You are on page 1of 364

Natural Science in Archaeology

Ervan Garrison

in Archaeological
Second Edition
Natural Science in Archaeology

Series editors
Günther A. Wagner
Christopher E. Miller
Holger Schutkowski
More information about this series at
Ervan Garrison

Techniques in
Archaeological Geology

Second Edition
Ervan Garrison
Department of Geology, University of Georgia
Athens, Georgia, USA

ISSN 1613-9712
Natural Science in Archaeology
ISBN 978-3-319-30230-0 ISBN 978-3-319-30232-4 (eBook)
DOI 10.1007/978-3-319-30232-4

Library of Congress Control Number: 2016933149

# Springer-Verlag Berlin Heidelberg 2016

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or
part of the material is concerned, specifically the rights of translation, reprinting, reuse of
illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way,
and transmission or information storage and retrieval, electronic adaptation, computer software,
or by similar or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are
exempt from the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in
this book are believed to be true and accurate at the date of publication. Neither the publisher nor
the authors or the editors give a warranty, express or implied, with respect to the material
contained herein or for any errors or omissions that may have been made.

Printed on acid-free paper

This Springer imprint is published by Springer Nature

The registered company is Springer Verlag GmbH Berlin Heidelberg
This book is dedicated to my grandchildren, Kelsey and Hunter.
ThiS is a FM Blank Page

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Organization of This Book . . . . . . . . . . . . . . . . . . . . . . 6
2 The Geomorphological and Geological Context . . . . . . . . . . . 11
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.2 Landscapes and People . . . . . . . . . . . . . . . . . . . . . . . . 11
2.3 Geomorphic Concepts . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.4 Geomorphic Setting . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.4.1 Fluvial Landforms . . . . . . . . . . . . . . . . . . . . . . 15
2.4.2 Coastal Landforms . . . . . . . . . . . . . . . . . . . . . 17
2.4.3 Submerged Landforms . . . . . . . . . . . . . . . . . . 19
2.4.4 Mountain and Glacial Landforms . . . . . . . . . . . 22
2.4.5 Ice Patches . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.4.6 Lacustrine Landforms . . . . . . . . . . . . . . . . . . . 28
2.4.7 Loessic and Glacial Landscapes: Midwestern
United States . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.4.8 Desert/Arid Landforms . . . . . . . . . . . . . . . . . . 33
2.4.9 The Sahara: Geoarchaeology of Paleolakes and
Paleoclimate . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.4.10 Karst/Cave Landforms . . . . . . . . . . . . . . . . . . . 36
2.4.11 Combe Grenal . . . . . . . . . . . . . . . . . . . . . . . . 37
2.4.12 Volcanic Landforms . . . . . . . . . . . . . . . . . . . . 40
2.5 Earthquakes: Volcanic or Otherwise . . . . . . . . . . . . . . . 42
2.6 Mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.7 Map Scale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
2.8 Data Sources for Mapping . . . . . . . . . . . . . . . . . . . . . . 46
2.9 LiDAR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
2.10 Structure from Motion (SfM) . . . . . . . . . . . . . . . . . . . . 48
2.11 Drones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
2.12 Making the Map: ArcGIS to Google Earth . . . . . . . . . . 50
2.13 Other Types of Maps . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3 Sediments, Soils, and Stratigraphy in Archaeological
Geology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.2 A Brief Review of Sediments and Soils . . . . . . . . . . . . . 56

viii Contents

3.3 The Soil Catena . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

3.4 The Soil Chronosequence . . . . . . . . . . . . . . . . . . . . . . . 63
3.5 Describing Archaeological Sediments and Soils
in Profile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
3.6 The Geological Stratigraphic Section . . . . . . . . . . . . . . 72
3.7 Nomenclature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
3.7.1 Allostratigraphy . . . . . . . . . . . . . . . . . . . . . . . 74
3.7.2 Sequence Stratigraphy . . . . . . . . . . . . . . . . . . . 75
4 Techniques for Archaeological Sediments and Soils . . . . . . . . 77
4.1 Sampling Sediments and Soils: Monoliths to Sediment
Grabs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
4.2 One Method for Constructing a Soil Monolith . . . . . . . . 79
4.3 Handling and Description of Cores: Some General
Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
4.4 Standard Operating Procedure for Collection of Sediment
Samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
4.5 “1700 Sondages”: Geological Testing of the Plateau of
Bevaix, Neuchâtel (Switzerland) . . . . . . . . . . . . . . . . . . 89
4.6 Analytical Procedures for Sediments and Soils . . . . . . . 90
4.7 Particle Size Analysis . . . . . . . . . . . . . . . . . . . . . . . . . 92
4.7.1 Hydrometer Method . . . . . . . . . . . . . . . . . . . . 94
4.7.2 The Pipette Method . . . . . . . . . . . . . . . . . . . . . 94
4.7.3 The Modified Pipette or “Fleaker” Method . . . . 95
4.7.4 The Imhoff Cone Method . . . . . . . . . . . . . . . . 95
4.7.5 Organic Content Determination Methods . . . . . 98
4.8 Determination of Total Phosphorus by Perchloric
Digestion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
4.9 Absolute Phosphate Analysis Versus Qualitative Color
Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
4.10 Colorimetry and Spectrophotometry . . . . . . . . . . . . . . . 102
4.11 Micromorphology: Describing Archaeological Sediments
and Soils with the Microscope . . . . . . . . . . . . . . . . . . . 103
4.12 Palynology: A Micromorphological Study of
Archaeological Pollen . . . . . . . . . . . . . . . . . . . . . . . . . 104
4.13 Phytoliths for Archaeology . . . . . . . . . . . . . . . . . . . . . . 109
4.14 Phytolith Identification and Morphology . . . . . . . . . . . . 110
4.15 Phytolith Extraction and Counting . . . . . . . . . . . . . . . . 111
5 Geophysical Techniques for Archaeology . . . . . . . . . . . . . . . 115
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
5.2 The International Society for Archaeological Prospection
(ISAP) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
5.3 Electrical Methods: Resistivity . . . . . . . . . . . . . . . . . . . 117
5.3.1 Resistivity Arrays . . . . . . . . . . . . . . . . . . . . . . 119
5.4 Vertical Sounding Methods in Archaeology . . . . . . . . . 122
5.5 Electrical Methods: Electromagnetic/Conductivity . . . . . 123
5.5.1 Application of EM Methods . . . . . . . . . . . . . . 124
Contents ix

5.6 Conductivity (EM) Survey of a Burned and Buried

Cherokee Seminary Building, Tahlequah, Oklahoma,
Cherokee Nation, USA . . . . . . . . . . . . . . . . . . . . . . . . . 124
5.6.1 Magnetic Methods . . . . . . . . . . . . . . . . . . . . . 124
5.6.2 Magnetometers . . . . . . . . . . . . . . . . . . . . . . . . 127
5.6.3 Gradiometers . . . . . . . . . . . . . . . . . . . . . . . . . 129
5.6.4 Magnetic Anomaly Interpretation: Basic
Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
5.6.5 Magnetic Prospection: Practice . . . . . . . . . . . . 130
5.6.6 Date Acquisition and Display . . . . . . . . . . . . . 131
5.6.7 Advantages and Disadvantages . . . . . . . . . . . . 131
5.7 An Abandoned Nineteenth-Century Cherokee Cemetery,
Park Hill, Oklahoma, Cherokee Nation, USA . . . . . . . . 132
5.7.1 Ground-Penetrating Radar: GPR . . . . . . . . . . . 132
5.7.2 Field Survey Methods . . . . . . . . . . . . . . . . . . . 137
5.7.3 Digital Post-Processing of GPR Data . . . . . . . . 137
5.7.4 Multiple Geophysical Techniques . . . . . . . . . . 139
5.7.5 Underwater Geophysical Survey Techniques in
Archaeology . . . . . . . . . . . . . . . . . . . . . . . . . . 139
5.8 The Sabine River Paleovalley, Northern Gulf of Mexico,
USA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
5.9 Doggerland: Mesolithic Landscapes of the Southern
North Sea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
6 Petrography for Archaeological Geology . . . . . . . . . . . . . . . . 145
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
6.2 Major Rock Types and Archaeology . . . . . . . . . . . . . . . 146
6.3 Sedimentary Rocks . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
6.3.1 Arenite or Arenitic Sandstone . . . . . . . . . . . . . 148
6.3.2 Arkose or Arkosic Sandstone . . . . . . . . . . . . . . 149
6.3.3 Graywackes or Clay-Rich Sandstone
(“Lutitic”) . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
6.4 A Roman Quern Production and Smithy Works . . . . . . . 154
6.4.1 Clastic or Calcarenite Limestones . . . . . . . . . . 154
6.4.2 Aphanitic Limestones . . . . . . . . . . . . . . . . . . . 155
6.4.3 Dolomite and Dolomitic Limestones . . . . . . . . 156
6.4.4 Cherts, Gypsum, and Ironstones . . . . . . . . . . . . 157
6.5 Igneous Rocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
6.5.1 Granite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
6.5.2 Diorite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
6.5.3 Andesite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
6.5.4 Basalt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
6.5.5 Rhyolite and Obsidian . . . . . . . . . . . . . . . . . . . 162
6.5.6 Gabbro/Diabase . . . . . . . . . . . . . . . . . . . . . . . 163
6.5.7 Tuff . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
6.6 Metamorphic Rocks . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
6.6.1 Metamorphic Rocks: Foliated . . . . . . . . . . . . . 165
6.6.2 Metamorphic Rocks: Non-foliated . . . . . . . . . . 166
x Contents

6.6.3 Metamorphic Rocks of Archaeological

Interest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
6.6.4 “Greenstone” . . . . . . . . . . . . . . . . . . . . . . . . . 170
6.7 Techniques: Optical and Otherwise . . . . . . . . . . . . . . . . 173
6.7.1 Hand-Specimen Macroscopic Analysis of Rocks
and Minerals . . . . . . . . . . . . . . . . . . . . . . . . . . 174
6.7.2 Optical Properties of Minerals . . . . . . . . . . . . . 175
6.7.3 Quantifying Minerals in Thin Sections . . . . . . . 177
7 Clay Minerals and Ceramics . . . . . . . . . . . . . . . . . . . . . . . . . 179
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
7.2 Clays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
7.2.1 Clays in Paleosols . . . . . . . . . . . . . . . . . . . . . . 181
7.2.2 Clay Deposits . . . . . . . . . . . . . . . . . . . . . . . . . 183
7.3 Ceramics and “Ceramics” . . . . . . . . . . . . . . . . . . . . . . . 185
7.3.1 Ceramic Properties of Clay . . . . . . . . . . . . . . . 187
7.3.2 Ceramic Ethnotechnological Study: Making
Navaho/Diné Pottery . . . . . . . . . . . . . . . . . . . . 193
7.3.3 Geoarchaeological Study of Ceramics . . . . . . . 195
7.4 Provenance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
7.5 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
7.6 Optical Petrography . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
7.7 X-Ray Radiography . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
7.8 Chemical Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
7.8.1 An Example of Chemical Analysis Using X-Ray
Diffraction (XRD) . . . . . . . . . . . . . . . . . . . . . . 202
7.8.2 Diagenetic Changes . . . . . . . . . . . . . . . . . . . . . 204
7.8.3 Analysis of Residues . . . . . . . . . . . . . . . . . . . . 205
8 Instrumental Analytical Techniques for Archaeological
Geology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
8.2 Analytical Techniques and Their Pluses (and Minuses)
for Archaeology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
8.3 X-ray Diffraction (XRD) . . . . . . . . . . . . . . . . . . . . . . . 213
8.4 X-ray Fluorescence (XRF) . . . . . . . . . . . . . . . . . . . . . . 215
8.5 Is pXRF a Revolution or Just the Latest Development in
XRF? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
8.6 Electron Microprobe Analysis and Proton-Induced X-ray
Emission (EMPA and PIXE; SEM) . . . . . . . . . . . . . . . . 218
8.7 Examples of Chemical Analysis of Nineteenth Century
Transfer Printed Whitewares Using EMPA . . . . . . . . . . 222
8.8 Atomic Absorption, Inductively Coupled Plasma/Atomic
Emission Spectroscopy (AAS, ICP/AES) . . . . . . . . . . . 224
8.9 Mass Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
8.10 Aphrodite? An Example of Isotopic Analysis Using Mass
Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
Contents xi

8.11 Inductively Coupled Plasma: Mass Spectroscopy

(ICP-MS; LA-ICP-MS) . . . . . . . . . . . . . . . . . . . . . . . . 230
8.12 Neutron Activation Analysis (NAA/INAA) . . . . . . . . . . 232
8.12.1 Provenance Determination . . . . . . . . . . . . . . . . 234
8.13 An Example of Neutron Activation Analysis: Coinage
and a Celtic Mint: The Titelberg, Luxembourg . . . . . . . 234
8.14 Electron Spin Resonance (ESR) . . . . . . . . . . . . . . . . . . 237
8.15 Magnetic Susceptibility . . . . . . . . . . . . . . . . . . . . . . . . 239
8.16 Cathodoluminescence Microscopy (CL) . . . . . . . . . . . . 240
8.17 Infrared and Raman Spectroscopy . . . . . . . . . . . . . . . . . 241
8.18 Instrumental Geochemical Techniques and Their
Availability to Archaeology . . . . . . . . . . . . . . . . . . . . . 243
8.19 A Closing Quote or Quotes . . . . . . . . . . . . . . . . . . . . . 245
9 Metallic Minerals, Ores, and Metals . . . . . . . . . . . . . . . . . . . 247
9.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
9.2 Early Metallurgy: Copper . . . . . . . . . . . . . . . . . . . . . . . 247
9.3 Metallurgy: Copper to Bronze . . . . . . . . . . . . . . . . . . . 249
9.4 Iron Metallurgy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254
9.5 A Metallurgy for the Common Man: African Iron . . . . . 257
9.6 Metallic Minerals and Ore Genesis . . . . . . . . . . . . . . . . 260
9.7 Model Studies for the Analysis of Metal in
Archaeological Geology . . . . . . . . . . . . . . . . . . . . . . . . 265
9.7.1 Slag: “The Archaeometallurgists’s Friend” . . . . 265
9.7.2 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268
9.8 Artifacts: “When Fortune Smiles” . . . . . . . . . . . . . . . . . 268
9.8.1 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268
10 Statistics in Archaeological Geology . . . . . . . . . . . . . . . . . . . 271
10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271
10.2 Descriptive Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . 272
10.2.1 Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
10.2.2 Frequency Distributions, the Normal
Distribution, and Dispersion . . . . . . . . . . . . . . 273
10.2.3 The Normal Distribution . . . . . . . . . . . . . . . . . 273
10.2.4 Central Tendency . . . . . . . . . . . . . . . . . . . . . . 275
10.2.5 Arithmetic Mean and Its Standard Deviation . . . 275
10.3 Inference: Hypothesis Testing, Types of Error, and
Sample Size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276
10.3.1 Hypotheses and Testing . . . . . . . . . . . . . . . . . . 276
10.3.2 Probability . . . . . . . . . . . . . . . . . . . . . . . . . . . 276
10.3.3 Types of Error . . . . . . . . . . . . . . . . . . . . . . . . 277
10.3.4 Confidence Interval and the Z-Score . . . . . . . . . 277
10.3.5 Sample Size . . . . . . . . . . . . . . . . . . . . . . . . . . 278
10.3.6 Example 1: Cranial Capacity in H. Erectus . . . . 278
10.3.7 Scale and Size: Adequacy of the Sample in
Geomorphology and Pedology . . . . . . . . . . . . . 280
10.4 Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
xii Contents

10.4.1 Example 2: Single Variable Data-Phosphorus in

Soils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
10.4.2 Data Analysis of More than One Variable . . . . 281
10.4.3 Covariation-Correlation, Causality or “Not” . . . 281
10.4.4 Least-Squares Analysis and Linear
Regression . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
10.4.5 Example 3: Alloys in Iron Age Metallurgy . . . . 284
10.4.6 “SHE” Measures of Variability of
Assemblages . . . . . . . . . . . . . . . . . . . . . . . . . . 285
10.4.7 Paired Comparisons or Paired Difference
Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . 286
10.5 Exploration: Multivariate Approaches to Large
Data Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
10.5.1 Exploring Correlation Among n Greater than
2 Variables: Some Multivariate Statistical
Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
10.5.2 Discriminant Function Analysis (DA) and
Principal Components Analysis (PCA) . . . . . . . 288
10.5.3 Cluster Analysis . . . . . . . . . . . . . . . . . . . . . . . 291
10.5.4 Chi-Square: “Nonparametric” Calculation of
Association . . . . . . . . . . . . . . . . . . . . . . . . . . . 292
10.5.5 The Next “New” Thing?: Bayesian Analysis and
Archaeological Age Determination . . . . . . . . . 293
11 Theory and Practice in Geoarchaeology: A Brief Discussion
with Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
11.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
11.2 Geoarchaeology and Landscapes . . . . . . . . . . . . . . . . . 298
11.3 Geoarchaeology at the Scale of Sites and Artifacts . . . . 299
11.4 Geoarchaeology and the Origins of the Deliberate Human
Use of Fire . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 300
11.5 Geoarchaeology and the Study of Natural Disaster . . . . 301
11.6 Geoarchaeology and a Brief History of Archaeological
Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301
11.7 Enter Probability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304
11.8 Geoarchaeology, Theory, and Practice: Examples . . . . . 305
11.8.1 The Medieval Norse Colonies of Greenland:
What Were They Thinking? . . . . . . . . . . . . . . . 305
11.8.2 Medieval Iceland to AD 1783–1784: Tephra and
Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306
11.9 Example 2: The Scull Shoals Mill Village Site, Oconee
River, Georgia, USA . . . . . . . . . . . . . . . . . . . . . . . . . . 307
11.10 Concluding Thoughts . . . . . . . . . . . . . . . . . . . . . . . . . . 309
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311
List of Figures

Fig. 2.1 Map of nearly 200 springs around Menngen Station (in red),
Northern Territory, Australia (Map drawn by Senior Elder
Yidumduma Bill Harney in 2010)
Fig. 2.2 Variety of fluvial terraces, paired and unpaired, created by ero-
sional and depositional cycles
Fig. 2.3 The Hjulstrom diagram. Flow velocity is juxtaposed to sediment
size to indicate “transport” (suspension) versus “sedimentation”
(deposition) for various sediment grain sizes
Fig. 2.4 Jekyll Island, Georgia, with overlay of a paleochannel or “Old
Channel” of the Brunswick River, occupied, in part, by Jekyll
Creek, today (Chowns et al. 2008, Fig. 9)
Fig. 2.5 Present-day view of Jekyll Creek looking eastward from the “Old
Channel” or “paleochannel” of the Brunswick River toward the
“Pleistocene core” per Chowns et al. (2008) (cf. Fig. 2.4).
Modern oyster reefs line the shoreline (Photograph by the
Fig. 2.6 (L) Illustration of the English Channel River (Redrawn from
Osburn 1916, Fig. 26); (R) early Holocene Doggerland (Redrawn
from Coles 1998)
Fig. 2.7 Lower. Maxilla of adult male mammoth dredged from the North
Sea off the Netherlands (Photograph by author)
Fig. 2.8 Paleochannels of the “Transect River,” Georgia Bight, Atlantic
Ocean, offshore South Carolina (see inset, lower left), as imaged
using multibeam sonar data (Stubbs et al. 2007; Harris et al.
2013). Inset, upper right, shows incision of the paleochannel
into Pliocene lithology
Fig. 2.9 (a) Shaded map showing the AD 563 turbidite deposits
(“déplacements des sediments,” in red) and the supposed propa-
gation of the tsunami (“raz-de-marée,” blue arcs) (Source. 24
Heures, Lausanne, March 8, 2011). (b) Shaded relief map (DEM)
(Swiss Federal Office of Topography) of the Lake Geneva. Water
depths are indicated by bathymetric contour lines (meters). The
563 AD turbidite is color shaded (orange-blue) as to extent and
sediment thickness (meters). Vessel tracks for seismic lines are
indicated. Sediment core locations are indicated by black

xiv List of Figures

squares. The inset bathymetric profile, X-Y, shows the gradient

of sediment thickness
Fig. 2.10 Schnidejoch (site circled) in the Bernese Alps of Switzerland
(Fig. 2, Hafner 2012)
Fig. 2.11 The first underwater archaeology dive—August 24, 1854, Lac
Léman (Lake Geneva). François Marie Etienne Forel, father of a
then 13-year-old François Alphonse Forel, is seated in the boat
with his back turned, presumably operating the air pump for the
diver (Adolphe Morot), while Frédéric Troyon holds a safety line
(Water color by Adolphe von Morot)
Fig. 2.12 Water color portrait of François Alphonse Forel in the room he
used as his study and laboratory, at his home in Morges on the
shores of Lake Geneva (Lac Léman), Switzerland (Watercolor
by Ernest Bieler, no date, private collection)
Fig. 2.13 Map by Bettis et al. (2003) showing loess distribution and
thickness in the Great Plains of the United States
Fig. 2.14 The Brady paleosol (Photograph courtesy of Charles
Fig. 2.15 The area of the Sahara Lakes (Adapted from: Saudi Aramco
World, vol. 65 (3), 2014)
Fig. 2.16 The Rift Paleolakes: NA Northern Awash, CB Chew Bahir, WT
West Turkhana, TH Tugen Hills, and LM Lake Magadi
Fig. 2.17 Combe-Grenal stratigraphy—Mousterian levels 24–54 are
shown in photo and inset. The rear portion of the inset fence
diagram (B) is correlated with the photograph, from Bordes
(1968), above. Levels 56–64 are shown in the elevation section,
below, which corresponds to the far left of the fence diagram
(Bordes 1972)
Fig. 2.18 West profile at Pech de l’Aze´ IV. This profile typifies the
deposits in this region. It consists of variously sized blocks of
limestone éboulis fallen from the roof and walls of the cave
within a sandy, iron-stained matrix. The sand is derived ulti-
mately from breakdown of the limestone and accumulated via
different sources from outside and mainly within the cave,
including slope wash, creep, and solifluction (Photograph by
Philippe Jugie)
Fig. 2.19 Tsunamigenic deposits for the Santorini eruption or eruptions,
17th c. BC, at East Beach cliff, Crete. The white line indicates the
erosional contact, termed the basal unconformity, between the
chaotic archaeological debris layer and the underlying geological
strata. Wall remains and dislocated building stones are present as
reworked material in the lower part of the archaeological layer
Fig. 2.20 Block diagram, with strata and folding illustrated (a) and
geological map (b) further illustrating synclinal, anticlinal
folding, strike-dip directions (From Allard and Whitney,
Environmental Geology Lab Manual, Wm. C. Brown, Fig. 4
Fig. 2.21 The first Landsat 8 image. Fort Collins, Colorado (Courtesy
List of Figures xv

Fig. 2.22 Three views of a historic slave cabin, Wormsloe Plantation,

Savannah, Georgia, USA. The upper two views illustrate
LiDAR views, and the lower view is a photograph of the cabin
with the LiDAR unit shown (LiDAR and photo images courtesy
Geospatial Mapping Center, the University of Georgia)
Fig. 2.23 Illustrating SfM. The upper view of an excavation was produced
using the program 123D Catch. The lower view is a digital-surface
view of the same unit using another program, Smart3D. Both
views are scalable and rotable (Images courtesy Geospatial
Mapping Center, the University of Georgia)
Fig. 2.24 A small, commercial drone (eBee by SenseFly) being preflight
programmed for a photographic survey. The photomosaic shown
on the right is the result of this particular flight (Photographs
courtesy of Chet Walker, shown programming the eBee drone)
Fig. 2.25 A Geo.pdf with the capability for embedding geographic coordi-
nate information. The red circles are location of selected test
locations across the survey area (Image courtesy of ARI,
Tallahassee, Florida)
Fig. 2.26 Linear band ceramic (LBK) sites (black dots) plotted relative to
loess soils (gray) in northwestern Germany (Adapted from Kuper
et al. 1977)
Fig. 3.1 Soil profile with commonly found horizons
Fig. 3.2 The Munsell color system. Using this diagram, an example soil
color such as “2.5YR5/4” can be understood as 2.5 yellow-red on
the hue circle, 5 on the vertical scale for value, and 4 on the
horizontal section for chroma
Fig. 3.3 Above, sand grains, roundness, and sphericity (Modified from
Powers 1953), below, sorting of sand grains (Modified from
Compton 1962, Manual of Field Geology)
Fig. 3.4 Example ternary diagram plots for, left, the United States and,
right, Great Britain and Wales
Fig. 3.5 Bounding in a spodosol. The A horizon is bounded as “clear to
abrupt”; that between the “A” and “E” is “gradual to diffuse”
(Photograph by the author)
Fig. 3.6 Disconformable (allo)contact between the Central Builders
member of the Wyoming Formation and the coarse gravel in
the cutbank of Cayuta Creek in Sayre, Pennsylvania, at river km
452 (Thieme 2003)
Fig. 3.7 Olduvai Gorge, Tanzania. Upper, the red sandstone of Bed III is
clearly visible as well as the top of Bed II, lower, stratigraphy of
Olduvai (Photograph courtesy of Dr. Sandra Whitney)
Fig. 4.1 Cutting out a soil monolith
Fig. 4.2 Rotary corer. Shown deployed in the excavation of a Roman era
cemetery in urban Tunis, Tunisia, the auger rapidly bores
through the clastic sediments. Photograph by Nina Šerman
Fig. 4.3 A Giddings hydraulic coring system. The operators are exchang-
ing a “Kelly (extension) bar” from the adjacent rack, to increase
the depth of the core. The vertical “mast” contains a hydraulic
xvi List of Figures

“ram” powered by a small engine (not shown) to push and pull,

utilizing the lever controls (red), the core barrels used to collect
sediments. The system is easily deployed using a vehicle as
shown. A typical core takes less than 30 min to collect
Fig. 4.4 Ewing piston corer
Fig. 4.5 Sediment sampling log
Fig. 4.6 Sediment box corer
Fig. 4.7 The sediment grab sampler
Fig. 4.8 (a) Above, shallow sediment corer; (b) right, a sediment grab
deployed at Jekyll Creek, Atlantic Intracoastal
Waterway (AIWW), Georgia, USA
Fig. 4.9 Sediment grab sample on a screen box
Fig. 4.10 Typical sequence of sondages used to systematically characterize
the landscape of the Plateau of Bevaix (Courtesy of the Archaeo-
logical Service, Canton of Neuchâtel)
Fig. 4.11 Typical stratigraphic sequence seen in the profile for a test unit.
The “complexes” listed represent sediment deposits of colluvium
and soils formed thereon
Fig. 4.12 The Hjulstrom diagram relating particle movement to size and
stream velocity
Fig. 4.13 Imhoff cone and soil auger sample. Suspected B horizon shown
separating in Imhoff cone
Fig. 4.14 Phosphate intensity plot for Great Zimbabwe using the
colormetric method (Courtesy of Paul Sinclair)
Fig. 4.15 The Neolithic site of Arbon-Bleiche 3 demonstrating the use of
micromorphology; (a) site map with the ground floors of the
houses and with the eight phases of the village. The pilings of
the houses are marked as black dots. The position of the profile
section B and the profile photograph (c) are marked with white
arrows and highlighted in red. (b) The west profile through house
A showing the typical sequence with two main organic cultural
layers divided by a sandy layer. (c) The Arbon-Bleiche 3 thin
sections of the column M 1030 (profile B) with the micromor-
phologically identified phases of installation, organic
accumulations beneath the house’s floor and inwash of sand
from the hinterland surrounding the village (Fig. 4.2 in Ismail-
Meyer et al. 2013, used with permission)
Fig. 4.16 Typical triporate pollen grains: left, beech; right, oak
Fig. 4.17 Various phytolith types. Arrows indicate “cross-body”
chlorodoid types, whereas a festicoid type may be present, center
of the photomicrograph. The photomicrograph courtesy Dr.
Liovando Marciano da Costa. The two close-ups of chlorodoids
and festicoid types are from Ball et al. (1996)
Fig. 5.1 General multielectrode resistivity system. This system can be
programmed for both Wenner and Schlumberger array
List of Figures xvii

Fig. 5.2 Resistivity pseudosection of Mound A, Etowah Mounds State

Park, Cartersville, Georgia, USA. This 20 m-high Mississippian
period temple mound was built sometime prior to 1250 AD. It
has been the subject of several geophysical studies beginning in
1993 with the author’s initial radar studies of this and adjacent
mounds B (Garrison 1998). This figure is the result of a later
electrical study (Garrison et al. 2005). The near-surface features
in red are well-documented “robber trenches” but the deeper
anomalies, in green, are prehistoric structural features
Fig. 5.3 The Cherokee female seminary, ca. 1,848. Courtesy of the
Oklahoma Historical Society
Fig. 5.4 Three plots of geophysical data from surveys of the site of the
Cherokee Female Seminary, Tahlequah, Oklahoma. The upper
plots are, respectively, a plan view of quadrature phase data
(330 Hz) and the isometric view of the same data. The lower
image is a GPR “time slice” of the same area showing good
correlation between the two sensors. Architectural interpretation
has been added to the radar plot. The area outlined in red is
buried ruin of the building
Fig. 5.5 Rose Cottage, from a painting of the house done from memory.
Oklahoma Historical Society
Fig. 5.6 A GeoPDF showing the location of UGA survey grids and other
landmarks at the Rose Cottage site
Fig. 5.7 Multisensor representations—left and above, magnetic; right,
ground radar—of an abandoned Native American cemetery,
Park Hill, Oklahoma, Cherokee Nation, USA. The magnetic
anomalies correspond to rows of grave pits while the radargram
indicates three sides of a buried stone wall. The latter view is
1.74 m depth and clearly shows the rectangular subsurface fea-
ture interpreted as a remnant of a stone or masonry wall. The
magnetic features (orange red) in the plots show linear patterns.
The separation between the anomalies indicate individual graves
Fig. 5.8 A profile view of radar data. The image is of a Chumash pit
dwelling, roughly 2 m in width, as illustrated in the artist con-
ception adjacent. This is an idealized view of an actual archaeo-
logical site on the island of Santa Cruz, California, USA, where
the author conducted research on the late twentieth century
Fig. 5.9 The effects of refraction on the radar wave as the dielectric
increases (upper) or decreases with depth (lower) (Adapted
from Goodman 1994, Fig. 5.6)
Fig. 5.10 A typical survey grid and GPR units; foreground, a control unit
and antenna using a survey wheel; background, another control
unit mounted, along with an antenna, on a survey cart. Tapes
indicate transect lines (Photograph by the author)
Fig. 5.11 Examples of post-processed radar data. Both represent a 2004
radar survey of a Roman villa site in western Switzerland. The
plot on the left shows the burial depth at almost a meter, while the
plot on the right highlights architectural details of the villa’s
principal building, the pars urbana
xviii List of Figures

Fig. 5.12 Study area, in red, for the Sabine River paleovalley study, north-
ern Gulf of Mexico, USA
Fig. 5.13 Acoustical profile and schematic view of sonar images produced
by an acoustic sub-bottom profiler of a drowned paleochannel in
15 m water depth in the Gulf of Mexico. Sediment vibracore
locations (Fig. 9.14) are indicated. This ancient estuary is
illustrated its drowning at transgression ca. 6000 BC (Pearson
et al. 2008)
Fig. 5.14 Left, sampling the drowned Sabine estuary shown in Fig. 9.13
with a vibracore; right, analyzing the sediments in a vibracore
taken from the estuary (Pearson et al. 2008)
Fig. 6.1 Ancient and historic jewelry (Source: National Geographic
Magazine). 1 Pendant of perforated shell ca.1000 BC; Syria.
2 Seal, shell, 7000–4000 BC; Syria. 3 Tubular bead, bone,
3000 BC; Syria. 4 Eye pendant, glass, 2000 BC; Egypt. 5 Pen-
dant, glass, at least exterior; 600 BC; Punic-Phoenician? 6 Object
made of carnelian, 600 AD; Central Asia. 7 “Bat wing” object—
“parietal”? Jasper; Mesoamerica. 8 Bead, lapis lazuli, 1000 AD;
West Asia. 9 Object fashioned from fossilized shell, 1000 AD;
West Asia. 10 Incised disk, shell, 1900 AD; Mauritania. 11 Kiffa
bead, powdered glass, 1930 AD; Mauritania. 12 Glass art
depicting Marilyn Monroe, 2004; USA
Fig. 6.2 Dimension stone. Middle Jurassic sandstones. Dun Beag Broch,
west coast of the Isle of Skye, Scotland. Historic Scotland
Fig. 6.3 The rock cycle
Fig. 6.4 Ternary diagram using Folk’s sandstone classification (1948,
1980) (Reproduced from Zahid and Barbeau (2011)). Arenitic,
arkosic, and lithic categories are subdivided according to the
percentage of each within the diagram. Subcategories, such as
lithic arkose and feldspathic litharenite, are illustrated following
a 75–50–25 % breakdown
Fig. 6.5 Sandstone Pictish standing stone. Perthshire, Scotland. So-called
Class 1 Stone by virtue of only “pagan” Pictish symbols used in
its decoration (Photograph by author)
Fig. 6.6 Two thin-section views of arenitic sandstones. Left, XPL
showing undulatory extinction of the quartz grains (black
grains). Crystals of sparite cement are clearly seen along the
edges of holes (blue voids) in the thin section. The quartz grains
show shattering due to sawing. Right, XPL image showing
“twinning” of feldspar grains (Left image by author, right
image courtesy the University of North Carolina-Chapel Hill)
Fig. 6.7 Neolithic Period stelae carved from sandstone, from
the megalithic site of Petit-Chasseur, Sion, Valais, Switzerland
(Musée d’histoire du Valais, Sion, photograph by R. Hofer)
Fig. 6.8 Rear, a siltstone, green “molasse” dimension stone; foreground, a
“grès coquille” or bioclastic limestone (Photograph by the
List of Figures xix

Fig. 6.9 Thin-section views of Grès coquille bioclastic limestone, western

Switzerland. The high density and lamination of shell and quartz
are clearly visible (Image courtesy of the Archaeological Ser-
vice, Canton of Fribourg, Switzerland)
Fig. 6.10 The Roman millstone quarry at Châbles-les-Saux, Switzerland
(Photograph courtesy of Dr. Timothy A. Anderson (shown in
lower left))
Fig. 6.11 Backscatter electron (BEI) image showing dolomite rhombs in
thin section. A large rotalliid foraminifera (center) is surrounded
by dolomite rhombs; large apatite grain appears in the upper,
center portion of the image (SEM image by author)
Fig. 6.12 Novaculite. Left, in thin section; right, projectile points
(Flenniken and Garrison 1975; Used with permission of Maney
Fig. 6.13 Pillow lava basalts in road cut, Curaçao, the Netherlands
Antilles. Clipboard for scale (Photograph by author)
Fig. 6.14 Bluestone rhyolite quarries, Fishguard Volcanic Group at Pont
Saeson, North Pembrokeshire (Map modified from Fig. 1, Bevins
et al. 2011)
Fig. 6.15 (a) Thin-section views of strongly foliated rhyolite with small,
dark (1 mm) lenses elongated in the plane of foliation. Sample
142/1947.22 from Wilsford, described originally by Stone (1948)
and subsequently by Ixer and Bevins (2010) (Courtesy of Profes-
sor Michael Parker Pearson). (b) Photomicrograph of strongly
foliated rhyolite with small, dark (2 mm) lenses elongated in
the plane of foliation. Sample SH08/265 from the 2008
Stonehenge excavations (Courtesy of Professors Tim Darvill
and Geoffrey Wainright). (c) Photomicrograph of strongly foli-
ated rhyolite with small, dark (0.5–1.5 mm) lenses elongated in
the plane of foliation. Sample PS8 from the Fishguard Volcanic
Group at Pont Saeson, North Pembrokeshire (cf. Fig. 6.14)
(Photomicrographs per Richard E. Bevins; the National Museum
of Wales and Wiltshire Studies)
Fig. 6.16 Metamorphic facies by pressure and temperature. “Low” tem-
perature or “blueschist” facies are typified by steatite (talc);
marbles are more “greenschist” facies
Fig. 6.17 The “almond-eyed korai” at the Acropolis Museum, Athens
(Photograph by the author)
Fig. 6.18 Livia. Marble portrait (Courtesy the Estate of Dr. Norman Herz)
Fig. 6.19 Left, Botticino Agglomerate limestone (Italy); right, Rojo
Alicante limestone (Spain)
Fig. 6.20 Limestone used for sculptural relief of a Roman tower tomb,
Aventicum, Germania Superior (modern-day Switzerland). The
subject is that of Greek mythology, Triton abducting a sea nymph
or Nereide. Roman Museum of Avenches (Photograph by the
Fig. 6.21 Soapstone Ridge, Georgia, USA. Upper, bowl “blank” in situ;
left, thin section showing high birefringence of talc; right, bowl
Fig. 6.22 Petroglyphic boulder. Granitic gneiss
xx List of Figures

Fig. 6.23 Menhir carved from schist discovered in the Canton of Neuchâtel,
in 1997, during excavations along the then proposed autoroute A5,
in the vicinity of the village of Bevaix. The stone is clearly carved
as stelae and is dated to the Neolithic Period, ca. 5000 BP. The
height is 3.2 m and the weight is 2,800 kg. On exhibit, Latenium
Museum, Neuchȃtel, Switzerland (Photograph by the author)
Fig. 7.1 Weathering of silicate minerals (After Horger 2000). Granites
gneisses of the Appalachians produce kaolinite with gabbros
yielding smectite and olivine-rich rocks yielding montmorillon-
ite via the mechanisms illustrated below
Fig. 7.2 The silica and alumina tetrahedral/octahedral units
Fig. 7.3 Closed kiln
Fig. 7.4 Phase changes in pottery
Fig. 7.5 Forming the “wet” pot. Note appliqué
Fig. 7.6 Modern Navaho pottery (Photograph by Robert Hill)
Fig. 7.7 Two thin sections of late prehistoric earthenwares, Georgia,
USA. In the left, firing differences—black indicating reducing
conditions and red-brown, oxidizing. Such differences are not
unusual for pit kiln-fired wares. The thin section on the right
illustrates more uniform firing conditions as well as a more
uniform grain-temper sizes. The temper in both is primarily
quartz-dominated medium sand. Both photomicrographs were
in PPL (Courtesy of Mark Williams)
Fig. 7.8 Sandhills regions, North Carolina. Clay sample locations are
marked (Fig. 7.3, p. 34, Herbert and McReynolds 2008)
Fig. 7.9 Various pottery shards from the Sandhills regions. Pottery
samples from the Kolb site (38 Da75): (a) JMH051, Yadkin
Fabric Impressed; (b) JMH052, Yadkin/Hanover Fabric
Impressed; (c) JMH053, Yadkin/Hanover Cord Marked; (d)
JMH054, New River Cord Marked; (e) JMH055, Yadkin Cord
Marked; (f) JMH056, New River Fabric Impressed (flex warp);
(g) JMH057, New River Cord Marked; (h) JMH058, Cape Fear
Fabric Impressed; (i) JMH059, Cape Fear Fabric Impressed; (j)
JMH060, Hanover I Fabric Impressed (Fig. 7.8, p. 28, Herbert
and McReynolds 2008)
Fig. 7.10 Ternary plots of Sandhills’ clay groups (Fig. 7.2, p. 117, Herbert
and McReynolds 2008)
Fig. 7.11 The total ion chromatogram (TIC) of maize lipids. This shows
primarily the plant and/or fish fatty acid ratio (more unsaturated
C18:1 than saturated C16:0), plant biomarkers (sitosterol), and
acylglycerols. Triacylglycerols are referred to as TAGs,
diacylglycerols as DAGs, and monoacylglycerols as MAGs
(Reproduced from Reber and Evershed 2004, Fig. 7.1)
Fig. 8.1 Schematic of X-ray diffraction. The diffractometer measures the
“d-spacing” between the crystal planes following Bragg’s Law, n
λ = 2 d sin φ, where n is the integer representing the peak of
interest, d is the lattice spacing of the specific crystallized min-
eral or element, y is the angle between the scattered X-ray and the
incident radiation, and l is the wavelength of the incident light
List of Figures xxi

Fig. 8.2 Micro-X-ray fluorescence spectrometer examining painted and

unpainted areas on the Orpheus Relief (Photograph by Jeff
Fig. 8.3 Typical handheld pXRF instrument. Bruker Tracer pXRF model
Fig. 8.4 Electron and X-ray emission for EMPA showing the relationship
between penetration depth, energy, and the sample’s atomic
number (Z)
Fig. 8.5 Left, green transfer printed whiteware shards; right, flow-blue
transfer printed whiteware shard (scale bars = 1 cm)
Fig. 8.6 Schematic of atomic absorption spectrometer
Fig. 8.7 Schematic for a modern thermal ionization mass spectrometer
Fig. 8.8 Lead isotope rations of lead curse tables and those of potential
ore sources in Greece, England, Spain, Tunisia, and Wales
(figure 5, Skaggs et al. 2012, p. 980)
Fig. 8.9 National Gallery of Art, Aphrodite
Fig. 8.10 Distribution of isotopes for various Aphrodites relative to known
isotopic fields for ancient Mediterranean marble quarries (Herz
and Pike 2005)
Fig. 8.11 Schematic of an inductively coupled plasma (ICP) spectrometer
Fig. 8.12 Schematic of a laser ablation inductively coupled mass spec-
trometer (LA-ICP-MS) system
Fig. 8.13 Upper, plan of Dalles Floor (green); lower, Dalles Floor (green)
in stratigraphic section
Fig. 8.14 Coin mold fragments and coin flans from the Dalles Floor,
Fig. 8.15 ESR spectra of marble showing typical hyperfine splitting of
manganese (Mn)
Fig. 9.1 Central Vinča culture area (left); detail showing early Balkan
mines (right)
Fig. 9.2 Chalcopyrite ore with a replica of Neolithic copper axe shown
(Photograph by the author)
Fig. 9.3 Early metal site of Vucedol, located overlooking the Danube
River, in eastern Croatia (Photograph by author)
Fig. 9.4 Experimental archaeology. Melting bronze ingots for casting in
the stone mold shown in the foreground (Courtesy Latenium,
Hauterive, Switzerland)
Fig. 9.5 Experimental archaeology. Cast duplicate of a bronze spear
(Courtesy Latenium, Hauterive, Switzerland)
Fig. 9.6 Experimental iron metallurgy. A pit forge, stone hammer, and
anvil (Photograph by author)
Fig. 9.7 A hypothesized crucible steel—“wootz”—process
Fig. 9.8 The Ncherenje mine on the Malawian Nyika Plateau (Photograph
by David Wenner)
Fig. 9.9 One of the deep pits used to extract ore from the Mbiri iron
mines, northern Malawi (Wenner and van der Merwe 1987)
(Photograph by David Wenner)
xxii List of Figures

Fig. 9.10 Malawian iron manufacturing, 1982, Chulu, Malawi; above, left,
furnaces prior to firing; above, in operation; above, right, micro-
graph (1000) showing iron (bright spheres) in bloom matrix;
right, bloom showing large (macroscopic) iron inclusions pro-
duced by the smaller, forced-draft furnace (Photographs by
Nicholas van der Merve)
Fig. 9.11 Characteristics of an ophiolite complex: axial magma chamber;
pillow lava basalts; sheeted basaltic dikes; layered gabbro;
Fig. 9.12 European metal deposits with central European mines
Fig. 9.13 Copper ore, azurite (left); native copper (right)
Fig. 9.14 Schematic of Laurion’s geology of alternating schist and lime-
stone with metal bearing contacts (1–3) between them. The ore
contacts, 1 and 3, the richest. Argentite: “silver ore” is associated
with the lead ore, cerussite (PbCO3). The principal silver ore
mineral is silver chloride, cerargyrite (AgCl), and silver sulfide
(Ag2S). The latter is mixed with galena, PbS
Fig. 9.15 Back-scattered electron image of iron slag
Fig. 10.1 The normal distribution. (a) The area under the curve is, by
definition, 1; (b) the measure of central tendency is the mean,
which is, again, by definition, zero; (c) the curve is perfectly
symmetrical about the mean; (d) the standard deviation, along
with mean, completely defines the normal curve; (e) the mean,
mode, and the median all have the same value. The probability
p(x) of occurrence of any observed values from 1 to 2 and 8 to 9
(a two-tailed σ is shown in shaded areas). The range of values in
this example is 0–10 with the mean (x) = 5.2; 1.5σ indicates the
standard deviations from the mean (Adapted from Waltham
1994, fig. 7.7, p. 119)
Fig. 10.2 Linear regression. Two correlations between the observed alpha-
recoil track (ART) density in mica inclusions used as a tempering
agent for prehistoric Hohokam pottery (“Sacaton Red,” “Vahki
Plain,” etc.) versus age in years before the present (BP). The line
width for each pottery type track density value is the standard
deviation (Garrison 1973; Garrison et al. 1978)
Fig. 10.3 Pair-wise comparison of mtDNA sequences of observed
differences in human-Neanderthal sequences and human-chim-
panzee sequences
Fig. 10.4 Bivariate plot of elements within Pentelic marble. Two ellipses
denote specific marble groups based on discriminating elements
(Courtesy Scott Pike)
Fig. 10.5 Left: Spode blue printed design. Right: back-scattered electron
(BSE) image of dye, with brighter areas those of high cobalt
Fig. 10.6 Binary diagram of PbO and SiO2 data for all Spode samples.
Clusters of “open” samples manufactured before and after 1833,
correlating with a change in ownership at the Spode factory
List of Figures xxiii

(cf. Table 10.3); the “unknown” (closed) samples are highlighted

in gray and occur in their respective manufacturing group
Fig. 10.7 Three chi-square distribution curves. Note the higher degrees of
freedom (df = 19) approximate the normal distribution
Fig. 10.8 For an assumed set of 6 samples #1 to #6 from the Hallstatt
period (750–400 BC) with ages of 750 BC (#1), 745 BC (#2), 740
BC (#3), 735 BC (#4), 730 BC (#5), and 725 BC (#6) indicated
by vertical thin lines, the corresponding 14C ages were looked up
in the calibration curve. Due to the flatness of the calibration
curve, we get the same 14C age of 2455 BP for all six samples.
After adding a random scatter of 40 year, we obtain the follow-
ing 14C ages: 2546 BP, 2490 BP, 2402 BP, 2446 BP, 2386 BP,
and 2491 BP. By individual calibration, the samples can no more
be assigned to distinct regions. The resulting probability
distributions (gray curves) rather cover the whole Hallstatt
period. These probability distributions correspond to our
simulated 14C measurement data. After the Bayesian sequence
algorithm is applied, one can see its tendency to divide the period
into six parts of equal size (black curves). Due to the flatness of
the calibration curve, the general shape of the individually
calibrated and of the “sequenced” probability distributions is
the same which true ages ever are assumed. In our example, the
posterior 95 % confidence intervals of samples #4, #5, and #6 are
not in agreement with their assumed true ages. All the
calculations (single calibration and sequencing) were performed
with OxCal v2.18 (Ramsey 1995) using the INTCAL98 14C
calibration curve (Stuiver et al. 1998). The program normalizes
the individual and the “sequenced” probability distributions to
the same maximum (probability) value (Italics added for
Fig. 11.1 Painting of the Scull Shoals mill, in the nineteenth century; at its
Fig. 11.2 A 1.5 m sediment core, using a cryoprobe at the Scull Shoals site,
right (Photographs by the author)
ThiS is a FM Blank Page
List of Tables

Table 3.1 Designations of master soil horizons and subordinate symbols

for soils (Adapted from Olson 1981)
Table 4.1 The standard grain-size scale for clastic sediments
Table 4.2 Sample preparation for pollen analysis
Table 4.3 Palynological composition of samples from Gray’s Reef
Table 4.4 Process of phytolith extraction
Table 5.1 Resistivity of common materials
Table 5.2 Magnetic susceptibilities of common rocks and minerals (cgs
Table 5.3 Magnetic properties of soil minerals
Table 5.4 Typical values of radar parameters of common materials
Table 6.1 Classification of sedimentary rocks
Table 6.2 Common igneous rocks
Table 7.1 Climate and weathering of silicate precursor minerals
Table 7.2 Consistency water %
Table 7.3 Characteristics of ceramic bodies
Table 8.1 a: Green sherd, WDS data. b. Blue sherd, WDS data
Table 8.2 Isotopes utilized in NAA analysis of coins and molds from the
Table 8.3 CL and other characteristics of selected white marbles
Table 8.4 Instrumental techniques and their detection limits
Table 9.1 Major mineral processes and deposits by origin (Carr and Herz
Table 9.2 a: Whole-rock chemical analyses of samples from the Mbiri
mine. b. Whole-rock chemical analyses of samples from the
Ncherenie mine
Table 9.3 Principal metallic ores
Table 10.1 Scales and statistical measures
Table 10.2 Compositional data (in %) for Iron Age artifact
Table 10.3 Spode pottery statistical analysis (Douglas 2000)
Table 10.4 Contingency table for winged versus socketed axes


This book is a survey of techniques used in Pleistocene and Holocene; the interrelation of
archaeological geology or as it is more widely diet, culture, and environment; and the relation
known, today—geoarchaeology. It is less a dis- of artifact raw material to site selection, trade
cussion of theory or methodology with regard to route development, and ancient technology.
the various geological techniques that are The acceptance of earth science techniques
presented. It is not an exhaustive presentation of into archaeological study has been gradual but
the diversity of earth science methods that can be has grown to a point that we now have
utilized in the service of archaeology. Earth sci- university-level programs in archaeological sci-
ence can be used in many ways in archaeology. ence, geoarchaeology, and/or archaeological
Certainly there are other ways to approach the geology. As Norman Herz, in Geological
multiplicity of archaeological questions that con- Methods for Archaeology (1998), stressed, it is
front the student of the past, and this book does not less what one calls the use of earth science
presume to suggest that the techniques presented techniques in archaeology than what one realizes
herein are the only ones appropriate for addressing from their application. The practice of modern
the many facets of archaeological inquiry. archaeology demands an understanding of earth
The Penrose Conference, in 1987, sponsored science. One does not have to look any further
by the Geological Society of America (GSA), than that of the importance of, first, Sir John
archaeological geology was a rapidly growing Evans and Joseph Prestwich’s and then Sir
interdisciplinary field in which archaeologists Charles Lyell’s 1859 visits to the gravel beds of
and geologists attempt to use the tenets of both the Somme River. In the Somme gravels,
their sciences to solve problems of mutual inter- Boucher de Crève Coeur de Perthes, in 1837,
est. This involves the application of the and, later, Jèrôme Rigollot had painstakingly
principles of geomorphology, sedimentation, iso- demonstrated the contemporaneity of ancient
tope geochemistry, and petrology in the solution stone artifacts and extinct paleontology. Their
of problems of interest to both. As examples, the visits brought the newly developed insights of
following may be cited: the effect of climate on the science of geology to the interpretation of
sites and site selection; the extent to which the Somme deposits by which greater scientific
archaeological sites may subsequently be validity was a result. This interplay of archaeol-
modified by erosion, sedimentation, and tectonic ogy and geology has resulted in similar advances
activity; the application of geophysics in the in our understanding of our human past.
search for and dating of sites; the application of Charles Robert Darwin, himself, wrote three
dating techniques especially as related to the separate treatises on the nature of archaeological

# Springer-Verlag Berlin Heidelberg 2016 1

E. Garrison, Techniques in Archaeological Geology, Natural Science in Archaeology,
DOI 10.1007/978-3-319-30232-4_1
2 1 Introduction

materials in the soil (1837, 1844, and 1881). now vastly longer than the other two ages. John
Charles Darwin’s so-called worm book Lubbock, of Great Britain, proposed dividing the
influenced many early researchers who, follow- Stone Age into two, “New” and “Old,” hence the
ing his lead, demonstrated how soil biota term paleo (old) and lithic (stone). By extension,
mechanically generate new strata and soil the “New” Stone Age was Neolithic or “New
horizons, as well as blur or destroy them (Leigh Stone.” Lubbock and other workers used differ-
2001; Johnson 2002). The title of the first was On ing criteria to define the two: the Neolithic was
the Formation of the Vegetable Mould and that defined in terms of industry, while the Paleolithic
of the last was The Formation of Vegetable was based on fauna—“reindeer period,” “ele-
Mould, which is misleading to the modern reader phant period,” etc. Although a British prehisto-
unless they understand that “vegetable mould” rian coined the term, the French created the
was a nineteenth-century term for topsoil chronology.
(Johnson 1999). Darwin’s essays examined the Edouard Lartet (1801–1871) developed the
relation of soils, earthworms, and artifacts. first useful chronology for the Paleolithic, using
Darwin’s question centered on why archaeo- paleontology as a guide. He divided Quaternary
logists invariably dug for materials that, in the (Pleistocene) into the Cave Bear Period, the Ele-
case of those of Late Antiquity, such as Roman, phant Period, the Rhinoceros Period, and the
should not be buried at all. We would call his Reindeer Period, from oldest to youngest.
study, today, a study of bioturbation or how the “Upper” Paleolithic cultures lived in the Rein-
action of soil-dwelling organisms, such as the deer Period, for example. DuPont reduced these
earthworms, creates soils. Bioturbation is not to two—“the Mammoth, Cave Bear Period” and
limited to worms (Johnson 1990). Larger “the Reindeer Period.”
burrowing animals such as gerbils, rodents, and Henri Breuil and Gabriel de Mortillet pro-
insects play a large role in this phenomenon. In duced subsequent chronologies that quickly
fluvial, lacustrine and marine environments’ replace Lartet’s at least in France, but even in
other burrowing organisms such as shrimp, the early twentieth century the British were using
clams, and worms play the same role. To Darwin, Lartet’s scheme.
bioturbation worked to bury artifacts (Leigh Lartet based his chronology on fauna ala
2001). His observations, however well worked Lubbock—not artifacts. This is a paleontological
out, were lost on antiquarians and later archaeo- approach. Lartet’s model, also, was regional in
logists mainly due to miscommunication. scope—not universal for all areas or regions. In
Geologists and soil scientists “spoke” a different this Lartet was accidently more correct than later
language or jargon from that of the archaeo- workers and their Paleolithic chronologies—
logists. In today’s geoarchaeology, we recognize which had differing timing of the Paleolithic
the importance of both natural burial processes phases in differing parts of the world.
along with those of erosion. By these two pro- Gabriel de Mortillet (1821–1898) recognized
cesses, two archaeological surfaces are created. the great temporal length encompassed by the
Following Darwin’s 1859 publication of Origin Paleolithic. He did not know exactly how much
(of Species), Charles Lyell, one of the founders time but geology clearly inferred this. He
of modern geology, published The Geological published his ideas in 1869. His divisions of the
Evidences of the Antiquity of Man (1863)— Paleolithic were simple: Acheulean, Mousterian,
conceivably the first book on geoarchaeology. Solutrean, and Magdalenian. With two additions,
After the great time depth of human antiquity in his lifetime, these basic divisions are used
was established in Europe after 1859, today.
archaeologists were faced with integrating this This simple system rested on some complex
greater temporality into the “3-Age System” of theoretical underpinnings. De Mortillet rejected
Denmark’s Christian Thomsen—Stone, Bronze, Lartet’s “faunal” ages—“Reindeer,” etc. He did
and Iron. It was obvious that the Stone Age was so for two reasons: first, paleontologists did not
1 Introduction 3

agree on any single scheme for faunal An integrative framework that incorporated
successions, and, secondly, instead of using the geology into the overall explanatory perspec-
presence-absence of species, Lartet used the rel- tive of archaeological research was necessary,
ative abundance of species. How to determine and it was a product of the efforts of numerous
this? Types of fauna will vary according to the workers across the globe in the twentieth century.
site and context. De Mortillet’s solutions: Stone Working at the Galilean cave sites of Skhul and
Tools. His divisions were materially based Tabun, Dorothy Garrod, the British archaeolo-
instead of paleontological. De Mortillet’s scheme gist, and Dorothy Bate, the British paleontolo-
was cultural-historical in concept—culture his- gist, combined to work out the sedimentological
tory through typology (of stone tools). French sequences for these Paleolithic deposits. Both
prehistorians, like many of the nineteenth- sites, excavated by Garrod between 1929 and
century researchers, were trained in geology. 1934, revealed the richness of Middle and
De Mortillet simply declared it “(prehistory) to Upper Paleolithic archaeology in the Levant.
be an extension of geology” (Daniel 1967). Into Garrod found both Neanderthal and
the mid-twentieth century, French prehistorians “Anatomically Modern Humans” (“AMH”) in
continued to hue to a “paleontological” paradigm the caves but not in the same cave. Tabun was
for the conceptualization of their excavations. “Neanderthal,” while more modern yet contem-
Strata were viewed biostratigraphically but mate- porary hominin occupations were identified at
rial finds—tool types ( fossile directeurs), aka the Skhul (McCown and Keith 1939). These
“biota”—separated antecedent and subsequent researchers recognized the importance of the
units (Sackett 1999). sedimentological context relative to the interpre-
Stratigraphy was used by North American tation of these important sites.
researchers in the mid-to-late nineteenth century. The dichotomies, specialty and gender based,
Rapp and Hill (2006) give a thorough review that existed in prewar scientific studies of Paleo-
of this in their textbook. Some of this historical lithic sites are illustrated at the Mount Carmel
context bears repeating here. In addition, like the caves. Garrod and Bate were female, albeit Uni-
French pioneers of Paleolithic archaeology, versity trained, whereas the paleontological
Americans used paleontological principles as description of the Skhul and Tabun hominins
well in their work. George Frederick Wright fell to Theodore McCown and Arthur Keith, at
had advocated for an American Paleolithic Garrod’s behest, we can be sure. These two
based in part on the geological contest of artifacts anatomists found the nearest affinities of the
(Wright and Haynes 1892). The question of a Skhul hominins, who they termed the Mount
geological context for this American Paleolithic Carmel people, with those found at Krapina in
at the time of Wright hinged in large part on Croatia “. . .through the anatomy of the Mount
working out the glacial sequence for North Carmel people there runs a substratum of
America at a level equivalent to that of Europe. characters which link them to the Neanderthal
The European glacial sequence owed much of its type.” The Tabun type, plainly Neanderthaloid,
precedence to the pioneering work of Louis may be placed at one extreme and the Skhul type
Agassiz in the 1840s. Along with Wright, at the other (McCown and Keith 1939). In these
W.H. Holmes, successor to John Wesley Powell conclusions, they were proven correct. This said,
at the United States Geological Survey the divisions between archaeology, geology, and
(1902–1910), recognized the importance of this paleontology were evident. Goldberg and Bar-
sequence as a reliable context for archaeological Yosef (1998); Meignen et al. (2001); Vita-Finzi
research (Holmes 1892, 1893). Key to working and Stringer (2007) represent three major
out the last glaciations sequence was Newton modern-day geoarchaeological studies of both
Horace Winchell, State Geologist for Minnesota, Tabun, Skhul, and, also, Kebara Caves.
later integrating this knowledge in his book, The Even in Tanzania, where Louis and Mary
Aborigines of Minnesota (Brower et al. 1911). Leakey toiled for years before finally unearthing
4 1 Introduction

the early hominins that made them, perhaps, the “geoarchaeology,” while Rapp and Rapp and
most recognized paleoanthropologists of the Gifford (1985) helped coin the alternate term
twentieth century, geology was necessary. Hans “archaeological geology.” As Hassan pointed
Reck worked with Leakey early on roughly and out in his 1979 review, geoarchaeology had
at the same time Garrod and Bate were moved away from a set of methods in support
excavating in Galilee, in the early 1930s. By of archaeology to one more focused on the
being trained in both volcanology and paleontol- relationships of local and regional geology to
ogy, Reck was well suited for working out the settlement locations together with the nature of
stratigraphy of the Olduvai Gorge in two site formation processes (Hassan 1977,1978;
expeditions to the site in 1913 and, again, with Schiffer 1982). Butzer (1971) proposed a “con-
Leakey in 1931. textual archaeology” of which the geological and
While much later in terms of evolutionary geomorphic played no small part. He—Butzer—
time, discoveries in New Mexico, in 1927, at in 1975, wrote that the geologist should be well
the Folsom site, extended America human antiq- acquainted with the goals and aims of archaeol-
uity into the Late Pleistocene. This discovery of ogy, while the archaeologist should be fully
extinct megafauna together with distinctive lan- aware of the potential of geological
ceolate tools propelled a renewed cooperation investigations. Specific geoarchaeological
between geology and archaeology (Howard questions should arise from a consideration of
1935). Edgar Howard integrated several lines of geological variables in man-earth interactive
research—paleoenvironmental, archaeological, systems (Hassan 1979).
and geological—into an early synthesis of what Indeed, this textbook simply reinforces simi-
became to be known as the “Paleo-Indian lar calls by others within the geology of archae-
Period.” Better known than Howard, and Elias ology. These include Karl Butzer (2008), George
Sellards, who had examined the association of (Rip) Rapp, Vance Haynes, Jack Donoghue,
human remains and megafauna in Florida, was Brooks Ellwood, Reid Ferring, William Farrand,
Kirk Bryan. Joel Schuldenrein, and Julie Stein to name but a
Bryan, according to Rapp and Hill (1998), few of these scholars. As previously noted,
along with Ernst Antevs, dominated the interface above, Rapp, in his books written with John
between geology and archaeology in the 1950s. Rapp and Gifford (1985), and Christopher Hill
Their work continued primarily in the American (1999, 2006), together with works by Davidson
Southwest at Clovis, Chaco Canyon, and other and Shackley (1976); Gladfelter (1977); Waters
locations such as Midland, Texas, where defini- (1992); Stein and Linse (1993); Stein and
tive Pleistocene-aged human remains were Farrand (2001); Bettis (1995); Pollard (1999);
excavated (Wendorf et al. 1955). In the Goldberg and MacPhail (2006), have provided
mid-twentieth century, in the United States, the clear models for modern geoarchaeological
use of geological methods at the Midland site by inquiry and interpretation. Archaeological geol-
Wendorf and his co-workers, Claude Albritton, ogy has become more than simply “method ori-
Alex Krieger, and John Miller, in the 1950s, set ented.” While a methodology is clearly a
an important and enduring example of how geol- technique-driven enterprise in the service of
ogy and archaeology buttressed each other to both archaeology and Quaternary studies, it has
mutual advantage (Wendorf et al. 1955). developed a theoretical contextualization
Karl Butzer is generally recognized as one of (Goldberg and MacPhail 2006). This has been
the first, in the mid-twentieth century, to inte- principally a recent development. In the first
grate geology and archeology (Butzer 1971; edition of this textbook, a draft chapter devoted
Hassan 1979). Gladfelter proposed a definition to a discussion of theory in archaeological geol-
for geoarchaeology in 1977. He, along with ogy was deleted in the editorial process.
Butzer together with Davidson and Shackley As a consequence of that editorial decision,
(1976), was the first to call the field the first edition of Techniques for Archaeological
1 Introduction 5

