Tropic. Agric. 44(3), 167–170.
Onduru, D. D., Gachini, G. N., de Jager, A., and Diop, J. M. (1999). Participatory research on compost
and liquid manure in Kenya. “Managing Africa’s Soils”, Vol. 8, p. 23. Drylands Programme,
International Institute for Environment and Development, London, UK.
Ong, C. (1994). Alley cropping — ecological pie in the sky? Agrofor. Today 6, 8–10.
Ong, C. K., Wilson, J., Deans, J. D., Mulayta, J., Raussen, T., Wajja Musukwe, N., Turner, N. C., and
Ward, P. R. (2002). Tree– crop interactions: manipulation of water use and root function. Agric.
Water Manag. 53(1–3), 171–186.
Ouédraogo, E., Mando, A., and Zombré, N. P. (2001). Use of compost to improve soil properties and
crop productivity under low input agricultural system in West Africa. Agric. Ecosyst. Environ. 84,
259–266.
Padwick, G. W. (1983). Fifty years of Experimental Agriculture in the maintenance of soil fertility in
tropical Africa: A review. Exp. Agric. 19, 293–310.
Pagiola, S., and Holden, S. (2001). Farm household intensification decisions and the environment.
In “Tradeoffs or Synergies?” (D. R. Lee and C. B. Barrett, Eds.), pp. 73–88. Agricultural Inten-
sification, Economic Development, and the Environment. CAB International, Wallingford, UK.
Palm, C., Wangari, N., and Gichuru, M. (1996). Selection of organic inputs for soil fertility
management. In “Soil Fertility Network for Maize-based Farming Systems in Selected Countries
of southern Africa. Research Results and Network Outputs in 1994 and 1995” (S. R. Waddington,
Ed.), pp. 19 –24. CIMMYT, Harare, Zimbabwe.
Palm, C. A., Myers, R. J. K., and Nandwa, S. M. (1997a). Combined use of organic and inorganic
nutrient sources for soil fertility maintenance and replenishment. In “Replenishing Soil Fertility
in Africa” (R. J. Buresh, P. A. Sanchez, and F. Calhoun, Eds.), pp. 193–217. Soil Science Society
of America, Madison, Wisconsin, USA.
Palm, C. A., Myers, R. J. K., and Nandwa, S. M. (1997b). Organic–inorganic nutrient
interactions in soil fertility replenishment. In “Replenishing Soil Fertility in Africa”
(R. J. Buresh, P. A. Sanchez, and F. Calhoun, Eds.), pp. 193–218. Soil Science Society of
America, Madison.
Palm, C. A., Gachengo, C., Delve, R., Cadisch, G., and Giller, K. E. (2001). Organic inputs for soil
fertility management in tropical agroecosystems: application of an organic resource database.
Agric. Ecosyst. Environ. 83, 27–42.
Pande, R. S. (1997). Fodder and Pasture Development in Nepal. Udatya Research and Development
Services, Kathmandu.
Paningbatan, E. P. Jr. (1990). Alley cropping for managing soil erosion in sloping lands, Transactions
of the 14th International Soil Science Society Congress, Kyoto, Japan. pp. 376 –377.
Pantanali, R. (1996). Lesotho: a note on the Machobane System. Report No. TCI Occasional Paper
Series No. 7. Food and Agricultural Organisation, Rome, Italy.
Patra, B., and Poi, S. C. (1998). Influence of intercropping on the nodulation in legumes. Environ.
Ecol. 16(2), 418–423.
Pieri, C. (1995). Long-term soil management experiments in Semi-arid Francophone Africa. In “Soil
Management Experimental Basis for Sustainability and Environmental Quality. Advances in Soil
Science” (R. Lal and B. A. Stewart, Eds.), pp. 225 –266. CRC, Lewis.
Pilbeam, C. J., Tripathi, B. P., Munankarmy, R. C., Gregory, P. J., Ellis Jones, J., Mason, T., and
Keatinge, D. (1999a). Productivity and economic benefits of integrated nutrient management in
three major cropping systems in the mid-hills of Nepal. Mountain Res. Develop. 19(4), 333–344.
Pilbeam, C. J., Tripathi, B. P., Sherchan, D. P., Gregory, P. J., and Gaunt, J. (1999b). Nitrogen
balances for the households in the mid-hills of Nepal. Agric. Ecosyst. Environ. 79, 61–72.
Place, F., and Dewees, P. (1999). Policies and incentives for the adoption of improved fallows.
Agrofor. Syst. 47, 323 –343.
Pound, B., Webb, M., Carreno, B., and Flores, A. (1999). Weed Management for Sustainable
Agriculture in Forest Margins; Investigacion Adaptiva en Margenes de Bosque. (Joint Final
Technical Report). Natural Resources Institute, London, and CIAT, Cali Bolivia.
Powell, J. M., and Williams, T. O. (1993). Livestock, nutrient cycling and sustainable agriculture in
West African Sahel. International Institute for Environment and Development, London,
Gatekeeper Series No. 37.
Pradeepmani, D. (1988). Introducing multipurpose trees on small farms in Nepal. In “Proceedings of
an international workshop held November 2–5, (1987) in Pattaya, Thailand” (D. Withington,
K. G. MacDicken, C. B. Sastry and N. R. Adams, Eds.), pp. 197–203. Winrock International
Institute for Agricultural Development, Arlington, Virginia, USA, and International Develop-
ment Research Centre, Canada.
Prasad, P. V. V., Satyanarayana, V., Murthy, V. R. K., and Boote, K. J. (2002). Maximizing yields in
rice–groundnut cropping sequence through integrated nutrient management. Field Crops Res.
75(1), 9– 21.
Pretty, J. (1995). Regenerating Agriculture: Policies and Practice for Sustainability and Self-reliance,
p. 336. Earthscan, London.
Pretty, J. (1999). Can sustainable agriculture feed Africa? New evidence on progress, processes and
impacts. Environ. Develop. Sustainability 1, 253–274.
Pretty, J., and Koohafkan, P. (2002). Land and Agriculture: a compendium of recent sustainable
development initiatives in the field of agriculture and land management, p. 76. Food and
Agriculture Organization of the United Nations, Rome.
Pretty, J. N., Morison, J. I. L., and Hine, R. E. (2003). Reducing food poverty by increasing
agricultural sustainability in developing countries. Agric. Ecosyst. Environ. 95, 217– 234.
Prudencio, C. F. (1993). Ring management of soils and crops in the West African semi-arid tropics:
the case of the mossi farming system in Burkina Faso. Agric. Ecosyst. Environ. 47, 237 –264.
Quansah, C., Drechsel, P., Yirenkyi, B. B., and Asante-Mensah, S. (2001). Farmers’ perceptions and
management of soil organic matter — a case study from West Africa. Nutrient Cycling in
Agroecosystems 61, 205–213.
Rachmat, R., Stur, W. W., and Blair, G. J. (1992). Cattle feeding systems and limitations to feed
supply in South Sulawesi, Indonesia. Agric. Syst. 39(4), 409 –419.
Ramasamy, S., ten Berge, H. F. M., and Purushothaman, S. (1997). Yield formation in rice in response
to drainage and nitrogen application. Field Crops Res. 51, 65–82.
Rao, M. R., and Mathuva, M. N. (2000). Legumes for improving maize yields and income in semi-arid
Kenya. Agric. Ecosyst. Environ. 78(2), 123–137.
Rao, M. R., Nair, P. K. R., Ong, C. K., and Latt, C. R. (1997). Biophysical interactions in tropical
agroforestry systems. Agrofor. Syst. 38(1–3), 3 –50.
Reddy, D. D., Rao, A. S., and Rupa, T. R. (2000). Effects of continuous use of cattle manure and
fertiliser phosphorus on crop yields and soil organic phosphorus on a vertisol. Bioresour.
Technol. 75, 114– 118.
Reed, J. D. (1986). Relationships among soluble phenolics, insoluble proanthocyanins and fibre in
East African browse species. J. Range Manag. 39, 5–7.
Reid, J. D. Jr. (1976). Sharecropping and agricultural uncertainty. Econ. Develop. Cult. Change 24(3),
549–576.
Reynolds, L., Di Domenico, C., Atta-Krah, A. N., and Cobbina, J. (1991). Alley farming in south-
western Nigeria: the role of farming systems research in technology development. In “Planned
Change in Farming Systems: Progress in On-farm Research” (R. Tripp, Ed.), pp. 85–108. John
Wiley & Sons Ltd, London.
Rocheleau, D., and Rocheleau, D. E. (1990). Gender complementarity and conflict in sustainable
forestry development: a multiple user approach “19th IUFRO World Congress”, pp. 432– 448.
Canadian IUFRO World Congress Organizing Committee, Montreal.
Rodell, M. G. W., Hopley, J. D. H., and Boaltwood, J. N. (1980). Effects of applied nitrogen, kraal
compact and maize stover on yields of maize growth on a poor granite soil. Zimbabwe Agric. J.
77, 229 –232.
Roose, E., and Ndayizigiye, F. (1997). Agroforestry, water and soil fertility management to fight
erosion in tropical mountains of Rwanda. Soil Technol. 11, 109 –119.
Roose, E., Kabore, V., and Guenat, C. (1999). Zai practice: a West African traditional rehabilitation
system for semiarid degraded lands, a case study in Burkina Faso. Arid Soil Res. Rehabil. 13(4),
343–355.
Ruthenberg, H. (1980). Farming Systems in the Tropics, p. 424. Clarendon Press, Oxford.
Saka, A. R., Phombeya, H. S. K., Itimu, O. A., Jones, R. B., Minae, S., and Snapp, S. S. (1995).
Agroforestry Commodity Team Annual Report. Department of Agricultural Research, Malawi
Ministry of Agriculture, Lilongwe.
Salako, F. K., Hauser, S., Babalola, O., and Tian, G. (2001). Improvement of the physical fertility of a
degraded Alfisol with planted and natural fallows under humid tropical conditions. Soil Use
Manag. 17(1), 41– 47.
Salam, M. A., Noguchi, T., and Koike, M. (2000). Understanding why farmers plant trees in
homestead agroforestry in Bangladesh. Agrofor. Syst. 50, 77–93.
Sanchez, P. A. (1999). Improved fallows come of age in the tropics. Agrofor. Syst. 47, 3–12.
Sanchez, M. D., and Rosales Mendez, M. (1999). Agroforestry systems for intensifying sustainable
livestock production in tropical Latin America. “FAO Anim Prod Health Paper”, 143, pp. 1 –13.
Sanchez, P. A., Jama, B., Niang, A., and Palm, C. (2001). Soil fertility, small-farm intensification, and
the environment in Africa. In “Tradeoffs or Synergies? Agricultural Intensification, Economic
Development, and the Environment” (D. R. Lee and C. B. Barrett, Eds.), pp. 325–344. CAB
International, Wallingford, UK.
Sanginga, N., Vanlauwe, B., and Danso, S. K. A. (1995). Management of biological N2 fixation in
alley cropping systems: estimation and contribution to N balance. Plant Soil 174(1–2), 119– 141.
Schaber, J. (1997). “FARMSIM: a dynamic model for the simulation of yields, nutrient cycling and
resource flows on Philippine small-scale farming systems”, MSc Thesis. p. 115. Universität
Osnabrück, Osnabrück.
Scoones, I., and Toulmin, C. (1998). Soil nutrient balances: what use for policy? Agric. Ecosyst.
Environ. 71, 255–267.
Seif El Din, A. G., and Raintree, J. B. (1987). Gum hashab and land tenure in western Sudan, Land,
trees and tenure. In “Proceedings of an international workshop on tenure issues in agroforestry,
Nairobi, May 27 –31, 1985”, pp. 217–224. International Council for Research in Agroforestry
(ICRAF), Nairobi, Kenya, and Land Tenure Center (LTC), Madison, Wisconsin, USA
Selvarajan, K. N., and Krishnamoorthy, S. (1990). Cost benefit analysis of potato in the Nilgiris
district, Tamil Nadu. Artha Vikas 26(1–2), 59– 66.
Senesi, N. (1989). Composted materials as organic fertilisers. Sci. Total Environ. 81/82, 521 –542.
Shepherd, K. D., and Soule, M. J. (1998). Soil fertility management in west Kenya: a dynamic
simulation of productivity, profitability and sustainability at different resource endowment levels.
Agric. Ecosyst. Environ. 71, 131–145.
Sherchan, D. P., Pilbeam, C. J., and Gregory, P. J. (1999). Response of wheat-rice and maize/millet
systems to fertilizer and manure applications in the mid-hills of Nepal. Exp. Agric. 35(1), 1–13.
Shumba, E. M., Dhilwayo, H. H., and Mukoko, O. Z. (1990). The potential of maize–cowpea
intercropping in low rainfall areas of Zimbabwe. Zimbabwe J. Agric. Res. 28, 33–38.
Siaw, D. E. K. A., Kang, B. T., and Okali, D. U. U. (1991). Alley cropping with Leucaena
leucocephala (Lam.) de Wit and Acioa barteri (Hook f.). Engl. Agrofor. Syst. 14, 219–231.
Singh, R. P., and Desai, G. M. (1991). Farmers’ fertilizer practices in the semi-arid tropics of India.
(Progress Report). Report No. 100, p. 54. Economics Group, Resource Management Program,
ICRISAT, Patancheru, Andhra Pradesh, India.
Singh, R. P., and Jodha, N. S. (1990). Determinants of intercropping in the semi-arid tropics. (Progress
Report). Report No. 95, p. 14. Economics Group, Resource Management Program, ICRISAT,
Patancheru, Andhra Pradesh, India.
Singh, K. N., and Sharma, D. P. (2001). Response of wheat to nitrogen and potassium in saline soils.
Exp. Agric. 37(3), 417 –427.
Singh, R. P., Ong, C. K., and Saharan, N. (1989). Above and below ground interactions in alley
cropping in semi-arid India. Agrofor. Syst. 9, 259–274.
Singh, K., Rao, B. R. R., Singh, C. P., Bhattacharya, A. K., Kaul, P. N., and Rajeswara Rao, B. R.
(1998). Production potential of aromatic crops in the alleys of Eucalyptus citriodora in semi-arid
tropical climate of south India. J. Medicinal Aromatic Plant Sci. 20(3), 749–752.
Slingerland, M. A., and Stork, V. E. (2000). Determinants of the practice of Zai and mulching in North
Burkina Faso. J. Sustainable Agric. 16(2), 53– 76.
Snapp, S. S., Mafongoya, P. L., and Waddington, S. (1998). Organic matter technologies for the
integrated nutrient management in smallholder cropping systems of southern Africa. Agric.
Ecosyst. Environ. 71, 185 –200.
Somda, Z. C., Powell, J. M., Fernadez-Rivera, S., and Reed, J. (1995). Feed factors affecting nutrient
excretion by ruminants and fate of nutrients when applied to soil. In “Livestock and Sustainable
Nutrient Cycling in Mixed Farming Systems of sub-Saharan Africa” (J. M. Powell, S. Fernadez-
Rivera, J. Reed, S. Williams, and C. Renard, Eds.), pp. 227–246. ILCA, Addis Abbaba.
Stoop, W. A., Uphoff, N., and Kassam, A. (2002). A review of agricultural research issues raised by
the system of rice intensification (SRI) from Madagascar: opportunities for improving farming
systems for resource-poor farmers. Agric. Syst. 71, 249 –274.
Stur, W. W., Horne, P. M., Gabunada, F. A. Jr, Phengsavanh, P., and Kerridge, P. C. (2002). Forage
options for smallholder crop–animal systems in Southeast Asia: working with farmers to find
solutions. Agric. Syst. 71(1–2), 75–98.
Styger, E., Rakotoarimanana, J. E. M., Rabevohitra, R., and Fernandes, E. C. M. (1999). Indigenous
fruit trees of Madagascar: potential components of agroforestry systems to improve human
nutrition and restore biological diversity. Agrofor. Syst. 46, 289 –310.
Sutihar, D. N. (1984). An econometric analysis of wheat production in Nepal: a case study of Saptari
district. Econ. J. Nepal 7(1), 56– 66.
Swift, M. J., Seward, P. D., Frost, P. G. H., Qureshi, J. N., and Muchena, F. N. (1994). Long-term
experiments in Africa: developing a database for sustainable land use under global change.
In “Long-term Experiments in Agricultural and Ecological Sciences.” (R. A. Leigh and A. E.
Johnson, Eds.), pp. 229 –251. CAB International, Wallingford, UK.
Swinkels, R., and Franzel, S. (1997). Adoption potential of hedgerow intercropping in maize-based
cropping systems in the highlands of western Kenya. 2. Economic and farmers’ evaluation. Exp.
Agric. 33, 211–223.
Swinkels, R. A., Franzel, S., Shepherd, K. D., Ohlsson, E., and Ndufa, J. K. (1997). The economics of
short rotation improved fallows: evidence from areas of high population density in western
Kenya. Agric. Syst. 55(1), 99–122.
Szott, L. T. (1987). “Improving the production of shifting cultivation in the Amazon Basin of Peru
through the use of leguminous vegetation”, Ph.D. Thesis. North Carolina State University,
Raleigh, North Carolina, p. 168.
Szott, L. T., and Palm, C. A. (1996). Nutrient stocks in managed and natural humid tropical fallows.
Plant Soil 186, 293–309.
Szott, L. T., Palm, C. A., and Buresh, R. J. (1999). Ecosystem fertility and fallow function in the
humid and subhumid tropics. Agrofor. Syst. 47, 163–196.
Tanner, J. C., Holden, S. J., Winugroho, M., Owen, E., and Gill, M. (1993). Feeding livestock for
compost production: a strategy for sustainable upland agriculture on Java. In “International
Conference on Livestock and Sustainable Nutrient Cycling in Mixed Farming Systems of Sub-
Saharan Africa” (J. M. Powell, S. Fernandez-Rivera, T. O. Williams and C. Renard, Eds.), pp.
115–128. Addis Ababa.
Tarawali, G., Manyong, V. M., Carsky, R. J., Vissoh, P. V., Osei-Bonsu, P., and Galiba, M. (1999).
Adoption of improved fallows in West Africa: lessons from mucuna and stylo case studies.
Agrofor. Syst. 47, 93–122.
Thomas, G. V., Rohini, I., Bopaiah, B. M., and Iyer, R. (1991). Beneficial microbes in the nutrition of
coconut. J. Plantation Crops 19(2), 127–138.
Tian, G., Hauser, S., Koutika, L. S., Ishida, F., Chianu, J. N., Tian, G., Keatinge, D., Carsky, R., and
Wendt, J. (2001). Pueraria cover crop fallow systems: benefits and applicability, sustaining soil
fertility in West Africa. In “Proceedings of a symposium sponsored by the Soil Science Society of
America and the American Society of Agronomy, 5– 9 November 2000”, pp. 137–155. Soil
Science Society of America Inc. Madison; USA, Minneapolis, USA.
van Tienhoven, N., Icaza, J., and Lagemann, J. (1982). Farming Systems in Jinotega Nicaragua.
CATIE Technical Series Technical Report No. 31. CATIE, Turrialba, Costa Rica.
Todaro, M. P. (1999). “Economic Development”, Addison-Wesley Publishing, Boston, USA. p. 783.
Turner, M. (1995). The sustainability of rangeland to cropland nutrient transfer in semi-arid west
Africa: ecological and social dimensions neglected in the debate. In “Livestock and Sustainable
Nutrient Cycling in Mixed Farming Systems of sub-Saharan Africa”. Volume II: Technical
papers. Proceedings of an International Conference held in Addis Ababa, Ethiopia, 22 –26
November (1993) (J. Powell, S. Fernandez-Rivera, T. Williams and C. Renard, Eds.), pp. 435 –
452. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia.
UNEP, (1999). “Use of Mucuna cover cropping to control Imperata cylindrica”. United Nations
Development Programme, New York, NY 10017, Accessed 14-Nov-2002 online at: http://www.
undp.org/tcdc/bestprac/agri/cases/benin15.htm.
UNPD, (2001). World Population Prospects: The 2000 Revision “Highlights”, Population Division,
Department of Economic and Social Affairs, United Nations, New York. p. 22.
Uphoff, N. T. (1999). Agroecological implications of the system of rice intensification (SRI) in
Madagascar. Environ. Develop. Sustainability 1(3–4), 297– 313.
Valverde, B. E., Merayo, A., Reeder, R., and Riches, C. R. (1999). Integrated management of itchgrass
(Rottboellia cochinchinensis) in maize in seasonally-dry Central America: facts and perspectives.
“Proceedings of International Crop Protection Conference, Brighton, UK, 15–18 November

1999”, pp. 131– 140. British Crop Protection Council, Farnham, UK.
Vandermeer, J. (1989). “The Ecology of Intercropping”. University of Cambridge, Cambridge.
Vanlauwe, B., Aihou, K., Houngnandan, P., Diels, J., Sanginga, N., Merckx, R., Jensen, E. S., and
Recous, S. (2001). Nitrogen management in ‘adequate’ input maize-based agriculture in the
derived savanna benchmark zone of Benin Republic. Plant and Soil 228(1), 61–71.
Versteeg, M. N., Amadji, F., Eteka, A., Gogan, A., and Koudokpon, V. (1998). Farmers’ adoptability
of Mucuna fallowing and agroforestry technologies in the coastal savanna of Benin. Agric. Syst.
56(3), 269 –287.
Vosti, S. A., and Witcover, J. (1996). Slash and burn agriculture — household perspectives. Agric.
Ecosyst. Environ. 58, 23– 38.
Vosti, S. A., Witcover, J., Oliveiara, S., and Faminow, M. (1998). Policy issues in agroforestry:
technology adoption and regional integration in the western Brazilian Amazon. Agrofor. Syst. 38,
195 –222.
Wakle, P. K., Khandare, V. S., Kalaskar, S. S., Karanjikar, P. N., and Daunde, A. T. (1999). Constraints
faced by farmers in adoption of banana cultivation technology. J. Soils Crops 9(2), 182 –184.
Waldie, K. J. (1990). Cattle and concrete: some aspects of social organisation among the Fula in and
around Kabala, northern Sierra Leone, Ph.D. Thesis, Manchester University, Manchester, UK.
(Available online at http://csac.anthropology.ac.uk/CSACMonog/Waldie/).
Wandera, J. L., Kiruiro, E. M., Chemitei, V. C., and Bulimu, B. O. (1993). Factors limiting the use of
fodders by smallholder dairy farmers in Kenya: experiences from on-farm, farmer-managed,
animal feeding trials. In “Sustainable Feed Production and Utilisation for Smallholder Livestock
Enterprises in Sub-Saharan Africa” (J. Ndikumana and W. de Leeuw, Eds.), AFRNET, Nairobi,
Kenya.
Webster, C., and Wilson, P. (1998). “Agriculture in the Tropics”. Blackwell Science, Oxford.
Wendt, J. W., Jones, R. B., and Itimu, O. A. (1994). An integrated approach to soil fertility
improvement in Malawi, including agroforestry. Soil fertility and climatic constraints in dryland
agriculture. In “Proceedings of ACIAR-SACCAR Workshop, Harare, Zimbabwe, 30 August–1
September 1993” (E. T. Craswell and J. Simpson Eds.), pp. 74–79. Australian Centre for
International Agricultural Research (ACIAR), Canberra, Australia.
Wheeler, T. R., Qui, A., Keatinge, J. D. H., Ellis, R. H., and Summerfield, R. J. (1999). Selecting
legume cover crops for hillside environments in Bolivia. Mountain Res. Develop. 19(4),
318 –324.
Whittome, M. P. B. (1994). “The Adoption of Alley Farming in Nigeria and Benin: the On-Farm
Experience of IITA and ILCA.” Ph.D. Thesis. University of Cambridge, Cambridge, England.
Wolf, E. C. (1986). “Beyond the green revolution: new approaches for third world agriculture”,
Worldwatch Institute, Washington DC, USA. p. 46.
Wortmann, C. S., and Kaizzi, C. K. (1998). Nutrient balances and expected effects of alternative
practices in farming systems of Uganda. Agric. Ecosyst. Environ. 71, 115– 129.
Yamoah, C., Agboola, A. A., and Wilson, G. F. (1986a). Nutrient contribution and maize performance
in alley cropping systems. Agrofor. Syst. 4, 247 –254.
Yamoah, C. F., Agboola, A. A., and Mulongoy, K. (1986b). Decomposition, nitrogen release and
weed control by prunings of selected alley-cropping shrubs. Agrofor. Syst. 4, 239 –246.
Yaseen, T. M., Kothari, S. K., Singh, U. B., Singh, K., Sattar, A., and Roy, S. K. (2001). Bio-economic
evaluation of scented rose (Rosa damascena) based intercropping systems in north Indian plains.
Journal of Medicinal and Aromatic Plant Sciences 23(2), 69–74.
Young, A. (1987). The potential of agroforestry for soil conservation and sustainable land use,
(ICRAF Reprint No. 39). In “Collected papers” (J. Kozub, Ed.), p. 16. Economic Development
Institute of the World Bank, Washington, DC, USA.
Young, A. (1989). “Agroforestry for Soil Conservation”, CAB International, Wallingford, UK. p. 284.
AMMONIA EMISSION FROM MINERAL
FERTILIZERS AND FERTILIZED CROPS
Sven G. Sommer,1 Jan K. Schjoerring2 and O.T. Denmead3
1
Department of Agricultural Engineering, Danish Institute of Agricultural Sciences,
Research Centre Bygholm, PO Box 536, DK-8700 Horsens, Denmark
2
Plant Nutrition Laboratory, Department of Agricultural Sciences, The Royal
Veterinary and Agricultural University, Thorvaldsensvej 40, DK-1871
Frederiksberg C, Denmark
3
CSIRO Land and Water, GPO Box 1666, Canberra ACT 2601, Australia
I. Introduction
II. Mineral Fertilizer Consumption
III. Ammonia Volatilization from Mineral Fertilizers
A. Production and Transport of NH3 in the Soil– fertilizer –
atmosphere Interface
B. Temporal NH3 Loss Pattern
C. Hydrolysis of Urea
D. Soil Hþ
E. Soil CEC
F. Solid Phase Processes
G. Climate and Infiltration
H. Microbial Processes (nitrification/immobilization)
IV. Ammonia Emission from Crop Foliage
A. Transport of NH3 Between Leaves and the Atmosphere
B. Magnitude of NH3 Losses
C. Physiological Processes Involved in NH3 Emission from
Crops
V. Management Strategies
A. Techniques for Reduction of NH3 Emission
B. Fertilizer Composition
C. Flooded Fields (rice paddies)
D. Injection of Anhydrous Ammonia
E. Crop Emissions as Affected by Fertilizer Application
F. Ammonia Emission from Decomposing Plant Material
G. Absorption by Crops
VI. Measurement Techniques
A. Tracer Techniques
B. Enclosures
C. Micrometerological Methods
D. Gradient Diffusion Methods
E. Eddy Correlation
F. Relaxed Eddy Accumulation or Conditional Sampling
G. Lagrangian Dispersion Models
557
Copyright q 2004 by Academic Press. All rights of reproduction in any form reserved
0065-2113/03 $35.00
558 S. G. SOMMER, J. K. SCHJOERRING AND O. T. DENMEAD
VII. Conclusions and Perspectives