Geology had no body of theory to discuss. While studies through its focus on landscapes and
geoarchaeology has no great pantheon of past formational processes over a wide range of spa-
practitioners such as older disciplines like phys- tial and temporal scales.” Special issues of
ics and chemistry (Garrison 2001), it has begun important journals such as Geodinamica Acta
to develop a theoretical approach to its study of (2007), Geomorphology (2008), and Catena
the intersection of earth processes and human (2011), all devoted to geoarchaeology, empha-
agency. It is no great stretch for size its global reach (Tourloukis and Karkanas
geoarchaeologists to pose questions posed as 2012) as well as the field’s multidisciplinarity as
hypotheses nor to use induction as a tool to a major feature and methodological strengths.
generate ideas of causation or affiliation among Hassan in his 1979 American Antiquity article
an observed set of variables (Buchdahl 1951). “Geology and the Archaeologist” may have
Wilson (2011) characterizes geoarchaeology as suggested a more narrow view of the discipline
evolving into a coherent discipline no longer than he intended. Geology as an earth science is
limited to geologists helping out archaeologists mostly synonymous with lithology of the Earth,
nor archaeologists simply with an interest in while geomorphology is associated with the sur-
geology. Geoarchaeology is seen as discipline face of the Earth. Geology concerns itself with
unto itself bridging its two parent disciplines of the Earth surface, but geomorphology is more
geology and archaeology (supra). Human life has likely found within Department of Geography.
never been distant from natural influences. Both geography and geology are branches of
Archeologists have always been skeptical of earth science.
environmental determinism in explanation and Over the latter part of the twentieth century,
would have no counsel for a geological determin- the incorporation of the methods of other areas of
ism. Modern geoarchaeology tends to heed earth science—geochemistry, geochronology,
Butzer’s (1982) injunction to treat neither culture and geophysics—has expanded the reach and
nor environment as static nor separate from one utility of archaeological geology. As Kraft notes
another. Charting intersections of culture and in a review of archaeological geology, the
nature tend to provide insight into the role of workers in this area are becoming more aware
each as they influence each other in the complex of the nuances of the historical and classical
coevolution of natural and anthropogenic record (Kraft 1994). Legend, oral tradition,
systems in prehistory (Haynes 1995; Redman mythology, and the body of cultural knowledge
1999; Levine 2007; Benedetti et al. 2011). for a prehistoric culture are being used in the
As Gladfelter, in his acceptance of the Geo- search for information about ancient events—
logical Society of America’s “Ripp Rapp” paleoenvironmental and paleogeographical
Archaeological Geology Award in 2001, implies, (supra). This is the essence of a “multidiscipline”
it (geoarchaeology) is an honorable enterprise archaeological geology—the use of geological
(GSA Today 2002). In the past years, this and analytical techniques in concert with archae-
observation has been made of archaeology itself ological expertise to gain knowledge of the past.
(Daniel 1967). The application of perspectives This book echoes many of these tenets set forth
and methods born in disciplines like geomor- by these other workers. In so doing, this work
phology now are recognized as crucial to the will serve to amplify that call for technical and
understanding of site location, function, duration, methodological rigor in the use of earth science
and like questions (Jones 2007). Quoting Beach in the practice of archaeology as a whole.
et al. (2008), “Geomorphology deals with Earth One last note as to what archaeological geol-
surface processes and landforms; archaeology ogy is about, simply put geoarchaeology. In a
deals with what artifacts in the landscape can sense, Hassan, Gladfelter, Rapp, and Gifford
tell us about the human past. Because geomor- may have inadvertently proscribed geography
phology as a discipline was born through a mix- and the important subfield of geomorphology
ture of physical geography and surficial geology, by using the terminology “archaeological geol-
geomorphology can inform archaeological ogy.” Geoarchaeology is the more inclusive
6 1 Introduction

terminology and that probably explains its more (the general case) is important (Evans 1972).
frequent usage today. This text will simply use From the geological standpoint, the identification
“geoarchaeology” throughout with the under- of structural framework and the characterization
standing that the author recognizes its more com- of the area’s lithology from outcrops and
mon usage as well as that the discipline’s exposures must be done as well. To do this, the
signature journal is entitled Geoarchaeology. archaeological geologist relies on traditional as
well as contemporary methods.
Studies show that animals throughout the
1.1 Organization of This Book Holocene are indicators of both changing
climates and cultural evolution and that coastal
This work is comprised of 11 chapters. The sites are strongly influenced by changing sea
contents of these chapters cover major elements level; the history of these changes can be
within archaeological geology. One significant recreated by sediment coring studies. An under-
aspect of earth science that is omitted is that of standing of site selection criteria is important for
geochronology. This was consciously done. archaeological studies in of fluvial, estuarine, and
Given the present-day literature on the use of coastal systems throughout the United States and
dating techniques in archaeology, another survey the world. Earth science techniques can be
of this area hardly seemed warranted. The reader deployed to quickly identify areas with a high
is directed to the premier volume of this series potential as ancient habitats. Geoarchaeological
by Springer-Verlag, Age Determination of Young evidence can be used to document neotectonism,
Rocks and Artifacts (Wagner 1998). The focus sea-level changes, and site location.
of the book will be on a systematic approach In the area of traditional geological mapping,
to the archaeological site, its geomorphic/ the researcher relies on their training in landform
geological features, and anthropogenic materials and soil recognition, basin hydrology, and struc-
from an earth science perspective. While geo- tural geology. The basic tools are maps—geo-
chronological techniques are central of the under- graphical, geological, and hydrological—
standing of stratigraphic contexts of sites, their together with the compass, the inclinometer, the
in-depth discussion will remain secondary if shovel, the rock hammer, and the hand lens.
discussed at all. Individual chapters are discussions Today, there are newer tools such as satellite
of those areas of earth science that bear directly on images, satellite location or global positioning
the nature of archaeological deposits and materials system (GPS), laser technology as seen in the
as interpreted from the geological perspective. increased use of LiDAR, laser transits, and levels
Chapter 2 will address the geomorphological coupled with laptop computers and geographic
and geological context of the archaeological site. information systems (GIS). The low-cost auton-
Immediately obvious is the question of scale to omous vehicle or “drone” has revolutionized the
be considered by the researcher. This devolves, mapping of landscapes and archaeological
in turn, on the archaeological objectives of the features. Whatever the technique or tool, the
specific research. Regional scale mapping would aim is the same: characterize the geomorphic
be called for by a research design that seeks to and geological context of the archaeological site
determine the spatial context (location) of a or sites. Once this is done, the archaeologist has a
cohort of sites, their cultural and temporal place- better understanding of the “why” of the site’s
ment important but secondary at this stage of the location in a particular landscape. The past land-
inquiry. Below regional scale would be the local scape or landscapes generally are masked by
environment of the individual site. This is, per- modern earth system processes such that the
haps, the most familiar type of geological determination of the prehistoric landscape
mapping used in archaeology. From a geomor- becomes an archaeological inquiry in and of
phological standpoint, the identification of itself.
landforms (the specific case) and landscapes
1.1 Organization of This Book 7

Chapter 3 principal focus is the description of materials achieving reliable results. Studies of
sediments and soils and their use in archaeolog- sites should include micromorphology, hearths,
ical geology. The chapter, in some detail, floors, phytoliths, and chemical analysis, espe-
examines ways to carry out more detailed inves- cially of Al, organic C, Fe, available phosphate,
tigation into the sedimentological context of the and pH. Sedimentation plays a key role as a
archaeological site and its environment beyond counter to erosion in site preservation and also
that of the present-day or near-surface allows dating, e.g., volcanic ash layers in
components. The prehistoric landscape can be, Santorini and, of course, Pompeii, where recent
and ofttimes is, more deeply found than what is studies have provided new insights into the
capable of recovery with the shovel or even timing and kind of eruptions that destroyed this
mechanical excavation. The recovery of sedi- city and nearby Herculaneum in AD 79. Micro-
ment columns from hand-operated and mechani- morphology is a mainstream methodology for the
cal coring devices is a favored means of study of deposits created or altered by human
geoarchaeological study and has a successful occupation; thereby, allowing for distinctions
history in application. Cores produce not only between cultural and natural transforms must
sediments but macrobotanical, macrofaunal, and be made.
artifactual remains as well. Cores taken in a In Chap. 5, shallow geophysics is discussed as
systematic campaign across a site, and its a modern outgrowth of the need for techniques
surrounding area, can yield significant amounts that can rapidly and effectively characterize near-
of environment, depositional history, and spatial surface geology. Overviews are cited and used to
extent. Mechanical excavation, in exposing soil update the earlier edition (Clark 2003; Linford
and sediment profiles, are superior to coring for 2006; Gaffney 2008). This is a geophysics that is
obtaining the “big picture” as regards the geo- basically non-seismic in nature or one that does
morphic mapping of the lateral continuity of not use sound as the principal means to measure
sedimentary deposits. Both hand and mechanical subsurface rock and sediment thickness. When
methods have their place in archaeological geol- applied to archaeological problems, this form of
ogy. It is difficult to visualize the deployment of geophysics has been called “archaeological geo-
a large mechanical excavator or even a drill rig physics” or more simply “archaeogeophysics.”
on a prehistoric burial mound given the rarity and Weymouth (1996) has called this use of geophys-
uniqueness of these features. Gone are the days ics “digs without digging.” Beginning in the
of total excavation of these types of structures. 1960s with the application of electrical methods
Careful and judicious sampling of them requires such as resistivity profiling, the field has expanded
methods that give the researcher greater control with the addition of electromagnetic techniques
(and analytical returns) over the technique to be such as conductivity meters and metal detectors,
used. Noninvasive, shallow geophysical the use of magnetics in the form of magnetometers
methods, discussed in Chap. 5, are proving to and gradiometers, and, since the late twentieth
be essential in the sensitive sites. century, the increased use of ground penetrating
Chapter 4 deals with analytical procedures for radar (GPR). Coupled with computer software
examining sediment and soil samples acquired that is basically derivative from seismic-based
by excavation or coring methods. The five geophysics, the expansion of geophysical
“p’s”— particle size, point counting, phosphorus prospection into archaeology has produced an
content, pollen, and phytoliths—are discussed. extensive body of research in recent decades that
Mineralogical identification of sediments will is recognized within both archaeology and
be discussed in the context of petrographic geology.
methods in Chap. 6. Many of the methods Chapter 6 deals with the use of petrographic
presented are hardly “cutting edge,” but they methods in the analysis of rocks, minerals,
have had significant development and applica- sediments, and archaeological artifacts or
tion across a broad range of archaeological features made up of all or some of the foregoing.
8 1 Introduction

The principal discussion centers on the utility of clays and their mineralogical transformations
of the geological thin section and the petro- at high temperatures provides insight into manu-
graphic microscope in analyses of rock and min- facturing processes. Geoarchaeological and geo-
eral items of archaeological interest. Beyond the chemical protocols for ceramic analyses will be
hand lens this protocol is the geologist’s most examined.
reliable and time-honored methodology. As Herz Chapter 8 deals with some of the more preva-
and Garrison (1998) has pointed out, it is also one lent instrumental techniques that have come to be
of the most economical and reliable ways of used in the study of archaeological problems.
determining the geological nature of artifacts. It Notable among these techniques are neutron
is certainly to be considered before moving to the activation analysis (NAA) and mass spectros-
next level of analysis—geochemical analytical copy of stable isotopes. Added to these are the
techniques. What is disquieting is how often the X-ray methods of fluorescence and diffraction
archaeologist does not use petrography before (XRF and XRD). The latter method has been
moving to more costly methods. In the case of styled as “the archaeologist’s first line of defense
ceramic studies, the use of petrography is a major in mineralogical identification.” Together with
and precise method of studying ceramics and the electron microprobe (EMPA), its high-energy
their clay sources (Vaughn 1991; Velde and cousins—various ion microprobe methods, induc-
Druc 1998; Courty and Roux 1995; Quinn 2013). tively coupled plasma spectroscopy (ICP), and
Determining sources of raw material for lithic atomic absorption spectroscopy (AAS)—the
artifact fabrication was discussed in the context archaeological chemistry of artifacts, and,
of manufacturing techniques, reasons for use of increasingly, their residual contents, food,
the particular raw material, source depletion, liquids, etc., are the subject of major contempo-
usefulness, and change of social patterns. Analy- rary studies across all aspects of archaeology:
sis of data obtained on physical properties such prehistoric and classical.
as mineralogy, petrology, and chemistry must be Because of the usefulness of stable isotopes in
evaluated in the light of method of production, geoarchaeology, Chap. 7 will examine their
use of the artifact, and effects of long-term worth for provenance in paleoenvironmental
burial. Finally, interpretation of the relationship and paleodiet studies. Sulfur isotopes are partic-
between the artifact and the ancient cultural con- ularly useful in the studies of handmade artifacts.
text must also take into account the expertise of The strontium isotope ratio, which depends upon
the ancient artisan. sea or land source, is useful in studies of marbles,
Chapter 7 reintroduces the importance of clay as are oxygen and carbon isotopes. Carbon, along
minerals and cultural materials, such as with nitrogen, sulfur, and strontium, is also use-
ceramics, made from these unique minerals. ful in determining diet and diet changes through
Our discussion will center on recognition and the study of the isotope content of bone. The
characterization of clay minerals both as primary nitrogen content of bone can distinguish herbi-
and secondary minerals and in particular their vore from carnivore and, along with carbon and
use by prehistoric and historic cultures. From sulfur content, is useful for determining the
both geological and mineralogical perspectives, nature of isotope cycling on a global scale. To
clays represent a robust contemporary area the question of whether trace element signatures
of study. Much like stone, “Clay is where you could be helpful in determining sources of raw
find it.” We’ll examine this proposition further material used in the manufacture of artifacts
as well as the basic geochemistry of how clay worldwide, the answer was that an extensive
forms as a weathering deposit (primary) of library of samples and a wide range of
silicate minerals and secondarily as deposits archaeometric techniques would be required.
of transported sediments. Consideration will Chapter 9 recognizes the importance of the
be given to the wide range of ceramics study of ores, manufacturing debris, and slags,
from earthenwares to porcelains. A discussion together with research on metallic artifacts as an
1.1 Organization of This Book 9

increasingly important aspect of archaeological ogy as it exists today: radiocarbon, fission track,
geology. The metal ages—bronze to iron— luminescence, archaeo- and paleomagnetism,
comprised two-thirds of Thomsen’s nineteenth- U-series, argon-argon, and aministratigraphy
century classificatory system for European pre- methods are only discussed in relation to exam-
historic cultures. The late twentieth and early ple applications and do not occupy a chapter to
twenty-first centuries have seen advances in themselves. Likewise, fission-track methods are
archaeometallurgy (Roberts and Thornton, eds. especially useful in dating volcanic rocks and,
2014). An example of the intersection metallur- potentially, pyroclastic deposits. For ages up to
gical techniques, used by Bantu tribesmen of 350,000 years, dating by U-series isotopes has
Malawi, illustrates how primitive iron-making been shown to be particularly useful. Remanent
methods necessitated the extraction of iron from magnetization and optical/thermal luminescence
low-grade ore when higher-grade ore was nearby (OSL/TL) can both be used to date sedimentary
and known. Trace element variety, content, and rocks, on the bases of time of deposition for the
ratios may also provide important clues to the former and time of burial and exposure to heat or
source of raw material, but application of the data sunlight. In the case of OSL, the growth of its use
must take into account, in the case of metallic in geoarchaeology has been exponential.
artifacts at least, the possible effects on the fin- It is a tribute to the increasing maturity of the
gerprint patterns of any flux used in the smelting field, as a discipline, that there is such breadth
process. and diversity in its subject matter. What I shall
Chapter 10 reprises the first edition’s review attempt to in the following chapters is expose the
of some basic statistical methods used in archae- student, and any interested colleague, to practical
ological geology. Updates relative to the concerns and earth science methods that are use-
advances in the application of multivariate ful in the consideration of archaeological
methods to large data sets (Speakman et al. problems. Archaeology identifies the results of
2011) and the increased use of Bayesian methods this interaction between humans and nature in the
will be presented. material remains found in the Earth. Linking the
Chapter 11 is a concluding chapter as well as a trajectory of past human cultures to that of the
brief discussion of theory and practice in earth system requires geological science. The
geoarchaeology. In some ways, it simply synthesis of this interaction provides a diversity
redresses the lack of such a discussion in the of roles for the archaeological geologist/
first edition. This textbook is not a comprehen- geoarchaeologist using the techniques outlined
sive treatment of the field of archaeological geol- in this textbook.
The Geomorphological and Geological
Context 2

2.1 Introduction of an archaeological locale provides crucial infor-

mation as to the why of a site’s specific location
As noted in the preceding chapter, geomorphol- and possible function.
ogy is a discipline more often found in the larger
field of physical geography, at least in the United
States and Great Britain, while sedimentology is a 2.2 Landscapes and People
speciality area of geology. While some in archae-
ology may wonder at a partition in what is basi- In Native American mythology, the landscape is
cally earth science, the location of geomorphology often characterized as either indifferent or even
within geography has had some unexpected antagonistic to humanity. Descriptions of the
benefits, most notably the recognition and use of ancient landscape by these and other cultures,
the burgeoning techniques of geographic informa- Australian, for instance, ascribe it as resulting
tion systems better known as GIS. Likewise from the action of powerful forces often in zoo-
geographers were quick to incorporate their tradi- morphic form. Humanity inherited this landscape
tional interest in remote sensing into GIS studies from more primal times (and species,
as well. Digitally based, these large-scale spatial “monsters,” etc.), and it was only made more
analysis systems have proven to be a boon to habitable by time and heroic figures such as the
geoarchaeological researchers. GIS frameworks, Hero Twins of Diné cosmology of the American
devised principally within geography, have made Southwest. Today, most geologists view the
it possible to converge many different lines of data Earth’s surface is a transient feature changing in
into synthetic map projections that can be a kaleidoscopic fashion through eons. This was
manipulated and compared so as to yield new not the case for many earth scientists into the
insights into the nature of archaeological sites early-to-mid-twentieth century (Frankel 2012).
and their locations. Had geomorphology been For much of the geology’s history, the Earth
housed exclusively within geology departments, was a “fixed” planetary sphere whose exterior’s
it is unlikely that GIS methodology would have surface was the only changing feature. This
entered the arsenal of archaeological geology as “fixist” view was eclipsed by the theory of plate
quickly as it has. Sedimentology and its close tectonics by the mid-1970s (Frankel, supra). The
relationship to paleoenvironmental study, particu- interior as well as the exterior of our planet is
larly in quaternary research, have continued to be subject to constant change.
a major component of archaeological geology. Non-western, nativist views of the Earth often
Likewise, the mapping the surface and subsurface reflect this dynamic view of the Earth. Certainly

# Springer-Verlag Berlin Heidelberg 2016 11

E. Garrison, Techniques in Archaeological Geology, Natural Science in Archaeology,
DOI 10.1007/978-3-319-30232-4_2
12 2 The Geomorphological and Geological Context

aboriginal descriptions of the early Earth incor- The Cherokee Indians name the double peak,
porate change as a key process. The forces for in the Great Smokey Mountains of North
much of this change were seen as corporeal Carolina, known to others as Chimney Tops,
supernatural beings or “monsters.” How else Duni’skwalgun’i. The name means “forked antler”
could the Earth being changed without recourse for the great stone deer concealed there (Camuto
to power beyond the kin of humans? Knowledge 2004a). It is either a great stone animal or a moun-
of the land surface was critical to human sur- tain of flesh and bone. Perhaps both (supra). This
vival. Prehistoric and some modern tribal and other ridges of the mountains are ancient
peoples set great store their knowledge of their storehouses and council places for animals
geographies. Modern cultural geographers and (as well as humans) (Camuto, supra). The Chero-
ethnographers have recorded this aspect of kee nouns that describe these mountains are
non-Western knowledge time and again. archaic and stony things. Nugatsa’ni, high ridge
Basso (1996) in his award-winning treatment with a gradual slope, is similar in its specificity as
of the “place making” of the western Apache that of the Western Apache, Tséé Chiizh Sidilé,
noted that “entire regions and local landscapes “Coarse-Textured Rocks Lie Above in A Compact
are where groups of men and women have Cluster” (Basso 1996).
invested themselves . . . and to which they Exacting landscape descriptiveness and its
belong.” The late Lakota scholar, Vine Deloria, relevance to non-Western cultures is further
Jr., observes that “most American Indian names illustrated by the hand-drawn map below. Cre-
embrace spatial constructions of history” ated, in less than 2 h, from memory, by a
(supra). Pulitzer prize-winning author, N. Scott Wardaman Aboriginal Elder, it shows the
Momaday writes: “From the time the Indian first locations of nearly 200 springs near his house
set foot on this continent, he centered his life in in the Northern Territory of Australia (Fig. 2.1).
the natural world. . . ..The sense of place was The twenty-first-century view of a malleable
paramount. Only in reference to the earth can and changeable Earth surface echoes that of the
he persist in this identity (Momaday 1994:1).” ancient non-Western models. At geological time

Fig. 2.1 Map of nearly

200 springs around
Menngen Station (in red),
Northern Territory,
Australia (Map drawn by
Senior Elder Yidumduma
Bill Harney in 2010)
2.3 Geomorphic Concepts 13

scales, the surface-forming processes of erosion that any change in a landscape implies a change
and deposition are secondary to the larger, dra- in process. Other concepts include:
matic forces of plate motion, tectonism, and orog-
eny. Such processes operate within time intervals, 1. Landscape evolution can be gradual or abrupt
outside human experience, approaching that of as thresholds are exceeded.
Wilson Cycles wherein oceanic basins rift open 2. The interaction of landscape changes with
and subsequently close after millions of years thresholds results from complex processes
(Levin 1988). and is termed complex response.
The geomorphic surface is a mixture of 3. Various tectonic fault varieties—normal,
landforms that are an expression of (1) modern reverse, and slip-strike—are associated with
geological processes and (2) geological pro- characteristic landforms.
cesses that are no longer active. Landscapes
reflect present and past processes of erosion and In general these geomorphic concepts apply to
deposition. The topography of a given landscape that part of the Earth’s history called the Quater-
is the result of climate-driven processes ofttimes nary. The Quaternary spans, exactly, the last
directly coupled to tectonic activity. Rinaldo 2.588 million years and encompasses the largest
et al. (1995) presented these relationships in a part of hominid evolution in the later Pliocene
mathematical model of landscape evolution and Pleistocene (Riccardi 2009; Gibbard and
which underscored the importance of tectonic Head 2009).
uplift to the formation and persistence of geo- One of the oldest models for landscape evolu-
morphic signatures. Seismicity is not a new con- tion is that of William Morris Davis who articu-
cern in geomorphology, but its character and lated the “Cycle of Erosion” (1899). In this
expression—earthquakes, faulting, and volca- model, brief periods of uplift are followed by
nism—are important to unraveling the nature much longer periods of inactivity and erosion.
and origin of various relicts and modern In Davis’ view, the landscape had a “life cycle”
landforms. from (1) youth through (2) maturity to (3) old
In the northern hemisphere, climate-driven age (supra). Young landscapes are characterized
processes have given rise to glaciations which by stream incision, and deep valleys are
the geomorphologist must always be cognizant followed, in mature landscapes, by less dramatic
of the special character of glacial landform. The stream gradients and valley down-cutting.
interactions of landform and environment are Mature landscapes have broad floodplains and
dramatically demonstrated in North America meandering streams characterized by high dis-
and Europe due in large part to their glacial charge and sediment loads. In old age, the land-
past. Concomitant with Ice Ages were marine scape has eroded to a planar or peneplain surface
regressions, and consequent transgressions, with low relief awaiting rejuvenation or the start
which alter coastal and fluvial systems and play of a new cycle. Davis put a time scale of a million
much the same role as uplift in non-glaciated years on his cycle—long but archaeologically
regions. Coastal landforms—past and present— more relevant than the longer Wilson Cycles.
are the expression of glacial processes acting In this, and, other process models, the land-
globally through sea-level change. scape can, and will, equilibrate for thousands of
years. Equilibrium is clearly a relative concept
with regard to landscapes. As we will discuss in
2.3 Geomorphic Concepts Chap. 3, paleosols typically mark these equilib-
rium periods of climate. Uplift can go on for
Keller and Pinter (2002) review the basic millions of years however imperceptively. Like-
principles used in the study of geomorphic wise, subsidence and/or erosion will occur over
landscapes. The most basic principle, in terms the same time scales. In many instances, the
of modern, process-oriented geomorphology, is interplay of these forces is a “zero-sum”
14 2 The Geomorphological and Geological Context

situation. In the contest between uplift and ero- surfaces are modified, daily, by gravity-driven
sion, the latter always wins in the cycle of ero- processes—creep, erosion, etc. Because of this
sion. Small wonder that archaeology is an dynamic and changing surface, the archaeologist
endeavor that deals primarily with that which is should never be lulled into thinking that surface
buried. Yet without uplift, the geomorphic sur- is immemorial. It is not.
face of the Earth would be that of the relatively The landforms that we propose to study and
unchanging surface of Mars. In the case of Mars, map are the result of three factors: (1) geomor-
landers and mapping satellites have identified phic processes, (2) stage of evolution of land-
what Morris would have termed “old age” with form, and (3) geological structure. To map
only erosion—primarily by wind—modifying a archaeologically interesting topography, one
desert surface. Still, what we term “deserts” on must (a) identify the geological processes by
Earth, such as Antarctica, the Atacama, or the which the landform were shaped, (b) recognize
Sahara, cannot begin to match those of Mars for the stages of development of landform and their
stark barrenness and lack of overall change. evolution through time, and (c) recognize the
Some geomorphologists define a type of equi- topographic expression of geological
librium termed “steady state” where the land- structures—dip, strike, clinal variation, etc. In
scape exhibits a relative constancy over a geoarchaeology, the consideration of topography
period of 100–1,000 years. This conceptual must include another factor—human-induced
view of the landscape can be called the “average” geomorphic change.
of a land surface at this particular time scale.
From a process-oriented perspective, one could
say the limits of equilibrium or thresholds are
relatively “high” such that abrupt or dramatic 2.4 Geomorphic Setting
landscape change is precluded at this scale.
In a tectonic sense, landscapes can be “dor- A brief review of geomorphic settings reminds us
mant” with little actual or observable seismic that there are basically only fluvial, desert/arid,
activity. Generally the most active landscapes coastal, glacial, volcanic, and karst/caves. Most
are those associated with seismically dynamic archaeology is done in relation to fluvial
zones such as found at crustal plate boundaries. landforms. Human groups have always
Conversely the most inactive areas, again, in the gravitated to available water, and this is readily
tectonic sense, are furthest from plate margins apparent in any study of agrarian cultures. Crops
such as continental interiors. It is easy to locate need reliable water supplies and these are most
tectonically active areas. They either exhibit often found in drainage systems. Besides avail-
volcanoes or earthquakes or both. Archaeological able water, there are rich, deep alluvial soils
examples come easily to mind—the Apennines, which ancient farmers were quick to identify
Naples/Pompeii, Thera/Akrotiri, Anatolia, the and exploit. Geomorphic settings are the result
Dead Sea region, the East African Rift Valley, of geomorphic processes—fluvial, aeolian, vol-
and Pacific Rim to name the most obvious. canic, and glacial processes—which in turn are
The “take-home message” for the archaeolo- special cases of the two primary processes: ero-
gist using modern process geomorphology is that sion and deposition. In this book, we are
the geomorphic surface on (and under) which concerned primarily with the identification of
archaeology is found is not a static surface. the results of geomorphic processes than the pro-
Even in a geological sense, the geomorphic sur- cesses themselves. The study of the latter is that
face changes at rates of decadal to annual scale. of surficial process geology. In this chapter, we
True, the mountain belts form at Wilsonian rates recapitulate the importance of these landforms to
as do their valleys and adjacent basins, but their archaeology and their attributes.
2.4 Geomorphic Setting 15

2.4.1 Fluvial Landforms Streams incise, creating terraces and valleys,

and aggrade, creating floodplains and new chan-
Mapping fluvial systems involves both the hori- nel courses (Fig. 2.2). When stream base levels
zontal and vertical dimensions of the riverine are lower, the stream incises even in its lower
topography. In creating floodplains and valleys, reaches. When the base level rises, due to chang-
streams alternately downcut and deposit ing climate and elevated lake/sea levels, a stream
sediments. In montane areas, alluvial fans are aggrades. Base level was first defined by John
key features that contribute to the available sedi- Wesley Powell (Leopold and Bull 1979). Powell
ment balance as well as the distribution of that correctly identified sea level as the ultimate arbi-
sediment. Another key landform feature in ter—exclusive of mediating factors (geological
stream valleys is terraces. It is on these topo- framework, etc.)—of upstream incision. An
graphic features that most archaeological sites incising stream is more of a conduit for sediment.
are located. Determining their number and Only when the discharge increases to flood levels
order is a primary objective in any mapping of does the sediment escape the channel to deposit
a fluvial system. The youngest terrace (T0) is new sediments in the floodplain. A stream carry-
adjacent to the floodplain, and the oldest is the ing large volumes of sediments, such as an out-
highest terrace (Tn) (Fig. 2.2). wash stream, can fill its channel to the extent that
Between these terraces, the channel moves an arbitrarily higher base level is established.
alternately eroding (cutbanks) and depositing Streams can then regularly overflow and signifi-
(point bars) alluvial sediments, generally oppo- cantly increase the volume of sediment in its
site to one another. Depending upon elevation valley.
and gradient (slope), which are direct In any discussion of fluvial systems, “load”
expressions of uplift, stream and valley morphol- and “form” are often paramount. Bed, wash, and
ogy vary from region to region. The hydrologic suspended “load” characterize the finer-grained
cycle (precipitation-evaporation) determines the sediment entrained in a river and deposited in its
amount of available water to the drainage sys- floodplain. Included with discussions of load is
tem, in turn influencing the incision and growth “flow.” Flow can be described as laminar, turbu-
of smaller-order streams within the drainage lent, chaotic, etc. Flow is quantified using
basin. The delta represents a major fluvial land- parameters such as Reynolds number, Froude’s
form historically significant to human cultures number, etc. Flow and sediment load are com-
throughout antiquity. bined in the Hjulstrom diagram (Fig. 2.3).
Fluvial systems, along with volcanoes and Recent studies (Pizzuto 2012) have shown the
glaciers, have the role of renewing “old” movement of fine sediment is less “suspended”
landscapes and building new ones. The move- as load than it is moved by saltation analogous to
ment of eroded sediments is principally the role aeolian processes. Where bed and suspended
of the stream. In doing so, it transports the sedi- load have a genetic relationship, “wash” load
ment through its valley creating terraces, levees, has none, being, rather, that portion of sediment
and floodplains and, ultimately, modeling coastal carried entirely by stream flow with no deposi-
landforms—estuaries and beaches. Depending tion. The architecture of sediment facies within
upon the factors of precipitation, vegetation or adjacent to streams can be characterized as
cover, lithology, and gradient, the sediment (1) massive, (2) cross-bedded, (3) laminar, and
load a stream carries varying with its discharge. (4) interbedded, to name the more commonly
That latter factor is dependent on climate. For used terms. Likewise, cross-bedded forms can
geoarchaeology, the correct reading of type, tex- be “trough-like,” “planar” within channel forms
ture, and distribution of fluvial sediments allows described as “m-scale,” “epsilon,” and “low-to-
the retrodiction of past climate—the controlling high angle” (Larsen and Brock 2014). This archi-
factor in paleoecologies and a primary mediating tecture is a direct reflection of stream flow
element for past human cultures. and sediment depositional regimes (Tilston
16 2 The Geomorphological and Geological Context

Fig. 2.2 Variety of fluvial

terraces, paired and
unpaired, created by
erosional and depositional

et al. 2015). When coupled with sediment size thus a direct record of their parent lithologies
and velocity estimates, retrodiction and and, in some part, the processes active on these
modeling of systemic dynamics are possible. rocks and minerals. For instance, if we observe
Soils formed or forming on the valley floors resistant materials, such as quartz and rutile, in
are buried, becoming paleosols, often together large amounts in the sediments we can speculate
with archaeological facies associated with these on the nature and intensity of the weathering
land surfaces (West and Dubos 2013). Streams, climate. By contrast, the finer sediments of clay
in incision, and meander, likewise, can exhume and silts suggest a much different climate sce-
past land surfaces. nario as well as that of deposition and stream
Weathering of the uplands, coupled with and dynamics.
caused in part by precipitation, is the ultimate Human changes of landscapes can alter runoff
source of the basin sediments. The sediments are and the sedimentological character of fluvial
2.4 Geomorphic Setting 17

climate became much wetter. Elevated levels, in

the Alpine lakes, which submerged the lake
dwellings of earlier cultures, attest to this.
At Petit-Chasseur, the increased precipitation,
together with denudation of the surrounding
slopes, led to greater erosion and runoff which,
in turn, led to increased flooding by the small
river. Coupled with the aggradation of the river,
colluvial debris flows inundated the site as well.
The archaeological site disappeared under a cone
of alluvium and colluvium.

2.4.2 Coastal Landforms

The fluvial delta marks the entry of the river to

Fig. 2.3 The Hjulstrom diagram. Flow velocity is
juxtaposed to sediment size to indicate “transport” (sus-
the sea. The delta is the alluvial fan deposited in
pension) versus “sedimentation” (deposition) for various the sea rather than in stream valley. Coastal
sediment grain sizes landforms are among the most changeable of all
geomorphic settings. This is due to sea-level
systems. Dramatic changes in erosion, due to the change wherein a rising sea level creates a
clearance of forests, increase the sediment loads submergent coast and conversely a falling sea
of streams, causing aggradation and increased level produces an emergent coast. In the United
recurrence of floods. Since the Neolithic, these States, the bulk of the eastern shoreline is
changes have increased principally because of submergent, while that of the west is largely
the introduction of plant and animal domestica- emergent. In the Gulf of Mexico, the shoreline
tion. In extreme cases, the soils of the watershed of eastern Texas and that of Louisiana together
are eroded to bedrock, but more often the upper with Alabama and Mississippi are submergent
soil horizons (cf. Chap. 3) are deflated, leaving due the great sediment loading of the continental
the resistant soil layers such as clay-rich B shelf by the Mississippi River. Coastal down
horizons on which new soils and sequences warping due to sediment loads mimics that due
can form. to little or no tectonic activity producing conti-
Alluvial and colluvial fans—Petit-Chasseur, nental margins like the Eastern United States.
Sion, Switzerland—50 years ago, during the con- Coastal landforms can challenge fluvial for the
struction of a new road between Sion and Petit- density of archaeological sites. Wherein the flu-
Chasseur, workers discovered the remains of vial system provided humans with rich riparian
over a dozen Late Neolithic (ca. 2500 BC) dol- zones, many coasts have comparably biologi-
men tombs (Osterwalder and Andre 1980). cally active estuaries. Human cultures have
Buried under deep (6 m+) alluvium, the elaborate taken advantage of these estuarine zones, and
tombs and 30 menhirs (standing stones, cf. their settlement remains dot the shorelines of
Chap. 5, Fig. 5.7) were well-preserved along a the Earth—past and present.
small tributary of the Rhône (Sauter 1976; Besse The most prevalent coastal landform is the
2014). During the construction, and reuse, by beach, whether on a barrier island or simply the
later peoples, of the tombs, the Late Neolithic margin, above sea level, of a continental land-
and Early Bronze Age climate only produced mass. The beach is a transient topography chang-
nominal runoff and erosion. Sometime after ing over time with eustatic sea level and over the
2200 BC and into the Middle Bronze Age, the year according to storms and season. Ancient
18 2 The Geomorphological and Geological Context

beaches are of archaeological interest. Henry de

Lumley, in his landmark excavation of Terra
Amata, demonstrated the presence of Pleistocene
hominids—Homo erectus—on now uplifted,
exposed beaches at Nice, France (Lumley
1969). Beaches and dune systems, often found
back of them, provide direct clues to sea level
and climate, fauna and flora found within their
Estuaries form back of the beach and repre-
sent the juncture of the river and the ocean. They
are unique landforms that are the richest, in terms
of biomass, along shorelines (Odum 1995).
Estuaries are transient landforms like beaches.
Unlike the beach, the estuaries can penetrate
inland and cover large areas of the coast. These
landforms are rich in archaeological sites from
the Mesolithic to the historic era. The deltas of
large fluvial systems attracted early human
groups to their rich alluvial deposits as well as
abundant marine resources.
Geoarchaeological sediment coring of the Fig. 2.4 Jekyll Island, Georgia, with overlay of a
coastal plain and estuarine deposits along the paleochannel or “Old Channel” of the Brunswick River,
Eastern North American margin has gone occupied, in part, by Jekyll Creek, today (Chowns
on apace, in the late twentieth and, continued, et al. 2008, Fig. 9)
into the early twenty-first century (Pavich
et al. 2010). Coupled with new data from the followed the thalweg of the modern Jekyll
offshore, the southeastern United States has Creek, thereby emptying into the Atlantic south
provided a higher-resolution view of the Pleisto- of a paleo-Jekyll Island (“Pleistocene core”)
cene and early Holocene. The penultimate glaci- rather at its modern tidal delta at the north end
ation (MIS 4) and interglacial (MIS 5) of the modern island. Figure 2.5 shows this pro-
paleoenvironments have been characterized posed paleochannel in its modern setting.
using pollen and macrobotanical remains from The Jekyll Creek area was vibracored by
sediment cores into estuarine sediments of the Chowns (supra), and bedforms, so described,
Chesapeake Bay (supra). suggested the remnant of a buried paleochannel
As shown in the following, Chowns of the Brunswick River. A later 2012 survey
et al. (2008) have presented reimagined Pleisto- (Panamerican Consultants, Inc.) mapped
cene age paleochannels now infilled behind mod- subbottom features as stratigraphic sections of
ern barrier islands (Fig. 2.4). Chowns and his alternating beds, channel parabolic “sag”
students have posited the presence of an “inlet- features indicating channel facies. A high back-
dominated” coast for the Georgia Bight rather scatter sonar contact, along the channel margin,
than a “tidal-dominated” coast of today. In this upon diver investigation, was confirmed to be a
inlet-dominated model, coastal streams, such as deposit made up of whole and fragments of
the Brunswick, Ogeechee, and other coastal plain oysters. All specimens were disarticulated, with
rivers, did not debouche at their modern evidence of oxidation (rust color) on many
locations but ran parallel to the coast behind valves. The oyster shells were large, and there
barrier island systems. Figure 2.4 illustrates this were equal numbers of left and right valves
scenario where the paleo-Brunswick River indicating a natural deposit of undetermined
2.4 Geomorphic Setting 19

Fig. 2.5 Present-day view

of Jekyll Creek looking
eastward from the “Old
Channel” or
“paleochannel” of the
Brunswick River toward
the “Pleistocene core” per
Chowns et al. (2008)
(cf. Fig. 2.4). Modern
oyster reefs line the
shoreline (Photograph by
the author)

age. Based on size, alone, the oyster specimens River” and its surrounding terrain (Fig. 2.6).
were larger than modern populations encoun- This Quaternary Channel River had its formation
tered during subsequent studies (Garrison on Oligocene times but occupied a repeatedly
et al. 2012a). incised Pleistocene era paleo-valley (Lencolais
et al. 2003) (Fig. 2.6).
Submerged landforms and ofttimes buried,
2.4.3 Submerged Landforms prehistoric archaeological sites thereupon are
difficult to discover and study. These submerged
Jonathan Benjamin (2010) summarizes much of sites are extremely relevant to a fuller under-
the modern archaeological emphasis on standing of human colonization and expansion
submerged coastal landforms of the continental in Asia, Europe, Africa, and the Americas.
shelf and estuaries in Europe. Specifically, focus These now drowned landforms were generally
on the inundated prehistoric terrain of the North subaerial coastal plains for several millennia
Sea Basin which remains one of the most enig- prior to and after the Last Glacial Maximum
matic archaeological landscapes in northwestern (LGM). Recent studies have amply demonstrated
Europe. This region was submerged by the sea that there are recoverable cultural and subfossil
for 11,000 years following the last glacial maxi- faunal materials (Fig. 2.7) left behind on
mum (LGM), and this change in sea levels submerged former coastal plains worldwide
resulted in the loss of an area larger than the (Al-Suwaidi et al. 2006; Anuskiewicz and
modern United Kingdom (Coles 1998, 2000). Dunbar 1993; Dunbar et al. 1989; Erlandson
The region therefore contains one of the most et al. 1998, 2011; Faught 2004a, b; Garrison
extensive and, presumably, best preserved pre- et al. 2012c; Grøn and Skaarup 2004; Grøn
historic landscapes in Europe (Gaffney et al. 2006, 2007; Kelley et al. 2013; Lavin 1988;
2007) and has been named “Doggerland” Lübke 2002; Merwin 2010; Rick and Erlandson
(Fig. 2.6; cf. Chap. 9). 2009; Rick et al. 2013).
This submergence of the present English The search for prehistoric submerged archae-
Channel and North Sea was not unknown to ological sites has gained increasing attention in
early twentieth-century archaeologists. Writing both the scholarly world and in the popular imag-
in 1916, H.F. Osburn specifically presents strik- ination and is, in part, driven by modern research
ingly accurate maps of the “English Channel that has demonstrated that climate change and
20 2 The Geomorphological and Geological Context

Fig. 2.6 (L ) Illustration of the English Channel River (Redrawn from Osburn 1916, Fig. 26); (R) early Holocene
Doggerland (Redrawn from Coles 1998)

geologically “recent” (2012). Helland-Hansen

et al. (supra) point out the fact: “Because we
live during a glacial highstand during which pre-
vious coastal plains are now flooded, our percep-
tion is colored by the present state. For long
periods of Earth history (especially greenhouse
intervals), shelfal bathymetric profiles were
likely not prominent features” (emphasis added).
Submerged, non-eroded, sites on continental
shelves often have a much higher potential to
preserve organic materials than similar, terres-
trial sites, and one can reasonably conclude that
prehistoric submerged sites have the potential to
Fig. 2.7 Lower. Maxilla of adult male mammoth
dredged from the North Sea off the Netherlands (Photo-
offer significant insights into human behaviors
graph by author) during any period of lowered sea levels and spe-
cifically badly needed insight into human
marine transgression are not gradual events to behaviors during periods of rapid sea-level rise
which human groups gradually adapt themselves. and climate change. Perhaps even more impor-
Instead, global research has shown that ice sheet tantly, archaeological research focused on the
collapse and subsequent sea-level rise often pro- end of the Pleistocene through the middle of the
ceed in rapid pulses (Berner and Berner 2012). Holocene (around 20,000 BP until 5,000 BP—
Likewise, the continental shelves themselves are note that all years are given in radiocarbon years,
2.4 Geomorphic Setting 21

not calendar) can provide insights into the colo- indications of ice free areas in the northeast end
nization and settlement of the prehistoric of, Graham Island. During, or just after a glacial
Americas. Without the inclusion of data sets maximum, decreased sea level, may have
gathered from submerged sites, our understand- exposed the floor of Hecate Strait and opened a
ing of coastal occupations during these early migration route via an exposed shelf along the
periods remains incomplete, and it is impossible eastern end of Dixon Entrance to the Alaskan
to understand the fuller spectrum of early human Islands, and perhaps as far north as the Bering
use of changing coastlines and coastal landforms Strait region. Certainly a 2500 drop in sea level
(Bailey and Flemming 2008; Benjamin and Hale would connect the Charlottes to the off-shore
2012; Faught and Donoghue 1997; Faught and islands and mainland” (Fladmark 1979). Davis,
Gusick 2011; Fischer 2011). in 2006, further notes that the evidence for a
The search for prehistoric underwater archae- coastal migration route is not new, citing
ological sites in North America has used, primar- Fladmark as well as Gruhn (1994, 1998).
ily, a methodology that focuses on the search for Submerged and buried prehistoric archaeo-
terrestrial evidence of early human presence and logical sites are, otherwise, difficult or nearly
then extrapolates that information to the near and impossible to detect without the use of “heroic”
offshore (cf. Fladmark 1979; Pearson et al. 1982; or expensive and/or time-intensive marine sur-
Benjamin 2010; Dixon 2013; Flatman and Evans vey methods. Typically, approaches to detection
2014). Some of the earliest dates for human of submerged sites on the continental shelves
occupation of the New World come from the have used geological and geoarchaeological
southeastern United States (Anderson and Gillam methods. The earliest examples examined seis-
2000; Faught 2008: 676, 683), while research mic data gathered for oil and gas exploration
within the last 30 years has demonstrated that such as in the Gulf of Mexico. In Europe, seismic
coastal areas are extremely productive zones data of a similar origin have also been used in
with a wide variety of food resources available larger, regional scale landscape reconstructions
year round (Reitz 1982; Rietz and Reitz 1988). termed “Doggerland,” now-submerged within
Taken together, these findings strongly suggest the North Sea (Fitch et al. 2005; Gaffney
that coastlines were probably occupied virtually et al. 2007) (Fig. 2.6).
as soon as human groups entered the Americas, Harris et al. (2005, 2013); Baldwin
following the so-called “coastline” hypothesis et al. (2006) have used multibeam bathymetry
that colonizing groups from northeastern Siberia surveys, side-scan sonar mosaics, high-resolution
followed the now-submerged coastline during subbottom profiles, and ground-truth surveys
entry instead of the so-called Ice-Free Corridor from 250 m to the modern tidewater region to
between the Laurentide and Cordilleran ice characterize the well-preserved but scattered
sheets (Erlandson et al. 2011). remnants of a submerged paleolandscape. By
Fladmark (1979) well prior to the period of coupling the data from sediment cores (Garrison
the general acceptance of a “Pre-Clovis” occupa- et al. 2008, 2012a, b; Chowns et al. 2008; Pavich
tion of the Americas observed, “A wide range of et al. 2010; and earlier studies), with these other
data suggests that a mid-continental corridor was methods, a drowned coastal plain can be
not an encouraging area for human occupation, if reimagined and related to Paleo-Indian cultures
even basically inhabitable, during the climax and who exploited that landform (Russell et al. 2009)
initial retreat of the Woodfordian ice.” In his (Fig. 2.8).
1970 report on the archaeology of the Queen Shallow features such as Pleistocene river
Charlotte Islands, now Haida Gwaii, he noted terraces buried by Holocene estuarine sediments
that “An interesting possibility is presented by were examined in the Gulf of Mexico using sedi-
the presence of cultural remains on the islands at ment coring methods. Cores extracted at specific
least 8,000 years old, and the geological seismic targets were examined for geochemical
22 2 The Geomorphological and Geological Context

Fig. 2.8 Paleochannels of the “Transect River,” Georgia (Stubbs et al. 2007; Harris et al. 2013). Inset, upper
Bight, Atlantic Ocean, offshore South Carolina (see inset, right, shows incision of the paleochannel into Pliocene
lower left), as imaged using multibeam sonar data lithology

and/or sedimentological markers for human 2.4.4 Mountain and Glacial Landforms
activities such as charcoal, bone, burnt shell,
and lithic debitage, and elevated Mn, Zn, and When Louis Agassiz established glacial theory
inorganic phosphate (Pearson et al. 1982, 1986, for good and all with his 1840 book, Etudes sur
2008, 2014; Stright 1986a, b, 1995). les Glaciers, archaeology, as well as geology,
In the southeastern United States, models adopted a powerful explanatory tool for
have been predicated on sea-level position in landforms over much of the northern hemisphere
tandem with regional geohydrology, and (1940, 1967). Geomorphic evidence of former
locations of lithic resources have been employed ice ages are often dramatic, notably erosional
successfully in the Big Bend of Florida (Dunbar features such as sculpted mountain terrains of
et al. 1989; Anuskiewicz and Dunbar 1993; cirques, arètes, horns, and hanging valleys.
Faught 2004a, b). In Europe, Benjamin (2010) Depositional features, such as moraines, while
proposes a generalized, stepwise methodology less picturesque, mark the limits of advances
for underwater research known as the “Danish and regressions of the ice sheets. In special
Model” that is extremely effective, with over an cases like the Salisbury Plain, the great stones
80 % success rate in locating underwater sites in of Avebury and Stonehenge likely originated as
Denmark where it was developed. This methodi- erratics carried by the movement of the
cal approach, or variants of it, has been well Finnoscandic ice sheet. It is their linkage to cli-
received due to its great adaptability to regional mate that makes the glaciers and landforms
or local contexts and conditions (Benjamin derived from them important to human prehis-
2010:258). tory. Like the volcano and river, the glacier has
2.4 Geomorphic Setting 23

shaped landscapes, started new soils, and, there- what led to these cold periods or their duration
fore, created new ecologies over 30 % of the was only be answered in the mid-twentieth cen-
Earth’s land surface. tury by the general sequence of glaciations which
The development of glacial theory, in the was established in Europe and America by the
mid-nineteenth century, paralleled that of unifor- start of the twentieth century by Albrecht Penck
mitarianism theory and that of the antiquity of the in Germany and T.C. Chamberlin in the US
Earth. Its impact on archaeological thinking then Midwest (Penck and Bruckner 1909). While
and now cannot be underestimated. What every presented as confirmation of the uniformitarian-
school child learns as the Ice Age was not pat- ism principle (Tarbuck and Lutgens 1984), it was
ently obvious to the geologists of the past century. perceived, at that time as anything but. Lyell’s
Prior to the development of glacial theory, glacial postulated uniformity of the Earth’s processes
features-moraines, outwash deposits, kettle lakes, from the beginning to “today” (1833, 1840) had
and erratics were simply the result of catastrophic no room for glacial theory (Hsű 1995). To make
floods aka Noachian. Lyell was more realistic room for recent (Pleistocene) glaciations,
about fluvial processes and suggested that Agassiz, with the help of Lyell’s mentor,
icebergs during the Noachian flood dispersed William Buckland, took Lyell to glacial
rock and sediment across the Earth’s surface. If moraines within two miles of his home. Glacial
this had been true, Lyell should have wondered theory, and “Ice Ages,” was accepted by the
why the bulk of these glacial landscapes only British, and Agassiz carried his message to
occurred in the northern hemisphere. America visiting at Harvard for some time.
Alpine farmers and mountain guides knew Glacial theory has immediacy for archaeology
what Lyell did not by way of daily experience. that stems from its description of an Earth, in the
Glaciers, even in their reduced modern forms, most important period of humanity’s develop-
advance and retreat, subject to climatic variation, ment, the Pleistocene, so unlike that of the modern
pushing rocks, sediments, and trees before them era. For the European archaeologists, no reading
and then leaving this as debris when they subse- of fossil sites or material culture could be made
quently melt. Ignatz Yenetz, a Swiss Civil Engi- without an eye toward how they fit into the ebb
neer, listened closely to a mountain guide, Jean- and flow of glaciations. For Americanists, the
Pierre Perraudin, and published, in 1821, a paper great continent-wide ice sheets presented a barrier
on the theory of glacial transport of exotic or to colonization of North and South America.
erratic boulders (also called “foundlings” beyond Human settlement of the Americas, by migration
the existing margins of Alpine glaciers (Hsű from Asia, had to wait for the right convergence of
1995)). In 1829, he went further to propose an natural opportunity and human necessity. This, as
Ice Age wherein glaciers covered nearly all of generally believed, came in the terminal Pleisto-
Switzerland as well as most of Europe. Most cene sometime after 20,000 years ago.
geologists doubted such a theory with the excep- The most dramatic archaeological discovery
tion of Jean de Charpenthier and his skeptical in the last decade of the twentieth century came
protegé, Louis Agassiz, Professor of Geology at at the Tyrolean glacier, on the Alpine border of
the University of Neuchâtel. Because the scien- Austria and Italy, with the discovery of the famed
tific method works best when used to falsify a “Iceman” mummy (Spindler 1994). Found at an
hypothesis or theory, Charpenthier and Agassiz altitude of 3,000 m, this intact prehistoric man,
set out to gather data to do just that. Their work and his accoutrements, reminded archaeology of
did just the opposite. In 1840, Agassiz published the importance of glacial landforms to human
his seminal work on glacial theory (supra). society as well as their potential for study. The
Implicit in Agassiz’s newly developed theory sacrificial mummies of the high Andes, at even
was the variability of the Earth’s climate-cold higher altitudes—6,100 m—are a testament to
periods leading to glacial advances—warm the importance of alpine regions to ancient
periods leading to their retreat. The question of cultures (Carey 1999).
24 2 The Geomorphological and Geological Context

Mountain landscapes, while can host glaciers, Rhône, itself, its course through the gorges
can be, and very often, unstable landforms in being blocked, not finding any exit or course
and of themselves. One recent example of to flow, until breaking free and creating great
geoarchaeological study is a tsunami of 563 devastation. After that, 30 monks came to
AD—Lake Geneva (SUI/FR). the location where the town (Tauredunum)
In 563 AD, the mountain above the town of was overwhelmed and searched the earth
Tauredunum collapses in a huge “landslide”/ (deposits) the mountain had brought there,
debris flow, burying the town and “running out” finding (only) bronze and iron.”
across the valley of the Rhone river just above
where it enters the Lake Geneva. Two contem- Finding the geoarchaeological evidence for
porary accounts exist for tsunami of 563 AD: this Medieval disaster fell to Katrina Kremer,
doctoral student at the University of Geneva,
(a) Bishop Marius of Avenches: “This year a who first conducted a sonar/echo sounder
great mountain at Tauredunum, in the dio- “pinger” survey in the fall of 2010 and then
cese of Valais, collapsed suddenly and took 10 m sediment cores into the lake sediments
destroyed the town and the nearby villages at 300 m water depths (Fig. 2.9).
and at the same time, their inhabitants. The Shown above, from a 2012 article (Kremer
landslide also moved the length of the lake, et al. 2012), in Fig. 2.9, and the cartoon,
60 miles long and 20 miles wide, flooding Fig. 2.9a, Kremer mapped, cored, and sampled
both shores, and destroying very old a huge sand deposit 15 km long and 7 km wide;
villages along with people and animals. 250 million cubic meters of sediment, 25 m
The lake demolished, at the same time, thick, was ejected into the lake, creating a giant
many churches and those that served them. wave or tsunami. Wood recovered in the cores,
Finally, the (wave?) struck the bridge at dated by radiocarbon, placed the deposit in AD
Geneva, the mills and its people entering 563. Kremer confirmed Gregory of Tours’
into the town killing many inhabitants.” account which clearly stresses the blockage of
(b) Bishop Gregory of Tours: “A great event the Rhone river and a subsequent inundation
occurred at the town of Tauredunum. This caused by the “rapid lowering” (se répandit
town is near/above the Rhône river. After impétuositié plus bas) of the accumulated water
not having heard for 60 days and more, so behind the landslide deposits, which subse-
loud the deaf could hear, the mountain, split quently failed. The tsunami was not the result
finally, from the other nearby mountain, and of a direct debris flow into the lake but a later,
crashed into the river, carrying along in the secondary discharge of a catastrophic flood of
landslide the people, the churches with their water and sediment into the lake, creating the
valuables which were in the buildings; the “bulge” of water necessary for a tsunami.
two banks (of the river) were obstructed. The In modeling studies of the event (2012),
water rose in the rear (of the blockage), the Kremer notes: “We have not accounted for how
place constricted, both shores of the valley, the subaerial rockfall actually initiated failure of
in whose throat the river ran. Having the delta, or the transformation of the subsequent
inundated the larger portion of the valley, it subaqueous landslide into a mass flow, for which
covered and destroyed that found on its we have no observational data.” Further, “. . is
banks. The waters that had accumulated sud- important to recognize that our results are still
denly dropped, drowning the people, strongly dependent of several key variables that
destroying their homes and the beasts of are uncertain, particularly the slide volume, and
burden perished; violently flooding all that the slide velocity. The slide volume used in
was found on both shores of the lake unto simulations is a minimum of that actually
Geneva. The mass of water spread as afore- measured and could potentially be even larger,
mentioned and passed over the walls. The that would generate even larger wave heights.
2.4 Geomorphic Setting 25

Fig. 2.9 (a) Shaded map

showing the AD
563 turbidite deposits
(“déplacement des
sediments,” in red) and
the supposed propagation
of the tsunami (“raz-de-
marée,” blue arcs) (Source.
24 Heures, Lausanne,
March 8, 2011). (b) Shaded
relief map (DEM) (Swiss
Federal Office of
Topography) of the Lake
Geneva. Water depths are
indicated by bathymetric
contour lines (meters). The
563 AD turbidite is color
shaded (orange-blue) as to
extent and sediment
thickness (meters). Vessel
tracks for seismic lines are
indicated. Sediment core
locations are indicated by
black squares. The inset
bathymetric profile, X-Y,
shows the gradient of
sediment thickness

The slide velocity could potentially be smaller of intense archaeological study in the twenty-first
than our used estimate, for example, due to inter- century—“ice patches”—in North America and
action with the water column resulting in loss of Europe, most notably Scandinavia and the Alps
momentum. Reducing the slide velocity from (Hafner 2012; Reckin 2013).
46 to 23 km/h decreased the maximum wave
height observed in Geneva from 8 to 3 m.”
More recently (2014, 2015), Kremer et al. have 2.4.5 Ice Patches
proposed earthquake-triggered mass movements
and tsunami as an alternative explanation for the With the onset of twenty-first-century melting,
observed occupation gap between 1872 and 1608 alpine areas, once glaciated, are yielding archae-
BC for Bronze Age settlements on Lake Geneva. ological evidence of prehistoric use (Reckin
Less dramatic, perhaps another important 2013). Prehistoric finds from four Swiss sites
aspect of glacial landscapes, has been the subject dating from the Neolithic period, the Bronze
26 2 The Geomorphological and Geological Context

Fig. 2.10 Schnidejoch

(site circled) in the Bernese
Alps of Switzerland (Fig. 2,
Hafner 2012)

Age, and the Iron Age have been recovered from et al. 2009), ice patch hunting focused on caribou
small ice patches (Schnidejoch, Lötschenpass, that sought out—and still do today—these
Tisenjoch, and Gemsbichl/Rieserferner). The features to escape biting insects. By “yarding
Schnidejoch, at an altitude of 2,756 m a.s.l., is a up” in the convenient ice patches, hunters were
pass in the Wildhorn region of the western aware of this behavior and used it to take game.
Bernese Alps (Fig. 2.10). It has yielded some of Evidence of this hunting behavior—two dart
the earliest evidence of Neolithic human activity shafts are from the Northwest Territories and
at high altitude in the Alps. The abundant assem- date from 2,310 to 2,410 cal BP and three other
blage of finds contains a number of unique artifacts from the Yukon include a dart shaft
artifacts, mainly from organic materials like (2,350–2,210 cal BP), a walking stick fragment
leather, wood, bark, and fibers. The site clearly (2,280–2,100 cal BP), and an antler point
proves access to high-mountain areas as early as (2,755–2,727 cal BP)—have been recovered.
the 5th millennium BC, and the chronological Bow-and-arrow technology dating between
distribution of the finds indicates that the approximately 800 and 150 cal BP has been
Schnidejoch pass was used mainly during found, including two complete arrows and two
periods when glaciers were in retreat (Hafner distal arrow shaft fragments, along with a very
2012). rare item: a partial willow bow dating to
In addition to the Neolithic finds, an Early 465–317 cal BP. Researchers also found
Bronze Age (EBA) round-headed pin was found modified feathers from a large bird, presumably
as well as an Iron Age coin together with signifi- used as fletching, in direct association with one
cant numbers of iron nails from Roman era shoe of the arrows. Lithic materials including a large
soles. biface and a projectile point were also located
In Scandinavia and the Yukon of North (Andrews, et al. 2012).
America, prehistoric and early historic finds High mountains were often difficult
have been found near ice patches (Reckin landforms for humans. Quoting the Swiss writer,
2013). In Norway and the North American Arc- Charles-Ferdinand Ramuz (1983), “Les
tic, animals and their hunters were actively seek- Montagnes, ils sont des milliers et vous? Vous
ing ice; ice patches with finds come in a wide etes seul.” (The mountains are thousands and
variety of forms. In the Yukon (Andrews you? You are alone.) While Ramuz is writing
2.4 Geomorphic Setting 27

primarily about Swiss mountain peasant attitudes In the Italian portion of the Southern Alps,
of the eighteenth and nineteenth centuries, he geoarchaeological studies in Trentino region
portrays the very real culturally based wariness have used stratigraphic and geomorphological
of mountain terrains. Still, archaeologists have methods to characterize landscape evolution of
demonstrated early peoples armed with stone the Trentino alpine region and its surroundings
technologies and hunting skills exploited high- from the Last Glacial Maximum (LGM) to the
mountain areas even if they did so seasonally Middle Holocene, approximately between 22,000
(Benedict 1985). In many areas of the world, and 4,000 years cal BP. The oldest evidence of
mountains are the integral portion of the land- the late glacial human occupation along the
scape. Archaeologists, working in these areas, southern margin of the Alps is recorded in the
have demonstrated that reconstruction of montane valleys. One of the key sites is Riparo
regional prehistory cannot be successfully Tagliente, in Valpantena (Verona, 250 m a.s.l.:
achieved without consideration of the processes Bisi et al. 1983; Bartolomei et al. 1985; Fontana
by which local populations adapted to mountain- et al. 2002; Table 2). This site was occupied from
ous ecosystems. Bender and Wright (1988) fur- the early late glacial, with dates ranging from
ther note that an archaeological disregard of 13,430  180 14c BP (15,360–16,548 cal BP)
mountains probably derives from our own cultur- to 12,040  170 14c BP (13,448–14,469 cal BP).
ally embedded notions of mountains as inacces- The site of Riparo Dalmeri (Dalmeri
sible and marginal (cf. Ramuz). Archaeologists, et al. 2005) with an earliest date of
nor geoarchaeologists for that matter, cannot a 12,900–13,303 cal BP shows clear evidence of
priori decide that mountains functioned as mar- the environmental situation during the
ginal environments in prehistory (Bender and Lateglacial Interstadial. The rock-shelter is
Wright, supra). located along a N-oriented wall at ca. 1,250 m
Research in alpine areas have conclusively a.s.l. Micromorphological studies have identified
demonstrated that prehistoric hunters and anthropogenic processes recorded in an
gatherers seasonally scheduled occupations of Epigravettian paleosurface (EDF stratigraphic
mountainous areas in order to procure the wide complex) that were related to the implementation
variety of resources available there (Bender of different activities: the preparation of the
1983; Wright et al. 1980). Local annual or even occupation surfaces by means of the accumula-
decadal scale variations in climatic conditions tion of silty sediment that was collected from
might curtail or prohibit occupation in any loess deposits off the site, the accumulation of
given year; the important factor is that the fine organic waste, tool production, trampling,
mountains, in general, were another stop on the and the remobilization of organic debris
round of annual movements made by Late Paleo- (Angelucci and Basetti 2009).
lithic and Mesolithic hunting and gathering In the Bighorn Mountains of the American
bands. The mountains thus become intrinsic, west, rock-shelters, specifically, 48BH1827
rather than marginal, to procurement strategies. (Two Moon Shelter) and 48BH1065
The foregoing example of prehistoric exploita- (BA Cave), have been the subject of
tion of ice patches more than illustrates this fact. geoarchaeological studies (Finley et al. 2005).
The geographer Carl Sauer defined a cultural Two Moon Shelter bears stratified Paleo-Indian
landscape which is one that has been “. . . fash- deposits that include a Folsom projectile point
ioned from a natural landscape by a cultural fragment, a 10,060  60 BP. radiocarbon date,
group. Culture is the agent, the natural are the and an undated Paleo-Indian component posi-
medium, the cultural landscape is the result” tioned stratigraphically below the Folsom occu-
(Sauer 1925). Like other natural landscapes, pation. BA Cave is a well-stratified and well-
mountains were the medium for cultures that dated archaeological deposit, yielding cultural
utilized them. occupations spanning the last 4,000 years.
28 2 The Geomorphological and Geological Context

Geoarchaeological investigations at BA Cave or otherwise—lacustrine landforms are primary

describe changes in sediment depositional loci for human settlement and natural resources.
regimes that may reflect significant late Holo- François Alphonse Forel, whose father, a histo-
cene trends in centennial-to-millennial-scale cli- rian, is shown in Fig. 2.11 making what is
matic changes in the Middle Rocky Mountain believed to be the first archaeological dive in
region (supra). history, is the “father of limnology” (Edgerton
In Nepal, geologically grounded appraisal of 1962; Vincent and Bertola 2012).
the paleoenvironment and geological data of the François Alphonse Forel (Fig. 2.12) never
Plio-Pleistocene and early Holocene periods ignored the relationship of human cultures to
allows reconstruction of the early history of lacustrine systems. In his three volume treatise
man before the beginning of the historical period (1892–1904), he synthesized his physical-chem-
in the Himalayas of Nepal (Pandey 1987). In the ical-biotic-human-coupled approach to the study
American cordillera, yet again, workers like of lacustrine systems. Because of his early
Cortegoso (2005), Ramiro et al. (2011); involvement in his father’s study of submerged
Sandweiss et al. (2009) have used prehistoric sites in Lake Geneva, F-A. Foral or
geoarchaeological methods to illuminate the “FAF” as he was known to his peers included the
relationship of montane prehistoric populations historical importance of a lake to people, ancient
to Holocene environmental change. to modern (Vincent and Bertola, supra). Because
of his lifelong interest in these ancient sites, he
devoted 78 pages to them in volume III of this
2.4.6 Lacustrine Landforms Monograph on Limnology (1892–1904). In this
work, he updated the 1860 map of the prehistoric
This author first wrote on lacustrine or limnolog- sites of Lake Geneva listing 47 sites. Most
ical processes in relation to drowned prehistoric importantly, Forel was the first to propose an
sites in lakes in 1975. That article dealt primarily integrated geological and sedimentological
with man-made lakes or reservoirs. With the approach to understand the location of these
exception of the artificial dam, artificial and nat- sites on the drowned terraces of the lake bottom
ural lake systems are very much alike. In the (Corboud 2004:28).
Lake Geneva example, above, and in the discus- Lochs are natural lacustrine systems in
sion of African lakes later in this chapter—paleo Scotland. There are over 30,000 lochs, large

Fig. 2.11 The first

underwater archaeology
dive—August 24, 1854,
Lac Léman (Lake Geneva).
François Marie Etienne
Forel, father of a then
13-year-old François
Alphonse Forel, is seated in
the boat with his back
turned, presumably
operating the air pump for
the diver (Adolphe Morot),
while Frédéric Troyon
holds a safety line (Water
color by Adolphe von
2.4 Geomorphic Setting 29

Fig. 2.12 Water color

portrait of François
Alphonse Forel in the room
he used as his study and
laboratory, at his home in
Morges on the shores of
Lake Geneva (Lac Léman),
Switzerland (Watercolor
by Ernest Bieler, no date,
private collection)

and small (T. Nicholas Dixon, Personal Commu- for study. Certainly geoarchaeology has much to
nication 2012). Much like the prehistoric contribute to their study as well.
settlements in Alpine lakes, in Scotland and A recent example of geoarchaeology’s contri-
Ireland, prehistoric peoples utilized the lochs bution to prehistoric use of a lacustrine environ-
for resources and settlement (Dixon 2004) from ment is the Paleolithic site of Schöningen
the Neolithic into the Medieval Periods. A (Thieme 1997; Stahlschmidt et al. 2015). The
unique aspect of this human use was the creation original excavators of this important Middle
of pile-dwellings known as crannogs (Munro Pleistocene site proposed that hominin activity,
1882; Morrison 1985; Fredengren 2002; Dixon likely that of H. heidelbergensis, included hunt-
1983, 2004). Their importance of crannogs to ing and butchery and occurred on a dry lake
prehistoric archaeology is uncontested, but they shore followed by a rapid sedimentation of
are understudied. One of their obvious values is organic deposits that embedded and preserved
the enhanced preservation of organic remains not the artifacts. Recent geoarchaeological analysis
seen on most terrestrial sites. Likewise, these challenges this model. Evidence supports a dif-
lacustrine sites are underappreciated as their ferent scenario where the sediments of
overall importance in prehistoric settlement Schöningen 13 II-4 aka “spear horizon” were
systems. Unlike the shoreline villages of the deposited in a constantly submerged area of a
Neolithic and Bronze Ages in Alpine Europe, paleolake. In one suggested depositional sce-
crannogs were family-based structures built nario, cultural activity occurred on a shallow
upon artificial islands of stone and clay (Dixon frozen lake surface and during thaws and warm
2004). The very difference in the social origin of seasons of both remnants—butchery tools and
the crannog and that of larger settlements begs the remains of the butchered simply fell to the
30 2 The Geomorphological and Geological Context

bottom—following the geological concept of parabolic, longitudinal, and linear. When active,
dropstones, while in another the materials are these features were not likely to have been the
relocated by geogenic processes (Stahlschmidt locus of much human settlement, but after the
et al. 2015). establishment of a vegetation cover, they have
While Olduvai Gorge is used as an example of occupied by both pastoralists and agriculturalists.
the geological stratigraphic section, it could as Geoarchaeological studies of artifacts and buried
easily be an exemplar for geoarchaeological stud- surfaces in aeolian deposits can provide signifi-
ies of lacustrine paleoenvironments of the Pleisto- cant information for the archaeologist (Ivester
cene. The Olduvai Basin in northern Tanzania is a et al. 2001; Leigh 1998).
Plio-Pleistocene sedimentary basin containing a Highway construction in Northwest Missouri
succession of fluvial and lacustrine deposits that exposed a stratified archaeological site that was
are largely derived from volcaniclastic material found to contain a range of lithic materials
(Hay 1976, 1990). These well-known fossiliferous extending into the Paleo-Indian Period (Reagan
deposits contain proxy records of the evolution, et al. 1978). The excavations carried out in the
behavior, and paleoecology of early hominins mid-1970s were characterized by the extensive
such as Homo habilis and Zinjanthropus boisei use of geoarchaeological techniques to describe
(Leakey 1971; Tobias 1991). Lowermost Bed II the deposits. The site, now destroyed, was called
(LMBII) is 1.75–1.85 myr old based on ages of the Shriver site and was located in the northwest-
the basalt “lithological floor.” The Olduvai Land- ern prairie section of Missouri on a plain
scape Paleoanthropology Project (OLAPP) has characterized by loess and glacial drift deposits
sought to understand synchronic, basinwide overlying shale and limestone bedrock (Dort
patterns of land use by the Oldowan hominins. 1978). Its discovery and its distinctive lithic
Paleoenvironmental reconstructions (Hay 1976, assemblage raised the possibility of it being a
1996; Bonnefille 1984; Cerling and Hay 1986; so-called “Pre-Clovis” site. This assignation
Sikes 1994; Hay and Kyser 2001; Deocampo and was based on the artifactual stratification wherein
Ashley 1999; Ashley and Driese 2000; Deocampo a Folsom period occupation (ca. 11,000 BP) was
et al. 2002) have provided a context for interpreting superposed on a deeper (~40 cm) assemblage of
associated archaeological finds (Blumenschine bifacial and unifacial flint artifacts containing no
and Masao 1991) and test aspects of recently pro- lanceolate projectile points (Reagan, supra).
posed hominid land-use models (Peters and Because of the unusual nature of the archaeolog-
Blumenschine 1995, 1996; Blumenschine and ical assemblage, the excavators conducted exten-
Peters 1998). As noted in Chap. 3, Hans Reck sive geoarchaeological studies to include
and later Hay distinguished these beds within the geological/geochemical, soil/sediment, pollen,
Olduvai section based upon their color, sedimen- and geochronological procedures. The author
tological texture, and biostratigraphy. was involved in the geochronological aspect of
the study utilizing thermoluminescence (TL) to
investigate flints suspected of being heat-treated
2.4.7 Loessic and Glacial Landscapes: (Reagan, supra).
Midwestern United States From a lithostratigraphic perspective, Dort
reduced the 13 units to a section of four geologi-
In the Midwestern United States are landforms cal units which he called: (i) Unit 1, the Bignell
that dramatically display the interplay of wind- loess, 8,000–13,000 BP; (ii) Unit 2, Peoria Loess,
born deposition of both sand and silt (loess). 13,000–18,000 BP; (iii) other glacial and ero-
These are the sand hill regions of the state of sional materials; and (iv) Upper Pennsylvanian
Nebraska along with the great loess/sand prairies bedrock (Reagan, supra). He placed the three
of Nebraska, Kansas, and the Missouri River cultural stratigraphic units into the upper two
Valley (Fig. 2.13). Most identifiable of aeolian geological units. The older of the culture strati-
landform are dunes—barchan, transverse, graphic units, “Pre-Clovis,” was located on the
2.4 Geomorphic Setting 31

Fig. 2.13 Map by Bettis et al. (2003) showing loess distribution and thickness in the Great Plains of the United States

sub-loess (Peoria) erosional surface implying 215 cm. Dort identified an erosional surface and
an antiquity of at least 18,000 years. Soil chemi- probable paleosol at this location as well. No
cal trends, particularly the phosphorus and archaeological significance was attached to this
potassium concentrations, showed a correlative lower geomorphic surface, but its paleoclimatic
discontinuity at this stratigraphic location relevance is obvious.
(65–75 cm). A similar break in the geochemical The most well-described post-LGM paleosol
signatures was seen deeper in the section at for the Midwestern United States area is the
32 2 The Geomorphological and Geological Context

Fig. 2.14 The Brady

paleosol (Photograph
courtesy of Charles

so-called Brady Soil. Peoria Loess deposition in buried mollisol which has the pH profile increasing
the central Great Plains appears to have occurred with depth together with high silt (63–76 %)
between approximately 20,000 and 10,500 BP values typical of loesses. The observed clay ratio
(Martin 1993) (Fig. 2.14). The Brady Paleosol (0.3) in the mollisol’s Ap horizon increases to a
is identified as carbon-rich, organic layer across value of 1.4 in the Bt2 horizon (supra). Argillic
most of Kansas and Nebraska. Formation of the horizons in mollisols have fine-to-coarse clay
Brady soil initiated between 15,500 and ratios ranging from 1.0 to 1.5 (Wascher
13,500 cal year BP, during a time period of et al. 1960). Likewise, in well-drained examples,
much slower dust deposition (0.1–0.2 mkyr1), the clay mineral alteration is minimal. Follmer
which allowed for pedogenesis and the accumu- noted some properties, the presence of incipient
lation of SOM (Marin-Spiotta et al. 2014). E horizon, in the mollisol that indicated the possi-
Increasing rates of loess deposition over time, bility of polygenesis or overprinting caused by the
estimated at 0.72–0.80 mkyr1, buried the presence of a later forest cover (supra). The C
Brady soil approximately 10,500–9,000 years horizon is the Henry Formation. Follmer does not
ago. specifically discuss the archaeological materials
The work of Follmer (1982) at the central Illi- found at the Rhoads site, but they are later
nois archaeological Rhoads site identified a Peoria (Archaic Period) than those excavated at the
Loess deposit on the till plain of the Illinoian Paleo-Indian Shriver site.
(penultimate) glaciation just south of the morainic Balco and Rovey II (2010) have used cosmo-
margin of the later Wisconsin glaciation. At the genic isotopic dating methods and magnetostra-
Rhoads site, the Peoria Loess caps sand and gravel tigraphy to assess the chronology of the major
outwash deposits termed the Henry formation, Pleistocene ice advances in North America. The
whose aggradation peak was about 17,000 first two of these advances were found to corre-
BP. Follmer gives the Peoria Loess a wider time late well with the newly proposed Pleistocene—
range than Dort—12,600–23,000 years—based on Pliocene boundary of 2.58 Ma (~2.4 Ma) and the
radiocarbon dates for its base, near St. Louis, of “mid-Pleistocene transition” at 1.3 Ma. The ini-
23,110  280, 23, 930  280 BP (McKay 1979). tial advance reached 39o N latitude or what is
On the Peoria Loess was identified a paleudal or today the Missouri River Trench in central
2.4 Geomorphic Setting 33

Missouri (Fig. 2.6). Lithological evidence for the complex fluvial history in a now hyperarid and
five advances lay in individual glacial tills. In uninhabited region (supra). Well-defined
dating the tills, the authors of this study were in networks of broad alluvial valleys, braided
fact dating underlying paleosols which had streams together with bedrock-controlled
formed from the overlying tills. The oldest channels, appear in the radar images and have
paleosols were named the Atlanta and Moberly been confirmed on the ground by archaeologists.
Formations and were magnetically reversed, Along these “radar rivers” have been found evi-
making them older than the Brunhes-Matuyama dence of some of the earliest inhabitants of north-
Boundary (~780 Ka). The younger tills were east Africa and northwest China (Wendorf n.d.;
magnetically normal, indicating ages less than Holcomb 1996). Allogenic streams exist in many
that of the Brunhes-Matuyama event. These tills arid lands—the Nile in Ethiopia and Egypt, the
were the Fulton, Columbia, and Macon Jordan in the valley and hills of the Dead Sea, the
Formations. All predated MIS 6 (~150 Ka). The Truckee ending in the Nevadan desert at Pyramid
Columbia is till dated to before this time Lake, and the Colorado of Grand Canyon to the
(0.22  0.16 Ma), whereas the Macon Forma- Gulf of California. These fluvial systems can be
tion was not well constrained, suggesting only sensitive indicators of paleoclimate because of
that it was younger, but uncertainties in the cho- their aggradation and erosion due to rainfall out-
sen cosmogenic method (26Al/10Be) prevented side the bulk of their drainages (Hassan 1988).
reliable dating of the final till (supra).

2.4.9 The Sahara: Geoarchaeology

2.4.8 Desert/Arid Landforms of Paleolakes and Paleoclimate

Just as nearshore continental shelves were one The Sahara is the largest of the world’s hot
coastal plain, drylands and deserts were often deserts (the polar regions are deserts, too, and
verdant savannahs and forests. Drylands form larger). C.B.M. McBurney, writing in 1960,
35 % of the Earth’s land surface (Oviatt stated, “The central fact of the North African
et al. 1997). Within these drylands are true environment is the Sahara (p.62).” McBurney
deserts (16 %). Water plays an important role in went on to say, “. . many the name conjures
shaping dryland topography but a more signifi- up a vision of vast area covered almost entirely
cant factor is the wind. The reason the wind is with rolling sand dunes and scattered here and
important is self-evident from an inspection of there with oases of almost incredible fertility.” It
the term “dry” land. is, as he points, the largest single tract of aridity
Dryland landforms created by water vary in the world spanning Africa from the Red Sea to
from arroyos (“wadis” in Old World parlance) the Atlantic. That said, natural vegetation,
to eroded mesas and buttes. Playas form in though rare, is not totally absent and takes the
deserts where precipitation is insufficient to form of widely spaced clumps of desert shrub
maintain permanent lakes. Where lakes do exist and, toward the desert’s margins, clumps of
in dryland areas such as the Great Salt Lake bunch grass extending over perhaps 60 % of the
(USA), the Dead Sea (Israel/Jordan), and Pyra- Sahara’s vast area (supra, 64). The southern mar-
mid Lake (USA), snow melt and other runoff gin is a zone of steppe and savannah some 400+
feed the basin from outside the immediate area km wide grading into tropical forest. The north is
of the lake itself. Ancient fluvial systems or steppe as well but narrower. The western half is
paleo-drainages have been revealed in the shal- mountainous—the Atlas Range—with conifer-
low subsurface of the eastern Sahara in the driest ous scrub and forest due to higher rainfall.
region of southern Egypt and northern Sudan In the central Sahara lie two important, in
(McCauley et al. 1982). Space shuttle-based prehistory, highland areas—the Hoggar and
radar provided us with evidence of a long and Tibesti massifs. A third, lesser massif—the
34 2 The Geomorphological and Geological Context

Air—lies southeast of the Hoggar. Peaks in the during the last 3,000 years of hyperaridity, rain-
Hoggar often exceed 2,500 m and two, in the fall has been measured only in millimeters a year,
Tibesti, are 3,000 and 4,000 m. These heights so the deficit must be made up by the aquifer. In
are equal to many well-known peaks in the the prehistoric Green Sahara era, an ancient lake,
Alps such as Schilthorn and Jungfrau. In prehis- the Mega-Chad, 50 or 100 times—an order of
tory, these heights are thought to have influenced magnitude—more extensive than the present-day
climate and cultures across the heart of the Lake Yoan, may have existed. The Green Sahara,
Sahara (Williams 1984b). Nonetheless, vast known as the African Humid Period, lasted
dune fields do exist that give the Sahara its roughly from 11,000 to 5,000 years ago. A
well-deserved reputation for inhospitality to all North African savannah—now the parched
but the hardiest. Sahara—existed for elephants, giraffe, hippos,
Nubian Sandstone Aquifer System underlies antelopes, and aurochs that once made their
much of the eastern Sahara. This is the world’s homes there.
largest fossil water aquifer, and it spreads In a 2008 research paper, geoarchaeologist
roughly 2,000,000 km2 (772,000 sq. mi.) beneath Stefan Kröpelin challenged the hypothesis that
Chad, Sudan, Egypt, and Libya to a maximum the Sahara underwent rapid desertification in the
depth of 4,000 m (12,8000 ). The Lakes of Holocene. Advanced by P.B. deMenocal (1995,
Ounianga, 18 interconnected, freshwater, saline, 2000), the North Africa dried from savannah to
and hypersaline lakes (Fig. 2.15), located more desert over the course of only a few centuries.
than 800 km (500 mi) from the nearest other lake, DeMenocal et al. used the record of a marine
Lake Chad, are supplied from below by this sediment core of Late Pleistocene-Holocene ter-
aquifer. Sahelanthropus tchadensis, dated to rigenous (aeolian) sediment at ODP Site 658C,
over 6 m.y., was discovered in 2001 and 2002 off Cap Blanc, Mauritania, to document very
in the Chadian desert by Michel Brunet abrupt, large-scale changes in subtropical North
et al. (2002) nearby. As the result of a hydrologi- African climate. This terrigenous record,
cal system, all but 1 of the 11 lakes of Ounianga according to deMenocal, exhibited a well-
Serir are fresh. defined period of low sediment influx between
The intense Saharan sun drives the evapora- 14.8 and 5.5 cal. ka BP associated with the Afri-
tion at nearby Lake Yoan, lowering the water can Humid Period, when the Sahara was nearly
level by 6 m (190 ) a year. Now, and presumably completely vegetated and supported numerous

Fig. 2.15 The area of the

Sahara Lakes (Adapted
from: Saudi Aramco World,
vol. 65 (3), 2014)
2.4 Geomorphic Setting 35

perennial lakes and an arid interval corresponding How did climate and tectonic change interact
to the Younger Dryas Chronozone. Using new during critical intervals of human evolution? To
data, from Saharan lake sediments, Kröpelin answer this fundamental question, recently,
asserted that it may have been a far longer, 2012–2013, scientific drilling teams have
much more gradual process of desertification, extracted long cores from five high-priority
one taking not centuries but millennia. areas in Ethiopia (Northern Awash River and
One biomineralogical piece of evidence for Chew Bahir areas) and Kenya (West Turkana,
ancient lakes is diatomite. Diatomite is alight Tugen Hills, and Southern Kenya Rift/Lake
sediment made up of the microscopic skeletal Magadi) where highly resolved, continuous
remains of diatoms, single-celled plants that lacustrine paleoclimate records can be collected
have sunk to the bottom of lakes and oceans. through important time intervals in the same
When the water of a lake recedes, diatomaceous basins that contain fossils and artifacts (Pennisi
earth can often be found on the walls of the 2013) (Fig. 2.16).
sandstone escarpments that surround the The extraction of high-resolution (centennial-
Ounianga Kebir and Serir basins. Sampled, the millennial scale) sediment cores allow acquisi-
elevation of deposits noted and carbon dated, the tion of datable, geochemical, paleoecological,
approximate depth and the extent of the Chadian and sedimentological records. These proxy data
“paleolakes” can be postulated. from the paleolake sediments include tephra,
In 2010, Kröpelin and Martin Melles of the pollen, phytoliths, macrobotanical remains,
Quaternary and Paleoclimatology Group at the diatoms, algae, macro- and micro-invertebrate
University of Köln directed the recovery of a and vertebrate remains, as well as any
16 m (52½0 ) sediment core form Lake Yoan, in artifacts—whole or fragmentary—fashioned by
pursuit of layers from the Holocene Era. The hominins.
laminated sediments have yielded a continuous, The HSPDP seeks, much like the Sahara lakes
10,940-year continental record of climate and study, to use lake sediment records, for continen-
environmental change for the south-central tal sites, to address paleoenvironment. Where
Sahara. Kröpelin et al. summarized, in 2008, the Saharan research focuses on the Late
that the Lake Yoa(n) record supports archaeolog- Pleistocene-early Holocene climatic variation,
ical and geological data from the eastern Sahara the Rift’s paleolake study delves more deeply
as well as palynological data from the West Afri- into the period of hominin evolution, the Plio-
can Sahel that the record of Saharan dust deposi- Pleistocene. With the recalibration of the base
tion in the tropical Atlantic Ocean is not date for the Pleistocene at 2.588 m.y., more
representative for landscape history throughout hominin genera, such as Homo erectus/ergaster
dry northern Africa. The dispute based on two and habilis, are now exclusively within this
geoarchaeologically oriented sediment studies geological era. Australopithecus species and
points out the disparity that can accrue between pre-Australo genera such as Orrorin,
even the best of data sets. Ardipithecus, and Sahelanthropus are firmly
In another, recent, large-scale geoarchaeological rooted within the Pliocene and late Miocene.
drilling project, the Hominin Sites and Paleolakes The Rift’s study addresses what is termed the
Drilling Project (HSPDP) has begun the study of variability selection hypothesis (Conroy and
six ancient lakebeds in the African Rift (Pennisi Pontzer 2012). This hypothesis attempts to wide-
2013) (Fig. 2.15). In the Baringo Basin, the project spread Plio-Pleistocene climatic and associated
has drilled to a depth of 228 m, extracting sediment environmental change to important events in
cores. This study, like that of the Lakes of hominin evolution (Potts 1998). HSPDP data
Ounianga research, seeks to examine the synchro- are hoped to be able to directly identify increased
nicity of deep sea cores’ findings with that of the environmental variability, allowing evaluation in
African continent. regard to hominin adaptations.
36 2 The Geomorphological and Geological Context

Fig. 2.16 The Rift

Paleolakes: NA Northern
Awash, CB Chew Bahir,
WT West Turkhana, TH
Tugen Hills, and LM Lake

2.4.10 Karst/Cave Landforms Witwatersrand, the massive dolomite strata con-

tain vast quantities of water that are of concern to
Nowhere is the linkage of hydrology and geology miners (Garrison 1999).
so apparent than in these types of landforms and Precipitation is critical to the production of
features. In some cases, the situation is one of the karst but only if the water table is high enough
underground drainages carving often vast solu- to facilitate the dissolution process. Therefore,
tion features rather than great fluvial valleys. development of karst is, like the fluvial landform,
Karst landforms are generally considered to orig- linked directly to climate and, importantly, the
inate on carbonate rocks, in most cases, geological structure (Palmer 1991). Faults,
limestones. This limestone can be composed of joints, and fractures form the conduits for water
50 % or more calcite or aragonite, both and entrained chemicals, such as carbonic acid,
pseudomorphs of CaCO3. In the Ozark Mountains that dissolve the limestone. Where sulfide
of the Central United States, karst forms in deposits exist along fault planes, the production
dolomites—CaMgCO3—that is more than 50 % of sulfuric acid will accelerate the growth of
this species, but more often limestone is the rock karst.
system (Easterbrook 1998:194). Springs are very The most recognized karst landform is the
common in these landforms. In Missouri, heart- sinkhole or doline. These closed depressions are
land of the Ozark Mountains, the dolomites are solution/collapse features that vary around a
rich in caves and springs (Vineyard and Feder basic, circular or elliptical, funnel-like shape. In
1982). In the gold reefs of South Africa’s some of these, the local hydrology forms a
2.4 Geomorphic Setting 37

spring. For archaeology, these sources of water Chauvet Cave, located along the Ardeche River,
have concentrated both fauna and human a tributary of the Rhône, lies on the south flank of
cultures, forming deposits of both within and the Massif Central and dates to the same period
around the spring. Where the water has aban- as Cosquer.
doned the subterranean feeder conduits of the
ancient springs, caverns and caves now exist.
Collapses occur over time that form traps for 2.4.11 Combe Grenal
animals and humans. In Spain, collapsed lime-
stone caves have produced some of the most As noted in the introductory chapter, the term
spectacular evidence of early humans on the “Paleolithic” was a British invention, but the
continent of Europe. The collection of sites is French applied it to their extensive variety of
known as Atapuerca that lies within the northern prehistoric sites. One such site is Combe-Grenal.
sierra of Spain. The caves, known since 1976, The excavation and significance of Combe-
have, since the beginning of serious excavations Grenal is most closely associated with Francois
in 1990s, produced over 1,600 human bones Bordes. Francois Bordes was the leading French
ranging in age from greater than 780,000 to Paleolithic archaeologist for most of the mid-part
127,000 years ago. This chronology has been of the twentieth century. Bordes was encouraged
derived from work in what is termed “the by Peyrony to excavate there (Dibble
Gran Dolina” or Big Sinkhole (Renfrew and et al. 2009a, b). At Combe-Grenal, Bordes and
Bahn 1996, 2000, 2004). his students, from 1953 to 1965, revealed a
The nearby site Sima de los Huesos or “Pit of sequence of almost 13 m of rich archaeological
Bones” has yielded the bulk of the large collec- deposits, apparently spanning the period from the
tion of human remains found to date and final stages of the penultimate glaciation (MIS 6)
revolutionized our thinking of the colonization through to almost the end of the Mousterian
of Europe by early humans. Indeed, in Europe succession, from 125,000 BP to 40,000 BP
and elsewhere, some of our earliest fossil humans (Fig. 2.17) (Bordes 1972). They excavated
and their material culture has originated in caves. 64 stratigraphic levels, and using these data,
The first of these was Neanderthal, discovered in Bordes refined the Middle Paleolithic stone tool
1856, in its eponymous valley. A continent typology used today by most workers in this
away, Homo erectus known as “Peking Man” period (Dibble et al. 2009a, b).
was found in the debris of the “Cave of the Bordes was trained as a geologist. As
Dragons”—Zhoukoutien. discussed in the introductory chapter, into the
In 1990 and, shortly thereafter, in 1994, two mid-twentieth century, French prehistorians
dramatic cave art discoveries were made in continued to hue to a “paleontological” paradigm
France. These solution features, in massive for conceptualizing the stratigraphy of their
limestones, south of the Massif Central, were, excavations (Fig. 2.17). Strata were viewed
respectively, Cosquer Cave and Chauvet Cave. biostratigraphically but material finds—tool
Both contain some of the oldest Upper Paleo- types ( fossile directeurs)—the “biota,”
lithic art ever discovered—29 to 35,000 years separating antecedent and subsequent units
(Clottes and Courtin 1996; Clottes 2001). Both (Sackett 1999). This approach to stratigraphy,
are located outside the heartland of Paleolithic Bordes discarded at Combe-Grenal. Bordes was
cave art, the Loire River valley and northern quick to recognize the fossile directeur approach
Spain. Cosquer Cave is unique, by virtue of its had two implicit assumptions: the first was the
discovery below the surface of the Mediterranean conflation of the Paleolithic record with that of a
near Marseilles. The cave entrance was found paleontological record—a one-to-one correlation
30 m below the sea surface by scuba divers. between archaeological levels and “natural”
Its entry, by ancient artists, occurred sometime stratigraphic units—and the second was the
during the last glaciation (Würm/Weichsel). excavated material record, tool traditions,
38 2 The Geomorphological and Geological Context

Fig. 2.17 Combe-Grenal stratigraphy—Mousterian photograph, from Bordes (1968), above. Levels 56–64 are
levels 24–54 are shown in photo and inset. The rear portion shown in the elevation section, below, which corresponds
of the inset fence diagram (B) is correlated with the to the far left of the fence diagram (Bordes 1972)
2.4 Geomorphic Setting 39

features, etc., which would reflect a specific Paleolithic research over the last half century
block of time in the archaeological record (supra).
(Sackett, supra). In short, major episodes of Pech de l’Azé IV is another of a complex of
depositional history were reflective of similar four Lower and Middle Paleolithic sites located
episodes of cultural history. in the Perigord region of southwest France.
The Bordesian approach, as it has been Bordes did spend extensive efforts at this site
termed, used well-defined artifact assemblages (Bordes 1972). Bordes discovered and first tested
segregated into “occupational horizons” and the the site in the spring of 1952 (Bordes 1954) and
minimal sedimentological units that were then excavated there continuously from 1970 to
discernable within these. Bordes viewed the rel- 1977. It and Combe-Grenal were two of the more
ative frequency of tool types, ensemble, with important examples of Bordes’ methodology—
their paleoenvironmental contexts as critical to the “Bordesian approach”—and his lifelong
developing a true chronostratigraphic description attempt to bring order to the Middle-to-Upper
of a site. One glance at his stratigraphic section Paleolithic in France. Goldberg, in 1979, and
for Combe-Grenal will confirm this. Combining then later (Goldberg and Sherwood 2006) did
the definition of polymorphic tool assemblages groundbreaking micromorphological studies of
with simple statistical measures, Bordes was able the deposits at Pech de l’Azé IV (Fig. 2.18). At
to wring greater complexity out of the Paleolithic this Middle Paleolithic cave, well-rounded
record. This said, Combe-Grenal was one of pebbles recovered during Bordes’ original
Bordes’ first applications of new proveniencing excavations were initially thought to be hammer-
methods, and it required managing these stone artifact. However, they, now, appear to be
methods over a very large-scale excavation. fluvially transported Pleistocene endokarstic
Because of this, the Combe-Grenal collection deposits that significantly predate the occupation
has some serious problems, especially in terms of the cave (Goldberg and Sherwood, supra).
of the stratigraphic integrity of the assemblages Bordes did not excavate Fontéchevade
(Dibble et al. 2009a, b). This is troublesome (Charente) cave, but perhaps he should have
because of the long sequence and also the rich- given the recent geoarchaeological reassessment
ness of its lithic and faunal assemblages at of the site formation processes and taphonomic
Combe-Grenal. The site has played a central factors at play there (Dibble et al. 2006; Chase
role in many of the dominant issues in French et al. 2009). The Tayacian, as an archaeological

Fig. 2.18 West profile at

Pech de l’Aze´ IV. This
profile typifies the deposits
in this region. It consists of
variously sized blocks of
limestone éboulis fallen
from the roof and walls of
the cave within a sandy,
iron-stained matrix. The
sand is derived ultimately
from breakdown of the
limestone and accumulated
via different sources from
outside and mainly within
the cave, including slope
wash, creep, and
solifluction (Photograph by
Philippe Jugie)
40 2 The Geomorphological and Geological Context

Paleolithic culture, was first recognized by the documented diagenetic and mineralogical pro-
French archaeologist Denis Peyrony at the site of cesses and their ramifications for correctly
La Micoque (Peyrony 1938), located in the interpreting anthrogenic deposits (Goldberg and
Dordogne in SW France. Not long afterwards, Sherwood 2006.
excavations at another site 80 km to the north- Along the coast of South Africa are numerous
west, in the Charente, yielded an even larger Paleolithic cave sites where geoarchaeological
Tayacian (Henri-Martin 1957). It was for this studies have made salient contributions. Klasies
reason, this latter site, Fontéchevade, ultimately, River Mouth cave has produced skeletal evi-
become the reference site for the Tayacian (Dib- dence of anatomically modern humans (AMH)
ble et al. 2006). as early as 90,000 years BP in what is termed the
Fontéchevade illustrates the importance of Middle Stone Age (MSA) (Ambrose and
revisiting and reevaluating earlier excavations Lorenz 1990). At Die Kelders Cave east of
with new techniques and methodologies. Chase Cape Town, MSA deposits have been shown to
et al. (2009) and Dibble et al. (2006) present postdate a last interglacial (OIS 5) high sea level
comprehensive reviews, based primarily on that overlies earlier deposits (Grine et al. 1991).
geoarchaeology, of the sedimentological context, Farther east Blombos and Sibudu Caves have
as well as the reality of the Tayacian as a Paleo- produced evidence of MSA stone and bone
lithic stone industry. When the site was origi- tools in association with early fire (Backwell
nally excavated, the association of bones and et al. 2008; Henshilwood et al. 2001; Schmidt
stone tools in a cave was uncritically accepted et al. 2012). The South African caves, in general,
as evidence of human agency in their accumula- are important because of the demonstrated
tion. Re-excavation and reanalysis in the 2000s signatures of modern human behavior in the use
have led to the conclusions that Tayacian of tool, ornamentation, and control of intentional
assemblages at Fontéchevade are primarily the fire (Klein 2000; Shimelmitz et al. 2014).
result of natural formation processes. These pro- With regard to early fire and hominins, results
cesses probably introduced some modified from micromorphology and FTIR refute many of
artifacts into the site and clearly produced the commonly held notions regarding the inten-
pseudo-artifacts that were originally interpreted tional use of fire by humans at Zhoukoudian
as artifacts. It is this combination of damaged and Cave (Goldberg et al. 2009; Goldberg and
pseudo-artifacts that became the basis for the Sherwood 2006). A similar type of study was
assemblages known as Tayacian (supra). The conducted at Kebara Cave, where numerous
authors concluded that the use of the term combustion features are also preserved. As at
Tayacian should be dropped. In this regard, it Tabun, higher concentrations of phytoliths
could join the Osteodontokeratic “culture” (Dart occurred in micromorphological samples from
1957) as an example of an industry that simply within the cave than in those from outside. The
isn’t. results indicate that human transport of fuel was
Beyond Europe, most notably the above the major source of phytoliths into the cave
examples from southwestern France, cave sites (Goldberg and Sherwood, supra).
on all continents have been extensively studied
by geologists. The study of caves is, perhaps, one
of the most mature aspects of modern 2.4.12 Volcanic Landforms
geoarchaeology. One of the most dramatic
advances in the realm of cave sediment studies Not every geomorphologist considers this type of
involves diagenesis and mineralogical research landform within the scope of basic geomorphic
and their implications for understanding site for- settings. This is somewhat surprising given the
mation processes, human use of caves, and dat- prevalence of volcanism and its impacts on
ing of cave sediments. Their systematic study at humanity over time. Islands are almost always
Kebara and Hayonim Caves during the 1990s volcanic landforms. Continental terrains are
2.4 Geomorphic Setting 41

imprinted with the remains of volcanic events. archaeological/geological significance to cata-

The history of archaeology would have been clysmic eruptions and earthquakes. Pompeii,
much different if it were not for the discovery and its neighbor Herculaneum, have the place
of the lost Roman cities of Pompeii and of honor in the pantheon of archaeologically
Herculaneum in the later 1700s (Corti 1951). In (and historically) documented catastrophes.
the modern world, we are reminded almost annu- From August 24 to August 25, 79 AD,
ally of the force of volcanoes and their disruptive Vesuvius erupted in, what has been named for
impacts on nearby human groups. At the global its chronicler, Pliny the Younger, a Plinian erup-
scale, we can consider volcanism a factor in tion. Hal Sigurdsson (1999) has published a
shaping climate (Bryson 1988). detailed discussion of Pompeii’s end. From an
Volcanic landforms range from the signature archaeological standpoint, Pompeii’s discovery
cone shape to craters, caldera, tabular mountains, in 1748 led to at first unsystematic antiquity
basalt flows, and ash beds to exposures of collecting, but by 1860, and the appointment of
intrusions of magma such as plutons, dikes, Giuseppe Fiorelli, well-documented excavations
batholiths, sills, cuestas, and pinnacles. There began and have continued to the present.
are floodplains called llanos created by lava Tourists, including Lyell, came to Pompeii both
dams of drainages (Inbar et al. 1994). Volcanic as antiquarians, and, in Lyell’s case, as a geolo-
landforms can, like the volcano itself, be gist. Sigurdsson uses the Latin texts and his own
categorized as active, dormant, or extinct. Iceland analysis of the tephra deposits, in what is an
is both an island and an ongoing display of a archaeological study of the fatal eruptive
variety of volcanic landform. In Western Europe, sequence. Herculaneum was destroyed on the
past eruptions of Iceland’s volcanoes have acted first of the 2-day eruption.
as “clocks” with widespread tephra deposits¼dY_
blown from these sources. Geoarchaeologists 3ggKg0Bc
can map these deposits and gain an understanding Thera, the volcanic island, also known as
of the timing of human occupation, a location far Santorini, buried the Minoan town of Akrotiri.
remote from the eruption site. Deposition pro- Like Pompeii, Akrotiri was buried under several
cesses associated with volcanoes are among the meters of volcanic ash. It was discovered in
most dramatic in nature. After deposition, these 1967. The eruption of Thera occurred along the
volcanic deposits are exposed to erosional pro- Hellenic Island arc, a back-arc volcanic region
cesses common to other geomorphic settings. formed in Miocene-Pliocene times (Hsű 1995).
The following statement from Luciana and Wrapped up in the eruption of Thera, a
Tiziano Mannoni (1984) provides adequate justi- Krakatoan-type blast has also been alternately
fication for archaeological geology’s interest in blamed for the eclipse of Minoan civilization as
volcanoes and tectonic processes: “The move- well as being both the inspiration and plupical
ment of the plates and continental platform location for the Atlantis Myth.
within the covering crust causes volcanic As to the former charge, Thera has been found
eruptions and earthquakes. Crustal subduction, not guilty by both archaeology and earth science.
orogenesis, and the metamorphism associated Spyridon Marinatos, first excavator of Akrotiri,
with crystal deformation, under great pressure speculated on the timing of the Theran eruption
and temperature, at plate boundaries create being coincident with the archaeologically
earthquakes and the formation of volcanoes.” observed destruction of Minoan Palace culture
All of this is subsumed under the rubic of the on Crete 100 km to the south of Thera. Placing
“Wilsonian Cycle of plate displacements” (Hsű the eruption at around time of the Late Minoan
1995; Wilson 1966). One has to go no further 1B period (ca. 1450 BC) helped solve questions
than the names of Pompeii, Troy, Thera, Port of the demise of high Minoan culture but posed
Royal, Jericho, and Ceren to assign landmark other problems for later archaeologists at
42 2 The Geomorphological and Geological Context

Fig. 2.19 Tsunamigenic

deposits for the Santorini
eruption or eruptions, 17th
c. BC, at East Beach cliff,
Crete. The white line
indicates the erosional
contact, termed the basal
unconformity, between the
chaotic archaeological
debris layer and the
underlying geological
strata. Wall remains and
dislocated building stones
are present as reworked
material in the lower part of
the archaeological layer

Akrotiri. No Late Minoan 1B ceramics have been the result of fault movements, both alter
found there, only earlier Late Minoan 1A period landscapes—natural and cultural. Amos Nur,
wares (ca. 1500 BC). Stanford geophysicist, has investigated
Based on this discrepancy, many earthquakes in the Holy Land (Nur and Ron
archaeologists see no direct connection between 1997). Earthquakes are the second consequence
the Thera eruption and Cretan problems caused of the tectonism—volcanoes being the other.
by tsunamis and ash falls. The recent publication Indeed, where there is volcanism, there will be
of geochemical data from the Greenland ice sheet earthquakes, but the opposite is not necessarily
cores (Fiedel et al. 1995) indicate a sharp so. In the collision of continental plates, volca-
increase in volcanically derived concentrations nism or magmatic arcs are not present or occur
of NO3 and SO42 in the atmosphere in the only early on in the process. The collision need
interval between 1646 and 1642 BC. Coupled not be “head on,” a convergent boundary, but a
with dendrochronological data for the same “side-swipe” type of collision, called a transform
period, 1628–1626 BC, a major volcanic erup- fault boundary. The boundary between the
tion occurred, but was it Thera? The evidence Pacific and North American plates is a transform
argues against a fifteenth-century BC eruption boundary as in that between the Mediterranean
and suggests an early mid-seventeenth century and Arabian Plates. Both are the seismically
BC for the Theran eruption (Bruins et al. 2008). active zones particularly in California and the
Recent geoarchaeological studies by Bruins Jordan-Dead Sea Transform fault. Along the
et al. (2008) have produced definitive evidence Dead Sea-Sea of Galilee axis, Nur has identified
for a tsunami or tsunamis associated with the a 14 m strike-slip movement in the past
Santorini eruption as shown in Fig. 2.19. 2,000 years. The observed motion of the fault
accounts for only 3–4 m of this amount. Why
the discrepancy? Nur suggests the occurrence of
2.5 Earthquakes: Volcanic or magnitude 8.0 or greater (Richter scale) event
Otherwise (s) in the past.
Using the literary sources, histories, and the
While they can be a direct result of volcanism, Bible, Nur has accounts of major earthquakes in
earthquakes, whether associated with volcanic 370 AD, 31 BC, and 756 BC. Judea, half of the
eruptions, or, as more generally the case, simply dual monarchy of Israel, and later the name of the
2.6 Mapping 43

Roman province, experienced a devastating world, this is where the value of GIS is most
earthquake mentioned in Josephus as the same apparent. Once the geographic information is in
time as the Battle of Actium (31 BC). The digital form, GIS programs can be used to make a
370 AD quake destroyed the Roman city of base geomorphic map (the primary surface in
Beth Shean and is mentioned by Ammianus GIS parlance) and then as series of subsequent
Marcellinus (Nur and Ron, supra, 55). This maps (secondary surfaces) showing elevation
event was recorded in Cyprus, 360 km to the data, stream locations, soil types, landform
southwest. Zachariah describes a major earth- types, and landform units such as alluvial fans,
quake, along with the classic strike-and-dip colluvial deposits, terraces, etc. By breaking up
movement, at Jerusalem in 756 BC. The stron- the geomorphic data into multiple displays, one
gest candidate for earthquake destruction is can avoid the problem of complicated and
ancient Jericho where “the walls come tumbling difficult-to-interpret maps. In the United States
down.” Jericho is no stranger to earthquakes, and many other developed countries, it is possi-
being struck in July 11, 1927 by a major event ble to obtain geographic and geological data
that destroyed major portions of Jerusalem, kill- encoded into computer-compatible form. Again,
ing hundreds. in the United States, a primary source would be
the US Geological Survey (USGS). Digital ter-
rain maps are available from that agency, while
2.6 Mapping coastal maps are available from the National
Oceanic and Atmospheric Administration
Identification of particular landform is typically (NOAA). Digitized topographic maps now exist
done by the use of (a) planimetric maps, at a scale commonly used in archaeology—
(b) aerial or satellite imagery, and (c) ground 1:24,000—that of the 7.5 min quadrangle map.
reconnaissance. The objective is to map the prin- Many countries have similar maps available. The
cipal geomorphic landform, from remotely GIS base maps of Mexico, for example, are
sensed data onto maps of scales appropriate to exemplary.
the archaeological questions being asked. Broad Much of the discussion involving geomorphic
geomorphic features, at local and regional scales, mapping is appropriate to that of geological
can be deduced from aerial/satellite images. In mapping. The use of aerial and other remotely
regions with extensive vegetative cover, conven- sensed data is important to the mapping of sur-
tional visible spectrum photographs are not as face rock units. The geological map thus shows
effective for evaluating landforms as those the distribution of rock units as they appear on
using infrared and microwave regions of the the surface of the Earth. Like the geomorphic
electromagnetic spectrum. These latter types of map, the data is overlain on existing maps at a
imaging can aid in landform classification by scale chosen by the investigator. The two types
differing levels and types of vegetative cover, of maps are similar, but the geological map uses
surface hydrology, and, with newer types of a suite of symbols that are commonly agreed
high-resolution imaging spectrometers, mineral- upon by geologists. Not only does the map
ogy and rock types can be examined. Ground- show rock types, but it also shows elevation,
level studies include walkover surveys and sub- strike, and dip of the geological structure. Strike
surface studies using both hand and mechanical or line of strike is the direction or trend of a
means. formation relative to north or south. Dip is the
From these images and ground investigations, angle between the horizontal and the tilt of the
the geomorphic map should describe formation. On geological maps, these appear as
(a) morphology, (b) landform, (c) drainage t-shaped symbols ({) where long lines indicate
patterns, (d) surficial deposits, and (e) tectonic strike and short lines indicate dip. A number can
features together with geomorphic/geological be shown by the symbol to indicate the angle of

processes active in the mapped area. In today’s dip (e.g., 25 ). As all maps must be oriented to the
44 2 The Geomorphological and Geological Context

Fig. 2.20 Block diagram, with strata and folding Allard and Whitney, Environmental Geology Lab Man-
illustrated (a) and geological map (b) further illustrating ual, Wm. C. Brown, Fig. 4 (1994))
synclinal, anticlinal folding, strike-dip directions (From

north, the angle of strike is generally deduced Topographic profiles, cross-sectional

from the map itself. illustrations, or sections can be produced of
Foliations are indicated by teethlike symbols, outcrops of formations to show stratigraphy in
while multiples of this symbol indicates a reverse the mapped area. These are not maps in the true
fault and lines with adjacent directional arrows sense but can be used in conjunction with the
illustrate slip-strike faults (Fig. 2.20). Chrono- surface geological map to gain a better under-
logical designations of the rock units are given standing of the rock units within the area. Struc-
by abbreviations such as “Q” for quaternary, “T” tural features such as folds—synclines and
for tertiary, etc. Color-coding is often used to anticlines, faults and unconformities—can be
good effect in geological maps. Contour lines shown in section across the geologically mapped
allow the inference of three-dimensional area. The bedding of formations—thicknesses
arrangements of rocks and sediments. and angularity—is also presented in this format.
2.7 Map Scale 45

The profile is indicated on the geological map as Contacts of rock units with a dip greater than
a straight line such as “AB.” On the profile, the the stream’s slope or gradient “V” down-
horizontal maintains the scale of the surface map, stream, while those with a dip less than the
but the vertical scale is more often than not gradient, “V” upstream.
exaggerated to several times than that of the 15. Vertical beds do not “V” or migrate with
horizontal scale. The vertical exaggeration is a erosion.
ratio of the horizontal scale divided by the verti- 16. Contacts migrate down-dip with erosion.
cal scale such as a scale of a topographic map of 17. Fault blocks erode according to their up- or
1:24,000 (2,000 ft to the inch) divided by a scale downthrown aspect—upthrown eroding
of 200 ft to the inch yielding a ratio of 2000/20 or faster than downthrown fault blocks.
a 10 vertical exaggeration. 18. Contacts between horizontal rock units are
Simple rules to follow involving topographic parallel to the topographic contours along
and geological maps include: those contacts.
19. The top of a map is always north.
1. All points on any given contour have the
same elevation.
2. Contours of different elevation never inter-
sect each other. 2.7 Map Scale
3. Contour lines never split.
4. Contour line spacing—close versus wide— Generally, geomorphology, in the service of
indicates steep versus gentle slopes; cliffs archaeology, is not practiced at the global scale.
are indicated by closely spaced or seemingly We are concerned, principally, with topography
coincident lines. and its control of surface hydrology and past
5. Contour lines close on themselves, either human ecosystems. Scale is important in the
within or without the confines of a map. conduct of geomorphic and geological mapping.
6. A contour that closes on itself within a map In the remote sensing of topography, from aerial-
indicates a hill. and space-based platforms, the National Aero-
7. A contour that closes on itself, with nautics and Space Administration (NASA) has
hachures, indicates a depression—the recognized three scales of interest: global scale,
hachures point inward. 1 km horizontal, 10 m vertical resolution; inter-
8. Contour lines curve up or “inward”— mediate scale, 100 m horizontal, 1 m vertical
pointing “up” a valley. The lines cross resolution; and local scale, 10 m horizontal,
streams at right angles. 0.1 m vertical resolution. These scales apply to
9. Bold or darker contour lines reflect simple both visual and multispectral imaging.
multiples of one another. Some maps indi- Their usefulness in geomorphic mapping is
cate a multiple of 5, 25, etc. This is termed fairly obvious. When mapping we refer to high
the contour interval (Fig. 2.20). resolution as small scale. In terms of map scale,
10. Anticlines have their oldest beds or rock this nomenclature is confusing to many. “Large-
units in the center. scale” maps of 1:150,000 to over 1:1,000,000 are
11. Synclines have their youngest beds in the “low resolution.” High resolution, at the local
center. scale, is represented by ratios of 1:20,000 or
12. Anticlines incline (“plunge”) toward the smaller. Observations made at different scales
closed end of a geological structure. are useful depending, of course, on the types of
13. Synclines incline toward the open end of the questions being asked. One must be careful, as
geological structure. Stein cautions (1993), in geoarchaeology, to
14. Monoclinal folds incline in only direction. address scale from the human perspective. This
Terraces act as monoclines. is to say that in archaeology, landscape
46 2 The Geomorphological and Geological Context

reconstructions must be focussed on their rele- thermal infrared band which discriminates
vance to human cultures, their processes, and between land and water; band 7 is the near-
consequent stability or change over time. In this infrared spectral band which penetrates haze,
regard, geoarchaeology is differentiated from tra- emphasizing water and boundaries in landform.
ditional geomorphology and other geosciences Landsat 3 did not have bands 5–7. These were
by its use of earth science methods to make incorporated into Landsat 4. It did have a Multi-
interpretations about human groups interacting Spectral Scanner (MSS).
with landscape, climate, vegetation, and hydrol- Landsat 4 was the first satellite in the Landsat
ogy (supra). In mapping archaeology onto topog- program to incorporate the Thematic Mapper
raphy, it is done more often at local or (TM) sensor. The Landsat 4 TM sensor is able
intermediate scales. to gather seven bands of data as opposed to the
four bands of data collected from the previous
Multispectral Scanner (MSS). In addition to hav-
2.8 Data Sources for Mapping ing three more bands of data to work with,
scientists are able to view the TM data at
If one starts with satellite imagery such as that of a much higher resolution than with MSS. Bands
Landsat—land + satellite—(hereafter “Landsat”), 1–5 and 7 each have a spatial resolution of
3/4, then the scales of this format varied from 30 m, while the MSS is only offered in 79 m
1:250,000 to 1:1,000,000. At the larger scale, the and 82 m resolutions. Band 6 (which is a thermal
images are 185 km on a side with data on several infrared band) has only a maximum spatial reso-
spectral bands that are available upon request. The lution of 120 m. The present-day Landsat
most commonly used bands for geomorphic 8 provides nine bands of spectral data (including
mapping are the visible and infrared bands 4 to panochromatic), seven of which are ETM
7 with composites available. Band 4 is the green (Enhanced Thematic Mapper) bands, with an
spectral band and is good for delineating areas of image resolution or pixel size of 30 m
water in coastal areas; band 5 is the red band and (Fig. 2.21). The panochromatic images have
is good for the vegetation mapping; band 6 is the 15 m pixel resolution. This is an improvement

Fig. 2.21 The first Landsat 8 image. Fort Collins, Colorado (Courtesy NASA/USGS)
2.8 Data Sources for Mapping 47

of original Landsat models whose resolution was Again this is due to the scale of interest. Features
60 m. Landsat 8 has two long wavelength infra- of geomorphic interest are not necessarily those
red bands (10, 11) with 100 m resolution. of archaeological interest, and this must be con-
Commercial satellite imaging by the IKONOS stantly kept in mind when selecting remote sens-
and SPOT satellite services, launched in the ing data for mapping studies. Archaeological
1990s, regularly produces 1 m pixel visual band sites, themselves, are rarely detected from satel-
images. The original image cost was about lite images (McMorrow 1995), so site scale stud-
$1,000/negative, but digitization has driven ies should rely on the advantages of aerial images
those costs down significantly. This availability of lower scale, higher resolution.
of high-resolution, online imagery has been a Some cautionary notes on using remote sens-
boon to archaeo-geomorphic mapping efforts. ing imagery as a substitute for planimetric maps
The High Resolution Infrared Spectrometer should be made here. When using these media
(HRIS) on Landsat 8 has a continuous sampling either to create a geomorphic/geological map or
range from 0.4 to 2.5 μm region with a capability augment a planimetric map, there are inherent
of 10 nm spectral sampling intervals, thus problems both in spatial perspective and eleva-
obviating the “band” restrictions of lower spatial tion. If an image has not been acquired at the
resolution systems such as Landsat MSS and TM truly vertical, then the image lacks spatial fidelity
as well as the French SPOT systems. The contin- across its height and breadth. The only point on
uous spectral capability of the HRIS gives the the image of spatial accuracy is directly at the
researcher the ability to assess sediments, rock nadir or point directly below the lens. All other
types, and minerals in exposed terrain (EOS areas of the image are spatially distorted with
1987:2–3). For example, the spectral reflectance respect to their true positions. That is to say, the
of sedimentary rocks, which constitute about ground distance between geomorphic features is
75 % of the Earth’s surface (Pettijohn 1975), not in a true proportion with the photo distance
varies to a degree that important facies such as between those features no matter what scale is
calcite, which absorbs in the short wavelength used. Additionally, the height of features is
infrared, can be readily identified along with distorted such that relief is exaggerated except
limestones, shales, and sandstones. Using in the flattest of terrains.
Landsat ETM bands 4, 5, and 7, Carr and Turner When photomosaics are made from a collec-
(1996) have demonstrated different spectral tion of vertical images, the same problems still
profiles for rock, soil, and weathered bedrock in ensue. Only the area around the nadir or principal
western Montana. point will have reliable spatial relationships.
Both aerial photographs and multispectral Those toward the edges of the images are the
data commonly had scales of 1:80,000 down to most distorted. This distortion can be removed
1:2,000. In practice, these can be co-registered by the process of rectification where all points on
with satellite data at comparable scales or simply the image plane are corrected to the vertical.
used singularly. Aerial images are of three Vertical exaggeration is harder to remove and
types—vertical, oblique, and mosaic. The most depends a great deal on the focal length of the
commonly used are the vertical images which lens making the image. Longer focal length lens
when co-registered with adjacent images form produces less vertical exaggeration as a rule.
the mosaic type. Both are used in geomorphic When both image distortion—spatial and verti-
mapping depending upon the scale of interest. cal—have been removed, the image is known as
Oblique aerial photographs provide wide-angle, an orthophotograph or ortho-image. Mosaics
perspective view of terrain and assist in deter- made of these are truly picture maps and can
mining low-relief features in relatively flat ter- have contour lines overprinted on them. When
rain. In rough terrain, the value of oblique images these mosaics are overlain on planemetric maps,
is reduced. For the detection of archaeological the best of both worlds is achieved in the creation
features, the aerial image is still unsurpassed. of the orthophotoquad map.
48 2 The Geomorphological and Geological Context

2.9 LiDAR triangulation, or interferometry), white-light

digitizers, and any other technique that scans an
In the most recent past, primary data for mapping area and returns x, y, z coordinates for multiple
was obtained from a variety of sources such as discrete points (commonly called “point clouds”).
traditional maps to satellite images. In the late Photos can clearly define the edges of
twentieth and early twenty-first century, a quiet buildings and other features when the point
revolution has occurred involving the generation cloud footprint cannot. A 3D visualization can
of large digital data sets using airborne and ter- be created by geo-referencing the aerial photos
restrial LiDAR Systems (Light Detection And and LiDAR data in the same reference frame;
Ranging) for mapping. “DEMs” or digital eleva- ortho-rectifying the aerial photos, as discussed
tion models have become the “standard” for above; and then draping the ortho-images on
intermediate-global scale mapping. At local top of the LiDAR grid. It is also possible to create
scale, terrestrial LiDAR (TLS) has proven to be digital terrain models and thus 3D visualizations
more and more the method of choice for using pairs (or multiples) of aerial photographs
archaeology. or satellite imagery, transformed through a cam-
In terms of functional use, LiDAR provides era model program such as 123D Catch
users with detailed terrain models which, after (AUTODESK) to produce a dense array of x, y,
digital processing, can yield bare-earth, vegeta- z data which can be used to produce digital
tion, features, and structures that are either natu- terrain model (DEM) and ortho-image products.
ral or anthropogenic in origin. Both airborne and
terrestrial laser scanners provide the LiDAR data
sets. Both can be expensive but costs have fallen 2.10 Structure from Motion (SfM)
dramatically over the past decade with increased
use of this technology. Likewise the size of the Structure from Motion (SfM) has increasingly
scanners has been reduced as well, making the become a low-cost, easy method of accomplishing
technology more useful on a host of smaller much of what, previously, had required imagery
platforms ranging from backpack to small aerial obtained from satellite and rotary/fixed-wing
drones. platforms. Using drone-mounted and handheld
Photogrammetry is long been used in different cameras, high-resolution images can be quickly
fields, such as topographic mapping, architecture, obtained and processed using open-source or
engineering, manufacturing, quality control, licensed software programs. The results are strik-
police investigation, and geology, as well as by ingly comparable to those from LiDAR, for
archaeologists to quickly produce plans of large instance, digital-surface, hill-shade renderings
or complex sites. Photogrammetric data combined and digital elevation models (DEM). (Fig. 2.22).
with the dense range data from LiDAR scanners Figure 2.23 illustrates a simple application of SfM
complement each other. Photos can clearly define on an archaeological excavation. The upper image
the edges of buildings when the point cloud foot- shows the various camera positions and the
print cannot (Fig. 2.22). It is beneficial to incorpo- sequence of photographs taken to render the iso-
rate the advantages of both systems and integrate metric view. The lower image is the digital-surface
them to create a useful product for rendering of the same view. Compare this with the
geoarchaeology. Photogrammetrically derived LiDAR views in Fig. 2.22. Both SfM renderings
LiDAR combines the more accurate in the x and are scalable and capable of 3D rotation. Two dif-
y directions of photographs with LiDAR range ferent softwares were used in this example. The
data which are generally more accurate in the first is by AUTODESK and available at no cost. It
z direction. This range data can be supplied by is known as 123D Catch. The second software
LiDAR, laser scanners (using time of flight, program used is Smart 3D offered by Acute3D.
2.11 Drones 49

Fig. 2.22 Three views of a historic slave cabin, view is a photograph of the cabin with the LiDAR unit
Wormsloe Plantation, Savannah, Georgia, USA. The shown (LiDAR and photo images courtesy Geospatial
upper two views illustrate LiDAR views, and the lower Mapping Center, the University of Georgia)

Luhmann et al. (2013) provide an overview of the scanning would make photogrammetry obsolete
diversity of close-range photogrammetry and 3D in archaeological studies. However, digital
imaging available. image matching and image-based modeling, as
illustrated in Figs. 2.20 and 2.21, have become
alternatives to laser scanners. The use of photo-
2.11 Drones grammetry in conjunction with unmanned aerial
vehicles (aka “drones”) offers effective low-cost
The utilization of unmanned, powered aircraft tools for the generation of 3D models and ortho-
has flourished in the early twenty-first-century images using image-based modeling techniques
beginning, first, with military uses. These without the use of laser scanners. This new flexi-
vehicles are either remotely piloted or autono- bility and access to both open-access and propri-
mous (Figs. 2.23 and 2.24). etary software make the drone a useful tool for
As Fernández-Hernandez et al. (2015) have the imaging and modeling of complex archaeo-
pointed out, even in the early years of the logical sites. Using techniques taken for DEM
twenty-first century, it was thought that laser methodology and presentation formats such as
50 2 The Geomorphological and Geological Context

Fig. 2.23 Illustrating

SfM. The upper view of an
excavation was produced
using the program 123D
Catch. The lower view is a
digital-surface view of the
same unit using another
program, SmartD. Both
views are scalable and
rotable (Images courtesy
Geospatial Mapping
Center, the University of

“hillshade,” photographs taken by low-cost commonly downloaded from Google Earth. Dig-
drones offer a rich and broader utility for the ital photographic images of the area of an archae-
larger geoarchaeological community as a whole ological site can be reconciled to planimetric
when compared to the LiDAR and other more maps of that same location through the use of a
costly and equipment-intensive approaches recent creation termed the Geo.pdf (Fig. 2.25).
(cf. discussion on SfM). The first concerns were of scale and geodetic
position. These two parameters often differ sig-
nificantly, in pre-digital, “Google Earth” days
2.12 Making the Map: ArcGIS such that they had to be ratioed where the two
to Google Earth formats were in a realistic numerical proportion.
Another concern was when the maps and the
In the digital age, a geomorphic map or geologi- images were made, e.g., the date of their produc-
cal map can be produced with media most tion. Photograph and other imagery differ
2.12 Making the Map: ArcGIS to Google Earth 51

Fig. 2.24 A small, commercial drone (eBee by is the result of this particular flight (Photographs cour-
SenseFly) being preflight programmed for a photo- tesy of Chet Walker, shown programming the eBee
graphic survey. The photomosaic shown on the right drone)

Fig. 2.25 A Geo.pdf with the capability for embedding geographic coordinate information. The red circles are
location of selected test locations across the survey area (Image courtesy of ARI, Tallahassee, Florida)

significantly from maps in their dates. A map is, handheld receivers, one can quickly map the
therefore, a synthetic creation that generalizes geomorphic and geological features.
the temporal frame with that of the spatial or Mobile phone applications have also been
geodetic. Working with digitally based media developed for as GPS, database management,
such as the Geo.pdf (Fig. 2.25) allowed us to and GIS use. Examples of such applications are
utilize another recently developed of the space ArcheoSurvey (
age—Global Positioning Satellites or GPS—for php/field-methods-and-techniques-mz) and
location and creating reference points within the Polaris Navigation GPS (http://www.appsapk.
map frame (red dots in Fig. 2.25). Using com/polaris-navigation-gps/).
52 2 The Geomorphological and Geological Context

2.13 Other Types of Maps when core data is available, a profile representa-
tion is done with the stratigraphic data overlain at
Paleogeographic—These maps are typically specific sounding. When doing an isopach map,
regional or broader in scale, showing the distri- the upper surface is used to portray the strati-
bution of ancient land and sea. In Europe and graphic unit. Isopach maps are very useful in
North America, the most common paleogeo- determining the shape, direction, and depth of
graphic maps are, perhaps, those showing the basin-like features and buried channels.
extent of the Pleistocene ice sheets relative to Lithofacies—A lithofacies is defined as that
unglaciated terrain. Another commonly seen rep- aspect or attribute of a sedimentary unit that is a
resentation is past sea levels and coastlines in the direct consequence of a particular depositional
Pleistocene. Unlike paelogeographic maps from environment or process. An obvious example is a
more ancient epochs, which rely on the plotting fluvially deposited unit wherein the grain shape,
of rock units that are time synchronous, paleo- mineralogy, surface texture, size, etc., of the
geographic maps of archaeological interest are sediment reflect stream transport. Another way
based on more geologically recent mappable sur- to visualize a lithofacies map is to think of it as
ficial features such as moraines and strandlines. an isopach map without contouring. Sometimes
Isopach—These maps are used to illustrate archaeologists will use a variant of the lithofacies
the thickness of particular strata. Unlike profile map in showing the association of archaeological
illustrations, the stratigraphic unit is mapped and sites with specific landforms or soils. In the latter
has contoured some selected interval. The ice case, Kuper et al. (1977) correlate Bandkeramik
sheet thicknesses illustrate a typical isopach rep- (LBK) sites with loessic soils in northwestern
resentation. Isopach maps generally are done for Germany, west of the Rhine River valley. The
buried strata and, hence, are based on core data. lithofacies map is a qualitative representation of
The use of isopach maps is infrequent in the surficial and subsurface strata known from
presentation of archaeological data. As a rule, isolated soundings (Fig. 2.26). One can assume

Fig. 2.26 Linear band

ceramic (LBK) sites (black
dots) plotted relative to
loess soils (gray) in
northwestern Germany
(Adapted from Kuper
et al. 1977)
2.13 Other Types of Maps 53

the distribution of some strata will not be demonstrated, then the lithofacies mapping of
uniform or continuous. This is true for most the unit, based on pedological/sedimentological
intra-site mapping and limits the applicability features, assists in bounding the site for a spe-
of the technique. Where soundings are of suffi- cific occupational period. This method works
cient density, the use of lithofacies maps has very well in shallow (<1 m) depth sites that
merit. Where a clear association of archaeolog- can be quickly sampled with hand coring equip-
ical material, with a stratigraphic unit, is ment (see Chap. 4).
Sediments, Soils, and Stratigraphy
in Archaeological Geology 3

3.1 Introduction important dimensions of the geological study of

archaeological sites (Benedetti et al. 2011). The
Since the principal focus of this chapter is sediment and soil context of the archaeological
the description of sediments and soils and their site is the essential matrix for the material
use in archaeological geology, a programmatic remains of archaeological interest. An adequate
approach to their study is used in modern scientific description of this matrix is basic to the
geoarchaeology. Many of the methods, discussed study of any site. Without it, artifacts and
in this chapter, and their use in both terrestrial and features lose much of their temporal and environ-
marine/lacustrine environments will be within mental points of reference. The pioneering
their specific applications to paleoenvironments. Russian soil scientist, V.V. Dokuchaev, at the
For the assessment of subsurface deposits and end of the nineteenth century, recognized that
their spatial/stratigraphic extent, these techniques soils resulted from (a) the climate of a locality,
provide the most reliable and efficient means of (b) the nature of the parent material, (c) the
doing so. From the geomorphological perspec- amount and character of the vegetation, (d) the
tive, any paleosol, unless found to outcrop in the age of the landscape, and (e) the relief of that
vicinity of interest, is exposed and described landscape (Clarke and Beckett 1971). By a close
using them. Likewise, for ground-truth purposes study of this context of the archaeological site,
of geophysical data, they provide a reliable means one can retrodict many of these important
of assessing depth of features such as subsurface factors. For example, clay minerals are sensitive
contacts and anomalies. These field techniques to most weathering environments and reflect
allow the investigator to recover representative changes in those environments—rainfall, tem-
and depth-constrained sediment, soil, and rock perature, etc.—when measured in soils. Likewise
samples for purposes of laboratory studies. these minerals are indicative, in many cases, of
The study of sediments is the branch of geol- the parent lithic substrate (Conklin 2013:47–48;
ogy termed sedimentology. Sediment beds accu- Rye and Holland 1998). Additionally the particle
mulate on top of another, whereas soils form in size of these minerals can help the researcher to
place. The study of soils is the field of pedology. determine the past depositional environment
A “ped” is the smallest, recognizable element of such as that of a backwater lake (finer clays/
soil structure. The solum is defined as the upper- silts) versus that of a fluvial nature (coarser
most part of the Earth’s surface and is composed silts/clays as well as sands).
of the O, A, E, and B horizons (Richter and Much of the modern study of sediments and
Yaalon 2012). Both sediments and soils are soils is done in the laboratory and we shall