Appendix A
Acknowledgment
References
A thorough understanding of the physical and chemical processes involved

in NH3 emission from inorganic N fertilizers and fertilized crops is required
if reliable and operational NH3 emission factors and decision support systems
for inorganic fertilizers are to be developed, taking into account the actual
soil properties, climatic conditions and management factors. For this reason,
the present review focuses on processes involved in NH3 volatilization from
inorganic nitrogen fertilizers and the exchange of ammonia between crop
foliage and the atmosphere.
The proportion of nitrogen lost from N fertilizers due to NH3 volatilization
may range from < 0 to .50%, depending on fertilizer type, environmental
conditions (temperature, wind speed, rain), and soil properties (calcium
content, cation exchange capacity, acidity). The risk for high NH3 losses may
be reduced by proper management strategies including, e.g., incorporation of
the fertilizer into the soil, use of acidic fertilizers on calcareous soils, use of
fertilizers with a high content of carbonate-precipitating cations, split
applications to rice paddies or application to the soil surface beneath the
crop canopy. The latter takes advantage of the relatively low wind speed
within well-developed canopies, reducing the rate of vertical NH3 transport
and increasing foliar NH3 absorption. Conversely, NH3 is emitted from
the leaves when the internal NH3 concentration is higher than that in the
ambient atmosphere as may often be the case, particularly during periods
with rapid N absorption by the roots or during senescence induced
N-remobilization from leaves. Between 1 and 4% of shoot N may be lost
in this way. q 2004 Academic Press.
I. INTRODUCTION
Agriculture is recognized as a major source of atmospheric ammonia (NH3),

contributing 50% of global NH3 emissions (Schlesinger and Hartley, 1992).
Recent inventories have shown that mineral fertilizers and plants account for
about 20% of the total emission of NH3 in Europe (ECETOC, 1994; Pain et al.,
1998; Hutchings et al., 2001) and 23% of global emission of NH3 is derived from
fertilizers and field-applied manure (Bouwman et al., 2002).
Ammonia is a chemically active gas and readily combines with NO2 3 and SO4
22
þ
in acid cloud droplets to form particulates (Asman et al., 1998). The NH4 particles
can be transported over long distances before being dry- or wet-deposited, while
AMMONIA EMISSION 559
gaseous NH3 usually is deposited much closer to the source (Asman and van
Jaarsveld, 1991). The deposited NH3 may cause acidification and eutrophication of
natural ecosystems (Schulze et al., 1989). Therefore, the United Nations has
included NH3 in the Convention on Long-range Transboundary Air Pollution
(CLRTP), and in addition, the EU Commission has set a limit — the NH3
ceiling — to the emission of NH3 from European countries (EEA, 1999).
For farmers, the loss of fertilizer nitrogen (N) due to NH3 emission may
significantly reduce N-fertilizer efficiency, contributing to a rather low overall
efficiency of applied N, i.e., , 50% in the tropics and , 70% in temperate areas
(Malhi et al., 2001). In order to prevent potential negative consequences of
gaseous NH3 losses, farmers may also apply fertilizer-N in excess of crop
requirements, which will increase the loss of N to the environment and
production costs. Although the use of fertilizers increased dramatically during the
20th century (Fig. 1), the recent trend is towards decreased use, as awareness of
the economic consequences and environmental impacts of N losses is
encouraging more efficient N utilization.
In the 1980s, emission factors were introduced to calculate NH3 emission from
European agriculture (Buijsman et al., 1987). Country-specific emission factors
were introduced in inventories in the 1990s (Misselbrook et al., 2000; Hutchings
et al., 2001). Major uncertainties are still associated with the use of NH3 emission
factors for inorganic fertilizers, because they in many cases are highly empirical
or have been derived from measurements carried out under conditions that
deviate considerably from modern management practices associated with
handling and applying fertilizers. As an example, the generally used emission
factor for urea is 15% for Europe and 25% for the tropics (see e.g., Bouwman
et al., 1997), which contrasts with the fact that NH3 emission from urea can be
completely avoided if the fertilizer is incorporated into the upper soil layers
Calcium Nitrate Ammonium
Ammonium phosphates
Developed world 35
Ammonium sulphate
Consumption, Tg N year–1
Nitrogen sollutions
Ammonium nitrate
80 Developing world
30
Million tonnes per year
Anhydrous-NH3
World
Other straight N
25
Other NP-N
60
20
NPK-N
40
Urea
15
10
20
5
0 0
1960 1970 1980 1990 2000
Year
Figure 1 Left: Global consumption of mineral fertilizers (IFA Statistics, 2002); Right:
Distribution of the fertilizer consumed in late 1990 (Bouwman et al., 1997).
(Harrison and Webb, 2001), or reduced to well below 10% if applied to a well-
established crop (Schjoerring and Mattsson, 2001).
More reliable NH3 emission estimates could potentially be derived from
mathematical models based on the physico-chemical processes controlling
NH3 emission from fertilizers and their interactions with soil, canopy and
atmospheric variables. Complex, mechanistic models of NH3 emission exist
(Avnimelech and Laher, 1977; Fleisher et al., 1987; Kirk and Nye, 1991;
Genermont and Cellier, 1997), but there are still difficulties in describing the
controlling processes and their interactions. The substantial requirement for
input data makes these models difficult to apply in decision support systems.
The present review focuses on the processes underlying NH3 volatilization
from inorganic N fertilizers. The influence of various soil and climatic parameters
on the production of NH3/NHþ 4 during dissolution of the fertilizer is described and
related to the temporal variation in NH3 losses and their accumulated values. The
existing knowledge on the ability of growing crop plants to reduce NH3 losses from
fertilizers lying on the soil surface beneath the canopy is outlined, as is the case for
the magnitude and mechanisms of NH3 emission from the crop foliage itself.
Finally, a description of measurement techniques for quantifying NH3 losses under
field conditions is included, since methodological aspects are important for
assessing loss rates reported in the literature as well as for planning new
experiments aiming at obtaining more and better knowledge on NH3 volatilization.
Taken together, the information in the review provides an integrated picture of
fertilizer-derived NH3 emissions, facilitating the development of decision support
systems intended to limit the volatile loss of fertilizer-N from fertilizers and plants.
II. MINERAL FERTILIZER CONSUMPTION
Figure 1 shows the consumption of mineral fertilizers up to the end of the 20th
century. In the 1950s, mineral fertilizers became cheap and the use of mineral-N
increased, reducing dependence on leguminous N fixation and animal manure as
sources of plant nutrients. In the developed world, the consumption of N
fertilizers increased until the late 1980s, but has declined since then due to stricter
environmental legislation and a growing number of set-aside policies in the
European Union. The consumption of fertilizers in central and Eastern Europe as
well as in the Commonwealth countries (EFMA, 1997) fell significantly during
the 1990s due to the economic crisis that emerged after the fall of the Berlin Wall.
Nitrogen fertilizer consumption has increased steadily in the developing world
since 1960, mainly due to the increased use of mineral fertilizers in Asia
(IFA Statistics, 2002).
Urea constitutes about 38% of the total consumption of nitrogenous fertilizers,
which is reported to be about 77 Tg N year21 (Bouwman et al., 1997). Urea is an
organic compound (amide) with a very high concentration of N (46%). After

application to soil, the urea is hydrolyzed to ammonium; therefore, urea is
included in the category of mineral fertilizers.
Other straight N fertilizers than urea include ammonium bicarbonate
(ABC) and ammonium chloride. Ammonium bicarbonate comprises 90% of
the other straight N fertilizers, and is the second most used N-containing
mineral fertilizer after urea (Fig. 1). It is widely used in China, where it is
produced locally by small- and medium-sized manufacturers (Andreas
Pacholsky, 2000; personal communication).
Apart from urea and ABC, the usage of straight N fertilizers, in order of
consumption, is ammonium nitrate (AN) . anhydrous ammonia (AA) .
nitrogen solutions . calcium –ammonium – nitrate (CAN). Nitrogen solutions
often consist of urea — ammonium – nitrate solutions (UAN), which contain
28– 32% N (EFMA, 1997).
Examples of multi-nutrient N-containing fertilizers are NPK, NP and
ammonium sulfate. These fertilizers are either produced by chemical reactions
(complex or compound fertilizers) or by mechanical blending of relevant
minerals (EFMA, 1997). The nitrogen content of the fertilizers may be very
variable (5 – 26%) and the content of ammonium and nitrate may vary among
fertilizers.
III. AMMONIA VOLATILIZATION FROM MINERAL

FERTILIZERS
A. PRODUCTION AND TRANSPORT OF NH 3 IN

THE SOIL – FERTILIZER – ATMOSPHERE INTERFACE
Emission of NH3 from applied fertilizers is consequent on transport of NH3

from the surface of an ammoniacal solution to the atmosphere. The rate of
emission is determined by the concentration gradient and resistance to NH3
transport between the surface and the atmosphere as controlled by atmospheric
transport processes, the chemical composition of the solution, and transform-
ations of Total Ammoniacal Nitrogen (TAN) (NH3 2 N þ NHþ 4 2 N) in the soil.
The air above the surface can be envisaged as a laminar or turbulent-free layer
close to the surface and, above this, a turbulent layer. Ammonia gas at the liquid –
air interface (x) is transported through the laminar layer by molecular diffusion
and then through the turbulent layer to the free atmosphere by turbulent diffusion.
The instantaneous rate of NH3 loss (Fg) may be given by the following equation,
as shown by Rachhpal-Singh and Nye (1986a) and van der Molen et al. (1990):
NHþ
4 $ NH3 " þH
þ
ð1Þ
Fg ¼ Kb £ ðx 2 NH3;a Þ ð2Þ
where Kb is a bulk transfer coefficient, and x and NH3,a are, respectively, the
concentration of NH3 in the air at the soil – air interface and in the free
atmosphere. The transfer coefficient is highly dependent on wind speed, but also
depends on atmospheric stability.
x is determined by the concentration of TAN and equilibrium processes in the
solution (Freney et al., 1983; Sherlock and Goh, 1985):
TAN
½NH3;L ¼ ð3Þ
1 þ 10ð0:09018þ2729:92=T 2 pHÞ
x ¼ ½NH3;L £ 101477:7=T21:69 ð4Þ

where NH3,L is the concentration of NH3 in the solution, and x is the equilibrium
concentration of NH3 in the atmosphere in immediate contact with the soil
solution.
The change in NH3,L concentration of a solution of TAN and bicarbonate at
different pH and temperatures is shown in Fig. 2. It is seen that below pH 7, the
concentration of NH3,L is low. Therefore, the increase in NH3,L with temperature
is most noticeable when pH is . 7. Soil buffers and cations and anions added with
fertilizers will affect soil pH and, hence, NH3 emissions. Furthermore, NH3
Figure 2 Effect of temperature and pH on NH3 concentrations in a liquid solution of TAN and
TIC under conditions with no emission of NH3 and CO2 gases. It is assumed that 100 kg urea is mixed
homogeneously in the top 1 cm soil layer and the soil has a water content of 30% (vol/vol); thus after
hydrolysis, TAN concentration is ca. 0.1 mol l21 and TIC is 0.05 mol l21.
emissions will be reduced through reduction of TAN as a result of NH3 loss,

infiltration of TAN or fertilizers below the soil surface by rainfall, reaction with
the cation exchange complex, immobilization, uptake by plants or oxidation by
nitrifying organisms.
Transport processes at the gaseous interface and in the air layers above the soil
surface can be modeled by use of a bulk transfer coefficient (Eq. (2)). The transfer
coefficient (Kb) is parameterized by a series of resistances that are additive
and are related to the resistances to transport by diffusion and convection.
The resistances comprise (I) an aerodynamic resistance (ra) representing the
resistance of the turbulent air layer between observation height and the
aerodynamic roughness length of the surface (van der Molen et al., 1990;
Genermont and Cellier, 1997), (II) a laminar resistance (rb) representing the
resistance to transport across the laminar boundary layer near the surface, which
is dominated by molecular diffusion, and (III) an interfacial resistance (rc)
representing the resistance to transport within the soil or plant solution layer to
the air interface (van der Molen et al., 1990).
1
Kb ¼ ð5Þ
ra þ rb þ rc
A similar set of resistances describes NH3 loss from water bodies, including rice
paddies (Denmead and Freney, 1992).
The resistance ra in the turbulent layer can be calculated from aerodynamic
theory using developments by, for example, Paulson (1970) and Monteith and
Unsworth (1990):
ln½ðz 2 dÞ=z0 2 cg ðz 2 dÞ
ra ¼ ð6Þ
kup
where z is the height of measurement within the boundary layer of the emitting
field, d is the zero plane displacement (the height at which wind speed appears
to go to zero), z0 is the roughness length of the surface, cg is a function that
accounts for the effects of atmospheric stability on gas transport and is defined
in Appendix, k is von Karman’s constant (< 0.4), and up is the friction velocity,
a measure of the flux of momentum between the atmosphere and the ground.
The friction velocity can be measured directly by eddy covariance or
estimated from the profile of horizontal wind speed u:
ku
up ¼ ð7Þ
ln½ðz 2 dÞ=z0 2 cm ðz 2 dÞ
The stability function cm is given in Appendix. The roughness length and the
zero plane displacement vary with surface characteristics. z0 and d are assumed to
be linear functions of the canopy height (h) (Monteith and Unsworth, 1990):
z0 < 0:1h: ð8Þ
d < 0:6h: ð9Þ

The resistance of the laminar boundary layer rb above the crop canopy or soil
surface can be estimated using the empirical relationship of Thom (1972):
rb ¼ 6:2u20:67
p ð10Þ
Van der Molen et al. (1990) estimate the resistance rc within the soil surface
layer as
0:5Ll
rc ¼ ð11Þ
Daq KH þ Dg
where the numerator represents the average distance over which the ammoniacal
N has to be transported to reach the soil surface, and Daq and Dg are the soil –
liquid and soil – gas diffusion coefficients for ammoniacal N. Sutton et al. (1998)
and Nemitz et al. (2001) present formulations for the resistances to NH3 transport
in plant leaves and canopies, and Denmead and Freney (1992) do the same for
NH3 transport across the viscous sublayer below the air– water interface in water
bodies.
Soil surface roughness leads to an increase in turbulence, which leads to an
increase in friction velocity (Brutsaert, 1982), and accordingly to an increase in
the exchange between the soil surface and the atmosphere. Thereby, surface NH3
gas becomes better mixed and dispersed in the near atmosphere, and NH3
emission is increased. Consequently increasing wind speed and temperature will
increase NH3 emission from applied fertilizers (Eqs. (1) –(10)), but on days with
high global radiation, increasing wind speed will reduce soil and plant surface
temperatures and thereby the emission potential of the solution. Thus, a number
of studies of NH3 emission from manure applied to fields have shown that NH3
emission is not always related to wind speed (Bussink et al., 1994; Beauchamp
et al., 1978; Sommer et al., 1997).
The effect of the internal boundary layers may be illustrated by
considering a field amended with ammoniacal fertilizers and surrounded by
untreated fields. Downwind from the leading edge, the atmospheric NH3
concentration will gradually increase due to the input of NH3 volatilized from
the soil solution, but the effect of the increase on emission rate will depend
on the surface boundary conditions. The driving force for NH3 emission is
the concentration difference between the soil interface containing TAN and
the atmosphere (Eq. (2)). If NH3 is freely available at the surface, as
following surface spreading of slurries, the surface concentration will remain
more or less unchanged and the NH3 emission will gradually decline
downwind. The dynamics of the emission will resemble those for a constant
concentration boundary condition at the surface, or a radiation boundary
condition if the wind speed is varying in time. If, however, NH3 production
is controlled by soil processes, such as the hydrolysis of urea incorporated
4.0
3.5
3.0
F/F1000 2.5
2.0
1.5
1.0
0 200 400 600 800 1000
Length of field, m
Figure 3 Ammonia emission rate (F) as a fraction of the emission rate (F1000) at 1000 m
calculated using the model developed by Philip (1959).
into the soil, the emission rate will be virtually unaffected by plot size or fetch.
Ammonia concentrations in the soil air will simply increase in response to the
increase in the concentration of the air above and the dynamics will resemble
those for a constant flux condition at the surface. In the case of slurry spreading,
increasing the area to which fertilizer is applied at a constant application rate will
be expected to reduce the emission rate of NH3 as a percentage of the amount of
TAN applied (Genermont and Cellier, 1997).
The initial NH3 emission rates after fertilizer application will tend to be larger
from small plots than big fields, and to decline faster because TAN has been
reduced due to the larger emission during the first few hours. The influence of
field size on NH3 loss has been demonstrated by results of simulations for smaller
plots (0 – 25 m) calculated by Vlek and Craswell (1981) for a flooded soil using a
diffusion model (Bouwmeester and Vlek, 1981). Figure 3 shows the dependence
of emission rate on plot size, calculated from Philip’s (1959) analysis for a
constant concentration boundary condition at the surface. Figure 3 indicates that
the average emission rate from a small treated plot with a fetch of only 1 m would
be around 2.5 times that from a large field with a fetch of 1000 m, and even for a
plot with a fetch of 100 m, the emission rate would still be about 30% higher than
that from the large field.
B. TEMPORAL NH 3 LOSS PATTERN
NH3 emission from a soil applied fertilizer is a function of dissolved NH3 in

equilibrium with the atmosphere either directly or through soil pores. According to
the dissociation of NHþ4 and Henry’s constant (Eqs. (3) and (4)) the emission is
related to the concentration of TAN and Hþ(pH ¼ 2 log(Hþ)). In consequence,
18
Cumulative NH3 volatilization (% of N)

16 DAP
14 Urea
12
10
8
6
4
2
0
0 2 4 6 8 10 12
Days from application
Figure 4 Ammonia volatilization from urea and diammonium phosphate (DAP) applied to a
sandy loam measured with a wind tunnel (after Sommer and Jensen, 1994).
doubling the TAN concentration will double NH3 emission, whereas doubling the
Hþ concentration will halve the emission.
Typical patterns of NH3 volatilization over time for urea and diammonium
phosphate are shown in Fig. 4. Since urea fertilizer contains no TAN, no NH3 is
emitted immediately after urea application to the soil. Subsequently, the urea in
contact with the soil will absorb water, and urea will hydrolyze, producing TAN
and bicarbonate (HCO2 3 ). The rate of hydrolysis is related to the amount of soil
water absorbed and temperature; therefore, the initial lag phase with no NH3
volatilization may vary. The NH3 loss rate declines after 5– 10 days due to a
reduction in NH3 concentration as TAN is volatilized, dissolved in increasing
volumes of soil water, leached by rain into the soil, absorbed to soil, or
transformed to nitrate (Black et al., 1985; Haynes and Williams, 1992). The rate
of NH3 volatilization from urea has been described by a logistic equation by
Stevens et al. (1989), showing that maximum loss rates may occur within one
to 10 days after application, depending on soils and environmental conditions.
A sigmoidal model was used by Sommer and Ersbøll (1996) to relate cumulated
loss of NH3 from urea to the time from application, showing that for loamy soils
half the total loss may be reached 2 –7 days after urea application (Table I).
The pattern of NH3 volatilization from urea is affected by whether urea is
applied in pellets, in solution, or crushed to a fine powder, because dissolution of
and diffusion of urea into the soil solution will be different. From urea applied as
a powder or in solution, the emission will occur earlier than after application of
prilled or granulated urea. This delay in emission from prilled or granular urea is
related to hydrolysis of the applied urea (see Section III.C on hydrolysis of urea).
Other ammoniacal fertilizers are composed of NHþ 4 salts of phosphate, sulfate,
or nitrate; these salts are readily dissolved in the soil water after application of
Table I
Models for Predicting NH3 Volatilization from Surface-Applied Ammoniacal Fertilizer
Fertilizer Equation Definition of variables References

2ct
Urea F ¼ a=ð1 þ be Þ 2 að1 þ bÞ F is the NH3 volatilization (% of urea-N or % of TAN), Stevens et al.
Nmax ¼ ab=ð1 þ bÞ t is the time in days from application of urea and a, b, and (1989)
tmax ¼ lnðbÞ=c for b . 0 c are time-related parameters
AMMONIA EMISSION
Nmax is total cumulated NH3 volatilization (% of urea-N)
tmax ¼ 0 for b # 0
tmax is the time in days where the loss rate is at maximum
Urea Sommer and Easlòll

F ¼ Nmax ð1 2 e2bt Þc
(1996)
NHþ
4 mineral fertilizer F ¼ Nmax ð1 2 e2bt Þ Sommer and Easlòll
(1996)
Urea F ¼ 3:67 þ 22:28ðDMÞ DM ¼ dry matter per unit surface (25 £ 65 cm2) Hoult and McGarity
(1987)
Urea F ¼ 12:1 2 0:97CEC þ 0:02CEC 2 CEC ¼ cation exchange capacity (cmol kg21), McGarry et al.
SMC ¼ soil moisture content (g g21) and (1987)
2 0:011SMC þ 0:049T
T ¼ temperature (8C)
Urea F ¼ 20:84 2 0:0452TA TA is total acidity (mmol kg21) Stevens et al.
(1989)
Data were taken from Stevens et al. (1989), Sommer and Ersbøll (1996), Hoult and McGarity (1987), McGarry et al. (1987), Stevens et al. (1989). In the study of
McGarry et al. (1987), a closed chamber was used and emission was measured by absorbing NH3 in a solution of acid, in the other studies emission was measured
with dynamic chambers in the laboratory.
567
the fertilizer, and NH3 volatilization will start within a short time after their
application. The NH3 volatilization rate will decline due to the same processes as
mentioned previously for the decline in volatilization rate after application of
urea. In consequence, the NH3 volatilization pattern shows no lag phase after
application and the cumulated loss may therefore be described by a simple
exponential equation (Table I).
C. HYDROLYSIS OF UREA
Urea is the diamide of carbonic acid and as such an organic fertilizer. Like
other amides, urea in solution is hydrolyzed to NH3 and bicarbonate (HCO2 3 ).
No NH3 is lost from urea that has not been transformed and emission of NH3
from urea applied in the field is, therefore, closely related to hydrolysis of urea
by the enzyme urease:
COðNH3 Þ2 þ 2H2 O $ NH3 þ NHþ 2
4 þ HCO3 ð13Þ
This reaction produces a mixture of NH3, ammonium and bicarbonate (NHþ
4)
(HCO2 3 ). The rate of hydrolysis of urea is related to urease activity, the
availability of water (Eq. (13)), pH, and temperature (Bremner and Douglas,
1971; Fig. 5).
Very dry soils with water content below permanent wilting point will delay or
inhibit urea hydrolysis; thus at soil water potential , 21.5 MPa, hydrolysis has
been shown to be insignificant (Al-Kanani et al., 1991). Increasing water content
of the soil will increase the rate of hydrolysis (Vlek and Carter, 1983; McInnes
et al., 1986a; Reynolds and Wolf, 1987). Air humidity influences hydrolysis of
urea because prilled and granular urea are hygroscopic and will absorb water at
Idasoil, pH 7.5
Activity, micro mol g–1 soil h –1
12 40
Marshallsoil, pH 6.8
10 Edinasoil, pH 6.1
30
8
6 20
4
10 Soil pH 4.9
2 Soil pH 7.8
Soil pH 6.4
0 0
7 8 9 10 0 2 4 6 8 10
pH
Urea, mol N l–1 (soil solution)
Figure 5 Rate of urea hydrolysis related to soil pH (left: Tabatabai and Bremner, 1972) and
concentration of urea (right: Nye, 1992).
high air humidity. In consequence, NH3 emission may be significant from urea
applied to a dry soil if air humidity is high (Bouwmeester et al., 1985; Reynolds
and Wolf, 1987; Black et al., 1987a). Hydrolysis of urea may be delayed in soils
low in pH or after the addition of acidifying anions mixed with the urea (Ouyang
et al., 1998); optimum pH for soil urease activity has been reported to range from
pH 8 to 9. Figure 5 shows that urease activity varies between soils and that for
some soils the rate of hydrolysis is not much affected by pH. This may explain
why Rachhpal-Singh and Nye (1986a) found that the rate of hydrolysis only
varied insignificantly when soil pH was in the range from 6.6 to 8.6. In soils
where water content is not limiting, urease activity increases with temperature
(Vlek and Carter, 1983). Urease activity also increases with increasing urea
concentration (Tabatabai and Bremner, 1972, Fig. 5).
The hydrolysis of urea applied in pellets is slow compared with that applied in
solution or as a fine powder, due to the slow diffusion of urea into the soil where
urease is abundant (Vlek and Carter, 1983). Therefore, hydrolysis may reach a
maximum within 3 –5 days for pellets and 24 –48 h for urea in solution or as a
fine powder (Lyster et al., 1980; Black et al., 1987a; Thomas et al., 1988;
Whitehead and Raistrick, 1990). Recous et al. (1988) showed that urea applied in
the field was hydrolyzed with a half-life of 1.9 days at 3.58C; on the other hand,
urea well mixed with the soil was hydrolyzed with a half life of 22, 15 and 6 h
when incubated in the laboratory at 4, 10 and 208C. This indicates that hydrolysis
rate increases significantly with temperature, and that urea mixed with the soil is
hydrolyzed rapidly.
At urea concentrations below 1 mol N per liter (soil) the rate of hydrolysis
initially increases linearly with concentration, then reaches an optimum value,
eventually decreasing at higher urea concentrations (Fig. 5). This may be
explained as substrate inhibition, which can be described by a Michaelis –Menten
equation (Nye, 1992). For the purpose of predicting hydrolysis in soils low in
urea and with pH in the range from 6.5 to 8.4, urease activity related to pH has
been described using the Michaelis– Menten equation at substrate concentrations
up to 0.2 mol urea-N per liter of soil followed by a straight line at urea
concentrations above 0.2 mol N per liter (Rachhpal-Singh and Nye, 1986b). It is
important to realize that after application of urea fertilizer in the field, the
maximum concentration of urea may be 10 mol N per liter (soil) near the
fertilizer granules (Nye, 1992), and that urea hydrolysis has been observed to
occur at concentrations below 1 mol N per liter (soil) in most of the studies cited;
therefore, the algorithms given cannot predict hydrolysis near the applied
granules.
Hydrolysis of urea applied to flooded paddy soils for rice production may be
affected by redox potential as well as pH. Thus at low redox potential (anoxic),
half-time of the hydrolysis (t1/2) has been shown to be 11.2 –24.8 h compared
with 7.4 –13.8 h for oxidized suspensions of acid soils (Lindau et al., 1989). In a
suspension of calcareous soil, hydrolysis rate is not affected significantly by
differences in redox potential (Lindau et al., 1989); however, the rate of

hydrolysis in calcareous suspension was greater (t1/2 of 1.7 –1.9 h) than in the
suspension of acid soil.
The effect of organic and inorganic compounds added to urea with the
function of delaying hydrolysis of urea and thereby reducing NH3 emission is
discussed in Section V on management strategies.
D. SOIL H1
One of the most important rate controlling factors of NH3 emission from
applied fertilizers is the Hþ concentration. At pH values higher than 7, the
concentration of NH3,L (see Eqs. (3) and (4)) is at a level that may produce
significant emissions from the applied fertilizer (Fig. 2). The fertilizers may be
grouped as acidic and alkaline with, respectively, a low and high potential for
NH3 emission (Table II). Soil alkalinity will also interact with fertilizers and
affect the concentration of Hþ. Furthermore, the fertilizer application will create
“hot spots,” and soil Hþ concentration will typically show a trend from the
fertilizer granule microsite to surrounding unaffected soil (Rachhpal-Singh and
Nye 1986a,b).
After application of ammonium salts (phosphate, nitrate, sulfate, etc.), the
emission of NH3 will produce acids of the anions (Eq. (14)), as shown for
ammonium sulphate (AS) (AS ¼ NH4HSO4) in the following example:
NH4 HSO4 þ H2 O $ NH3 " þHþ þ HSO2
4 ð14Þ
Each mole of emitted NH3 will increase the concentration of Hþ by one mole.
Therefore in acidic soils or poorly buffered soils, the pH will decline due to
the NH3 emission and the NH3 emission will gradually decline, because little
NH3,L will be available for volatilization, consequently, the total loss of TAN
will be low.
In calcareous soils, emission will be related to the solubility of the anion (Fenn
and Hossner, 1985; Du Preez and Burger, 1988). The solubility of nitrate (NO2 3 ),
Table II
Fertilizers Grouped in Relation to the Acidity or Alkalinity
They Produce When Dissolved in Soil Water
Acid Moderate acid Alkaline
(NH4)2SO4 (NH4)H2PO4 (NH4)HCO3

NH4NO3 (NH4)2CO3
NH4Cl NH3
(NH4)2HPO4 (NH3)2CO
for example, is high compared to SO22 22

4 and hydrogenphosphate (HPO4 ), which
2þ
may precipitate with Ca (Fenn and Hossner, 1985). Following precipitation of
the anions HPO22 4 and SO22 22
4 , carbonate (CO3 ) will dissolve and this will
increase the pH of the soil solution (Larsen and Gunary, 1962; Fenn et al., 1981).
The pH in calcareous soils will be between 7 and 8 and most of the total inorganic
carbonate (Total inorganic carbon (TIC) ¼ HCO2 3 þ CO3 þ CO2) will be in
22
2
the form of HCO3 ; thus the reaction in the soil will be as follows:
ðNH4 Þ2 X þ CaCO3 $ NH3 " þ NHþ 2
4 þ HCO3 þ CaX # ð15Þ
ðNHþ 2
4 Þ þ HCO3 $ NH3 " þCO2 " þH2 O ð16Þ
ðNH4 Þ2 X þ CaCO3 $ 2NH3 " þCO2 " þCaX # þH2 O ð17Þ
As mentioned above, HCO2 3 is an alkaline anion and the precipitation of the
acid anions (X ¼ HPO22
4 or SO 22
4 ) and dissolution of CaCO3 will increase pH,
thereby increasing the potential for NH3 emission. It is, therefore, a general rule
that emission is lower after application of fertilizers containing soluble anions
than from fertilizers containing anions that may precipitate.
Hydrolysis of urea will produce a mixture of NH3, NHþ 2
4 , HCO3 and CO3
2
22
and this may increase pH, because NH3 and CO3 are conjugate bases
(pKa ¼ 9.48 for NH3/NHþ 2
4 and pKa ¼ 10.4 for HCO3 /CO3 ). Application of
22
ammonium bicarbonate (ABC ¼ (NH4)HCO3) or ammonium carbonate

(AC ¼ (NH4)2CO3)) may also increase soil pH due to the alkaline properties
of NH3 and CO22 3 . In soil amended with urea or ABC, the pH in the microsites
affected by the fertilizers will be . 8; at this pH a high proportion of TAN will
be in the NH3 form and most of the TIC will be in the form of HCO21 3 .
In consequence, each mole of emitted NH3 will increase the concentration of Hþ
by one mole (Eq. (1)), and emission of CO2 will reduce the Hþ concentration by
one mole (Eqs. (18), (19), Fig. 6).
þ 2
3 þ H3 O $ HCO3 þ H2 O
CO22 ð18Þ
HCO2 þ
3 þ H3 O $ CO2 " þH2 O ð19Þ
Due to the combined NH3 and CO2 emission from these fertilizers, the NH3
emission may be very high after application of urea, ABC, and AB to soils
(Roelcke et al., 2002).
Alkaline fertilizers such as ABC, urea, and diammoniumphosphate (DAP)
generally cause a much higher NH3 emission than neutral or acidic fertilizers
such as AS, CAN, or monoammonium phosphate (MAP), because pH is a
dominant factor controlling NH3 emission. Inorganic ammonium compounds
containing anions that are not producing precipitating calcium salts (NO2 2
3 , Cl )
will reduce pH and the emission will be significantly lower than the emission
from fertilizers precipitating with Ca ions (Fenn et al., 1981). The addition of, for
example, calcium chloride, potassium chloride, sulfate, or triple superphosphate
10
pH of soil solution
9
5
0.025
CO2: NH3 =1:1
0.020 CO2: NH3 =2:1
NH3, mol ltr.–1
CO2: NH3 =0:1