# Springer-Verlag Berlin Heidelberg 2016 55

E. Garrison, Techniques in Archaeological Geology, Natural Science in Archaeology,
DOI 10.1007/978-3-319-30232-4_3
56 3 Sediments, Soils, and Stratigraphy in Archaeological Geology

discuss several of these techniques in Chap. 4. Sediments and soils result from the chemical
Nonetheless much information can be gained in and mechanical weathering of rocks and their
the field. The description of a sediment or soil minerals. The breakdown of the lithic material
profile is typically done in the field as many and its subsequent redeposition is part of what
elements of the fundamental nature of those Stein (1988) terms the “history of a sediment.” A
deposits would be lost even with the most careful sediment’s history is a function of four factors:
of sampling for later laboratory description. (1) source, (2) transport, (3) depositional envi-
Beyond the macromorphological information ronment, and (4) postdepositional environment.
obtained from profile studies, the collection of A study of the physical and chemical properties
subsamples for later study is essential. How one of a sediment or soil can evaluate the nature of
goes about this involves standardized and well- these factors. By using the principle of uniformi-
known tools and procedures taught at college- tarianism (Hutton 1788)—the soil-forming pro-
level courses of soil science, sedimentology, cesses active today are those of the past—and
and geomorphology. Still it is important to reit- those of sedimentation, one can compare a sedi-
erate the description of the use of these tools and ment or soil of a known history and infer the
techniques as this generally omitted by journal processes responsible for the properties and
editors as too tutorial in nature. characteristics of those under study. In archaeo-
logical geological studies of sediments and soils,
the processes involved are generally both natural
3.2 A Brief Review of Sediments and anthropogenic (Butzer 1982). Again, by
and Soils examining archaeological sediments, and com-
paring them to reference examples from other
Geologists usually define soil as a layered mixture well-studied sites, one can infer the specific pro-
of mineral and organic components that are phys- cesses involved in their formation (Schiffer
ical and or compositionally different from an 1987). From the standpoint of geomorphology,
original, unweathered parent material (Smith there are two basic processes—erosion and depo-
and Pun 2006). To agriculturally trained soil sition—which determine the location and variety
scientists, sediments differ from soils in the of sediment and soil:
capacity of the latter to support plant growth.
As we shall see in the following chapter, a
Many definitions of these materials focus on the sediment’s source can be evaluated by examining
organic content of a soil as compared to that of a its textural makeup – grain-size, grain morphology
sediment. Soil ultimately originated in a rock and surface features, and composition (mineral-
ogy). These parameters give insights into the trans-
substrate referred to as bedrock. Clastic sediment
port history of the sediment with fluvial sediments
is a solid fragmental material that originates from demonstrating a well-known rounding due to water
weathering of the bedrock and is transported or transport. Aeolian sediments can be recognized by
deposited by air, water, or ice. The unconsolidated grain shape and by the grain surface “polish” or
glossiness. Size gives clues to the transport mech-
portion of the Earth’s surface that is not soil is the
anism – large, poorly rounded gravels are good
regolith. The regolith is weathered rock and/or indicators of glacial transport in terrains of the
debris that overlies solid bedrock. In many places, northern hemisphere. Small grained sediments are
such as the Piedmont of the Appalachian found in arid environments such as deserts and
drylands as well as fan and fluvial sediments. In
Mountains, it forms a several-meter-deep cover
the Great Basin of the western United States fine-
referred to as saprolite. Soils form on over an grained sediments are common in large slope-wash
extended period of time. Except in the most colluvial fans (Nelson 1992).) Clays from these
exceptional cases, soils require thousands of fans were often used by prehistoric potters
(Speakman et al. 2011).
years to form in most regions of the world such
as the formation of Bt horizons in the Appalachian Depositional features of sediments are deter-
Mountains of the southeastern United States in mined from samples either in situ or recovered en
less than 4,000 years (Layzell et al. 2012). bloc. The structures, of these deposits, include
3.2 A Brief Review of Sediments and Soils 57

the orientation of bedding planes and laminae of recognize that its influence on geomorphology is
various grain sizes and textures. These deposi- complex, the study of climate, in archaeological
tional features are observed in profiles and cores. contexts, begins with the soil and its role as a
The last phase of a sediment’s history is that of proxy for paleoclimate.
postdeposition. In archaeological contexts, zona- In the absence of biologic effects, subaerial
tion develops either due to soil formation or weathering would have essentially involved the
cultural activities such as building, disposal of reaction of rocks with the ambient atmosphere.
materials, and the simple act of living which Subaerially weathered soils sometimes dry out.
produces surfaces that can be identified by their As they dry, desiccation cracks may form on the
micro- and macro-characteristics (Gè et al. 1993; top portion of the profile. If these cracks are filled
Courty et al. 1989; Goldberg and Berna 2010). by sediment and are later consolidated and
Soils: From the perspective of archaeology, compacted, they provide a clear testament to
soils and buried soils—paleosols—perhaps the subaerial origin of the soil.
more so than sediments per se, provide clearly Soils that form in place on a homogeneous
delineated markers of past landscapes and the substrate such as an igneous or high-grade meta-
factors responsible for their formation (Valentine morphic rock tend to exhibit unidirectional
and Dalrymple 1976; Mack et al. 1993: Johnson changes in texture, mineralogy, and bulk chemi-
1998). Archaeological cultures exploited soils cal composition from the unweathered parent
particularly after the advent of plant and animal rock to the top of the soil profile (Holland and
domestication. Beyond being the important sub- Zbinden 1988). At their base, such soils typically
strate for the “agricultural revolution” (Childe grade smoothly into the parent material such as
1951), soils, as paleosols, serve as stratigraphic the ultisol example. The top of the mature soil
horizons as well as proxies for temporal/cultural generally retains few, if any, of the parent rock’s
periods in the archaeological record. It is more textural or mineralogical features. There is usu-
and more common to read of a past cultural ally a continuum of intermediate stages of
facies associated with a particular paleosol, e.g., decomposition of the original minerals and
a paleolithic cultural association with a loess- or textures. As the original minerals of the igneous
tuff-derived soil, and the association of the or metamorphic substrate break down, they are
Bandkeramik culture with light, well-drained replaced by soil minerals. In a mature ultisol,
alluvial soils. These early Neolithic sites are kaolinite, gibbsite, and amorphous silica often
found along the Rhine and Danube drainages— dominate the mineralogy near the top of the
specifically on gravel river terraces overlooking soil. Smectite and vermiculite are common in
medium-sized drainages in areas of loessic soils the lower portion of the profile. Soils often
(Fig. 2.25) (Howell 1987). develop on sedimentary rocks. Most sedimentary
As Dokuchaev (1899) correctly observed, rocks are vertically heterogeneous. Soils that
soils are the result of in situ formative factors. develop on flat-lying clastic sediments and
These factors have been arranged into the acro- other heterogeneous substrates tend not to
nym CLORPT first introduced by Hans Jenny develop textural, mineralogical, and chemical
(1941). Jenny includes the same five state profiles that distinguish them from other clastic
factors—(1) climate, (2) relief or topography, sediments. It is important to recognize that not
(3) organic(s), (4) parent rock, and (5) time as all, if indeed, many, soils are the result of
Dokuchaev—but related them into a quantitative in-place weathering. Surficial processes such as
equation: s ¼ f (cl, o, r, p, t,. .). In the particular colluviation, erosion, redeposition, etc. all con-
case of granite, a common soil formed under tribute to the polygenetic origins of soils (Driese
forests is the ultisol. Using the parameters of et al. 2011).
the CLORPT acronym, climatic geomorphology Soils are described by the use of pedostrati-
is an area of study that has direct relevance to graphic nomenclature. This is in contrast to the
archaeological inquiry. While it is important to nomenclature of lithostratigraphy which follows
58 3 Sediments, Soils, and Stratigraphy in Archaeological Geology

the NACSM code (see below). One must con- within the Hamara were described as “subunits,”
sider both in many descriptions of profiles as the still maintaining a valid priority in sequence
former “overprints” the latter in all cases. Waters terms. The subunits are separated by the use of
(1992) discusses this in detail, noting that soil-sequence boundaries as just described. This
because a paleosol is superimposed onto a usage of sequence-stratigraphic terminology
lithostratigraphic unit (alluvial deposit, dune, differs from its more common application to
ash deposit, etc.), it is always younger than the marine sequences where “sequences” are made
unit on which it develops. Likewise, because the up of parasequences separated by erosional
soil-forming processes alter original particle surfaces. In the marine context, these reflect
sizes and create new pedologic structures, the basin facies changes, whereas in a soil-sequence
original stratigraphic features such as contacts setting, the soils form on the “erosional” surface
are often obscured or otherwise made difficult (s). To use sequence-stratigraphic nomenclature
to interpret (supra). Soil horizons have no exclusively would require that terms like “forma-
relationship to the geological deposition in the tion” be dropped. In a lithostratigraphic context,
lithostratigraphic units on which they occur. The the contacts are the boundaries or erosional
description of soil horizons and sub-horizons will surfaces. Later in this chapter, we shall discuss
be discussed in the next paragraphs. These are the role of the erosional surface for developing
typically done in the field, but sample blocks can allostratigraphic sequences. Again, it must be
and are described in a laboratory setting. Cer- plain that one must take care in the use of these
tainly this is true for “deep” core samples (>1+ nomenclature systems and be very clear in
m) obtained using devices like the Giddings-type describing their use in a particular study.
and other mechanical coring devices Horizonation: A solum or the upper part of
(cf. Chap. 5). the soil horizons—A and B—of a buried soil,
Gvirtzman and Weider (2001) present an termed a paleosol, like a sediment, can give us
excellent discussion of a combined field and lab- information on the five factors that produced
oratory description of a 53 ka eastern Mediterra- it. Paleosols are the grist of most soil studies in
nean soil-sequence stratigraphy. In a section with archaeological geology. Buried soils are charac-
at least six paleosols, these researchers have used teristic of the mid-latitudes and are rare in the
a combination of descriptive protocols, which tropics and arctic. Russian scientists such as
can be used as a model for similar studies. In Dokuchaev and Sibirtsev first described soils as
their study, pedostratigraphic, lithostratigraphic, consisting of zones (supra). The American C. F.
and sequence-stratigraphic nomenclature are Marbut elaborated the Russian scheme giving
combined in full description of the soil sequence. names to differing soil types. Today we continue
For pedostratigraphy, the soil sequence is the use of names for specific types of soils, but
described from the top down, while the strati- the term zone, to distinguish soil profile
graphic sequence is described from the bottom divisions, has been replaced with soil horizon
up (supra). Moreover these researchers use the nomenclature.
terminology, and methodology, of modern The generalized soil profile (Fig. 3.1) consists
sequence stratigraphy as applied to paleosol of the “master” horizons: O, A, B, E, C, and
units such that the sequence boundary (cf. this R. Their formation, over millennia, occurs
chapter) is a surface between the top of an older through the process of horizonation. O horizons
buried soil or sediment and the base of a younger are those layers above the mineral surface and
overlying soil or sediment. consist of decomposed organic material such as
Following the standard practice in lithostra- leaf litter and the like. It, together with the A
tigraphic description, the paleosols were consid- horizon, is also termed the “topsoil.” The A hori-
ered “lower” in stratigraphic hierarchy than zon is the topmost mineral horizon located just
“member” and were described as say a “Hamara below the surface. These horizons generally have
Member” of the “Hefer Formation.” Paleosols a high organic content and have lost iron,
3.2 A Brief Review of Sediments and Soils 59

Fig. 3.1 Soil profile with

commonly found horizons

aluminum, and clays, resulting in a concentration rock or parent material. In undisturbed soils, the
of siliceous material. The E horizon is now gen- R horizon lies conformably with the superposed
erally recognized as a separate horizon although sediments and soils above it.
it is still common to see it considered, by some The relationship of artifacts to soil horizons is
soil scientists, as a leached A or B and even a central to any archaeological geology of a site.
“transitional zone” between the A and B but As reiterated in the upcoming discussion,
having no specific horizon name. E horizons are there is no sufficient phenomenological relation-
those of maximum leaching or eluviation of soil ship between artifacts, their occurrence on sites,
minerals, most notably iron, alumina, and clays. and soils per se. Cremeens and Hart (1995)
The E horizon is bleached or lighter colored provide a good review of this important relation-
in appearance compared to that of adjacent ship in their consideration of chronostratigraphy
horizons. The B horizon is the illuvial layer and pedostratigraphy relative to the archaeolog-
where the translocated mineral complexes from ical context. In that paper, the authors note the
the A to E are concentrated. Depending upon difficulty of using the North American Code of
precipitation, a B horizon will vary in the amount Stratigraphic Nomenclature (NACSN) pedostra-
of redeposited minerals as well as development tigraphic term of “geosol,” which stipulates that
in profile thickness. B horizons are mistakenly a paleosol must be overlain by a lithostra-
referred to as the “subsoil,” by definition, making tigraphic unit—a condition rarely met by Holo-
the O and A the topsoil no matter their composi- cene soils (supra, 15). Likewise, no subdivisions
tion and thickness. The C horizon is sometimes of a geosol are admitted which is a problem for
termed the regolith, but this term, as implied archaeological description to say the least. Adja-
above, is better reserved for unconsolidated cent soil horizons—A, E, B, etc.—are spatially
material overlying rock (Brady 1990). The C discrete but not temporally so. In this they differ,
horizon may appear as rock but it has undergone fundamentally, from the lithostratigraphic
pedogenic processes. Saprolytic soils are good description of sediments. In even Holocene
examples of this with deeply weathered C contexts, paleosols can be of three types—
horizons several meters thick in some regions. (1) relict or a surface, never buried paleosol;
The R horizon is unaltered, consolidated country (2) exhumed or a buried paleosol now subaerially
60 3 Sediments, Soils, and Stratigraphy in Archaeological Geology

exposed; and (3) buried or “classic” paleosol. based on bounding discontinuities or

Paleosols are recognized on the basis of their unconformities; as well as pedostratigraphic
organic, textural, and stratigraphic features differences, based on one or more differentiated
(Collinson 1996). Because of their antiquity, the soil horizons.
increased exposure to diagenetic processes can In this book, the taxonomic level of soil order
make original pedogenic differences between is used for discussion purposes. As noted above,
horizons more pronounced when compared to there is no compelling methodological reason to
analogous modern soils. This is observed in associate a soil with any form or period of archae-
“red bed-type” paleosol successions such as ological site. Archaeological sites can and do
the example given for the “Hamra” soils. The occur on land surfaces that have the various soil
authors of that study note this in their observation orders discussed below. Still there is no particular
that “the older hamra developed (and) enhanced type of cultural site that demands a particular type
the red color and increased clay content of soil. Prairie soils, such as mollisols, could be
(supra).” In the following discussion of soil associated with archaeological sites of an agricul-
chronosequence, this “deepening” of red hues— tural nature, such as in the American Midwest
10YR to 2.5 YR—is synchronous with increas- characterized by late Quaternary lithostra-
ing clay content (Layzell et al., supra). tigraphic units such as the Loveland, Roxana,
In the Gvirtzman and Weiner study, the soil and Peoria formations (Leigh and Knox 1994).
sequence demonstrates a common aspect of These three younger formations date as follows:
paleosols which is an overprinting of a later soil Roxana (55,000–25,000 year B.P.) and Peoria
onto the preceding one. This is referred to as (25,000–12,000 year B.P.) loesses recognized
polygenesis and leads to what are termed “com- elsewhere in the midcontinental United States.
posite” paleosols. In particular the production of But is it, likewise, necessary for a paleosol
red pigments may take away from existing pedo- like the Brady soil (Fig. 2.12) to have formed
genic features (Collinson, supra). When the soil on a widespread paleolandscape like the Peoria
sequence is not so mixed and the separation of Loess on those same prairies? Archaeological
paleosols is such that soil-forming episodes sites ranging from the earliest periods in North
are distinct, then one must examine the nature of America are found on both the Brady Soil and the
the geological processes that “reset” the sequence Peoria Loess (Reagan et al. 1978). An associa-
to primary stages of development. Vitousek tion of a particular form of archaeological site
et al. (1997), in their study of Hawaiian paleosols, with a soil could be called a necessary condition,
suggest that glaciations and volcanic eruptions in terms of logic, but hardly a sufficient condi-
are the most likely to do this. Gvirtzman and tion. Early agriculture and early entrant hunting-
Weiner (supra) observed that coastal aeolian foraging cultures are as easily found on ultisols
deposition was the mechanism at their location. (typically found in formerly forested landscapes)
In the study of the geomorphology of the and on young entisols or inceptisols found on
Quaternary, most workers use Birkeland’s active fluvial landforms.
cross-disciplinary approach (Birkeland 1984). As such the stratigraphy of archaeological
For purposes of archaeology where one may be deposits can be formalized using the North
describing a paleosol or a sediment stratum, American Stratigraphic Code (NACSN 1983) or,
it is important to remember that the objectives perhaps, better still by the International Strati-
of that description may be very different, as graphic Guide (Salvador 1994). Workers such as
mentioned earlier in this section, from that of a Stein (1993) endorse the use of cultural items such
pedologic or edaphic characterization. Archaeo- as artifacts or features to specify ethnostratigraphic
logical sediments can have lithostratigraphic differences in archaeological sediments.
differences, based on texture and composition; Taxonomy: Soils are classified according to
chronostratigraphic differences, based on tempo- systematic taxonomic naming systems. The sys-
ral distinctions; allostratigraphic differences, tem developed in the United States is termed the
3.2 A Brief Review of Sediments and Soils 61

7th Approximation or Soil Taxonomy. This sys- the soil is characteristic of a subgroup, it is
tem was developed by soil scientists of the US described a “typic,” and with the addition of
Department of Agriculture and those of many textural designators “loamy,” “sandy,” etc., then
other countries and was introduced in 1965, in a particular soil could be described as a loamy,
the United States, and has been adapted, to vary- typic agriaquoll. The final level—series—is the
ing degrees, by over 45 other countries. In this least informative as the series name departs
system, the A and B horizons become epipedons from the logical taxonomic pattern using, simply,
or surface horizons, and those horizons below a name such as “Brookston,” “Westland,”
the B are subsurface horizons. There are six “Fayette,” etc. to designate the soil. A buried
categories in the classification system arranged soil or paleosol is indicated by the prefix paleo-
in a descending order of scale—order, suborder, to the great group designation such as a buried
great group, subgroup, family, and series. The aquoll (a wet mollisol) termed a paleoudoll.
most inclusive category is that of order. There In fluvial contexts, multiple paleosol sequences
are only 12 soil orders but there are around are very possible (see “Allostratigraphy”
16,800 soil series recognized in the United States (Sect. 7.1)).
alone. The 12 soil orders in the American soil taxo-
The soil taxonomic classification system was nomic classification system are:
developed to describe and map soils found on the
surface of the Earth today. As Fig. 3.2 illustrates, 1. Entisols: soils without any distinctive layers
this is not the exclusive soil classification or profile development.
scheme. The American system, the FAO, and 2. Inceptisols: soils with poorly developed
the French schemes are the most commonly layers or weak B horizon, e.g., mineral trans-
used (Collinson, supra). The American system location is not present or is poorly advanced.
relies on the organic content of the near-surface 3. Mollisols: soils with a dark A horizon or
layer and the chemical characteristics of the epipedon and that are base rich and have
deeper horizons. The FAO scheme, somewhat strong B horizons, typical of grasslands.
simpler, recognizes basic soil types, predicated 4. Alfisols: soils with a well-developed argillic
on maturity of the soil type and climate. The or natric (high sodium (Na)) B horizons and
French system relies less on climate criteria and are well watered forming under deciduous
more on soil-forming processes. forests as a rule; E horizon is usually present.
Soil taxonomic nomenclature allows a quick 5. Ultisols: soils with well-developed argillic B
and thorough description of the properties of a horizons, but low in base minerals (e.g., acid
soil. The names of the classification units are soils) typical of soils formed under native
simple combinations of syllables that are derived pine forests in humid areas such as the
from Latin or Greek root terms, each of which American South; E horizons can be present.
conveys some key aspect of the nature of the soil 6. Oxisols: soils typical of the tropics with
such as the Latin root aridus for “dry” appears in deep, highly weathered B horizons that are
the order name Aridisol. The Latin root mollis, very reddish in appearance.
meaning “soft,” forms the order term Mollisol. 7. Vertisols: soils that are clay rich but have
The suffix sol is the Latin for “soil” so a mollisol no B horizons and are characterized by an
is a “soft soil” and likewise an aridisol is a “dry annual “shrink-swell” cycle producing large,
(land) soil.” As one descends the naming system, deep cracks in their profiles.
a wet, soft soil is described by the suborder term 8. Aridisols: “dry soils” characterized by ochric
aquoll which is the conjunction of “aqu(a)” and or pale profiles due to lack of organic matter;
“(m)oll.” Note the higher taxonomic designator found in dry areas of the world.
“(m)oll” appears last in the sequence. Continuing 9. Spodosols: soils with acidic B horizons rich
the description of the soil the level of the great in iron/aluminum oxides; A horizons are
group, a clay-rich, wet mollisol is a agriaquoll. If highly leached.
62 3 Sediments, Soils, and Stratigraphy in Archaeological Geology

Fig. 3.2 The Munsell

color system. Using this
diagram, an example soil
color such as “2.5YR5/4”
can be understood as 2.5
yellow-red on the hue
circle, 5 on the vertical
scale for value, and 4 on the
horizontal section for

10. Histosols: soils characterized by high Acrisols: clay rich and is associated with
organic matter (>20 %) originating in bogs, humid, tropical climates (ultisols—all orders)
marshes, and swamps; no B horizons. Alisols: poorly drained soils with a dense subsur-
11. Andisols: soils of volcanic origin with face clay layer (vertisols)
weakly developed or no real profile; Andosols: volcanic soils, containing glass and
dominated by silicate mineral—no clays— amorphous colloidal materials (andisols)
and with humic epipedon; previously Anthrosols: formed or heavily modified by long-
included with inceptisols. term human activity (various orders)
12. Gelisols: soils that have permafrost and many Arenosols: an entisol consisting of unconsoli-
are cryoturbated (frost heaving, solifluction, dated sand deposits (entisols)
etc.). These soils consist of minerals or Calcisols: a soil with a substantial secondary
organic soil materials or both. They com- accumulation of lime (aridisols)
monly have layers of gelic materials and histic Cambisols: horizonation is weak; young soil
(high organic %) or ochric (pale) epipedons. (inceptisols)
Chernozems: black-colored soil containing a
The Food and Agriculture Organization high percentage of humus (boreal mollisols)
(FAO) developed a “World Soil Classification Ferralsols: soils containing at all depths <10 %
System.” weatherable minerals (oxisols)
In the revised FAO system, there are Fluvisols: a genetically young soil in alluvial
28 “World Classes” formed by 106 “Soil Units” deposits (entisols)
(FAO 1988). Gleysols: wetland soils (hydric soil) that
USDA Soil Taxonomy equivalents in develop a characteristic gleyic color
parentheses: (inceptisols)
3.4 The Soil Chronosequence 63

Greyzems: boreal or wetland soils (USDA equiv- same soil type such as an ultisol with only a
alent—mollisols; gelic or aqualls) variation in the horizonation of the down slope
Gypsisols: soils with substantial secondary accu- examples. At the top and upper slope, the
mulation of gypsum (aridisols) horizons may be thicker in comparison with
Histosols: a soil consisting primarily of organic those at mid-slope where erosion has reduced
materials (histisols) the thickness. At the base of the slope the soil,
Kastanozems: rich in humus, formed from early there could have deep horizons similar to those of
maturing native grasslands (mollisols) the top but, in contrast, could have a much differ-
Leptosols: soil over hard rock or highly calcare- ent character due to more minerals moving down-
ous material; lithosols (no USDA equivalent) slope. In many cases, the lower slope soil could be
Lixisols: soil developed under intensive tropical another type altogether such as an inceptisol or
weathering (alfisols) entisol. Slopes without forested cover will develop
Luvisols: clay-rich soils in humid-subhumid for- soil types consistent with edaphic conditions. A
ested areas (ultisols) south-facing slope, well watered, could produce a
Nitisols: deep, red, well-drained soil >30 % clay parkland or upland prairie mollisol.
content; blocky structure (alfisols/ultisols) Bushnell (1943) links the concept of the
Phaeozems: dark soil; high base content, but catena to that of a catenary wherein the
without a calcareous soil horizon (mollisols) differences observed at points along a catenary
Planosols: light-colored, coarse-textured, surface such as a valley slope result from environmental
horizon; clay subsoil (alfisols/mollisols) forces at work along it. He stresses topography as
Plinthosols: iron-rich soils characterized by the a controlling variable. Parent material is another
presence of plinthite (oxisols) key factor.
Podzols: derive from either quartz-rich sands and Bushnell’s catenas are made up of a series of
sandstones (spodsols) types made up of profiles that are characterized
Podzoluvisols: A and E horizons leached of as to number, thickness, arrangement, color, tex-
organic material and soluble minerals ture, structure, composition, etc.—all the
(spodsols) categories discussed in the following section. In
Regosols: very weakly developed mineral soil in the ultisol example, Bushnell would call this a
unconsolidated materials (entisols) “simple catena” wherein the singular soil series
Solonchaks: pale or gray soils found in arid to are homologous in all features except for those
subhumid, poorly drained areas (aidisols/ due to variation in drainage. “Series” is used
mollisols) strictly in the taxonomic sense, wherein a spe-
Solonetz: soils with a so-called natric horizon, in cific soil is so designated.
the upper 100 cm of the profile (alfisols)
Vertisols: a soil in which there is a high content
of expansive clay (see alisol)
3.4 The Soil Chronosequence

Unlike a soil catena, a soil chronosequence is “a

sequence of soils developed on similar parent
3.3 The Soil Catena materials and relief under the influence of con-
stant or ineffectively varying climate and biotic
The strict definition of a catena is a sequence of factors, whose differences can thus be ascribed to
types of soil down a hill slope. These types reflect the lapse. . .of time” (Stevens and Walker 1970).
slope processes such as precipitation, evapora- If the ages of the soils of a chronosequence
tion, erosion, deposition, and parent material that are known, then rates for soil formation as
are key variables that influence soil formation well as variation on soil-forming minerals and
and stability. A catena can consist of simply the conditions can be deduced.
64 3 Sediments, Soils, and Stratigraphy in Archaeological Geology

In a study of a southeastern US soil 3.5 Describing Archaeological

chronosequence of Paleudults, formed on five Sediments and Soils in Profile
fluvial terraces of the Catawba River, NC,
Layzell et al. (2012) estimated terrace ages Strata and Horizons: At a cleaned exposure or
based on elevation from 3 to 42 m above the excavated profile, the first step in description is to
modern river. The oldest terrace (Qt1) was determine the number and kind of strata or
estimated to be 1,470  180 ka, while the horizons present. If the deposit is not a soil, one
youngest terrace Qt5 was dated to 4  0.5 ka. still determines its lithostratigraphic nature. Each
In this study, soil morphological properties such stratum or horizon, whichever the case, should be
as color hue, clay content, and iron mineralogy measured from a surface datum or downward and
showed positive trends in concert with upward from the surface of a mineral horizon
terrace ages. if present. These strata are sketched and
Specifically, color hue increased from 10 YR characterized. If the stratum is a soil, then the
on Qt5 to 2.5 YR on Qt1. Likewise, clay content presence of master soil horizons should be noted.
increased, particularly in the Bt horizon from Arabic numerals, in the master horizon designa-
~25 % on Qt5 to ~63 % on Qt1. West and tion, indicate certain key characteristics such as
Dubos (2013) relate this difference in hue to “O2,” an O horizon with accumulated organic
increasing clay and iron. Maximum clay content materials which in the past (and in certain
values reached an asymptote on terrace Qt2 countries today) were indicated by lowercase
(~610 ka). This compared well with Engel letters such as “Oa.” Lowercase letters are still
(1996) who measured ~53 % clay content in used today in the description of soil profiles such
early Pleistocene soils (~770–970 ka). as the important designation for archaeological
Iron minerals converted over the geologists—the lowercase “b” which indicates a
chronosequence from amorphous hydroxy buried soil or paleosol such as “Bb.” The accu-
forms (goethite, hematite, ilmenite, etc.) to crys- mulation of translocated clays is commonly
talline forms (oxides) (Layzell et al., supra). denoted as a “Bt” horizon. A more complete
West and Dubos (2013) term the Blue Ridge listing of these designators is shown in the fol-
and acidic upland soils as parasequic which lowing table. Master horizon categories and sub-
recognizes soils with larger amounts of the iron ordinate distinctions are made depending upon
oxides such as goethite, hematite, and gibbsite in the detail desired at the field site (Table 3.1).
the clay fraction. Dithionite-extractable pedo- Lithological discontinuities or contacts are
genic iron (crystalline forms) increased, progres- indicated by Roman numerals such as a change
sively, in the chronosequence from 3.63 % on within an A horizon from a loessic sediment
Qt5 to 6.37 % on Qt1, again, like clay content, to a quartzitic sand matrix which would
reaching a maximum value on the Qt2 terrace. designated as “IA” and “IIA.” In the case of
These minerals, together with the color and clay buried soils, the second or lower soil sequence
variables, point to a plateau in their rate of devel- is noted as follows: “A’2, B’2.”
opment peaking at ~128 ka. The authors of this Color: This parameter is important to be
particular chronosequence suggest that the observed and recorded as systematically as pos-
effects of parent material and climate were sible. The method most used today is the Munsell
overshadowed by changes in sedimentation and Soil Color Chart. The color chart is a collection
fluvial processes (supra). These changes of color chips organized in notebook form. The
represented a major shift in the landscape evolu- color chips are arranged according to hue, value,
tion for this region. An example of this shift in and chroma. Hue is the principal spectral color.
rates of soil formation is the rapid formation of The pages of the color chart take their names
clay-rich Bt horizons in as little as 4,000 years. from this attribute such as “2.5YR,” indicating
3.5 Describing Archaeological Sediments and Soils in Profile 65

Table 3.1 Designations of master soil horizons and subordinate symbols for soils (Adapted from Olson 1981)
Master horizons Subordinate symbols
01 Organic undecomposed horizon b Buried horizon
02 Organic decomposed horizon ca Calcium in horizon
A1 Organic accumulation in mineral soil horizon cs Gypsum in horizon
A2 Leached, eluviated horizon cn Concretions in
A3 Transition to B horizon f Frozen horizon
AB Transition to B horizon; like A in upper part g Gleyed horizon
A and B A2 w/ less than 50 % occupied w/ spots of B h Humus in horizon
AC Transition horizon; not dominated by A or C ir Iron-rich horizon
B and A B w/ less than 50 % occupied w/ spots of A2 m Cemented horizon
B B horizon w/ accumulation of clays, iron, cations p Plowed horizon
Humus; residual concentration of clays; coatings (cutans) sa Salt-rich horizon
Or alterations in original material forming clay and structure si Silica-cemented
B1 Transition horizon more like B than A t Clay-rich horizon (Bt)
B2 Maximum expression of B horizon (clay rich) x Fragipan horizon
B3 Transitional horizon to C or R
II, III, IV Lithological discontinuities
C Altered parent material from which A and B horizons are presumed
to have formed
A’2, B’2 Second soil sequence in bisequel soil
R Consolidated or competent bedrock

a variation from gray to black, pale red to very Recording colors using the Munsell system
dusky red, reddish brown to dark reddish brown, allows for a relative standardization of the
and light red to dark red. “5YR” differs in hue reporting of soil/sediment color but it is not an
from white to black, pinkish white to dark reddish infallible tool. Too often the beginning student is
brown, and reddish yellow to yellowish red. not instructed properly in the use of the color
There are ten hues in the Munsell system—the chart and what the three parameters signify.
two already mentioned along with 5R, 7.5R, Light conditions at the time of the measurement
10Y, 7.5YR, 10YR, 2.5Y, 5Y, and gley. On can vary significantly in the field. Measurements
each page, the color chips are arranged biaxially made in the shadows of an excavation trench will
according to value (vertical axis) and chroma differ from those made just outside of the trench.
(horizontal axis). Value refers to the relative Direct sunlight is generally avoided and many
lightness (darkness) of the color, while chroma use indirect or reflected light. Colors read in the
designates the relative purity of the color (red, morning and evening, aside from being qualita-
red-orange; yellow-red, etc.) and its spectral tively darker, will tend to appear more red, e.g.,
strength increasing in numerical value with 10YR is less red than 2.5 YR. Moist and dry
decreasing grayness. Figure 3.2 illustrates the readings are taken of the sample. The simplest
relationships among the three variables of procedure is to take a small sample of the mate-
Munsell color. An example of a Munsell color rial on the end of a trowel and hold it behind the
descriptor for a sediment or soil could be “5YR2/ appropriate page of the color chart notebook.
1.” The descriptor reads: a hue of 5YR, a value of Color matches are seldom exact and the reader
2 and a chroma of 1. This color is black. How- must make the best judgment possible perhaps
ever, a one-unit shift in the chroma to 5YR2/ indicating the quality (poor to good) of the color
2 makes this color dark reddish brown. match. A few archaeologists still eschew the use
66 3 Sediments, Soils, and Stratigraphy in Archaeological Geology

of the Munsell chart, preferring to simply desig- names, showing values and chromas for each
nate the color as “brownish red,” “light tan,” etc. color name (Kelly and Judd 1976).
While his/her system may be readily intelligible Texture: This parameter, also known as
to them individually, it lacks that quality for granulometry, along with color, is the most
others. While two workers may disagree on salient of a sediment or soil. Texture is size and
whether a horizon is 10YR8/3 or 10YR7/4, the proportion of the particles that make up a deposit.
Munsell color designation is still “very pale The four basic categories of size are gravel, sand,
brown” which in another more unstandardized silt, and clay. In the United States, these size
description could be “tan,” “light brown,” or categories are divided as follows: <0.002 mm,
even “buff colored”! Figure 3.2 illustrates the clay; 0.002–0.05 mm, silt; 0.05–2 mm, sand; and
three components of a Munsell color designation. >2.0 mm, gravel. The International Society of
Recently, Ferguson (2014) has examined the Soil Science scheme differs in the size range
use of the Munsell color chart in archaeology. He for silts and sands: 0.002–0.02 mm, silt, and
notes that archaeologists have limited their color 0.02–2.0 mm, sand (cf. Chap. 5). Likewise,
descriptions to those of the Munsell Soil Color other countries vary somewhat in their size
Chart. While giving a fuller discussion of the categories so it is important to note which size
variety of Munsell color charts—Plant Tissue, classification scheme is reported (Fig. 3.1).
Bead Color, and Rock Color charts—he Another size scale still in use in the United States
concluded that, given the familiarity and length is the phi scale (Φ). This classification scheme
of use of color names in the Soil Color Chart, it ranges from  8 Φ (gravel) to + 8 Φ (clay).
would be counterproductive to replace the sys- Texture or particle size analysis is done with
tem (supra). The point of Ferguson’s review is sieves to mechanically separate the gravel and
not to question a relatively standardized use of sand grains into finer and finer divisions. Silt and
colors and color names in the description of clay sizes and amounts are determined by their
archaeological materials, but rather to suggest a settling rates in water.
broader “palette” of recognized colors and color A sandy soil consists of at least 70 % sand and
names for use in archaeology. His point is well 15 % or less clay. A silty soil contains at least
taken as the Soil Color Chart is basically appli- 80 % silt and 12 % or less clay. Clay soils must
cable to just that—soil. Not all archaeological contain 35 % or more clay. In the field, a rough
materials—lithic, ceramic, or organic—fall com- approximation is done by “feel” or examining the
fortably within the range of soil colors. In fact, texture by manipulating a sample of the material
only recently have the “tropical soil colors” between the fingers—sands are gritty to the
pages been added to the Soil Color Chart as an touch, silts are less so, and clays display a char-
extra cost option—now downloadable—mainly acteristic plasticity and stickiness when wet. This
for archaeologists since oxisols do not exist in latter parameter is known as consistence and is
continental North America nor Europe. In Asia, used to describe soils (see below). Visually, sand
the tropics, South America, and portions of particles are readily apparent to the eye. These
Africa, yes—but for descriptive purposes of are described in terms of “rounding” and
earthenwares, in the Americas, these two pages “sorting” relative to grain assemblages (Fig. 3.3).
expand the color palette as Ferguson has Silt, in its finer particles, and clay cannot be
discussed. Those archaeologists who prefer not seen without an electron microscope.
to use the Munsell Soil Color Chart for descrip- Texture or particle size is typically
tive purposes might prefer the use of a broader represented, graphically, in ternary diagrams
color name system as seen in the other Munsell wherein the apices are 100 % sand-clay-silt
charts as well as other color naming systems such (Fig. 3.4).
as the Inter-Society Color Council on Soil Color- Structure: Aggregates or peds of soil, separated
National Bureau of Standards (ISCC-NBS) sys- by planes of weakness, have four principal types:
tem which provides a full range of Munsell (1) spheroidal, (2) platelike or platy, (3) prismatic,
3.5 Describing Archaeological Sediments and Soils in Profile 67

Fig. 3.3 Above, sand grains, roundness, and sphericity (Modified from Powers 1953), below, sorting of sand grains
(Modified from Compton 1962, Manual of Field Geology)

100 100
a 10
b 10
90 90

20 20
80 80



30 30

70 CLAY 70


40 40


60 60






50 50










30 LOAM 70 30
80 LOAM 80
100 90 80 70 60 50 40 30 20 10 100 90 80 70 60 50 40 30 20 10
Percent sand (50m–2 mm) Percent sand (60m–2 mm)

Fig. 3.4 Example ternary diagram plots for, left, the United States and, right, Great Britain and Wales

and (4) blocky. Categories 1 and 2 most often aspect of peds is their size. Like texture, there is
occur in A horizons, while prismatic and blocky a relative scale for the size of peds in a horizon
peds are generally found in B horizons. A key from very fine (<1 mm) to very coarse (>10 mm)
68 3 Sediments, Soils, and Stratigraphy in Archaeological Geology

with spheroidal and platy structures. The class of dominate in alkaline sediments and soils. These
spheroidal structures is subdivided into granular chemical species replace the aluminum and
and crumb. Most sediment grains are not perfectly hydrogen ions available for reactions. The chem-
round so they will most likely fall into one or ical nature of a sediment or soil has a direct
the other of the two subclasses. Blocky structures bearing on the diagenetic effects on archaeolog-
vary in size from very fine (<5 mm) to coarse ical materials such as bone, wood, or other
(20–50 mm). Prismatic structures vary from organic artifacts. For instance, bone preservation
<10 mm to 50–100 mm. As a rule, the age of the is significantly reduced in acidic soils. Precipita-
soil influences the development of structure with tion, the ultimate source of water—hence hydro-
older soils having more well-developed or strong gen and hydroxyl (OH-) ions—in sediments and
peds, and, conversely, young soils have soils, varies and mediates their acidity. Where
weak peds. soil moisture is low, then one should expect a
Consistence: This property of soils is low acidity as well. Ultisols, for example, formed
measured under three conditions—dry, moist, in semitropical and temperate regions should be
and wet. It describes a soil’s resistance to and indeed are acidic in nature.
mechanical stresses and manipulation. Most dry Stoniness: This category was developed by
soils have little coherence and readily fall apart. British workers. The principal parameters looked
Those that have not have been cemented by for are quantity or abundance, lithology, size,
carbonates, iron/aluminum oxides, and silica and shape. A common way of reporting stoniness
minerals other than the clays. Plinthite is an is as percentage of volume. This, in itself, is
example of the cementation of sediments by somewhat misleading as most determinations of
iron that was translocated by groundwater. percentage are done from a profile rather than a
Calcium carbonates and gypsum (CaSO4 2H2O) true volume. If one extracted a volume of soil or
produce cemented petrocalcic and petrogypsic sediment, it is easy to see that the size of the lithic
horizons and are so-named in the master horizon materials has a significant influence on the mea-
nomenclature. This type of cementation is seen surement which is to say that larger stones will
in aridisols. Generally, soil, when moist, has a require larger volumes. A straightforward scale
coherence that varies from loose to extremely for a vertical section’s percentage of rocks is
firm. Wet soil has a consistence that varies from slight stony (<7 %), stony (7–30 %), and very
nonsticky to very plastic as one would expect stony (>30 %) (Hodgson 1978). Again, large
that the more plastic the material, the higher the areas of a soil profile should be used to make
clay content. this determination.
pH or Reactibility: Chemical solutions can As pointed out in Chap. 1, Darwin was the
provide indications of the acidity or alkalinity first to remark on what we call bioturbation and
of a sediment or soil. Differential weathering of its importance to the formation of the upper soil.
minerals translocates cations such as Ca2+, Mg2+, If not offset by what is termed “floralturbation”
K+, H+, Na+, NH4+, and Al3+, creating a great (Leigh 2001; Thoms 1995), e.g., root growth,
variety in the geochemistry of sediments and tree throws, animal burrows, and “stump burns”
soils. The most common chemical parameter (the latter in forest fires), soil biota will bring
tested for in soils is the presence of the hydrogen fine-grained sediments to the surface. This loos-
ion commonly expressed as pH. The acidic ening of sediments, gravity and the infiltration of
(pH <7) or basic (pH >7) character of a sedi- water to allow larger clasts – stones/cobbles – as
ment or soil typically depends on the concentra- well as artifacts – to settle downward in the upper
tion of hydrogen (H+) and aluminum (Al3+) ions soil just as Darwin observed (1896). “Stone
present. The exchange of cations and the relative lines” form at the base of this upper biomantle
amount of base formation is dependent on (Johnson 1990).
pH. In low pH (acidic) situations, cation In the Ozark Mountains of Missouri, USA,
exchange is reduced. Base-forming cations Henry R. Schoolcraft remarked in 1818 on the
3.5 Describing Archaeological Sediments and Soils in Profile 69

general stoniness of the landscape. Schroeder to black—color such as the typical mollisol.
(1967) describes three types of stoniness in the Organic horizons can be evidence of anthropic
eastern Ozark Mountains: (1) stone-free; (2) mod- deposition of organic debris in cultural middens.
erately stony, and (3) extremely stony. The last Mode: More usually used to describe the
category describes a soil mass and surface com- nature or distribution of minerals in a rock, it is
posed of 50–90 % rock fragments, mostly cherts, less commonly used in describing sediments and
that can reach a thickness of 7–15 m (supra). soils. In the context of paleosols, the preservation
Organic Matter: For the soil scientist, this of primary minerals is directly linked to time and
component defines the soil from simple sedi- diagenesis. When we refer to “mature” beach
ment. In the description of master soil horizons, sands, we infer that only the most resistant
it is the key parameter in determining the O minerals, e.g., quartz, have survived to make up
horizon. It is a somewhat transitory element of the textural/mineralogic assemblage. The com-
soils in that its residence time rarely exceeds a mon silicate minerals ranging from quartz,
few hundred years unless it is depositional through the feldspars, to olivine respond very
(anaerobic) environments that preserve it— differently to weathering with the latter species
inundated deposits such as wetlands and subma- the first to disappear, while zircon, structurally
rine landforms. In these environments, microbial similar to olivine, persists longer in the soil
and biochemical activity is suppressed or absent. (Ritter et al. 1995). By and large, the heavy
Organic matter has a great deal to do with soil minerals can be used to gauge the exposure of
moisture. Dark, organic soils tend to indicate the paleosol to normal meteoritic waters,
greater water retention, whereas soils with lesser percolating fluids such as soil acids. Certainly
amounts of organic matter allow water penetra- their presence or absence is a barometer of
tion to deeper depths. The amount of organic the “history” of that paleosol; therefore, their
matter in soils is relatively low, being on the mode should be noted just as we measure the
order of 6 % or less in topsoils and even less in nature of the secondary minerals—clays, hydrous
the subsoil (Brady 1990). Organic matter is a rich oxides—gibbsite, hematite, goethite, etc.
source of ions from the breakdown of plant and Boundaries: Transitions between strata and
animal matter into organic complexes. horizons are characterized as to their distinctness.
Outside of the field setting, organic matter is In describing a soil, its mineralogy, texture, color,
categorized as humus—raw, mild, intimate, or and chemistry all contribute to the determination
mechanically incorporated varieties (Clarke and of horizons. In archaeological settings, these
Beckett 1971). In the Soil Taxonomy, soils attributes, plus those of cultural origin—anthro-
containing appreciable quantities of organic mat- pogenic—help in the delineation of stratigraphic
ter are described as slightly humose, humose, differences. Where archaeological deposits occur
very humose, and organic soil (supra, 46). It is as buried paleosols, one must be aware of the
difficult to assess these categories in the soil likelihood of overprinting by subsequent soil-
horizon. To the touch, organic matter can make forming processes. Likewise many of the changes
a soil feel like silt, but the texture is different—a observed in archaeological sediments may not
silt smears evenly, while an organic soil sample be stratigraphic in nature. This is to say that not
will fragment or fray. every solum in a site profile is necessarily
Histosols are soils rich in organic matter important to its stratigraphic description. Archae-
(>20 %). Peats, by definition, are histosols but ologically interesting levels, of like chronostra-
are more closely classified by British workers tigraphy, may vary significantly in textural
into three classes—pseudo-fibrous, fibrous, and color and chemical characteristics. In general,
amorphous (supra, 72–73). Organic matter that is stratigraphic differences are denoted by contacts,
incorporated into the soil mass is termed intimate which in turn are most usually determined
humus and can be dispersed throughout the A by changes in color and texture, together with
and B horizons, giving these a darker—brown inclusions—anthropogenic or otherwise.
70 3 Sediments, Soils, and Stratigraphy in Archaeological Geology

Fig. 3.5 Bounding in a

spodosol. The A horizon is
bounded as “clear to
abrupt”; that between the
“A” and “E” is “gradual to
diffuse” (Photograph by the

Micromorphological features such as lamellae These contacts are generally defined on the
may indicate archaeologically interesting strati- basis of lithostratigraphy where breaks in the
graphic relationships, and then they may not color and textural quality of the sediments are
(Anderson and Schulderein 1985). obvious (Fig. 3.5). In geological sections, one
Using strictly pedostratigraphic terminology, is always cognizant of temporal gaps in a depo-
the vertical boundaries between horizons are sitional sequence which are designated as
described as either: conformities or unconformities. These are the
boundaries between adjacent strata. Figure 3.6
(a) Abrupt: if less than 3 cm wide illustrates pedogenic contact terminology
(b) Clear: if 3–5 cm wide described below. Conformities, in a geological
(c) Gradual: if 5–10 cm wide sense, have little temporal breaks in the deposi-
(d) Diffuse: if more than 10 cm wide tional record, whereas unconformities represent
temporal hiatuses where deposition has halted
The example in Fig. 3.5 shows a spodosol and even been eroded (disconformities). Con-
with clear bounding between horizons. formable contacts can be described by all of the
3.5 Describing Archaeological Sediments and Soils in Profile 71

Fig. 3.6 Disconformable (allo)contact between the Central Builders member of the Wyoming Formation and the
coarse gravel in the cutbank of Cayuta Creek in Sayre, Pennsylvania, at river km 452 (Thieme 2003)

terms listed above. Unconformities are more archaeological stratigraphy. Much of this is
likely to be described as abrupt due to significant attributed to floral bioturbation, but this may be
changes in the subsequent sediments. In the hori- overstating its influence on site stratigraphy.
zontal dimension, lateral variation in strata is These effects are not generally so extensive as
described somewhat differently than in the case to obliterate the archaeological stratigraphy.
of their vertical relationships. These are: While Thoms and Leigh (supra) have studied
haploidization in sandy soils, their arguments
1. Abrupt: the boundary between adjacent units could be extended to upland ultisols where forests
is distinct. can produce haploidization. A common soil cate-
2. Pinch-out: one unit thins progressively to gory in the Appalachians is the “typic hapluadult”
extinction; not uncommon in archaeological or clay-rich, haploidized ultisol.
deposits. Goldberg (1995), following Stein, reduces the
3. Interfingered: the adjacent units split into checklist for soils to the following:
smaller divisions that interdigitate with each
other. 1. Color (moist and dry) using the Munsell color
4. Gradual: adjacent units blend into one another notation
without abrupt lithological changes 2. Lithological composition (e.g., detrital
vs. chemical_
Haploidization: In contrast to sharp or 3. Geological vs. anthropogenic components
clear boundaries between the stratigraphic 4. Texture (e.g., gravel, sand, etc.) and the mor-
components of a solum, haploidization describes phology of the particles
what some term as “regressive pedogenesis” 5. Internal organization, such as bedding
(Johnson and Watson-Stegner 1990). The E, and 6. Consistence (e.g., loose, riable, firm, hard)
sometimes upper B soil horizons, can be substan- 7. Voids and porosity (types and amounts)
tially disrupted. Some geoarchaeologists (Thoms 8. Boundaries between units (e.g., abrupt, grad-
1995, 2007; Leigh 2001) have commented on soil ual, etc.)
haploidization as it impinges on cultural/
72 3 Sediments, Soils, and Stratigraphy in Archaeological Geology

3.6 The Geological Stratigraphic For the early part of humanity’s own story, the
Section rock record was the only reliable means of
assessing that antiquity with any assuredness.
The mapping of geological sections and Following the protocols of stratigraphy, early
structures has been a traditional capstone tech- antiquarians and later archaeologists were able
nique taught to almost all earth science students. to deduce precedence, succession, and simulta-
By understanding the horizontal and vertical neity of fossils and ancient tools found within the
relationships of geological strata across space, members of a formation. Necessarily, the pri-
the extent and nature of depositional and ero- mary relation was that simply the oldest strata
sional sequences are realized. The principal were at the bottom of the section—Smith’s prin-
differences in geological and geomorphological ciple in its most simplistic expression of the Law
mapping are the scale and the nature of what is of Superposition. As mentioned previously, Sir
being mapped. Scale is one thing while land- John Evans and Joseph Prestwich’s visit to
scape and lithology are quite another. Both use Boucher de Perthes’ excavations in the Somme
the same basic protocols on different scales of River terraces provided the geological frame-
lithostratigraphy. Both scales have relevance for work for assigning the true antiquity to the
archaeological inquiry. For instance, the location latter’s claims for the relationships between
of resources such as toolstone or metal ores will ancient fauna and tools within the gravels.
generally signal the likelihood of nearby habita- The relative age of a lithostratigraphic
tion and/or extraction sites. Conversely the pres- formation’s bed or member (a bed being a sub-
ence of these materials in the archaeological unit of a member) can be determined by super-
context requires the identification of the source position and the Principle of Inclusions which
location. Either way the mapping of the geology holds that inclusions or fragments in a rock unit
of an area is required. are older than the surrounding unit itself. For
instance, in the case of lava or tephra deposits,
those inclusions, geological or cultural, are, by
definition, older than the volcanic matrix. In
3.7 Nomenclature the case of volcanic flows or ash falls, such as
those of Pompeii and Ceren, cultural debris was
Formations make up geological sections that entrained or enclosed in the deposits and
in turn are composed of rock strata or beds provides what are termed a terminus post
organized into members which, unless they are quem—that is, the deposit cannot be older than
of igneous origin such as basalt or granite, were the inclusions. Before the advent of chronometric
laid down according to the principle of original dating techniques, archaeology relied on this
horizontality which presumes a depositional principle of stratigraphy without specifically
history of a uniform lenticular nature. Intrusive stating it. In East Africa, the finds of early
features—dikes, sills, etc.—are the result of vol- hominids have come from sedimentary
canic processes which, in turn, are the result of sequences of ancient lakebeds such as Olduvai
the tectonic activity due to the restless nature of Gorge (Fig. 3.7).
the continental plates. Structural features within Hans Reck, a member of the Geological Sur-
sections such as faults, syn-and-anti-clines, folds, vey of German East Africa, and later Richard
and the like are indicative of Earth seismicity as L. Hay (1976) untangled these sequences for
well. Stratigraphic relationships such as confor- the Leakeys—Mary and Louis—enabling the lat-
mity and the lack thereof express the broad ter to more confidently assign the Pleistocene
patterns of the rocks in the Earth’s crust. A geol- ages to finds such as Zinjanthropus (now classi-
ogist reads the geological section as a series of fied as Australopithecus) and Homo habilis
pages of the Earth’s history with all the twists (Fagan 1994). The Olduvai sequence is a model
and turns of that record imprinted therein. of the use of stratigraphic principles with the
3.7 Nomenclature 73

Fig. 3.7 Olduvai Gorge,

Tanzania. Upper, the
red sandstone of Bed III is
clearly visible as well as the
top of Bed II, lower,
stratigraphy of Olduvai
(Photograph courtesy of
Dr. Sandra Whitney)

straightforward terminology of Bed I, Bed II, forms—macro and micro—to help assign a
Bed III, and Bed IV. chronostratigraphic sequence to a section.
A Bed V, relatively thin and recent, was Lithostratigraphic units are defined by
identified at the top of the Olduvai section physical attributes of rocks where similarity in
being what is called the end member as well. lithology implies a geographic, and genetic,
Reck distinguished these beds within the Olduvai association in the units. A lithostratigraphic cor-
section based upon their color, sedimentological relation does not necessarily imply a temporal
texture, and biostratigraphy. The latter proce- equivalence. At Olduvai, biozones based on fos-
dure relies on the identification of fossil sil genera can provide a rough temporal control.
74 3 Sediments, Soils, and Stratigraphy in Archaeological Geology

“Roug,” in a temporal sense, implies that the 3.7.1 Allostratigraphy

presence of a particular fossil type (index fossil)
only indicates the time range of a plant or animal Goldberg’s eighth category, boundaries between
species—fine for geology but not very useful for (soil) units, has taken on increased importance in
archaeology. At Olduvai biostratigraphic corre- mapping paleosols. Soils are soft. Typically, they
lation was useful because of the millions of years are eroded prior to the deposition of younger
that the strata represented. units producing unconformities. Allostratigraphy
At Olduvai Beds I and II represent the most uses unconformities in the form of soil, soils, and
ancient lithified remains of sand and clay-rich geomorphic surfaces to define allounits which
lacustrine environments of a late Pliocene/early define discrete depositional (and erosion)
Pleistocene. Bed III represented a dramatic episodes (Morrison and Davis 1984; Morrison
change in color compared to the lower two 1991). Alloformations are defined as “mappable
beds. It is strikingly red in color and contrasts stratiform bodies of sedimentary rock defined on
sharply with its whiter stratigraphic neighbors. the basis of bounding discontinuities.” This defi-
The next upper member was named Bed nition follows Article 58 of the Stratigraphic
IV. Reck recognized the sedimentary nature of Code of the North American Commission
the rocks but noted the record of volcanism as on Stratigraphic Nomenclature (1983). In short
well as indicated by the tuff deposits. The Rift an allosurface is a period of relatively stable
Valley, an active seismic zone of extension environment where sediment deposition/soil
where the African and Eurasian continental formation occurs. Subsequent burial of that sur-
plates are pulling apart, has had volcanic activity face indicates a possible shift away from envi-
throughout prehistory. The volcanic deposits ronmental and/or climatological stability. An
provide lithological and geochemical signatures allosequence presumes a polycyclic series of
of individual eruptive events identified today allounits in section. Allosequences can be found
through tephrochronology. The volcanic rocks in almost any landform but they are most com-
also allow techniques such as radiogenic argon monly found in lacustrine and alluvial landforms
to be used to directly date them. Olduvai Gorge’s (Negrini 2002; Lichtenstein 2003). During
geological sequence was delineated by Reck and periods of stream down-cutting or lower lake
then confirmed by the early use of the then new levels, soils can form and mature on former
potassium-argon direct dating technique. Bed I, levees and terraces. With the return of fluvial or
containing both Australthropines and habilines, lacustrine sediment deposition, these surfaces
was found to date to 1.95 million years (Fagan, and their soils can become paleosols.
supra). Reck’s and Daly’s geological mapping of
Olduvai demonstrated the utility of stratigraphy Example: The Allostratigraphy
for elucidating the provenance or context of
of the Wyoming Branch
important archaeological facies.
of the Susquehanna River,
Archaeostratigraphy, as we have seen in the
Pennsylvania, USA
section on caves, describes the sequence of the
A recent study of ten large river basins located in
materials of antecedent or subsequent cultural
the Appalachian Piedmont of the southeastern
deposits found in an archaeological site. This is
United States has quantified background erosion
illustrated in Paleolithic site deposits. In recent
rates (8 m/m.y.), using Be10 dating of basin
years, the interstratification of “archaeo-facies”
sediments (Reusser et al. 2015). Additionally
in Paleolithic sites has led to animated
that study examined the transport of eroded
discussions about their exact nature and their
sediments by these rivers and found that only
interpretation. An excellent example is that of
6 % of the eroded sediments were transported.
Middle Paleolithic Mousterian technofacies—
The non-transported sediments remain stored at
Quina, Denticulate, Mousterian of Acheulian
the base of hillslopes and in valley bottoms
Tradition (MAT), etc.
3.7 Nomenclature 75

(supra). This includes the significant amount of Pleistocene outwash gravels. While a few
sediments eroded after historic land and forest hundred kilometers north of those rivers in the
clearances began in the Colonial Period includ- Piedmont study, the Susquehanna is a major
ing, perhaps, an earlier increase due to late pre- Appalachian draining river, matching or exceed-
historic maize agriculture. In the following two ing the scale of the Savannah River. It intersects
examples, we shall examine how geoarchaeology the more narrow northward extent of the Pied-
provides data to examine stratigraphy as well as mont in Pennsylvania on its way to the
landform history. The first example is from Chesapeake Bay. Unlike the more southerly
Pennsylvania, USA, and the second from the Savannah, the Susquehanna drained the
Piedmont of North Georgia. Laurentide Ice Front.
The alluvial deposits of the North Branch of Thieme (2003) describes an Early Archaic
the Susquehanna River represent an archive of component archaeological site on the Central
environmental change as well as a repository of Builders member in the upper 20 cm of a buried
artifacts and cultural features made by prehis- soil (3AB horizon) with a hearth containing
toric people (Thieme and Schuldenrein 1998; wood charcoal dated to 9165 +210/205 BP
Thieme 2006). While many local factors affected (A-10053), which calibrates to 11,070–9,699
how frequently the river flooded and when it BP at two sigma. We must not expect this date
stored sediment within the valley, stratified to be that of the unit itself; while the contact
archaeological sites exhibit remarkably consis- represents a chronological discontinuity, the
tent discontinuities that bound discrete packages gap between the outwash gravel and the Central
seen in both culturally sterile alluvium and Builders’ finer-grained sediments is not well
archaeological site profiles. These include some constrained. It is reasonable to assume the depo-
Paleoindian as well as Early Archaic sites sition of the finer-grained materials began in the
(Thieme, supra). early Holocene, sometime before the establish-
Thieme (2003) proposed an allostratigraphic ment of the archaeological occupation. What the
unit, the Wyoming Valley formation, for the Central Builders unit does provide is clear cor-
alluvial deposits which underlie the first terrace roboration for Reusser et al. (supra) with regard
in the North Branch of the Susquehanna River to the long-term storage of eroded materials.
Valley (Fig. 3.6). He correlated this formation These materials have been stable surface for the
throughout the valley from Binghamton, millennia needed for soils to form.
New York, downstream to Harrisburg,
Pennsylvania. The lower bounding surface is a
contact with outwash or ice-contact stratified 3.7.2 Sequence Stratigraphy
drift at the base of the alluvial terrace. Four
members are distinguished within the Wyoming In modern geology, the nomenclature of stratig-
Valley formation on the basis of buried soils raphy has seen a new terminology introduced
which formed on former terrace surfaces. under the rubric of “sequence” stratigraphy to
The members were traced through 23 strati- describe basin edge sea-level rise and fall. Here
graphic sections in the 250 km study area on the term formation is replaced by a new term that
the basis of field observations, thin-section of sequence. Sequence stratigraphy recognizes
photomicrographs, and laboratory analyses of that sedimentary rock units, adjacent to basins,
grain size, soil chemistry, clay mineralogy, and have generally resulted from sea-level change in
magnetic susceptibility. A total of 139 radiocar- transgression or regression. In the former, the
bon dates constrained the age of bounding shoreline moves landward, and in the latter, it
surfaces within an allostratigraphic framework. moves seaward. A vertical representation of
In Fig. 3.6, two sediment (storage) units are these processes shows different facies super-
exposed, one a member of the Wyoming Forma- posed according to the particular depositional
tion (Central Builders) and the other the late environment. For example, a transgression
76 3 Sediments, Soils, and Stratigraphy in Archaeological Geology

of nearshore sands (sandstones) underlies et al. 1999). These dramatic shifts in sea level,
mid-shelf facies (shales) which are, in turn, during regressions, opened corridors and “land
overly deep sea muds (limestones) in the section. bridges” between continents such as Asia and
The sequence of transgression, highstand, regres- North America, the Bering Strait land bridge,
sion, and lowstand strata, as typically seen in and between Europe and the British Isles,
seismic sections, is designated as tracts. Doggerland (Fig. 2.6); http://instaar.colorado.
A system tract is made up of parasequences, edu/QGISL/bering_land_bridge/). Both humans
which, in turn, are made up of sets of facies like and other biota used these connections for colo-
those just described (Reading and Levell 1996). nization purposes. In the lithostratigraphy of
Erosional surfaces separate parasequences. A coastal areas, we see this recorded in the sedi-
regression or abrupt basinward shift in sedimen- mentary sequences for those periods.
tary facies exposes a depositional sequence to Lericolais et al. (2003); Törnquist et al. (2003)
subaerial weathering and creates what traditional use sequence stratigraphy to elucidate the
stratigraphic nomenclature would term an late Pleistocene sedimentary record for the
unconformity. In a sequence-stratigraphic Channel River and the Rhine-Meuse delta (the
description of a depositional system, the Netherlands).
“formations” are replaced by “sequences” with By subdividing the sedimentary deposits into
gaps or unconformities interpreted as flooding unconformity bounded sequences, the timing of
(ravinement) intervals of a significant temporal sea levels in the late Pleistocene can be exam-
duration. ined. Lericolais (supra) identify a lowstand his-
As we have discussed in the context of the tory for the Channel River at MIS stages 22, 16,
coastal landforms and prehistoric Doggerland, 6, and 2 (LGM). Törnquist (supra) date the for-
sea level has played a significant role in the mation of the last sequence boundary using five,
archaeology of the Pleistocene, alternately over 35 m in length, sediment cores, of the
exposing or flooding vast areas of continental Rhine-Meuse delta, to shortly after 80 ka or
margins. Worldwide or eustatic sea-level change, roughly MIS 5a or a falling stage of sea level,
in the Pleistocene, has been driven by multiple thereafter (MIS 5/4) with subaerial erosion and
periods where large volumes of the oceans have sediment bypass. The subsequent lowstand for
been sequestered or released during glacials, MIS 4 was either not evident in their sediment
interglacials, and more transient Heinrich and data or was not as pronounced as that of the MIS
Dansgaard-Oeschger (D-O) events (Bond 5/4 interval.
Techniques for Archaeological Sediments
and Soils 4

4.1 Sampling Sediments and Soils: hand, but in deeper sites, mechanical excavation
Monoliths to Sediment Grabs is required. At the profile face, one can assess
many of the primary parameters used in a
Undisturbed Versus Disturbed Sampling—Pro- description of a sediment or soil—color, textural
file Samples, Soil Monoliths, and Solid Cores: changes, contacts, etc. The principal drawback in
Undisturbed sampling of an archaeological sedi- simple hand excavation is one of time. A longi-
ment can be as straightforward as the description tudinal study of an excavation profile such as at
of an exposed profile in the field. It can also mean an ongoing, relatively long-term archaeological
the recovery of a representative fraction of those excavation is the most appropriate way to gain an
sediments for studies to be performed in a labo- adequate field description of that profile. Such is
ratory rather than in a field setting. Disturbed the case in many more famous studies such as
samples are those taken from individual horizons cave sites such as Abri Pataud (Farrand 1975;
or strata in which structure and morphology Stein and Farrand 1998), Combe Grenal (Bordes
information is sacrificed. The sense of the term 1972), Tabun (Jelinek et al. 1973), and Rogers
“sample” used in this discussion is not one com- Shelter (Wood and Wood and McMillan 1975) to
monly associated with statistical evaluations but name but some more well-known examples.
simply the acquisition of a portion of the sedi- Mechanical excavation or exposure of a soil/
ment or soil for analytical purposes. Sampling sediment profile is the most commonly used
(statistical) procedures can, in many cases, be method in geomorphology today. The most com-
used to assure the representativeness of the field mon mechanical means is the “backhoe”—a trac-
or laboratory fraction of the sediment or soil tor or tracked hoe-type excavator capable of
being examined. This issue will be addressed in digging to depths of 7 m with lengthy trench
the following section. In this section, we shall exposures. If one is interested in investigating
examine a variety of techniques commonly used fluvial valley, this type of excavator can expose
to acquire data on archaeologically interesting entire terrace and flood plain systems within
sediments. it. Unlike discontinuous sampling, as typical of
Profile Exposures—The simplest method of coring studies, the continuous exposure, by use
examining a profile or section is to find an out- of the mechanical hoe, provides a more synthetic
crop or exposure. Failing this most researchers view of the depositional stratigraphy—in both
create their own by excavating one. This can be vertical and horizontal dimensions.
done by either hand or mechanical methods. Tra- Long-term studies of sediment profiles are the
ditionally shallow profile exposures are dug by exception rather than the rule in today’s

# Springer-Verlag Berlin Heidelberg 2016 77

E. Garrison, Techniques in Archaeological Geology, Natural Science in Archaeology,
DOI 10.1007/978-3-319-30232-4_4
78 4 Techniques for Archaeological Sediments and Soils

archaeology. Researchers are more often than not readings of the profile face are below this level.
required to get the information on an excavated The result is the same—a representation of the
exposure in a few days or less. Working under profile face to scale. Where feasible, a photo-
such time constraints, the field description of the graph of the profile should be taken to assist in
exposure is important but so is the recovery of a the interpretation of the profile after the excava-
representative sample of those sediments for tion has been closed. Color photography is pre-
more thorough study in the laboratory. Still the ferred over black and white so as to record the
field exposure provides the most faithful picture important parameter of soil or sediment color.
of the exposure in terms of scale and setting. It is Photography of a profile face is not an easy
critical that the profile face be properly prepared task. When they are available, a professional
before beginning its description. Simply photographer should perform this important
excavating the profile does not provide the best task. In some cases, as with an excavation trench,
view of the face. One must dress the profile face to the focal distance to the face is such that several
adequately bring out the features of the deposits. photographs must be taken to develop a compos-
The tool of choice for this task is the ite image—mosaic—of the profile. Lighting is
archaeologist’s friend—the trowel. If archaeology almost always difficult on a profile, and the best
has mastered one thing in stratigraphic studies, as results are obtained in complete shade rather than
first mandated by Sir Mortimer Wheeler, it is the in direct sunlight just as with the case of
meticulous care taken with the vertical profile. obtaining Munsell colors. Digital photography
The nature of a deposit is best seen when the has provided significant improvement, and ease,
profile face is cleaned and made uniformly even. of documenting field profiles.
After the face is adequately prepared, Profile Samples: Soil monoliths are small ver-
measurements must be taken. The establishment tical columnar sections of a profile. This technique
of a reference datum level is required. Most was developed by Russian soil scientists in the
researchers approach this step in one or two late nineteenth century and introduced to the rest
ways—(1) establish a surface datum level or of the world at the Chicago International Exhibi-
(2) establish a relative datum level on the profile tion in 1893 (Vanderford 1897). The monolith can
face. This writer has no preference in the selec- function as both a permanent record of the deposit
tion of either method. When applied with care, and temporary reference section from which
either technique works equally well. In this world measurements can be made. The profile can be
of “automatic” laser levels, the use of a surface reconstructed from a series of these smaller
datum is manifold easier than in the past when sections. This procedure is laborious and can be
workers used string levels to transits to control accomplished almost as well by the quicker
verticality in an excavation. Even with use of the method of coring. The advantage of the monolith
relative datum level at some time, the stratigra- is in its use in the laboratory in assessing, under
phy must be reconciled to the actual topographic more controlled conditions than those of the field,
setting or geomorphic surface. diagnostic parameters in detail and under more
Using the relative datum level technique, the convenient time constraints.
placement of the datum can be made at any point To take a monolith, a container is constructed
on the profile face as long as it—the datum line— for the monolith from either metal or wood. The
is level. The datum line is usually a simple string size of monoliths varies but commonly is about
level strung between two points on either end of 1 m in length, 10–20 cm wide, and a few
the profile face. From this level line, the stratig- (<10 cm) cm thick. Methods vary, in detail and
raphy of the profile can be measured up and objectives, but the basic method is to use a
down. The result is an accurate picture of the wooden or metal boxlike tray that is open on
profile to the scale that one has chosen. The one side. The monolith is excavated to the
surface datum level line is “zero,” and all dimensions of the box and cleared on all sides
4.2 One Method for Constructing a Soil Monolith 79

including the base of the section. Once this is 3. With two people in the pit holding the frame
accomplished, the box is placed over the mono- and one at the top of the pit, the person at the
lith and pressed firmly against it to assure a firm top of the pit begins shoveling down parallel
juncture between the container and the section. to and 4–500 away from the frame.
Cutting the monolith away from the profile face 4. As the shovel removes the vertical layer of
is perhaps the most difficult task requiring soil, the people in the pit holding the frame
patience and some skill with a sharp shovel or begin leaning the frame back as sections of the
trowel to uniformly disengage the section. Some sample are added to it. (In very sandy soils
workers simply press a metal frame into the during the dry season, thoroughly wet the soil
profile face and then excavate the monolith out overnight and remove small sections at a time
along with the box. Either way collects a rather into the frame.) This part may have a few
faithful representation of the profile being sam- failed attempts. Multiple samples can be
pled. Of course the exact nature of the deposit taken from the same pit.
will condition the ease of removing the monolith 5. Once the entire sample is collected, carry the
particularly in rocky sediments or archaeological frame and sample outside of the pit and trim
deposits with many cultural inclusions. Smaller the lip down to the frame edge.
monoliths for purposes of micromorphological 6. Add adhesive to presentation board and invert
studies can be taken using the Kubiena frame over extracted sample in frame. Very care-
(Kubiena 1953). These small metal boxes are, fully press frame into presentation board and
again less scrupulously but expeditiously, mim- invert till sample is transferred to presentation
icked by the use of domestic metal food board. Remove frame.
containers whose lids have been removed. 7. Carefully transport the sample. Clean face of
sample. Dry.
8. Add full strength Future Floor Wax and thor-
oughly saturate soil. Apply with a spray bot-
4.2 One Method for Constructing tle. Will dry clear.
a Soil Monolith 9. Add finishing touches to presentation board.

Materials Once the monolith, of whatever dimensions,

• One metal frame: 50  4–600  1–200 . is freed from the profile, it must be stabilized
• One presentation board: made to fit sample from before transport. Thicker monoliths require less
frame, backboard must be a 2  800 , adding treatment before transport. The monoliths, par-
quarter round at top helps hold the sample. ticularly the larger 1 m varieties, are heavy—
• Glue for adhering sample to presentation 25 kg or more. All specimens should be carefully
board. logged, photographed, and described before
• Future Floor Wax placing a coverlid over them. They can be safely
• Spray bottle transported to the laboratory where they are
• Backhoe uncovered and more elaborate studies are
• Three people begun. In some instances, the monoliths are
• Sharpshooter with flat shovel head impregnated with plastic resins for reasons rang-
ing from the removal of thin sections to the
Method fabrication of displays. From the standpoint of
1. Excavate a pit with a backhoe to over 5 ft deep scientific studies such as textural, chemical, and
with a clean, vertical face on one wall and age determination, it is best to leave impregna-
room for two people to stand comfortable in tion of the monoliths to last (Fig. 4.1).
front of the smoothed face. One use of impregnation of soil profiles has
2. Hammer the frame into the smooth face from been used to obtain “peels” of the face of a
surface to the desired depth. deposit (Orliac 1975). In this method, a thin
80 4 Techniques for Archaeological Sediments and Soils

devices such the “box corer” used in sampling

marine and lacustrine sediments. Coring devices
range from gravity corers, piston corers, percus-
sion corers, and drive corers to rotary corers. The
first two types, like the box corer, are used for
inundated or wetland sediments. Percussion,
drive, and rotary corers are typically used for
terrestrial coring although a variant of the percus-
sion type, the vibracorer, is routinely used in wet-
land and inundated contexts. Vibracorers are
particularly effective in penetrating unconsoli-
dated sandy sediments. A recent, rather innova-
tive, “coring” device is the freezing corer or
“cryoprobe” which has proved effective in both
inundated and terrestrial settings (Garrison 1998).

Rotary Corers These are mechanical coring

devices that extract solid columns of sediment
or soil that is relatively uncompacted. Unlike
larger rotary drilling systems such as used in
hydrocarbon or hydrological exploration, rotary
corers used in archaeological geology do not
Fig. 4.1 Cutting out a soil monolith usually use drilling fluids which are used to
lubricate the drill bit in rock. As almost all
archaeological deposits are sedimentological in
layer of synthetic latex is applied to the profile nature, the rotary corer can penetrate their rela-
and allowed to harden. When dry, the latex or tive shallow depths (a few meters as a rule)
epoxy preserves the profile to the depth of the without lubricants, which, in many instances,
impregnation—a few mm to a cm or so. This prove inconvenient for accurate chemical and
image is adequate to describing color and tex- textural tests. As shown in Fig. 4.2, the rotary
tural differences as well as providing a facsimile corer is most often vehicle mounted for ease of
of the profile that can be kept for future refer- transport to the archaeological site. The coring
ence. In Switzerland, Arnold and his co-workers tool is an open cylinder a few centimeters in
(1978) have used epoxy peels of entire archaeo- diameter that has a drill bit or straight cutting
logical surfaces both for study and for more edge of hardened steel alloy. The barrel of the
elaborate particle size, chemical, or other studies corer often features a screw-like exterior creating
such as palynological ones; the peel technique is the appearance of a hollow auger that allows the
inadequate in terms of sample size as well as central section to remain open for the acquisition
adulterating the sediments of the profile. of a continuous, solid sediment/soil column in all
but the stoniest or wet conditions.
Core Samples This type of undisturbed sedi- The advantage of the rotary corer, or for that
ment or soil column is acquired by a variety of matter any corer type, is the recovery of an
methods and results in specimens of varying undisturbed column of sediment or soil. The
dimensions of length and diameter. The key recovered monolith is reflective of the macro-
word in describing cores is diameter. The coring and microscale stratigraphy without a great deal
devices used to obtain soil and sediment samples of compaction or mixing of sediment/soil struc-
are cylindrical as a rule. There are square coring ture and features such as lamellae. For
4.2 One Method for Constructing a Soil Monolith 81

utilizes an eccentric or offset gearing system to

develop a steady vibratory motion on the core
barrel enabling it to “settle” into a deposit.
Vibracorers can be pneumatic, hydraulic, or elec-
tric powered. Percussion corers can vary in size
from a meter in length to over 10 m. Core barrel
diameters can vary from a 50 to 75 mm as a rule.
The most popular terrestrial drive coring device
is the Giddings corer. This device applies
hydraulic pressure to a core barrel of 75 mm
diameter extracting solid columns of over a
meter in length (Fig. 4.3).
Hand-driven percussion corers are the most
popular and economical type of the coring
types. These corers typically consist a solid or
split-core sampler barrel, a series of solid metal
extensions that terminates in a swage- or screw-
type adaptor for a slide hammer. The individual
hammers the core into the ground, in set
increments, determined by the length of the cor-
ing tool, to the ultimate depth of interest and then
extracts the solid core. A type of coring tool
called a gouge auger can be used. This “auger”
is a cylindrical open device for use in soft
sediments. Hand-driven corers are rarely over
Fig. 4.2 Rotary corer. Shown deployed in the excavation
of a Roman era cemetery in urban Tunis, Tunisia, the 25 mm in diameter and obtain relatively small
auger rapidly bores through the clastic sediments. Photo- samples of the sediments/soils. Compared to
graph by Nina Šerman some screw augers of 100 mm diameter, there
is a clear advantage in overall sample size for the
measurement of parameters such as porosity, screw auger even if the sediment is a disturbed
moisture content, and consolidation of clays sample. Core liners of clear plastic can be used
solid, undisturbed core samples are mandatory inside all types of corers from mechanical
(Shuter and Teasdale 1989). models to hand types. For maximum protection
and preservation of the in situ context of the
Percussion and Drive Corers These types of cored deposit, these sampling devices are
coring devices are excellent for the recovery of recommended. This is important for the recovery
soft or compressible sediments (supra). Likewise of fragile macrobotanical or macrofaunal
they work well in unconsolidated sandy materials as well as pollen remains.
sediments particularly the percussion corer. The An important factor in percussion corers, most
difference between the percussion and drive notably those that are hammer driven, is the
corers is the application of the pressure on the disturbance and compaction of fine-grained
coring tool. The percussion corers apply inter- sediments. This can occur in vibrating corers as
mittent impact pressure either by a hammer or by well. Quantitative data is lacking on this point,
raising and dropping the entire coring tool. The but it is suspected that an increase in the driving
drive corer applies pressure in a constant manner rate, measured in blows per foot, increases the
usually with a hydraulic booster system. The potential for disturbance and compaction in the
vibracorer is a percussion type of corer that recovered sediment column. The author has
82 4 Techniques for Archaeological Sediments and Soils

Fig. 4.3 A Giddings

hydraulic coring system.
The operators are
exchanging a “Kelly
(extension) bar” from the
adjacent rack, to increase
the depth of the core. The
vertical “mast” contains a
hydraulic “ram” powered
by a small engine (not
shown) to push and pull,
utilizing the lever controls
(red), the core barrels used
to collect sediments. The
system is easily deployed
using a vehicle as shown. A
typical core takes less than
30 min to collect

observed this in small diameter, rigid barrel hand Kullenberg corer, and the Ewing type. Figure 4.4
corers when using a 5 kg slide hammer in resis- illustrates the principal components of a Ewing-
tant deposits such as stiff muds or sands. If a type piston corer used, by the author, to recover
sample requires a driving rate in excess of riverine and marine sediments. A smaller, auxiliary
25 blows per foot, then one should be on guard corer, called a messenger corer, is added to a grav-
for sample disturbance. If the corer’s hammer is ity corer design whose barrel assembly has
quite heavy—say, 50 kg or so—then this is par- incorporated a piston coupled, by a cable, to the
ticularly true with regard to compaction of the messenger corer and the hoisting mechanism for
recovered core sample. both elements of the piston corer. As shown in the
illustration, the messenger corer is cantilevered off
Gravity (“Drop”) and Piston Corers These the larger gravity-piston corer such that it will
corers, as previously said, are more often used contact the sediments first, raising the cantilever
to core marine and lacustrine sediments. Both arm such that a cam device, in turn, releases the
types are raised and then released to rely on cable at the point of attachment on the hoisting
gravity for the force required to penetrate device along with the main corer. The heavier,
inundated or soft wetland sediments. In the main corer drops in free-fall mode, into the
contexts that they are deployed in, they are rarely sediments, while simultaneously the fixed cable
coring into soils with the exception of soils remains taut along with the piston. Thus, as the
developed on peat bog deposits such as paleo- core barrel falls deeper into the deposits, the piston
histosols. To aid in penetration of the sediments, moves up the core barrel drawing sediments into
these corers use heavy weight loads attached to the partial vacuum created by the rapid “rise” of the
the upper part of the coring barrel with as much piston up the barrel. These corers use plastic core
as a ton of weight added to corers of 2–3 m in liners to aid in creation of the vacuum and to
length. Needless to say, a corer of this weight is preserve the sample. Piston corers are sometimes
not recovered by hand. prone to problems such as foreshortening or
The piston corer is a gravity corer design that stretching of sedimentary units (Buckley et al.
uses the creation of a partial vacuum in the recov- 1994).
ery barrel to recover wet, unconsolidated
sediments. The piston corer works well in multi- Freeze Corers Hammer-driven freeze corers or
bedded deposits of silts and sands. Various types “cryogenic soil probes” are a fairly recent addi-
are used today such as the Livingstone corer, the tion to sediment sampling in archaeology
4.2 One Method for Constructing a Soil Monolith 83

Fig. 4.4 Ewing piston


(Fig. 3.8) They have been used to sample wet to relatively standardized field models (Hochuli
sediments for biological purposes as well as 1994; Garrison 1998). The principle of the freez-
those of archaeology (Grebothé et al. 1990; ing corer is simple—insert a closed-tip metal
Lassau and Riethmann 1988; Miskimmin tube into the earth, fill it with a cryogenic fluid,
et al. 1996). Early in the 1990s, European and such as nitrogen, and freeze the adjacent
North American workers quickly moved the sediments to the metal tube. Extraction of the
design of these corers forward from prototypes tube brings with it all the frozen sediments
84 4 Techniques for Archaeological Sediments and Soils

adhering to the core barrel. The diameters of the 4.3 Handling and Description
freezing corer barrel have varied with develop- of Cores: Some General
ment as has the freezing fluids used in these Considerations
designs. The present-day archaeological designs
have opted for 75 mm diameter steel barrels Depending on the type of core sample one has—
about the length of rotary and drive corers such auger, drill, etc.—some general observations
as the Giddings type, e.g., 1 m or so in length. should be made about the sample as quickly as
Those used in lacustrine settings are cylindrical possible, either in the field or immediately upon
or wedge-shaped containers of slurries of alcohol arrival at a laboratory. All samples allow the
and dry ice (Lotter et al. 1997). determination of the basic parameters of a sedi-
Figure 3.8 illustrated the design and use of a ment or soil—color, general texture, consistency,
typical freezing corer. The corer is hammered into general organic content, etc. Some things must
the sediments, and then a transfer hose is attached wait for more rigorous measures to be applied to
to valving on the top of the corer. Liquid nitrogen, the sample such as the quantitative determination
at a temperature of 197  C, is circulated through of size classes, total organic matter content,
the corer for a period of about 20 min. A mechan- macro- and microbotanical makeup, chemical
ical extraction assembly, composed of a tripod attributes, magnetic properties, etc. Nonetheless,
and lifting device, are attached to the frozen core a researcher must recognize that the freshly
sample and it is withdrawn. The description of the recovered sample contains information that can
recovered sample is described much the same as be readily assessed. These data, however qualita-
any sediment or soil sample—color, texture, tive, in some cases, provide the investigator with
structures, etc. Moisture content is important to firsthand information as to the general nature of
the efficient use of the freezing corer, but it is the sediment or soil. Additionally, these data are
surprising at the relatively small amount of mois- not to be replicated as the simple passage of time
ture needed for adequate freezing in most subtly alters the core sample through changes in
sediments (5–10 %). moisture content and chemical changes such as
oxidation and the corrosion of mineralogic or
Augers and Drills Undisturbed Sampling organic remains. Rapid storage of the core sam-
Techniques—Rotary augers and drills have a ple, in a cold room or freezer, can do much to
long history in soil and sediment sampling. ameliorate or retard these changes, but even in
Recently, two noted geoarchaeologists have these cases, workers typically record as much
discussed the relative merits of small-diameter data about the sample as possible before placing
corers (25 mm) versus larger-diameter (100 mm) the core in storage. If the core sample cannot be
soil augers (Stein 1986, 1991; Schuldenrein archived in such facilities, then it is imperative
1991). Both types are commonly used today, that the material be sampled and described as
and their use will continue for both practical quickly and as fully as possible.
(speed, costs) and scientific reasons. Rotary Rotary, Percussion/Drive, Box, and Piston
augers and drills recover disturbed samples in Cores: These devices can extract whole core
terms of pedologic structure and micromorphol- samples such that the structure of the peds and
ogy. If the samples are in soil horizons, the strat- the lithostratigraphy are generally preserved. The
igraphic integrity of the deposits are degraded or sample can be extruded from the core barrel, a
destroyed, but much information is retained for sample holder, or the subsample can be
both field and laboratory studies. Larger sample contained within a core tube liner made of a
sizes are routinely recovered using soil augers plastic material. The core sample can, in some
and mechanical drill systems. Mechanical dril- cases, be maintained in a type of container that
ling systems can recover sediments as effectively can be removed or detached to allow transport,
as comparable solid column corers. storage, and examination of the sample.
4.4 Standard Operating Procedure for Collection of Sediment Samples 85

Depending on the length and diameter of the estimates of the sedimentation rate and mixing
core sample, analysis and storage considerations depth for the GLNPO Lake Michigan Mass Bal-
can vary. In the case of small-diameter (25 mm/ ance Study.
1 in.), shortish (30 cm/12 in.) core samples,
handling and storage protocols are not very com- 2.0 Summary of Procedure
plex. In the case of long, larger diameter, 75 mm Before any box cores are taken, test grabs,
+ (3 in. +), cylindrical cores, handling, sampling, using a ponar, will be taken to determine the
and storage become a more serious issue. With suitability of the sediment for coring. If coring
the larger core samples, it is not uncommon to is possible, then the box corer will be deployed.
section or cut the sample into shorter, more man- Once the box coring is completed and box core is
ageable lengths such as 50 or 100 cm. Again, the back onboard the ship, then four 10 cm
core may, in the case where the liner or barrel is (ID) plastic tubes will be inserted by hand into
used for handling, be sampled and then archived the Master box core, thereby creating four
without too much disturbance. Cores that are not subcores (A–D). Each of the subcores will be
immediately placed in climate-controlled storage sectioned and these subsamples stored for future
such as cold rooms can alter, biologically and analysis.
chemically such that the peds and inclusions are
adulterated in terms of shape, size, etc. Textural 3.0 List of Equipment
and color differences are typically maintained.
The general nature of lithostratigraphic
relationships, in the core, can still to be seen, Item Quantity
but the absolute, metric nature of the Modified box corer (Soutar corer) 1
Box corer extraction rigging 1
relationships can be compromised.
Set of critical spare parts for box corer and 1
Whatever the core type, the infield and labo- extraction rigging
ratory description of the sediment section follows Hydraulic extruding stand for 400 diameter 1
a commonsense procedure in most cases. A subcores
standardized protocol, such as the following log Set of core sectioning gear 1
of the box sediment corer, shown in Fig. 4.6, is 125 ml polyethylene bottles/pre-labeled and As
commonly used. This example is used by the tared needed
Ponar grab sampler 2
Great Lakes Environmental Research Laboratory
Winch for Ponar deployment 1
and the Great Lakes Water Institute (Edgington
10 cm/400 diameter subcore butyrate tubes
and Robbins 1991). The categories used in the 12 vacuum-extractor caps 2
core description can vary and be easily Portable vacuum pumps with Tygon tubing 2
customized to the particular objectives of a proj-
ect. The categories shown in this example were
used in the description of lacustrine/marine box 4.0 Sampling Procedure
core samples, but they could as easily be used for 4.1 Test grabs, using a Ponar grab sampler, will
sediment cores for most any application. be taken to determine the suitability of the
sediment for coring. If three grabs return
without a sample, then the site will be
4.4 Standard Operating Procedure vacated. If the Ponar grabs are obtained but
for Collection of Sediment coring is not feasible, then surface samples
Samples from the grabs will be obtained. If coring is
possible, then box coring will be undertaken
as long as there appears to be a limited risk of
1.0 Scope and Application damage to the box core.
The application of this sampling procedure is for 4.2 Once the box core has been retrieved and
the collection of sediment cores, using a box is back on the ship’s deck, then the core
corer, for the analysis of radionuclides to provide is examined for acceptability. This
86 4 Techniques for Archaeological Sediments and Soils

examination is done by using the viewing 5.0 Sample Custody

window on the front side of the box core. If After the sectioning of each core, the Co-PIs will
the core is unacceptable, then the contents of verify that all the samples are accounted for and
the box core will be released and the box that they are transferred to proper storage. After
corer redeployed. sampling has been completed and the samples
4.3 Acceptable box cores are subcored by transported to the lab, the CO-PIs will again
carefully inserting a 10 cm diameter butyrate verify that all samples have properly transferred
tube into the core. Distortion of the sediment and stored. The location of all samples is noted in
during the tube insertion is minimized by the sample log.
the application of a partial vacuum to the
tube top. By continuous manual adjustment 6.0 Sample Labeling and Logs
of the vacuum as the core is inserted, the Prior to each sampling event, a complete set of
interface within the tube remains in align- sample bottle labels will be prepared. The num-
ment with the interface of the surrounding ber and type of these labels will depend on the
sediment in the box core. length of the sediment core recovered and the
4.4 Sediments within the tube are hydraulically estimated sedimentation rate. An example of a
extruded and sectioned onboard the ship. typical label is seen in Fig. 4.5.
Extrusion is done by the application of Box cores are so-named because sediment
water pressure from the ship’s hose line to samples reside in metal “boxes” (Fig. 4.6). As
a rubber stopper inserted into the base of the indicated in Sect. 4.2, the acceptable sediment
core tube. Fine control of water flow allows core is removed from the corer, another “box”
slow movement of the core upward into a sampler is loaded, and the corer is redeployed.
separate short section of tube (the collar) Also, as noted the core can then be subcored for
placed in-line with the core tube top. The various purposes ranging from textural analysis
collar is scribed in cm intervals so as to to age determination.
define the amount of core section to be Percussion, drive, or otherwise cylindrical
displaced laterally into an aluminum sediment cores, upon recovery the core will gen-
receiving tray. erally be “split” into equal halves using power
4.5 Subcore taken for the analysis of radio- saws. The halves were measured and described
nuclides will be sectioned with plastic as to color and sediment texture. They were
utensils. drawn on standardized recording sheets, as
4.6 Subcore samples are stored in conformity shown, and photographed along their entire
with EPA QA/OC requirements. sections. Subsampling, of the core, for textural,
4.7 A backup core is taken in case of unexpected geophysical, chemical, and botanical analyses, is
problems in analyzing the first core or if an relatively straightforward.
interest in analysis of additional material The recovery of samples for use in age dating
develops. is relatively routine as well, but in certain cases,
4.8 Core lengths are expected not to exceed special handling is required. For instance, the
50 cm in length and should more than paleomagnetic/archaeomagnetic evaluation of a
cover the entire post-settlement history of core sample can only be done on intact samples.
deposition. The dating is typically done in situ along the core
4.9 A detailed record of the sediment charac- using a magnetometer designed for this. Most
teristics, as a function of depth, as well as a archaeological dates are not obtained from the
notation of any unusual properties (i.e., large magnetic dating of cores unlike geological core
wood chips) will be entered in the sampling samples where the procedure is more common.
log. An example of the sampling log form is For one thing, paleomagnetism measures field
shown in Fig. 4.5. reversals across a broad time range, whereas
4.4 Standard Operating Procedure for Collection of Sediment Samples 87

Lake Michigan Mass Balance and EMAP Study Sediment Sampling Log
Station No:
Core No:



Latitude: Longitude:

Section Description Section Description

0–1 29 - 30

1–2 31 - 32

2–3 32 - 33

3–4 33 - 34
4–5 34 - 35

5–6 35 - 36
6–7 36 - 37

7–8 37 - 38

8–9 38 - 39
9 – 10 39 - 40

10 - 11 40 - 41

28 - 29 59 - 60

Sectioned by Recorded by
Samples Checked by Received & Checked by

Samples Stored at Date

Fig. 4.5 Sediment sampling log

archaeomagnetic dating requires declination and As optically stimulated luminescence (OSL)

inclination measurements which are compared to dating of sediments has become common, the
regional reference dating curves (Eighmy and need for special handling of the particular core
Sternberg 1990). Radiocarbon dating of organic samples will be needed. Basically this entails the
inclusions is the most typical means of age deter- use of darkness and photographic darkroom
mination of cores within the practical range procedures to prevent undue exposure of the
(~50,000 years) of the respective carbon dating core sediments, such that the remnant OSL signal
techniques—beta counting or accelerator mass is not affected (Godfrey-Smith et al. 1988). From
spectroscopy (Blackwell and Schwarz 1993). one-half of each core sediment, samples were
88 4 Techniques for Archaeological Sediments and Soils

as most archaeological deposits are within this

range, their utility will remain high. The hand
augers and drills allow very rapid sampling for
small, yet representative samples. Color, tex-
tural, or even chemical analyses require only a
few grams of sample in most cases, so the vol-
ume of sample needed is not that great. Again, it
is the continuity and cohesiveness of the sample
that is at issue, and if this not a methodological
priority, then the use of the devices is feasible.
Like the more intact samples from the rotary/
push/percussion corers, handling, analysis, and
storage depend a lot on the sample size. Long,
continuous sediment samples may be extracted
using the hand drill or auger. Mechanized units
certainly can do this. The length of the sample is
only limited by the number of extensions that are
used in the recovery of the core. Samples from
the core may be placed in drill core boxes or
simply placed in plastic bags. These latter items
are labeled as to the location and weight of the
Fig. 4.6 Sediment box corer sample as a well as the other identifying infor-
mation—core sample #, date, etc. One problem
encountered by this author, in the use of a
drawn at set intervals (ca. 20–30 cm). Depending mechanized drill, was the difficulty in monitor-
on the lengths of the cores, typically 3 m, this ing and recovering the sediment sample during
resulted in about a dozen sediment samples of the course of the coring operation. This particular
100–300 g each. These sediment samples were unit did not have a variable speed control on the
placed in plastic bags, marked and catalogued as drill, so sample was simply extruded from the
to location and time of recovery as well as depth borehole as the drill descended. This type of drill
within the individual core. The other halves of is to be avoided if controlled sampling is a goal
the cores were covered with mylar and aluminum of the study.
wrap for their entire lengths. The sampled halves Cryogenic or Freeze Cores: While relatively
will be treated the same way as well. Both sets of rarely used in coring studies today, more and
cores—sampled and unsampled—were more references to the use of these devices have
transported to laboratories for further study and appeared in a wide variety of literature, mostly to
subsequent storage in controlled environmental do with the sampling of wetland and/or
conditions and reserved for future research. Stan- inundated deposits. The freezing of soils
dard drill core boxes, made of heavy corrugated originated in the Soviet era (Khakimov Akh
Kraft cardboard or corrugated plastic, are avail- 1957). The author’s use of these devices has
able from most geological/soil science supply demonstrated some handling problems that are
houses. One box can hold up to 3 m (10 ft) of unique to them. One major difference in the core
core sample. taken by a freeze corer versus that recovered by
Rotary Augers and Drills: The core samples other types is temperature. Whether the coolant
from these devices have, by their very nature, used is carbon dioxide(78  C) or colder nitro-
lost pedological and lithostratigraphic informa- gen (197  C), the recovered core remains fro-
tion. The use of these devices is typically limited zen to the core barrel for some time—up to over
to shallow sediment/soil sampling (<2–3 m), but an hour in many instances. The core sample is
4.5 “1700 Sondages”: Geological Testing of the Plateau of Bevaix, Neuchâtel (Switzerland) 89

of box coring, is a commonly used sampler, in

fluvial, lacustrine/estuarine, and marine settings,
that is very versatile for all types of hard bottoms
such as sand, gravel, and clay. Shown in Fig. 4.7
is the Van Veen-type self-tripping sampler with
center hinged jaws and a spring loaded pin that
releases when the sampler makes impact with the
bottom. A simple metal pin prevents premature
As long as the grab sampler is hanging freely
from the hydrographic wire or line (Fig. 4.8b),
the trigger mechanism will keep the jaws open.
Fig. 4.7 The sediment grab sampler
However, as soon as there is slack in the winch
line, the trigger will be released. When the winch
rock hard until it finally melts, and even then the starts to raise the grab sampler, the jaws will
sample rarely loses its consistency and cohesion. close, thereby taking a “bite” (sample) from the
This is not a bad thing. Indeed it is one of the real bottom. Grab samples, brought aboard, are
advantages of the freeze corer that relatively lowered into a rectangular stainless steel or
uncompressed samples can be recovered intact wooden frame box that has a fine-to-very fine
over considerable lengths (>1 m/core). Too screen on the bottom (Fig. 4.9).
many times, in inundated sediments, even with
standard coring devices with sample catchers in
place, some or all of the samples have been lost
during recovery of the core sample. Freezing 4.5 “1700 Sondages”: Geological
cores do not suffer from this shortcoming. Testing of the Plateau
Another interesting aspect of a frozen core is of Bevaix, Neuchâtel
the necessity, upon occasion, to literally defrost (Switzerland)
the sample. This is done using a portable propane
torch. The objective is not so much as to thaw the A geological survey using 1700 sondages (“trial
core but to remove the frost that forms on the excavations”) was undertaken from 1991 to
surface of the sample obscuring the color of the 1997, along the construction right of way for
material. Once thawed, the battery of analytical the A5 autoroute, between Cortaillod and
techniques that can be applied to freeze core are Bevaix, supplemented by a palynological study
no different from those used on other core of the swamp known as the Bataillard, which has
samples. The removal of sample requires the allowed the reconstruction the paleoenvir-
cutting away of the core from the barrel to onmental history of the Plateau of Bevaix, since
which it adhered. This can be easily done with the withdrawal of the Würmien glacier until
a trowel where the length of the core is sectioned, today. This work has allowed a better under-
to the barrel, and the sediments simply peeled standing of the role of climate and human factors
back along the length. The sample can be trans- in the history of the landscape of the plateau.
ferred to core boxes or other sample containers. Figure 4.10 shows a typical set of these system-
atically deployed soundings across an agricul-
The Grab Sampler The collection of bottom tural field along the highway route.
and subbottom sediment samples sediment grab In addition to the large number of excavated
sampling is a well-used technique. Grab samples tests, the study dug 28 cross-sectional trenches to
are taken using easily deployed over-the-side characterize the relationship of the sedimento-
apparatus (Figs. 4.7 and 4.8a, b). The Ponar- logic and pedologic units to the underlying lithol-
type grab sampler, just mentioned in the context ogy. Across the plateau section, the paleosols
90 4 Techniques for Archaeological Sediments and Soils

Fig. 4.8 (a) Above, shallow sediment corer; (b) right, a sediment grab deployed at Jekyll Creek, Atlantic Intracoastal
Waterway (AIWW), Georgia, USA

Plateau of Bevaix study and nearby Neuchȃtel

(Hadorn 1994). The paleoecological, sedimento-
logical, and artifactual information must be col-
lected level by level, evaluated quantitatively, to
provide, at best, a fragmentary synthesis of a
prehistoric milieu. As Gall and others have
noted, the paucity of that fragmented record,
plus its inherent alienness, makes understanding
elusive and speculative (Gall 1983, 2012). None-
theless, it is the record we, as archaeological
scientists, must use, and the sedimentary fraction
is no small part of the picture. In geology, the
Fig. 4.9 Sediment grab sample on a screen box study of sediments is the field of sedimentology.
Because of the linkage between sedimentary
were observed to form on colluvium derived basins and hydrocarbons, there has been a strong
from either quaternary morainic or tertiary focus on this aspect of earth science both in terms
molasse units. The most prominent paleosol of theory and practice. Because of their impor-
was the “Atlantic period soil” (4900/ tance in the natural rock cycle, sedimentary rocks
4750–3500/3300 BC) (Weber-Tièche and and their genesis—mineralogy, depositional
Sordoillet 2008). In Fig. 4.11, this paleosol is histories, and environments—have been exam-
represented by complex 3. The study delineated ined to the extent that much information salient
what are termed biostartic versus rhexistatic to archaeological inquiry can be deduced from
periods when the landscape was either stabilized the sedimentary matrix or context in the site
by vegetation or destabilized by the lack of veg- (Soukup et al. 2008). Today, with the develop-
etation and other paleoenvironmental factors. ment of optical stimulated luminescence (OSL)
technology, and cosmogenic nuclides, 14C, 10Be,
and 26A1, the “chronometric” or “numeric” dat-
4.6 Analytical Procedures ing of soils and sediments is possible (Wagner
for Sediments and Soils 1998).
While the possibility of directly dating an
As most depositional environments, geological archaeological sediment facies is exciting, this
or archaeological, result in stratigraphic only adds to the other well-established
sequences or horizons such observed in the techniques used in the analysis of the sediment
4.6 Analytical Procedures for Sediments and Soils 91

Fig. 4.10 Typical

sequence of sondages used
to systematically
characterize the landscape
of the Plateau of Bevaix
(Courtesy of the
Archaeological Service,
Canton of Neuchâtel)

Fig. 4.11 Typical

stratigraphic sequence seen
in the profile for a test unit.
The “complexes” listed
represent sediment deposits
of colluvium and soils
formed thereon Complex 1

Complex 2

Complex 3

Complex 4

Complex 5

such as what I have called “the 5 Ps,” particle sedimentary particles are found in structural
size, point counting, palynology, phytolith, and materials—daub or brick—or ceramic artifacts,
phosphate procedures. As Julie Stein, Gall, and additional understanding of their sources, trade,
others pointed out, the study of sediments allows or other aspects of these materials can be studied.
us to deduce their “history” such as origin, trans- The study of those smallest of sediments—the
port, and deposition (Stein 1988). The specific clays—occupies a subdiscipline of its own and
study of a particle’s mineralogy (petrography of has proven quite pertinent to the study of archae-
sediment particles, particle size, and point ological ceramics. The mean size and percentage
counting) is discussed in the two following of sediments together with the calculation of
chapters. straightforward coefficients, such as (a) grain
Simply studying the morphology and variety size, (b) sorting, and (c) skewness, provide sig-
of sedimentary particles allows a rather reliable nificant information on transport, deposition, and
retrodiction of processes such as geological diagenetic processes. For instance, the nature of
source, climate, and diagenesis. If the cultural versus natural deposits in cave sediments
92 4 Techniques for Archaeological Sediments and Soils

Table 4.1 The standard grain-size scale for clastic

Name (mm) (μm) φ
G Boulder 4,096 12
R Cobble 256 8
A Pebble 64 6
V Granule 4 2
E Very coarse sand 2 1
L Coarse sand 1 0
Medium sand 0.5 500 1
S Fine sand 0.25 250 2
A Very fine sand 0.125 125 3
N Coarse silt 0.062 62 4
D Medium silt 0.031 31 5
Fine silt 0.016 16 6
M Very fine silt 0.008 8 7
U Clay 0.004 2 8
Fig. 4.12 The Hjulstrom diagram relating particle move- D
ment to size and stream velocity As devised by Johann A. Udden and Chester
K. Wentworth (Blair and McPherson 1999), the φ scale
can be examined by particle size, shape, and was devised to facilitate statistical manipulation of grain-
distributional (mean size, kurtosis, etc.) as well size data and is commonly used. φ ¼ –log2 D(mm) where
D is the grain diameter. It is a geometric scale where the
as other micromorphology procedures (Courty succeeding size class is twice the size of the former class
et al. 1989; Courty 1992; Goldberg 1995). or size grade
Bulk density and carbonate percentage are
important parameters determined in the labora- particles by chemical, mechanical, or ultrasonic
tory setting. The latter technique is important in means and their separation into size categories or
the characterization of sediments and soils where grades by sieving and sedimentation (Gee and
the buildup of calcium carbonate is common- Bauder 1986). As these particles range from
place and a key parameter in soil genesis as rocks to clays, various methods of their classifi-
well as sediment deposition. cation have been developed using arbitrary size
In fluvial geomorphology and sedimentology, classes or limits as illustrated in Table 4.1.
the Hjulstrom classification or diagram After color determination, grain-size determi-
(Fig. 4.12) is an important nomogram to under- nation is the most fundamental test performed on
stand the relationship between particle size, sus- soil/sediment samples. Three general size or tex-
pension, and stream velocity. The size of the tural classes are used by most investigations to
particle determines, in large part, whether a spe- show the textural divisions in a particular sample.
cific size range of bed materials will become These are plotted on ternary diagrams whose axes
suspended load or entrained in a current. are 100 % clay, silt, and sand (Fig. 4.1). In the
Table 4.1 lists specific size classes for United States, the standard protocol is to use the
particles: 1.0 is sand; 10 is small gravel, etc. US Department of Agriculture (USDA) classifica-
tion of the texture of sediments. In Europe, and
elsewhere, the particle size classes vary slightly
4.7 Particle Size Analysis from that of the USDA scheme. An in-depth rec-
onciliation of the various classifications is
Particle size analysis (PSA) is the measurement described elsewhere (Hodgson 1978). The USDA
of the size distribution of the individual particles terminology has gained wide acceptance outside
of a sediment or soil. It involves the destruction the United States, but even in countries such as
of dispersion of sediments into their constituent Belgium and the Netherlands, where the USDA
4.7 Particle Size Analysis 93

terminology is used for other soil features— (pedocals), iron oxides (pedalfers), clays, and
horizons, contacts, mineral content, etc.—these organic matter all disperse well in ultrasound
countries have set up their own particle size baths. Still Mikhail and Briner (1978) insist that
classifications (Supra, 54). While this heterogene- a stepwise pretreatment of a sample with H2O2,
ity looks to be the case for the foreseeable future, weak acid wash, and sodium hypochlorite satu-
the use of the ternary diagram to represent the ration before ultrasonic treatment produces the
variety of particle size classes is relatively a stan- best results.
dard practice. Figure 5.1 shows the USDA particle Mechanical sieving is the first step in PSA.
size scheme (a) together with examples from Screen sieve sizes have long been standardized,
(b) Belgium, (c) France, (d) Germany, and the problems and limitations inherent to this
(e) Switzerland, and (f) England and Wales. method are well understood (Day 1965). Reliable
Typical PSA methods require the sediment to sieving requires the use of mechanical shakers
be broken down in to its individual particles by that are sets and forgets such that consistent
both chemical and physical means. This two-part results can be expected. Pansu et al. (2001) pro-
procedure is necessary in most sediments where vide a comprehensive review of grinding and
clay minerals comprise a significant portion. The sieving equipment in their text on soil analysis.
larger size particles (>2 mm) are sorted and If one does not use ultrasonic dispersion, sedi-
sized by the use of wire-mesh screen sieves in mentation techniques follow most sieving using
mechanical shakers (Fig. 4.9). Of course, the pipette or hydrometer methods. The settling of
sediment must be dry before either mechanical small particle grains such as the clay minerals
or chemical sorting is to be attempted. The ease follows Stokes’ law in the form:
of sorting or dispersion depends on the nature of
the sediments. Because of this fact, there is no v ¼ gðρs  ρi ÞX2 =18η
one “standard” PSA technique.
Sediments rich in organics and clays are clas- where v (mm/s) is the settling velocity of the
tic and require substantial chemical pretreatment particle; ρs (kg/m3), the particle’s density; ρi
with agents such as sodium hexameta-phosphate (kg/m3), the liquids density; X (mm), the particle
(HMP), hydrogen peroxide, sodium hypochlo- size; and η (kg/s/m), the viscosity of the liquid.
rite, sodium hypobromite, and potassium per- g (9.8 m/s2) is the gravity. Stokes assumed
manganate. Acid treatments are generally to be particles to be spherical, an assumption that
avoided because these reagents, such as HCl, can does not affect the results too drastically given
destroy the lattice of clay minerals. HMP is the that many particles do not meet this criterion,
most commonly used dispersant chemical and is notably the clays. The diameter of the particle,
recommended in microbotanical separations of and hence its size, is obviously more critical than
pollen and phytoliths (Yaalon 1976; Rovner sphericity. To do PSA of the sediment fraction,
1996a). two methods or apparatus are generally used:
For small-size particle aggregates such as the (1) a pipette method and (2) a hydrometer
clays, ultrasound is used to cause dispersion. method. Both have variants and are described in
Ultrasonic baths operate on the transmission of detail in most works on PSA technique. Much
the high-frequency sound waves through the sed- like one’s choice of dispersal techniques, the
iment or soil suspension. Bubbles form and burst specific method is the one that most effectively
creating a “sonic cavitation” that disperses the accomplishes the goals of one’s analysis of a
most highly aggregated particles without damage specific type of archaeological sediment. All are
to the individual grains. Ultrasound can be used somewhat time consuming with typical times of
without chemical dispersants. Simple 20 h for clay dispersal (Rockwell 2000).
suspensions of distilled water are routinely Examples of each sedimentation method are
used. Sediments or soils rich in calcium presented below.
94 4 Techniques for Archaeological Sediments and Soils

4.7.1 Hydrometer Method is the sedimentation pipette method as described

by Day (1965) and Green (1981). The sample
This technique is commonly utilized to assess the sediment suspension is placed in a 1 l cylinder
clay fraction of sediments samples. The tech- after 10 ml of HMP has been added together with
nique allows for rapid, multiple measurements distilled water to make the 1 l volume. Using a
from the same solution/suspension. The standard Lowy pipette, a 10 ml sample is withdrawn from
hydrometer used in the United States is ASTM the cylinder at a depth of 200 mm at set time
152H model. The upper stem of the hydrometer intervals. A typical sequence, suggested by Catt
is graduated to determine the settling depth, h, and Weir (1976), is as follows:
for particles of diameter, X. The settling depth
(mm), h, is directly related to the graduated 57 s ¼ 0.063 mm or 4 φ fraction
reading, R, (ml) on the hydrometer stem. For an 3 min 48 s ¼ 0.031 mm or 5 φ fraction
ASTM 152H, the empirical calculation is 15 min 12 s ¼ 0.016 mm or 6 φ fraction
1 h 48 s ¼ 0.008 mm or 7 φ fraction
h ¼ 0:164R þ 16:3 4 h 3 min 12 s ¼ 0.004 mm or 8 φ fraction

To calculate X, the variant of Stokes’ law used is

These aliquots are drained into pre-weighed
X ¼ θt1=2 aluminum evaporating dishes. These weighing
dishes are placed in an oven at 105  C to dry.
t ¼ temperature ( C), and θ ¼ 1,000 (Bh) ½ The dishes are then weighed and the weight
where B ¼ 30η (ρs  ρi); the units are discussed percentage per fraction calculated. The silt size
above. fractions are determined similarly using
One can continue to obtain the concentration established settling times from tables. Pearson
of the clay fraction in g/l from the simple calcu- et al. (1982), in a successful study of inundated
lation C ¼ R  RL where RL is the hydrometer archaeological sediments, used the following
reading of a blank solution. These readings are particle size analysis on an average 150 g sample,
taken at the same time intervals (e.g., 1.5 h, 24 h, following Folk (1968), which were analyzed in a
etc.). The term, summation percentage, P, is two-step procedure—coarse fraction then fine
equal to C/Co  100 where Co is the ovendry fraction:
weight of the total sediment sample. For
Course Fraction (Steps 1–7)
instance, the percentage of the clay fraction in
the sample is given by
1. Coarse fraction samples soaked in water for
Pclay ¼ m lnð2=XT Þ þ PT 24 h to disperse clays then dried (air).
2. Weigh sample.
T is the time interval and m is the slope of a 3. Sample soaked, in water, to further disaggre-
summation percentage curve between two time gate clays.
intervals T1 and T2. The hydrometer may be used 4. Sample wash through a 4 φ (0.067 mm) wet
to determine the sand fraction percentage which, sieve screen to remove all the fine (silt and
plus the clay percentage, may be subtracted from clay) fraction.
100 to obtain the silt percentage in the sample; % 5. Sample dried and weighed.
silt ¼ 100  (% clay + % sand). 6. Subtraction of this weight (5) from the
prewash sample weight yields the fine frac-
tion weight.
4.7.2 The Pipette Method 7. Depending upon the particular size classes of
interest, e.g., five to eight 1 φ divisions,
The method most commonly used to determine further separation of the coarse fraction is
silt and clay size fractions, in a sediment sample, done by sieving.
4.7 Particle Size Analysis 95

Fine Fraction (Steps 8–11) remainder of the liquid remaining in the fleaker,
after the respective size samples have been with-
8. The fine fraction sample is dispersed with drawn, is rinsed into a 63 μm sieve. Any sand
calgon at a concentration of 2.1 g/l. recovered in the sieve is rinsed with deionized
9. The sample is placed in a 1,000 ml beaker water and placed in a container to be dried over-
and vigorously stirred. night as the other samples. The containers are
10. Pipette samples, with a standard (Lowy) then weighed to 0.0001 g and the fractions com-
pipette, are drawn at various times (up to pared to the particular particle size scheme used
8 h) at various withdrawal depths for a total in the procedure (U.S. Department of Agriculture
of seven withdrawals/samples. or GAO) (Fig. 4.1).
11. Each withdrawal is dried, weighed, and
analyzed at selected size division (e.g.,
4 φ–100 φ).
4.7.4 The Imhoff Cone Method

Volumetric Procedure: Fill an Imhoff cone to the

1 l mark with a well-mixed sample.
4.7.3 The Modified Pipette or Settle for 45 min.
“Fleaker” Method Record volume of settleable solids in the cone
as ml/l.
A new procedure, reported by Indurante Sediments: Fill the Imhoff cone to the 1 l
et al. (1990), has the advantage over the standard mark with a well-mixed sample of 100 g.
pipette method described by Day (1965) and The size fractions—sand, silt, and clay—can
Green (1981). The principal difference is the be estimated using the graduations on the cone.
use of a single vessel, called a “fleaker,” to dis- Settle for 0.25–0.50 h, regardless of soil tex-
perse and pipette the sample. The fleaker is a ture. Sofka et al. (1992) observed that even for
300 ml vessel that is a combination of the beaker clayey soil, settling is fairly rapid, since most of
and Erlenmeyer flask forms. Aliquots are taken the suspension occurred as aggregates rather than
from the fleakers in a manner similar to that individual primary particles. Furthermore, the
described above. These are dried and weighed amount of additional soil settling out of suspen-
to arrive at estimates of the silt and clay factions sion from 0.5 to 1.0 h is negligible (1–2 %) and
in the following manner. causes no serious errors if a 0.50-h settling inter-
Five grams of soil is placed in the 300 ml val is inadvertently exceeded (Carter et al. 1985,
fleakers that are tared to 0.001 g. Five ml of unpublished data).
sodium metaphosphate solution (50 g/l) is Sediment remaining in suspension after 0.50 h
added to the fleaker with a pipette. The mixture is the nonaggregated clay fraction (Gee and
is then diluted to the 100 ml level, and the fleaker Bauder 1986) (cf. Fig. 4.13). The entire contents
stoppered and put on an Eberbach shaker for of the Imhoff cones are then decanted, with rins-
8–12 h at slow speed. After shaking, the mixture ing, the volume of sediment in each sample can
is diluted further to 250 ml, stoppered again, and be measured, and each sample dried in an oven at
vigorously shaken by hand for 1 min. The sample 60  C, weighed, ashed in a furnace at 500  C, and
is then allowed to settle for a 1.5–2 min interval. weighed again to determine both ash-free dry
A 20–25 ml aliquot is withdrawn at the 5 cm mass and the mass of inorganic matter in each
depth with the time noted. The mixture is, thus, sample. The mean of replicates of both granular
repeatedly sampled in this manner with the (sand fraction) and nongranular (silt-clay)
respective samples being drained into weighing fractions is used to calibrate the dry-weight
cups, typically of lightweight aluminum, that are densities of each fraction. Still, after settling,
tared to 0.001 g and placed in a drying oven. The the volume of granular and nongranular sediment
96 4 Techniques for Archaeological Sediments and Soils

Fig. 4.13 Imhoff cone and

soil auger sample.
Suspected B horizon shown
separating in Imhoff cone

in the cone can be assessed because this interface standard parametric statistical estimators to be
is generally obvious, separating sand from clay- applied to the data they produce. These
silt fraction. estimators include mean, variance, and standard
In many respects, the use of the Imhoff cone deviation together with parameters such as skew,
mimics a settling column. Its principal sorting, etc. Some of these measures will be
advantages are its ease of use, portability, and discussed in some depth in Chap. 8.
speed. The Imhoff cone method cannot replicate
the settling column nor fleaker methods for exac- Point Count Analysis (PCA)
titude in determining particle size fractions Point counting analysis is time-honored tech-
nique developed in sedimentary petrography for Statistical Parameters and PSA the determination of the proportions of various
Particle size analysis and all of the other four minerals in a thin section of a rock specimen
techniques, presented in this chapter, require (Carver 1971). The assumption is that the results
4.7 Particle Size Analysis 97

of a point count will accurately reflect the per- of random processes at play in all populations.
centage of a particular component—feldspar, About PCA, much ink, and rightly so, has been
quartz, pyroxene, pyrite, etc., within the sam- spilled over the appropriate procedures to be
ple—and, to a larger degree, the percentage of used in arriving at reliable estimates of the
that mineral (or particle size) within the source constituents of rock or sediment aggregates as
from which the sample was taken (the Delesse counted from a thin section or facsimile thereof.
Relation). PCA as implied was developed by These discussions extend down to the grid type,
petrographers to assess the mineral content in size, and microscope stage to be used. The aim is
whole rocks using what is a sampling technique. the same—count enough “points” on the slide to
Shackley (1975) has applied point counting to estimate the target population correctly.
archaeological sediments. In PCA studies, much What is the sample size, n, that best estimates
discussion is given to the number of grains that grains smaller than a cobble and larger than a
should be counted. This is to be expected as PCA clay? Assuming that most samples are either rock
is a sampling, n, of a larger population, N, or sediment specimens with grains of sand size or
wherein common statistical estimators such as less, then the grain count will fall somewhere
average, mean, standard deviation, and variance between 100 and 1,000. Sample counts of
are used to characterize the relationship of the between 200 and 400 grains are fairly standard
sample to the population. Chapter 8 will discuss in studies of foraminifera, pollen, and sedimen-
statistical procedures commonly used in archae- tary rocks (Martin 1970; Bryant and Holloway
ological geology rather than embed that discus- 1983; Krumbein and Pettijohn 1938). Krumbein
sion in this topic. Without diverging too far from (1938) has graphically illustrated the estimation
this goal, it is, perhaps, appropriate to discuss the of probable error in varying sample sizes used in
relationship between particle size and that of the PCA (Fig. 4.3). According to that analysis, it can
sample size to be counted, n. It is almost a mantra be seen that the probable error stabilizes between
in statistics, that is, as one increases the sample 5 % and 15 % at about 300 grains. It is easy to see
size, n, the approximation of estimators for N, the why this sample count is so commonly used in
population, is more assured by the Law of Large pollen studies.
Numbers (cf. Chap. 8). In microbotanical (pollen, phytoliths) and
One cautionary side of this is that the result microfaunal (forams) studies, the sample is dis-
one obtains is uninformative as it represents esti- persed, in solution, across a counting slide which
mator values that are so “average” as to be of no can be gridded or can be covered by a glass
real value. This is to say that if one counts coverslip, appropriately divided for counting
enough of anything—mineral grains, pottery purposes. Such standardized items exist for
sherds, people, etc.—the statistical estimates of these particular analytical procedures. In geology
those large samples will be so homogenous as to the most commonly point-counted sample is the
unreflective of important subpopulation rock thin section. More often today, the
differences such as the difference in igneous sediments examined by PCA are in their section
rock types—basalt versus rhyolite, phases of pot- mounts as well (Goldberg 1999). Still, it is not
tery, and different cultural groups. This is a uncommon to see sediments dispersed, as evenly
conundrum present in all sampling studies— as possible, onto the gridded card or glass slide
what value of n is large enough to detect impor- covered with a mounting spray cement (Gagliano
tant population differences yet not so large as to et al. 1982).
mask the very differences we seek to detect and Most PCA is done using stereomicroscopes of
reliably estimate? Another way to express this relatively low-power (7X–30X) magnification.
problem is to cast it in terms of randomness Depending upon the specific goals of the PCA,
versus nonrandomness. Too large a sample, n, the microscope can be either petrographic or not.
allows for the suppression of nonrandom trends If mineralogical identification is important, then
or differences such that the results are reflective a petrographic capability is required. This would
98 4 Techniques for Archaeological Sediments and Soils

be the case in studies of archaeological stone packages at fee-service laboratories. Of these the
such as marbles, granites, or other building or organic content can be determined in most
artifactual rock. It would be useful in the identi- laboratories without too much difficulty. The
fication and characterization of temper found in mineral oxides are generally done with analytical
earthenwares. For instance, in a thick (100 μm) methods (XRF, etc.) that may or may not be
thin sections, one should readily differentiate the readily available in most labs. A discussion of
quartz and feldspars on the basis of bright these analytical methods is presented in Chap. 7.
second-order interference colors under crossed
polars (XPL). At standard thin section thickness, Total Organic Carbon (TOC) Total organic
30 μm, polarizing filters allow the full gamut of carbon is the sum of organic carbon in the soil.
mineralogical identification (cf. Chap. 6). When While new instrumental methods have been
examining slides for pollen and the smaller developed in recent years, most soil testing
phytoliths, polarized microscopy is not required laboratories routinely use one or the other the
and, in most cases, would impede the PCA. This following: (1) wet digestion and (2) colormetric
is because the transparency of the grains is (Soil and Plant Analysis Council, Inc 2000).
important to making a reliable identification The wet-oxidation digestion method uses
based on particle shape and texture (Rovner potassium dichromate (K2Cr2O7) with external
1996b). A raster-type microscope stage (X–Y heat (hot plate) and back titration to measure
direction) is used to move the appropriate grid the unreacted dichromate which serves as the
distance across the slide, counting grains at the proxy for the carbon content. Typically
grid intervals, until the selected grain sample 0.1–0.5 g. of soil is added to a 500 ml Erlenmeyer
count (n) is reached. along with 15 ml of dichromate reagent. The
As Van der Plas and Tobi (Van der Plas and mixture is then boiled for 30 min. After cooling,
Tobi 1965) have pointed out, in their discussion three drops of the indicator solution,
of the reliability of point counting results, it is n-phenylanthranilic acid, is added and the mix-
determined not only by the number of grains ture titrated against Mohr’s salt solution, ferrous
counted but on the distance between points as ammonium sulfate, until a color change from
well. Their point is that there is an obvious rela- violet to bright green is observed. The % organic
tionship between a sample’s grain size and the matter ¼ {[meq K2Cr2O7 – meq FeSO4  0.3]/g
point distance chosen for its modal analysis. For soil}  1.15. Here meq. is milliequivalents.
instance, it would be patently foolish to choose a The colormetric procedure uses the property
point counting distance of 0.2 mm if all or most of absorbance as discussed, in detail, below, in
of the grains were 2 mm in diameter. The grain/ reference to phosphorus determination. Here,
point sizes will vary, and therefore, the point 2.0 g of soil is placed in a test tube to which
counting protocol must be flexible to match ana- 20 ml of sodium dichromate solution is added.
lytical requirements of the specific case. Van der The reaction is exothermic, and after the mixture
Plas and Tobi (1965) stress that “the point dis- is cooled (40 min to 1 h), it is allowed to stand for
tance chosen should be larger than the largest 8 h. Aliquots of the clarified solution are then
grain fraction that is included in the analysis.” transferred to the spectrometer for absorbance
measurements at 645 nm wavelength (supra,
178). The values so obtained, when plotted in
4.7.5 Organic Content Determination comparison to standard curves, yield the organic
Methods content.

Other important parameters—TOC (total organic Loss on Ignition (LOI) This method has
carbon) and LOI (loss on ignition) along with become comparable in reporting to that done by
mineral oxides (FeO, MgO, etc.) together with the methods just discussed. It is equally effective
water content—are provided in most analytical for the analysis of soils containing low to very
4.7 Particle Size Analysis 99

high organic matter (OM) (supra). It has a sensi- straightforward manner of basic mineral determi-
tivity of 0.2–0.5 % organic matter. A sample of nation, of the introductory geology laboratory,
5–10 g soil weight (W) is placed in an ashing the use of hydrochloric acid is required. In this
vessel. The vessel is first placed in a drying oven more rigorous determination, the HCl release of
at 105  C for 4 h. Once cooled, the sample is CO2 gas is collected and measured by the use of
weighed to the nearest 0.01 g and then placed in a sealed flask and graduated column apparatus
muffle furnace at 400  C for 4 h. The sample is called the Chittick method (Singer and Janitsky
removed, cooled, and weighed again. To calcu- 1986). The evolved gas displaces a fluid (colored
late the % organic matter, one uses the formula: water) into the graduated column. This value is
divided by the original sample weight to deter-
%OM ¼ ½ðW105  W400 Þ  100=W105 mine the carbonate content generally reported as
percent (%) (supra). Bulk Density Phosphorus Analysis (PA)

Bulk density is the average dry weight per unit As Joseph Schuldenrein (1995) states “P (phos-
volume of soil. Depending on the nature of the phorus) appears to be the most generic indicator
sample, fine or coarse grained, the principal of anthropogenic input. . .” for archaeological
method is to determine the matrix bulk density by sediments. The importance of phosphorus analy-
addition of paraffin and subsequent measurements. sis in archaeology was first recognized by
The sample is first dried and weighed then soaked Arrhenius (1931), but it was R. C. Eidt who
in paraffin, dried, and weighed again. This established the systematic use of the phosphate
determines the weight (and volume) of the paraf- method in evaluating anthropogenic sediments
fin. Finally, the sample is immersed in water and (Schuldenrein 1995). Eidt, his co-workers, and
weighed. Bulk density, in g/cm3, is calculated thus: students established the processes, chemistry and
Bulk density ¼ ðρw ÞðWCd Þ=WCp  WCpw chronology of phosphorus, as phosphate
   compounds, in archaeological sediments (Eidt
þ W p ðρw Þ=ρp 1973, 1977, 1985; Sjöberg 1976; Woods 1977).
The assumption is that the magnitude of human
where ρw is the density of water; WCd is the dry
impact on a landscape can be demonstrated by
weight in air; WCp is the waxed weight in air;
the total phosphorus concentration such that
WCpw is the waxed weight in water; Wp is the
archaeological sites and inter-site activity areas/
weight of the paraffin; and ρp is the density of the
features can be isolated. Phosphorus occurs in
sediments as either bound (non-labile) or
For fine-grained samples (size <<2 mm), the
unbound/free (labile) phosphate compounds.
procedure is straightforward. With coarser, gravel-
The labile forms correspond to Al and Fe
rich samples, some error can be introduced by the
phosphates. As we have discussed, bound or
size of the large clasts. A sample should be broken
occluded forms of Al and Fe phosphates develop
apart to determine the amount of these clasts
as inorganic compounds in soil horizons.
within a sample. As a result, the bulk density so
Inorganic compounds are (a) those containing
determined is more an average value and less a
calcium and (b) those containing iron and alumi-
specific value for an individual sample. Elevated
num. The solubility of inorganic calcium
bulk density values are typical of older soils with
compounds increases with the increased simplic-
significant clay and carbonate volumes (supra).
ity of the compound, but the simpler compounds
are only available in very small quantities, and Carbonate Content they easily revert to insoluble states. The fixation
The amount of calcium carbonate in aridisols, of phosphates, when the pH value is in the range
marine, and lacustrine sediments can be signifi- of a little less than 8.0 and 8.5, is mostly in the
cant and must be accurately measured. Like the form of calcium phosphates. In alkaline soils,
100 4 Techniques for Archaeological Sediments and Soils

phosphates will react with the calcium ion The method extracts total phosphorus, i.e., both
and with calcium carbonate. The next step organic and inorganic phosphorus compounds.
includes a conversion of the calcium phosphate The accuracy level of the method of the analysis
(Ca3 (PO4) 2) to insoluble compounds, such as ranges between 95 and 98 % recovery. All
hydroxy-, oxy-, carbonate-, and fluorapatite samples are dried overnight in an oven at a tem-
(insoluble apatite compounds are formed at pH perature of around 50  C. Samples consisting of
levels of above 7.0) compounds, the latter being very fine particle sizes are ground directly after
the most insoluble compound known. An exces- having been dried, while samples with coarser
sive amount of calcium carbonate (CaC03) particle sizes have to be screened. The choice of
produces very insoluble compounds. At a pH sieve size was made so that the sieved sample
level of between 6.0 and 7.0, phosphate fixation would have the same particle size as the ground
is at its minimum. one. A half gram of each sample is put into a
It is not within the scope of this book to fully graduated digestion tube, to which is added 2 ml
describe the extremely complex and incom- of perchloric acid (HClO4) and nitric acid
pletely investigated phosphorus cycle in (HNO3), respectively.
sediments. It is important to mention the poten- The tubes are then placed in an aluminum
tial sources of both anthropogenic and heating rack under a perchloric acid fume hood.
non-anthropogenic sources such as geology and The HNO3 is boiled off at a temperature of
soils. After mineral sources in sediments, and 170  C. After a stepwise increase to 225  C, the
soils, the bulk of phosphorus is from organic remaining HClO4 is left to boil for 1 h. The tubes
matter such as dead plants and animals that are checked regularly to prevent being boiled
decompose in the soil. Phosphorus is an impor- dry. After the digestion, the volume of each
tant energy source in organisms particularly as a tube is raised to 50 ml with distilled H2O, stirred,
component of the ATP (adenosine triphosphate) and left to rest for a period of at least 12 h. A
molecule. As a result, all plants and animals stock solution is made consisting of (a) 9.0 g
contain significant amounts of phosphorus in ammonium molybdate ((NH4) 6Mo7O24.4H2O)
their tissues which is incorporated into sediments and 0.2 g potassium antimony tartrate (K(SbO)
after their deaths. In addition, humans and C4H4O6.1/2H2O), which are mixed in 500 ml of
animals excrete large amounts of phosphorus as H2O) and (b) 65 ml sulfuric acid (H2SO4) and
both feces and urine. Plant and animal decay, 335 ml of H2O. Initially mixed separately, the
burials, refuse coupled with urine, and excrement (a) and (b) reagents are mixed together and vol-
increase the phosphorus content of soils above ume raised to 1,000 ml with H2O.
ambient levels. The accumulation of this phos- A primary standard solution, with a concen-
phorus in anthropogenic soils, if it can be fixed tration of 1,000 μg/g P, is made by dissolving
into insoluble forms, can have long residence 4.387 g potassium phosphate monobasic (KH2
times—millennia in many cases—and provide a PO4) in sufficient H2O to yield 1,000 ml. A
relatively reliable geochemical marker of archae- secondary standard solution, with a concentra-
ological interest. tion of 10 μg/g P, is made by adding 40 ml
HClO4 to 10 ml of the primary standard solution
and dilution with H2O to a volume of 1,000 ml.
4.8 Determination of Total Working standard solutions ranging from 1 to
Phosphorus by Perchloric 6 μg/g P are made from the secondary standard
Digestion solution. The color-developing solution, freshly
made for every batch of 100 samples, consists of
Information on this method is found in Eidt and 2.0 g ascorbic acid mixed with 200 ml of the
Woods (1974) and Woods (1977). The methodo- stock solution and raised to 1,000 ml with
logical procedure in the chemical analysis is H2O. Five milliliters of the color-developing
described through the steps noted in Fig. 4.4. solution is added to 1 ml of each of the working
4.9 Absolute Phosphate Analysis Versus Qualitative Color Tests 101

standard solutions as well as to 1 ml of the Intercomparison of their data—relative color

samples. After 1 h, their absorbencies are read intensity and the absolute colorimetric determi-
on a colorimeter at a wavelength of 740 nm. The nation—showed similar findings in the measure-
absorbance of the standards is plotted against the ment of site extent (Fig. 4.14).
known concentration. Like Sinclair (supra), Barba used the phos-
The chemical analysis, as just described, is phate spot test. On house floors at Manzanilla
accurate and has good reproducibility. A linear (2007), Barba and Lazos (2000) detected nonran-
regression of P content and absorbance gives a dom, linear patterns on floors in the phosphate
correlation coefficient (r) of 0.9998 (cf. Chap. 10 data that suggested heavy traffic on stairs and
for a detailed discussion of linear regression). The around the altar (Barba 2007).
regression line is a linear relationship between Cavanagh et al. (1988) utilized a “cheap,
absorbance and concentration judging from the robust” colorimetric phosphate procedure devel-
r value of 0.9998 and the straight line. Beer’s law oped by Craddock et al. (1985) to assess site
states the following: “the fraction of light absorbed parameters at Laconia, Greece. Craddock
is directly proportional to the concentration of the et al. (1985) developed the molybdenum blue
colored constituent.” The phosphate determination method to analyze multiple samples in a rela-
conforms closely to Beer’s law: the slope of the tively short time. Like the simpler, qualitative
regression line is just below optimal linearity as is tests used at Great Zimbabwe, the procedure
expressed by the regression equation y ¼ 0.0957 used at Laconia examined the total phosphorus
x + 0.0028. A direct proportionality, as stated by present without the extraction of the bound
Beer’s law, would mean y ¼ 0.1x. (P2O5) “total” component as accomplished by
the perchloric acid technique described. The
molybdenum blue method is described as
4.9 Absolute Phosphate Analysis follows: (1) Beginning with excavated/cored
Versus Qualitative Color Tests 10 g soil/sediment samples that are (2) air-dried
and screened through 1 mm metal sieves, (3) 1 g
Total phosphorus-bound fraction analysis versus subsamples are digested in (4) 5 ml of 2 N HCl
those of labile or less bound phosphates has little boiled for 10 min. (5) A 0.2 ml aliquot
or no comparative value. The latter phosphates is removed and 10 ml of molybdenum blue
can be qualitatively detected by methods similar reagent is added, and (6) the liquid is analyzed
to the familiar litmus strip test for acidic/basic with a colorimeter standardized to zero
solutions. In the qualitative phosphate test, one optical density with a phosphorus-free reagent.
measures the intensity of color produced in a spot Like the perchloric approach, a set of eight or
test on paper or in a small tube. While somewhat more reference samples of known concentrations
subjective and decidedly not quantitative, the are used to produce a calibration regression
color test has economy and ease of application line. If percent transmittance (T ) is measured,
in a field setting on its side. At the site of Great this can be converted to optical density by the
Zimbabwe, researches from Lund University formula:
used both quantitative, bound phosphate analysis Optical density ¼ log10 T
and the field color tests with good effect (Sinclair
1991, 1996, Personal Communication; Sinclair The molybdenum blue reagent is mixed from
and Pétren n.d.). The Swedish team, in extensive 65 ml H2SO4 (6N), 37.5 ml ammonium molyb-
coring tests, extracted phosphates in column date (40 g/l), 12.5 ml potassium antimonyl tar-
samples to measure the extent, vertical and hori- trate (2.743 g/l), and 1.32 ascorbic acid (solid) as
zontal, of the famous African Iron Age site. suggested by Murphy and Riley (1962).
102 4 Techniques for Archaeological Sediments and Soils