0.015
0.010
0.005
0.000
0.00 0.01 0.02 0.03
NH3 emission, mol l –1
Figure 6 Effect of emission of NH3 and CO2 from a liquid solution of TAN (NH3 þ NHþ 4 ), TIC
(CO2 þ HCO2 2
3 þ CO 3 ), and Cl . The initial concentrations are TAN 0.1 mol l
22 21
, TIC
0.05 mol l21, and Cl2 0.06 mol l21, and pH is 6.85. It is assumed that either equal amounts (mol)
of TIC and TAN (CO2:NH3 ¼ 1:1), 2 units of TIC to 1 unit of TAN (CO2:NH3 ¼ 2:1) or 1 unit of
TAN (CO2:NH3 ¼ 0:1) is emitted.
reduces pH and in consequence will reduce NH3 emissions (Sloan and Anderson,
1995; Christianson et al., 1995; Fan et al., 1996; Ouyang et al., 1998; Goos et al.,
1999). The typical pattern of accumulated NH3 emission from fertilizers
will generally show NH3 emission rates in the following order ABC . urea .
DAP . AS . CAN ¼ MAP due to the alkaline properties of the fertilizers and
interaction of precipitation of anions (Du Preez and Burger, 1988; Whitehead and
Raistrick, 1990; Sommer and Jensen, 1994; Ouyang et al., 1998; He et al., 1999;
Zia et al., 1999; Roelcke et al., 2002).
From soils with limited buffer capacity, the NH3 volatilization will
increase at increasing application rates of ABC and urea, because pH will be
high for a long period (Black et al., 1985; Fenn et al., 1987). However,
2
emission of NH3 from acid fertilizers (SO22 32
4 , PO4 , Cl etc.) will decrease at
increasing concentrations of TAN in fertilized soils with limited buffer
capacity (Avnimelech and Laher, 1977), because NH3 emission will
immediately cause a reduction in pH locally (Eq. (1)). From soils high in
buffer capacity, soil pH is of importance irrespective of the concentration of

TAN (Avnimelech and Laher, 1977).
NH3 volatilization from granulated fertilizers and AA has been correlated
with the total acidity and titratable acidity of soils (Izaurralde et al., 1987; Stevens
et al., 1989; Sommer and Ersbøll, 1996). On non-calcareous soils, the accumulated
NH3 losses (Nmax) from urea and CAN are positively related to soil pH and
inversely to total acidity (Fig. 7), i.e., NH3 losses are closely related to both soil
pH and Hþ buffering capacity of the soil (Avnimelech and Laher, 1977; Ferguson
et al., 1984; Stevens et al., 1989). Therefore, for the purpose of predicting losses
40
F=6.12pH-19.7, R2=0.70
Cumulated NH3 volatilization
30
(% of N)
20
10
0
3 4 5 6 7 8 9
pH(H2O)
30
25
Cumulated NH3 emission
F= –0.05TA+22.4 R2=0.52
20
(% of N)
15
10
0
0 100 200 300 400 500
TA, mmol kg–1
Figure 7 Cumulative NH3 volatilization related to soil pH and total acidity (TA) of the soil
(adapted from Stevens et al., 1989; Sommer and Ersbøll, 1996; Whitehead and Raistrick, 1990).
60 F=0.245exp(0.5858 pHmax)R2 = 0.607
50
Cumulated NH3 volatiliztion

40
(% of N)
30
20
10
3 4 5 6 7 8 9 10
pH max
Figure 8 Cumulative NH3 volatilization from urea and mineral fertilizers related to maximum
pH after soil amendment (adapted from Whitehead and Raistrick, 1990; Sommer and Ersbøll, 1996).
from urea applied to the field, the relation between Nmax and soil pH may be used,
as shown in Fig. 7; this relationship explains about 70% of the variation in three
different studies. The volatilization from ammoniacal mineral fertilizers may also
be predicted on the basis of the relation between NH3 volatilization and pHmax
(Fig. 8), but the pHmax of the soil surface after fertilizer application should be
measured or estimated. More complicated mechanistic models presented by
Avnimelech and Laher (1977); Rachhpal-Singh and Nye (1986a,b); Fleisher
et al. (1987); and Kirk and Nye (1991) may be used for predicting the risk of
volatilization when applying different types of fertilizers to the field; the problem
with these models is that the input data for the models may often be missing.
E. SOIL CEC
TAN in soil will be partitioned between the three phases — liquid, solid, and
gas phase. The NHþ 4 component in solution will be in equilibrium with NH4
þ
on the solid phase exchange site (Fleisher et al., 1987). At pH , 8, more than 95%
of the TAN (Fig. 2) will be of NHþ4 form and can be exchanged with exchangeable
cations (ex-C). Agricultural soils contain the divalent cations Ca2þ and Mg2þ
with a high affinity for adsorption; the exchange of NHþ 4 on exchange sites
therefore can be defined by use of the activity ratio law (Russel, 1977):
NHþ
4 þ ex 2 C $ ex 2 NH4 þ C
þ
ð20Þ
NHþ Ex 2 NH4
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
4
¼K ð21Þ
Ca þ Mg
2þ 2þ Ex 2 Ca; Mg
It can be seen from Eqs. (20) and (21) that increase in cation exchange capacity
(CEC) will increase NHþ 4 adsorbed to the soil (Avnimelech and Laher, 1977;
McGarry et al., 1987; Whitehead and Raistrick, 1993). At low cation exchange
capacity, adsorption of NHþ 4 will not affect NH3 emission; at CEC ca.
250 mmol kg21 the effect of absorption tends to be insignificant (O’Toole et al.,
1985), whereas when CEC is above this level, NH3 volatilization is reduced
significantly. Further, NHþ 2þ
4 does not exchange easily with Ca , thus high Ca
2þ
concentrations in the soil may reduce the effect of a high soil CEC. This will be the
case in calcareous soils, where high concentrations of exchangeable Ca2þ will
reduce the fraction of NHþ 4 absorbed.
The consequence of the ratio law is that increasing soil water content due to
rain will change the equilibrium and the divalent cations in solution will be
exchanged with NHþ 4 (Chung and Zasoski, 1994). Conversely, if the solution
is concentrated by water being removed due to drying, NHþ 4 will exchange
with divalent cations on the CEC. Thus, during a drying event after fertilizer
application, the concentration of NHþ 4 in solution will not increase linearly
with evaporation of water, and in consequence the exchange process will
reduce the rate of NH3 emission expressed relative to initial TAN content per
time unit and increase the duration of the period with significant emission
rates; this effect is most pronounced in soil high in CEC, as can be deduced
from the work of Fleisher et al. (1987). This retention to CEC during drying
events may, in addition to the solid phase theory (see below), contribute to
an explanation of why NH3 emission may be low from a soil that has been
dried (Fenn and Kissel, 1976). Furthermore, Fenn and Kissel (1976) have
shown that in very dry soils there is a physical adsorption of gaseous NH3
to soil.
F. SOLID PHASE PROCESSES
Ammonia emission from DAP applied to timed soils tend to be lower than the
emission from AS applied to soils high in calcium (Sommer and Ersbøll, 1996);
the lower NH3 emission from DAP in Ca-rich soil has been ascribed to
precipitation of calcium ammonium phosphate or magnesium ammonium
phosphate (struvite) (Larsen and Gunary, 1962; Whitehead and Raistrick, 1990).
The conditional stability constants for the solid phases of struvite may be
calculated according to principles described in detail by Ringbom (1963) and
Stumm and Morgan (1981); the reaction is as follows:
Mg2þ þ NHþ
4 þ PO4 $ MgNH4 PO4ðsÞ
32
ð22Þ
Thus, in the microsites with fertilizers, precipitation of struvite is mainly

controlled by Mg2þ, phosphate, and pH, as NHþ 4 is present in large amounts.
Struvite precipitates and is stabile at pH . 6, and will consequently be present as

stable solid phases above this pH level. Thus, struvite can be formed after
addition of urea and ABC to most soils and in some soils after addition of acid
fertilizers; the amount of struvite precipitation will, as mentioned above, be
related to Mg2þ, phosphate, and pH. The precipitation of solid phases of
ammonium will increase during a drying event, due to increasing concentrations
of the ions in solution.
Salts with hydrated water may in absence of “free” water, i.e., in a dry soil,
donate a proton to NH3, transforming it into NHþ 4 (Evangelou, 1990). The loss of
a proton produces a negative charge on the salt crystal, which facilitates the
absorption of NHþ 4 . Thus besides the precipitation reactions proposed by Fenn
et al. (1981), this mechanism may explain the reduction in NH3 emission from
surface-applied ammoniacal fertilizers applied together with Ca or Mg salts.
G. CLIMATE AND INFILTRATION
The effect of climate on NH3 volatilization from urea differs significantly from
mineral fertilizers, because urea has to be hydrolyzed before volatilization will
start, and hydrolysis is related to water availability and temperature. Thus after
application of urea to a dry soil there will be no hydrolysis of the urea and no NH3
volatilization. A rain event immediately after surface application will dissolve the
urea and, as urea movement lags only slightly behind the waterfront, the urea will
infiltrate the soil (Fenn and Miyamoto, 1979). In the soil, urea is transformed to
TAN, which will be absorbed by soil colloids; therefore, TAN is not as easily
transported as urea (Black et al., 1987a). The effect of rain may therefore be
variable and related to soil humidity and air dryness prior to the rain and also to
the amount of water added during the rain event.
More NH3 volatilizes after urea application on a wet soil than on a dry soil,
because the humidity will initiate urea hydrolysis (Fenn and Miyamoto, 1979).
An example of the interaction between application of urea and climate is shown
in Fig. 9; in late March 1993 with no rain after application, ca. 9% of the N was
lost, while only about 2% was lost in 1994 with a rain event of 13 mm
immediately after application (Schjoerring and Mattsson, 2001). The 13 mm of
rain in 1994 transported urea into the soil. The second round of urea application
in late April resulted in a loss of 7 –8% of the applied N in both years (Fig. 9).
However, due to very dry conditions in 1993, the NH3 emission was delayed for
more than 2 weeks until a light shower was received mid-May, accelerating
the dissolution of the fertilizer granules and the hydrolysis of the urea.
The total amount of NH3-N lost was 13 and 9 kg N ha21 in 1993 and 1994,
respectively.
The effect of convective and diffusive transport of urea in the soil is important
when predicting NH3 emission. Diffusion may reflect changes in TAN
NH3 emission, kg NH3-Nha–1 season–1

14 Winter wheat applied urea 14
1993
12 1994 12
10 10
8 8
6 6
4 4
2 2
0 0
rch ril y
Ma Jun
e
Jul
y
gus
t
Ma Ap Au
Figure 9 Cumulated NH3 emissions ^ SE from urea applied to growing winter wheat crops.
Vertical arrows denote time of urea application. The reason for the low loss in 1994 was that 13 mm
of rain fell immediately after the urea was applied so that the urea was transported into the soil. Very
dry conditions in 1993 delayed the NH3 emission for more than two weeks until a light shower
was received mid-May accelerating the dissolution of the fertiliser granules and the hydrolysis of
the urea.
concentration and pH around the fertilizer granules and convection the gross
transport. Complex models predicting diffusive and convective transport of
incorporated urea have been developed by Rachhpal-Singh and Nye (1986a,b)
and by Kirk and Nye (1991). These models indicate that convective transport into
the soil as affected by rain and transport to the soil surface during drying
conditions are important for the prediction of the magnitude of NH 3
volatilization. Less than 10 mm of rain has little effect on the magnitude of
volatilization (McInnes et al., 1986a; Ryden et al., 1987), and at rainfall greater
than 10 –16 mm, losses are reduced if urea remains in a non-hydrolyzed form
(Black et al., 1987b) and no loss is determined after 20 –25 mm of rain (Fenn and
Miyamoto, 1979; Bouwmeester et al., 1985). For ammoniacal fertilizers, one
may assume that about 20 mm of rain is sufficient to reduce NH3 volatilization
significantly.
Furthermore, transport of TAN into the soil may also be affected by soil
humidity. Model simulations of Nye (1992) showed that emissions tended to
increase when the moisture level was at about field capacity and also when soil
was drier. At high soil water content, diffusion of TAN in the liquid phase
increases with increasing soil water, and in dry soil diffusion of NH3 in the vapor
phase increases with reduction in water content compensating for the reduced
TAN diffusion in the liquid phase.
The rate of NH3 volatilization from urea will be affected by temperature
because both hydrolysis rate and NH3 transfer from the liquid to the atmosphere
increase with increasing temperature (Black et al., 1985; McGarry et al., 1987).
Thus, the lag phase will be shorter and the initial loss rates higher at high
soil temperatures. Total losses may not be affected by changes in temp-
erature, because volatilization will continue for a longer period at low than at
high temperature (Fenn and Kissel, 1974; Harper et al., 1983); consequently
total loss may be related to the potential of losses given by soil and other
variables (McGarry et al., 1987). A low temperature in combination with rain
may reduce losses significantly, especially because hydrolyses is slow and urea
may infiltrate the soil during showers (Fenn and Miyamoto, 1979). Thus, Harper
et al. (1983) concluded that the amount and distribution of rain after urea
application appeared to control the total NH3 volatilization from urea
application.
Ammonium phosphate or sulfate, which may precipitate with calcium, has a
different relation to temperature than AN, which will not precipitate (Fenn and
Kissel, 1973). The precipitate-forming fertilizers showed no differences in
volatilization due to change in temperature at high (. 275 kg NH4-N ha21) and
low (, 66 kg NH4-N ha21) application rates. Volatilization rates increased from
26 to 45% of TAN with a temperature increase from 12 to 328C, when the
precipitate-forming fertilizers were applied to a calcareous soil at application
rates of 110 kg NH4-N ha21. Volatilization from AN increased from 14 to 45%
of applied TAN with a temperature increase from 12 to 328C and was not affected
by changes in application rates (Fenn and Kissel, 1974). These patterns may be
related to interaction of soil buffer capacity and concentrations of TAN as
mentioned above; i.e., in soils at high buffer capacities and low NHþ 4
concentrations, the soil buffer will control emission, and at high NHþ 4
concentration, a fraction of the anions applied will be in solution and control
emission of NH3.
Wind will affect losses of NH3 from fertilizers; if humidity is high enough
to keep the salts in solution, volatilization is expected to increase with wind
(Eqs. (1) – (10)). In the few field studies using non-interfering micrometeoro-
logical techniques it has, however, been shown that wind cannot be identified as
the most important variable and wind has only been shown to affect the
volatilization from urea significantly during short periods (Harper et al. (1983), or
the effect of wind speed has not been significant during the study (McInnes et al.,
1986a). These experiments may back up the assumption that the interaction
between wind and soil temperature may confound the effect of wind.
H. MICROBIAL PROCESSES (NITRIFICATION/IMMOBILIZATION)
Emission of NH3 from TAN supplied to the soil in form of urea or ammonium
salts will compete with depletion of the TAN by microorganisms through
immobilization to soil organic N or nitrification to NO2 3 (Malhi and McGill,
1982; Recous et al., 1992). The rate of TAN transformation by microorganisms
depends on a range of factors, including the population density of microorgan-

isms, soil temperature, water and oxygen concentrations of the soil and inhibition
due to high NH3 concentrations. This means that it is difficult to generalize the
contribution of microbial activity to the emission pattern of NH3, but the
following general assessment is based on the initial period of 0– 14 days after
application of ammoniacal fertilizers, i.e., in the period with significant NH3
emission from the fertilizer-treated soil.
Microbial transformation of TAN to organic N (assimilation/immobilization)
takes place immediately after TAN is dissolved in soil solution. Rate of
assimilation is related to availability of carbon in the soil rhizosphere, TAN
concentration, soil water content and temperature (Recous et al., 1988; 1999).
Few studies have quantified the relationship between environment and
immobilization; however, it is assumed that immobilization is less responsive
to temperature than, for example, mineralization (Murphy et al., 2003).
Immobilization has been reported to account for 5% of the TAN applied
within the first 2– 3 days after spreading and 8– 10% after 14 days at an average
soil temperature of 3.58C (Recous et al., 1988) and 3.5% during the initial 7 days
at 238C (Sørensen, 2001). In humid sandy tropical soils, the rate of microbial
immobilization was 12% of TAN applied during 50 days of incubation of urea or
(NH4)2SO4 (Atwell et al., 2002), confirming that the response to temperature is
not very significant and the assimilation by soil microorganisms is relatively low
both in temperate and tropical soils. Furthermore, soil type seems not to affect
immobilization, as similar immobilization rates have been reported for different
soils in rotation after a wheat crop, probably because immobilization is limited by
readily assimilable carbon available for the soil microflora (Recous et al., 1992).
Nitrification of TAN mixed with soil is largely affected by soil pH, being
negligible at soil pH values lower than ca. 4 and increasing linearly with pH
increasing from 4 to 6.2 in a sandy soil (Winter and Eiland, 1996). Increasing
TAN (substrate) concentrations increases nitrification until either salt effects or
NH3 toxicity reduces nitrification (Malhi and McGill, 1982). Thus, at a level
higher than 250– 300 mg NHþ 4 -N g
21
(soil), nitrification has been shown to be
inhibited after mixing (NH4)2SO4 in soils having a pH of 5.8– 7 (Malhi and
McGill, 1982; Flowers and O’Callaghan, 1983); the inhibition was primarily a
salt effect, as the NH3 fraction of TAN is low in soils at pH , 7. At high pH, the
concentration of the NH3 component of TAN will increase and reach levels toxic
to the microflora (Malhi and McGill, 1982). Nitrification can be expected at
“permanent wilting point” and increases with increasing water content (Malhi
and McGill, 1982; Flowers and Callaghan, 1983). Near water saturation,
nitrification is absent due to a shortage of oxygen (Malhi and McGill, 1982). The
nitrification rate increases linearly with soil water content (zero order process;
Flowers and O’Callaghan (1983) and is ca. 2.8 times higher at 0.033 MPa than at
21.5 MPa, respectively, at field capacity and permanent wilting point (Malhi and
McGill, 1982).
Nitrification rate, micog–1g(soil) d–1

5 2.5˚C 7.5 Soil I
25˚C Soil II
4 7.0
Soil III
Soil pH
3 6.5
2 6.0
1 5.5
0 5.0
0 10 20 30 40 50 60 0 100 200 300
Temperature, ˚C Concentration of TAN,
microg TAN-N g–1(soil)
Figure 10 Left: Measurements of nitrification in Alberta soils from an area with a mean annual
air temperature of 2.58C (Malhi and McGill, 1982) and North Australian soil with a mean annual air
temperature of 258C (Myers, 1975). The figure is inspired by Malhi and McGill (1982). Right: Change
in soil pH after completion of nitrification of (NH4)2SO2 mixed with the soil and no NH3 emission.
Soil I: gray luvisol; Soil II: dark gray chernozemic, and Soil III: black chernozemic (Malhi and
McGill, 1982).
Temperature dependence of nitrification is different in warm or tropical soils

and temperate soils (Fig. 10), however 15N pool dilution studies have shown that
nitrification may be very rapid in most climatic regions (Watson et al., 2000;
Watson et al., 2002 quoted in Murphy et al., 2003). Daily nitrification rates of
2.2 – 7.9% for an acid sandy soil and a sandy loam at 238C have been reported by
Sørensen (2001) and 4– 6% at 58C by Flowers and O’Callaghan (1983), at
substrate concentrations of 50 – 100 mg NHþ 4 -N g
21
(soil). These rates
correspond to the nitrification rates calculated by Malhi and McGill (1982)
using a simple first-order model with input of substrate concentration,
temperature, and water content of the soil. A generally applicable model will
need inclusion of the effects of pH and changes in the population of nitrifiers
(Gilmour, 1984; Grant, 1994).
Immobilization may cause reduction in NH3 emission, but the effect is
considered to be negligible because the process is relatively slow compared
with volatilization from applied mineral fertilizers, i.e., daily immobilization
rates is 0.5 – 1.3% of the TAN or urea applied compared to NH3 emission
rates of 5 –10% immediately after hydrolysis of applied urea or dissolution of
ammonium salts (see Fig. 4). The studies also indicate that immobilization of
NO2 3 is very slow (i.e., about 5% of the immobilization rate of TAN; Mary
et al. (1998)) and may therefore not be accounted for in NH3 emission
studies. Contrary to immobilization, nitrification should be accounted for
when interpreting NH3 emission experiments or when developing models to
predict NH3 emission.
Nitrification may affect NH3 emission through reduction in TAN and by

reducing pH, as nitrification of 1 mol NHþ þ
4 produces 2 mol H , according to the
following equation:
NHþ 21
4 þ 2H2 O $ NO3 þ 2H3 O
þ
ð23Þ
The study of Flowers and O’Callaghan (1983) showed that nitrification of

100 mg NHþ 4 -N g
21
(soil) resulted in a reduction of 0.5 pH units and nitrification
þ
of 250 mg NH4 -N g21 (soil) resulted in a reduction of 1 pH unit. Nitrification,
therefore, reduces NH3 emission due to reduction in both concentration of TAN
in soil solution and a reduction in the NH3 component of TAN. Figure 10 depicts
the change in soil pH measured in the study of Malhi and McGill (1982),
showing that the change in pH is related to amount of added TAN and
differences in soil buffering capacity. The addition of fertilizers may cause a
local change in soil pH, the magnitude of the change being related to the
concentration of fertilizer, soil pH –buffer capacity, and fertilizer type. In
consequence of the change in pH due to application of the fertilizers, the
nitrification rates was in the following order urea . DAP . AS in the study of
McInnes and Fillery (1989).
IV. AMMONIA EMISSION FROM CROP FOLIAGE
A. TRANSPORT OF NH 3 BETWEEN LEAVES

AND THE ATMOSPHERE
In plants, the major source of NH3 is TAN dissolved in the water film in the
mesophyll cell walls of leaves (the so-called apoplastic solution; Husted and
Schjoerring, 1995). The concentration of TAN and Hþ is affected by uptake of N,
translocation, and transformation of N, which varies with plant developmental
stage, climate, and fertilization. The NH3 flux, FNH3, between a single plant leaf
and the atmosphere can be described as:
F NH3 ¼ gleaf ðx 2 NH3;a Þ ð12Þ
where gleaf is the conductance to diffusion of NH3 between the atmosphere and
the interior of the leaf, and x is the NH3 concentration of the air in the substomatal
cavities and intercellular air spaces within the leaf. Whether a leaf will act as a
sink for or a source of atmospheric NH3 depends on the difference in internal and
external NH3 concentration. If NH3,a exceeds x, NH3 will be absorbed, while in
the opposite case emission will occur. When NH3,a equals x, no net NH3
flux occurs between the leaf and the atmosphere. The internal NH3 concentration
at which FNH3 is zero is called the stomatal compensation point for NH3
(Eq. (12); Farquhar et al., 1980; Husted et al., 1996). As evident from Eq. (12),
the rate and direction of NH3 fluxes between plant leaves and the atmosphere at a
given atmospheric NH3 concentration are controlled by the conductance to NH3
transfer and the internal NH3 concentration.
The leaf resistance (inverse of conductance) to NH3 transfer includes a
stomatal and cuticular resistance in parallel and usually also includes a boundary
layer resistance in series with the two other terms. The latter is a function of the
aerodynamic properties of the leaf, wind speed, and canopy turbulence, and will
generally be an order of magnitude smaller than the maximum stomatal
resistance. The cuticular resistance (Rcut) is extremely high for NH3, probably in
the range of 2000 – 40,000 s m21 (van Hove et al., 1989). This means that hardly
any NH3 will pass through the cuticle. However, NH3 can readily be deposited on
the cuticular surface due to the presence of a surface water film. Since NH3 is
highly soluble in water, moist leaf surfaces can act as a storage compartment for
atmospheric NH3, making the leaf surface a temporary sink for NH3.
B. MAGNITUDE OF NH 3 LOSSES
In general, NH3 is emitted from intensive agricultural ecosystems (Fig. 11),

while semi-natural ecosystems act as NH3 sinks (Sutton et al., 1993a; Schjoerring
et al., 1998). The wild understory plant, Luzula, and three native grass species
showed low NH3 compensation points of 0.5 – 2 nmol mol21 (Hanstein et al.,
1999; Hill et al., 2001). Crop plants such as oilseed rape or barley usually show
USA, 1985
Ammonia emission,kg NH3-Nha–1
16
14
12
DK, 1993-1994
DK, 1994
DK,1993
10
DK,1989-1990
8
DK,1994
DK,1993
6
4
UK
2
UK
0
Barley Wheat Oilseed rape
Figure 11 Seasonal NH3 emissions from agricultural crops. Data from Harper et al. (1987);
Mattsson and Schjoerring (2001); Sutton et al. (1993b); Schjoerring et al. (1993); Yamulki et al.
(1996).
NH3 compensation points between 2 and 6 nmol mol21 (Husted and Schjoerring,
1995, 1996; Husted et al., 1996; Mattsson et al., 1997). In barley, the NH3
compensation point changed in relation to the developmental stage although the
plants were grown under constant N limitation (Husted et al., 1996). In a field
experiment in the Netherlands, intensively managed ryegrass showed NH3
compensation points varying over the season between 1 and 7 nmol mol21
(van Hove et al., 2002). For the same crop species grown in Scotland, Loubet
et al. (2002) monitored NH3 compensation points ranging from 0.02 mg
NH3 m23 in periods between fertilizations up to 10 mg NH3 m23 just after
fertilizations. Under laboratory conditions, both ryegrass and Bromus erectus
showed very high NH3 compensation points up to 18 nmol mol21 particularly
when supplied with high levels of NHþ 4 to the growth medium (Mattsson and
Schjoerring, 2002).
Quantification of the NH3 exchange between the atmosphere and the canopy
of barley, wheat and oilseed rape over two growing seasons show that the crop
foliage is a net source of NH3 to the atmosphere, with emissions ranging between
1 and 5 kg NH3-N ha21 year21 (Fig. 11). Very high NH3 emissions of up to
15 kg NH3-N ha21 per season were reported for winter wheat in the United
States by Harper et al. (1987). Harper et al. (1996) and Plantaz (1998) also
measured high daily NH3 emissions from grassland in the Netherlands during
spring and summer and from these emissions, high NH3 compensation points
were derived. In contrast, based on measurements of stomatal NH3 compensation
points by apoplastic bioassay, van Hove et al. (2002) concluded that plants in
intensively managed grasslands in the Netherlands would not contribute to
atmospheric NH3 loadings.
For wheat, oilseed rape, and barley, the accumulated NH3 loss over a growing
season constituted between 1 and 4% of the applied N, or between 1 and 4% of
the total shoot N (Schjoerring and Mattsson, 2001). The loss increased under
conditions with a high N concentration in the foliage and was positively
correlated with the above-ground crop N content at anthesis, but not with that at
final maturity. There were no indications that NH3 emissions were larger under
conditions unfavorable for nitrogen remobilization from vegetative plant parts
(low N harvest index). Nevertheless, a distinct peak in NH3 emission occurred
during senescence. NH3 emissions from plant stands, measured under simulated
environmental conditions in wind tunnels, ranged between 0.8 and 1.4% of the
N content of the shoot, equivalent to 1.1– 2.9 kg NH3-N ha21 (Mannheim et al.,
1997). The highest emissions were observed in faba beans, whereas the emissions
in winter wheat, spring rape, and white mustard were lower. The total NH3
emissions were not affected by removing a part of the ears (sink reduction), but
emissions occurred earlier, as did the plant senescence. This suggests that the
NH3 emissions are closely related to senescence (Schjoerring et al., 1993;
Mannheim et al., 1997). Emission of NH3 has also been suggested to contribute to
the decline in shoot nitrogen content that is often observed in agricultural crops
during the generative growth stage (Wetselaar and Farquhar, 1980; Schjoerring
et al., 1989; Francis et al., 1993). In this context, it is important to emphasize that
assessments of NH3 losses based on measurements of changes in 15N-labeled
or total above-ground N are indirect and other pathways of nitrogen loss
may influence the results, resulting in overestimation of losses (see below in
Section VI on Measurement techniques).
It can be concluded that plant communities on agricultural cropland represent
a net source of NH3 to the atmosphere. Net emissions range from below 1 up to
15 kg NH3-N ha21 per season, depending on plant species, crop nitrogen
economy status, and climatic conditions. Crops will in many areas represent a
significant input of NH3 to the atmosphere and NH3 losses may become large
enough to significantly affect crop N budgets.
C. PHYSIOLOGICAL PROCESSES INVOLVED IN NH 3 EMISSION