Fig. 4.14 Phosphate intensity plot for Great Zimbabwe using the colormetric method (Courtesy of Paul Sinclair)

1992). Pretreatment with an acid (sulfuric, nitric)

4.10 Colorimetry and heat provides the conditions for hydrolysis of
and Spectrophotometry the inorganic fraction: (a) organic phosphates are
converted to orthophosphate by heating with acid
For the geoarchaeologist, there exist relatively (acid oxidation) and persulfate (b).
“quick and dirty” methods for ascertaining total
phosphorus without recourse to elaborate labora- (a) Na4 P2 O7 þ 2H2 SO4 þ H2 O
tory techniques such as perchloric digestion. This
latter technique is essential for a thorough, quanti- ! 2H3 PO4 þ 2Na þ 2SO2
tative determination of the bound and “condensed”
phosphate fractions. One other component of the (b) R  O  P  O  R0 þ K2 S2 O8 þ H2 SO4
total phosphorus is that of the labile or orthophos- ! H3 PO4 þ 2Kþ þ 3SO2
phate fraction. This is helpful to archaeologists as it
can be readily determined in the field.
The most common procedure is acid hydrolysis In (b) R and R0 represent various organic
of the organic and condensed in organic forms groups. If one determines the total phosphorus
(meta-, pyro-, and other polyphosphates) to reactive by the second method (b), it can then be followed
orthophosphate (Hach Water Analysis Handbook by the reactive phosphorus test. By subtracting
4.11 Micromorphology: Describing Archaeological Sediments and Soils with the Microscope 103

that amount from the total phosphorus result, one structure of spatial patterns in natural and non-
can determine the organically bound phosphate natural layers using microscopic techniques.
which is a good indicator of cultural deposits of Goldberg (1995) has used resin impregnation
archaeological interest. The analysis of reactive of soil/sediment blocks as a precursor to produc-
phosphorus—the orthophosphate—can be done ing thin sections of the clastic materials for use in
using a field spectrophotometer. Orthophosphate micromorphology. Undisturbed sample soil/sed-
absorbs at the 890 nm wavelength, in the Hach iment blocks are taken from the field profile.
field spectrophotometer, and produces reliable These samples are then impregnated, in a vac-
measurements, in mg/l of PO43. Using the uum chamber, with epoxy resin. In this matrix,
same device, one can assess the total phosphorus after hardening, the sample can be treated like a
(PO4) using either the acid persulfate digestion rock sample and sawn to produce viewable thin
method or the ascorbic acid method (Hach water sections of standard dimensions. This procedure
analysis handbook, 2nd edn 1992). is finding increasing application in the
Colorimetry and spectrophotometry rely on geoarchaeological studies of site deposits
straightforward principles where the color of a (Courty, Goldberg, and Macphail 1989). At the
molecule in solution, such as phosphorus, thin section scale, undisturbed, oriented samples
depends on the wavelength of light it absorbs. are examined by the microscope—petrographic
When a sample solution is exposed to white light, or otherwise (SEM; CL) (allowing for observa-
certain wavelengths are absorbed, and the tion of composition in mineralogical terms, tex-
remaining are transmitted to the eye or a photom- ture, and fabric) the geometric relationships
eter. In our discussion of the determination of among constituents (supra).
total phosphorus by perchloric acid digestion, Courty describes soil micromorphology as the
the final measurement is to test absorbance in a integrated use of various microscopic techniques
light cell where the absorbance and concentra- to study the arrangement and nature of the
tion of the unknown is determined from an absor- components that comprise sediments and
bance versus concentration curve, for the soils (1992).
element of interest, in our case, P, constructed She, particularly, discusses the micromor-
by the use of standards. The absorbance, A, is phologic study of sediments and soils as an
generally defined as the log (Po/P) where Po ¼ essential component of environmental recon-
incident light intensity and P ¼ intensity after struction. Likewise, it helps identify the various
absorption. Using Beer’s law, A ¼ a∙ b ∙ c, where kinds of domestic and specialized activities that
a ¼ molar absorptivity (%), b ¼ path length are involved in the formation of living floors and
(μm), and c ¼ concentration (g/ml). We hold a other anthropogenic surfaces (supra). Microstra-
and b constant, and measure c concentration by tigraphic changes can reflect changes in deposi-
the spectrometer. tional processes.
Additionally, soil micromorphology has been
applied to the study of stone tools (Fladmark
1982). Microdebitage is defined as all stone flak-
4.11 Micromorphology: Describing ing residue less than 1 mm in maximum dimen-
Archaeological Sediments sion (Susino 2010). Lithic microdebitage is
and Soils with the Microscope produced during lithic tools and rock engraving
manufacture. For tools or rocks containing
Kubiëna (1938) first introduced the study of quartz, small particles persist as microwaste
undisturbed soil samples using microscopic initially on the ground surface but later as
techniques. He became the founder of soil micro- particles that are incorporated into sedimentary
morphology and published in 1938 his book deposits associated with the occupation site.
Micropedology. Soil micromorphology can be Experimental replication indicates that it is pro-
defined as follows: the in situ study of the duced in great quantities by stone tool
104 4 Techniques for Archaeological Sediments and Soils

manufacture and can permeate site matrices as a (1982) as totally removed from their original
permanent signature of past cultural activity. settings and, in many cases, reused for other
If microwaste is distinguished from naturally purposes. The most obvious example of this
occurring sediments, then microwaste and/or its would be the digging of a berm or terrace
surrounding sediment provides a mechanism for where fill is “robbed” from its original deposition
establishing lithic use in archaeological deposits. locale and redeposited in another.
Identification of microwaste is undertaken by Some skill in petrography is suggested by
comparing the shape of microwaste particles Courty (supra). This is illustrated by Angelucci’s
with those from surface sediments, taking partic- study of knapped lithic materials in thin section
ular note of angularity and irregular shapes, and (Angelucci 2010). In contrast to other types of
by analyzing quartz or other toolstone grain cultural materials, microdebitage is usually not
morphologies. On the basis of surface features weathered to any degree. From a micromorpho-
identified on experimental microwaste, particles logical standpoint, micromaterials, such as lithics,
identified as microwaste can be separated from fall into the category of “coarse” components. As
the sedimentary grains (Susino, supra). Scanning Angelucci points out, Courty et al. (1989) con-
electron microscopy (SEM) and light micros- sider lithic materials as a “mineralogic” compo-
copy are used to determine surface differences nent. For instance, “Flint is easily distinguished by
between pedogenic grains and anthropogenic its petrographic properties, which include low
material from archaeological deposits. birefringence and crytocrystallinity. . ..and is
It has been noted that microwaste may lie exceedingly hard to distinguish natural from
within or close to the age boundaries of the worked fragments. . .” (Courty et al. in Angelucci
surrounding strata/sediments but not necessarily 2010). This is a bit of understatement as flints or
where larger lithic material is found. Microwaste cherts are non-birefringent to the point of optical
is incorporated into sediment and may not be extinction in most petrographic settings. This
subjected to the same degree of movement as caveat, notwithstanding, several authors have
larger lithic waste and stone tools within the reported identification of various lithic artifacts
deposit. Indeed, Baker (1978) noted smaller in thin section (Angelucci, supra; Bergadà
artifacts from the surface are proportionately 1998), most notably chert or quartzite. The present
fewer (53.3 %) compared to the proportion of author has identified toolstone microdebitage
smaller artifacts within a total cultural deposit from siliceous sandstones and arenites (Garrison
(61.5 %). One possible reason for this, per et al. 2012a).
Baker, size effect, is that smaller artifacts might Figure 4.15 illustrates the utility of micromor-
tend to be trampled into the sediment or soil and phological study of Neolithic settlements. Micro-
larger artifacts are not resulting in an increased morphology was used extensively at the site of
amount of “micro” materials in comparison to Arbon-Bleiche 3, a lakeshore site on the
larger items—cores, grinding stones, etc.— Bodensee, Switzerland (Ismail-Meyer et al. 2013).
within a deposit.
Courty (1992) describes three types of cul-
tural deposits. The first or “primary” deposits 4.12 Palynology: A
result from utilization of an activity area (e.g., Micromorphological Study
flint knapping; food preparation). Ash layers are of Archaeological Pollen
primary cultural deposits. “Secondary” deposits
are primary deposits that have suffered Jean-Claude Gall (1983, 2012), in his book
modifications in their original settings. Examples Ancient Sedimentary environments and Habitats
of secondary deposits are “dumped” ashes of Living Organisms, has recognized the kinship
removed from a hearth area, shells consumed of archaeology and geology, in the fact that both
and dumped outside the eating area, etc. The face a “fossilized world.” If we are interested in
“tertiary” deposit were defined by Butzer the paleoenvironment of that world, then the only
4.12 Palynology: A Micromorphological Study of Archaeological Pollen 105

Fig. 4.15 The Neolithic site of Arbon-Bleiche main organic cultural layers divided by a sandy layer. (c)
3 demonstrating the use of micromorphology; (a) site The Arbon-Bleiche 3 thin sections of the column M 1030
map with the ground floors of the houses and with the (profile B) with the micromorphologically identified
eight phases of the village. The pilings of the houses are phases of installation, organic accumulations beneath
marked as black dots. The position of the profile section the house’s floor and inwash of sand from the hinterland
B and the profile photograph (c) are marked with white surrounding the village (Fig. 4.2 in Ismail-Meyer
arrows and highlighted in red. (b) The west profile et al. 2013, used with permission)
through house A showing the typical sequence with two
106 4 Techniques for Archaeological Sediments and Soils

recourse is “through the organisms and the 3. The morphological features of pollen and
sediments” (supra). In the case of the archaeolo- spores remain consistent within a species; dif-
gist, one must add cultural (artifactual) remains ferent species produce their own unique
as well. The study of archaeological sediments forms.
and what they contain as “fossils”—organic 4. Each pollen- or spore-producing plant is
remains such as floral and faunal materials, restricted in its distribution by ecological
artifacts and non-portable cultural features, and conditions such as moisture, temperature,
textural and chemical indices—can provide a and soil type.
great deal of evidence about a paleoenvironment 5. Most wind-blown pollen falls to the earth’s
within which a past human population together surface within a small radius, roughly
with plant/animal communities coexisted. 50–100 km from where it is dispersed.
Soil archives are rich. Plants and trees pro-
duce plenty of pollen, invisible to the naked eye As von Post observed, pollen is often produced
in soil layers. Plants can be recognized by their in significant quantities that survive, in sediments,
pollen, and when well preserved per layer, a list for long periods and have distinctive, often
of species can be determined (Kooistraa and species-specific forms and is distributed over sig-
Kooistra 2003). Fossil pollen grains and grass nificant areas, mostly by the wind as “pollen rain.”
spores are preserved in many soils and sediments The pollen of trees is termed arboreal or AP, that
that are not overly acidic. These elementary of shrubs and herbs, together with some grasses,
reproductive particles are produced in large non-arboreal (NAP). Bryant (1978) lists three
amounts by both trees and shrubs and the grasses. factors that influence the preservation of pollen
Some trees, like the pines, produce millions of grains in sediments and soils: (1) mechanical,
pollen grains annually. Other plants, by contrast, (2) chemical, and (3) biological. Mechanical dam-
are more meager in their output such as clover age occurs during dispersal and deposit ion gener-
(200 grains per anther). The grains themselves ally by crushing or abrasion of the grain.
are larger than clay particles (10–100+ μm) and Chemical corrosion of the exine is related to the
are amenable to optical microscopic study. Pol- pH of the depositional environment such that
len grains are tough customers being preserved sediments of 6.0 or higher alkaline degrades pol-
by a very resistant exine made of sporopollenin. len and alkaline soils that fix phosphates degrade
The exine defines the basic morphology of the pollen. Fungi and bacteria can breakdown the
pollen grain and is distinctive of most higher pollen. Susceptibility of pollen to corrosion results
plants. The palynologist inspects the grains for in the differential preservation of various species.
shape, size, surface texture, and apertures. As a Pine, although produced in large amounts, along
result of over a hundred years of taxonomic stud- with oak (Fig. 4.15), elm, and birch, can be over-
ies of pollen, beginning with Lagerheim and his represented in pollen spectra by a factor of
student, von Post (Bryant Jr 1989), in the early 4, whereas alder and beech (Fig. 4.16) are direct
1900s, we enjoy a broad and reliable repertoire of reflections, percentage-wise, of their occurrence
pollen identification keys (Faegri and Iverson in forest (Andersen 1986).
1975; Bryant, supra). Archaeologists, as with other scientists inter-
Von Post set forth five principles for pollen ested in past environments, must recognize the
studies which remain valid today: various factors that obscure recovered pollen
percentages as representational of vegetation
1. Many plants produce pollen or spores in great composition. Key factors are the biological pro-
quantities, which are dispersed by wind ductivity of the respective plants, dispersal
currents. modes and patterns, and grain survivability. For
2. Pollen and spores have very durable outer the purposes of archaeological geology, it is less
walls (exine) that can survive long periods important to this discussion to reprise the
of time. specialized discussions of varieties of pollen,
4.12 Palynology: A Micromorphological Study of Archaeological Pollen 107

Fig. 4.16 Typical triporate pollen grains: left, beech; right, oak

their identification, and the various methodologi- sediment/soil being sampled. Cores are taken in
cal approaches to reconciling pollen spectra to wetlands such as bogs, sinks, and shallow lakes
vegetational reality in the past than it is to exam- where sediments are conducive to pollen preser-
ine methods for the recovery and extraction of vation. Short, open barrel corers can simply be
pollen grains. pushed into soft sediments. Deep cores require
Pollen, itself, can and should be considered as a mechanical or hydraulic assists. In drier
sediment. It is an organic component of the sedi- sediments, the rotary, percussion, or drive corers
mentary matrix and sensu stricto a defining ele- are useful. After recovery, the sediment column
ment of “soil.” In paleosols, pollen is preserved is subsampled at selected intervals (ex. 10 cm).
subject to soil geochemistry. Within the peds of a If the sediment is in the form of a core, then
soil, the pollen grains are found with the silt-sized some care must be taken to note not only the
particles. As such, pollen grains are not recover- sampling interval but where that interval is in
able with the methods typically used to separate relation to the top of the core. Most investigators
the coarse sediment fraction. Pollen grains are stress that the sample be labeled, in terms of
separated by dispersal and digestion procedures. distance (cm, mm, etc.), from the top of the
Unlike the techniques used in the separation and core. Until just recently, the author was in com-
preservation of the clay, those used for pollen plete concurrence with this procedure. Having
recovery utilize more caustic reagents such potas- used geological coring, as a sediment sampling
sium hydroxide, hydrochloric acid, and the high method technique for some time, it now seems
reactive acid, hydrofluoric. Acquiring sediments that using the top of core, as the point of refer-
for pollen extraction and analysis can be done in ence, may not be as appropriate as previously
conjunction with the study and description of a thought. In situations of single cores, using the
soil/sediment profile. Similar to our example of top of the core as “zero” may have merit, but in
obtaining a soil monolith, a bloc of material is the case of multiple cores, across an uneven
removed by the use of a trowel or small shovel. topography, the stratigraphic relationship
Such a bloc or vertical sequence of blocks can be between cores, in an absolute sense, may be
returned to the laboratory for subsequent, system- difficult to determine without some measure of
atic removal of subsamples every few centimeters surface elevation, such that each core can be
as a general rule. placed in correct relationship to its neighbor.
Coring is a common means of retrieving con- One core must be a reference elevation datum
tinuous sediment columns. Of the various coring and the others placed in relation to it.
procedures described, the one most appropriate Where elevation data is lacking, it is
for the specific study depends in great part on the suggested that the measurement of “zero,” in
108 4 Techniques for Archaeological Sediments and Soils

the core, may be more useful, if taken from the Table 4.2 Sample preparation for pollen analysis
bottom or tip of the core. There is nothing arbi- (i) Boil sediment with 10 % NaOH or kOH to
trary about the bottom point of a core. In sam- remove soluble humic acids and to break the
pling wetland or submerged sediments, it is not sediment down
(ii) Sieve to remove coarse debris
uncommon to lose portions of upper sediments
(iii) Treat with cold 1 0 % HCl to remove carbonates
due to suspension/dispersion of light fractions or
(iv) Treat with either hot 48 % HF for a short time
simple unconsolidation of the sediment by the (up to 1 h) or cold 48 % HF for a long time (up to
coring device. In these cases, one does not have 24 h) to remove silicates (silt, clay, diatoms, etc.)
a true surface measurement that is coincident (v) Treat with hot acetolysis mixture (9 parts acetic
with the observed top of sediment in the core anhydride: 1 part concentrated H2SO4) to
hydrolyze cellulose
sample. By measuring from the core tip, some
(vi) After neutralization, stain the residue with
consistency is assured. In the case of varying safranin or basic fuchsin
surface topography, this procedure “normalizes” (vii) Mount the residue in a suitable medium with a
the disparity in strata by removing the offset in refractive index (1.55 which is that of pollen
altitude such that these strata will be observed at grains. An excellent medium is silicone oil
their same relative positions, in the core sample, (R. I. ¼ 1.4)
regardless of surface elevation. Courtesy George Brook, University of Georgia Geogra-
phy Department
Additional caution should be taken in the
removal of sediment whether from a core or
profile. Scrape the sampling site’s surface
Observed spike grainsðaÞ
removing the surface material with a spatula or
trowel that is cleaned between samples. Take Total spike grainsðbÞ
uniform samples of 0.5–1.0 cm3 (a few grams) Observed grains of taxon XðcÞ
in size. A typical extraction procedure is shown Total particles of taxon XðdÞ
in Table 4.2. The sequence of various regents and
dispersals removes the sediment matrix where one simply solves for (d) total particles of
containing the pollen and spores. After the pollen taxon X. Of course, using a spike of Eucalyptus
grains and spores are extracted, stained, and pollen is precluded in areas where the species
mounted, a typical sample area, such as occurs naturally or has been introduced (e.g.,
22  22 mm, is called “strew” (Rich 1999). southern California). Shane (1992) suggests
As with point counting, the pollen grain count using 7–10  104 spike grains/ml such that by
results in a relative sample frequency. This sam- adding 0.5–1.0 ml of this spike to a 0.5 cm3
ple pollen frequency must be converted to abso- sediment, sample will yield an “ideal” 2 grains
lute pollen frequencies to compensate for of the pollen spike to the sample (Maher 1981).
disparities in pollen influx at different times In counting a spiked sample, one procedure is to
such as the difference between fluences of 1,000 use a hemocytometer, a thick, specialized slide
and 100,000 cm2 although both represent 90 per- with apertures (2) for introduction of aliquots
centile samples. Davis (1969) suggested a novel into two, separate grids. Once the sample aliquot
method to do this by amending or “spiking” the is placed in the hemocytometer, the grains along
samples with a tracer pollen type in known the parallel sides of those grid squares marked by
amounts. Using this method, the number of pollen “X” are counted and listed in pairs. A total of
grains of each pollen type can be ascertained 30–40 grid squares are counted. The volume/
relative to known ratio of the tracer species. square is about 0.1 mm3 so 10 squares equal
The spike procedure may use something like 1 ml by volume. One hopes to observe at least
exotic pollen grains like those of Eucalyptus. The 300 grains in a count of at least 5 squares to
concentration of any pollen taxon can then be assure representation statistics. If the counts,
measured in ratio-proportion to this “exotic” done by a palynology laboratory, show a paucity
spike by the simple relation: of pollen grains (~300 grains/sample), they will
4.13 Phytoliths for Archaeology 109

Table 4.3 Palynological composition of samples from Gray’s Reef

Taxon Gray’s Reef clay Gray’s Reef core 1 Gray’s Reef core 2
Alnus 2.3 1.3 2.1
Ambrosia 0.89 0.97 0.62
Asteroideae 1.5 – –
Betula 0.59 0.97 0.98
Carya 3.6 1.9 1.2
Castanea 0.29 1.6 0.62
Chenopodiaceae/Amaranthaceae 15.2 8.1 5.6
Corylus 0.32 0.62 –
Cyperaceae 1.2 0.32 –
Fagus 0.59 0.32 –
Fraxinus 0.29 0.32 0.62
Gramineae 2.7 6.8 0.98
Iva 0.59 0.65 –
Liquidambar 1.8 – 0.30
Myrica 1.2 0.32 0.30
Osmunda 0.59 0.32 –
Ostrya/Carpinus – 0.32 –
Picea 2.1 – –
Pinus 41.7 56.5 67.2
Polygonum 0.59 – –
Polypodium – 0.65 –
Pteridium – 0.32 –
Quercus 13.4 12.7 13
Salix 0.59 – 0.62
Sambucus – 0.32 –
Sphagnum 0.29 0.32 –
Stellaria 0.29 – –
Taxodium 2.1 – –
Tsuga 0.59 – –
Ulmus 0.59 0.97 0.62
Woodwardia 2.1 0.97 –
Indeterminate 2.1 2.6 3.7
Totals 100 100 99.4
Values are expressed as percent of the total identifiable pollen and spores (minimum number of 300 grains: F. J. Rich,
unpublished data (in Russell et al. 2009; Table 4.4, Garrison et al. 2012a, b; Weaver 2002)

generally advise against extensive analyses complementary. In an integrated approach, their

(Table 4.3). application results in a more complete recon-
Micromorphology can identify whether struction of paleoenvironments and past land
pollen may occur in situ, analyze processes of usage (Hadorn 1994; Kooistraa and Kooistra
redistribution, or explain why pollen are not 2003).
found. Palynology can determine pollen and
organic materials present in thin sections to
the species level, identify the crops grown in 4.13 Phytoliths for Archaeology
arable fields, and explain deposition patterns. In
general, micromorphology identifies abiotic Rovner (1983, 1988) describes phytoliths as
conditions, palynology the biotic ones, both botanical microfossils that can provide significant
with such an overlap in that they can be paleoenvironmental and archaeobotanical
110 4 Techniques for Archaeological Sediments and Soils

information. Phytoliths are mineral deposits (cal- the identification of plant taxonomy. Phytolith
cium, silica) that form in and between plant cells populations are complex and stochastic—
(Mulholland and Rapp 1992). Of the two mineral- suffering from multiplicity of shapes, subtle and
ogic varieties of phytoliths—calcium and silica or not so subtle morphological variation, and dupli-
“opal”—the latter has formed the bulk of most cation (redundancy) of basic forms among taxa
modern phytolith studies. While both calcium (Rovner 1996b). In this form of systematics,
and opal phleboliths exist as either amorphous differences between phytoliths are ignored in
forms or cell morphology (“cystoliths”), the opal favor of their similarities. Phytoliths are grouped
phytoliths exist longer in sedimentary contexts. rather than differentiated into smaller taxonomic
Fossil opal phytoliths have been reported in sedi- categories such as genus and species. Further-
mentary rock millions of year old (Jones 1964). more, most measurements that are taken are
Bryant (1974) has reported both calcium and opal length and width since they are the easiest to
phytoliths in ancient coprolytes, but the former are measure manually (Rovner 1996b). Opal
not frequent in sediments. Bukry (1979) has phytoliths are produced by angiosperms,
reported opal phytoliths preserved in deep ocean gymnosperms, and pteridophytes (Piperno
sediments. 1988). They are produced by both
Christian Gottfried Ehrenberg began in 1836 monocotyledons and dicotyledons. Plant families
the systematic study of phytoliths or producing opal phytoliths include both Poaceae
phytolitharia which he published in 1845, and Gramineae grasses, Cyperaceae (sedges),
10 years after the first published work on Ulmaceae (elm), Fabaceae or Leguminosae
phytoliths by another German scientist named (beans), Cucurbitaceae (squashes), and
Struve (1835). His subsequent, extensive collec- Asteraceae or Compositae (sunflower)
tion and study of phytoliths from Europe, Africa, (Mulholland and Rapp 1992). Still phytolith pro-
Asia, and America led to Ehrenberg’s duction is not universal, and even when plants do
Mikrogeologie (1854). Ehrenberg’s lead inspired so, the phytoliths are amorphous and are not
other German researchers to continue phytolith distinctive.
studies in sediments and expanded observations The immense variety in shape stems both
of phytolith sources (Grob 1896). From the from intraplant taxonomic differences and sim-
beginning, the link between phytolith studies ple cellular differences between different parts of
and that of paleoenvironments, a key interest of the plant from a specific taxa. Additionally,
modern archaeology, was articulated by plants produce vast amounts of these cell casts
Ehrenberg and those who came after him. as they die and decompose in the soil. It should
Increasingly more common, the phytolith studies be noted that the two major differences exist
of archaeological sediments have yet to reach the between phytoliths and pollen besides long-
level of acceptance as palynological studies. This term preservation: (1) phytoliths are less taxo-
situation is rapidly changing, and the use of phy- nomic specific than pollen, and (2) they are more
tolith studies has proven to be of diverse benefit likely to be deposited in situ rather than broad-
to both archaeology and geology (Twiss 2001). cast over large areas like many forms of pollen.
When phytoliths can be reliably identified, they
make up stable components of archaeological
4.14 Phytolith Identification sediments (Pearsall and Trimball 1984; Pearsall
and Morphology 1989).
The size and shape of phytoliths vary
Traditionally, the identification of phytoliths has depending on the plant in which they are depos-
been based on typologies. The most typical ited and the area of deposition within the plant.
shapes found have been given significance for Monosilic acid is carried into the tissues of the
4.15 Phytolith Extraction and Counting 111

plant in groundwater and is usually deposited in siliceous residue (supra). Most of the residue
stems, leaves, and inflorescences. Eventually the used in a phytolith sample is composed of the
silica is precipitated forming opaline silica (SiO2 relatively nondiagnostic epidermal cells.
.nH2O) in the shape of the area that contains Fragments of the epidermis of leaves, stems,
it. These areas lie within or between cells. This and husks (inflorescence bracts—hair cells,
forms the cystalith or morphocast of the cell. papillae, etc.) offer today’s worker improved
Upon decay of this plant, the phytoliths become chances for identification success.
a component of the soil which can then be studied In terms of agriculture, many European
to determine the past floral assemblages of the grasses fall into the festucoid category, while
immediate area as Ehrenberg had foreshadowed. maize contains panicoid lobate shapes. Rice
Guntz and Grob (1896), more than Ehrenberg, contains a type of phytolith known as bulliforms,
identified and taxonomically classified more phy- while parts of other cultigens such as the rind of
tolith varieties than their predecessor. Guntz, squash, the leaf of bean, and the sunflower as
alone, studied 130 species of grass in attempt to well as maize form unique shapes that were
link cell morphology of leaves to the area’s cli- identified by Bozarth (1986, 1993). Squash, for
mate in which they grew (Powers 1992). example, forms spheroid phytoliths covered with
The grass family is a monocotyledon taxon contiguous concavities. Weed species that thrive
that produces an abundance of phytoliths and is in areas of agricultural abandonment often con-
also a good indicator of climatic changes. These tain chloridoid short cells. The identification of
phytoliths have been subdivided into general phytoliths from these species remains important
long and short cell categories. Distinctive grass to the study of archaeology. Trees and shrubs
short cells have been further subdivided into also produce phytoliths. The shapes vary greatly
festucoid, panicoid, and chloridoid classes and need further study to identify their taxa and
(Twiss et al. 1969). Festucoids include grasses origin. Grasses are largely unidentifiable below
that prefer cool and moist conditions and include family level.
Festuceae, Hordeae, Aeneae, and Agrostideae.
Their shapes range from trapezoid, rectangular,
elliptical, and acicular to crescent, crenate, and 4.15 Phytolith Extraction
oblong. Chloridoid grasses grow under drier and and Counting
warmer conditions. Their shapes are less diverse
being a basic saddle shape. Chloridoid grasses The technique of phytolith extraction set forth by
occur in Chlorideae, Ergrosteae, and Rovner (1996a) was used on the soil samples
Sporoboleae. Panicoids include the other grass collected from the site. It represents a relatively
tribes of Andropogoneae, Paniceae, Maydeae, simple but labor intensive and time-consuming
Isachneae, and Oryzeae which are located in procedure which is necessary to separate
warm, moist environments. Their short cells con- phytoliths from the other components of the soil
tain the most distinct shapes such as bilobates, (Table 4.4). Approximately thirty grams of soils
cross bodies, and crenate shapes. The grass short are necessary for processing using this
cells and their morphology are illustrated and procedure.
discussed in detail by Brown (1984). Twiss has Once the phytoliths have been successfully
pointed out that for grass phytolith classification, extracted, one proceeds to identify and count
it is the costal short cells that are the most diag- as many recognizable forms as possible at
nostic (2001). Unfortunately, these cells com- family or subfamily taxonomic levels. Most
prise only a small fraction (~5 %) of the total researchers prior to the 1990s relied on light
112 4 Techniques for Archaeological Sediments and Soils

Table 4.4 Process of phytolith extraction

Step Reason
Distilled water Removes organic material by flotation
Centrifuged Allows phytoliths to settle to bottom
Decanted Removes excess water
Repeat 3 times
Bleach Dissolves remaining organic material
Soaked overnight
Distilled water Rinses off bleach
Repeat rinse 3 times
HCI Removes CaCO3
Distilled water Rinses off HCl
Repeat rinse 3 times
Na metaphosphate Removes clay minerals
Distilled water Rinses off Na metaphosphate
Repeat rinse 3 times
Dried overnight Removes moisture
Placed in vials
Heavy liquid (ZnBr or Na metatungstate) sp.g.2.35 Floats off phytoliths
Distilled water
Repeat rinse 3 times
Placed in vials
After Owens (1997)

microscopy to identify and count phytoliths. The identification of phytolith morphology

method most commonly used is the “quick scan” (Fig. 4.17). Both technologies are superior imag-
under magnifications of 300–500X. The slide, ing systems compared to traditional optical
containing the phytoliths, is raster scanned with microscopy. SEM is well established, and confo-
the observer tallies all phytoliths present as to cal laser microscopy has the like advantage of
known and unknown types. The use of scanning imaging the cell cast in great detail in storable
electron microscopy (SEM) and confocal digital, isometric images. The drawback to these
technologies is changing phytolith analytical higher-resolution imaging techniques is scale
procedures particularly with regard to the and speed.
4.15 Phytolith Extraction and Counting 113

Fig. 4.17 Various

phytolith types. Arrows
indicate “cross-body”
chlorodoid types, whereas a
festicoid type may be
present, center of the
photomicrograph. The
photomicrograph courtesy
Dr. Liovando Marciano da
Costa. The two close-ups of
chlorodoids and festicoid
types are from Ball
et al. (1996)

Because the SEM and confocal systems focus optical procedures. Russ and Rovner (1989) have
on micron-submicron scales, the investigator suggested the use of computer-assisted image
does not examine as many phytoliths during a analysis to achieve reliable, less subjective phy-
scan. The identifications are more confidently tolith identification as well as more rapid assess-
made, but the overall number of individual ment of total numbers of phytoliths per sample.
“grains” counted is minuscule compared to the
Geophysical Techniques for Archaeology

5.1 Introduction archaeology occurred during the 1960s and

1970s (supra). Using the same hyperbole, per-
This chapter will provide a brief but relatively haps, a “second golden age” is upon us.
comprehensive overview of geophysical In this survey we shall examine basic
techniques used in archaeology today ranging principles of the specific geophysical technique
from electrical—resistivity and conductivity— under discussion and then proceed to a brief
to magnetic and radar methods. The methods to discussion of the technique’s importance to
be discussed are the ones most commonly used archaeology. When illustrative, a brief example
in archaeological geophysics. Less frequently or two will be given. Mathematical formulae and
used methods—seismic, gravity, thermographic, equations are part and parcel of geophysics but,
induced polarization, and self-potential—to in this chapter, we shall only examine those most
name some of the other methods used in “con- germane to the method. There are a host of
ventional” geophysics, will not be examined in works, some referenced in the following para-
any detail. Many of these other methods have graph which do provide in-depth summaries of
achieved interesting results, in regard to archae- the physics and formulae involved in the use of
ological prospection, but they remain, for a these geophysical methods. Excellent general
variety of reasons, of marginal interest to archae- textbooks, on geophysics, such as those by
ology. As with all the geophysical methods Prem Sharma (1997), Kearey and Brooks
utilized by archaeology, none were developed (1991), and Burger (1992) together with field
with archaeological prospection in mind. As is manuals (Milsom and Eriksen 2011) are readily
true with most, if not all, of the various methods available to those readers who seek a more
used by archaeology, these methods have been in-depth understanding of the principles.
borrowed and adapted to fit archaeological goals. Textbooks (Scollar 1990; Clark 1990, 2003;
As Gaffney (2008) observes, the increased use Conyers and Goodman 1997; Gater and Gaffney
of geophysical methods in archaeology is at 2003; Witten 2006; Aspinall et al. 2009; Conyers
an all time high. Linford (2006) notes that this 2013) and major reviews (Wynn 1990;
was presaged in the 1980s. Gater and Gaffney Weymouth 1996; Kvamme 2001; Aspinall
(2003) suggested, perhaps, correctly, that the et al. 2008; Linford 2006; Gaffney 2008),
demise of the influential journal Prozpezoni together with a dedicated journal—Archaeolog-
Archaologiche, in that same period, contradicts ical Prospection (1994 to present)—are all
Linford’s contention. The so-called golden age in indicators of the increased importance of geo-
development and application of geophysics to physical techniques for archaeology. Excellent

# Springer-Verlag Berlin Heidelberg 2016 115

E. Garrison, Techniques in Archaeological Geology, Natural Science in Archaeology,
DOI 10.1007/978-3-319-30232-4_5
116 5 Geophysical Techniques for Archaeology

historical background reviews appear in some The suite of geophysical instrumentation

of these works as well as elsewhere. Without developed or adapted for archaeological
digressing into historical perspective, it is impor- prospection is indeed impressive. With the
tant to remember that the use of geophysics in concurrent usage of geophysics in mineral
archaeology began in earnest after World War II, prospection, environmental studies, and geotech-
first in Europe and then elsewhere. The sequence nical engineering, the development of new
of development proceeded from the use of soil systems and iterations of existing ones has,
resistivity methods by British workers (1950s) to indeed, been impressive. Linford (2006) and
magnetics by the Germans and Italians (1960s) Gaffney (2008) have both noted that while the
and the development of ground-penetrating radar basic physical nature of geophysical methods
(GPR) by Americans (1980–1990s). Other geo- have not changed, digitalization has
physical techniques have found their way into revolutionized their design, application, data
archaeological use but to a lesser degree than processing, and presentation. Along with this
those just mentioned. These latter methods welcome, increased utilization of geophysics in
include induced polarization (IP), conductivity archaeology has come to an increase in a more
(EM), magnetic susceptibility, microgravity, unwelcome aspect similar to that described for
thermography, seismic reflection/refraction, and the chemical analysis of archaeomaterials
electrical tomographic imaging. (Artioli and Angelini (2011)—the proliferation
Beginning in the 1960s, the International of data that should never appear in scientific
Symposium on Archaeometry has prominently forums or journals.
featured archaeological prospection in its Beyond venues like that of the International
programs. The increasing inclusion of geophys- Archaeometry Symposia, papers on research and
ics in archaeology has seen the use of interesting application of archaeological prospection have
terms to describe this. Wynn (1987, 1990) received more and more time at less technically
proposed the term “archaeogeophysics” for all focused professional meetings. Professional
ground and airborne geophysical methods conferences in the earth and soil sciences,
applied to archaeology. In the strictest sense, archaeology, and remote sensing commonly
most geophysical methods used today are terres- include papers on the theory and use of geophys-
trial (“ground”) in nature although the use of ical methods in archaeology. Add to this favor-
airborne or space-based sensing technology (syn- able trend the use of workshops on the “hands on
thetic aperture radar (SAR), light distance and application of these techniques” such as the US
ranging (LiDAR)) has produced significant National Park Service’s annual workshop on
archaeological finds ranging from buried, “des- archaeological prospection techniques entitled
ert” rivers (Wendorf et al. 1987; Wiseman 1996) “Current Archaeological Prospection Advances
to Mayan centers hidden under tropical forests for Non-Destructive Investigations in the 21st
(Chase et al. 2011; Comer and Harrower 2013). Century.”
There is an additional use of geophysical Academic training in archaeological
methods that Wynn does not specifically include prospection has evolved as well. Important
in his 1990 definition although he discusses it examples can be found in European (Oxford,
in his review article at some length—acoustic Bournemouth, Bradford, Zürich, CNRS-France,
reflection and sonar in marine archaeology. Warsaw, Vienna, and Rome) academic settings
Today, seismic/acoustic profiling coupled with together with the North American Universities:
magnetometers and side-scanning sonars are the UNAM (Mexico City), University of
standard marine search procedures conducted Arkansas, University of Georgia, University of
on archaeological sites ranging from shipwrecks Mississippi, Louisiana State University, and
to drowned terrestrial landforms of archaeolog- University of Pennsylvania. As today’s students
ical interest (Pearson et al. (2008); Gaffney become skilled in theory and practice, the appli-
et al. 2007). cation of geophysics in archaeology is assured.
5.3 Electrical Methods: Resistivity 117

Professionals, in the past decade, have formed a minerals, basalt, gabbro, diorite, and magnetite,
professional organization focused exclusively on are better than the sedimentary or metamorphic
the use of geophysics in archaeology. rocks. Table 5.1 lists resistivity values of com-
mon sediments, rocks, and minerals. Of the
sediments, clays are the best conductors, and
5.2 The International Society clay-rich soils have lower resistance to electrical
for Archaeological Prospection flow than, say, quartzitic or carbonate types.
(ISAP) The difference in the resistance or resistivity,
as it most commonly called, as measured by
The inaugural meeting of the Society was held on galvanometers or ohm/voltmeters, is used to
September 13, 2003, during the fifth Interna- detect and characterized buried features of an
tional Conference on Archaeological Prospection archaeological nature. The fundamental physical
in Cracow, Poland. “The object of the Society relation involved in resistivity is Ohm’s law.
shall be to advance the education of the public in
R ¼ V=I
archaeology (including the man-made landscape
and the built-environment) through the promo- In Ohm’s law, R is the resistance (measured in
tion of high standards of research, application ohm Ω), V is the voltage (measured in volt V ),
and communication in the field of archaeological and I the current (measured in ampere A). One
prospection and related studies. The Society’s ampere is equal to one coulomb of (electron)
scope shall be international, both in activities charge moving in 1 s through a potential of 1 V.
and membership” (Armin Schmidt, Personal For a current to flow, there must be a difference
Communication 2015). in the voltage potential, also termed bias. On the
microscopic level, Ohm’s law can be written, in
vectorial form, as
5.3 Electrical Methods: Resistivity
E ¼ σj
Resistivity methods have “pride of place” in
archaeological geophysics by virtue of being E is the electric field, σ is conductivity, and j is
the first geophysical method used by the current density.
archaeologists (Aitken 1974). These early Resistivity and resistance are not totally syn-
efforts, after World War II, consisted of much onymous terms. Resistivity is specific resistance,
trial and error, with multiple, galvanic electrodes expressed as ρ in the unit of ohm-meter Ωm. It is
used as a rule. The first applications were what the resistance to current flow, over some dis-
we will call “profiling” studies in that the tance, and is proportional to the area. In resistiv-
electrodes—simple steel rods in most cases— ity survey we measure the apparent resistivity,
are advanced over the site at a fixed interval ρa, which is the “true” resistance across a nonho-
between the probes, Wenner and Schlumberger mogeneous media like the earth. The apparent
methods (supra). A rough rule of thumb, still resistivity is expressed with a geometric correc-
followed today, holds that the electrode separa- tion factor that is particular to a particular
tion is equivalent to the profiling depth or that a electrode spacing or array. The geometric factor
1 m electrode separation will detect buried for the Wenner array, the most commonly used
features at 1 m depth in the site. array, is 2πa where a is the electrode spacing and
Electrical resistivity or resistance survey 2π considers the half-space volume over which
methods measure electrical flow through soils the current flows. With this in mind, we rewrite
or sediments. Most earth materials conduct Ohm’s law, for the Wenner array as
currents, poorly at best, but vary nonetheless, ρa ¼ 2π a V=I
according to mineralogic makeup and water con-
tent. Rocks are poor conductors. Some of the A material with significant resistance to current
metallic or igneous varieties, such as the sulfide flow is an insulator, while one that allows current
118 5 Geophysical Techniques for Archaeology

Table 5.1 Resistivity of common materials

Material Frequency at measurement, CPS (if not DC) Resistivity, ohm-cm
Galena 0.5–5.0
Pyrite 0.1
Magnetite 0.6–1.0
Graphite 0.03
Rock salt (impure) 3  103–5  105
Serpentine 2  104
Siderite 7  103
Igneous rocks
Granite 108
Granite 16 5  105
Diorite 106
Gabbro 107–1.4  109
Diabase 3.1  105
Metamorphic rocks
Garnet gneiss 2  107
Mica schist 16 1.3  105
Biotite gneiss 108–6  108
Slate 6.4  104–6.5  106
Sedimentary rocks
Chattanooga shale 50 2 x 103–1.4  105
Michigan shale 60 2  105
Calumet and hecla
Conglomerates 60 2  105–1.3  106
Muschelkalk sandstone 16 7  103
Ferruginous sandstone 7  105
Muschelkalk limestone 16 1.8  104
Marl 7  103
Glacial till 5  104
Oil sand 4  102–2.2  104
Rock type Resistivity range (\-m)
Consolidated shales 20–2  103
Argillites 10–8  102
Conglomerates 2  103–104
Sandstones 1–6.4  108
Limestones 50–107
Dolomite 3.5  102–5  103
Unconsolidated wet clay 20
Marls 3–70
Clays 1–100
Alluvium and sands 10–800
Oil sands 4–800

flow is a conductor. Glass and silicate minerals water is a good insulator whereas tap water is a
are insulators, while the metals are conductors. conductor. Why? Distilled water is free of any
The conductance of current through soil is mineralogic salts and other ions where tap water
greatly influenced by water content. Water has charge carrier ions such as sodium, chlorine,
alone is not necessarily a conductor as distilled etc., that conduct electrical current. This property
5.3 Electrical Methods: Resistivity 119

is termed ionic conductivity. In sediments, the array instruments. The multielectrode (25+) cable
number of ionic elements is much greater than arrays used with computer-driven control units
simple water. It is the water that mobilizes the allow the rapid collection of field data in a host
ions to form the electrolytic soil/water solution. of array configurations—Schlumberger, Wenner,
When an electrical potential is applied across a double-dipole, etc.
soil volume, with electrodes, the material and
moisture will act first as charge carriers. After a
period of time, the soil will impede the motion of 5.3.1 Resistivity Arrays
ions so that they accumulate at grain boundaries
wherein the ground acts like a battery. This is In archaeological prospection four-electrode
particularly true where direct current (DC) is arrays are most commonly used. We now review
applied to the electrodes. To avoid unwanted in order of popularity. The Wenner array is the
polarization effects, resistivity devices use “standard” against which other arrays are
alternating current (AC). An alternating current assessed (Milsom and Eriksen 2011).
source, regulated by an oscillator, varies the sign
of the current in +0.5 to –0.5 cycles such that the Wenner Array
ground “battery” cannot build up charge without The high impedance measurement circuit used
subsequent reversal and discharge. As a result for Wenner potential electrodes (P1, P2) reduces
there can be no net polarization. The early contact resistance by orders of magnitude
two-electrode arrays were prone to another because it draws little current. It has been found
problem known as “contact resistance” which that electrode size, depth, and geometry are of
comprises a large part of the total resistance. little importance but electrode spacing, as we
Workers overcame this problem by adding two have said, is more so. With the Wenner array,
additional electrodes, equally spaced apart, and this relationship is fairly direct with the depth of
the four-electrode array was born. profiling roughly equivalent to electrode spacing.
As stated, resistivity is the most venerable of Depth and spacing are generally treated as
geophysical techniques used for archaeology equivalent.
(Aitken, supra); resistivity will continue to The diagram in Fig. 5.1 illustrates the spacing,
be used into this millennium as a productive a, is equal and set to the depth of interest for the
prospection method. Over a half-century of expe- specific traverse. To traverse a site, the array is
rience, in both archaeology and geology, resistiv- advanced by increments equal to a with readings
ity has provided a solid conceptual as well as made. The direction of the survey will influence
applied basis. To remain a productive geophys- the appearance of most anomalies. While it is
ical methodology for archaeology, advances time consuming, it is sometimes useful to resur-
in electrode design, such as non-polarizing vey at right angles to the direction of the original
electrodes, together with current reversal cir- survey. For sounding surveys the electrode
cuitry, low power AC batteries, and low induction spacing is expanded by increasing increments
cables have all contributed to a more efficient of a ¼ 1, 2, 4, 8. . .n. It is the easiest of the four-
modern resistivity technology. These hardware electrode arrays to conceptualize. One simply
innovations have been coupled with the advances lines up the four electrodes, at an equal interval
in computer design to make modern instruments a, drives a current, reads the resistance, and
more compact and flexible. Data is in digital then moves the electrodes, in sequence, by
formats making storage manipulation and predetermined interval a. This is repeated until
modeling by geophysical software all the more the area of interest has been completely traversed
routine. The most often heard problem—lack of on an X-Y grid.
speed in field survey—has been addressed, at All of the arrays discussed in this section can
least in archaeology, by the modern double- be used for archaeological profiling studies and
dipole array instruments and the multielectrode indeed they have. The profiling method simply
120 5 Geophysical Techniques for Archaeology

ent line
nt me asurem
Th e curre
pin 43
pin 44
pin 45
pin 46
pin 47
pin 48 line
xt measur
The ne
pin 43
pin 44
pin 45
pin 46
pin 47
pin 48

An electrode selector A resistance meter A personal computer

a 12 bit A/D converter

Fig. 5.1 General multielectrode resistivity system. This system can be programmed for both Wenner and
Schlumberger array configurations

requires the traverse of an archaeological site of profiling in the discussion of the Schlumberger
using a preselected offset and sampling interval. array.
This is termed the “constant-spread” traverse.

Resistivity Sounding This category of resistiv- Double-Dipole Array

ity measurement is typically described as a In principle, the double-dipole array mimics
sounding method (Burger 1992) but that is only induced polarization (IP) techniques in array
if the center of the array is kept at a fixed location configuration and measurement. The principal
and the electrode spacing, a, in the Wenner difference in the two techniques is the method
method, are sequentially increased. This com- of applying and measuring potentials—the
mon variation is to simply repeat the constant- double-dipole array applies a continuous poten-
spread traverse at increased increments of (a). tial, while IP relies on induced, remnant
Rather than construct a resistivity pseudosection, potentials transient in the soils. For a discussion
described in the following sections, each of IP methods, see Wynn (1988). The double-
(a) spacing traverse’s data, for that level, can be dipole array has the potential electrodes
viewed independently of preceding or separated, at some distance, from the current
subsequent traverses. This data can be modeled electrodes (C1, C2), while the electrode spacing,
in an areal contour plot rather than in a depth a, is the same for both sets of electrodes) involves
section. Multi-level plots can be made treating a geometric factor that differs from Wenner and
each level as a “resistivity surface” (Herbich Schlumberger arrays. Here, the parameter n is the
1993). We shall provide an example of this type distance between the current electrodes.
5.3 Electrical Methods: Resistivity 121

ρa ¼ πa n ðn þ 1Þ ðn þ 2Þ V=I twice the separation between the electrodes. The

thing that commends the twin-electrode array to
Williams (1984a, b, 1993, Personal Communica- archaeological prospection is its compact nature.
tion) touts the double-dipole array over that of The so-called mobile probes are on a fixed frame
the Wenner array, in archaeological prospection, with a cable connection to the other electrode
primarily because of the location of the measured pair located at a fixed point some distance away
maximal value of apparent resistivity. With the (greater than 30 times the electrode spacing as a
Wenner array, the profile of a resistive feature rule). Another benefit of the large separation in
takes on a double-peak shape and is reminiscent the array halves is a reduction in background
of the magnetic dipole shapes discussed later in resistance. Current practice in Great Britain,
this chapter with the maximal value for ρa, over and elsewhere, in Europe, favors the use of the
the feature, reduced because of electrode spacing twin-electrode array. It is quick—up to 6,000
and type. readings per day—and coverage is comparable
This “double-peak signature” is not seen in to that of magnetometer survey, a hectare
plots of double-dipole data as the most sensitive per day.
location for measuring apparent resistivity is
measuring at (ηa). Two disadvantages of the
double-dipole array are (1) overall sensitivity Schlumberger Array
and (2) comparability of reading to those of Conceptually, the Schlumberger array is as well
other arrays. This is true for either profiles or understood as the Wenner array and is, perhaps,
soundings. The double-dipole array has become more used than that array for sounding work. On
popular in both archaeological and mineralogical the first inspection, the two arrays appear similar
prospection using sounding data to produce 2D with the potential electrodes in the array’s center
inversions and pseudosections (Griffiths and and the current electrodes on the ends. The
Barker 1994; Milsom and Eriksen 2011). Both potential electrodes are fixed in their spacing
Wenner and Schlumberger arrays produce more such that a minimum factor of 2.5 exists
reliable, deep soundings. between their interelectrode distance, 2 l, com-
pared to that of the current electrodes, L. The
array can be moved as an ensemble for a profile Two-/Twin-Electrode Array
survey or the current electrodes moved sepa-
Milson and Eriksen (supra) characterize the
rately increasing L, by logarithmic steps, to pro-
modern variant of this array (double-dipole)
duce sounding data. The inner electrodes are not
as the one most popular in archaeological
moved until voltage levels are not measurable.
prospection today. The successful modern vari-
The apparent resistivity for this array is calcu-
ant is a DC system with a fixed interelectrode
lated as follows:
distance, a, that allows the unit to be operated by
a single person without cables to caddy. The use ρa ¼ π L2  l2 V =2 l
of DC, with its polarizing effect, has been
ameliorated by use of a switching circuit to Williams (1982, 1998) considers a 1 m depth/
minimizing charging. electrode spacing adequate for most archaeolog-
The twin-electrode array was especially ical resistivity work. This being the case, then the
designed for archaeological prospection (El- use of a rigid frame resistivity meter mounting
Gammili et al. 1999). It is basically a Wenner fixed-distance electrodes (a) and a data logger or
array divided into two parts, with a very large portable computer is in keeping with most
separation between the current and potential profiling needs of an archaeological nature.
electrode pairs (supra). Aspinall and Lyman Calculation of the apparent resistivity for the
(1970) have shown that this arrangement allows two-electrode array is the same as that for the
for the same penetration of a Wenner array with Wenner array.
122 5 Geophysical Techniques for Archaeology

5.4 Vertical Sounding Methods the past processing, filtering or smoothing of

in Archaeology the data could not seem to overcome this fact
(Szymański and Tsourlos 1993), but the use of
Electrical sounding has been called “electrical inversion techniques have improved results
drilling.” As such it can be thought of as a type significantly (Barker 1992; Sasaki 1989;
of coring device to get at the depth as well as Tripp et al. 1984.
the distribution of archaeological features, i.e.,
pits, hearths, and remains of structures. The Today, with double-dipole data, inversion of
Schlumberger technique provides excellent the pseudosection data gives a 2D tomographic
profile data. It can also be used for vertical image (Fig. 5.2). It is now possible to get accu-
soundings, involving the expansion of the spread rate depth and location using the inversion pro-
of current probes AB around a central point with cess (Sharma 1997). By starting with an initial
constant position of potential probes MN, giving model—the pseudosection—which is essentially
information on the number of horizontal resistiv- a “guess” as to what the subsurface looks like, the
ity interfaces at that point. The result reflects the inversion then changes or “iterates” the model
properties of a vertical section or pseudosection data until the model agrees with the collected
of the subsurface, but: data within accepted error. These are called
“error-minimization” routines. It is analogous to
• The quality of the interpretation decreases the technology used in medical MRI and CAT
radically with depth, since the properties of scans. Barker (1992) developed an algorithm,
larger volumes of soil are being sampled by based on the finite difference method, that inverts
each expanding reading. the pseudosection and converts the apparent
• Each datum point will include significant resistivity measurements to “true” resistivity—
contributions from volumes of soil not directly an immediate gain is the adjustment of depth
under the array—the pseudosection compresses variation. For a detailed discussion of the inver-
all such effects onto a two-dimensional display sion method, see Szymański and Tsourlos (1993)
of apparent resistivities onto an arbitrary and Noel and Xu (1991) as well as Barker (1992).
vertical grid. Having given the reader these caveats, many
• The pseudosection, whether displayed as a current articles describe the use of electrical
block of numbers, a gray-scale image, or in sounding and, in particular, suggest the use of
contoured form, is not a direct representation pseudosections to aid in the interpretation of the
of the subsurface and any interpretation as an subsurface archaeology (Dalan et al. 1992;
image can be extremely misleading—the Aspinall and Crummet 1997; Szymański and
depth scale can be completely arbitrary. In Tsourlos 1993). Griffiths and Barker (1994)

Fig. 5.2 Resistivity pseudosection of Mound A, Etowah initial radar studies of this and adjacent mounds B (Garrison
Mounds State Park, Cartersville, Georgia, USA. This 20 m- 1998). This figure is the result of a later electrical study
high Mississippian period temple mound was built some- (Garrison et al. 2005). The near-surface features in red are
time prior to 1250 AD. It has been the subject of several well-documented “robber trenches” but the deeper
geophysical studies beginning in 1993 with the author’s anomalies, in green, are prehistoric structural features
5.5 Electrical Methods: Electromagnetic/Conductivity 123

have used a computer-controlled electrode array secondary field is a direct consequence of

along a fixed-length cable such that the sequence, the magnetization induced by the primary field.
spacing, and array configuration are selected and High-frequency waves attenuate first and
manipulated by the software resident in the con- quite rapidly with depth. As a consequence
trol unit (Fig. 5.1). frequencies above 1,000 Hz are rarely used in
devices for archaeological prospection. A semi-
empirical formula relates soil resistivity to fre-
quency as follows:
5.5 Electrical Methods:
Electromagnetic/Conductivity pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
h f =ρ ¼ 10
The reciprocal of resistivity is conductivity and is h is the depth, f the frequency, and ρ the resistiv-
written with the Greek sigma, σ, just as in the ity. For instance, if we solve for f with ρ equal to
microscopic form of Ohm’s Law, so 104 Ω-m and h equal to 10 m, then f should be
100 Hz.
σ ¼ 1=ρ
With the continuous wave (CW), frequency
Electromagnetic (EM) induction, avoided in domain EM devices such as the Slingram,
resistivity measuring systems, is utilized to mea- Geonics EM 31 and EM 38, and the primary
sure the conductance in the subsurface. Like transmitter coil put out a sinusoidally varying
resistivity, differences in conductivity measured current at a set frequency which serves as a
in the subsurface may have as their source reference for the secondary (receiver) coil
objects or features of archaeological interest. whose current characteristics are compared to
Induction of eddy currents is a consequence of the primary. The amplitude and phase angle
a primary dipole field whose magnetic field lines components of the current are compared, e.g.,
flow through a conductor producing the weak the inphase and quadrature (out-of-phase)
secondary currents (electrical). Conductive components. These are read directly on a digital
minerals like clays and metallic oxides in soils meter on the device. The principal advantage of
enhance induction by what is termed mutual the conductivity meter is speed. These units can
inductance (Tite and Mullins 1969; Milsom and cover a site’s area as quickly as the magnetome-
Eriksen 2011). ter. Even as impressive as the new twin-electrode
A Slingram-type conductivity meter uses a resistivity units are, they are no match for the
source current flowing through a wire coil loop ease of survey that the conductivity and magne-
producing an alternating low-frequency mag- tometer provide to archaeological prospection. It
netic field (i.e., between 200 and 20 kHz), while should be noted that the noncontact nature of the
a second coil, at a distance L, measures two conductivity measurement usually results in a
different components of the magnetic field—the reduced sensitivity, ordinarily by a factor of
inphase and quadrature components. The inphase 5, when compared to resistivity measurements,
component is a measurement of the magnetic but this is usually not important in archaeological
susceptibility of the soil and that of quadrature cases.
a measurement of the soil conductivity. Inphase
measurements are like those of a “metal detec- Multifrequency Devices These instruments,
tor,” only much more sensitive. Quadrature such as the GEM-2, made by Geophex, Ltd.,
measurements are those of electrical properties and the Profiler EMP-400, manufactured by
of the soil. GSSI, Inc., operate over a range of frequencies.
The response of the conductivity meter is For the Profile EMP-400, it operates on three
complex: the secondary field is not inphase with frequencies from 1,000 to 16,000 Hz. The fre-
the primary field and this increases with the con- quency range for the GEM-2 is 300–96 kHz.
ductivity of the ground. The presence of this They are “time domain” instruments. As we
124 5 Geophysical Techniques for Archaeology

recall the effect of frequency on the soil, gener- measuring a weak, secondary electromagnetic
ally speaking, the higher frequencies “see” all field/current in the ground. The ability of the
soil as conductive, and, therefore, the depth of subsurface materials to conduct electromagnetic
survey is quite shallow. Circuitry and coil energy can then be measured and used to
geometries allow the use of multiple frequencies interpret the anomalies found within the subsur-
which are particularly good at isolating shallow face. EM surveys that can measure multiple
conductive features such as those interesting to frequencies can correlate these frequencies to
archaeology (Figs. 5.2 and 5.4). These multifre- different depths. Lower frequencies are able to
quency devices allow for the acquisition of large, penetrate deeper into the ground due to their long
digital data sets that can be plotted and compared wavelengths but cannot resolve finer details.
as to the presence of conductive features and Higher frequencies may be able to resolve those
their “depths.” True depth estimations can be finer details in the data but are unable to pene-
problematic for EM survey results. With units trate as deeply into the ground. By using a range
like either the EM-31 or EM-38, the rough rela- of six frequencies during the Cherokee Female
tionship between coil separation (1.2  coil sep- Seminary survey, it was possible to explore dif-
aration) and depth can be used. The E-31 has a ferent depths from just below the ground surface
maximum depth of survey of about 6 m, while down to deeper geological and embedded archae-
that of the EM-38 is only 1.5 m. This problem ological features.
will crop up again in our discussion of another A GSSI GEM-300 conductivity (EM) meter
EM-based device: the ground-penetrating radar was used for this survey. Data were collected at
(GPR). stations placed every 2 ft along transects that
were spaced 3 ft apart. These intervals were
used because the building was constructed
5.5.1 Application of EM Methods using feet-inch units. EM data were downloaded
from the GEM-300 instrument and processed
EM surveys are done as traverses. Like resistivity using Geosoft’s Oasis MONTAJ software. Each
data, the results are typically presented as of the six frequencies was used to provide infor-
(a) profiles or (b) surface plots, either plan view mation about anomalies at different depths.
or isometric. The following example illustrates Depths were roughly calculated using the for-
the use of EM deployed over the site of a historic mula for skin depth, but more accurate correla-
school building. tion for depth was generally found by cross-
comparing EM data plots with those using GPR
data (Figs. 5.3 and 5.4).

5.6 Conductivity (EM) Survey of a

Burned and Buried Cherokee
Seminary Building, Tahlequah, 5.6.1 Magnetic Methods
Oklahoma, Cherokee
Nation, USA The degree to which a material becomes
magnetized in an applied magnetic field H is
In a survey conducted over the area of a burned called its magnetic susceptibility. Later in this
and buried nineteenth-century seminary building chapter, we will discuss the use of this property
(Fig. 5.2), conductivity data was taken using of rock minerals and sediments for archaeolog-
the predecessor of the Profiler EMP-400, the ical purposes, but for the moment we will
Gem-300, a multifrequency instrument capable use it to examine how magnetism is used in
of up to six frequencies in a range from 330 to archaeological geophysics. How a material is
20,000 Hz. Electromagnetic induction, as noted magnetized is termed magnetic induction. Mag-
above, works by inducing and then subsequently netic susceptibility, χ, is defined as
5.6 Conductivity (EM) Survey of a Burned and Buried Cherokee Seminary Building. . . 125

Fig. 5.3 The Cherokee female seminary, ca. 1,848. Courtesy of the Oklahoma Historical Society

χ ¼ M=H magnetism can be ferromagnetic or anti-

ferrimagnetic depending on orientation of
where M is the induced magnetism and H is the magnetic moments. Iron (Fe), cobalt (Co), nickel
total magnetic intensity of the applied field. (Ni), their alloys, or minerals are powerfully
χ is a dimensionless quantity in the cgs system attracted to magnetic fields (“magnetic induc-
and generally measured for a volume or mass. tion”). As a result, ferromagnetic and ferrimag-
The above relation describes volume susceptibil- netic materials are the most prevalent in
ity and when divided by the material’s density, producing strongly magnetic features in archaeo-
χ/d, the mass susceptibility is obtained. Magnetic logical sites. We, therefore, characterize
materials vary in susceptibility from weakly neg- materials on their ability to “acquire” magnetism
ative to strongly positive. Magnetic prospection or their magnetic susceptibility. Typical
depends on an observed contrast in the suscepti- susceptibilities are shown in Table 5.2. The
bility of materials. Almost all magnetic features reader will note that χ, in this table, is written in
in archaeological sites are differentiated by the Gaussian or cgs units. Volume magnetic suscep-
amount of the magnetic mineral magnetite, tibility is generally regarded as dimensionless,
Fe3O4 present. Magnetism is characterized as while the magnetic susceptibility per mass
either diamagnetic, paramagnetic, or ferri(o) is a more useful quantity for use in
magnetic. Briefly these various terms describe magnetochemistry of archaeological materials
the magnetic response at the atomic level. and is given the units of reciprocal density.
Diamagnetism arises from the distortion of an Rocks, artifacts such as ceramics, can acquire a
atom’s electron orbits in a magnetic field. permanent net magnetism termed remanent
Always present, this form of magnetism is very magnetism. There are various ways this can
weak (about 1 nT). When an atom has an unfilled occur such as heating, chemical activity, and
electron shells, valence or otherwise, they will pressure. The remanent magnetism so acquired,
demonstrate paramagnetism individually. This by whichever physical means, is relatively per-
latter point is important. By behaving individu- manent, over geological time, and forms the
ally, in bulk arrangements (crystals, glass, basis for paleo- and archaeomagnetic dating
liquids), there is no net magnetic behavior or methods. Interaction of the induced and rema-
moment. Paramagnetic materials demonstrate nent magnetic components yield the net magne-
weak magnetism in applied fields. tism of a material we observe in the earth’s field.
When a material’s atoms, with unfilled outer To measure this net magnetism, magnetometers
shells, behave collectively or coupled, the are utilized.
126 5 Geophysical Techniques for Archaeology

Fig. 5.4 Three plots of geophysical data from surveys image is a GPR “time slice” of the same area showing
of the site of the Cherokee Female Seminary, good correlation between the two sensors. Architec-
Tahlequah, Oklahoma. The upper plots are, respec- tural interpretation has been added to the radar
tively, a plan view of quadrature phase data (330 Hz) plot. The area outlined in red is buried ruin of the
and the isometric view of the same data. The lower building
5.6 Conductivity (EM) Survey of a Burned and Buried Cherokee Seminary Building. . . 127

Fig. 5.4 (continued)

Iron-bearing minerals such as goethite, hema- USS Cairo (Bearrs 1966). Geophysicists have
tite, and magnetite (cf. Tables 5.2 and 5.3) are developed several magnetometers for use in
common in many soils that contain archaeolog- prospection. These include the fluxgate magne-
ical features. Till et al. (2015) have described soil tometer, proton precession magnetometer, and
mechanisms that increase the presence of the the optically pumped magnetometer.
more magnetic of these minerals—magnetite Magnetometers measure the sum of the
and maghemite. Conversion, by chemical/recrys- induced and remanent magnetization in some
tallization pathways, of less magnetic goethite absolute or relative ratio to the earth’s field
and hematite forms together with either natural depending on the instrument (LeBorgne 1960).
or anthropogenic fires to magnetite/maghemite Fluxgate models measure this along a single axis,
has been documented for soils where organic while the other types measure the total or abso-
matter acts as a reductant (supra). lute field—the sum of all axes of the magnetic
field at a location. The single axis fluxgate
readings can be subtracted from the total field
5.6.2 Magnetometers to obtain the remnant component that is
often directly representative of archaeological
Magnetometers, unlike conductivity and resistiv- features. Absolute or total field devices, proton
ity devices, which actively generate electric precession or optically pumped systems, are
fields are “passive” instruments. By comparison also used to detect local magnetic materials—
the simplest device for measuring magnetic fields artifacts and architectural remains, for instance.
is the compass. While this hardly seems to
qualify this important directional tool as a Fluxgate Magnetometer This magnetometer is
prospection instrument, it was used to success- simply a magnetic metal around which two coils,
fully relocate the sunken Civil War ironclad, primary and secondary, are wound to form a
128 5 Geophysical Techniques for Archaeology

Table 5.2 Magnetic susceptibilities of common rocks be measured. The magnetometer can measure
and minerals (cgs units) all three components of the total magnetic
Top soil 104 field’s X, Y, and Z orientation. Because of the
Clay (unfired) 105 complexity of the sensor(s) output, a micropro-
Altered ultrabasic rocks 104–102 cessor is needed. In practice the x-axis is fixed to
Basalt 104–103 north and the Y and Z fluxgate sensors held in
Gabbro 104
some fixed orientation to X. The data can be
Granite 105–103
displayed as to (a) the north-south component
Andesite 104
(X) and the east-west component (Y) and the
Dolerite 102–0.1
Rhyolite 105–104 vertical component (Z).
Shale 105–104
Schist and other Proton Precession Magnetometer This type of
Metamorphic rocks 104–106 magnetometer is perhaps the most common field
Most sedimentary rocks 106–105 system used in archaeological prospection.
Limestone and chert 106 Conceptually, it is somewhat more complex
Pyrrhotite 103–1 than the fluxgate design. As the name implies,
Magnetite 103–2.0 the measurement of the magnetic field strength is
Chromite 104–1 through use of proton (the hydrogen nucleus)
Hematite 106–105 precession. Precession is a property of electrons
Pyrite 104–103 and protons due to its spin (Leute 1987). An
individual proton will align, as its spin generates
Table 5.3 Magnetic properties of soil minerals a very weak magnetic dipolar moment, along the
Mineral Formula Magnetism direction of a magnetic field such as the geomag-
Magnetite Fe3O4 Ferrimagnetic netic field of the earth.
Hematite α-Fe2O3 Antiferromagnetic Protons, such as in a fluid—water, gasoline, or
Maghemite γ-Fe2O3 Ferrimagnetic alcohol—can act in a collective manner, simi-
Goethite α-FeOOH Antiferromagnetic larly to that of ferromagnetic domains. A proton
Lepidocrocite γ-FeOOH Antiferromagnetic precession magnetometer has as its sensor, a
Ilmenite FeTiO3 Antiferromagnetic container of a proton-rich fluid, encircled by
Pyrrhotite Fe7S8 Ferrimagnetic wire windings that carry a current. The current,
Pyrolusite MnO2 Antiferromagnetic 1–1.5 A, generally, can generate a locally brief
yet strong magnetic field that “polarizes” the
transformer. In operation the introduction of AC protons. After the field is turned off, the protons
in the primary coil produces (induces) a mag- precess together until their motion decays into
netic field in the core metal as a secondary volt- random modes, about the field lines (of magnetic
age or waveform. Any external field along the force) of the geomagnetic field. There is a known
axis of the core can alter the waveform propor- frequency for this precession about the geomag-
tional to the strength of the magnetic field at a netic field that is termed the Larmor frequency,
sensitivity of 1 part in 50,000 nT, i.e., 1 nano- using an average value of the earth’s field of
Tesla (nT). If the field vector is off the axis of the 50,000 nΤ; it is equal to 2,100 Hz. The protons
fluxgate device, then it is not as readily detected. processing in unison, at or near the Larmor fre-
Fluxgate devices are popular in Great Britain quency, induce a weak electrical current in
and the continent (Milsom and Eriksen 2011; the sensor’s windings which, when amplified,
Marshall 1999; Neubauer and Eder‐Hinterleitner converts the voltage’s frequency to direct mea-
1997). sure of the magnetic field strength to about
A three-component fluxgate magnetometer 0.02 Hz or 0.5 nΤ (Leute 1987).
has been described by Kamei et al. (1992). Cou-
pling two of these devices creates a gradiometer Optically Pumped Magnetometer In the geo-
wherein both the total and gradient fields can magnetic field, electronic levels will have their
5.6 Conductivity (EM) Survey of a Burned and Buried Cherokee Seminary Building. . . 129

orbital “degeneracy” removed such that a dis- feature more strongly than the upper sensor. The
crete energy, of a few electron volts, exists upper sensor can be thought to be affected more
between the + and  energy states of an atomic by the magnetic field of the earth rather than the
electron. This effect was discovered by Pieter feature. By subtracting the readings of the paired
Zeeman in the nineteenth century and was sensors, the gradient is obtained. In modern
named for him. The use of the Zeeman effect instruments this is done electronically and digi-
created a third type of magnetometer called the tally displayed. To get a similar result, two
optical-pumped or absorption cell magnetome- magnetometers are often used with one
ter. The electrons of a specific element such as remaining at a fixed location, the base station,
sodium, cesium, or rubidium can move between and a second unit, the mobile unit, used to con-
the Zeeman levels, generally from ground (low- duct the survey. At each location the readings,
est) state to higher energy states by absorption of taken simultaneously, are subtracted to give what
appropriately polarized light protons (hν). is termed the differential magnetic reading. This
Optical or absorption cells, filled with a vapor differential reading subtracts the magnetic field
or sodium, for example, will experience reading of the base station from the mobile
“pumping” of ground state electrons into higher reading. It is similar but not quite the same as
Zeeman energy states until the ground state is the gradiometer reading.
depleted. Absorption stops as the cell becomes
opaque. Absorption is re-initiated by applying a
weak high-frequency radio field to the cell,
5.6.4 Magnetic Anomaly
and the higher electron energy states will
Interpretation: Basic Principles
“depopulate” and cascade back down to the
ground state. Optically pumped magnetometers
Magnetometers measure the field that results
have been in use almost as long as proton preces-
from the sum of the magnetic field strength
sion and fluxgate models. A recent use of the
induced by the earth’s field (average value ¼
Zeeman effect in the proton precession design
50,000 nT) and any additional permanent mag-
takes advantage of a resonant proton-election
netism or remanence. The remnant field compo-
coupling in an organo-electrolyte (paramagnetic)
nent of the total field, at the measurement point,
that increases the magnetometer’s sensitivity to
is often off axis to that of the earth’s field. Mag-
0.01 nT. The effect is termed the Overhauser
netic features are termed “anomalies” in the par-
effect and the device based on this effect is the
lance of geophysics. Unless the anomaly has a
Overhauser-Abragam magnetometer (Milsom
strong remnant component, such as a hematite
and Eriksen 2011).
ore body or a buried clay hearth or kiln, most
features found by archaeological prospection are
the result of magnetic induction due to higher
5.6.3 Gradiometers than average amounts of ferri- and ferromagnetic
minerals and their increased magnetic suscepti-
Almost any type of magnetometer—proton, flux- bility. Magnetic prospection relies on the con-
gate, and optical—can be configured to function trast in terms of magnetic susceptibility among
as a gradiometer. Two sensors are paired at a buried materials that is expressed as k and total
fixed distance apart such that a horizontal or magnetism is M. Less rigorous treatments reduce
vertical gradient exists between the pair. This to semiempirical characterizations of the
magnetic gradient or separation is small— anomaly’s strength such as
0.5–1.0 m—compared to most magnetic features.
T ¼ k M=r n  dimension=shape factor
Gradiometers react to near-surface anomalies,
with the lower sensor, if the instrument’s sensors Here T is the anomaly strength, r is the distance
are aligned in a vertical mode, detects the buried to the anomaly from the sensor, and the
130 5 Geophysical Techniques for Archaeology

exponent, n, is the rate of decay of magnetic to the south of the actual feature at a distance
strength or falloff factor and a shape or dimen- equal to one-third the depth of the feature.
sion factor. M and k are as before. This expres-
sion is reduced to
5.6.5 Magnetic Prospection: Practice
T ¼ M=r n
Many “successful” magnetic surveys have been
Field inclination is indirectly considered by its
done in archaeology without a great deal of atten-
effect on total magnetic intensity such that:
tion paid to first principles. In those surveys the
T ¼ 2 M/r 3 for a dipolar anomaly along its results were judged on the presence or absence
axis; T ¼ M/r 3 for a dipolar anomaly, off axis; together with strength of anomalies. Indeed with
and T ¼ M/r 2 for a monopolar anomaly, any the first magnetometer models, this was about all
orientation. Field measurement of I is difficult one could expect before reliable digital measure-
and is most commonly estimated, e.g., 60 incli- ment of the field strength was added. Martin
nation at 30 N latitude. If I is known or Packard and Russell Varian are credited with
estimated, a simple vectorial solution for the the development of the proton precession instru-
anomaly strength can be obtained. ment in 1954 (Weymouth 1996). By 1956 John
The falloff factor is important. The field of a Belshé demonstrated that a replica of an ancient
dipole falls off as the third power of the distance. kiln produced a magnetic moment. Martin
For a monopole this is only by the square of the Aitken and Teddy Hall, of Oxford, constructed
distance. In practical terms this implies that an a proton-based unit and located an actual kiln
anomaly’s strength, measured at 81 nT, directly buried in a Roman site (supra). Linnington, at
over the feature, will fall to 3 nT at a distance of the Lerici Foundation in Rome, surveyed the
1 m. By way of comparison, active EM devices, Greek site of Metapontum in 1968 and began
such a metal detectors, have a total falloff factor the modeling of magnetic data. By the 1960s,
of r6! workers, such as Scollar, in Bonn, were doing
One, even if assuming a purely vertical or automatic data collection and applying
horizontal field direction, can ascertain much computers to the processing and display of that
about the nature of an anomaly by observing data (Scollar et al. 1986). Workers such as
the rate of falloff and the sign of the anomaly Aitken, Linnington, Tite, Scollar, Ralph, Beven,
along with its strength. Anomaly shapes change Weymouth, and others rapidly moved magnetic
with an increase/decrease in sensor height and prospection, in archaeology, forward.
field direction. Depth estimation is likewise Understanding the susceptibility contrasts of
important. One can use the straightforward rela- materials greatly enhances the predictability of
tionship seen in the above formulae or use “rules what one can hope to detect in differing field
of thumb” or nomograms. Milson estimates the situations. For instance, a contrast of 102 versus
accuracy of these latter methods to within 30 % one of Δ k ¼ 104 results in strikingly different
(Milson and Eriksen 2014). One method is anomaly strengths, i.e., 32 vs 0.32 nT. Another
the “full-width half-max” rule (FWHM). The factor, not mentioned in the discussion of
anomaly’s maximum value for Z (field strength) principles, is the daily or diurnal variation in
is used to determine those points on the x-axis the earth’s field. This is a real and fairly regular
where the value falls to Z/2. The distance effect observed over the course of the day. This
between the two points of Z/2 is proportional to change is observed; field strength is a result of
the approximate depth of the anomaly. For a ionospheric currents that grow in strength over
spherical dipole, the depth is ~x/2 and for a the day, with solar heating, and subside over-
monopole the depth is ~1.3 x/2. In the night. The pattern, in mid-latitude areas, follows
mid-latitudes of the Northern Hemisphere, the an increase in field strength to 11 AM decrease of
maximum value of an anomaly tends to occur ten or more nanoTeslas toward dusk and then a
5.6 Conductivity (EM) Survey of a Burned and Buried Cherokee Seminary Building. . . 131

steady increase up to midnight. Again the As just discussed geomagnetic and ionospheric
increase is in the low terms of nT. Long-duration factors are accounted for but not chosen before-
surveys must account for the diurnal variation. hand. Unlike other geophysical surveys, archae-
This is typically done by using the differential ological prospection involves more grid-area
arrangement discussed previously. These survey on closely spaced (0.5–2 m+) lines. The
changes are subtracted from the data to normal- spacing of grid intervals is largely determined by
ize the AM and PM components. If a gradiometer the size of the feature(s) being sought and sus-
instrument is being used, then such a correction ceptibility contrasts. For small, discrete objects
is less necessary. Another common factor that of low-to-moderate magnetic contrast, the grid
intrudes on long traverses (>100 m) is the earth’s should rarely exceed 1 m in interval/offset
field gradient. This is a consequence of running spacing. For larger features such as structural
most profiles in a N-S direction. Those readings remains, roadways, etc., the interval can be
most northward will gradually increase as a func- relaxed to 5 and even 10 m spacing if the area
tion of the normal increase in the earth’s field to survey is large.
strength as one approaches the poles. It is easily On the average, one team of two to three
observed as slight change of profile slope and individuals can survey a 20  20 m area on a
subtracted from the data. 1 m interval in a 6–8 h period. Speed in these
The two principal magnetic methods, magne- surveys is determined by the cycling time of the
tometer and magnetic susceptibility, have proven instrument which is basically a nonfactor with
successful in the search for archaeometallurgical any modern instrument given that the operator
sites and their features (Powell et al. 2002). can occupy a point and measure, move, and
The former, magnetometer, “proton,” repeat the sequence in a typical one-half second
“optically pumped,” or “fluxgate,” is considered cycling time of a precession device. Optically
“passive” in that the instrument produces no pumped and fluxgate systems read continuously
physical perturbation of the soil or feature and so one just walks, reads, etc. This aspect of mag-
only detects the magnetization by means of sen- netic survey is its great advantage. Compared
sor coils. Magnetic susceptibility systems—sus- even to the two-probe resistivity array, magnetic
ceptibility meters and electrical conductivity surveys in terms of time make it a tortoise-hare
meters—are “active” instruments in that they situation. This time the hare wins.
use an alternating current in a coil to induce or Other techniques comparable, in terms of
measure secondary magnetic fields in a feature speed, are some EM survey devices and, in
(Florsch et al. 2011). As we discussed in the more and more instances, GPR. Data produced
foregoing treatment on conductivity, induction by magnetic prospection is three-dimensional,
is measured as a function of the applied fre- that is to say, x, y, z, or “triplet form.” The z
quency or frequencies. In an excellent example value is that of the magnetic field. Because
of one lesser used geophysical methods, Florsch mapping software today commonly uses triplet-
et al. (2011) utilized induced polarization for the form data, it is an easy thing to substitute mag-
study of slag heaps. netic field strength for altitude producing both
(a) contour and (b) isometric plots of the data.

5.6.6 Date Acquisition and Display

5.6.7 Advantages and Disadvantages
Magnetometer surveys, as with most of the other
techniques discussed in this chapter, are done on Speed of survey, relative ease operation, rapid
orthogonal grids of straight headings and fixed computer-based display, and manipulation of
offsets between (traverse) lines. Other than sen- large data sets have made magnetic prospection
sor height, these are the only real variables in a one of the most widely used of today’s geophys-
magnetic survey that an investigator can control. ical techniques. The physical principles are well
132 5 Geophysical Techniques for Archaeology

understood drawing on the depth of research into was undertaken by the University of Georgia’s
magnetism at both the micro- and macrolevels. Institute for Native American Studies (INAS),
The cost of instruments has become less of a which conducted a multisensor geophysical sur-
problem with the increased volume of devices vey of the historic Rose Cottage site and its
available. vicinity in August 2013.
Of the most prominent types of The most dramatic feature seen in the radar
magnetometers—proton precession, optically data is shown in Fig. 5.6. The rectangular feature
pumped vapor cells, and fluxgate devices—only that is just below the present ground surface
the latter are relatively immune to external noise continues to some depth, over a meter, and is
and strong local field gradients. Even these clearly a “man-made” or anthropogenic feature.
devices may have problems in some volcanic It is 40 m in overall length and more than 15 m in
terrains. With the exception of the fluxgate width. The north or upper portion of the feature
systems, most magnetometers cannot be used too extends beneath the elevation termed the “Old
close to power transmission lines or large Road.” The survey grid did not extend beyond
concentrations of iron and steel. A problem, but the edge of this latter topographic feature.
not one of the magnetometer itself, is the interfer- Within the boundaries of the large rectangular
ence caused by modern metal commonly found on radar feature, large and sharply delineated mag-
more ancient archaeological sites. In these cases netic features are clearly seen (Fig. 5.7).
the use of EM devices, in conjunction with
magnetometers, has merit. Neither surface or
near-surface metal debris nor strong field gradients 5.7.1 Ground-Penetrating Radar: GPR
significantly impact resistivity survey techniques.
The publication of textbooks (Goodman and
Conyers 1998; Conyers 2012) exclusively focus-
5.7 An Abandoned Nineteenth- ing on the use of fairly recently developed geo-
Century Cherokee Cemetery, physical technique in archaeology highlights the
Park Hill, Oklahoma, Cherokee rapid incorporation of GPR into archaeological
Nation, USA prospection. At the end of the century, after only
slightly over two decades, GPR has risen to a level
Rose Cottage (Figs. 5.5 and 5.6) was the resi- of recognition among archaeologists previously
dence of Chief John Ross, principal chief of the reserved for magnetic or resistivity instruments.
Cherokee Nation, 1828–1866, after the tribe’s This newfound popularity of GPR in archaeology,
forced removal from tribal lands in Georgia, in environmental studies, geotechnical and con-
Tennessee, and North Carolina, in 1838–1839. struction areas of engineering, as well as geology
Located somewhere near the site, the has led to new and older journals alike, devoting
nineteenth-century home, the Rose Cottage, of entire issues to the applications and methods of the
Principal Chief John Ross, the “Forgotten Ceme- technique (Geophysics 1986).
tery” grid was suspected to lie under mown,
grassy pasture bordered by a low elevation on Principles GPR is an electromagnetic geophys-
its north side by the so-called Old Road which ical technique where a radar beam (wave) is
connected the Rose Cottage with the present-day transmitted into the subsurface and, depending
Park Hill Road. Mentioned in a 1937 Elizabeth on different electrical properties of materials in
Ross interview as “One of the oldest burying the subsurface, that beam can be diffracted and
grounds in Cherokee County, now altogether reflected at objects and interfaces (Leute 1987)
obliterated, once lay near the north bank of the (Fig. 5.7). The strength and the travel time, the
Park Hill branch, several hundred yards northeast time delay, between transmitted and returning
of the Campbell Spring.” Relocating the Rose pulses, tt, of radar reflections can be plotted by
Cottage site as well as the nearby burial ground recorders and converted to depth and dimension.
5.7 An Abandoned Nineteenth-Century Cherokee Cemetery, Park Hill, Oklahoma. . . 133

Fig. 5.5 Rose Cottage, from a painting of the house done from memory. Oklahoma Historical Society

Fig. 5.6 A GeoPDF showing the location of UGA survey grids and other landmarks at the Rose Cottage site

GPR “sees” through freshwater, ice, soil, rock, plus 50 % above and below the center frequency
brick, and paving although highly conductive makes up the chirp’s range. For example, a
materials such as seawater, clays, and metal atten- 100 MHz center frequency includes frequencies
uate or occlude radar propagation. The radar down to 50 MHz and up to 150 MHz. The general
pulse is transmitted as a short burst of EM energy rule in estimating GPR depth of penetration is the
over a compact range of frequencies called a lower the frequency, the deeper the potential sur-
“CHIRP.” The center frequency of the antenna vey depth. Because GPR operates at relatively
134 5 Geophysical Techniques for Archaeology

Fig. 5.7 Multisensor representations—left and above, view is 1.74 m depth and clearly shows the
magnetic; right, ground radar—of an abandoned rectangular subsurface feature interpreted as a rem-
Native American cemetery, Park Hill, Oklahoma, nant of a stone or masonry wall. The magnetic
Cherokee Nation, USA. The magnetic anomalies cor- features (orange red) in the plots show linear
respond to rows of grave pits while the radargram patterns. The separation between the anomalies indi-
indicates three sides of a buried stone wall. The latter cate individual graves

high frequencies 20–1500 MHz, compared to sites, this range is quite adequate as the bulk of
seismic frequencies where “high” is considered most habitation sites seldom exceed this depth
to be in the range of 100 Hz, penetration beyond a range with the exception of tells and other deeply
few tens of meters in all but the most favorable found settlements. Even in these sites of excep-
material—rock generally—is uncommon. For tional depth, the antenna with a frequency appro-
most antennae, used in archaeology, the depth priate to the survey can be found.
range is 5 m or less, the working limit of The resolution of a specific radar frequency is
400–500 MHz models. For most archaeological a direct function of wavelength, which, in turn, is
5.7 An Abandoned Nineteenth-Century Cherokee Cemetery, Park Hill, Oklahoma. . . 135

a function of the velocity of radar wave in a Table 5.4 Typical values of radar parameters of com-
material. Velocity relies on the value of electrical mon materials
permittivity or more termed, the dielectric “con- Material ε σ (mS m1) v (m ns1) χ
stant” (coefficient), ε. With EM techniques, it Air 1 0 0.30 0
governs attenuation in both cases where induc- Ice 3–4 0.01 0.16 0.01
tion currents extract energy from the radar pulse. Freshwater 80 0.5 0.033 0.1
The conductivity of materials generally follows Saltwater 80 3,000 0.01 1,000
Dry sand 3–5 0.01 0.15 0.01
an increase in the dielectric constant, but this is
Wet sand 20–30 0.01–1 0.06 0.03–0.3
not so direct a relation as to have seeming
Shales and 5–20 1–1,000 0.08 1–100
contradictions where fresh- and seawater share clays
the same value (ε ¼ 80) yet have extreme Silts 5–30 1–100 0.07 1–100
differences in conductivity (σ ¼ 0.5 vs 3,000). Limestone 4–8 0.5–2.0 0.12 0.4–1
This contradiction is better understood when the Granite 4–6 0.01–1 0.13 0.01–1
electrolytic properties—ionic vs nonionic—are (Dry) salt 5–6 0.01–1 0.13 0.01–1
considered. Permittivity is a direct reflection of
the molecule’s ability to polarize. The typical R, the reflection coefficient, is related to the
values of ε, σ, and velocity (v) as well as attenu- soil permittivity, ε, of the upper media, ε1, and
ation factors are shown in Table 5.4. the lower media, ε2, by
The wavelength and velocity are given by the  
R ¼ ε1 ð1=2Þ  ðε2 Þ1=2 =½ðε1 Þ1=2 þ ðε2 Þ1=2
This equation can be used to determine the
υ ¼ 1= pμε
ffiffiffi pffiffiffi
strength of reflections at interfaces of materials
λ ¼ υ=f ε and v ¼ c= ε of different lithology or composition.
λ¼υ=f An easy way to increase penetration and sig-
nal return is to increase power on the outgoing
c is the velocity of light in free space (0.30 m ns1), pulse. Increasing power (or gain) of the radar
υ is the velocity of the EM wave in the medium, f is pulse can create other problems such as loss of
the frequency, ε is the dielectric constant (permit- detail in the upper areas of the trace of profile and
tivity), and μ is relative magnetic permeability refraction multiples from objects and secondary
which in most materials (except metals) is close images such as hyperbolae and extraneous noise
to l. Most calculations assume it to be unity. Veloc- in the record. Many digital GPR units as well as
ity and wavelength change with material giving post-processing softwares allow the operator to
radar its ability to image the subsurface. For exam- enhance gain losses in a GPR pulse somewhat
ple, in air, a 100 MHz pulse wavelength is 3 m; in selectively. The GPR trace is radar waveform
rock with a velocity reduced to 0.1 m ns1, it is that is perturbed as it penetrates through subsur-
only 10 cm. In wet clay this will drop to about face materials such that the returning waveform’s
1–2 cm. As we have said, resolution is a function shape provides us information about the earth’s
of wavelength so features, in rock, can be resolved shallow near surface (Fig. 5.7).
to 10 cm but no smaller as that dimension is inside A GPR profile display (Fig. 5.8) is simply a
the wavelength and thus subject to refraction and series of traces, in series, taken over a traverse of
blurring. In this the radar wave mimics other EM some fixed distance using a set timing interval.
waves. Another general rule for GPR resolution is The height of the GPR profile is a set time inter-
that the size of the object should be not less than val in nanoseconds. Typically, one estimates the
one-tenth of its depth (Milsom and Eriksen 2011). average velocity for the soil at the site to be
As a result a 10 cm object cannot easily be detected surveyed and this is used to set the time “window”
at over 1 m unless the reflection, in terms of signal or interval. Dry sand’s velocity is 0.15 m ns1
return, is exceptionally good, e.g., a good dielec- so a 100 ns (tt) window represents about 9.5 m
tric contrast and low attenuation. depth. Before presetting an unrealistic depth for
136 5 Geophysical Techniques for Archaeology

Fig. 5.8 A profile view of radar data. The image is of a of an actual archaeological site on the island of Santa
Chumash pit dwelling, roughly 2 m in width, as illustrated Cruz, California, USA, where the author conducted
in the artist conception adjacent. This is an idealized view research on the late twentieth century

e s
1 0.00000
5 0.00100
10 0.00100
15 0.00100
depth 25 0.00100
40 0.00100
81 0.00010

0 range (m) 5.0

81 0.00010
40 0.00100
25 0.00100
15 0.00100
10 0.00100
5 0.00100
1 0.00000

0 range (m) 5.0

Fig. 5.9 The effects of refraction on the radar wave as the dielectric increases (upper) or decreases with depth
(lower) (Adapted from Goodman 1994, Fig. 5.6)

the particular antenna type/frequency, one can depth. Here, σ is the conductivity emphasizing,
estimate the realistic depth of attenuation called again, its important role in attenuation of the GPR
the skin depth by the empirical rule, 35/σ ¼ skin signal (Fig. 5.9).
5.7 An Abandoned Nineteenth-Century Cherokee Cemetery, Park Hill, Oklahoma. . . 137

5.7.2 Field Survey Methods direction, a GPR must be rapidly recalibrated for
comparable data across the complete grid. This is
GPR, in seismic parlance, is a “zero or fixed routine and requires little in the way of resetting
offset” profiling technique, “zero” because in of instrumental parameters. Generally the surface
some antennae, using the so-called bow-tie zero set and gains are all that require adjustment.
dipole, the transmitter and receiver are in the Once a routine is established, and environmental
same antenna. conditions allow, survey progress can be steady
Two antennae may be coupled, in bipolar or and consistent. One GPR survey conducted in the
bistatic configuration, such that one is the trans- later 1990s by a University of Georgia team
mitter and other the receiver yielding a small covered over eight linear km of traverses in a
(1–2 m) fixed offset. These offsets are trivial 5-day period. This time would be significantly
compared to those used in seismic reflection or less today—1 day of less—using multichannel/
refraction survey where offsets can be in antennae arrays.
hundreds of meters across a geophone array. A basic grid area is 100  100 m surveyed
The antenna is either carried, stepped, or pulled using a single frequency antenna on a 1 m offset
across the ground along the preset traverse. In interval (Fig. 5.10). A factor that has led to
some cases where exceptional S/N is required, GPR’s rapid acceptance in archaeology is the
the antenna may be “stepped” along the survey immediate return on survey effort. GPR is a
line at preset intervals. This allows “stacking” of “WYSIWYG” technique—what you see is what
the data to be done at each station, along the you get. The GPR record is a digital facsimile of
traverse, thus increasing the S/N. Most field sur- the subsurface immediately below the antenna
vey requires a minimum of two personnel—one along the direction of the traverse shown
the control unit operator and the other in charge on-screen and most often recorded as a file.
of the antenna (“the mule”). Such records are seen in Fig. 5.10. For more
Grid size is important, particularly today, with detailed analyses, these files are evaluated with
varieties of spatial profile and three-dimensional the aforementioned post-processing routines
plots available on computer programs. Unless back in the lab.
specific parameters are maintained, the applica- The linear feature along the left axis of both
tion of this data processing tools is obviated. One plots is interpreted as a water channel or aque-
meter offset, for transverses, have become a duct serving the site (Garrison et al. 2013, 2014).
default interval for most archaeological GPR
surveys. Adjacent traverse lines have a good
probability of crossing features at or near this 5.7.3 Digital Post-Processing
size within the grid. Halving the interval doubles of GPR Data
survey time, but almost all features at or above
this size will be detected. The depth of survey is As illustrated in Fig. 5.11, much of the enhanced
preset as we have discussed. If the site is shallow, interpretability of GPR records have come
1–2 m, in overall thickness, then it is important to through the use of digital processing similar to
set the time window to be below this, say that developed for acoustical seismic methods.
2.5–3.0 m range in ns. For dry sand, this is roughly Once the GPR data is logged in this format, it can
130–200 ns, for a velocity of 0.15 ns m1. This is be readily manipulated to remove background
not an impractical setting for antennae frequencies noise and enhance gain in weak returns by
of 100–500 MHz. reverse time-varying gain amplification, together
Orthogonal grid directions can be surveyed. If with the use of other filters—FIR (finite impulse
time permits, it should be attempted. Linear response)—as well as those to compensate for
features, which the antenna may just parallel on spherical effects in propagation. Advanced
a traverse, can be bisected by a crossing pattern seismic-style algorithms such as “boxcar filter-
survey. At the start of each traverse, no matter the ing” and “Kriging” together with Fourier
138 5 Geophysical Techniques for Archaeology

Fig. 5.10 A typical survey grid and GPR units; fore- antenna, on a survey cart. Tapes indicate transect lines
ground, a control unit and antenna using a survey wheel; (Photograph by the author)
background, another control unit mounted, along with an

Fig. 5.11 Examples of post-processed radar data. Both depth at almost a meter, while the plot on the right
represent a 2004 radar survey of a Roman villa site in highlights architectural details of the villa’s principal
western Switzerland. The plot on the left shows the burial building, the pars urbana
5.7 An Abandoned Nineteenth-Century Cherokee Cemetery, Park Hill, Oklahoma. . . 139

transform of frequency to time series are localized archaeological features such as tombs
available which, in turn, prepares the data or burials may require the use of more than one
for deconvolution—which removes peg-leg geophysical technique to adequately relocate
hyperbolae and multiple reflections from the them in many instances.
record. Migration, a common seismic technique,
can be applied to profiles that are adjacent and
are closely enough spaced (Meats 1996). The 5.7.5 Underwater Geophysical Survey
reader is directed to Goodman and Conyers Techniques in Archaeology
(1997, 2013) for a good introductory review,
while the advanced student is directed to their The use of geophysical techniques in underwater
bibliography and journals such as Geophysics, archaeology began in the 1960s (Tite 1972;
Applied Geophysics, Journal of Archaeological Aitken 1974). These rapidly moved to assemble
Science, Archaeometry, Environmental and multiple sensor arrays as these became feasible
Engineering Geophysics, and more recently, over the subsequent decades. Today, in geophys-
Archaeological Prospection. ical surveys of shipwrecks and landforms, these
A useful iteration of GPR data is the so-called ensembles typically include (1) a towed marine
time slice technique (cf. Goodman and Conyers magnetometer, (2) a 100–500 MHz side-scan
1997, 2013). This processing method combines sonar, and (3) an acoustic “sub-bottom” profiler,
data from an adjacent profile at fixed intervals of more commonly of a CHIRP, swept-frequency
time such as 30 ns, 45 ns, 60 ns, etc. Using a type. In some instances, a towed pulsed EM
protocol similar in form to GIS methodology system is used in place of or in conjunction
where sequential overlays or “slices” are built with the magnetometer. Survey tracks are
up gives the viewer an overhead view through pre-plotted and tracked using high-precision
the site as visualized in the GPR data line-of-sight radar transponder systems or more
(cf. Figs. 9.7 and 9.10). and more frequently with reference to GPS. Posi-
tional accuracy in dynamic survey mode regu-
larly approach 3–5 m in accuracy. Anomalies,
5.7.4 Multiple Geophysical acoustic, magnetic, and electrical, are logged
Techniques along-track to strip charts and digital files. The
multiple data sets can be post-processed and
Hesse (1999) has written succinctly about reconciled just as those done terrestrially.
the combined use of multiple geophysical Two techniques that are unique to underwater
prospection techniques (“sensors”) for archaeol- survey are acoustic in nature. Deep seismic stud-
ogy. Another excellent example is that of the ies of the offshore for minerals have always been
geophysical study of Roman Grand, in the acoustic. Their detail, or resolution, of the sea-
Lorraine region of France, presented in a theme floor and near sub-seafloor was too large scale for
issue of Les Dossiers Archeologie (1991). The archaeological purposes. The late Harold
point Hesse makes is that the efficacies and short- Edgerton, of the Massachusetts Institute of Tech-
coming of individual geophysical techniques nology (MIT), developed a viable acoustical
may to a large degree offset one another in the seismic profiler by directing a beam of a piezo-
latter case and enhance interpretability of electric transducer of an appropriate frequency,
anomalies in that of the former. Multiple sensor 2–7 kHz, straight down to allow the beam to
studies are to be attempted wherever possible. profile the upper sediments of sea bottom. With
Cost and availability are prime considerations. sound velocities in water, ca. 1,500 m s1, a
The scale of the study is a key factor. Deploying 3.5–5 kHz transducer produces a waveform in
expensive multiple sensor on relatively small lengths of 30–50 cm. Such wavelengths can reg-
sites may be a case of instrumental “overkill.” ularly detect objects of archaeological interest
In seeming contradiction, important, very much like GPR does on land.
140 5 Geophysical Techniques for Archaeology

The waves are pulsed or pinged by the trans- on the continental shelf of North America will
ducer(s) activated by an applied voltage across archaeological sites dating from the Archaic
condensers. Acoustical impedance differences Period (8,000–3,000 B.P.) occur and be pre-
caused by the relative compactedness or served?” (Stright 1986a, b, 1990, 1995a, b). Her
looseness of the materials appear as returning comparative analysis of published relative
signals or reflections of differing gray scale— sea-level curves for North America indicated
dark for hard and light for unconsolidated that Archaic Period archaeological sites may
materials. The water is transparent to the sound exist as deep as 22 m below present mean sea
waves so the soil/water interface is easily seen. level in portions of the northern Gulf of Mexico.
The returning waves, although weakened, are This early effort at submerged archaeological
collected by the transducer and amplified for site discovery relied on a research strategy
graphical display on screen or recorder. Measur- based on actualistic and archaeological
ing the signal (reflector) and travel time (l ms analogues (Pearson et al. 1982) together with a
~2.7 m at 1,500 ms1), the depth and thickness of series of technological approaches that have
a sediment reflector can be measured accurately. since become standard in this field of study—
Like GPR the waves tend to refract and reflect sub-bottom profiling, side-scan sonar, sediment
more than once between the sea or lake floor and coring, etc.
the surface generating multiple reflections. These The study area was the now-filled and
are easily detected in the acoustical record. Indi- submerged stream valley of the Sabine River
vidual features—rocks and buried archaeological offshore of Louisiana and Texas (Fig. 5.12).
objects—appear as diffraction hyperbolae. If the The research involved the collection and synthe-
object is metal, its character can be rapidly noted sis of a large amount of high-resolution seismic
by use of other techniques, magnetic or EM. data and core records in order to reconstruct
Edgerton (1986) went on to quickly design the the prehistoric, subaerial landscape of the
side-scan sonar after the “sub-bottom” profiler in region. Models of prehistoric site distributions
the mid-1960s. The initial analogue models have derived from onshore analogues were extended
been supplanted by digital systems, some whose to the offshore landscape to identify the
design traces back to Edgerton. The side-scan now-submerged landforms with a high likelihood
sonar, by immediate reference to the output of containing preserved cultural remains. These
acoustical frequencies of 100–500 kHz, as are presages, by over two decades, are the so-called
commonly used in archaeological prospection, Danish model popular in the literature of prehis-
does not attempt to penetrate the sub-seafloor. toric underwater archaeology today (Benjamin
Rather, the purpose is to produce a horizontal 2010). Early Holocene climate change, reflected
terrain map on both sides of towed body mount- principally in sea-level change, was incorporated
ing parallel transducers depressed downward at into the predictive model of site distribution
angles of 3–10 . and preserved landform occurrence in the off-
shore area of interest. More than 70 vibracores
were used to collect sediment samples from
5.8 The Sabine River Paleovalley, several geophysically identified high-probability
Northern Gulf of Mexico, USA landforms (Fig. 5.13). Analyses indicated the
existence at least one location of cultural deposits
In 1984, over a quarter of a century ago, Charles dating to ca. 8,800 BP. Subsequent research in
Pearson, Sherwood Gagliano, Richard the region has employed a similar strategy in the
Weinstein, and David Kelley undertook a study search for submerged prehistoric deposits.
designed to locate submerged prehistoric archae- The Archaic Period can be called an “American
ological deposits on the continental shelf of the Mesolithic Period” because of the extensive
Gulf of Mexico. This research was predicated on similarities between the cultures of both Eurasia
the question, raised by Melanie Stright: “Where and the Americas. Hunting-collecting economies
5.9 Doggerland: Mesolithic Landscapes of the Southern North Sea 141

Fig. 5.12 Study area, in red, for the Sabine River paleovalley study, northern Gulf of Mexico, USA

dominate the post-Holocene. The start of the Kingdom and is confined in latitude between
Mesolithic, in the Early Holocene, is generally 55 and 51 north. The Southern Sector covers
regarded as synonymous with two major techno- 6250 km2 (approximately 1 %) of the total area
logical developments. The first is the development of the North Sea (Fitch et al. 2007). To examine
of small armaments such as the bow and arrow as such a large submarine area requires large
opposed to spears and spear throwers of the late geophysical data sets. The utilization of spatially
Paleolithic/Paleo-Indian period. Microlithic tools extensive oil industry data has allowed the
and the stone elements used to tip arrows appear recovery of information pertaining to the
for the first time at this period (Wickham-Jones actual Mesolithic landscape of the North Sea
2005, 2010; Sassaman and Anderson 1996; aka “Doggerland” (Gaffney et al. 2007)
Sassaman 2004; Thompson and Worth 2011). (cf. Figs. 2.6 and 2.7). This information reveals
The late twentieth-century underwater study of the diversity of this landscape and shows that
the prehistoric northern Gulf of Mexico has much much greater consideration of submerged
in common with the early twenty-first-century Mesolithic landscapes is now required of
study of the prehistoric North Sea Region (Fig. archaeologists. Indeed, Caroline Wickham-
5.14). Jones (2005) recognized this situation when she
suggested that “archaeologists of the Mesolithic
should now investigate the potential of the under-
5.9 Doggerland: Mesolithic sea world.”
Landscapes of the Southern The North Sea Palaeolandscapes Project
North Sea (NSPP) has demonstrated that recovery of
archaeological landscape information through
The Southern North Sea (SNS) is a marine basin extensive 3D seismic data is both possible and
that occupies a position between the European desirable. The information derived provides a
countries of Norway, Denmark, Germany, the unique opportunity to explore human activities
Netherlands, Belgium, France, and the United within a spatially extensive prehistoric landscape
142 5 Geophysical Techniques for Archaeology

Fig. 5.13 Acoustical profile and schematic view of sonar Mexico. Sediment vibracore locations (Fig. 9.14) are
images produced by an acoustic sub-bottom profiler of a indicated. This ancient estuary is illustrated its drowning
drowned paleochannel in 15 m water depth in the Gulf of at transgression ca. 6000 BC (Pearson et al. 2008)

and, more fundamentally, suggests that the land- archaeological prospection had been recognized
scape could have contained a significant Meso- (Kraft et al. 1983), this opportunity has not been
lithic population. The geophysical survey data realized until Gaffney et al. began their interdis-
cover more than >22,000 km2 in the Southern ciplinary study.
North Sea alone (Fitch et al. supra), and although The Central and Southern North Sea are
the potential of this data to inform submerged nowhere deeper than 60 m, and the flat-topped
5.9 Doggerland: Mesolithic Landscapes of the Southern North Sea 143

Fig. 5.14 Left, sampling

the drowned Sabine estuary
shown in Fig. 9.13 with a
vibracore; right,
analyzing the sediments in
a vibracore taken from the
estuary (Pearson
et al. 2008)

Dogger Bank rises to within 15 m of the surface, from massive blocks of raw data. Subsequent 2D
providing a total relief of 45 m over a distance of and 3D views illustrate river channels, estuaries,
about 500 km east to west. The objective of the beaches, marshes, lagoons, and braided/
NSPP was to analyze the upper part of the seis- overlapping drainage patterns much like the ear-
mic profile and to extract from that record the lier Sabine River paleovalley study. The success
sedimentary facies (“seismic facies”) and topo- of the project depended upon access to massive
graphic changes which indicate past landscape computing power available to the project team
surfaces during the Mesolithic period from through the Hewlett-Packard Visual and Spatial
10,000 to 5000 BP. Geographic information Technology Centre at Birmingham University
systems (GIS) were used to extract time slices (Gaffney et al. 2007).
Petrography for Archaeological Geology

6.1 Introduction axes or as a grinding stone for a Roman hand

mill. Dimension stones are used in construction
In the first edition of this textbook, it was asserted of structures—buildings, dams, canals, roads,
that the most common artifactual material of etc. Art stone will, of course, include portraiture,
prehistory is lithological—rocks and/or minerals. statuary, and monuments such as tombs, stelae,
This is still the case. Stone is, and was, the most etc. Because of stone’s durability and availabil-
durable of all materials available to early humans ity, it has fulfilled all three of these roles admira-
and, in most environmental settings, the most bly over eons of human existence.
readily available. Its durability made it desirable For much of human prehistory, beginning in
for a multitude of tasks as well as helping insure the Pliocene, “tools” were fashioned of durable
its survival in archaeological sites. For the stone; archaeology has been forced to focus on
archaeologist, the survival of ancient human the nature of these tools and the stone—
stone tools and artifacts has been both a blessing toolstone—from which they were made. In
and a source of unintentional biases in terms of 1894, W.H. Holmes outlined the steps for the
the reconstruction of past cultural behavior. Even analysis of stone tools:
with the earliest of human culture surely, there
were other implements other than those of stone, The first necessary step is their identification as
implements as distinguished from allied forms,
but by their durability, the latter have survived, natural and artificial, that have no claims to be
while other materials have long ago perished called such. The second step is a consideration of
from the archaeological record. their natural history, which embodies two distinct
The model in Fig. 6.1 is wearing only five lines of development, one of the individual from its
inception in the raw material, through a series of
items—6, 7, 8, and 9—that can be classified as steps of technical progress, to the finished result;
stone per se. Item 9 might be hedging the defini- the other the species or group from a primal culture
tion of stone, a bit, but fossilization is basically germ through countless generations of
lithification; thereby it is fair to classify this item implements; with these goes the evolution of
form and function. The third step consists in
as “stone.” More specifically, in this chapter, the regarding implements as historic records, first,
use of stone for adornment or decorative with reference to questions of time; second with
purposes will be classified as “art stone.” Two regard to questions of culture, and third, to the
other categories will be used—toolstone and history of peoples. (Holmes 1894: 121)
dimension stone. The former will encompass
those types of stone used in functional categories In terms of the first step, the stone tool indus-
related to tasks. A toolstone can be used for stone try from the lowest levels (Bed I) of Olduvai

# Springer-Verlag Berlin Heidelberg 2016 145

E. Garrison, Techniques in Archaeological Geology, Natural Science in Archaeology,
DOI 10.1007/978-3-319-30232-4_6
146 6 Petrography for Archaeological Geology

Fig. 6.1 Ancient and historic jewelry (Source: National 600 AD; Central Asia. 7 “Bat wing” object—“parietal”?
Geographic Magazine). 1 Pendant of perforated shell Jasper; Mesoamerica. 8 Bead, lapis lazuli, 1000 AD; West
ca.1000 BC; Syria. 2 Seal, shell, 7000–4000 BC; Syria. Asia. 9 Object fashioned from fossilized shell, 1000 AD;
3 Tubular bead, bone, 3000 BC; Syria. 4 Eye pendant, West Asia. 10 Incised disk, shell, 1900 AD; Mauritania. 11
glass, 2000 BC; Egypt. 5 Pendant, glass, at least exterior; Kiffa bead, powdered glass, 1930 AD; Mauritania. 12
600 BC; Punic-Phoenician? 6 Object made of carnelian, Glass art depicting Marilyn Monroe, 2004; USA

Gorge was among the first implements defini- This discussion begins with sedimentary rocks
tively identified as made by hominins (Conroy (Blatt 1982; Flygel 1982; Garrels and Mackensie
and Pontzer 2012: 329–330). These tool forms 1971; Pettijohn et al. 1973; Rapp 2002) (Fig. 6.2).
have been found elsewhere, in earlier deposits As illustrated in Fig. 6.3, the rock cycle
but their forms are in all cases, manufactured. represents the continuum of lithogenesis wherein
Pertinent, in particular to this chapter, is the the formation of one type of stone may entail
“natural history” that includes consideration of the destruction or alteration of another type
the geological formation of rock types and their such as the magma melt of accretionary sediment
sources, e.g., the stone itself. Finally, how these wedges in subduction and a subsequent re-
implements function as historic records will be extrusion as igneous rock types. Lithified
touched upon in the context of our modern shells—micro and macro—form the carbonate
analyses of them. One such example are the structure of limestones and other biogenic rocks
bluestones of Stonehenge. which, in turn, undergo metamorphic alteration
into that ubiquitous art stone, marble. Shales
become slates; sandstones alter to quartzites.
6.2 Major Rock Types
and Archaeology
6.3 Sedimentary Rocks
The three rock types, sedimentary, igneous, and
metamorphic, appear throughout prehistory as the We begin with sedimentary rocks. Sedimentary
materials for simple domestic tools to major rocks are formed by the consolidation of
masonry construction. In this section, most impor- sediments. In the rubric of geology, they are
tant archaeological rocks and minerals will be termed the “soft rocks” although the hardness
discussed with some illustrative examples from of some bioclastic limestones, as we shall see
archaeology as well as a listing of their key petro- later in this chapter, is formidable enough to be
graphic properties that can be used in their identi- used as grinding stone. Of the many sedimentary
fication using optical and instrumental methods. classification systems or schemes, most can be
6.3 Sedimentary Rocks 147

Fig. 6.2 Dimension stone.

Middle Jurassic
sandstones. Dun Beag
Broch, west coast of the
Isle of Skye, Scotland.
Historic Scotland

skeletal, and oolitic. This discussion will follow

the former classification terminology of clastic,
chemical, and biogenic (Table 6.1).
Earth scientists understand that organic
remains are most prone to diagenesis—the pro-
cess of the chemical recrystallization or replace-
ment in a substance, buried in the Earth, when
given enough time, will change almost any natu-
ral material, even stone. Diagenesis might be
confused with weathering which is the physical
and chemical disintegration of rock and Earth
materials upon exposure to atmospheric agents.
In the case of most rocks and minerals, the time
needed for diagenetic change to occur is much
longer than that which alters organic materials
such as the bone and botanicals. There are
instances, numerous in some geological epochs,
where even these latter materials become
Fig. 6.3 The rock cycle mineralized during diagenesis and are no differ-
ent from stone. Happily, for paleontologists,
reduced to three groups of rocks, principally most past life forms are fossilized and preserved
clastic, biogenic, and chemical (Andrews for study. For archaeology, subfossil and fossil
et al. 1997a, b), which in turn can be termed bone could become toolstone, art, and dimension
clastic, carbonate, and evaporites. Another way stone (Figs. 6.1 and 6.2).
of viewing these rocks is simply textural. Clastic
texture refers to the mineral fragments and rock Clastic Rocks Fragments of rocks are called
debris which comprise the material. Non-clastic clasts. Clasts are the product of the weathering,
texture is that produced by the chemical or bio- physical or chemical, of rocks. Clastic rocks
genic processes which form the rock. Within the are classified on the basis of grain size, ranging
non-clastic forms are the textures—crystalline, from gravel-sized fractions in conglomerates
148 6 Petrography for Archaeological Geology

Table 6.1 Classification of sedimentary rocks

Texture Composition Rock name
A. Clastic rocks
Coarse grained (over 2 mm) Rounded fragments of any rock type—quartz, quartzite Conglomerate
chert dominant
Angular fragments of any rock type—quartz, quartzite, Breccia
chert dominant
Medium grained (1/16–2 mm) Quartz with minor accessory minerals Quartz sandstone
Quartz with at least 25 % feldspar Arkose
Quartz, rock fragments, and considerable clay Graywacke
Fine grained (1/256–1/16 mm) Quartz and clay minerals Siltstone
Very fine grained (less than Quartz and clay minerals Shale
1/256 mm)
B. Chemical and biogenic (some of them are not chemical but biological)
Medium to coarse grained Calcite (CaCO3) Crystalline
Microcrystalline, conchoidal Micrite
Aggregates of oolites Oolitic limestone
Fossils and fossil fragments loosely Coquina
Abundant fossils in calcareous Fossiliferous
matrix limestones
Shells of microscopic organisms, Chalk
clay soft
Banded calcite Travertine
Textures are similar to those in Dolomite CaMg(CO3)2 Dolomite
limestone dolostone
Cryptocrystalline, dense Chalcedony (SiO2) Chert, etc.
Fine to coarse crystalline Gypsum (CaSO4* 2H2O) Gypsum
Fine to coarse crystalline Halite (NaCl) Rock salt

and breccias through sandstones, siltstones, 6.3.1 Arenite or Arenitic Sandstone

and fine-grained shales. The grain-size scale for
clastic rocks is exactly that for sediments General properties: While the name “arenite”
(Table 6.1). is a bit dated, simply implying a quartz-rich
sandstone, it still commonly appears in texts.
Sandstones Of the clastic rocks, the sandstones Similar terms, not used in this chapter, are
are familiar archaeologically as building stone. “rudite” or a conglomerate or gravel-rich
They are the easiest to recognize as they are stone and “lutite,” a clay-rich stone (claystone).
cemented sand grains as their name so aptly Of the general class of “sandstones,” quartz
implies. One scarcely needs more than a hand arenites are greater than 95 % quartz. The grain
lens to initially determine this class of rocks. matrix is well cemented. Orthoquartzite is an
Within the class of sandstones are (a) the arenite cemented with quartz. Other arenites
arenitic or “quartz sandstones,” composed mostly may be cemented with calcitic, dolomitic,
of quartz; (b) arkose, made up of feldspars rather clayey, or ferruginous material. Figure 6.4
than quartz; and (c) graywackes which contain illustrates a ternary diagram using Folk’s classi-
15 % finer-grained materials—clays and silts. fication for sandstones. As a rule, these hard,
6.3 Sedimentary Rocks 149

Fig. 6.4 Ternary diagram First dataset

using Folk’s sandstone
Second dataset
classification (1948, 1980) 95 Third dataset
(Reproduced from Zahid Fourth dataset
and Barbeau (2011)). Average first set
Arenitic, arkosic, and lithic Subarkose Sublitharenite Average second set
categories are subdivided Average third set
according to the percentage Average fourth set
of each within the diagram. 75
Subcategories, such as
lithic arkose and
feldspathic litharenite, are
illustrated following a
75–50–25 % breakdown

F 25 50 L
Arkose Lithic Feldspathic Litharenite
Arkose Litharenite

Sandstone Classification (Folk, 1980)

tough rocks are less well suited as archaeological source environments than depositional processes
toolstone and are better used as dimension stone. (Williams et al. 1955). In thin section, feldspars
Figure 6.5 illustrates sandstone used for mega- appear brown in PPL, colorless, otherwise (e.g.,
lithic monuments. not pleochroic). In XPL the twinning of the
In thin section, the quartz grains of an arenite feldspars can be observed (Fig. 6.6). The
are closely bound and relief is contingent on “Carlsbad twin” is common in alkali feldspars
orientation (Fig. 6.6). In plane-polarized light but is seen in plagioclase as well. Fine parallel
(PPL), single crystals are rarely euhedral. In twinning (“albite”) is seen in the anorthites and
cross-polarized light (XPL), quartz grains exhibit microcline varieties. Relief is low or negative in
classic undulatory extinction. Interference colors both alkali and sodium/calcium feldspars (oligo-
are low order but bright. clase, anorthite). Silcrete
6.3.2 Arkose or Arkosic Sandstone As chemical sedimentary rocks, silcretes are akin
to calcretes and ferricretes (the former being
General properties: Arkoses are defined by the cemented by calcite and latter by crystallized
abundant presence of feldspar (>25 %). The iron oxides). In South Africa where this rock
feldspar may be either plagioclase or orthoclase. type is common, it forms an indurated duricrust
Arkoses are considered to be more reflective of of silica—cemented sands and gravels—a
150 6 Petrography for Archaeological Geology

Age (MSA), 40–90 ka, in the Cape area, heat

treatment of this rock has been documented
(Schmidt et al. 2012; Brown et al. 2009).

6.3.3 Graywackes or Clay-Rich

Sandstone (“Lutitic”)

General properties: These are sandstones with

greater than 15 %—silts and clays—as a matrix.
Like the arenites, the graywackes vary, in name,
according to the dominant grain types—quartz-
itic, feldspathic, and lithic. These sandstones are
generally darker in color and poorly sorted. The
argillitic grains in the matrix are not visible to
either the eye or under the optical microscope
due to their small size (~2 μ).
A typical graywacke, under PPL, shows an
opaque fine-grained matrix with larger quartz,
feldspar, accessory mineral grains, and rock
fragments. As with other sandstones, the optical
properties of quartz and feldspars in PPL show
Fig. 6.5 Sandstone Pictish standing stone. Perthshire, coloration and shape according to their specific
Scotland. So-called Class 1 Stone by virtue of only mineralogy. The XPL views show extinction and
“pagan” Pictish symbols used in its decoration (Photo-
graph by author) twinning are apparent in quartz and feldspars.
Rock fragments and clays will show coloration,
in PPL, such as brown-colored iron oxides in the
resistant capstone. It is prevalent in Australia as cement, pale green chlorite and black in XPL,
well. Silcrete is found in Europe, notably and brownish-green of glauconite. In XPL,
England and France but less so in the Americas. amphibole will show an orange interference
Arakawa and Miskell-Gerhardt (2009) have, color. Lath-like mica, particularly, muscovite,
however, identified silica indurated sandstones will show the birefringence colors of blue-red
in the Four Corners region of the American and orange, much like the expanding clays (illite,
Southwest. This silcrete-like stone has been chlorite, etc., cf. Chap. 7) (Fig. 6.7).
incorrectly labeled as “quartzite” by many
archaeologists (supra). These authors have Siltstone, Mudstone, Claystone,
identified a silica indurated shale facies in the and Shale
same region. Both rock types were commonly These clastic sedimentary rocks are composed of
used for toolstone by prehistoric groups. Both silicate minerals, mainly quartz and clays less
rocks are lower Cretaceous in age. than 5 μm in diameter. Interestingly, they are
In thin section, silcretes are dominated by the most abundant of all sedimentary rocks but,
quartz clasts (Schmidt et al. 2012). These compared to the harder sandstones and
undergo extinction in XPL. The microquartz limestones, have less archaeological importance
cement, at least in South African samples, is than that of a geological nature with the excep-
relatively variable, ranging from fine to coarse tion of shale. Siltstones occur in basin-edge
grained (supra). As a toolstone, secrete usage successions between sands and muds (Emery
extends well into the upper Pleistocene in both and Myers 2009). These rocks also occur in fore-
South Africa and Australia. By the Middle Stone land basins as erosional depositional sediments.
6.3 Sedimentary Rocks 151

Fig. 6.6 Two thin-section views of arenitic sandstones. thin section. The quartz grains show shattering due to
Left, XPL showing undulatory extinction of the quartz sawing. Right, XPL image showing “twinning” of feld-
grains (black grains). Crystals of sparite cement are spar grains (Left image by author, right image courtesy
clearly seen along the edges of holes (blue voids) in the the University of North Carolina-Chapel Hill)

Fig. 6.7 Neolithic Period stelae carved from sandstone, from the megalithic site of Petit-Chasseur, Sion, Valais,
Switzerland (Musée d’histoire du Valais, Sion, photograph by R. Hofer)

Figure 6.8 shows a stone of the latter type—a of their habit of fracture along bedding planes,
“green molasse” siltstone from the Jura had been used from tabular artifacts to roofing
Mountains. material in the past. For structural purposes, even
In grainstones and mudstones, the former shale is too fragile for building proper.
variety is composed of calcite grains cemented Because of the small grain sizes common in
by spar. Mudstone is pretty much as the name these rocks, petrographic microscope techniques
implies—consisting mostly of small silt-sized can do little than to observe small-scale sedimen-
grains. The sediment is not overly compacted tary structure and describe larger-grained
and the grain structure or shapes, in grainstones, siliciclastic inclusions. PPL and XPL properties
are oolitic to peloidal. Bioclasts can occur as such as bifrefringence will show the presence of
well. Mudstone, also known as claystone, is clay minerals and micas, such as chlorite and
most commonly named shale. Shales, because muscovite. In XPL, rotation of the thin section
152 6 Petrography for Archaeological Geology

Fig. 6.8 Rear, a siltstone,

green “molasse” dimension
stone; foreground, a “grès
coquille” or bioclastic
limestone (Photograph by
the author)

will often demonstrate extinction. Also in PPL of the specific rock type does require a hand lens
and XPL, the differing optical properties of adja- and, in many cases, recourse to microscopes.
cent laminae will appear as lighter and darker This is due to the different depositional
bands. The authigenic crystals of calcite, dolo- environments in which the rocks originate. If
mite, siderite, etc. will appear highly birefringent the rock was deposited in a marine or lacustrine
as a rule (Fig. 6.8). setting, then the matrix can be calcareous and
clay rich together with significant amounts of Breccias and Conglomerates sand. If the rock is more the result of fluvial
If one considers grain size as a continuum, clasts action, the matrix will reflect this with interstitial
offer clues to the source of the sedimentary silts and clays bound by carbonate cements.
rocks—shales form from micron-sized minerals Their matrices can vary from limonitic, calcare-
and sandstones from millimeter-sized minerals, ous, and siliceous to lithified clay-rich cements.
while breccias and conglomerates are composed Breccias and conglomerates are important rocks,
of the largest clasts. Morphologically, the from an archaeological standpoint, with their use
breccias and conglomerates are differentiated in both artifacts (toolstone) and buildings
on the basis of surface roundedness—the latter (dimension stone) (Lucas and Harris 1962;
are more rounded than the more angular breccias. Mannoni and Mannoni 1984; 159; Winkler
The lack of angularity in conglomerates is con- 1994; Mazeran 1995).
sidered to be the result of fluvial action. Both
terms for these rock types denote their structure Carbonates: Limestones
and not their composition. As one would expect, and Dolomites
their mineralogical makeup can be, and is, quite Of the carbonate or calcareous sedimentary rock,
varied, ranging from carbonate to volcanic in limestones are the most common. This is because
composition. their origins are most frequently in marine
The identification of the breccias and basins. In the ocean, the shells or skeletons of
conglomerates rarely requires more than a hand small animals rain down constantly, forming
lens due to the size of the fragments ranging up to thick layers of carbonate “Globigerina” ooze
pebbles. The identification of the binding matrix (Thomson 1877). These muds lithify into the
6.3 Sedimentary Rocks 153

skeletal limestones. As a rule the limestones, limestones, can result from processes like
skeletal and otherwise, are authigenic in nature micritization but also can result from pressure
that is to say they have formed in place, at that that compacts the pore space between grains.
place of deposition. The lime carbonates or By simply calling the most abundant carbon-
limestones are made up, chiefly, of calcite ate rocks either limestones or dolomites belies
(CaCO3), dolomite (Ca(Mg)CO3), and ankerite their petrographic complexity with diverse
Ca(Mg,Fe,Mn)CO3. textures from clastic to organic accretion.
Rocks that are primarily calcite are termed Optically, it is difficult to differentiate the two,
limestone. Those with dolomite are dolostone or and the dye Alizarin Red S is used to do so much
simply dolomite; and mixtures can be dolomitic the same way sodium cobaltinitrite is used to
limestone. Some limestones have a large propor- distinguish the feldspars. Etching limestones
tion of the calcium carbonate pseudomorph, ara- with weak (10 % by volume) HCl or 2N formic
gonite. In most limestones, particularly those acid is another technique. Calcite is more soluble
exposed to meteoric water, the aragonite than the rhombohedral dolomite which stands out
exsolves and recrystallizes to calcite. Such in relief on either hand sections or uncovered thin
limestones are “old” limestones or compact sections. After extensive etching, only the
limestones. Returning to the allogenic forms of non-carbonate minerals remain such as chert,
limestones, these clastic rocks consist of calcite clay, feldspar, and quartz.
of either terrigenous, organic, or oolitic origin Limestone, as a rule, does not make good
(Williams et al. 1955). Most are of marine origin. toolstone, but as dimension or art stone, these
The fragments are cemented by additional car- rocks found great utility in antiquity. The most
bonate that precipitates into the pore space spectacular use of limestone as dimension stone
between the calcite grains. If the number of sili- is the pyramids. In the Americas, notably Central
cate grains increases, the rock is a calcareous America and southern North America, limestone
sandstone rather than a calcareous limestone or was commonly used for building purposes. The
calcarenite. Some limestones are made up of Maya use of limestone rivaled that of ancient
small (<2 mm) roundish grains that appear to Egypt.
be organic forms or tiny geoids. These rather Abundant as karst, the Yucatan, and areas
unusual grains are ooids that are composed of southward, Mayan cities and infrastructure—
carbonate formed around a detrital silica nucleus. temples, causeways, cisterns, etc.—were built
If these forms are larger than 2 mm, they are using limestone. As stelae, limestone recorded
called pisoids or oncoids. Bioclastic limestones the accomplishments of Mayan rulers from
are made of cemented shell and skeletal Copan to Tikal.
fragments. These clasts can consist of molluscs, Roman stone portraiture used limestone when
brachiopods, corals, forams to algae, and worms. and where marble was not available either physi-
Peloidal limestones resemble ooids but the grains cally or economically. Roman—both western
lack nuclei and are completely micrite. and Byzantine—mosaicists preferred limestone
Micrite or microcrystalline calcite has a diam- for tesserae.
eter of 5 μm or less (Adams et al. 1997). Sparites It was cheap, available, and found in a variety
are calcite grains over 5 μm in size (Fig. 6.6). The of colors from gray/green to yellow. If the lime-
presence of these forms of calcite in limestone stone colors were not acceptable, there was
gives them their respective names—micritic or always paint as the Romans used on
spary limestones. Micritization is the process that reproductions of Greek originals (Abbe 2011).
clastic limestones, particularly bioclastic Even when not used as a colorant on portraiture,
varieties, lose their original texture and form monuments such as Trajan’s Column has clear
fine-grained forms. Micrite is a major cement in evidence of the extensive use of paint and gild-
limestones along with carbonate mud sediments. ing. In historic America, limestone quarries in
Loss of grain texture or clastic inclusions, in the Missouri portion of the Ozark Mountains
154 6 Petrography for Archaeological Geology

produced dimension stone as well as polished hand-operated mills known as rotary querns.
limestone (“white marble”) (Rovey et al. 2010). The tool marks visible on the quarry faces and
on a series of aborted quern extractions recov- Organic or Bioclastic Limestone ered both in the quarry backfill and along the
Components of these rocks are diverse calcare- surface of the road allow reconstruction of the
ous organic materials characterized by (a) overall operational sequence of quern production. This
particle size and shape and (b) internal wall small quarry is beside a road 6 m wide and
structure of the organism. This latter quality is excavated over a length of about 300 m. This
one aspect where the use of XPL has a significant road was a major thoroughfare linking the
role in thin sections—level studies of these Roman cities of Yverdon with Avenches and
limestones (Adams et al. 1984). The age of served to export the querns. The smithy, installed
limestones can be relative to the percentage of on the other side of the road, a stone’s throw from
aragonite if it was there at the formation of the the quarry, included about 750 kg of smithy slag
stone. Over time, it will revert to granular calcite. and about 3,000 metal objects. A wooden con-
In oysters and corals, both calcite and aragonite struction is interpreted as both the house and
appear. The structural shape or form will remain workshop of the blacksmith. It is obvious that
after dissolution or recrystallization of the arago- the smithy repaired and maintained the tools of
nite unlike those forms which only secrete pure the quern maker. There is also evidence that the
calcium carbonate whose shells or skeletons smithy served more than the quarry and made
are lost. other tools for a local clientele. Finally a small
The identification of a pseudomorph of calcite wooden house, built with posts, was set along a
is beyond the hand lens and the microscope space between the road and the quarry. It is
unless it is an electron-based unit. Even then it possible that it served to lodge the quern maker.
is perhaps best to leave the identification to Because of the rock’s hardness and durability,
techniques such as X-ray diffraction or WDS it was an excellent stone for crushing grain to
analysis on an electron microprobe. Fine flour. During excavations along an autoroute in
sediments or micrite clement can fill in the western Switzerland, a millstone quarry, forge,
voids or molds of lost organisms leaving casts. and outbuildings adjacent to a Roman road were
Identification of the organisms and the overall uncovered (Anderson 2003). Petrologically this
porosity of cement is typically done in thin sec- particular stone is marine in origin with an
tion. Most examinations are done in PPL on extremely high density of lenticular shell and
stained specimens. As mentioned, XPL observa- quartz cemented in a micrite matrix (supra)
tion has advantages in examining organic (Fig. 6.9). Besides millstone, the rock was, and
structures. If the bioclasts are intermixed with still is, used for walls, curbing, steps, borders of
sediments, the crossed polars can differentiate windows, etc. In antiquity it served as the base
these mineral grains. Bioclastic limestones are for columns and protective caps or coping on
used for dimension stone (Fig. 6.7). Roman tower tombs at the Chaplix cemetery,

6.4 A Roman Quern Production

and Smithy Works 6.4.1 Clastic or Calcarenite Limestones

The main phase of the site of the Roman site of Made up of calcite fragments, these rocks are
Châbles (Canton of Fribourg, Switzerland) dates cemented by micrite or sparite. Sorting of the
to the turn of the first and second centuries AD composite grains is just as with sandstones. The
and comprises a quern quarry, a road, a smith, particles of calcarenites are either organic
and a house (Anderson et al. 1999, Anderson fragments, ooliths, or carbonate rock fragments.
2003). The quarry was exploited for small, Here the grains are predominantly carbonate
6.4 A Roman Quern Production and Smithy Works 155

Fig. 6.9 Thin-section

views of Grès coquille
bioclastic limestone,
western Switzerland. The
high density and lamination
of shell and quartz are
clearly visible (Image
courtesy of the
Archaeological Service,
Canton of Fribourg,

rather than silica. Nearshore beach deposits and cement. Micritic cement can appear greenish-
dunes materials form into calcarenites. These brown, in PPL, while sparite is colorless. If dolo-
deposits are rapidly cemented into rock by addi- mitization has affected the rock, then rhombs of
tional carbonate chemically precipitated in the this mineral can be seen in PPL or XPL. Recall
pore spaces. Cementation can be rapid forming that our previous discussion on the distinctive
what is termed “beach rock.” In areas of multiple optical properties of the carbonates—high bire-
transgressive-regressive cycles, micritization and fringence, sharp relief, and birefringence—can
dolomitization can both cement these carbonate be used to characterize the limestones.
rocks in submarine and subaerial conditions.
Hand lens examination of calcarenite samples
can lead to misidentification as bioclastic 6.4.2 Aphanitic Limestones
sandstones. In some instances, the difference is
truly one of degree so the difficulty is real General properties: Fine grained, by definition,
(Fig. 6.10). the grains are difficult if not impossible to see
Grain size and texture are the first parameters with the unaided eye or even a hand lens. Color is
easily distinguished in PPL. Porosity varies widely generally white-gray although the presence of
from tightly welded to poorly sorted varieties. In iron or organic carbon can create pink, red, and
XPL, where the calcarenite grains are tightly gray-black varieties. Fracture is chert-like, that
packed and cement over growth has occurred, is, almost conchoidal due to the very fine grain of
both the over growth and clasts will appear dark the stone. Fossils are not uncommon in these
due to a uniform extinction color for grains and rocks. Karstic terranes are often aphanitic
156 6 Petrography for Archaeological Geology

Fig. 6.10 The Roman

millstone quarry at
Switzerland (Photograph
courtesy of Dr. Timothy
A. Anderson (shown in
lower left))

limestones. Most limestones do not make good

stone tools due to their inherent low hardness.
Exceptions, such as the Roman mill stones, do
occur in the archaeological record.
Fine grained and dark in PPL and the
increased argillaceous content will enhance the
XPL image’s optical properties. Where clays are
concentrated in laminae, the thin section reflects
this texture as well. Intercalated calcite veinlets
will give the section a weblike appearance.