FROM CROPS
In addition to being taken up directly from the soil, NH3/NHþ 4 is produced in

plant tissues by a number of different metabolic processes. For example, NHþ 4 is
produced via nitrate reduction, through the fixation of atmospheric nitrogen by
root nodules, by photorespiration in leaves, and through the phenyl propanoid
pathway (Hirel and Lea, 2001). Ammonium may also be released during
reassimilation of nitrogen transport compounds (e.g., asparagine, glutamine,
arginine, and ureides) and through breakdown of other N compounds during
senescence and remobilization. These processes take place in different cell
organelles and in different tissues leading to both spatial and temporal variation
in tissue NHþ 4 concentrations. It is therefore difficult to predict the proportion of
the NHþ 4 present in plant leaves that at any stage will be contributing to the
emission of NH3. A further complicating factor is that environmental conditions,
particularly nitrogen supply, also affect tissue NHþ 4 levels (see below).
1. NH3 Compensation Point
The major route for NH3 exchange between plants and the atmosphere
is through the leaf stomates. In order for NH3 to be emitted through the stomates,
the concentration inside the leaf (i.e., compensation point) has to be higher than
the ambient concentration (see Section IV.A). The leaf apoplast constitutes the
interface between the atmosphere and the living leaf tissue, and the NHþ 4
concentration in the liquid phase of the apoplast (the water contained within the
cell walls) is therefore a critical parameter in determining the gaseous NH3
concentration inside the leaf. The stomatal NH3 compensation point can be
calculated from measurements of the pH and NHþ 4 concentration in the apoplastic

solution (Eqs. (3) and (4); Husted and Schjoerring, 1995; Mattsson and
Schjoerring, 2002). Apoplastic NHþ 4 concentrations normally range between
0.01 and 2 mM but have been shown to be highly dynamic and closely coupled
to the plant N metabolism (Mattsson et al., 1998; Nielsen and Schjoerring, 1998;
Mattsson and Schjoerring, 2002).
2. N Uptake and Translocation
There are different pathways for NHþ 4 to reach the apoplastic solution; either
þ
through xylem transport of NH4 from the root or through NH3 efflux from the
mesophyll cells. Xylem transport of NHþ 4 can be quite substantial, particularly if
the plants are grown on high levels of nitrogen. The concentration of NHþ 4 in
xylem sap increases with increasing supply of both NHþ 4 (Mattsson et al., 1998)
and NO2 þ
3 (Husted et al., 2000a). These increasing xylem sap NH4 concentrations
þ
are usually reflected in the apoplastic NH4 concentration (Mattsson et al., 1998;
Finnemann and Schjoerring, 1999; Husted et al., 2000a). Inside the leaf cells,
NHþ 4 can either be assimilated by cytosolic glutamine synthetase (GS1) or taken
up into the chloroplasts and assimilated by the chloroplastic form of the enzyme
(GS2). Depending on the capacity of these enzymes to assimilate NHþ 4 into
glutamine and then by the enzyme glutamate synthase (GOGAT) to convert
glutamine into glutamate, NHþ 4 may accumulate in the leaves. In general, high
NHþ 4 concentrations in xylem sap and apoplast also result in a high leaf tissue
extract NHþ 4 concentration (Mattsson et al., 1998), but there are also examples
where leaf tissue NHþ 4 did not rise despite increasing xylem and apoplast
concentrations (Husted et al., 2000a). In grasses grown with either NO2 þ
3 or NH4 ,
þ
significant correlations between leaf tissue NH4 concentration and both NH3
emission and apoplastic NHþ 4 concentration were observed (Mattsson and
Schjoerring, 2002).
Decreasing the activity of GS, either by using an inhibitor such as MSX or by
using mutants or transgenic plants with lower assimilation capacity, usually
results in increasing concentrations of NHþ 4 in various plant compartments.
In consequence, NH3 emission to the atmosphere increases within a few hours
after adding MSX to the growth medium in both barley and oilseed rape, because
GS activity is decreased and tissue TAN concentrations increase (Husted and
Schjoerring, 1995; Mattsson and Schjoerring, 1996). In barley mutants with
reduced activity of chloroplastic GS, higher leaf tissue and apoplastic NHþ 4
concentrations have resulted in higher NH3 emission compared with wild-type
plants (Mattsson et al., 1997). Emission of NH3 also seems to increase more with
increasing temperatures in the mutants than in the wild-type plants, suggesting a
higher sensitivity to photorespiration in GS mutants. A massive release of
NH3/NHþ 4 takes place during photorespiration, particularly at high temperatures

(Leegood et al., 1995), making it extremely important for the plant to have
an efficient reassimilation of this NH3. The influence of photorespiration on
apoplastic and leaf tissue NHþ 4 concentrations was investigated in antisense
GS2 oilseed rape plants, i.e., plants with reduced activity of the key enzyme
responsible for re-assimilation of photorespiratory NHþ 4 wild-type plants (Husted
et al., 2002). Despite a 50– 75% lower in vitro leaf GS activity in the antisense
plants, there was no tendency for these plants to have higher tissue NHþ 4
concentrations than wild-type plants. Antisense plants exposed to leaf
temperatures increasing from 14 to 278C or to a three-fold increase in the
O2/CO2 ratio did not show any change in steady state leaf tissue NHþ 4 con-
centration or in NH3 emission to the atmosphere. These results show that oilseed
rape has a large surplus of GS in the leaves, which makes the plants less sensitive
to increasing photorespiration than is the case for barley (Mattsson et al., 1997).
The influence of N2 fixation on the NH3 emission potential has been
investigated in white clover (Herrmann et al., 2002). Symbiotic N2 fixation and
mineral N acquisition were shown to be well balanced with respect to the
apoplast and plant tissue NHþ 4 concentrations, leading to equal NH 3
compensation points in plants grown with or without an external NO2 3 supply.
Field measurements over pea crops similarly indicate NH3 emissions comparable
to those for barley and wheat (Schjoerring and Mattsson, 2001).
Senescing plant material undergoes a sequence of biochemical and
physiological events including protein degradation, which eventually leads to
cell death. The massive release of NH3 in senescing material is thought to be a
consequence of amino acid deamination and catabolism of nucleic acids
(Brouquisse et al., 2001). Dark-induced senescence of barley plants and detached
leaves of oilseed rape both showed increased NH3 emission which was
synchronized with chlorophyll degradation and liberation of NHþ 4 in the leaf
tissue (Schjoerring et al., 1998).
V. MANAGEMENT STRATEGIES
Fertilizers are spread mainly as granules or prills. The fertilizers are

granulated and prilled because this reduces inhomogeneity and because these
fertilizers can be spread evenly. Furthermore, additives mixed with the fertilizers
can produce slow-release fertilizers or reduce transformations after application
as, for example, with urea. Fertilizers are either broadcast onto the soil surface or
injected into the soil. AA is usually injected or can be dissolved in irrigation
water, and urea is sometimes incorporated for the purpose of reducing NH3
emission. In Asia, farmers broadcast nitrogen fertilizer (mainly urea) into the
water in rice paddy fields.
A. TECHNIQUES FOR REDUCTION OF NH 3 EMISSION
Top dressing and partly covering urea may reduce emissions significantly
(Fig. 12). The effect of placing urea is related to the depth and soil characteristics
(Fenn and Kissel, 1976). Placing urea in a soil low in CEC and with low pH
buffering capacity may create a zone of high concentrations of dissolved TAN
and a high pH due to hydrolysis. In consequence, the concentration of NH3 may
be very high in the zone affected by urea. In these soils, shallow placement of
urea may have little effect on NH3 emission (Fig. 12) because the NH3 will be
transported by diffusion to the surface and be lost (Blaise et al., 1996). Thus there
is a high correlation between amount of urea applied and placement depth and
emission, so at high application rates urea should be placed at greater depths than
at low application rates. To eliminate NH3 emission from urea applied to
calcareous soils, the fertilizer may need to be placed as deep as 5 –7.5 cm
(Fenn and Miyamoto, 1981; Ismail et al., 1991).
Harrowing stubble before urea application may halve NH3 volatilization,
because cultivation forms cracks and small hollows where urea prills will be
protected from volatilization, and rain events or irrigation may leach urea into the
soil (Bacon et al., 1986; Bacon and Freney 1989). Mixing ammoniacal fertilizers
with the soil may be a less efficient reduction measure than injection to the same
depth because a part of the mixed-in fertilizer will be close to the surface and
TAN will be transported either by diffusion or convection upwards and be lost
(Nye, 1992).
Increasing the application rate may reduce the relative emission of NH3 from
urea and ammonium fertilizers applied to calcareous soils (Du Preez and Burger,
1988). On acidic soils the proportion of urea-N lost due to NH3 emission has been
Broadcast
Ammonia emission, NH3 pct. of applied N
35
Placed 1 cm
30
Broadcast
Mixed 2 cm
25
Mixed 3 cm
20
Placed 2.5 cm
15
Placed 2 cm
10
0
Wind tunnel Dynamic chamber
Figure 12 Ammonia emission from urea broadcast or placed at different depth to a bare soil
(Bouwmeester et al., 1985; Nye, 1992).
shown to increase by increasing the application rate (Black et al., 1987; Watson
and Kilpatrick, 1991). This discrepancy is due to interaction of fertilizer
alkalinity and soil acidity. In acid soils, the pH buffer may prevent high increases
in pH after addition of urea at low rates, whereas pH may increase in acid soils
amended with urea at high rates because the amount of alkaline hydrolysis
products will be higher than microsite acidity of the soil. Similarly, surface-
application of urea in bands may increase NH3 emission compared to broad-
spreading when applied to acid soils. Although, absorption of NHþ 4 may affect
transport of TAN in the soil, model simulations of Nye (1992) have shown that
soil pH buffer capacity is more important in influencing NH3 emission.
Application of pelleted fertilizers or fertilizers in solution may contribute to
local “hot spots” with high salt concentrations and pH different from the
surrounding soil. In a hot spot with urea, the pH will be high compared with the
pH of the surrounding soil, and in a hot spot with “acid” ammonium salts pH will
be low. Immediately after band application the high urea concentrations may
reduce urease activity due to substrate inhibition (Fig. 5) and as the uncharged
urea may infiltrate more easily to greater depths than NHþ 4 this may result in
reduced emission of NH3 (Fenn and Miyamoto, 1979, Bouwmeester et al., 1985).
Straw has a much higher pH buffer capacity and pH than soil and a 20-times
higher urease activity than the surface 10 mm of soil; consequently broadcasting
urea to straw spread on the soil may contribute to a high potential of NH3 emission
(McInnes et al., 1986b). Similar, applying urea to trash left from harvesting of
sugarcane may cause significant losses of NH3 (Fig. 13) because the urea is not in
Trash - dry
W. wheat seeding, Z 05
Ammonia emisson, NH3 pct. of applied N
W. wheat, tillering, Z 14.22
40
Banana plantation
Trash - heavy rain
W. wheat, ear initiation, Z 30

Calcareous
Fine silt loam
Coarse silt
30
Fine silt loam
20
Disced-scarified
10
Trash
0
UAN Urea Urea AS Urea Urea
Figure 13 Ammonia emission from UAN (urea ammonium nitrate), urea and AS (ammonium
sulphate) applied to bare soil, sugarcane trash-covered soil, to winter wheat, and banana plantation.
(Adapted from McInnes et al., 1986a,b; Bouwmeester et al., 1985; Bacon and Freney, 1989; Freney
et al., 1992; Prasertsak et al., 2001).
contact with soil and the trash has a high pH. Moving the litter aside on no-till
soil and applying urea to the soil surface may reduce losses significantly
(Touchton and Hargrove, 1982).
B. FERTILIZER COMPOSITION
Ammonia emission from fertilized soils has been reported to vary between
negligible and 40% when measured with the micrometeorological mass balance
technique (Fig. 13; Freney et al., 1992; McInnes et al., 1986a). The emission
from a specific fertilizer will vary with climate, soil pH – buffer capacity, and
CEC; therefore the average values shown in Figs. 13 and 14 cover considerable
variation, as indicated by the relatively high standard variation.
Ammonia emission is affected by the choice of fertilizer and rate of
application. Fertilizers applied to calcareous or limed soils should preferably be
acidic fertilizers with anions not forming calcium precipitates, e.g., AN or
NH4Cl. Soils low in Ca may be fertilized with MAP, AS, or AN, which will
reduce pH upon dissolution in soil water. Surface application of urea to acidic
soils with a high pH – buffer capacity or a high CEC may not cause high losses.
Urea should not be surface-applied to calcareous soils because TAN is at risk of
being lost due to NH3 emission; therefore, incorporation would be recommended.
40
Ammonia emission, NH3-N pct. of applied N
30
20
10
0
Urea DAP CAN AS
Figure 14 Ammonia emission from ammoniacal fertilizers, i.e., urea, diammoniaum phosphate
(DAP), calcium ammonium nitrate (CAN) and ammonium sulfate (AS) broadcast to crops, measured
with wind tunnels (Sommer and Jensen, 1994; Velthof et al., 1990; van der Weerden and Jarvis, 1997).
An example of adapting fertilizer application to soil and crop is banana

plantations in a tropical area in Queensland (Australia), where 17% of the applied
urea-N may be lost due to emission of NH3 (Prasertsak et al., 2001). The
fertilizers cannot be incorporated because this will damage the roots; therefore,
Australian banana producers are recommended to switch from urea to AN, with a
loss potential lower than that of urea (Prasertsak et al., 2001). On grassland,
incorporation of solid fertilizers may also be detrimental to the plants, and on acid
soils application of liquid urea ammonium nitrate instead of granular urea may
reduce losses because the extent of microsites with high TAN concentrations is
reduced. Lightner et al. (1990) showed that applying urea in liquid solution may
reduce emission, with 35% compared with granular-applied urea.
Application of inorganic salts with fertilizers may reduce NH3 emission
significantly. The salts used should be soluble, thus CaSO4 is not efficient for
reducing NH3 emission. Adding CaCl2 with urea will contribute to reduction in
2 þ
NH3 emission, because CO22 3 will precipitate as CaCO3 and Cl and NH4 will
reduce pH significantly as NH3 is lost. KNO3 or KCl may be used with urea to
reduce NH3 emission, the Cl2 may contribute to produce an acid environment,
reducing the NH3 emission potential. In addition, Kþ cation may exchange with
absorbed Ca2þ, which will precipitate the CO22 3 produced during hydrolysis
(Fenn et al., 1982).
Sulfur applied with urea will hydrolyze to SO22 4 and reduce NH3 emission, but
the effect may not be very significant in all environments, i.e., NH3 emission from
sulfur-coated urea was equal to or somewhat lower than from untreated urea
(10.1% as compared with 12.6%) in the study of Black et al. (1985), and natural
finely ground sediments of pyrite (FeS) have been shown to reduce emission from
urea by 54% (Blaise et al., 1996). Thiosulfate at 10% reduces hydrolysis rates
and, under drying conditions, CaCl2 may also reduce hydrolysis (Black et al.,
1985); the reduction is due to the low pH of the Cl2 (Fig. 5). In dry conditions, Cl
concentration will be high due to slow diffusion and the pH reduction will be
significant (Sloan and Anderson, 1995). Addition of H2SO4 or H3PO4 with urea
to calcareous soils is questionable due to precipitation of the anions with calcium.
In calcareous soils, adding acids that precipitate should, therefore, be avoided and
either HCl or HNO3 may be used with urea to reduce losses of NH3 (Fenn and
Richards, 1986). However, Fenn and Hossner (1985) do not recommend the use
of these acids because urea nitric acid is a cold explosive in solid form and HCl
may be noxious and injurious to plants. Furthermore, adding pyrite and H2SO4
may contribute an increase in the soil’s demand for lime, which will be an
additional cost for the farmer.
Urease inhibitors delaying hydrolysis of urea to TAN may be an option for
reducing NH3 emission. Slowing down the hydrolysis allows time for urea to
infiltrate into the soil by diffusion or convection after a rain event, because the
uncharged urea is more readily transported in soil than the charged NHþ 4
(Fenn and Miyamoto, 1981). Of the inhibitors tested, the phosphoryl di- and
triamides, (N-(n-butyl)thiophosphorictriamide (NBPT) or phenylphosphorodia-

midate (PPD) were found to be the most useful, and have proven effective to
varying degrees depending on the environment and management (Byrnes and
Freney, 1995). The efficiency is related to climate and soil conditions; i.e.,
applying the urea with urease inhibitor to a humid soil during a period with little
rain may not reduce NH3 emission, whereas application followed by a heavy
rainfall few days after application may prove very efficient. To become efficient,
NBPT must be converted to the oxygen analogue (N-(n-butyl)phosphorictria-
mide, NBPTO); on the other hand, PPD is efficient immediately but is
decomposed (departs from neutrality) rapidly (Bremner, 1995). In waterlogged
soil or in floodwater in rice fields, the oxidation of NBPT applied with urea will
be retarded and the inhibition of urea hydrolysis will be inefficient (Bremner,
1995). Therefore, Phongpan et al. (1995) applied urea with PPD and NBPT to
rice fields. It appeared that initially PPD inhibited urease activity, and during this
time at least part of the NBPT was converted to NBPTO; then as the activity of
PPD declined, NBPTO inhibited the hydrolysis of urea. This combined urease
inhibitor treatment reduced NH3 emission from 15 to 3% of the applied N.
Hydroquinone is a less effective urease inhibitor than the phosphoramides but
may have beneficial physiological effects on the plant (Bremner, 1995). Due its
low cost, hydroquinone has been tested and is formulated with urea for use in
China.
The concentration of a chemical inhibitor required to suppress hydrolysis
decreases with increasing granule size of the fertilizer, as seen with nitrification
inhibitors (Singh et al., 1994). Thus, increasing granule size of the urea fertilizer
may delay hydrolysis significantly because the contact to soil is reduced. Black
et al. (1987a) reported that urea concentration was significant for 5 days when
applying 8 mm granules, compared with 2 days after application of powder or
, 4 mm granules. The pH in the granules increased to values between 8 and 9
within 2 days in both the small and large granules. Consequently, urea was
hydrolyzed into a microsite with high pH and high TAN concentration for a
longer period in the large granule than in small granules, causing the cumulated
emission to be large. This was confirmed by the study of Watson and Kilpatrick
(1991), who found little difference in NH3 emissions between different sized
prills of applied urea.
C. FLOODED FIELDS (RICE PADDIES)
The fertilizer application to rice paddies is highlighted, because rice is the

world’s most important food crop and 75% of the world’s rice is grown in several
inches of water held in small dikes, i.e., paddies (Maclean, 1997). Management
of the rice paddies may briefly be characterized by flooding of the dry paddy,
puddling, transplanting of rice, vegetative and reproductive phases of rice,
and harvesting. Nitrogen uptake efficiencies range from 20 to 60% of applied N

(Maclean, 1997). This low efficiency reflects poor management and the use of
urea-N, leading to high NH3 emissions. Thus, reducing NH3 emission may act to
optimize N-use efficiency.
The water in paddies is buffered by bicarbonate (HCO2 3 ), which is the main
contributor to the alkalinity of the water. Therefore, consumption of CO2 during
daytime and respiration of CO2 during the night will cause large fluctuations in
the pH of the water in rice paddies with a high photosynthetic biomass, especially
algae (Fillery et al., 1986; Bowmer and Muirhead, 1987). The uptake of CO2 will
increase pH during the daytime and respiration of CO2 will reduce pH at night.
The diurnal variation in NH3 emission clearly reflects the variation in pH (Fillery
et al., 1986).
Alkalinity may be significantly higher in paddies irrigated with floodwater or
alkaline well water compared with paddies irrigated with rainwater (Vlek and
Crasswell, 1981; Fillery et al., 1984). The alkalinity and pH of the water will be
affected by the applied fertilizers; thus (NH4)2SO4 decreased soil water pH
fluctuation initially, but after 2 days pH fluctuations were similar to variations in
pH of floodwater fertilized with urea in the study by Fillery et al. (1984). This
indicates that after a few days, the HCO2 3 of the water, including CO2 from
respiration, had buffered the pH to the initial values. In contrast, emission of NH3
from (NH4)2SO4 applied to water with low alkalinity would have resulted in a
reduction in pH, which eventually would have reduced emission significantly
(Vlek and Crasswell, 1981). Hydrolyzed urea will contribute to the alkalinity
and may increase pH to 10 in weakly buffered water (Mikkelsen et al., 1978),
and the NH3 emission may be significant. Removing the algae biomass or
reducing the photosynthetic activity will reduce the diurnal variation in pH of the
water, and the addition of a suitable photosynthetic inhibitor (terbutryne) with
fertilizer may reduce the daytime pH of rice floodwater for up to 6 days and the
potential NH3 emission by 43% (Bowmer and Muirhead, 1987). Adding urease
inhibitors will delay hydrolysis of urea and reduce the concentration of TAN in
the water, thereby reducing emission of NH3 (Fillery et al., 1986). The effect of
urease inhibitors is variable and related to water chemistry.
Mixing of the water due to wind and changes in buoyancy due to heating and
cooling of surface water will vary during the day. Stratification of subsurface
water may reduce emission in periods with heating of surface water and little
mixing caused by wind. Radiation during the daytime will heat the surface water,
which will enhance stratification, and cooling in the afternoon will lead to
enhanced mixing (Leuning et al., 1984). In consequence, mixing may contribute
to the high NH3 emission rates in the afternoon.
Application of acidic fertilizers will reduce NH3 emission. Fig. 15 shows that
emission of NH3 from urea may be higher than emission from (NH4)2SO4 when
applied at transplanting. In Fig. 15, data from different studies have been used
and comparing results from different studies may bias conclusions due to the
60
NH3 emission, % of applied N 50
Incorporation dry soil, IRRI

Incorporation in water, IRRI
Incorporation in water, IRRI

At transplanting, NWS
At transplanting, IRRI
At transplanting, IRRI
40
30
Panicle, IRRI
Panicle, IRRI
20
10
0
Urea (NH4)2SO4
Figure 15 Ammonia emission from rice paddies, measured with the micrometeorological mass
balance technique at either IRRI (Philippines) or in New South Wales (Australia). (From Leuning
et al., 1984; De Datta et al., 1989, 1991; Fillery and De Datta,1986; Fillery et al., 1986; Freney et al.,
1981).
effect of different environmental conditions (e.g., effect of alkalinity, wind,

temperature, etc.), but the experiments were carried out on similar soils and
wind fluctuations were similar in the studies of Freney et al. (1981) and
De Datta et al. (1991).
Paddy rice in the tropics is fertilized with urea, which may be applied to the
rice field before transplanting rice, to the rice at transplanting, to the floodwater
ca. 4 weeks after sowing and at panicle initiation or booting (De Datta et al.,
1989, 1991; Son and Buresh, 1994). In the study of De Datta et al. (1991),
46– 54% of the urea-N applied immediately after flooding was emitted as NH3
due to the low buffer capacity of floodwater and a high pH after hydrolysis of the
urea. Incorporating the first application of urea into the dry soil before flooding
may reduce NH3 emission to about 10% compared with the emission of 46 – 54%
of urea-N after broadspreading of the urea to the rice field after flooding (Fig. 15).
This soil had a pH of 4.6 and a CEC of 250 mmol kg21 and because TAN was
absorbed on soil colloids as NHþ 4 , the potential of NH3 emission from the
incorporated urea was low. Incorporation of urea in flooded soils has little effect
on NH3 emission from urea but significantly reduces the emission from
(NH4)2SO4; probably urea incorporated in flooded soils is not adequately in
contact with the soil to reduce emission (Fig. 15).
In addition to incorporating urea before flooding rice fields, NH3 emission
from paddy fields may also be reduced by matching the application to crop
demand by split applications of urea to the floodwater. Urea application to rice
may be delayed 14 – 16 days after transplanting in the period of vigorous growth
and assimilation of the nitrogen applied (Son and Buresh, 1994). The plant uptake
of N will thereby reduce NH3 emission and contribute to a higher fertilizer

efficiency of the applied urea-N. Little NH3 is emitted from urea applied to the
rice field at panicle initiation or booting, because rice is rapidly assimilating the
nitrogen, and shading of the water reduces temperature and wind speed (Freney
et al., 1981; De Datta et al., 1989). Furthermore, shading of the water reduces pH
increase due to respiration of algae.
D. INJECTION OF ANHYDROUS AMMONIA
The emission of NH3 from injected AA will be related to application rate, depth
of injection, knife spacing, and soil buffering capacity (Izaurralde et al., 1990).
Penetration depth of NH3 increases with decreasing soil water content (Blue and
Eno, 1954; McDowell and Smith, 1958). Thus, from AA injected into a dry soil, the
NH3 emission was 20% because NH3 retention capacity was low and a part of the
injected NH3 could move through the air-filled pore space to the soil surface
(Sommer and Christensen, 1992). On the other hand, the crevice left after injection
into a wet soil was left open and the distance of NH3 penetrated into the soil to the
open crevice was short and emission was as high as 50% (Sommer and Christensen,
1992). Hopkins et al. (1963) showed that injection of AA may push water in the soil
ahead of the gas and thereby reduce gas movement in water-filled soils, which will
result in high TAN concentrations in soil water near the site of injection.
Anhydrous ammonia is usually injected into the soil after winter or a rainy
season. The injection will take place when soils are moist, as driving on wet soils
is either impossible or will compact the soil. Under these conditions, little NH3
will be emitted from AA injected to depths below 10 cm (Baker et al., 1959;
Denmead et al., 1977; Sommer and Christensen, 1992), because the furrow will
close and soil water will absorb the NH3.
Injection depth of AA where little or no NH3 emission will take place may be
predicted using TAN-transport models with input of amount of applied AA, soil
pH buffer capacity, diffusion as affected by soil water, and CEC (Izaurralde et al.,
1990). The model calculations have shown, for a fine sandy loam and a silt loam
soil, that emission is low at 15 cm injection and significantly higher (8%) at 5 cm
injection depth.
E. CROP NH 3 EMISSIONS AS AFFECTED BY FERTILIZER

APPLICATION
Volatile NH 3 losses from crops depend on seasonal variations in

climatic conditions affecting the growth and nitrogen economy of the crops
(Schjoerring and Mattsson, 2001). In general, losses are expected to increase with
the N concentration of the foliage and to the extent that this is controlled by other
growth factors than fertilizer N availability; there may be no connection between

NH3 emission and fertilizer application. As an example from a field study by
Schjoerring and Mattsson (2001), lower NH3 emissions were measured from
wheat plants fertilized according to optimum N-recommendations as compared
to plants applied a reduced amount of N-fertilizer (75% of optimum), because the
latter plants had a higher shoot N concentration during part of the growing
season. In the same study, the accumulated NH3 loss over a growing season was
positively correlated with the above-ground crop N content at anthesis, but not
with that at final maturity, and there were no indications that NH3 emissions were
larger under conditions unfavorable for nitrogen remobilization from vegetative
plant parts (low N harvest index).
Enhanced NH3 emission from leaves under conditions with excessive N
absorption by roots has been observed in several laboratory experiments (see e.g.,
Mattsson and Schjoerring, 2002). A similar relationship seems also to be valid
under field conditions; thus, as shown in Fig. 16, application of the acid fertilizers
DAP and AS to a grass ley or young wheat plants in April resulted in a peak of
NH3 emission between 5 and 12 days after fertilizer application. A similar peak
was not observed when the same fertilizers were applied in early spring (March),
i.e., before significant plant uptake of N, where the NH3 emission continuously
declined following high initial loss rates. The extra peak in April is
assumed to represent NH3 emission from the seedlings, following rapid
Ammonia emission, g NH3 m–2 day–1
0.5 DAP, April

DAP, March
0.4 AS, April
AS, March
0.3
0.2
0.1
0.0
0 5 10 15 20
Days from application of fertilizers
Figure 16 Wind tunnel measurements of NH3 emission from DAP and AS applied to grass (10–
15 cm) on 9 March 1992 and to winter wheat (5 cm) on 1 April 1992. In the March experiment, air
temperature was 4.38C, soil temperature 5.78C and wind speed in the wind tunnel 4.2 m s21. In the
April experiment, air temperature was 5.88C, soil temperature 7.48C, global radiation 10.9 MJ m22
and wind speed in the wind tunnel 4.2 m s21. In the April experiment, an atypical peak in emission
was measured from 5 to 13 days after fertilizer application; this peak is attributed to NH3 emission
from the weak seedlings not having capacity to transform NHþ 4 to amide/amines due to low global
radiation.
nitrogen uptake in a period with low global radiation and, thus, a low
photosynthetic activity.
F. AMMONIA EMISSION FROM DECOMPOSING PLANT

MATERIAL
Decomposing crop residues may constitute a significant source of atmospheric

NH3. In measurements by Mannheim et al. (1997), NH3 emissions from
decomposing sugar beet leaves, potato tops, and field-bean straw ranged from
0.9 to 3.7% of the N content. The highest emissions, reaching from 8.6 up to
12.6 kg N ha21, occurred from sugar beet leaves and potato shoots with high
water content, whereas the emissions from field-bean straw with high dry matter
and N content were relatively low (3.1 kg N ha21 or 0.9% of the N content). The
NH3 emission from sugar beet leaves was reduced 81% by plowing and 63% by
mulching (Mannheim et al., 1997). A marked influence of moisture on the NH3
volatilization during decomposition was also observed for ryegrass herbage,
where 20 –47% of the herbage N was lost over a 70 day period if kept moist,
while volatilization was less than 1% of herbage N during drying (Whitehead
et al., 1988). The quantity of NH3 lost increased with herbage N concentration
and temperature during decomposition (Whitehead et al., 1988). As an example,
herbage containing 3% N on a dry matter basis lost 10% of its N through NH3
volatilization over a period of 28 days, whereas no volatilization was detected
from herbage containing 0.9% N.
The profound increase in NH3 volatilization with moisture and nitrogen
content of plant residues, as well as with temperature, is consistent with the
stimulating influence of these parameters on protein degradation and liberation of
NH3/NHþ 4 in senescing leaves. An example of the influence of nitrogen status is
shown in Fig. 17, where bulk leaf tissue NHþ 4 concentrations increased more
during senescence of high-N leaves compared with low-N leaves, particularly
when plants had been growing at relatively low light intensity, resulting in
reduced C/N ratio relative to high-light grown plants (Fig. 17). The ratio between
[NHþ þ
4 ] and [H ], which reflects the NH3 emission potential from leaf litter, also
increased dramatically during senescence (Fig. 17).
Significant amounts of N may be lost from the plant by senescent leaves
falling off and decaying on the ground. Husted et al. (2000b), and Nemitz et al.
(2000) showed that in oilseed rape decomposing plant residues on the soil surface
made a large contribution to NH3 emission. Some of the NH3 emitted from these
leaves might be reabsorbed by leaves still attached to the stem (Nemitz et al.,
2000). The soil has also been suggested as an NH3 emission source. However,
only a small amount of NH3 emission from the soil was found and it was
proposed that the soil acts mostly as a sink for NH3 (Neftel et al., 1998).
40
A
Bulk tissueNH4+,μmolg–1 tissuewater
High light 0N
30 High light 3N
High light 6N
Low light 0N
Low light 3N
Low light 6N
20
10
0
70000
B
60000
Bulk tissue[NH4+]/[H+]ratio
50000
40000
30000
20000
10000
0
day 0 day 2 day 4
Days of senescence
Figure 17 (A) Bulk tissue extract NHþ 4 concentrations and (B) NH3 emission potential as
expressed by the [NHþ þ
4 ]/[H ] ratio in leaves of Lolium perenne grown for 4 weeks in 0, 3, or 6 mM
NO2 3 under high and low light conditions on the day of leaf excision and after 2 and 4 days of
senescence in darkness. C/N values in low-light grown plants were 52, 12, and 10 for plants grown in
0, 3, or 6 mM NO2 3 , respectively. The corresponding values for high-light grown plants were 63, 24,
and 23. Values are means ^ SE for three replicates. (M. Mattsson and J. K. Schjoerring, unpublished
results).
G. ABSORPTION BY CROPS
Plants are capable of absorbing atmospheric NH3 when the NH3 concentration
in the atmosphere exceeds that in the substomatal cavities. The absorption
responds linearly to ambient NH3 concentration over a very broad range even up
to around 500 nmol mol21 (van Hove et al., 1987; Whitehead and Lockyer,
1987). From urea applied to a grass pasture, the NH3 volatilization decreased
with increasing canopy density (Hoult and McGarity, 1987; Ping et al., 2000).
The crop may have changed the microclimate and the canopy may have absorbed
NH3 volatilized from the urea, thereby reducing the flux of NH3 from the soil –
plant system as shown in Fig. 13 for urea applied to winter wheat at increasing
physiological age. The maize canopy reduced NH3 emission from NH3 dissolved
in irrigation water (Denmead et al., 1982), due to a reduction in wind speed,
reduction of temperature and uptake of NH3 by the maize canopy.
In oilseed rape, significant amounts of N may be lost before flowering in
dropped leaves (Schjoerring et al., 1995), followed by significant NH3 emissions
from the decaying leaves (Sutton et al., 2000). However, a significant part of the
emitted NH3 is absorbed by leaves still attached to the plants, resulting in a very
high cycling of NH3 within the canopy, much higher than the net exchange with
the atmosphere above the canopy (Nemitz et al., 2000).
VI. MEASUREMENT TECHNIQUES
The following section gives an overview of the most widely used techniques
for measuring NH3 emission from fertilizer applied to the soil. McGinn and
Janzen (1998) and Harper (2003) have comprehensively reviewed techniques for
measuring NH3 fluxes to and from the soil.
A. TRACER TECHNIQUES
Ammonia emission has been estimated by crop response to fertilization or by

soil mass balance after application of fertilizer that in some studies have been
enriched with N15 (see Denmead et al., 1977; Farquhar et al., 1980; Nömmik,
1966; Moal et al., 1995; Morvan et al., 1997). Crop response has been shown to
be a very unreliable estimator due to variation in total N uptake and uptake of
soil-N, other loss pathways such as leaching or nitrification/denitrification, and
differences in the release patterns of N14 and N15 (Schjoerring et al., 1989).
Isotope techniques for estimating N losses have also been questioned, since they
seem to overestimate the loss, because NH3 can be both absorbed and released; a
plant grown with 15N-enriched fertilizer will tend to lose 15NH3 and gain 14NH3,
even if the net NH3 flux is zero (Francis et al., 1997). A significant transfer of 15N
between labeled and unlabeled plants has been shown to occur via the atmosphere
in controlled environment studies (Janzen and Gilbertson, 1994). This will result
in an overestimation of losses estimated by 15N analyses. Soil mass balances are
unreliable due to loss pathways other than NH3 emission and also because of
variations in concentration of TAN or organic N in the soil, i.e., for comparison,
recovery of bromide (Br2) added to soil in columns was between 78 and 116%
(de Jonge et al., 2003), showing the low precision of this technique.
Recently, Vandré and Kaupenjohann (1998) have described a method
whereby the transfer factor of NH3 from a source to a passive sampler on
experimental plots is determined by releasing NH3 at a known rate via a cylinder
and tubing on standard comparison plots. The transfer factor is then applied to
passive sampler measurements of concentration from manure-treated plots to
determine NH3 release rate (i.e., flux) from treated plots. A similar approach has
been used by Warland and Thurtell (2000) to infer rates of nitrous oxide
evolution from soil.
Sherlock et al. (1995) showed for bare soils that the emission of NH3 from
applied mineral fertilizers can be calculated as the product of wind speed, NH3
gas in equilibrium with NH3 dissolved in the surface soil layers, and a transfer
coefficient.
The equilibrium concentration technique (JTI method; Svensson, 1994) is a
micrometeorological method suitable for measuring NH3 emissions from small
plots. It involves sampling close to the soil surface to measure the driving force
for volatilization and the aerodynamic resistance to flux. The method has recently
been verified against the IHF method for applications of urea fertilizer and
manure to large plots (Misselbrook and Hansen, 2001).
The above-mentioned techniques have not been used extensively. The most
commonly used methods have been enclosures and the micrometeorological
methods described below.
B. ENCLOSURES
Enclosures are much used in both field and laboratory experiments. The
enclosures can be chambers placed on the soil surface with no air flow through
the head-space, i.e., static chambers, or they may be dynamic chambers with lids
through which the air is exchanged by means of ventilators or pumps. These
methods are useful when emission measurements are required over well-defined
areas (e.g., small plots) and for comparing treatments under identical
environmental conditions (Livingston and Hutchinson, 1995).
Chambers are popular because of their portability, versatility, relative
simplicity, and high sensitivity. They permit process studies and experiments
with many treatments in numbers that could not be contemplated with conventional
micrometeorological approaches because of the large land areas that the latter
require. Moreover, the large increase in gas concentration that occurs in the head-
space means that chambers can detect fluxes that are 100 times smaller than can be
detected by micrometeorological techniques (Denmead, 1994).
1. Static Chambers
The flux (F) is calculated from the rate of increase in gas concentration in the
enclosure just after the system has been closed.
F ¼ ðV=AÞdx=dt ð24Þ
where V is the volume of the head-space, A is the area of soil surface enclosed by
the chamber, x is gas concentration and t is time. The increase in gas
concentration is often measured by absorbing the NH3 in an acid solution and
storing in an open container or absorbing it on to a filter. However, the gas
concentration gradient from the emitting surface to the air beneath the enclosure
decreases as concentration in the air increases. Hence, the size of the enclosure
and the measurement period must be carefully selected to avoid negative
feedback on the rate of diffusion of the gas. Thus, the measured emission may be
lower than when using methods which pass air over the soil and fertilizer (Volk,
1959; Denmead, 1979; McGarry et al., 1987). Such chambers can be used when
gas emission rate is controlled by soil processes, as is the case for N2O and CH4,
but not in situations where air exchange has a large impact on the emission rate,
as may be the case for NH3.
2. Dynamic Chambers
Closed dynamic chambers have been used in the laboratory and in the field.
Air is drawn through the chamber, and the rate of NH3 emission is determined
from the NH3 enrichment of the air stream, using Eq. (25). In laboratory studies,
the surface of the soil in the chambers is generally about 0.01– 0.04 m2 and the
chambers are closed continuously (Kissel et al., 1977; Bacon et al., 1986;
Sommer and Ersbøll, 1996), whereas chambers with lids that automatically close
during short intervals of NH3 emission measurements have been used in field
studies (Kissel et al., 1977). The wind tunnel system described by Lockyer (1984)
is an example of a large dynamic, open chamber covering a surface area of about
1 m2. It employs a fan to draw air over the treated area. Emission (F) from the
area is calculated from:
F ¼ ðNH3;o 2 NH3;i Þv ð25Þ
where NH3,o and NH3,i are the NH3 concentrations in the outlet and inlet
air, respectively, and v is the volume of air flowing through the tunnel over
the sampling period. Average NH3 recovery of between 74 and 90% has been
found for wind tunnel systems (Sommer et al., 1991; van der Weerden et al.,
1996). It is suggested that the NH3 trapping efficiency of wind tunnel systems
should be checked on a regular basis to avoid errors in measurement.
As discussed already, emissions from the small plots covered by chambers
might be higher than those from a field to which fertilizers have been applied.
Ryden and Lockyer (1985) showed that NH3 emission measured with a wind
tunnel adjusted to the wind speed at 10 cm height in the open was 5% higher than
the emission measured with the micrometeorological mass balance technique.
Usually wind speed in small laboratory chambers is made deliberately high and
emissions correspondingly represent maximum losses (Kissel et al., 1977; Bacon
et al., 1986).
C. MICROMETEOROLOGICAL METHODS
The concept of this approach is to measure NH3 emissions from large, open
experimental areas, which can be plots or entire fields. Micrometeorological
methods have the great advantage of not being intrusive, and of integrating across
heterogeneities in the experimental area. They include mass balance methods,
gradient diffusion approaches, eddy correlation, and relaxed eddy accumulation
techniques, and methods based on Lagrangian dispersion.
1. Mass Balance Methods
These are probably the most widely used techniques for measuring NH3
emissions from larger plots manured with mineral fertilizers. This technology
does not affect emissions from the plots to which fertilizer has been applied, but it
should be remembered that emission from a small plot might be higher than from
a large field when scaling up emissions from the former. As discussed earlier in
the chapter, whether or not plot size is important in this respect depends on the
boundary conditions at the surface. Intercalibration studies have shown that
different mass balance techniques give similar results (Schjoerring et al., 1992;
Sommer et al., 1995; Wood et al., 2000; Sherlock et al., 2002).
2. Integrated Horizontal Flux (IHF) Methods
This mass balance method equates the loss of NH3 from the surface of a
treated plot with the difference between the amount of NH3 carried off the plot by
the wind and the amount carried on to it (Denmead et al., 1977; Denmead, 1983;
Wilson et al., 1983). It calculates the average surface flux density of NH3 in the
treated area, Fg, from the difference in the horizontal fluxes of NH3 across
downwind and upwind boundaries:
1 ðzp
Fg ¼ u ðxd 2 xu Þdz ð26Þ
X z0
where X is the fetch (the distance traveled by the wind across the plot), u is
horizontal wind speed, and xd and xu are the downwind and upwind
atmospheric NH3 concentrations. The integration limit zp is the height at which
the NH3 concentration is at background level. It should be noted that the
integral in Eq. (26) is in terms of instantaneous values of u and the NH3
concentrations. The integral is usually evaluated with mean wind speeds and
mean concentrations. This neglects turbulent terms implicit in the transport
equation, Eq. (26), and results in an overestimation of the true flux by 5 –15%
depending on the geometry (Raupach and Legg, 1984; Leuning et al., 1985;
Wilson and Shum 1992; R.L. Desjardins et al., unpublished, 2003). In most
studies, circular plots and making the “downwind” measurements at the plot
center are used. The wind will always blow towards the center regardless of
wind direction and the fetch will always be the same, viz., the plot radius.
Radii from 3.5 to 40 m have been employed (Beauchamp et al., 1978; Gordon
et al., 1988).
A large simplification in technique for circular plots follows from
developments by Wilson et al. (1982) and Denmead (1983), showing that for a
given surface roughness and plot radius, there exists one particular height of
measurement (ZINST) where the horizontal flux density is in a fixed ratio to the
vertical flux density, regardless of atmospheric stability. The emission can thus
be inferred from measurements of wind speed and atmospheric NH3
concentration (x) at a single height above the ground (Wilson et al., 1983).
The flux F is calculated from:
u £ x
Fg ¼ ð27Þ
Z
where ū and x̄ are the mean wind speed and mean NH3 concentration,
respectively, measured at ZINST. The term Z is the normalized horizontal flux
(ūx̄/F0), which is given in the form of nomograms for different surface
roughnesses and plot radii by Wilson et al. (1982). The method offers
considerable savings in labor and equipment, with results comparable to those
obtained with the IHF method outlined above. This approach has been made even
simpler by the development of passive samplers that measure the horizontal flux
directly (Leuning et al., 1985; Sherlock et al., 1995). These require no power, no
pumps, no anemometers, no data-loggers. A single sampler mounted at ZINST
can return the mean NH3 flux from the treated plot over periods from 1 day to
several weeks (for details see Freney et al., 1992).
3. Theoretical Profile Shape — Philip’s Solution (PTPS)
McInnes et al. (1985) have used Philip’s (1959) analysis for a constant flux
boundary condition to quantify the relationship between the surface flux and
horizontal fluxes at any height. Circular geometry is used. Philip’s analysis
predicts the concentration C1 generated at a particular height by a unit surface
flux density F1 for a given wind speed. The true emission (F) is calculated from
the concentration C measured at that height using the equation:
C
F¼ £ F1 ð28Þ
C1
Calculation of the fluxes at various heights requires measurements of air
temperature and wind speed at two heights, as well as soil temperature and
atmospheric stability. An advantage of the method is that it allows fluxes to be
calculated from measurements at a height where the error to magnitude of
measurement is smallest, but it is based on a constant flux surface condition,
which may not always be appropriate.
4. Perimeter Profile Method
The perimeter profile method is another mass balance method that employs
four masts placed perpendicular to each other around the perimeter of an
experimental area (Schjoerring et al., 1992). Arrays of flux samplers (Ferm,
1991) are mounted in pairs on masts around the boundary of a circular
experimental area. The horizontal fluxes of the inward and outward pointing
tubes are determined separately for each of several heights on each mast. The
vertical flux of NH3 is then determined by stepwise summation of the difference
between the inward and outward facing horizontal fluxes. This technique is
laborious but has the advantage that there is no demand for a homogeneous
surface around the plot to which fertilizer is applied. Denmead et al. (1998)
describe a somewhat similar technique in which air is sampled at several heights
along the full length of each boundary. It is designed particularly for situations in
which there are scattered point sources, such as grazed pastures where NH3 is
emitted from scattered dung and urine patches.
D. GRADIENT DIFFUSION METHODS
The vertical transport of gases in the surface layer is described by:

Fg ¼ 2Kg ›x=›z ð29Þ
where Fg is flux of NH3, x̄ is its mean concentration over a sampling period long
enough to encompass all the significant transporting eddies, and Kg is an eddy or
turbulent diffusivity for gases. Kg is determined by wind speed, height,

aerodynamic roughness and atmospheric stability. Equation (29) leads to
practical methods for calculating gas fluxes from knowledge of the flux of a
tracer gas above the surface and the corresponding gradients or differences in
mean atmospheric concentrations of the tracer and the gas. Aerodynamic
methods use momentum as the tracer flux and horizontal wind speed as the tracer
“concentration”. Energy balance methods use available energy as the tracer flux
and a linear function of temperature and humidity as the “concentration”. Both
methods are appropriate for large surface sources with fetches of hundreds of
meters, and have been widely used for measuring NH3 flux to and from crops.
1. Aerodynamic Methods
The basic equations are given in Appendix. If measurements are made at

only two heights, the gas flux is calculated from:
k2 ðu2 2 u 1 Þðx1 2 x2 Þ
Fg ¼ ð30Þ
{ln½ðz2 2 dÞ=ðz1 2 dÞ 2 ½c1 ðz2 2 dÞ 2 c1 ðz1 2 dÞ}{ln½ðz2 2 dÞ=ðz1 2 dÞ 2 ½c2 ðz2 2 dÞ 2 c2 ðz1 2 dÞ}
In Eq. (30), the subscripts denote the two measuring heights, the overbars
denote means over a suitable measuring period such as 20 or 30 min, and c1 and
c2 are corrections for stability effects, given explicitly in Appendix.
Aerodynamic methods have been much used for measuring NH3 fluxes from
crops, soils, and water bodies (e.g., Denmead et al., 1978; Denmead, 1983;
Harper et al., 1983, 2000; Sutton et al., 1993a,b, 2000; Genermont et al., 1998;
Griffith and Galle, 2000).
2. Energy Balance Methods
These methods are also discussed in Appendix. For concentrations measured

at two heights, the gas flux is calculated from:
ðRn 2 G0 2 SÞðx1 2 x2 Þ
Fg ¼ ð31Þ
rcp ½ðT 1 2 T 2 Þ þ ðe1 2 e 2 Þ=g
where Rn is the net radiation receipt at the surface (incoming short- and long-
wave radiation minus reflected and re-emitted radiation), G0 is the flux density of
heat into the soil at its surface, S is a storage term, T and e are air temperature and
vapor pressure, and g is the psychrometric constant.
Advantages of the method over the aerodynamic approach are that it does not
require calculation of z0 or d and is applicable in all stability conditions. As well,
the basic measurements can provide a measurement of evaporation rate, which
often has an important influence on NH3 production and emission. Its

disadvantage is that it can give erroneous fluxes at night because both Rn and
G are then small and difficult to measure, and net radiometers may not function
effectively since their domes can become covered with dew. The same problems
can arise in rainy periods. Aerodynamic methods are usually more reliable at
night. Like the aerodynamic approach, the energy balance method is much used
for measuring NH3 fluxes on a field basis. Some example applications are
provided by Denmead et al. (1982, 2003) and Freney et al. (1992). It should be
noted that sensors mounted at different heights above the surface have different
fetches, so that problems arise in the use of gradient diffusion methods if the
surface fluxes are spatially heterogeneous.
E. EDDY CORRELATION
This is the preferred micrometeorological method for measuring scalar fluxes,

because it is a direct measurement requiring no simplifying assumptions (about
similarity between the Ks) and no stability corrections, and it gives the vertical
flux at the point of measurement. There are no problems with different footprints
for different measurement levels as there are for gradient-diffusion methods. The
basic method requires simultaneous measurements of the vertical wind speed w
(usually by a sonic anemometer) and the gas concentration x at a sampling
frequency rapid enough to catch all the significant eddies, say 10 Hz. The
instantaneous vertical gas flux density ¼ wx, and the mean flux density is the
average of wx over a period long enough to encompass all the effective
transporting eddy sizes. There are problems in using the method with some trace
gases because it requires fast gas measurement (10 Hz sampling), which is not
yet possible for many trace gases, and density effects caused by simultaneous
fluxes of sensible and latent heat may necessitate large corrections to the apparent
flux (Webb et al., 1980). Recently, however, a fast-response gas analyzer
employing a tunable diode laser has been developed for measuring atmospheric
NH3 concentrations. Further, calculations by Wesely et al. (1989) indicate that
for NH3, the corrections for density effects will usually be negligible, so that eddy
correlation measurements of NH3 fluxes are likely to become more common.
F. RELAXED EDDY ACCUMULATION OR CONDITIONAL

SAMPLING
The relaxed eddy accumulation method retains many of the advantages of

eddy correlation inasmuch as it is a point measurement and requires no stability
corrections. In addition, it does away with the need for rapid gas measurement. It
still requires rapid measurement of the vertical wind speed w, but substitutes fast
solenoid valves for a fast-response gas analyzer. The anemometer controls a

simple valving system. Air is sampled at a point at a constant rate and is diverted
via the valves into “up” and “down” bins depending on the direction of the
vertical wind. The contents of the bins are then analyzed for the mean gas
concentrations of the ascending and descending eddies, but the analyzer can be of
slow response and does not have to be mounted in the field, or denuders may be
used to trap the NH3 in the air streams (Zhu et al. 2000). The flux of NH3 is
calculated from (Businger and Oncley, 1990):
F ¼ bsððNHþ þ þ 2
3 Þ 2 ðNH3 Þ Þ ð32Þ
where b is a coefficient close to 0.6, sw is the standard deviation of w, and

(NHþ3)
þ
and (NHþ3)
2
are the respective average NH3 concentrations of the
ascending and descending eddies (Businger and Oncley, 1990).
G. LAGRANGIAN DISPERSION MODELS
Lagrangian dispersion analyses adopt a coordinate system that travels with the
dispersing entity. This is in contrast to the Eulerian analyses that we have
described so far in this chapter, except for the trajectory-simulation model of
Wilson et al. (1982). Eulerian dispersion uses a fixed coordinate system and
considers the passage of scalars at a point fixed in space.
1. Backward Lagrangian Stochastic Dispersion Model
This model allows predictions of the strength of any surface source from one-
point measurements of wind speed and concentration at any location downstream
(Flesch et al., 1995). The plot geometry and the location of the measuring point
relative to the plot can be quite arbitrary, but must be known. The model
calculates trajectories of air parcels backward in time from the sensor location to
the source. It employs computer-simulated gas releases (about 10,000) to relate
the surface flux density F0 to mean concentrations developed at specified heights
and distances downwind. Required input information is z0, ū and gas
concentration xg,z in excess of background at one particular height and distance
downwind, plus atmospheric stability. Solutions have the form:
F0 ¼ nuz xg;z ð33Þ
where n is a coefficient calculated for the particular situation by the model. This
method differs from the ZINST approach because it caters for any geometry and
measurements can be made at any height downwind, but the stability needs to be
known. Examples of its use are given by McGinn and Janzen (1998).
2. Inverse Lagrangian Analysis
Soils and plants can be both sources and sinks for NH3, so that measurements
of the net NH3 exchange by a plant community will usually be insufficient for
full understanding of ecosystem functioning. Additional information will be
needed on the strengths and locations of the canopy sources and sinks. Recent
research, notably by Raupach (1989a,b,c), has led to the development of a
micrometeorological tool known as Inverse Lagrangian Analysis that allows the
identification of the sites of gas exchange in plant canopies in a non-disturbing,
continuous way from relatively simple observations of concentrations and
turbulence parameters within and above plant canopies. It is beyond the scope
of the present review to examine the analysis in any detail, but examples of
its application to NH3 exchange in crops of corn, oilseed rape, and sugarcane
can be found in Harper et al. (2000), Nemitz et al. (2000), and Denmead
et al. (2003).
VII. CONCLUSIONS AND PERSPECTIVES
The emission of NH3 from inorganic N fertilizers varies considerably,

depending on soil properties, fertilizer type and atmospheric conditions.
Particularly the pH and pH buffer capacity of the soil and dissolving fertilizer
salts may have a large impact on the NH3 emission. A simple classification
tool that can be useful in simple decision support systems may be to classify
soils as acid soils, calcareous soils, neutral soils, etc. and fertilizers as,
e.g., alkaline, acid fertilizers, and fertilizers at risk of causing precipitates
with Ca.
The effect of temperature, wind and rain may interact and these
interactions should be considered when estimating NH3 emission potentials.
Thus, a high wind speed on days with high radiation may on the one hand
reduce soil surface temperature, while on the other hand increase atmospheric
NH3 transport, resulting in no net effect on NH3 emission. Temperature may
significantly affect NH3 emission, but the total amount lost will not always
increase, because the emission may be higher for a longer period at low than
at high temperatures. Consequently, simple models relating emission to air
temperature and wind speed may be erroneous, and more complex models
will have to be developed, including calculation of surface temperature as
affected by global radiation and wind. The effect of rain on NH3 emission
from urea depends on timing and intensity — a light shower immediately
after application may increase emission and heavy rain will reduce emission.
Ammonia emission directly from the foliage of crops may on some, but not all,
occasions be related to N fertilizer application. There are indications that
seedlings rapidly absorbing N via the roots may emit NH3 because internal N
assimilation is too slow relative to N uptake. At later growth stages, the leaves
may absorb NH3 emitted from fertilizer on the soil surface beneath the canopy,
while during grain filling NH3 may again be emitted from the foliage as N is
remobilised from senescing leaves. Biological activity in the soil associated with
nitrification and N-immobilization may affect fertilizer NH3 emissions via effects
on pH and NHþ 4 concentration. In models of ammonia losses from fertilized
fields, both canopy –atmosphere NH3 exchange and nitrification should be
included.
Incorporation of fertilizers is among the most efficient techniques for reducing
NH3 emission from dry soil. Splitting fertilizer applications so that the canopy
more efficiently will be able to capture NH3 may also contribute to reduced
emissions. Additives may influence loss of NH3 from applied fertilizers but their
efficiency may be variable, e.g., depending on rain transporting urea into the soil.
Furthermore, additives inhibiting urea hydrolysis are costly and additives
reducing pH may either be a safety risk to farmers or may increase the demand for
lime. The trend towards no-till crop production will increase the area of land
covered with trash on which fertilizers are subsequently applied. Ammonia
emission will be high from these fields and techniques should be developed to
reduce NH3 emission without increasing soil tillage.
Dynamic modeling of NH3 emission in relation to soil properties is
challenging due to the many interacting processes. The physico-chemical and
biological processes (pH, hydrolysis, cation adsorption, precipitation, nitrifica-
tion etc.) affecting the gradient of ammonium and pH between the fertilizer
granules and the soil may vary spatially and be difficult to predict. Thus, mixed
models using empirical algorithms and mechanistic submodels may prove useful
for the purpose of developing reliable emission-models with an acceptable and
realistic need for input data. Atmospheric turbulence equations may predict
transport of NH3 from a bare soil surface to the atmosphere, but may fail to
describe NH3 emission from fertilizers on the soil beneath a plant canopy, i.e., the
models may not precisely take into account how the canopy will change the
vertical wind speed profiles and provide shade. Moreover, crops may absorb
NH3, with the assimilation being related to NH3 concentration.
Data are needed for validating models of ammonia emission and for the
purpose of establishing relationship between emission and the most important
emission factors. In the processing of data, systematic biases due to effect of
measurement technique and plot size must be taken into account, i.e., NH3
emission from small plots may be too high compared with emissions measured on
a large field scale. Therefore, results from small dynamic chamber studies should
only be used qualitatively when assessing the effects of different soil, fertilizer or
climatic factors on emission patterns. This bias should be corrected when scaling
up the results to field scale or when developing decision support systems and
calculating national inventories.
APPENDIX A. DERIVATION OF OPERATING

FORMULAE FOR MEASURING GAS FLUXES BY
AERODYNAMIC AND ENERGY BALANCE METHODS
1. ATMOSPHERIC STABILITY
This Appendix is prefaced with a short explanation about atmospheric stability.

Stability is concerned with the relative effects of buoyancy (positive or negative)
and wind shear on transport and mixing in the atmosphere. Atmospheres are
described as being neutral, stable, or unstable depending on buoyancy. In neutral
conditions, which occur for short periods around sunrise and sunset, there is
essentially no temperature gradient and hence no buoyancy. Parcels of air
displaced upwards are at the same temperature as their surroundings and are
neither accelerated nor decelerated. Transport is subject only to mechanical
turbulence. In stable conditions, which usually occur at night, the temperature
gradient is positive (air is cooler closer to the ground) and parcels displaced
upwards are heavier than their surroundings, so their motion is decelerated.
Transport is inhibited. In unstable conditions, the usual daytime condition, the
temperature gradient is negative (air is warmer closer to the ground) and parcels
displaced upwards are lighter than their surroundings and so their motion is
accelerated. Transport is enhanced. A common index of stability is the Monin –
Obukhov length, L, which can be calculated from measurements of the
temperature and wind speed gradients or from direct measurements of the heat
and momentum fluxes (Paulson, 1970).
2. AERODYNAMIC METHOD
The methodology has been described in a number of reviews (e.g., Thom,

1972; Monteith and Unsworth, 1990; Denmead, 1994), and only summary results
are given here. The general formula for calculating the gas flux from wind speed
and concentration measurements is

›u ›x ›u ›x
k 2 z2 k2
›z ›z ›lnz ›lnz
Fg ¼ ¼ ðA1Þ
wm w g wm wg
where the symbols are as used previously and wm and wg are functions to account
for the effects of atmospheric stability on momentum and gas transport,
respectively, and are given explicitly in the references above.
If wind speeds and concentrations have been measured at a number of
points in the surface boundary layer, their gradients can be calculated and
Eq. (A1) evaluated directly. Usually, however, wind speeds and concentrations
are measured at only two levels in the boundary layer. Then, Eq. (A1)
becomes:
k2 ðu2 2 u 1 Þðx1 2 x2 Þ
Fg ¼ ðA2Þ
{ln½ðz2 2dÞ=ðz1 2dÞ2½cm ðz2 2dÞ2 cm ðz1 2dÞ}{ln½ðz2 2dÞ=ðz1 2dÞ2½cg ðz2 2dÞ2 cg ðz1 2dÞ}
where cm and cg are integrated forms of the stability functions, wm and wg. In
neutral conditions:
cm ¼ cg ¼ 0 ðA3Þ
in stable conditions,
cm ¼ cg ¼ 25ðz2dÞ=L ðA4Þ
and in unstable conditions,

cm ¼ 2ln½ð1þxÞ=2þln½ð1þx2 Þ=222tan21 xþ p=2 ðA5Þ
and
cg ¼ 2ln½ð1þx2 Þ=2 ðA6Þ
where
x ¼ ð1216z=LÞ1=4 ðA7Þ
If the friction velocity up is available from eddy correlation measurements, an
alternative aerodynamic formulation is:
ku* ðx1 2 x2 Þ
Fg ¼ ðA8Þ
ln½ðz2 2dÞ=ðz1 2dÞ2½cg ðz2 2dÞ2 cg ðz1 2dÞ
3. ENERGY BALANCE METHOD
The energy balance at the surface is

Rn ¼ H þ lE þ G0 þ S; ðA9Þ
where Rn is the net radiation receipt at the surface (incoming short- and long-
wave radiation less reflected and re-emitted radiation), H is the flux density of
sensible heat from surface to atmosphere, lE is the flux density of latent heat, l
being the latent heat of vaporization of water and E the surface evaporation rate,
G0 is the flux density of heat into the soil at its surface, and S is a storage term,
e.g., the change in heat stored in water bodies or the air or biomass of a plant
community, or the solar energy fixed in photosynthesis. Rearranging Eq. (A9)
leads to
H þ lE ¼ Rn 2 G0 2 S ðA10Þ
Following Eq. (17),

›z
H ¼ 2Kh rcp ›T= ðA11Þ
lE ¼ 2Kg ðrcp =gÞ›e =›z ðA12Þ
and
Fg ¼ 2Kg ›x=›z ðA13Þ
where most of the symbols are as defined previously, Kh and Kg are the eddy
diffusivities for heat and gases or vapors respectively, e is vapor pressure, and
g is the psychrometric constant (< 66 Pa K21). Much micrometeorological
research has shown that in most field situations, Kh ¼ Kg. On that assumption,
ðRn 2 G0 2 SÞ›x=›z
Fg ¼ ðA14Þ
›z þ ð›e =›zÞ=g
rcp ½›T=
or, for measurements at two levels,
ðRn 2 G0 2 SÞðx1 2 x2 Þ
Fg ¼ ðA15Þ
rcp ½ðT 1 2 T 2 Þ þ ðe1 2 e 2 Þ=g
Thus, H, lE and Fg can be determined from measurements of the energy
fluxes at the surface and differences in temperature, humidity, and gas con-
centration between two heights above the surface. All the terms in Eq. (A15) can
be measured directly with appropriate instrumentation. Advantages of the
method over the aerodynamic approach are that it does not require calculation of
z0 and d and is applicable in all stability conditions.
ACKNOWLEDGMENT
Hydro Agri supported this review financially.
REFERENCES
Al-Kanani, T., MacKenzie, A. F., and Barthakur, N. N. (1991). Soil water and ammonia volatilization
relationships with surface-applied nitrogen fertilizer solutions. Soil Sci. Soc. Am. J. 55,
1761–1766.
Asman, W. A. H., and van Jaarsveld, H. A. (1991). A variable resolution transport model applied for
NHx in Europe. Atmos Environ. 26A, 445 –464.
Asman, W. A. H., Sutton, M. A., and Schjørring, J. K. (1998). Ammonia: emission, atmospheric
transport and deposition. New Phytol. 139, 27 –48.
Atwell, B. J., Fillery, I. R. P., McInnes, K. J., and Smucker, A. J. M. (2002). The fate of carbon and
fertilizer nitrogen when dryland wheat is grown in monoliths of duplex soil. Plant Soil 241,
259 –269.
Avnimelech, Y., and Laher, M. (1977). Ammonia volatilization from soils: equilibrium
considerations. Soil Sci Soc Am J 41, 1080–1084.
Bacon, P. E., and Freney, J. R. (1989). Nitrogen loss from different tillage systems and the effect on
cereal grain yield. Fert Res. 20, 59–66.
Bacon, P. E., Hoult, E. H., and McGarity, J. W. (1986). Ammonia volatilization from fertilizers
applied to irrigated wheat soils. Fert. Res. 10, 27–42.
Baker, J. H., Peech, M., and Musgrave, R. B. (1959). Determination of application losses of anhydrous
ammonia. Agron. J. 51, 361 –362.
Beauchamp, E. G., Kidd, G. E., and Thurtell, G. (1978). Ammonia volatilisation from sewage sludge
applied in the field. J. Environ. Qual. 7, 141–146.
Black, A. S., Sherlock, R. R., Smith, N. P., Cameron, K. C., and Goh, K. M. (1985). Effects of form of
nitrogen, season, and urea application rate on ammonia volatilization from pastures. N. Z. J.
Agric. Res. 28, 469– 474.
Black, A. S., Sherlock, R. R., and Smith, N. P. (1987a). Effect of urea granule size on ammonia
volatilization from surface-applied urea. Fert. Res. 11, 87 –96.
Black, A. S., Sherlock, R. R., and Smith, N. P. (1987b). Effect of timing of simulated rainfall on
ammonia volatilization from urea, applied to soil of varying moisture content. J. Soil Sci. 38,
679–687.
Blaise, D., Tyagi, P. C., and Khola, O. P. S. (1996). Ammonia volatilisation from urea as affected by
the addition of iron pyrites and method of application. Nutr. Cycl. Agroecosyst. 46, 97–101.
Blue, W. G., and Eno, C. F. (1954). Distribution and retention of anhydrous ammonia in sandy soils.
Soil Sci. Soc. Am. Proc. 18, 420–424.
Bouwman, A. F., Lee, D. S., Asman, W. A. H., Dentener, F. J., Van Der Hoek, K. W., and Olivier, J. G.
J. (1997). A global high-resolution emission inventory for ammonia. Global Biogeochem. Cycl.
11, 561 –587.
Bouwman, A. F., Boumans, L. J. M., and Batjes, N. H. (2002). Estimation of global NH3 volatilization
loss from synthetic fertilizers and animal manure applied to arable lands and grasslands. Global
Biogeochem. Cycl. 16(2) (DOI 10.1029/2000GB001389).
Bouwmeester, J. B., and Vlek, P. L. G. (1981). Wind-tunnel simulation and assessment of ammonia
volatilization from ponded water. Agron. J. 73, 546– 552.
Bouwmeester, R. J. B., Vlek, P. L. G., and Stumpe, J. M. (1985). Effect of environmental factors on
ammonia volatilization from a urea-fertilized soil. Soil Sci. Soc. Am. J. 49, 376–381.
Bowmer, K. H., and Muirhead, W. A. (1987). Inhibition of algal photosynthesis to control pH and
reduce ammonia volatilization from rice floodwater. Fert. Res. 13, 13 –29.
Bremner, J. M. (1995). Recent research on problems in the use of urea as a nitrogen fertilizer. Fert.
Res. 42, 321 –329.
Bremner, J. M., and Douglas, L. A. (1971). Decomposition of urea phosphate in soils. Soil Sci. Soc.
Am. J. 55, 1761–1766.
Brouquisse, R., Evrard, A., Rolin, D., Raymond, P., and Roby, C. (2001). Regulation of protein
degradation and protease expression by mannose in maize root tips: Pi sequestration by mannose
may hinder the study of its signaling properties. Plant Physiol. 125, 1485–1498.
Brutsaert, W. (1982). Evaporation into the Atmosphere. Reidel, Dordrecht, The Netherlands.
Buijsman, E., Maas, J. F., and Asman, W. A. H. (1987). Anthropogenic NH3 emission in Europe.
Atmos. Environ. 21, 1009–1022.
Businger, J. A., and Oncley, S. P. (1990). Flux measurement with conditional sampling. J. Atmos.
Ocean Technol. 7, 349–352.
Bussink, D. W., Huijsmans, J. F. M., and Ketelaars, J. J. M. H. (1994). Ammonia volatilisation
from nitric-acid-treated cattle slurry surface applied to grassland. Neth. J. Agric. Sci. 42,
293–309.
Byrnes, B. H., and Freney, J. R. (1995). Recent developments on the use of urease inhibitors in the
tropics. Fert. Res. 42, 251–259.
Christianson, C. B., Carmona, G., Klein, M. O., and Howard, R. G. (1995). Impact on ammonia
volatilisation losses of mixing KCl of high pH with urea. Fert. Res. 40, 89– 92.
Chung, J.-B., and Zasoski, R. J. (1994). Ammonium–potassium and ammonium–calcium exchange
equilibria in bulk and rhizosphere soil. Soil Sci. Soc. Am. J. 58, 1368– 1375.
De Datta, S. K., Trevitt, A. C. F., Freney, J. R., Obcemea, W. N., Real, J. G., and Simpson, J. R.
(1989). Measuring nitrogen losses from lowland rice using bulk aerodynamic and nitrogen-15
balance methods. Soil Sci. Soc. Am. J. 53, 1275– 1281.
De Datta, S. K., Buresh, R. J., Sampson, M. I., Obcemea, W. N., and Real, J. G. (1991). Direct
measurement of ammonia and denitrification fluxes from urea applied to rice. Soil Sci. Soc. Am. J.
55, 543–548.
de Jonge, L. W., Sommer, S. G., and Jacobsen, O. H. (2003). Infiltration of and ammonia volatilization
from pig and cattle slurry applied to harrowed and stubble soils. J. Environ. Qual. (In
preparation).
Denmead, O. T. (1979). Chamber systems for measuring nitrous oxide emission from soils in the field.
Soil Sci. Soc. Am. J. 43, 89–95.
Denmead, O. T. (1983). Micrometeorological methods for measuring gaseous losses of nitrogen in the
field. In “Gaseous Loss of Nitrogen from Plant–Soil Systems” (J. R. Freney and J. R. Simpson,
Eds.), pp. 133– 158, Martinus Nijhoff/Dr W Junk, The Hague.
Denmead, O. T. (1994). Measuring fluxes of CH4 and N2O between agricultural systems and the
atmosphere. In “CH4 and N2O: Global Emissions and Controls from Rice Fields and other
Agricultural and Industrial Sources” (K. Minami, A. Mosier, and R. Sass, Eds.), pp. 209–234.
National Institute of Agro-Environmental Sciences, Tsukuba, Japan.
Denmead, O. T., and Freney, J. R. (1992). Transfer coefficients for water–air exchange of ammonia,
carbon dioxide and methane. Ecol. Bull. 42, 31–41.
Denmead, O. T., Simpson, J. R., and Freney, J. R. (1977). A direct field measurement of ammonia
emission after injection of anhydrous ammonia. Soil Sci. Soc. Am. J. 41, 1001–1004.
Denmead, O. T., Freney, J. R., and Simpson, J. R. (1982). Dynamics of ammonia volatilisation during
furrow irrigation of maize. Soil Sci. Soc. Am. J. 46, 149–155.
Denmead, O. T., Harper, L. A., Freney, J. R., Griffith, D. W. T., Leuning, R., and Sharpe, R. R. (1998).
A mass balance method for non-intrusive measurements of surface–air gas exchange. Atmos.
Environ. 32, 357–369.
Denmead, O. T., Raupach, M. R., Leuning, R., Dunin, F. X., and Freney, J. R. (2003). Inverse
Lagrangian analysis of heat, vapor and gas exchange in plant canopies. In “Micrometeorological
Measurements in Agricultural Systems” (E. J. Sadler, J. M. Ham, B. Tanner, J. M. Baker, and J. L.
Hatfield, Eds.), American Society of Agronomy, Madison, (in press).
Du Preez, C. C., and Burger, R. du T. (1988). Ammonia losses from ammonium-containing
and — forming fertilizers after surface application at different rates on alkaline soils. Fert. Res.
15, 71–78.
ECETOC. (1994). Ammonia Emissions to Air in Western Europe, Technical Report No. 62. European
Centre for Ecotoxicology and Toxicology of Chemicals, Brussels, Belgium.
EEA. (1999). Overview of National Programmes to Reduce Greenhouse Gas Emissions. European
Environmental Agency, Final version, April 1999.
EFMA. (1997). The Fertilizer Industry of the European Union: The Issues of Today, the Outlook for
Tomorrow. European Fertilizer Manufacturers Association, Brussels, Belgium.
Evangelou, V. P. (1990). Regulation mechanisms and field implications of ammonia bonding with
various crystalline salts. Soil Sci. Soc. Am. J. 54, 394 –398.
Fan, M. X., MacKenzie, A. F., and Blenkhorn, H. D. (1996). A modified urea based NP fertilizer:
Urea — TSP–MAP combinations. Fert. Res. 45, 217–220.
Farquhar, G. D., Firth, P. M., Wetselaar, R., and Weir, B. (1980). On the gaseous exchange of
ammonia between leaves and the environment: determination of the ammonia compensation
point. Plant Physiol. 66, 710–714.
Fenn, L. B., and Hossner, L. R. (1985). Ammonia volatilization from ammonium or ammonium-
forming nitrogen fertilizers. Adv. Soil Sci. 1, 124 –169.
Fenn, L. B., and Kissel, D. E. (1973). Ammonia volatilization from surface applications of ammonium
compounds on calcareous soils. I. General theory. Soil Sci. Soc. Am. Proc. 37, 855–859.
Fenn, L. B., and Kissel, D. E. (1974). Ammonia volatilization from surface applications of ammonium
compounds on calcareous soils. I. Effects of temperature and rate of ammonium nitrogen
application. Soil Sci. Soc. Am. Proc. 38, 606 –610.
Fenn, L. B., and Kissel, D. E. (1976). The influence of cation exchange capacity and depth of
incorporation on ammonia volatilization from ammonium compounds applied to calcareous soils.
Soil Sci. Soc. Am. Proc. 40, 394–398.
Fenn, L. B., and Miyamoto, S. (1981). Ammonia loss and associated reactions of urea in calcareous
soils. J. Soil Sci. Soc. Am. 45, 537–540.
Fenn, L. B., and Richards, J. (1986). Ammonia loss from surface applied urea-acid products. Fert. Res.
9, 265– 275.
Fenn, L. B., Taylor, R. M., and Matocha, J. E. (1981). Ammonia losses from surface applied nitrogen
fertilizer as controlled by soluble calcium and magnesium: general theory. Soil Sci. Soc. Am. J.
45, 777 –781.
Fenn, L. B., Matocha, J. E., and Wu, E. (1982). Substitution of ammonium and potassium for added in
reduction of ammonia loss from surface-applied urea. Soil Sci. Soc. Am. J. 46, 771–776.
Fenn, L. B., Malstrom, H. L., and Wu, E. (1987). Ammonia losses from surface-applied urea as
related to urea application rates, plant residue and calcium chloride addition. Fert. Res. 12,
219–227.
Ferguson, R. B., Kissel, D. E., Koelliker, J. K., and Basel, W. (1984). Ammonia volatilization from
surface applied urea: effect of hydrogen ion buffering capacity. Soil Sci. Soc. Am. J. 48, 578– 582.
Ferm, M. (1991). A sensitive diffusional sampler. Report B-1020, Swedish Environmental Research
Institute, Gothenburg.
Fillery, I. R. P., and De Datta, S. K. (1986). Ammonia volatilization from nitrogen-sources applied to
rice fields. 1. Methodology, ammonia fluxes, and N-15 loss. Soil Sci. Soc. Am. J. 50, 80–86.
Fillery, I. R. P., Simpson, J. R., and De Datta, S. K. (1984). Influence of field environment and
fertilizer management of ammonia loss from flooded rice. Soil Sci. Soc. Am. J. 48, 915–920.
Fillery, I. R. P., Roger, P. A., and De Datta, S. K. (1986). Ammonia volatilization from nitrogen
sources applied to rice fields. II. Floodwater properties and submerged photosynthetic biomass.
Soil Sci. Soc. Am. J. 50, 86 –91.
Finnemann, J., and Schjoerring, J. K. (1999). Translocation of NHþ 4 in oilseed rape plants in relation to
glutamine synthetase isogene expression and activity. Physiol. Plant 105, 469– 477.
Fleisher, Z., Kenig, A., Ravina, I., and Hagin, J. (1987). Model of ammonia volatilization from
calcareous soils. Plant Soil 103, 205–212.
Flesch, T., Wilson, J. D., and Yee, E. (1995). Backward time Lagrangian stochastic dispersion models,
and their application to estimate gaseous emissions. J. Appl. Meteorol. 34, 1320– 1332.
Flowers, T. H., and O’Callaghan, J. R. (1983). Nitrification in soils incubated with pig slurry or
ammonium sulphate. Soil Biol. Biochem. 15, 337 –342.
Francis, D. D., Schepers, J. S., and Vigil, M. F. (1993). Post-anthesis nitrogen loss from corn. Agron. J.
85, 659 –663.
Francis, D. D., Schepers, J. S., and Sims, A. L. (1997). Ammonia exchange from corn foliage during
reproductive growth. Agron. J. 89, 941–946.
Freney, J. R., Denmead, O. T., Watanabe, I., and Craswell, E. T. (1981). Ammonia and nitrous oxide
losses following applications of ammonium sulphate to flooded rice. Aust. J. Agric. Res. 32,
37–45.
Freney, J. R., Simpson, J. R., and Denmead, O. T. (1983). Volatilization of ammonia. In “Gaseous
Loss of Nitrogen from Plant –Soil Systems” (J. R. Freney and J. R. Simpson, Eds.), pp. 1 –32,
Martinus Nijhoff/Dr. W. Junk, The Hague.
Freney, J. R., Denmead, O. T., Sod, A. W., Saffigna, P. G., Chapman, L. S., Ham, G. J., Hurney, A. P.,
and Stewart, R. L. (1992). Factors controlling ammonia loss from trash covered sugarcane fields
fertilized with urea. Fert. Res. 31, 341–349.
Genermont, S., and Cellier, P. (1997). A mechanistic model for estimating ammonia volatilisation
from slurry applied to bare soil. Agric. Forest. Meteorol. 88, 145–167.
Genermont, S., Cellier, P., Flura, D., Morvan, T., and Laville, P. (1998). Measuring ammonia fluxes
after slurry spreading under actual field conditions. Atmos. Environ. 32, 279–284.
Gilmour, J. T. (1984). The effects of soil properties on nitrification and nitrification inhibition. Soil Sci.
Soc. Am. J. 48, 1262–1266.
Goos, R. J., and Cruz, A. D. (1999). Effect of ammonium sulfate pretreatment on ammonia
volatilization after urea fertilization. Commun. Soil Sci. Plant Anal. 30, 1325–1336.
Gordon, R., Leclerc, M. Y., Schuepp, P., and Brunke, R. (1988). Field estimates of ammonia
volatilisation from swine manure by a simple micrometeorological technique. Can. J. Soil Sci. 68,
369 –380.
Grant, R. F. (1994). Simulation of ecological controls on nitrification. Soil Biol. Biochem. 26,
305 –315.
Griffith, D. W. T., and Galle, B. (2000). Flux measurements of NH3, N2O and CO2 using dual beam
FTIR spectroscopy and the flux-gradient technique. Atmos. Environ. 34, 1087–1098.
Hanstein, S., Mattsson, M., Jaeger, H.-J., and Schjoerring, J. K. (1999). Uptake and utilization of
atmospheric ammonia in three native Poaceae species: Leaf conductances, composition of
apoplastic solution and interactions with root nitrogen supply. New Phytol. 141, 71–83.
Harper, L. A. (2003). Ammonia: measurement issues. In “Micrometeorological Measurements in
Agricultural Systems” (E. J. Sadler, J. M. Ham, B. Tanner, J. M. Baker, and J. L. Hatfield, Eds.),
American Society of Agronomy, Madison, (in press).
Harper, L. A., Catchpoole, V. R., Davis, R., and Weir, K. L. (1983). Ammonia volatilization: soil,
plant and microclimate effects on diurnal and seasonal fluctuations. Agron. J. 75, 212–218.
Harper, L. A., Sharpe, R. R., Langdale, G. W., and Giddens, J. E. (1987). Nitrogen cycling in a wheat
crop: soil plant and aerial nitrogen transport. Agron. J. 79, 965–973.
Harper, L. A., Bussink, D. W., van der Meer, H. G., and Corre, W. J. (1996). Ammonia transport in a
temperate grassland. 1. Seasonal transport in relation to soil fertility and crop management.
Agron. J. 88, 614 –621.
Harper, L. A., Denmead, O. T., and Sharpe, R. R. (2000). Identifying sources and sinks of scalars in a
corn canopy with inverse Lagrangian dispersion analysis. II. Ammonia. Agric. For. Meteorol.
104, 75–83.
Harrison, R., and Webb, J. (2001). A review of the effect of N fertilizer type on gaseous emission. Adv.
Agron. 73, 65 –108.
Haynes, R. J., and Williams, P. H. (1992). Changes in soil solution composition and pH in urine-
affected areas of pasture. J. Soil Sci. 43, 323–334.
He, Z. L., Alva, A. K., Calvert, D. V., and Banks, D. J. (1999). Ammonia volatilization from different
fertilizer sources and effects of temperature and soil pH. Soil Sci. 164, 751–758.
Herrmann, B., Mattsson, M., Fuhrer, J., and Schjoerring, J. K. (2002). Leaf –atmosphere NH3
exchange of white clover in relation to mineral N nutrition and symbiotic N2 fixation. J. Exp. Bot.
53, 139–146.
Hill, P. W., Raven, J. A., Loubet, B., Fowler, D., and Sutton, M. A. (2001). Comparison of gas
exchange and bioassay determinations of the ammonia compensation point in Luzula sylvatica
(Huds.) Gaud. Plant. Physiol. 125, 476– 487.
Hirel, B., and Lea, P. J. (2001). Ammonia assimilation. In “Plant Nitrogen” (P. J. Lea and
J.-F. Morot-Gaudry, Eds.), pp. 79 –99. Springer-Verlag, Berlin.
Hopkins, D. F., Wells, G. L., and Butler, B. J. (1963). High-pressure injection of anhydrous ammonia,
Trans. ASAE, Paper no. 62–645 presented at the winter meeting of the American Society of
Agricultural Engineers in Chicago, December 1962, pp. 343–352.
Hoult, E. H., and McGarity, J. W. (1987). The influence of sward mass, defoliation and watering on
ammonia volatilization losses from an Italian ryegrass sward topdressed with urea. Fert. Res. 13,
199–207.
Husted, S., and Schjoerring, J. K. (1995). Apoplastic pH and ammonium concentration in leaves of
Brassica napus L. Plant Physiol. 109, 1453–1460.
Husted, S., and Schjoerring, J. K. (1996). Ammonia fluxes between Brassica napus plants and the
atmosphere in response to changes in leaf temperature, light intensity and air humidity: interactions
with stomatal conductance and apoplastic NHþ þ
4 and H concentrations. Plant Physiol. 112, 67 –74.
Husted, S., Mattsson, M., and Schjoerring, J. K. (1996). Ammonia compensation points in two
cultivars of Hordeum vulgare L. during vegetative and generative growth. Plant Cell Environ. 19,
1299–1306.
Husted, S., Hebbern, C., Mattsson, M., and Schjoerring, J. K. (2000a). A critical experimental
evaluation of methods used for analysing ammonium in plant tissue, xylem sap and apoplastic
fluid. Physiol. Plant. 109, 167 –179.
Husted, S., Schjoerring, J. K., Nielsen, K. H., Nemitz, E., and Sutton, M. A. (2000b). Stomatal
compensation points for ammonia in oilseed rape plants under field conditions. Agric. For.
Meteorol. 105, 371–383.
Husted, S., Mattsson, M., Möllers, C., Wallbraun, M., and Schjoerring, J. K. (2002). Photorespiratory
NHþ 4 production in leaves of wildtype and GS2-antisense oilseed rape. Plant Physiol. 130,
989–998.
Hutchings, N. J., Sommer, S. G., Andersen, J. M., and Asman, W. A. H. (2001). A detailed ammonia
emission inventory for Denmark. Atmos. Environ. 35, 1959– 1968.
IFA Statistics (2002). International Fertilizer Industry Association: http://www.fertilizer.org/ifa/
statistics/STATSIND/tablen.asp.
Ismail, K. M., Wheaton, F. W., Douglass, L. W., and Potts, W. (1991). Modeling ammonia
volatilization from loamy sand soil treated with liquid urea. Trans. ASAE 34, 756–763.
Izaurralde, R. C., Kissel, D. E., and Cabrera, M. L. (1987). Titratable acidity to estimate ammonia
retention. Soil Sci. Soc. Am. J. 51, 1050–1054.
Izaurralde, R. C., Kissel, D. E., and Cabrera, M. L. (1990). Simulation model of banded ammonia in
soils. Soil Sci. Soc. Am. J. 54, 917–922.
Janzen, H. H., and Gilbertson, C. (1994). Exchange of N-15 among plants in controlled environment
studies. Can J. Soil Sci. 74, 109–110.
Kirk, G. J. D., and Nye, P. H. (1991). A model of ammonia volatilization from applied urea. V. The
effects of steady-state drainage and evaporation. J. Soil Sci. 42, 103 –113.
Kissel, D. E., Brewer, H. L., and Arkin, G. F. (1977). Design and test of a field sampler for ammonia
volatilization. Soil Sci. Soc. Am. J. 41, 1133–1138.
Larsen, S., and Gunary, D. (1962). Ammonia loss from ammoniacal fertilisers applied to calcareous
soils. J. Sci. Food Agric. 11, 566 –572.
Leegood, R. C., Lea, P. J., Adcock, M. D., and Häusler, R. E. (1995). The regulation and control of
photorespiration. J. Exp. Bot. 46, 1397– 1414.
Leuning, R., Denmead, O. T., Simpson, J. R., and Freney, J. R. (1984). Processes of ammonia loss
from shallow floodwater. Atmos. Environ. 18, 1583–1592.
Leuning, R., Freney, J. R., Denmead, O. T., and Simpson, J. R. (1985). A sampler for measuring
atmospheric ammonia flux. Atmos. Environ. 19, 1117–1124.
Lightner, J. W., Mengel, D. B., and Rhykerd, C. L. (1990). Ammonia volatilization from nitrogen
fertilizer surface applied to orchardgrass sod. Soil Sci. Soc. Am. J. 54, 1478– 1482.
Lindau, C. W., Reddy, K. R., Lu, W. F., Khind, C. S., Pardue, J. H., and Patrick, W. H. Jr. (1989). Effects
of redox potential on urea hydrolysis and nitrification in soil suspensions. Soil Sci. 148, 184– 190.
Livingston, G. P., and Hutchinson, G. L. (1995). Enclosure-based measurements of trace gas exchange:
application and sources of error. In “Biogenic Trace Gases: Measuring Emissions from Soil and
Water” (P. A. Matson and R. C. Harriss, Eds.), pp. 14–51. Blackwell Science, Oxford.
Lockyer, D. R. (1984). A system for the measurement in the field of losses of ammonia through
volatilization. J. Sci. Food Agric. 35, 837–848.
Loubet, B., Milford, C., Hill, P. W., Tang, Y. S., Cellier, P., and Sutton, M. A. (2002). Seasonal
variability of apoplastic NHþ 4 and pH in an intensively managed grassland. Plant Soil 238,
97 –110.
Lyster, S., Morgan, M. A., and O’Toole, P. (1980). Ammonia volatilization from soils fertilized with
urea and ammonium nitrate. J. Life Sci. R. Dublin Soc. 1, 167–175.
Maclean, J. (1997). “Rice Almanac”, 2nd edn. IRRI, Los Baños, Philippines.
Malhi, S. S., and McGill, K. S. (1982). Nitrification in three Alberta soils: effect of temperature,
moisture and substrate concentration. Soil Biol. Biochem. 14, 393 –399.
Malhi, S. S., Grant, C. A., Johnston, A. M., and Gill, K. S. (2001). Nitrogen fertilization management
for no-till cereal production in the Canadian Great Plains: a review. Soil Tillage Res. 60,
101 –122.
Mannheim, T., Braschkat, J., and Marschner, H. (1997). Ammonia emissions from senescing plants
and during decomposition of crop residues. Z. Pflanzenernaehr. Bodenkd. 160, 125 –132.
Mary, B., Recous, S., and Robin, D. (1998). A model for calculating nitrogen fluxes in soil using 15N
tracing. Soil Biol. Biochem. 30, 1963–1979.
Mattsson, M., and Schjoerring, J. K. (1996). Ammonia emission from young barley plants: influence
of N source, light/dark cycles, and inhibition of glutamine synthetase. J. Exp. Bot. 47, 477–484.
Mattsson, M., and Schjoerring, J. K. (2002). Dynamic and steady state responses of inorganic nitrogen
pools and NH3 exchange in leaves of Lolium perenne and Bromus erectus to changes in root N
supply. Plant Physiol. 128, 742 –750.
Mattsson, M., Häusler, R. E., Leegood, R. C., Lea, P., and Schjoerring, J. K. (1997). Leaf–atmosphere
ammonia exchange in barley mutants with reduced activities of glutamine synthetase. Plant
Physiol. 114, 1307–1312.
Mattsson, M., Husted, S., and Schjoerring, J. K. (1998). Influence of nitrogen nutrition and
metabolism on ammonia emission from plant leaves. Nutr. Cycl. Agroecosyst. 51, 35– 40.
McDowell, L. L., and Smith, G. E. (1958). The retention and reaction of anhydrous ammonia on
different soil types. Soil Sci. Soc. Am. Proc. 22, 38–42.
McGarry, S. J., O’Toole, P., and Morgan, M. A. (1987). Effects of soil temperature and moisture
content on ammonia volatilization from urea-treated pasture and tillage soils. Ir. J. Agric. Res. 26,
173 –182.
McGinn, S. M., and Janzen, H. H. (1998). Ammonia sources in agriculture and their measurement.
Can. J. Soil Sci. 78, 139 –148.
McInnes, K. J., and Fillery, I. R. P. (1989). Modeling and field measurements of the effect of nitrogen
source on nitrification. Soil. Sci. Soc. Am. J. 53, 1264–1269.
McInnes, K. J., Kissel, D. E., and Kanemasu, E. T. (1985). Estimating ammonia flux: a comparison
between the integrated horizontal flux method and theoretical solutions of the diffusion profile.
Agron. J. 77, 884 –889.
McInnes, K. J., Ferguson, R. B., Kissel, D. S., and Kanemasu, E. T. (1986a). Field measurements of
ammonia loss from surface applications of urea solution to bare soil. Agron. J. 78, 192 –196.
McInnes, K. J., Ferguson, R. B., Kissel, D. S., and Kanemasu, E. T. (1986b). Ammonia loss from
applications of urea-ammonium nitrate solution to straw residue. Soil Sci. Soc. Am. J. 50, 969–974.
Mikkelsen, D. S., De Datta, S. K., and Obcemea, W. N. (1978). Ammonia volatilization losses from
flooded rice soils. Soil Sci. Soc. Am. J. 42, 752–730.
Misselbrook, T. H., and Hansen, M. N. (2001). Field evaluation of the equilibrium concentration
technique (JTI method) for measuring ammonia emission from land spread manure or fertiliser.
Misselbrook, T. H., van der Weerden, T. J., Pain, B. F., Jarvis, S. C., Chambers, B. J., Smith, K. A.,
Phillips, V. R., and Demmers, T. G. M. (2000). Ammonia emission factors for UK agriculture.
Atmos. Environ. 34, 871 –880.
Moal, J.-F., Martinez, J., Guiziou, F., and Coste, C.-M. (1995). Ammonia volatilization following
surface-applied pig and cattle slurry in France. J. Agric. Sci. 125, 245–252.
van der Molen, J., van Faassen, H. G., Leclerc, M. Y., Vriesema, R., and Chardon, W. J. (1990).
Ammonia volatilization from arable land after application of cattle slurry: derivation of a transfer
model. Neth. J. Agric. Sci. 38, 239–254.
Monteith, J. L., and Unsworth, M. H. (1990). “Principles of Environmental Physics”. Edward Arnold,
London.
Morvan, T., Leterme, Ph., Arsene, G. G., and Mary, B. (1997). Nitrogen transformations after the
spreading of a pig slurry on bare soil and ryegrass using 15N-labelled ammonium. Eur. J. Agron.
7, 181– 188.
Murphy, D. V., Recous, S., Stockdale, E. A., Fillery, I. R. P., Jensen, L. S., Hatch, D. J., and Goulding,
K. W. T. (2003). Gross nitrogen fluxes in soil: theory, measurement and application of 15N pool
dilution techniques. Adv. Agron. 79, 119.
Myers, R. J. K. (1975). Temperature effects on ammonification and nitrification in a tropical soil. Soil
Biol. Biochem. 7, 83–87.
Neftel, A., Blatter, A., Gut, A., Hogger, D., Meixner, F., Ammann, C., and Nathaus, F. J. (1998).
NH3 soil and soil surface gas measurements in a triticale wheat field. Atmos. Environ. 32,
499–505.
Nemitz, E., Sutton, M. A., Gut, A., San Jose, R., Husted, S., and Schjoerring, J. K. (2000). Sources and
sinks of ammonia within an oilseed rape canopy. Agr. For. Meteorol. 105, 385 –404.
Nemitz, E., Milford, C., and Sutton, M. A. (2001). A two-layer canopy compensation point model for
describing bi-directional biosphere–atmosphere exchange of ammonia. Q. J. Roy. Meteor. Soc.
127, 815 –833.
Nielsen, K. H., and Schjoerring, J. K. (1998). Regulation of apoplastic ammonium concentration in
leaves of oilseed rape. Plant Physiol. 118, 1361–1368.
Nömmik, H. (1966). Use of microplot technique for studying gaseous loss of ammonia from added
nitrogen materials under field conditions. Acta Agric. Scand. 16, 147–154.
Nye, P. H. (1992). Towards the quantitative control of crop production and quality. II. The scientific
basis for guiding fertilizer and management practice, particularly in poorer countries. J. Plant
Nutr. 15, 1151–1173.
O’Toole, P., McGarry, S. J., and Morgan, M. A. (1985). Ammonia volatilization from urea-treated
pasture and tillage soils: effects of soil properties. J. Soil Sci. 36, 613– 620.
Ouyang, D. S., Mackenzie, A. F., and Fan, M. X. (1998). Ammonia volatilization from urea amended
with triple superphosphate and potassium chloride. Soil Sci. Soc. Am. J. 62, 1443– 1447.
Pain, B. F., van der Weerden, T. J., Chambers, B. J., Phillips, V. R., and Jarvis, S. C. (1998). A new
inventory for ammonia emissions from UK agriculture. Atmos. Environ. 32, 309–313.
Paulson, C. A. (1970). The mathematical representation of wind speed and temperature profiles in the
unstable atmospheric surface layer. J. Appl. Meteorol. 9, 857–861.
Philip, J. R. (1959). The theory of local advection. Int. J. Meteorol. 16, 535–547.
Phongpan, S., Freney, J. R., Keerthisinghe, D. G., and Chaiwanakupt, P. (1995). Use of
phenylphosphorodiamidate and N-(n-butyl)thiophosphorictriamide to reduce ammonia loss and
increase grain-yield following application of urea to flooded rice. Fert. Res 41, 59 –66.
Ping, J., Bremer, E., and Janzen, H. H. (2000). Foliar uptake of volatilized ammonia from surface-
applied urea by spring wheat. Commun. Soil Sci. Plant Anal. 31, 165 –172.
Plantaz, M. A. G. H. (1998). “Surface/atmosphere exchange of ammonia over grazed pasture”.
Dissertation No. 2468. University of Wageningen, The Netherlands.
Prasertsak, P., Freney, J. R., Saffigna, P. G., Denmead, O. T., and Prove, B. G. (2001). Fate of urea
nitrogen applied to a banana crop in the wet tropics of Queensland. Nutr. Cycl. Agroecosyst. 59,
65–73.
Rachhpal-Singh, and Nye, P. H. (1986a). A model of ammonia volatilization from applied urea.
I. Development of the model. J. Soil Sci., 37, 1–9.
Rachhpal-Singh, and Nye, P. H. (1986b). A model of ammonia volatilization from applied urea.
II. Experimental testing. J. Soil Sci. 37, 37, 9–29.
Raupach, M. R. (1989a). A practical Lagrangian method for relating scalar concentrations to source
distributions in vegetation canopies. Q.J. R. Meteorol. Soc. 115, 609–632.
Raupach, M. R. (1989b). Stand overstorey processes. Phil. Trans. R. Soc. Lond. B 324, 175 –190.
Raupach, M. R. (1989c). Applying Lagrangian fluid mechanics to infer scalar source distributions
from concentration profiles in plant canopies. Agric. For. Meteorol. 47, 85 –108.
Raupach, M. R., and Legg, B. J. (1984). The uses and limitations of flux–gradient relationships in
micrometeorology. Agric. Water Manage. 8, 119–131.
Recous, S., Fresneau, C., Faurie, G., and Mary, B. (1988). The fate of labelled 15N urea and
ammonium nitrate applied to a winter wheat crop. Plant Soil 112, 205 –214.
Recous, S., Machet, J. M., and Mary, B. (1992). The partitioning of fertilizer-N between soil and crop:
comparison of ammonium and nitrate applications. Plant Soil 144, 101– 111.
Recous, S., Aita, C., Mary, B., and Faurie, B. (1999). In situ changes in gross N transformations in bare
soil after addition of straw. Soil Biol. Biochem. 31, 119–133.
Reynolds, C. M., and Wolf, D. C. (1987). Effect of soil moisture and air relative humidity on ammonia
volatilization from surface-applied urea. Soil Sci. 143, 144–152.
Ringbom, A. (1963). “Complexation in Analytical Chemistry”. Interscience Publishers, Chichester.
Roelcke, M., Li, S. X., Tian, X. H., Gao, Y. J., and Richter, J. (2002). In situ comparisons of ammonia
volatilisation from N fertilizers in Chinese loess soils. Nutr. Cycl. Agroecosyst. 62, 73– 88.
Russel, E. W. (1977). “Soil Conditions and Plant Growth”, 10th edn. Longman, London, New York.
Ryden, J. C., and Lockyer, D. R. (1985). Evaluation of a system of wind tunnels for field studies of
ammonia loss from grassland through volatilisation. J. Sci. Food. Agric. 36, 781 –788.
Ryden, J. C., Whitehead, D. C., Lockyer, D. R., Thompsen, R. B., Skinner, J. H., and Garwood, E. A.
(1987). NH3 emission from grassland and livestock production systems in the UK. Environ.
Pollut. 48, 173–184.
Schjoerring, J. K., and Mattsson, M. (2001). Quantification of ammonia exchange between
agricultural cropland and the atmosphere: measurements over two complete growth cycles of
wheat, oilseed rape, barley and pea. Plant Soil 228, 105– 115.
Schjoerring, J. K., Nielsen, N. E., Jensen, H. E., and Gottschau, A. (1989). Nitrogen losses from field-
grown spring barley plants as affected by rate of nitrogen application. Plant Soil 116, 167–175.
Schjoerring, J. K., Sommer, S. G., and Ferm, M. (1992). A simple passive sampler for measuring
ammonia emission in the field. Water Air Soil Pollut. 62, 13–24.
Schjoerring, J. K., Kyllingsbaek, A., Mortensen, J. V., and Byskov-Nielsen, S. (1993). Field
investigations of ammonia exchange between barley plants and the atmosphere. I. Concentration
profiles and flux densities of ammonia. Plant Cell Environ. 16, 161–167.
Schjoerring, J. K., Bock, J. H. G., Gammelvind, L., Jensen, C. R., and Mogensen, V. O. (1995).
Nitrogen incorporation and remobilization in different shoot components of field-grown winter
oil-seed rape as affected by rate of nitrogen application and irrigation. Plant Soil 177, 255–264.
Schjoerring, J. K., Husted, S., and Poulsen, M. (1998). Soil– plant– atmosphere NH3 exchange
associated with Calluna vulgaris and Deschampsia flexuosa. Atmos. Environ. 32, 507– 512.
Schlesinger, W. H., and Hartley, A. E. (1992). A global budget for atmospheric NH3. Biogeochemistry
15, 191–211.
Schulze, E.-D., De Vries, W., Hauhs, M., Rosén, K., Rasmussen, L., Tamm, C.-O., and Nilsson, J.
(1989). Critical loads for nitrogen deposition on forest ecosystems. Water Air Soil Pollut. 48,
451 –456.
Sherlock, R. R., and Goh, K. M. (1985). Dynamics of ammonia volatilization from simulated urine
patches and aqueous urea applied to pasture. I. Field experiments. Fert. Res. 5, 181 –195.
Sherlock, R. R., Freney, J. R., Bacon, P. E., and van der Weerden, T. J. (1995). Estimating ammonia
volatilization from unsaturated urea fertilized and urine affected soils by an indirect method. Fert.
Res. 40, 197–205.
Sherlock, R. R., Sommer, S. G., Khan, R. Z., Wesley Wood, C., Guertal, E. A., Freney, J. R., Dawson,
C. O., and Cameron, K. C. (2002). Ammonia, methane, and nitrous oxide emission from pig
slurry applied to a pasture in New Zealand. J. Environ. Qual. 31, 1491–1501.
Singh, Y., Malhi, S. S., Nyborg, M., and Beauchamp, E. G. (1994). Large granules, nests or bands:
methods of increasing efficiency of fall-applied urea for small cereal-grains in North America.
Fert. Res. 38, 61–87.
Sloan, J. J., and Anderson, W. B. (1995). Calcium chloride and ammonium thiosulfate as ammonia
volatilisation inhibitors for urea fertilizers. Commun. Soil Sci. Plant Anal. 26, 2425–2447.
Sommer, S. G., and Christensen, B. T. (1992). Ammonia volatilization after injection of anhydrous
ammonia into arable soils of different moisture levels. Plant Soil 142, 143– 146.
Sommer, S. G., and Ersbøll, S. K. (1996). Effect of air flow rate, lime amendments and chemical soil
properties on the volatilization of ammonia from fertilizers applied to sandy soils. Biol. Fert.
Soils 21, 53 –60.
Sommer, S. G., and Jensen, C. (1994). Ammonia volatilization from urea and ammoniacal fertilizers
surface-applied to winter wheat and grassland. Fert. Res. 37, 85– 92.
Sommer, S. G., Olesen, J. E., and Christensen, B. T. (1991). Effects of temperature, wind speed and
air humidity on ammonia volatilization from surface applied cattle slurry. J. Agric. Sci. 117,
91–100.
Sommer, S. G., Mikkelsen, H., and Mellqvist, J. (1995). Evaluation of meteorological techniques for
measurements of ammonia loss from pig slurry. Agric. For. Meteorol. 74, 169 –179.
Sommer, S. G., Friis, E., Bach, A., and Schjørring, J. K. (1997). Ammonia volatilization from pig
slurry applied with trail hoses or broadspread to winter wheat: effects of crop developmental
stage, microclimate and leaf ammonia absorption. J. Environ. Qual. 26, 1153–1160.
Son, T. T., and Buresh, R. J. (1994). Urea management for lowland transplanted rice as affected by
application of farmyard manure. Fert. Res. 39, 199 –203.
Sørensen, P. (2001). Short-term nitrogen transformations in soil amended with animal manure. Soil
Biol. Biochem. 33, 1211–1216.
Stevens, R. J., Laughlin, R. J., and Kilpatrick, D. J. (1989). Soil properties related to the dynamics of
ammonia volatilization from urea applied to the surface of acidic soils. Fert. Res. 20, 1–9.
Stumm, W., and Morgan, J. J. (1981). Aquatic Chemistry: an Introduction Emphasizing Chemical
Equilibria in Natural Waters. John Wiley and Sons, New York.
Sutton, M. A., Fowler, D., and Moncrieff, J. B. (1993a). The exchange of atmospheric ammonia with
vegetated surfaces. I. Unfertilized vegetation. Q. J. R. Meteorol. Soc. 119, 1023– 1045.
Sutton, M. A., Fowler, D., and Moncrieff, J. B. (1993b). The exchange of atmospheric ammonia with
vegetated surfaces. II. Fertilized vegetation. Q. J. R. Meteorol. Soc. 119, 1047–1070.
Sutton, M. A., Burkhardt, J. K., Guerin, D., Nemitz, E., and Fowler, D. (1998). Development of
resistance models to describe measurements of bi-directional ammonia surface-atmosphere
exchange. Atmos. Environ. 32, 473–480.
Sutton, M. A., Nemitz, E., Milford, C., Fowler, D., Moreno, J., San Jose, R., Wyers, G. P., Otjes, R. P.,
Harrison, R., Husted, S., and Schjoerring, J. K. (2000). Micrometeorological measurements of net
ammonia fluxes over oilseed rape during two vegetation periods. Agr. Forest Meteorol. 105,
351–369.
Svensson, L. (1994). A new dynamic chamber technique for measuring ammonia emissions from
land-spread manure and fertilizers. Acta Agric. Scand. B 44, 35–46.
Tabatabai, M. A., and Bremner, J. M. (1972). Assay of urease activity in soils. Soil Biol. Biochem. 4,
479–487.
Thom, A. S. (1972). Momentum, mass and heat exchange of vegetation. Q. J. R. Meteorol. Soc. 98,
124–134.
Thomas, R. J., Logan, K. A. B., Ironside, A. D., and Bolton, G. R. (1988). Transformation and fate of
sheep urine-N applied to an upland U.K. pasture at different times during the growing season.
Plant Soil 107, 173–181.
Touchton, J. T., and Hargrove, W. L. (1982). Nitrogen-source and methods of application for no-
tillage corn production. Agron. J. 74, 823–826.
Van Hove, L. W. A., Adema, E. H., Vredenberg, W. J., and Pieters, G. A. (1989). A study of the
adsorption of NH3 and SO2 on leaf surfaces. Atmos. Environ. 23, 1479–1486.
Van Hove, L. W. A., Heeres, P., and Bossen, M. E. (2002). The annual variation in stomatal ammonia
compensation point of rye grass (Lolium perenne L.) leaves in an intensively managed grassland.
Van Hove, L. W. A., Koops, A. J., Adema, E. H., Vredenberg, W. J., and Pieters, G. A. (1987).
Analysis of the uptake of atmospheric ammonia by leaves of Phaseolus vulgaris L. Atmos.
Environ. 21, 1759–1763.
Vandré, R., and Kaupenjohann, M. (1998). In situ measurement of ammonia emissions from organic
fertilizers in plot experiments. Soil Sci. Soci. Am. J. 62, 467–473.
Velthof, G. L., Oenema, O., Postmus, J., and Prins, W. H. (1990). In situ field measurements of
ammonia volatilization from urea and calcium ammonium nitrate applied to grassland.
Meststoffen 1/2, 41 –45.
Vlek, P. L. G., and Carter, M. F. (1983). The effect of soil environment and fertilizer modifications on
the rate of urea hydrolysis. Soil Sci. 136, 56–63.
Vlek, P. L. G., and Crasswell, E. T. (1981). Ammonia volatilization from flooded soils. Fert. Res. 2,
227 –245.
Volk, G. M. (1959). Volatile loss of ammonia following surface application of urea to turf or bare
soils. Agron. J. 51, 746–749.
Warland, J. S., and Thurtell, G. W. (2000). A micrometeorological method for in situ denitrification
measurements using acetylene inhibition. Agric. For. Meteorol. 103, 387 –391.
Watson, C. J., and Kilpatrick, D. J. (1991). The effect of urea pellet size and rate of application on
ammonia volatilization and soil nitrogen dynamics. Fert. Res. 28, 163– 172.
Watson, C. J., Travers, G., Kilpatrick, D. J., Laidlaw, A. S., and O’Riordan, E. (2000). Overestimation
of gross N transformation rates in grassland soils due to non-uniform exploitation of applied and
native pools. Soil Biol. Biochem. 32, 2019–2030.
Webb, E. K., Pearman, G. I., and Leuning, R. (1980). Corrections to flux measurements for density
effects due to heat and water vapour transfer. Q. J. R. Meteorol. Soc. 106, 85–100.
van der Weerden, T. J., and Jarvis, S. C. (1997). Ammonia emission factors for N fertilizers applied to
two contrasting grassland soils. Environ. Pollut. 95, 205–211.
van der Weerden, T. J., van der Moal, J. F., Martinez, J., Pain, B. F., and Guiziou, F. (1996).
Evaluation of the wind-tunnel method for measurement of ammonia volatilization from land.
J. Agric. Eng. Res. 64, 11– 13.
Wesely, M. L., Lenschow, D. H., and Denmead, O. T. (1989). Flux measurement techniques.
In “Global Tropospheric Chemistry: Chemical Fluxes in the Global Atmosphere” (D. H.
Lenschow and B. B. Hicks, Eds.), pp. 31– 46. National Center for Atmospheric Research,
Boulder, CO.
Wetselaar, R., and Farquhar, G. D. (1980). Nitrogen losses from tops of plants. Adv. Agron. 33,
263 –302.
Whitehead, D. C., and Lockyer, D. R. (1987). The influence of the concentration of gaseous ammonia
on its uptake by the leaves of Italian ryegrass, with and without an adequate supply of nitrogen to
the roots. J. Exp. Bot. 38, 818–827.
Whitehead, D. C., Lockyer, D. R., and Raistrick, N. (1988). The volatilization of ammonia from
perennial ryegrass during decomposition, drying and induced senescence. Annals Bot. 61,
567 –571.
Whitehead, D. C., and Raistrick, N. (1990). Ammonia volatilization from five nitrogen compounds
used as fertilizers following surface application to soils. J. Soil Sci. 41, 387– 394.
Whitehead, D. C., and Raistrick, N. (1993). The volatilization of ammonia from cattle urine applied to
soils as influenced by soil properties. Plant Soil 148, 43– 51.
Wilson, J. D., and Shum, W. K. N. (1992). A re-examination of the integrated horizontal flux method
for estimating volatilization from circular plots. Agric. For. Meteorol. 57, 281 –295.
Wilson, J. D., Thurtell, G. W., Kidd, G. E., and Beauchamp, E. G. (1982). Estimation of the rate of
gaseous mass-transfer from a surface source plot to the atmosphere. Atmos. Environ. 16,
1861–1867.
Wilson, J. D., Catchpoole, V. R., Denmead, O. T., and Thurtell, G. W. (1983). Verification of a simple
micrometeorological method for estimating the rate of gaseous mass transfer from the ground to
the atmosphere. Agric. Meteorol. 29, 183–189.
Winter, F. P., and Eiland, F. (1996). The effect of soil pH on nitrification in coarse sandy soil.
In “Progress in Nitrogen Cycling Studies” (O. van Cleemput, Ed.), pp. 655–658. Kluwer
Academic Publishers, The Netherlands.
Wood, C. W., Marshall, S. B., and Cabrera, M. L. (2000). Improved method for field scale
measurement of ammonia volatilization. Commun. Soil Sci. Plant Anal. 31, 581–590.
Yamulki, S., Harrison, R. M., and Goulding, K. W. T. (1996). Ammonia surface-exchnage above an
agricultural field in Southeast England. Atmos. Environ. 30, 109–118.
Zhu, T., Pattey, E., and Desjardins, R. L. (2000). Relaxed eddy-accumulation for measuring ammonia
volatilization. Environ. Sci. Technol. 34, 199 –203.
Zia, M. S., Aslam, M., Rahmatullah, Arshad, M., and Ahmed, T. (1999). Ammonia volatilisation from
nitrogen fertilizers with and without gypsum. Soil Use Manage. 15, 133 –135.
Index
A ammonia volatilization
climate interaction 576 –8
1:1– 2:1 mixed-layer clays 139 microbial processes 578–81
2:1 layer silicate clays 138– 9, 143, 144 –5 mineral fertilizers 561 –81
abiotic catalysis, humic substances formation soil CEC 574 –5
406 –10 soils–fertilizer– atmosphere interface
accelerated fallows 500, 501 561–5
acidity, volcanic soils 160–3 solid phase processes 575–6
activation tagging, rice 78 –9 TAN/pH interaction 565 –8
adhesion processes 9, 10, 31, 42, 44 –5 urea hydrolysis 568–70
adsorption, humic substances 397 –400 ammonium carbonate 561
aerodynamic gas flux measurement 609–10 ammonium chloride 561
AFM see atomic force microscopy ammonium nitrate 561
Africa, black pepper 343, 345 anaerobic chemical weathering 418–19
aggregates, soil stabilization 423– 6 andic horizons 125, 126
agricultural productivity 115–16, 500–1 Andisols
see also soil productivity; yields see also volcanic soils
agricultural sustainability 454 –6 acidity 160– 3
agroforestry 479 –84, 503 allophanic and nonallophanic 119–22,
see also trees 135–6, 143 –4, 149, 151
agronomy, black pepper 288–95 aluminum toxicity 160–2
algorithms, advanced 19 distribution 117
allele mining, rice 96 –7 fertility 163–7
allelic series, rice 95 nitrogen dynamics 151–5
alley cropping 479– 84, 505–6, 517, 524 organic matter accumulation 145 –7
allophanes, Andisols 119–22, 125, 135– 6, phosphorus dynamics 155–7
143 –4, 149, 151 properties 151
aluminum Soil Taxonomy 124, 125–7
Al3þ availability 120 andosolization 118
humus complexes 141, 143, 146, 149 Andosols, WRB scheme 124–5, 127
oxides in soil 400 –3 anhydrous ammonia 561
plant pathogens suppression 162–3 animal manure 496 –9, 525
toxicity in Andisols 160 –2 biomass 507– 8
volcanic soil genesis 147–50 labour requirement 521
ammonia emission 557–622 nutrient content 509– 10, 511
atmospheric 558–60 anticarcinogens 359–60
compensation point 584 –5 antiinflammatories 357–8
crop foliage 581– 6 antimicrobial agents 361–2
crops affected by fertilizers 594 –6 antioxidants 360–1
injected anhydrous ammonia 594 Arabidopsis
leaves-atmosphere transport 581 –2 genome 66 –8
magnitude of losses 582–4 insertional mutants 73–8
management strategies 586 –98 mutants 93
measurement techniques 560, 598–607, Asia, black pepper trade 274 –7, 344
609 –11 assays, DNA microarrays 183 –270
physiological processes 584 –6 atmosphere, TAN, soil-fertilizer interface
soil pH factor 570–4 561–5
623
624 INDEX
atmospheric ammonia 558–60 quantity 490, 491, 504–8

atmospheric stability 563, 609 transfer techniques 490–3
atomic force microscopy (AFM) 11 –27, 44– 5 biomineralization, metals 445 –6
advanced algorithms 19 biomolecule–mineral interface forces 27– 35
artifacts in force measurements 17– 18 ligand linkage schemes 28–30
bond chemistry and energies 19–25 siderophores and oxide surfaces 30 –5
colloidal tip 37 biotechnology
hysteresis 15–16 agricultural sustainability 454– 6
jumps 15 black pepper 282–5, 370
linkage protocol 38 rice 57 –9
optical lever collection system 11–12 biotin–avidin interaction 27
probes (cantilever/tip) 11 –18, 28– 30 black pepper (Piper nigrum) 271 –389
relevance to biological–inorganic interface see also Indian black pepper economy;
25–7 Indonesian black pepper economy
spring constant determination 16–17 agronomy 288– 95
tip shape 16 biotechnological research 282 –5, 370
whole cell technique 35–9 botany 279– 85
attachment strategies, DNA microarrays breeding 281–5
191–7 chemistry 285 –8, 322
Australia, black pepper trade 344 color 287–8
azotobactin DTPA extraction 302, 303–5
goethite interaction 31, 32, 33, 36 essential oils 285–7, 335 –9
mineral solution 35 fertilization 291– 5
oxide interaction 32, 34 future development 369–72
global economy 339– 47
B harvesting 322– 3
history of use 273–5
bacteria manuring 291, 294
black pepper 361 –2 nutrients
interaction with minerals 412, 414 availability 278–9, 288 –9, 290 –1,
soil formation role 419–20 371– 2
bacterium–mineral interface forces 2–4, 35–43 buffer power concept 296– 305, 372
Escherichia coli–muscovite 39 –41 deficiency symptoms 293, 295
Shewanella oneidensis –goethite/diaspore zinc buffer power 301 –5
41–3 pests and their control 311–21
whole cell microscopy technique 35–9 Foot Rot 278, 282, 284, 304, 311–15,
biodegradation, xenobiotics 436–8 349– 50
biodiversity mechanisms 451 insect infestation 320 –1
bioinformatics, rice 72–3 nematode infestation 315–19
biological control pepper weevil 321
nematode infestation 317 –19 Pollu beetle 308, 320–1
Phytophthera 314 pharmacopoeia 355–65
VAM 314–15, 317 plantation establishment 306 –11
biological evaluation, rice 71– 2 processing 322–8, 330 –9
biological nitrogen fixation (BNF) 513–16 cleaning and grading 325–7, 330 –1
biomass drying 324–5
banks 490, 492, 512 packaging and storing 327–8
labour requirement 521 production and trade 274–5, 276 –9
manuring 496– 7 products 365, 368 –9
nutrients 511 –12 propagation 281 –5, 306– 8
quality 490, 508– 11 root absorbing power 297– 8
INDEX 625
soils 289–90, 310–11 China, black pepper 274, 344

value added products 332–3, 334, 338, classification, volcanic soils 123–7
365 –9, 371 clays
yields 277–8, 339 –40 1:1–2:1 mixed-layer 139
BNF see biological nitrogen fixation 2:1 layer silicates 138–9, 143, 144 –5
bonds adsorption and pH 398– 400
chemistry 19–25 climate
polymers 9, 10 ammonia volatilization impact 576 –8
botany, black pepper 279–85 change and ecosystem effects 449–50
Brazil, black pepper supply 341 volcanic soil genesis 122– 3
breeding, black pepper 281 –5 clinical applications, black pepper 363
bridging polymers 8–11, 43 cluster analysis 228 –30
Br½nsted acidity 433–5 CNS see central nervous system
buffer power concept, zinc 301–5 CO2 weathering cycle 131 –3
bulk density, volcanic soils 164–5 colloids
burrowing nematodes 316–17 formation and transformation 141 –5
bush pepper, India 281, 351–2 soil environmental conditions 142
by-products, black pepper 368 –9 volcanic soil constituents 135–41
color, black pepper 287 –8
community genome arrays (CGAs) 253–6
C companion cropping, black pepper 309
competition, cover crops 485
calcium–ammonium –nitrate 561 complexation reactions, metals 441 –2
candidate gene approach, rice 96–7 composting 493–6, 525
carbon conservation, volcanic soils 166–7
global ion cycling 448 –9 consumer products, black pepper 365
mineral colloids influence 426–31 consumption, mineral fertilizers 559, 560–1
nitrogen ratio 509 contact printing, DNA microarrays 199–200
carbonic acid 115, 129, 130, 144 contamination, DNA microarrays 204
catalytic power, primary minerals 406– 9 continuous cultivation 500–1
catalytic transformations, organic pollutants coordination, nonhumic organics 396
431 –6 costs 524–5
cation exchange capacity (CEC) 574 –5 covalent bonding, DNA microarrays 192 –4
central nervous system (CNS) 357 –8 cover crops 484–90, 506
CFM see chemical force microscopy crop foliage, ammonia emission 581 –6
CGAs see community genome arrays crop improvement, rice 95–100
chavicine 287 cropping see alley cropping; companion
chemical control cropping; cover crops;
nematodes 317, 319–20 intercropping
Phytophthera 314 crops
Pollu beetle 320–1 see also agricultural productivity;
chemical force microscopy (CFM) 28, 45 black pepper; rice; yields
see alsoatomic force microscopy ammonia absorption 598
chemical weathering ammonia emission 557–622
anaerobic conditions 418– 19 cover crops 484 –90, 506
volcanic soils 115, 127–34 high-value 528–9
chemical-induced mutants, rice 83– 8 legumes 476–7, 513–16, 525
chemistry cross-species inference, rice 97–8
black pepper 285–8, 322 cryoground pepper 335
bonds 19– 25 crystalline soil components 134–5
volcanic soils 145–50 Cuba 529–30
626 INDEX
cultivation drying, black pepper 324 –5

black pepper 306 –11 DTPA extraction, black pepper 302, 303 –5
history, rice 56– 9 duration, cover crops 484–5
cultural traditions 533 dynamic force spectroscopy 25–7
D E
DAP see diammoniumphosphate ecology, soil contamination risk 456–9

data acquisition, DNA microarray image economic growth, black pepper impact 273,
processing 214 –15 346– 7
data analysis in microarrays 219–31 economics, LEIA 523 –6
see also microarrays ecosystems
data transformation 225 ammonia emission consequences 558–60
differentially expressed genes 225–7 climate change impact 449–50
experimental design 222–4, 231–6 soil, central role 453–4
genes used 221 –2 soil interactions impact 447–59
normalization 219–25 eddies, ammonia emission measurement 605–6
statistical methods 227 –31 electrostatic force 5–7, 40–1
systemic variation sources 219– 21 electrostatic interaction, DNA microarrays
data collection, forces 11 –27 192, 193
data mining 186 electroultrafiltration 303
data processing, forces 18– 19 emissions
decomposing plant material 596 ammonia 557–622
deferoxamine (DFO) 31 calculation 559–60
Deinococcus radiodurans 238 –9 EMS see ethyl methanesulfonate
deletion mutations 86–7 encapsulation, black pepper 338
demand, black pepper 278, 343 –5, 369 enclosures, ammonia emission measurement
dendrimeric linker coating 195 599– 601
deskinning, black pepper 333–4 energies, force measurements 19–25
developing countries, rice 98–100 energy balance method, gas flux measurement
DFO see deferoxamine 610– 11
diammoniumphosphate (DAP) 571, 575 entrapment tagging, rice 79–80
diaspore 32 environment, atmospheric ammonia 558 –60
Shewanella oneidensis forces 41–3 environmental microarray studies 236–45
diseases 284, 311–21 enzymes
insect infestation 320 –1 interactions, soil mineral colloids 415 –18
nematode infestation 315 –19 soil immobilization 438 –40
Phytophthera 278, 282, 284, 304, 311 –15 Escherichia coli–muscovite forces 39–41
distillation, black pepper 286, 335– 6 essential oils, black pepper 285–7, 335–9
diversification, intercropping 476– 7 ethyl methanesulfonate (EMS) 84
DLVO theory 40– 1 Europe, black pepper 343
DNA microarrays 183–270 exotoxicology 453 –4
data analysis 219 –31 experimental design, DNA microarrays
environmental studies 236–45 222– 4, 231–6
fabrication 190–205
genomic expression monitoring 231–42 F
hybridization/detection 184 –6, 205–14
image processing 214 –19 fabrication of microarrays 186, 190 –205
types/advantages 187– 9 arraying technology 197 –201
DNA pool screening, rice 81 contamination 204
drugs, black pepper 355–7 density 201–3
INDEX 627
nucleic acids, attachment 191–7 functional gene arrays (FGAs) 243–50

printing 199 –205 functional validation, rice genes 89–95
reproducibility 203 fungi
storage time 203– 4 interaction, minerals 412, 421
fallows 517 –18 nematode control 318 –19
accelerated 500, 501 Phytophthera 278, 282, 284, 304, 311 –15
evaluation 537 fungicides, Phytophthera 314
improved 499–504, 517, 525– 6 futures market, black pepper 353–5
farmers
perception 538
G
requirements 483 –4
FARMSIM model 535–6
gas fluxes formulae 609–11
FAs see fulvic acids
gel coating, DNA microarrays 195–6
ferrihydrite 140
gender issues 533
fertilization, black pepper 291–5, 371
gene chips see microarrays
fertilizers
genes
see also composting; manuring arrays, rice 71
alkaline 571–2 differential expression identification
ammonia volatilization 561–81 225–7, 231 –42
ammonium 560 environmental studies 242 –50
animal manure 496– 7, 498 expression analysis 89–91
composition 589–91 functional validation, rice 89–95
consumption 559, 560– 1 microarray data analysis 225–34
costs 524 –5 probes, FGAs 243– 4
crop ammonia emissions 557–622 replacement, rice 94 –5
FGAs see functional gene arrays silencing, rice 92
flavoured pepper products 366, 367 tagging, rice 74–80
flexible polymers 8–9 genetic diversity, Oryza spp. 59– 62
flooded fields 591 –4 genetic pathways and regulation, rice 97
fluorosis 159 genetic stocks, rice 70
Foot Rot (Phytophthera) 278, 282, 284, 304, genome sequencing 185
311 –15, 349 –50 genomics, rice 55–111
force–distance curves 12 –15, 18– 19, 20 geomedicine 452– 3
forces between mineral and biological surfaces germplasm, black pepper 284 –5
1 –54 global food production 474, 539
see also atomic force microscopy global market, black pepper 345, 355
bacterium– mineral interface 35– 43 global pepper economy 339–47
biomolecule–mineral interface 27–35 global population 473–4
bond rupture 23 –4 GLR see growth-light regime
chemical force microscopy 28, 45 glucuronidase reporter gene (GUS) 79
data collection 11 –27 goethite
data processing and statistics 18 –19 azotobactin interaction 31, 32, 33, 36
dynamic force spectroscopy 25– 7 Shewanella oneidensis forces 41–3
intermolecular 4–11 gradient diffusion methods 603 –5
spectroscopy 25–7 grading, black pepper 325–7, 330 –1
summary 10 grain legumes 476 –7, 525
formulae, gas flux measurement 609–11 grass family genetics 62–4
forward genetics, rice 73 –89 gravitational attraction 3
forward screening, rice 80–1 green manures 484
fulvic acids (FAs) 379 green pepper 332 –3, 365 –6, 367–8
fumigation, black pepper 332 Green Revolution 56, 101, 474
628 INDEX
ground pepper 334, 366 hydrophobic force 8

growth-light regime (GLR) 281 hysteresis 15–16
GUS see glucuronidase reporter gene
I
H
image processing, microarrays 214–19
halloysite 137–8, 144 imogolite 119, 120, 121, 125, 137, 143 –4
harvesting, black pepper 322– 3 see also allophanes
HAs see humic acids improved fallows 499–504, 517, 525 –6
health Indian black pepper economy 347–55
bush pepper 351–2
see also pharmacopoeia
Foot Rot 313, 314
black pepper impact 362
futures market 353–5
ecotoxicology 453–4
harvesting 322
hepatic toxicity 358–9
manuring 291, 294
hedgerows, alley cropping 480
marketing 352–3
HEIA see high external input agriculture
nematodes 315
hepatic toxicity, black pepper 358–9
Piper species 279– 82
heterologous biassays, rice 92–3
plantations 306 –10
high external input agriculture (HEIA) 474 –5,
production 349 –52
526–7
soils 289–90
high-troughput technologies, rice 70– 1
supply 340 –1
high-value crops 528–9
trade 275 –6
hisingerite 141
Indonesian black pepper economy
holistic approach, LEIA 531, 541 plantations 310 –11, 312
Honduras processing 328–30
Mucuna system 486 –7, 530– 1 supply 341 –2
Qezungual System 503–4 trade 276 –7
household models, LEIA technologies 534–5, industrial processing, black pepper 330–9
536–7 infiltration, ammonia volatilization 576 –8
human health 362, 453 –4 INM see integrated nutrient management
see also pharmacopoeia insect infestation, black pepper 320–1
human nutrition 451–4 insecticides 321, 364 –5
humic acids (HAs) 397, 408 –9 insertion sequence database, rice 81 –3
humic substances insertional mutants, rice 73 –83
binding mechanisms 397–400 integrated nutrient management (INM) 526–30
formation 406 –10 interactions
humification, Maillard reaction 410, 411 biodiversity mechanisms 451
humus 141, 143 ecosystem impact 447 –59
hybridization in microarrays 184–6, 205 –14 enzymes, soil minerals 415–18
detection 210 –14 humic– enzyme complexes 415–18
labeling 206 –10, 212 –13 metal transformation 440–7
practical issues 211– 14 mineral colloids 394–400
probe design and synthesis 205– 6 miocroorganisms, soil minerals 410– 14
probe DNA retention 211–12 organic pollutants 431–40
probe–target hybridization 210 soil mineral colloids 410–26
spatial resolution and cross-talk 213 intercropping 476–9, 517, 525
hydration force 7–8 interface forces see intermolecular forces
hydrogen bonding, nonhumic organics 396 intermolecular forces 1–54
hydrolysis, urea 568– 70 atomic force microscopy 11–27
hydrophobic bonding, nonhumic organics 397 bacterium–mineral interface 2–4, 35 –43
INDEX 629
biomolecule–mineral interface 27–35 Latin America, black pepper 344

electrostatic 5–7, 40 –1 legumes
solvation 7– 8 biological nitrogen fixation 513– 16
steric 8–11 grain legumes 476– 7, 525
van der Waals 3–5, 40 –1 LEIA see low external input agriculture
international collaboration, rice 98–100 Lennard–Jones potential 13–14, 15
International Rice Genome Sequencing Project Leptothrix, Mn oxide 420
(IRGSP) 58, 65 Lesotho, intercropping 478–9
International Society of Soil Science (ISSS) 393 Leucaena, alley cropping 480
ion cycling, ecosystem interaction 448 –9 Lewis acidity 433–5
ion exchange, nonhumic organics 395–6 ligands
ion– dipole interaction 396 linkage schemes 28–30, 44–5
IR29 rice 90 siderophores and oxide surfaces 30–5
IR64 rice 85 light conditions 597
IRGSP see International Rice Genome light-directed synthesis, photolithography
Sequencing Project 197–9
iron oxides 403–6 light-growth regime 281
iron–humus complexes 141, 146 linkages, ligands 28–30, 44–5
irradiation-induced mutants, rice 83–8 livelihood improvement technologies 473 –555
loading, molecular bonds 22–5
low external input agriculture (LEIA) 473– 555
J alley cropping 479 –84
animal manure 496 –9
Japan, black pepper 344 biomass transfers 490–3
composting 493–9
cover crops 484 –90
K
green manures 484
improved fallows 499 –504
kaolinite-smectite 139
integrated nutrient management 526– 30
Kerala, India, black pepper economy 275 –6,
intercropping 476–9
349 –51, 352 –3 modelling 534– 8
“King of spices” see black pepper socio-economic issues 516 –26, 538
soil fertility management 504–16
L systems perspective 530 –4
Lupinus sp., nitrogen fixation 129 –30
labeling of microarrays 206–10, 212–13
labour requirement 520–3, 540 M
biomass 493, 521
composting 494 –5 Machobane farming system 478 –9, 531
improved fallows 502– 3 macro-colinearity, rice 63–4
Machobane farming system 478– 9 Maillard reaction 410, 411
manuring 498–9, 521 maize, cover crop 486– 7, 489
rice intensification 522–3 Malaysia, black pepper 276, 277, 342
weed control 521– 2 mammalian cells 37
land management, volcanic soils 166–7
alley cropping 482–3 management of ammonia emission 586– 98
fragmentation 520 decomposing plant material 596 –7
Qezungual System 503 fertilizer application 594–6
requirement 516 –18, 540 fertilizer composition 589–91
tenure 483, 488, 518 reduction techniques 587 –9
Langrangian dispersion models 606– 7 rice paddies 591 –4
630 INDEX
manuring mineral–bacterium interface forces 35 –43

see also animal manure; composting mineral–biomolecule interface forces 27– 35
biomass 496 –7 mineralogy
black pepper 291, 294 volcanic rocks 127, 128
green manures 484 volcanic soils 134 –45
labour requirement 498– 9, 521 model genetic system, rice 59– 64
Machobane farming system 479 modelling, LEIA technologies 534–8
marketing, black pepper 352–3 molecular bond chemistry 19–25
measurement of ammonia emission 560, MPT see multipurpose tree techniques
598–607, 609–11 Mt. St. Helens volcano 129, 132, 157
eddy correlation and accumulation 605–6 Mucuna, cover crop 485 –9, 530
enclosures 599– 601 multipurpose tree (MPT) techniques 501 –2
gradient diffusion methods 603–5 muscovite–Escherichia coli forces 39–41
Langrangian dispersion models 606–7 mutagenic effects, black pepper 359–60
micrometeorological methods 601–3 mutants
tracer techniques 598 –9 chemical- and irradiation-induced 83 –8
medicines, black pepper 355–7, 363 deletion and point mutation stocks 85–8
melanization 118 mutagens 84 –5
MeNPOC 198
metals N
biomineralization 445–6
oxides, organic substances influence 400–6 natural genetic variation, rice 88–9
precipitation effects 444– 5 nematicides 319–20
reactions, soil 441–4 nematodes, black pepper 315 –19
transformation, biological processes 440–7 neural network analysis 230–1
micro-colinearity, rice 64 nitrification 578–81
microarrays 183– 270 nitrocellulose coating, DNA microarrays 196–7
advantages 188 –9 nitrogen
data analysis 219 –31 see also total ammoniacal nitrogen
environmental studies 236–45 accumulation and mineralization 151–3
fabrication 186, 190– 205 animal manure 509–10, 511
image processing 214 –19 biological fixation 129–30, 513–16
substrates 190–1 biomass 505, 508
surface modification 191–7 black pepper 291–2
types 187–8, 258 carbon/nitrogen ratio 509
microbial cells 38 fertilizer efficiency 154–5
microbial processes, ammonia volatilization global ion cycling 448– 9
578–81 immobilization 578– 81
microchips see microarrays leaching in Andisols 153–4
micrometeorological methods 601–3 mineral colloids influence 426 –31
micronutrients, black pepper 294 uptake and translocation 585 –6
microorganisms volcanic soil genesis 151–5
interactions, soil minerals 410–14, 421 –6 non-contact ink-jet printing 200–1
metal sorption 445 –6 non–porous microarray substrates 190–1
organic pollutants 431 –40 nonallophanic Andisols 119– 22
micropropagation, black pepper 284 noncrystalline material, volcanic soils 115, 118,
microspotting, DNA microarrays 199–200 127, 134
mineral colloid interactions 394–400 nonhumic organics, binding mechanisms 394,
humic–enzyme complexes 415 –18 395– 7
mineral fertilizers see fertilizers North America, black pepper trade 275, 344
mineral surfaces, humic substances 397–8 nucleic acid attachment, microarrays 191–7
INDEX 631
NUTMON project 529 oxides

nutrient buffer power concept 296–305, 372 aluminum 400– 3
measurement 298 –301 azotobactin interaction 32, 34
soil solution 296–7 iron 403 –6
zinc buffer power 301–5 manganese 420
nutrients organic substances influence 400– 6
see also manuring; nitrogen; organic matter; redox reactions 440 –1
phosphorus; potassium; zinc surfaces 30–5
animal manure 509– 10
biomass transfers 490 –3, 505 P
black pepper 278–9, 288–9, 290 –1, 371– 2
costs 524 packaging, black pepper 327–8
deficiency symptoms (black pepper) paddy rice 56, 591–4
293, 295 particulate organic matter (POM) 421–6
integrated management 526–30 see also organic matter
mining 511 –13 pepper
organic 527–8 see also black pepper
pepper soils 371–2 ancillary by-products 368– 9
root concentration 298–301 bush pepper 281, 351–2
subsistence agriculture 528 cryoground 335
dehydrated green 332–3
O green 365–6, 367 –9
ground 334, 366
odor, black pepper 287 harvesting schedule 322
oleoresin 281, 285, 326 –7, 329 –30, pharmacopoeia 355– 65
336 –8, 369 value added products 332–3, 334, 338,
oligonucleotide microarrays 183–270 365–8
oligonucleotide probes 206 white 329, 333– 5, 368
OM see organic matter pepper oil 329, 335 –9
opaline silica 139–40 pepper weevil 321
open reading frame (ORF) arrays 229 –30, pests
256 –7 see also diseases
organic ligands 27–30, 44–5 black pepper 308, 311–21
organic matter (OM) 507, 508–10, 511 pH
see also biomass; soil organic matter adsorption dependence 398
particulate matter 421 –6 clays 398 –400
volcanic soils 115, 118, 145 –7 soils, ammonia emission s 570 –4
organic pollutants, formation 431– 40 TAN interaction 565–8
organically grown black pepper 278, 291, pharmacopoeia, black pepper 355 –65
294, 370 phenotyping, rice 71–2
Oryza phosphorus
see also rice biomass 505, 508, 512, 521
biotechnology 57 –9 black pepper 292
gene and allele reservoir 62 global ion cycling 448–9
genomes 60 mineral colloids influence 426–31
improvement program species 60–1 volcanic soil genesis 155– 7
O. glaberrima 59–60 photolithography, light-directed synthesis 197–9
O. indica 60–1, 65, 68–9 phylogenetic oligonucleotide arrays 250– 3
O. japonica 60–1, 68–9 physical properties, volcanic soils 163 –7
O. nivara 61 physiological processes, ammonia emissions
O. sativa 59–60 584–6
632 INDEX
Phytophthera (Foot Rot) 278, 282, 284, 304, reproducibility, microarray quality 203
311–15, 349–50 research
phytosanitation 313 see also biotechnology; rice genomics
pin configuration, DNA microarrays 202–3 black pepper 370–2
Piper nigrum see black pepper rice 57 –9
Piperaceae 279 retrotransposon tagging, rice 77–8
piperine 281, 285, 287, 326 –7, 337, 356–7 reverse genetics, rice 73– 89
plant pathogens, suppression by aluminum rhizosphere
162–3 interactions, environmental 451 –2
plantations, black pepper 306 –11 metals, transformation 440–4
point mutations 87–8 rice
Pokkali rice 90 paddies 56, 591 –4
pollination, black pepper 280 System for Rice Intensification 512–13,
Pollu beetle, black pepper 308, 320–1 522– 3
pollution 436–40, 452 –3 Rice Biotechnology Program 57–8
polygenetic soil profile 119 rice genomics 55 –111
polymers, bridging 8 –11, 43 see also Oryza
POM See particulate organic matter allelic series 95
porosity, volcanic soils 164 –5 bioinformatics 72– 3
porous substrates, microarrays 190– 1 biological evaluation 71 –2
potassium, black pepper 292 candidate gene approach and allele mining
poverty, pathways out of 537, 538 96– 7
precipitation, metals transformation 444 –5 chemical- and irradiation-induced mutants
prices, black pepper 340 –1, 348 83– 8
principal component analysis (PCA) 228 crop improvement 95–100
printing technologies, microarrays 197–201 cross-species inference 97 –8
probe–target hybridization 189, 210 cultivation history 56–9
probes, AFM 11–18, 28–30 features and composition 66 –8
processing, black pepper 322–8, 330–9 functional validation 89 –95
productivity, volcanic soils 115 –16, 150, 160– 3 gene discovery ingredients 69–73
propagation, black pepper 281 –5, 306– 8 gene expression analysis 89 –91
protein profiles, rice 91 gene replacement 94–5
protonation, nonhumic organics 396 gene silencing 92
pruning, black pepper 308 genetic diversity 59– 62
Pyrococcus furiosus 239, 241–2 genetic stocks 70
heterologous bioassays 92–4
Q high-throughput technologies 70 –1
insertional mutants 73 –83
Qezungual System, Honduras 503–4 international collaboration 98–100
quality control IR64 variety 85
microarray fabrication 203 model genetic system 59–64
microarray image processing 215– 19 natural genetic variation 88–9
quantitative measurements, microarrays Oryza indica/japonica comparison 68–9
247–9, 255–6 pathways and genetic regulation 97
research 57 –9
R salient features 66–8
sequencing 64–5
rainfed rice 56 single genetic system 62–4
redox reactions, metal oxides, soils 440–1 risk assessment, soil contamination 456– 7
reduction techniques, ammonia emissions 587– 9 RNA, microarray assay 183 –270
reductionist approach, LEIA 531, 541 Rockefeller Foundation 57–8
INDEX 633
roots soil mineral colloids

absorbing power (black pepper) 297 –8 binding mechanisms 394–400
exudates 442–4 biogeochemical cycling 426– 31
fungus infection 312 interactions 410 –26
nutrient concentration 298–301 surface properties 394 –5
root knot nematodes 315–16 soil minerals
rupture forces, molecular bonds 23–4 humic substances, formation 406 –10
Russia, pepper imports 348 precipitation 421–3
weathering 418– 20
S soil moisture, Andisols 126
soil organic carbon fraction (SOC) 504–8
SBH see sequencing by hybridization soil organic matter (SOM) 115, 118, 145– 7,
scatter plots, microarray data 227 426–31, 504, 505, 506
self-organizing maps (SOMs) 230–1 soil productivity, volcanic soils 150, 160 –3
sensitivity, DNA microarrays 245, 246– 7, soil properties, alley cropping 481–2
255 –6 soil solution 296– 7, 441–2
sequencing by hybridization (SBH) 185 soil structure, volcanic soils 163 –4
share-cropping 518–20 Soil Taxonomy, Andisols 124, 125 –7
Shewanella oneidensis soil texture, volcanic soils 163–4
DNA microarray analysis 236–8 soils
goethite/diaspore forces 41–3 aggregation dynamics 423 –6
shifting cultivation 499–500 ammonia emissions, pH factor 570–4
siderophores 30–5, 45 atmosphere, TAN 561 –5
silanization, DNA microarrays 195 black pepper 289– 90, 310–11
silica CEC 574– 5
opaline 139–40 contamination risk 456–9
volcanic materials 127 ecosystem role 453–4
silicates, 2:1 layer 138–9, 143, 144 –5 formation, microorganisms 419–20
silicon metals, reactions 441 –4
nutrition in volcanic soils 160 restoration 457–9
soil solution activity 133 –4 structure stabilization 421–6
similarity measurement, microarray data 227– 8 TAN, microbial processes 578–81
SOC see soil organic carbon fraction volcanic 113–82
socio-economic issues of LEIA 516 –26, 538 solvation force 7–8
economics 523 –6 SOM see soil organic matter
labour 520– 3 specificity, oligonucleotide microarrays 244–6,
land 516 –20 251–2
soil environmental conditions, colloid formation spices see black pepper
142 spot quality, DNA microarrays 202, 214 –19
soil erosion, alley cropping 482 SRI see System for Rice Intensification
soil fertility 504–16 Sri Lanka, black pepper trade 276–7, 307, 342
animal manure 498 stabilization, soil structure 421–6
biological nitrogen fixation 513–16 statistical methods, microarray data 227–31
biomass quality 508 –11 statistics, forces 18 –19
biomass quantity 504 –8 steric force 8–11
biomass transfers 492 sterilization, black pepper 330 –2
composting 494 storage times, DNA microarrays 203–4
nutrient mining 511–13 structural force 7–8
pepper soils 371 struvite, precipitation 575–6
volcanic soils 157–8, 163 –7 subsistence agriculture 473 –555
soil genesis, volcanic materials 118– 23 see also low external input agriculture
634 INDEX
substrates, microarrays 190–1 vesicular arbuscular mycorrhizae (VAM)

sulphur 314– 15, 317
global ion cycling 448–9 volatilization see ammonia volatilization
mineral colloids influence 426 –31 volcanic rocks (ejecta) 114, 116, 124, 129 –33
supply, black pepper 339 –43 glass 127–9, 133–4
surface modification, microarrays 191–7 mineralogy 128
sustainable livelihoods approach 531 –2, 533, weathering 127–9, 132, 133–4
534 volcanic soils 113 –82
System for Rice Intensification (SRI) 512–13, see also Andisols; Andosols
522–3, 530–1 acidity 160–3
aluminum dynamics 147–50
T bulk density and porosity 164–5
charge characteristics 150– 1
T-DNA tagging, rice 74–5, 78 –9 chemical characteristics 145 –50
TAN see total ammoniacal nitrogen chemical weathering 115, 127–34
classification 123– 7
tephra, chemical weathering 129 –33
colloid formation and transformation
Thailand, black pepper 342
141– 5
Tithonia diversifolia 512, 517, 521
colloidal constituents 135–41
biomass transfers 491–2
distribution 114–15, 116–18
topography, black pepper plantations 307–8
fertility 157–8, 163–7
total ammoniacal nitrogen (TAN)
genesis 118–23
climate interaction 576–8
management and conservation 166 –7
microbial processes 578– 81
mineral nutrients 157–9
pH interaction 565 –8
mineralogical characteristics 134– 45
soil–fertilizer–atmosphere interface 561–5
nitrogen dynamics 151 –5
toxicity, aluminum 160 –2
noncrystalline material 115, 118, 127, 134
toxicological effects, black pepper 363–4
organic matter accumulation 145–7
tracer techniques, ammonia emission
phosphorus dynamics 155–7
measurement 598–9
physical properties 163–7
transcription profiling, rice 89–90
productivity 115 –16, 150, 160–3
transposon tagging, rice 75–7
properties 115–16
trees
silicon nutrition 160
alley cropping 481
texture and structure 163 –4
biomass production 505 –6
water retention and availability 165–6
cover crops 489 volcanism 114, 116 –17
multipurpose tree techniques 501–2
W
U
water retention, volcanic soils 165– 6
urea weathering
climate interaction 576–8 see also chemical weathering
consumption 560– 1 CO2 cycle 131 –3
hydrolysis 568 –70 soil minerals 418–20
temperature and moisture conditions 130
V volcanic soil genesis 119–20, 122 –3, 127
weed control 539
value added pepper products 332, 334, 338, alley cropping 481
365–9, 371 cover crops 485–6, 489–90
VAM see vesicular arbuscular mycorrhizae labour requirement 521 –2
van der Waals force 3–5, 40– 1 weevils, black pepper 321
INDEX 635
Western Ghats, India, black pepper 281 –2, Y

290
wheat streak mosaic virus (WSMV) 93–4 ‘yellows’, black pepper 317
white pepper 329, 333–5, 368 yields 539
whole cell force microscopy technique 35–9 alley cropping 480, 524
Whole-genome Open Reading Frame arrays animal manuring 497
256 –7 black pepper 277– 8, 339–40
wilt, black pepper 312, 313 intercropping 477
Working Group MO (1990) 393 organic matter fertilization 510
World Reference Base for Soil Resources rice 56–7
(WRB), Andosols 124–5, 127
WSMV see wheat streak mosaic virus Z
X zinc
buffer power in black pepper 301 –5
xenobiotics, bioavailability 432, 436–40 deficiency in black pepper 294, 295

Advances in Agronomy v.82

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Advances in Agronomy v.82

Uploaded by

Copyright:

Available Formats

VOLUME 82

John S. Boyer Paul Bertsch

Ronald Phillips Kate M. Scow

Emeritus Advisory Board Members

Kenneth J. Frey Eugene J. Kamprath

Prepared in cooperation with the

Lisa K. Al-Almoodi David M. Kral Diane H. Rickerl

q 2004 Elsevier Inc. All rights reserved.

Electronic Storage or Usage

First edition 2004

Library of Congress Cataloging in Publication Data

British Library Cataloging in Publication Data

Typeset by Alden Press, Oxford

FORCES BETWEEN MINERALS AND BIOLOGICAL

I. Introduction — Forces in Nature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

RICE FUNCTIONAL GENOMICS : LARGE -SCALE GENE

THE NATURE , PROPERTIES AND MANAGEMENT OF

III. Soil Genesis in Volcanic Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118

MICROARRAY TECHNOLOGY AND APPLICATIONS IN

B. Data Transformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225

THE AGRONOMY AND ECONOMY OF BLACK PEPPER

V. Establishing a Pepper Plantation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306

SOIL MINERAL – ORGANIC MATTER –MICROORGANISM

B.Complexation Reactions of Metals . . . . . . . . . . . . . . . . . . . . . . . . . . 441

LOW EXTERNAL INPUT TECHNOLOGIES FOR

AMMONIA EMISSION FROM MINERAL FERTILIZERS

G. AN (55), Department of Life Science, Pohang University of Science and

I. Introduction — Forces in Nature

I. INTRODUCTION — FORCES IN NATURE

II. FUNDAMENTAL FORCES AT THE INTERFACE

“All intermolecular forces are essentially electrostatic in origin” (page 11,

A. THE VAN DER WAALS FORCE

the time dependent ﬂuctuations in the electric dipole moment of a particle as it

B. THE ELECTROSTATIC FORCE

The electrostatic force arises through a variety of mechanisms leading to the

dielectric constant, causes the dissociation of surface functional groups. These

C. THE SOLVATION FORCE

intervening liquid ceases to behave as a structureless media resulting in a force

D. THE STERIC FORCE AND BRIDGING POLYMERS

FðxÞ ¼ ðkB T=bÞ½0:25ð1 2 x=LÞ22 2 0:25 þ x=L ð4Þ

where kB is the Boltzmann’s constant, T is the temperature, b is the persistence

Type of interaction Description

van der Waals Force between all particles due to

III. FORCE CURVE THEORY AND COLLECTING

Force measurements attempt to capture interactions representing the

Hooke’s law, F ¼ 2 ksd, then allows conversion of cantilever deﬂection, d,

A. FORCE – DISTANCE CURVES, CAPTURING A POTENTIAL

The focus is now turned to the region of interaction, which is of primary

EðDÞ ¼ 2A=D6 þ B=D12 ð5Þ

dE=dD ¼ FðDÞ ¼ 26A=D7 þ 12B=D13 ð6Þ

Figure 3 (a) Differentiated Lennard– Jones potential provided as an example interaction to be

an attractive force as a characteristic jump to contact on approach. These “jumps”

Certainly other sources besides the instability of the cantilever contribute to

Both phenomena facilitate additional bonding, larger hysteresis and higher

D. SPRING CONSTANT DETERMINATION

If a quantitative analysis of absolute force values is desired, determination of

derives the spring constant by measuring changes in the cantilever’s resonance

E. ARTIFACTS IN FORCE MEASUREMENTS

present in the region of no contact, common when making measurements in a

F. DATA PROCESSING AND STATISTICS

H. RELATING BOND CHEMISTRY AND ENERGIES

Force microscopy measurements intuitively have the potential to describe

phenomenon leads to an intriguing positive correlation between the rupture force,

loge ðRof Þ ¼ 2Eb =kB T þ loge ð fb =tD Þ ð7Þ