6.4.3 Dolomite and Dolomitic Fig. 6.11 Backscatter electron (BEI) image showing
Limestones dolomite rhombs in thin section. A large rotalliid forami-
nifera (center) is surrounded by dolomite rhombs; large
Ca(Mg)CO3 is a major component of limestone apatite grain appears in the upper, center portion of the
image (SEM image by author)
although of a secondary origin, replacing
preexisting carbonate minerals (Adams
et al. 1984) unless the euhedral, rhomboidal, termed a dolomite. Dolomitization is the replace-
and dolomite can be seen microscopically ment process that limestones undergo. As dolo-
(Fig. 6.11), it can only be distinguished by mitization is not observed in today’s oceans, by
staining, after etching, of the thin sections with the logical extension of the uniformitarianism
Alizarin Red. The presence of dolomite in a principle, it is hard to visualize it as a primary
limestone does not make it a dolomite. If the process in earlier times. Whether a direct chemi-
percentage is less than 10 %, it is limestone; if cal precipitate or diagenetic replacement min-
the percentage is less than 50 %, the rock is eral, dolomites and dolomitic limestones are
dolomitic limestone; and only when 50 % or more resistant and harder than calcitic
more dolomite is present is the rock strictly carbonates. For these properties, these rocks
6.4 A Roman Quern Production and Smithy Works 157

have been regularly used in building throughout forms are found in the chert varieties. The most
antiquity. common of these varieties are flint, jasper, and
Dolomite, as a mineral, is a major component novaculite. In cherts composed in large amounts
of many limestones (Adams et al. 1997). The of fibrous chalcedony, these are called by that
macroscopic appearance of dolomitic limestone name. Cherts are generally some admixture of all
is generally fine grained although not as fine as or some of these forms. “Flint,” in most English-
the aphanitic forms nor as coarse as some of speaking contexts, is the other common name for
calcarenites. In some thin sections, the dolomite chert. In a strict sense, it refers to gray-to-black
preserves the original texture or structure of the cherts composed mostly of chalcedony and/or
calcite protolith. Two key properties distinguish cryptocrystalline quartz (Williams et al. 1955).
dolomite from calcite—the first is the distinctive It occurs in both nodular and bedded forms. One
rhombohedral shape of the single dolomite crys- of the most famous archaeological localities for
tal and the second is its non-uptake of Alizarin flint is that of Grimes Graves, Norfolk, England,
Red stain. The euhedral rhombs of dolomite site of extensive mining activities in the Neo-
scattered through a limestone’s calcitic matrix lithic Period.
are singularly recognizable (Fig. 6.11). Calcite, Much like dolomite, chert, although silica, is
by comparison, is generally anhedral. In closely thought to be a metasome as well being less a
spaced dolomitic-calcitic matrices, the dolomite precipitate than a diagenetic production depos-
rhombs become subhedral or even anhedral, ited in shallow seas. The nature of cherts takes
making the necessity of staining obvious. three forms or “habits”—chalcedony, opaline,
Where dolomite has replaced calcite, the grain and crystalline. Most cherts are either crystalline
shape is calcitic. or chalcedony as opal is less stable over time.
Opaline chert is found, but it is more likely to
find a chert with amorphous opal in a matrix of
6.4.4 Cherts, Gypsum, and Ironstones quartz—fibrous (chalcedony) or crystalline
(chert/flint). Cherts may have abundant
These rock types, taken ensemble, make up only a impurities such as clay minerals, feldspars, dolo-
small fraction of the sedimentary rocks, but, from mite, and, in the case of the radiolarites, abundant
an archaeological point of view, some—notably microfossils. Conchoidal fracture is diagnostic of
chert and gypsum—have an importance beyond almost all chert. This property plus its durability
that of sheer geological abundance. Others like and sharp edges made this rock a favorite with
ironstones have had economic importance in the almost all prehistoric cultures.
industrial past. Because of its use as tools since As a silicate rock, the crypto- to microcrystals
great antiquity, we shall begin our review of these of quartz will demonstrate the optical properties
rocks with chert. Chert is a hard stone, very dura- of that mineral (cf. section on sandstone).
ble yet workable in the hands of artisan. Staining with Alizarin Red identifies any calcite
in the chert. Quartz replaces calcite (metaso-
Chert/Flint Perhaps the most pervasive of all matic) in limestones, and in thin section, this
archaeological lithic materials, these are rocks will be observed as anhedral grains exhibiting
composed of silica rather than carbonate the extinction and birefringence of quartz.
(Leudtke 1992). They form within aqueous Because the cherts are aggregates, the extinction
environments and are found within limestones pattern is “patchy.”
or interbedded with them in section. It can Novaculite is a material used in hone or
occur as either primary or secondary—the first whetstones and is gray-to-white chert that is
by precipitation and the second by the process of crypto- to microcrystalline whose texture is
replacement of a limestone matrix. The silica can almost granular to the touch. The true grain tex-
occur in three forms—amorphous, such as opal; ture can only be observed by a petrographic
cryptocrystalline; and microcrystalline. All these microscope (Fig. 6.12). Novaculite has been
158 6 Petrography for Archaeological Geology

Fig. 6.12 Novaculite.

Left, in thin section; right,
projectile points (Flenniken
and Garrison 1975; Used
with permission of Maney

studied as a subject of the thermal pretreatment anhydrite-anhydrous CaSO4 forms. At or near

of cherts to improve their working qualities for the Earth’s surface, and exposure to meteoritic
the production of flaked stone tools (Flenniken waters, gypsum replaces anhydrite.
and Garrison 1975). Classified here under the sedimentary rocks,
Jasper is a brown-to-red chert often difficult these forms are more correctly minerals. While
to differentiate from chalcedony (Fig. 6.1). The less durable as a dimension stone in antiquity
reddish color originates from ferric oxides in the (ex. Knossos), alabaster has been mined and
material. Chalcedony, as seen in X-ray studies, is used for veneer slabs as well as for sculpture.
microfibrous as a rule but contains amorphous Rock gypsum is a massive, coarse-grained, gray-
silica as well. This gives this chert a glossy, slick to-white rock composed almost exclusively of
feel to the touch and, in some varieties, an almost aggregates of gypsum. Where iron oxides are
gemstone quality. Chalcedony is an opaline chert incorporated into the matrix, it is colored with
and as a result is very hard. Opal is not stable and various shades of orange to red.
will recrystallize over time; thus, some cherts Gypsum, in thin section, is readily distin-
which may have been opaline have reformed guished from anhydrite. Gypsum exhibits low
into chalcedony and crystalline forms. Carnelian, birefringence and negative relief compared to
a reddish-orange variety of chalcedony higher birefringence and lighter relief of anhy-
(Fig. 6.1), was used as jewelry in the fifteenth drite. The interference color of gypsum is green,
to thirteenth century Bronze Age Canaanite sites while that of anhydrite has bright interference
such as Hazor, in modern-day Israel. colors.
Radiolarites are cherts with large proportions
of the skeletal remains of radiolarians, diatoms, Ironstone Deposited in marine contexts, this
and sponge spicules in the stone matrix of quartz carbonate rock is more than 15 % iron. The iron
and opal. These are very hard rocks and used in can be in the form of siderite (FeCO3 and
prehistory to make durable tools. Jurassic chamosite). The latter can oxidize to the iron
radiolarites are found in California’s prehistoric oxide limonite. Limonite can appear in oolitic
cultural middens. form. Hematite, Fe2O3, is more stable in terrige-
nous ironstones than either siderite or chamosite.
Rock Gypsum After carbonates, the hydrous Ironstone can be bioclastic and some are banded
sulfate mineral, CaSO4H2O, precipitates in the with chert layers (Adams et al. 1997).
evaporation of marine basins. The production of
gypsum or its chemical replacement anhydrite is Travertine Deposited in caves, this carbonate is
temperature dependent. Above 42  C, gypsum also known variously as “flowstone” or “drip-
will precipitate; below this temperature, stone.” It is distinguished by its lamellar or
6.5 Igneous Rocks 159

banded nature which reflects its origin in the The most common classification of igneous
waters percolating into solution features such as rocks is that of the International Union of Geo-
caves. Archaeologically it can occur as a finish logical Sciences (IUGS) Subcommission on the
stone because of its beauty, but more recently it Systematics of Igneous Rocks (Streckeisen
has become the object of paleoenvironmental 1979). Beyond, and before, the IUGS classifica-
studies wherein the bands of the rock are equated tion scheme, several others used in the study of
to the “growth rings” of trees. It has been igneous rocks have produced hundreds of names.
demonstrated that speleothems—stalactites and The names used in this text will be those of the
stalagmites—can be dated using techniques such IUGS system. The IUGS classification of plu-
as U-Th, ESR, and radiocarbon. Coupled with tonic or intrusive igneous rocks is less debated
examination of stable isotopic ratios in the growth than that for volcanic or extrusive igneous rocks
bands, it is possible to retrodict past climatic due to the fine-grained character of these rocks.
conditions (Hendy 1971; Shopov et al. 1994; The color index—light to dark—is commonly
Brook et al. 1996). In the situation of cave sites, used along with mineralogy and texture in
the presence of flowstone coating prehistoric describing igneous rock. In our discussion of
artifacts, skeletal remains, or art can reassure igneous rocks, we emphasize their modal miner-
researchers of their veracity and relative antiquity. alogic classification rather than their chemistry.
Examples of this are found at Petralona Cave The texture of igneous rocks ranges from
(Greece) and Cosquer and Chauvet Caves in coarse to fine grained or pegmatitic/phaneritic
France. At Petralona Cave, a skull of an archaic to porphyritic/aphanitic (Andrews et al. 1997a,
Homo sapiens was encrusted with travertine b). Texture refers to the size and shape of indi-
which attested to its age geologically and geochro- vidual mineral grains. Phaneritic texture is visi-
nologically (Schwartz et al. 1980; de Boris and ble to the unaided eye, while aphanitic texture is
Melentis 1991). At Cosquer and Chauvet, particu- not. The texture of an igneous rock is an indica-
larly at the former locations, the spectacular tion of the cooling rate of the parent magma.
ancient cave art has calcite encrustations over Large-grained or very coarse (pegmatitic) igne-
some paintings and engravings (Clottes and ous textures are found in the plutonic rocks,
Courtin 1996; Clottes 2001). while finer-grained textures are found in volcanic
rock. Obsidian is the finest “grained” of all igne-
ous rocks being in actuality a volcanic glass. In
the field, an archaeological geologist can usually
6.5 Igneous Rocks only evaluate color and texture if the grains are
phaneritic and estimate the mineralogy of an
Igneous rocks are formed by the solidification of igneous rock sample. The latter property, in its
molten magma or rock that has been melted in most generalized form, grades igneous rock min-
the upper mantle. This rock has been eralogy from felsic (feldspar-silica rich) to mafic
incorporated into the mantle from various (ferromagnesian rich) (Table 5.2).
sources, but in general they are the stuff of Igneous rocks are found in almost all areas
crustal rock which has been subducted at plate and are either intrusive, plutonic forms or the
margins after crustal plate collisions. As such, more localized, extrusive volcanic varieties. In
the rock to be melted can be sedimentary, meta- the former, we see more metamorphic rocks
morphic, or igneous in nature. The type of rock associated with the intrusive geology. Sedimen-
melted determines the nature of the chemical tary facies can blanket more ancient igneous
components of the magma. Igneous rocks are structures, or the igneous rocks can force their
mainly composed of the eight rock-forming way up through layers of older rock. Either pro-
minerals—quartz, plagioclase feldspar, K-spar cess results in the presence of igneous rock that
or orthoclase, muscovite mica, biotite mica, was recognized in antiquity for its durability and
amphibole, pyroxene, and olivine. utility, hence its presence in archaeological
160 6 Petrography for Archaeological Geology

deposits. In the case of obsidian, this volcanic petrological usage, this book will attempt to use
glass became one of the earliest economic rocks. the terms more familiar to the American student,
At Franchthi Cave, Greece, on the Argolid recognizing the equal status of the more interna-
coast, obsidian is present in low percentages tional terminology.
(~5 %) in Mesolithic levels increasingly through
the Neolithic to 95 % of the lithic archaeological
assemblage (Cherry 1988). This obsidian has
unique provenance. It was quarried on the island 6.5.1 Granite
of Melos 150 km distant meaning that sea-going
watercraft, in the Aegean, has great antiquity. At General properties: Granites are intrusive,
the Anatolian site of Catal Hüyük, which phaneritic rocks with a low or leucocratic color
flourished between 7500 and 6500 BC, obsidian index consisting of mainly quartz, feldspars, and
was a prominent trade mineral being mined in the micas. Of the feldspars, potassium feldspar is the
volcanic flows of the Konya region (Cunliffe most prevalent form. The granites are therefore
1988). The quantitative characterization relies classified as felsic. This ferromagnesian rock is
on many of the optical techniques already typically made up of suites of minerals such as
described in this chapter. To the archaeological biotite (mica), hornblende (amphibole), and
geologist examining igneous rocks, the questions augite (pyroxene). Granite has been a principal
are mainly of archaeological provenance—eco- building stone through antiquity. The Egyptians
nomic, cultural, cosmological, etc.—rather than used Aswan granites in the Giza pyramids for
geological provenance, e.g., magmatic type, tem- lining material of passages, for chambers, and
perature, etc. The composition of the rock will for doorframes (Lucas and Harris 1962). In con-
assist the researcher in determining the cultural trast to the Egyptians, granite was little used by
history of the material. the Classical Greeks (Higgins and Higgins 1996).
To do this, the investigator must determine the Where granite makes up the cores of Greek
mode (mineralogic content), texture, and color. islands, weathering of this rock results in fertile
The latter parameter, formalized originally by soils (Melas 1985).
Johanssen (1938), into the color index, is based Quartz appears featureless in PPL as a
on the percentages of light and dark minerals general rule. The feldspars will vary in the
present. When and where possible, the character- property of twinning ranging from microcline’s
ization should proceed to optical studies and, cross-hatched variety to simple twin patterns
where necessary, instrumental examination. in potassium feldspar. Plagioclase shows albite
Many of the following descriptions are based twinning. Accessory minerals such as the
on the excellent Atlas of Igneous Rocks and micas are recognized by their interference
Their Textures (Mackensie et al. 1982). Where colors in XPL and their pleochroism in PPL
terminology differs from that of American (Table 6.2).

Table 6.2 Common igneous rocks

SiO2, Na2 O, K2O; felsic minerals ! CaO, MgO, FeO; mafic minerals
Chemical type Acid Intermediate Basic Ultrabasic
Texture Rock names
Fine grained, Rhyolite Andesite Basalt
Extrusive Obsidian Dacite
Volcanic lavas Pitchstone
Medium to
Coarse grained Granite Diorite Gabbro Peridotitea
Plutonic Granodiorite Pyroxenitea
Metamorphosed equivalents are soapstone and serpentine
6.5 Igneous Rocks 161

6.5.2 Diorite representative of most andesites formed in vol-

canic arc areas.
Sometimes termed the “black granite,” diorites The PPL and XPL views of andesite clearly
are medium-to-coarse-grained igneous rocks. It demonstrate the microcrystalline ground mass of
has less than 5 % silica by definition and is plagioclase. The ferromagnesian crystals together
mineralogically 50 % ferromagnesian. Granodi- with larger phenocrysts “float” in the andesine.
orite, by contrast, has more than 10 % silica as
feldspar (orthoclase), along with amphibole and
pyroxene. Many of these intermediate color
6.5.4 Basalt
index rocks are melanocratic or dark colored,
hence their common name. The mafic crystals
Basalt is the most common extrusive igneous rock.
of pyroxene and hornblende along with biotite
Basalts form at the mid-ocean ridges, rifts, and hot
are mainly responsible for the coloration. Quartz
spots. Basalts are all the ocean’s floors. Basalt is
and potassium feldspar, if present, occur in trace
the most abundant fine-grained igneous rock.
amounts. Diorite was used in sculpture by the
One form of basalt is “pillow lava” commonly
Egyptians in the Neolithic into Pharaonic times
seen in ophiolites (Fig. 6.13). Basalt has a long
(Lucas and Harris 1962). The cosmetic palettes,
history of human use from grinding stones and
so popular with ancient Egyptians, were made
hammers to construction stone and sculpture. In
from diorite as were mace or axe heads, bowls,
the last category, the great carved heads of the
and vases (supra). Caton-Thompson and Gardner
Olmec culture were made from basalt (Winkler
(1934) reported diorite artifacts from the Fayum
1994). Basalt was quarried in the Old Kingdom
Neolithic sites as well. In Egypt, there occurs a
of Egypt in the Fayum, west of modern-day
dioritic-gneiss often called diorite by early
Cairo (Lucas and Harris 1962). Because this
rock does not demonstrate conchoidal fracture,
Plagioclase crystals are lath-like in thin sec-
some archaeologists have difficulty in
tion. Twinning is common with the single
recognizing intentional fabrication of the stone
Carlsbad Twin to multiples of the Albite variety.
as tools.
Potassium feldspar will be interstitial to the pla-
Commonly, basalts are composed of plagio-
gioclase. Interspersed through the thin section
clase and ferromagnesian minerals—magnetite,
will be the brown and green crystals of the
olivines, pyroxenes, etc. However, the relative
ferromagnesian minerals—hornblende (amphi-
abundance of these minerals is quite variable.
bole) and biotite.
Poldervaart and Hess (1951) stated the follow-
ing: “In most basalts of oceanic islands and in
many other alkalic basalts, orthopyroxene or
6.5.3 Andesite pigeonite is generally absent. Olivine appears as
early phenocrysts and continues to crystallize for
General properties: Andesite is considered to the a considerable interval.” The glass in basalts will
extrusive equivalent of diorite (Andrews appear as a dark groundmass—genuinely so in
et al. 1997a, b). Most andesites are porphyritic. XPL—and form the backdrop for the plagioclase
Colors range from gray to dark gray. The rock laths and ferromagnesian crystals. The optical
contains ferromagnesian minerals—pyroxene, properties of the plagioclase have been detailed
amphibole, and biotite—with feldspar earlier in this section. By far the most colorful
phenocrysts. Fine-grained nature andesite was mineral in a basalt thin section will be the highly
used for hand axes in East Africa. This “andes- birefringent olivine with blues, pinks, and reds
ite” is only so based on its chemistry—it is not prevalent.
162 6 Petrography for Archaeological Geology

Fig. 6.13 Pillow lava

basalts in road cut,
Curaçao, the Netherlands
Antilles. Clipboard for
scale (Photograph by

6.5.5 Rhyolite and Obsidian geologists renewed a long-running attempt to

provenance the famous bluestone rhyolite build-
General properties: An extrusive porphyritic vol- ing stones. Recent geological research by
canic rock but unlike basalt, it is felsic in nature. Rob Ixer (University College London) and
Some geologists carry the definition into the cat- Richard Bevins (National Museum of Wales)
egory of felsic volcanic glasses making obsidian suggests that many of the bluestones came from
a variety of “rhyolitic” rocks. Taking this the north side of the Preseli Hills (Ixer and
approach, the rhyolites can range from totally Bevins 2010). At Craig Rhos-y-felin in the
crystalline to totally glassy hyaline. Quartz and Brynberian valley, a tributary of the Nevern, the
feldspar phenocrysts are common in fine-grained quarry for one of the rhyolite monoliths whose
rhyolites and obsidians (Mackenzie et al. 1982). debitage has been found at Stonehenge
In antiquity, in igneous terranes, early humans (Fig. 6.14). The rhyolites are fine grained in
used rhyolitic rock for tools such as hand axes character and lack any obvious distinctiveness,
and crude flake implements. Obsidian, which in hand specimen or thin section (Fig. 5.14). This
does fracture conchoidally, was a favorite mate- makes determining their provenance difficult.
rial of early lithic technologists as discussed ear- Zircon mineral chemistry argues persuasively
lier. This volcanic rock has, by far, the sharpest that rhyolitic rocks cropping out in the Pont
edge of any cutting medium and was also Saeson area of north Pembrokeshire, located on
recognized for its beauty. Melrose green rhyolite, low ground to the north of Mynydd Preseli, are
from southeast New England, was used for Mid- the source of the informally termed “rhyolite
dle Archaic to Late Woodland Period artifacts with fabric” which occurs as field debris at
and has been described by Hermes et al. (2001). Stonehenge. This represents the first authorita-
This pale green rock was confused by several tive provenance of any of the rhyolitic compo-
archaeological workers before being nent of the Stonehenge bluestone assemblage.
characterized, mineralogically, using optical Because of their felsic nature, thin sections of
and geochemical methods (supra). these rocks can be pretty colorless in thin section.
Coincident with expanded archaeological Indeed, the only birefringent crystals will be
investigations at the site of Stonehenge, those of an accessory or accidental origin. The
6.5 Igneous Rocks 163

Fig. 6.14 Bluestone rhyolite quarries, Fishguard Volcanic Group at Pont Saeson, North Pembrokeshire (Map modified
from Fig. 1, Bevins et al. 2011)

microcrystalline to glassy ground mass will be antiquity only rarely for building stone because
brown to gray in PPL and quite dark in XPL. The of the friable nature of the rock matrix. It was
phenocrysts of potassium feldspar will show lit- not uncommon to see this material used in
tle twinning, while those of plagioclase will dem- smaller artifacts such as polished implements
onstrate abundant twins. The quartz phenocrysts and items such as cups, bowls, plates, and
will be of a higher relief than the groundmass but ceremonial items such as mace heads. Egyptian
appear clear to colorless in PPL and gray in XPL culture of prehistory used both diorite and gabbro
with some extinction (Fig. 6.15). in all these categories (Caton-Thompson and
Gardner 1934).
In the olivine gabbros, augite and plagioclase
6.5.6 Gabbro/Diabase enclose equant and round olivines (Mackensie
et al. 1982). Plagioclase generally occurs in
General properties: An intrusive, mafic igneous laths, particularly in intergranular forms although
rock seen in the dikes of ophiolites. A fine- the laths are not aligned or oriented as in trachytic
grained gabbro often found in ophiolites dikes gabbros where this is seen. The optical properties
and sills is given the name “diabase” by of the dominant plagioclase, augite, and olivine
European workers. In the section, the gabbroic are colorful under XPL. Diabases in thin section
layers lie between upper pillow lava basalts and vary from olivine to pyroxene-rich plates and laths
the lower mantle rock. Varieties range from of these minerals as well as ilmenite and magne-
olivine-rich gabbros to plagioclase-rich varieties. tite. Calcite is often seen with its bright colors in
Gabbro contains no quartz, by definition. These XPL. Diabases show wide textural variation, even
rocks are coarse grained (phaneritic) and are in thin sections, and some are notably porphyritic
generally dark. The gabbros were used in (Williams et al. 1955).
164 6 Petrography for Archaeological Geology

6.5.7 Tuff

Tuff is an extrusive felsic-to-intermediate

pyroclastic igneous rock composed of volcanic
ash and a variety of mineral grains and rock
fragments. Its color ranges from dull earth
tones of light brown-pink to grays. Crystals
and inclusions of sedimentary rock, lava
fragments, and other pyroclasts are common
within tuffs. Tuffs do appear in use for con-
struction in antiquity but rarely. In ground or
polished stone vessels, the use of tuff does
appear in early dynastic Egypt. A very small
fraction (4 %) of stone vessels found in The
Old Kingdom (First Dynasty) tomb of
Hamaka, at Saqqara, were of “ash” or tuff
(Emery 1996).
At low power magnification, the heterogenous
texture of this rock is quite apparent. Ash will
appear as dark inclusions, while the glass inclu-
sion will be colorless and opaque or translucent
in PPL. Welded tuffs will have a glassy ground
mass, with lamination apparent. The quartz and
feldspar crystals will be elongated or flattened
with little crystal shape evident except in the
unwelded tuffs.

6.6 Metamorphic Rocks

Metamorphic rocks represent rock that “changed

form”—recrystallized due to changes in tempera-
ture and/or pressure. These changes typically take
place along plate junctions and intrusive/extrusive
igneous boundaries. The most restricted form of
metamorphism is that of contact metamorphism
Fig. 6.15 (a) Thin-section views of strongly foliated rhy- directly adjacent to intrusive igneous bodies
olite with small, dark (1 mm) lenses elongated in the plane of where heat and hydrothermal fluids can cause
foliation. Sample 142/1947.22 from Wilsford, described
originally by Stone (1948) and subsequently by Ixer and
new crystallization (neomorphism) and precipita-
Bevins (2010) (Courtesy of Professor Michael Parker tion or replacement (metasomatism) of minerals.
Pearson). (b) Photomicrograph of strongly foliated rhyolite Most metamorphic rocks result from regional
with small, dark (2 mm) lenses elongated in the plane of metamorphism which occurs over large areas
foliation. Sample SH08/265 from the 2008 Stonehenge
excavations (Courtesy of Professors Tim Darvill and
associated with mountain building. Within moun-
Geoffrey Wainright). (c) Photomicrograph of strongly foli- tain belts are slates formed from shales, quartzites
ated rhyolite with small, dark (0.5–1.5 mm) lenses elongated from sandstones, gneisses from sediments and
in the plane of foliation. Sample PS8 from the Fishguard granites, serpentine and steatite from olivine-
Volcanic Group at Pont Saeson, North Pembrokeshire
(cf. Fig. 6.14) (Photomicrographs per Richard E. Bevins;
pyroxene-talc-rich rocks, marbles from limestones,
the National Museum of Wales and Wiltshire Studies) and greenstone from basalt and andesite. The time
6.6 Metamorphic Rocks 165

Fig. 6.16 Metamorphic

facies by pressure and
temperature. “Low”
temperature or “blueschist”
facies are typified by
steatite (talc); marbles are
more “greenschist” facies

scales are long in most instances ranging from low-grade metamorphic rock. Texture refers to a
thousands to millions of years. metamorphic rock’s fabric orientation and direc-
Metamorphic rocks are described by texture tion or the lack of it, e.g., foliated versus
and mineralogy. These rocks can have igneous, non-foliated. The bulk of metamorphic rocks dem-
sedimentary, or metamorphic origins and as such onstrate some form of foliated texture (Fig. 6.16).
retain much of the character of the original lithol-
ogy while being the changed species that they are.
Some of the most archaeologically important 6.6.1 Metamorphic Rocks: Foliated
metamorphic rocks are marble, quartzite, and
soapstone. Upon finding such rocks away from Slate: a very planar, parallel foliation.
an orogenic zone, such as on a coastal plain with Phyllite: Foliation of fine-grained sheetlike
nothing but sedimentary facies for hundreds of minerals such as muscovite and biotite.
kilometers, an archaeologist must seek cultural Phyllites are known for a glossy luster called
(trade, exchange, curation) reasons for their pres- “phyllitic sheen.”
ence in these sites. Metamorphic rocks are classi- Schist: Fine-to-coarse-grained platy minerals
fied according to their mineralogy, texture, facies, are aligned in parallel to subparallel fabric
or grade. In terms of their mineralogy, V. M. which is obvious to the eye. Like phyllite,
Goldschmidt, in 1912, correctly observed that such a rock is said to have “schistosity.”
most metamorphic rocks contain only a few— Schists fracture along the foliations. The vari-
four to five—major minerals (Philpotts 1989). ety of minerals in schists is diverse, e.g., chlo-
The mineral assemblage, in a metamorphic rock, rite, tourmaline, muscovite, biotite, garnet,
helps the petrologist to interpret its thermody- hornblende, staurolite, kyanite, and
namic history. This is less important to the archae- sillimanite.
ological geologist, but the relationship between Gneiss: Gneisses have alternating layers of dif-
metamorphic grade—low, intermediate, and ferent minerals such as quartz and biotite. In
high—and texture is good to know. Grade refers Europe, gneiss refers to coarse-grained, mica-
to temperature and pressure, with a high-grade poor rocks irrespective of fabric (Yardley
metamorphic rock more recrystallized than a et al. 1990).
166 6 Petrography for Archaeological Geology

6.6.2 Metamorphic Rocks: metamorphic origin, was used in making

Non-foliated sarcophagi and statuary.
Under the microscope, quartzite is a fabric of
Hornfels: The minerals in hornfels are micro- interlocking quartz of varying grain size. Smaller
crystalline giving the rock the property of grains are seen in quartzites that have undergone
conchoidal fracture like chert and obsidian. more intense deformation (Philpotts 1989). The
Granulite: This is an even-grained garnet or grains of quartz will show the optical properties
pyroxene-rich rock lacking the micas or of this mineral—low relief, little or no birefrin-
amphiboles. Its essential components are gence, colorless, and some undulatory extinction
orthoclase feldspar, plagioclase, and quartz. in more isotropic grains. Unlike a traditional
This texture is observed in marble, quartzite, sandstone, a quartzite can show a lamellar or
greenstone, serpentine/serpentinite, and parallel fabric to the grains.
skarn. It is plane foliated that can be laminated
parallel to the foliation. This characteristic Marble
leads to slablike breakage and usage in con- Marble is one of, if not, the most archaeologi-
struction in antiquity. cally important of the metamorphic rocks. Fine-
Amphibolite: This rock is medium to coarse grained varieties have been used for sculpture
grained. The rock is mainly hornblende and since antiquity in both the Old and New Worlds.
plagioclase. The parallel alignment of the In the classical period of Greco-Roman tradition,
hornblende crystals gives the rock its charac- marble became a preferred building stone and
ter but seldom produces a foliation. continues to enjoy a wide usage today as both
an interior and exterior stone. In the Aegean,
marble appears as a quarried stone and sculpture
media in the Cycladic islands third millennium
BCE (Cherry 1988). Cemeteries of islands such
6.6.3 Metamorphic Rocks as Melos, Naxos, Paros, and Andros have pro-
of Archaeological Interest duced thousands of marble figurines and vessels.
In the case of Melos, the marble must come from Quartzite adjacent islands as it has none of its own Melos
General properties: This metamorphic rock gen- as we have seen was an early (Neolithic) source
erally originates as a quartz-rich sandstone facies for obsidian.
such as an arkose or arenite. It is a product of In Egypt, marble was used in the Neolithic for
regional metamorphism. As a rule it is a white bowls and vases (Mannoni and Mannoni 1984).
rock weathering or stained to colors, such as By the Old Kingdom, at least at Saqqara, site of
cream or orange, in iron-rich sediments. Quartz- the first pyramid—the so-called Step Pyramid—
ite can appear gray when minerals such as biotite a royal tomb was constructed of calcareous mar-
or magnetite are present with the quartz grains. ble in the twenty-eighth century BC. In Archaic
Quartzite is very hard and durable but difficult to Period Greece (ca. seventh century BC), the use
work. As an archaeological material, quartzite of marble for statuary increases with its use in
appears as a variety of implements such as life-sized, and greater than life-sized,
spear points, knives, perforators, or almost any representations of young men (kouros/kurai)
stone tool form. Because of its texture and color- and young women (kore/korai) (Fig. 6.17). Ante-
ation, the whiter varieties were fashioned into cedent to classical statuary, these kurai were
adornment such as beads, pendants, and the almost always portrayed nude or nearly so,
like. In some of the earlier literature, on the use while the kore was elaborately swathed in
of quartzite in antiquity, such as Egypt, the rock gowns called peplos. Diadems were commonly
was described as “hard sandstone” (Lucas and seen on the heads of kore. The kouros and kore
Harris 1962). This “quartzite,” if indeed it is of were not the initial artstone products of Greece.
6.6 Metamorphic Rocks 167

Fig. 6.18 Livia. Marble portrait (Courtesy the Estate of

Dr. Norman Herz)

isotopic study of a Classical Period Aphrodite,

whose origin was likely Carrara, is presented.
Marble is a granular rock, composed princi-
pally of calcite and is derived from limestones.
This, of course, oversimplifies the diversity
encountered in marble texture and coloration
which, in the latter, ranges from the whites of
Greece’s Pentelic facies to the gray marbles of
Vermont (USA) and pinks to reds of Georgia and
Alabama, USA, and Carrara, Italy. Richly col-
ored banded or “streaked” varieties are found in
Italy and elsewhere and doubly metamorphosed
brecciated varieties (Carrara). The color and tex-
tual variety of those marbles is, of course, the
result of their precursor minerals and their meta-
morphism. The marbles of peninsular Italy
(Fig. 6.18) are generally formed at depth by
regional metamorphic processes, while those of
Fig. 6.17 The “almond-eyed korai” at the Acropolis the Aegean islands can be the result of contact
Museum, Athens (Photograph by the author) metamorphism.
The white marbles, like Pentelic, are almost
This honor falls to the Cycladic island figurines 100 % calcite while impure quartz, carbon,
first seen in the Early Bronze Age (3700–2700 clay, and iron give color to others—quartz,
BC). Contemporaneous with the marble statues clay, and iron produce yellow or pink color;
were marble bowls and marble-tempered carbon and other organics produce gray-to-
ceramics (Rapp 2002). black marbles; and greenish clay-rich marls,
Into the Classical and through the Hellenistic micaceous minerals, and serpentines produce
Periods of Greece, the demand for marble for gray-to-green colors. Marble colored by its com-
both building and sculpture opens quarries across ponent minerals is termed idiochromatic and
the Mediterranean adding to those of Egypt and those by inclusion pigments in cements between
Greece, with sites in Anatolia, Sicily, and Italy, grains or in veins and intertices are called
most notably Carrara. In Chap. 8, a detailed allochromatic.
168 6 Petrography for Archaeological Geology “Marbles” or When Is a Marble Not Tectonic breccia, such as that of Trets-
a Marble? Pourcieux (France), is a truly a metamorphosed
“The stone found in the Brescia quarries (Italy) is a rock. The metamorphism of this calcarenite
sedimentary limestone. It is called a marble shows features such as stylolitic (“tooth edged”)
because it is a hard limestone” (Giuliano Ghirardi hematite crystals and mylonization (Mazeran
quoted in Stone World, 2011). The particular “mar- 1995). The breccias of the French Maritime
ble” referred to by this geologist is the Botticino Alps are not metamorphic and are calcareous
Agglomerate (Fig. 6.19). Another famous “mar- and dolomitic avalanche breccias. The matrix is
ble” is Rojo Alicante marble (Fig. 6.19) that is a calcareous-argillaceous cement rich in hema-
petrographically a biomicritic limestone (Lopez- tite. The metamorphosed Italian breccias
Buendı́a et al. 2010). “Carthage marble” was quar- (“Dorata,” “Giallo Broccatello,” “Rossa”)
ried in the southwestern portion of Missouri (USA) have been studied petrographically with their
from before the 1860s and used extensively for provenance near Sienna (Bruno and Lazzarini
paving and dimension stone (Rovey et al. in 1995). The Breccia Rossa Appenninica, in thin
Hannibal and Evans 2010). It, too, is really a section, shows its marine origin with abundant
hard limestone classified as a “marble” stone. pelagic bivalves, radiolarites, and brachiopods
Four varieties of breccias and conglomerates embedded in a dark-brown micritic cement,
have been identified and characterized relative to hence its name. Breccia Dorata, a favorite
use in Gallo-Roman architecture in southeastern for Roman monuments, has clasts of micro-
France (Mazeran 1995). Within this suite of crystalline calcite, with quartz cemented with
materials was a range typical of the variety calcite. The cement varies from golden-yellow
rightly imputed to “marbles.” The “breccia” of to pink.
Vimines, used in Roman decorative plaques and Figure 6.20 is an excellent example of lime-
columns, has a carbonate cement of a micritic/ stone playing the sculptural role of marble in
microsparite nature. The “Bordeaux Conglomer- antiquity. The sculpture, from relief of a tower
ate,” classified as a “pudding stone,” is held tomb, found at the Germania Superior city of
together by a molassic-glauconitic-cement of Aventicum (modern-day Avenches, Switzerland),
strictly marine origin. It is termed “breccia as reconstructed, is typical of marine-themed
frutticulosa” by Italian archaeologists. In PPL, funerary sculpture. The particular subject in this
the grains of the cement show a characteristic work is Triton abducting a sea nymph or Nereide.
green of glauconite and, in XPL, weak birefrin- Discovered in 1989, during a salvage archaeo-
gence. The cement was rich in calcite that is logical project, the sculpture illustrates the artis-
given to bright, high-order colors. Oxidation of tic usage of high-quality local limestone for
ovoid glauconite grains will produce brown large-scale mortuary architecture (Castella
margins because of their ferrous iron content. 1998).

Fig. 6.19 Left, Botticino

Agglomerate limestone
(Italy); right, Rojo
Alicante limestone (Spain)
6.6 Metamorphic Rocks 169

Fig. 6.20 Limestone used for sculptural relief of a Roman Triton abducting a sea nymph or Nereide. Roman Museum
tower tomb, Aventicum, Germania Superior (modern-day of Avenches (Photograph by the author)
Switzerland). The subject is that of Greek mythology, Serpentinite and Steatite 1984). These rocks are serpentines with veins of
Serpentinite is a low-grade metamorphic rock carbonates.
originating from recrystallization of mafic Steatite or soapstone is a general term used to
minerals such as olivine and pyroxene (Williams describe metamorphic rocks that are composed
et al. 1955). primarily of talc, hydrous magnesium silicate,
but which may contain varying quantities of
Peridodite þ H2 O ! SerpentineðAntigoriteÞ other minerals, carbonate, amphibole, magnetite,
þ magnetiteðFe3 O4 Þ þ H2 and chlorite. Steatite deposits form as a result of
regional or contact metamorphism due to the
Olivine, in hydrothermal fluids, reacts to form metasomatic alteration of an original parent
serpentine at low temperatures, in the near sur- body. These parent bodies are normally
face (Yardley 1989). Some European texts ultrabasics, such as serpentinites, peridotites,
(Monttana et al. 1977) as well as most early dunites, and pyroxenites, or in some rarer cases,
archaeological workers have had difficulty the original deposit may be a carbonate sedimen-
differentiating serpentine, the ultramafic mineral, tary rock. In any provenancing study, it is impor-
from the metamorphic rock, serpentinite. Indeed, tant to recognize the nature and extent of
the terms serpentine and serpentinite appear utilization of the artifacts derived from the raw
interchangeably for the description of the meta- material of interest (Bray 1994).
morphic rock. Brecciated rock such as the Steatite, in the United States and Canada, is a
ophicarbonates of Northern Greece (Thessaly) talc or “olivine-talc-carbonate rock,” such as
produced from peridotites, mistaken or miss- actinolite, low in serpentinite and the accessory
named as “marbles,” gives interesting textural talc which grades into high talc (~10 %) ultra-
and color patterns (Mannoni and Mannoni mafic metamorphic rock. Steatite is also known
170 6 Petrography for Archaeological Geology

Fig. 6.21 Soapstone

Ridge, Georgia, USA.
Upper, bowl “blank” in
situ; left, thin section
showing high birefringence
of talc; right, bowl

as “soapstone” in the New World. It is also called are completely recrystallized with no trace of
“potstone” in the Old. The latter term reflects its relic grains. These are quickly differentiated
use in Stone Age cultures as a vessel material. In from the dark serpentine, in XPL, due to their
eastern prehistoric North America, soapstone/ relief, birefringence colors, and other optical
steatite vessels appear from the Late Archaic properties (Fig. 6.19). The talc, in steatite, will
Period to the Early Woodland Period cultures be especially birefringent showing bright yellow
(1800 BC–AD) of the Appalachian Mountains and other colors reminiscent of muscovite.
(Fig. 6.21) (Wells et al. 2015). In the prehistoric
sites of the Chumash and Gabrielino cultures of
California, soapstone was used in vessel manu- 6.6.4 “Greenstone”
facture. The famous pipestone steatite quarries of
eastern South Dakota were used by American “Greenstone” is defined as any altered basic
Indians to fashion elegant “pipestone” pipes for igneous rock which is green in color due to the
tobacco. In many steatites, chlorite appears as an presence of chlorite, actinolite, or epidote
important accessory mineral. The color of (Parker 1994); this rock is metamorphic and, as
serpentinite and steatite generally runs from a category, is used by archaeologists to describe
dark green to lighter gray-green varieties due, stone artifacts such as axes. The coloration and
in the latter case, to talc and chlorite. texture give the material its name rather than a
The serpentine crystals of serpentinite and strict reading of its mineralogical makeup. Actin-
steatite are generally replacement of preexisting olite schists have both the desired color—green
olivines and pyroxenes and the fabric may reflect to almost black—and the hardness that a tool or
this in PPL. The texture is lamellar in most cases implement must have. The rock polishes into
with crystals of pyrite, magnetite, and regrown what is termed “ground stone.” It is its hardness
olivine sometimes present. Some serpentinites and toughness which make this material so
6.6 Metamorphic Rocks 171

attractive to human groups and dominate the way western Switzerland’s perforated “battle axes” is
it is worked. While it takes a moderately good clearly metamorphic, and in most cases, the
edge, it withstands shock well and so can be used material is either a schist—specifically a schist
for artifacts which might shatter during use. In rich in actinolite, chlorite, and/or epidote—or a
the late Neolithic Europe, Switzerland, and serpentinite. The texture of the former is rough or
neighboring lands—Austria, France, etc.—both raw rock is schistose where, in the finished arti-
solid and shaft-hole axes are quite common fact, the polished stone is dominated by the color
(Desor 1873; Petrequin et al. 2005). of the actinolite and limonite grains. In thin sec-
These axes are made from a variety of “green- tion the amphibole actinolite appears green,
stone” or “roche verte” (Ramseyer 1992; while the quartz and feldspar are granoblastic
Petrequin et al. 2005). The shaft-hole axes are with attendant optical properties of these
also known as “battle axes” or “hammer axes” minerals. Actinolite grains will show a pro-
(Briard 1979; Champion et al. 1984). Also termed nounced pleochroism. In XPL, the epidote will
“jadeite,” which produces some confusion as jade- show bright birefringence colors. As we have
ite, per se, it is mineralogically a pyroxene. What previously said, serpentinite—a greenchist facies
we can say is the axes are uniformly made of a rock as well—is fine grained with low birefrin-
green “rock”—light to dark—that takes a good gence. In PPL jadeite crystals demonstrate a high
polish. Swiss geologists identify polished stone relief and they typically appear as platy and
materials ranging from serpentine and quartzite interlocked. Under XPL, the jadeite grains show
to diorite (Winiger 1981). Seen in western a low birefringence.
Switzerland and elsewhere in the Alps is a suite
of metamorphosed oceanic/pelagic sediments Slate
within which occur the serpentinized peridotites A common regional metamorphic rock of pelitic
(Hsu 1994). During subsequent glaciations of the or pelitic-arenaceous origin, slate’s most
Pleistocene, the ophiolitic rocks were deposited recognized precursor, is shale. It is fine grained
along the morainic margins of the Rhône Glacier and demonstrates a planar fabric that gives rise to
and readily available for the Neolithic artisans. the definition of slaty foliation. The slates are
Elsewhere, ophiolites occur along other sub- uniformly gray to black, and the principal
duction zones such as that of the Aegean basin. components of the fine-grained ground mass are
An example is the use of metamorphic rock, in chlorite, muscovite, quartz, and hematite
particular jade, in which, as we have seen, either (Yardley 1989). Other minerals are andalusite
jadeite or nephrite, is the dominant mineral and cordierite which can form elongated crystals
(Higgins and Higgins 1996). On the Cycladic (porphyroblasts) as do tourmaline, rutile, and
island of Syros, serpentinite and jadeite rocks pyrites (Williams et al. 1955). Graphite
occur in quantity and were used for axe manu- represents the lowest grade of metamorphism.
facture. “Greenstone” (nephrite and bowenite) This rock is found throughout antiquity with its
was a major resource for the human occupants usage as palettes in Egypt, for flooring and roof-
of New Zealand prior to contact with Europeans. ing, and in its most ancient archaeological
The sources for this material are found in a num- appearance as a slate “tablet” for Paleolithic
ber of locations within the South Island. etchings of mammoths and figures at
Given the disparity in the identification of Gönnersdorf (Bosinski 1992).
archaeological “greenstone,” it is likely that the In thin section, the fine-grained layering of the
choice of specific material was simply made to fit parent facies is often preserved in the ground-
local lithic resources and we, as archaeological mass. XPL illuminates the micaceous minerals
geologists, apply our specificity to the individual present with high birefringent colors.
case. As such the general and optical properties Porphyroblasts will demonstrate the appropriate
of the rock will follow that of the specific miner- optical properties for the specific mineral: in
alogy involved. For instance, the “roche verte” of reflected light pyrites appear yellowish-white in
172 6 Petrography for Archaeological Geology

PPL; rutile with red-brown pleochroic colors in Georgia, roughly 3  1 m in size. It is covered
XPL; and chlorite with high relief, more brightly with, on its upper portions, concentric circle
colored in PPL but less so in XPL. motif petroglyphs of unknown age. It was first
described by C.C. Jones in 1873 in his Antiquities
of the Southern Indians (Smith 1964). It has been Gneisses and Schists suggested that the stone originally stood upright
The varieties of these rocks are extensive usually in the manner of a menhir with the “upper” end
related to their mineralogy. Grain size is medium (left in the picture) being an effigy. Schists are
to coarse in both types with augen gneisses hav- diverse reflecting their parent and metamorphic
ing a porphyroblastic texture. Most gneisses mineralogy.
appear lineated or banded with alternating Commonly found schists include quartz-
quartz/feldspar and amphibole/biotite. Schists biotite schist, hornblende schist, muscovite
are foliated. Figure 6.22 shows a petroglyphic schist, chlorite schist, and talc schist. Most
granitic gneiss boulder found in Forsyth County, schists form from shales and the classic

Fig. 6.22 Petroglyphic

boulder. Granitic gneiss
6.7 Techniques: Optical and Otherwise 173

metamorphic sequence for this process is (chlo-

rite) slate, schist, biotite-garnet schist; staurolite
kyanite schist; and sillimanite schist. Previously
illustrated in Fig. 2.3, the dolmen tombs of Petit-
Chasseur used extensive amounts of gneiss and
schist for their construction of 13 megalithic
tombs of which the more imposing had stelae
erected around 3000 BC by Neolithic
communities living just to the east of the modern
city of Sion. After functioning for more than a
millennium, the cemetery was abandoned around
1600 BC, disappearing from human memory
until its rediscovery by archaeologists in 1961.
A stela made from schist is shown in Fig. 6.23.
The grade of metamorphism—low to high—is
indicated by the mineral assemblage with the
micas at the low end and sillimanite at the high
end. As we shall see in the last section of this
chapter, the firing of clays, in ceramic manufac-
ture, closely follow the metamorphic mineralogy
of shale with the low–high-grade minerals—
mostly A12SiO5 polymorphs—forming as the
firing temperature is increased from
earthenwares to porcelains.
Gneisses contain layers of quartz and feldspar
along with other minerals. In thin section, the
low relief and birefringence of the quartz and
feldspar contrast with the darker minerals. The
mineral assemblages of schist are related to the Fig. 6.23 Menhir carved from schist discovered in the
Canton of Neuchâtel, in 1997, during excavations along
grade of metamorphism. Where the schist has the then proposed autoroute A5, in the vicinity of the
formed from sandstone, the rock will have exten- village of Bevaix. The stone is clearly carved as stelae
sive quartz and feldspar in layers much like the and is dated to the Neolithic Period, ca. 5000 BP. The
more common micaceous varieties. Because of height is 3.2 m and the weight is 2,800 kg. On exhibit,
Latenium Museum, Neuchȃtel, Switzerland (Photograph
the deformation of these minerals, the grains will by the author)
be elongated and, in the case of the feldspars,
lath-like. The PPL and XPL properties will be
consistent with those we have listed previously. 6.7 Techniques: Optical
As the metamorphic grade increases, the XPL of and Otherwise
their thin sections produces a riot of highly
birefringent colors starting with the micas As rocks are collections of minerals, so their
continuing through the higher temperature identification is the principal goal of petrogra-
minerals of the sillimanite group—staurolite, phy. The most straightforward means is as just
andalusite, kyanite, and sillimanite. Most of mentioned—a 10 hand magnifier. Again, the
these minerals have a relatively high relief and key to one’s choice of identification method is
a few have distinctive colors. Staurolite is pleo- specificity. If one’s goal is a simple determina-
chroic yellow-pale brown. Sillimanite and kya- tion of the class of rock being examined, it is
nite have moderate birefringence and are fibrous reasonably sure that a distinction between sedi-
to prismatic with clear-cut extinction. mentary, igneous, and metamorphic rock types
174 6 Petrography for Archaeological Geology

can be accomplished with a good hand magnifier. color is that it should be used in concert with other
If one wishes to ascertain a specific type of sedi- criteria to assure confidence in identification.
mentary rock—clastic, carbonate, or evaporate—
then this too is possible for the trained eye. Streak Streak is the color of powdered
Beyond this level of classification becomes minerals. This property is a more reliable indica-
increasingly more difficult. The next level in pet- tor of color than that of the visual color of the
rographic identification requires the optical hand specimen. One can either rub the mineral
microscopes. This can be done with a low-power sample on a porcelain tile or powder the sample
binocular microscope using reflected light or a to obtain the streak. Hematite streaks red; galena,
more powerful, high (optical)-resolution petro- silver-gray; hornblende, green; talc, white; azur-
graphic microscope used to study thin sections of ite, blue; and magnetite, black.
rocks and minerals. Beyond the optical methods
are X-ray, electron, particle beam, and magnetic Luster This is the property that describes how
instrumental techniques, of which X-ray methods the mineral absorbs, reflects, or refracts light.
are the most common (Chap. 8). In this chapter are Broad categories of luster are “metallic” and
presented both hand-specimen and thin-section “nonmetallic.” Pyrite and galena have metallic
examples of rocks and minerals important in luster. Quartz has a “glassy” luster as does oliv-
archaeology. ine; soapstone (steatite) has a “greasy” or “resin-
ous” luster as do many flints.

6.7.1 Hand-Specimen Macroscopic Cleavage and Fracture The crystalline struc-

Analysis of Rocks and Minerals ture of the particular minerals is observable in
their geometric shapes showing clear member-
ship in one of the six crystal systems: cubic
Color This is one of the most obvious physical (pyrite), monoclinic (gypsum), triclinic (plagio-
properties of lithic materials. It can be diagnostic clase feldspar), orthorhombic (sulfur), tetragonal
but as a rule it is not the major identifying charac- (chalcopyrite), and hexagonal (apatite). A
teristic as it is in the studies of soils and sediments. euhedral shape is one where the mineral grain
The reason for this is relatively simple. There are has distinguishable crystal faces on all sides such
12 soil orders while there are over 2,000 minerals. as calcite. Some mineral grains have only a few
Add to this fact that many archaeological rocks crystal faces and are subhedral, while grains with
and minerals vary in their color (allochromatic), no crystal faces showing in a rock are anhedral.
even within a specific type. Quartz has “smoky,” Most minerals are subhedral as a rule. Minerals
“rose,” and “amethyst” color varieties even break or cleave along planes that reflect the crys-
though they are all quartz. The same thing is tal system. In general, minerals with low crystal
observed in flint or chert which has a host of symmetry (triclinic) exhibit distinctive cleavage
varieties in different colors (cf. section later in such as the feldspars. Other minerals with high
this chapter). Some minerals, such as the metallic symmetry are less likely to exhibit distinctive
ones, tend to have reasonably characteristic cleavage. Fracture in rocks must be distinguished
shades of one basic color (idiochromatic) such as from that seen in minerals. Likewise fracture is
copper ores: blue (azurite) and chrysocolla (blue not cleavage. Cleavage planes are bonding
to green)—and the very commonly occurring mal- surfaces of the crystal. Fracture can occur along
achite is green. Chalcopyrite, a common copper cleavage planes but it does not have to. Fracture
ore in antiquity, is yellow. Native varieties are surfaces can be nonplanar, nonparallel because a
generally more predictable—sulfur is yellow, cop- rock is a collection of minerals reflecting this
per is reddish brown, slate and graphite are gray to heterogeneity in its overall response to breakage.
black, etc. The best one can say of the property of A good example is the important archaeological
6.7 Techniques: Optical and Otherwise 175

material—flint—which does not exhibit cleavage the silicate minerals, all of the feldspar groups—
but rather fractures in a distinctive, conchoidal albite to sanidine—form twins. Calcite is distin-
manner. Obsidian, a volcanic rock, does not have guished, by twinning, from its close mineralogic
a crystalline structure and, hence, no cleavage cousin, dolomite.
property. Obsidian does fracture conchoidally. Twinning is observed at the hand specimen
The other common fracture forms are irregular level and seen under the petrographic microscope
and fibrous. An archaeological material that where it is used, in the latter case, to identify
fractures irregularly is basalt as does most minerals in thin sections. Twins result from
quartzite although the latter can exhibit conchoi- errors in crystal growth. These errors result in
dal fracture. characteristic twin forms and follow what are
called “twin laws”—the Spinel law, the Albite
Hardness Hardness is a reliable property of law, the Carlsbad law, etc. Orthoclase twins
minerals that have not been weathered or other- according to the Carlsbad law. Microcline—a
wise altered. Ten minerals have been arranged in potassium feldspar with some sodium—is recog-
a scale known as the Mohs hardness scale. The nizable for its characteristic “crosshatch” or
following list gives the range of hardness relative plaid-like twinning which follows the Albite law.
to some common minerals:
10: diamond—the hardest material; can’t be
scratched 6.7.2 Optical Properties of Minerals
9: corundum—can be scratched with diamond
8: topaz—can be scratched with emerald Minerals, optically, come in two forms—isotro-
7: quartz (also flint, olivine, garnet)—can be pic and anisotropic. The isotropic minerals have
scratched with steel the same optical properties in all directions
6: feldspar and opal—can be scratched with within the crystal structure (recall that one of
quartz the definitions of a mineral is it must have crys-
5: apatite (also hornblende)—can be scratched talline structure—obsidian does not, so it is sim-
with ordinary knife ply a “rock”). Anisotropic minerals vary in their
4: fluorite (also dolomite, many copper ores)— optical properties—ray paths and angles (Perkins
can be scratched with glass 2011). All minerals belong to six crystal systems
3: calcite (also mica, halite)—can be scratched with isotropic minerals comprising one system
with common wire nail or a penny and the other five anisotropic.
2: gypsum—can be scratched with the fingernail Light microscopy and quantitative mineralogy
1: talc—can be crushed between the fingers have been the cornerstones of any study in rocks
and mineral in geoarchaeology. Yual Goren
(2014) points out that several aspects of
Twinning This property is observed in a host of geoarchaeology require the optical microscope as
minerals in several colloquially named forms their primary tool. The instrument of preference is
such as the “fairy cross,” seen in staurolite, and usually the polarizing—or petrographic—micro-
the “fishtail” observed in some gypsum and the scope. The most common usages are, respectively,
“dovetail” in calcite. Other commonly used petrography and micromorphology. While
names for twinned minerals are “albite twins,” borrowed from earth sciences, these two areas
“penetration twins,” “iron cross,” “spinel twins,” have developed a slightly different meaning within
and “star twins.” This short list does not exhaust the archaeological context. While optical
the varieties of twins that are seen across a large microscopes are also used in numerous other
number of common minerals such as calcite, science-based fields of archaeology—such as
feldspar, galena, gypsum, pyrite, quartz, rutile, metallography, archaeobotany, palynology,
spinel, sphene/titanite, stibnite, and wurtzite. Of etc.—their use for micromorphology and
176 6 Petrography for Archaeological Geology

petrography for archaeological purposes has differing velocities. Upon recombination, the
increased over the last few decades. two light paths interfere and create interference
Petrographic thin sections are just that—paper colors. As seen in Fig. 5.16, talc is highly
thin (30 μm) slices of a rock sample that are birefringent. The index of refraction and birefrin-
mounted and polished on glass slides. These gence are consistent optical properties of
slides are then examined using the petrographic minerals used by petrographers. The law of
microscope invented in 1828 by William Nichol refraction, called Snell’s law (see Chap. 7,
(Ford 1918). The use of thin sections was XRD) after its discoverer, in 1621, Willebrord
pioneered by Henry Clifton Sorby in his exami- Snellius, proposed the following familiar form:
nation of Jurassic limestones from Yorkshire,
sin i= sin r ¼ n
England, in 1851 (Adams et al. 1997).
The petrographic microscope is distinguished where i is the angle of incidence for the light
from other types of optical microscopes by its beam and r the angle of refraction as the
reliance on polarized light. Polarized light is light light beam exits the material and n the index
that vibrates in one specific plane. Historically, of refraction or geometric ratio of i and r.
petrographic microscopes use a Nicol prism Values of n fall between 1.4 and 2.0 for most
(so-called for William Nicol) made of calcite to minerals. Typical values are apatite, 1.624–1.667;
produce light that vibrates in only one plane calcite, 1.486–1.740; fluorite, 1.434; gypsum,
(plane-polarized light or PPL). Modern petro- 1.519–1.531; olivine, 1.63–1.88; quartz,
graphic microscopes use an organic film as a 1.544–1.553; and zircon, 1.923–2.015. A few
filter to produce PPL. transparent minerals like diamond (2.418) and
PPL is produced by one polarizing filter rutile (2.605–2.901) exceed the typical range
located below the sample stage. A second, move- for the index of refraction.
able, polarizing filter, with vibration direction set The difference in the highest and lowest
perpendicular to the lower polarizing filter, is indices of refraction in a mineral results in its
located above the sample stage. The light ray birefringence. In petrography, it is termed weak,
emerging from the lower filter vibrates perpen- moderate, strong, very strong, and extreme.
dicular to the upper filter’s vibration direction. Using PPL, alone, it is generally possible for an
With both filters in the light path, we observe the experienced observer, to estimate the refractive
property of extinction. This arrangement is index and make distinctions between different
termed crossed polars or XPL where the lower birefringent minerals.
filter is the polarizer and the upper filter is the Another important aspect of a mineral’s color
analyzer. When a thin section of a mineral or a is pleochroism. This optical property is observed
rock with a variety of minerals is placed between in a few common anisotropic crystalline
the two crossed polars, one can observe various minerals, biotite and hornblende being examples.
intensities and colors because the light is passing Isotropic minerals do not exhibit pleochroism.
through a anisotropic material, such as a carbon- When the petrographic microscope’s stage is
ate, that bends the PPL such that not all the light rotated in PPL, anisotropic minerals will demon-
is extinguished by the upper polar. The resultant strate this property by changing colors. Biotite
light is called birefringence. mica, a common pleochroic mineral, is easily
Birefringence is a diagnostic property of identified by this property coupled with its high
anisotropic minerals—it is not seen in isotropic birefringence. Pleochroism is easily observed at
fluorite, garnet, and spinal or amorphous (glass, the hand specimen level in transparent varieties
man-made, or volcanic) minerals—those of cordierite and spodumene.
materials in which the light ray travels unaffected The final optical property, on grain mounts,
through the material. It is observed in XPL. sometimes mentioned in the identification of
Polarized light, upon entering a crystal, will minerals, that is only observed in thin sections
split (refract) into two components traveling at is relief. Relief, like the properties of
6.7 Techniques: Optical and Otherwise 177

birefringence and dispersion, is directly related Percentage mineral ¼ ðWA=WTÞ  100 ½1
to the refractive index. Relief is basically a sub-
jective and relative characteristic of a mineral This and other methods equated measured
(MacKenzie and Adams 1994). Here the differ- weights or areas to volume equivalents. To mea-
ence is not between the path difference and sure the volume of clay skins, Buol and Hole
wavelength of the light in the mineral but the (1961) projected the microscope image onto a
difference between the mineral, on the thin sec- ground glass, traced the image on translucent
tion, and its mounting medium—epoxy, Cana- paper, and then cut and weighed each piece.
dian balsam, etc. Most mounting media have a They measured clay skin volume as a function
well-known refractive index of 1.537 (1.54) such of horizonation and genesis of translocated clay.
that minerals with a higher refractive index than Johanssen (1938) and Fry (1933) used a camera
1.54 will demonstrate what is termed “positive” lucida attached to a microscope to project the
or moderate to high relief where the mineral image. The periphery of each mineral was traced
appears as bolder than media in PPL. Low or with a planimeter to yield relative proportions.
negative relief would be where the mineral Despite the awkwardness of area measurements,
appears less distinct than the mounting media. recent developments in computer image analysis
The refractive index is very useful in determining utilize this principle.
the identity of a mineral, particularly those iso- The Linear Method – In 1898, Rosiwal devel-
tropic varieties like fluorite. A general rule oped a much simpler method of measuring the
among petrographers is that felsic minerals— volume percentage of minerals (Holmes 1930).
quartz and feldspar—typically show low relief This technique measures lengths of line segments
and are almost colorless. By comparison, rather than areas and samples a portion of the
ferromagnesian (mafic) minerals exhibit high mineral area rather than the whole area, as in the
relief and color, e.g., olivine. Delesse Method. It measures the frequency with
which a mineral is encountered during the tran-
sect. The proportion of the length of the line
6.7.3 Quantifying Minerals in Thin crossing a mineral species (LA) is compared to
Sections the total length of the traverse (4). The percent-
age of a mineral in the specimen is given by
Rocks and minerals, where pertinent their gen- Percentage mineral ¼ ðLA 14Þ  100½2
eral—hand specimen or grain mount—and thin-
section properties, express definite optical The theoretical basis and statistical accuracy for
properties, like that of the macroscopic indices. this method are discussed by Lincoln and Rietz
One of the earliest and most common optical (1913).
techniques of the so-called areal method for
quantifying volume percentages of minerals Reflected Light Microscopy Many minerals,
was that of Delesse in 1847 (Holmes 1930; particularly the metallic ones, do not transmit
Jones 1977). Delesse studied polished blocks, light no matter how thinly they are ground. To
traced the outline of minerals onto paper, then examine such opaque minerals, in polished sec-
transferred the pattern to tinfoil, and cut around tion, plane-polarized light is reflected off the
the boundaries. The tinfoil pieces were sorted upper, polished surface. Many research-grade
into mineralogical classes and weighed to yield petrographic microscopes are equipped with
relative percentages. The volume percentage of a both transmitted and reflected optics. One can
mineral was determined by comparing the weight differentiate isotropic and anisotropic minerals
of each cut pattern representing a mineral species under reflected light. Most isotropic minerals
(WA) to the total weight of cut pieces (WT). will remain an unchanging dark color as the
microscope’s stage is rotated. Anisotropic
178 6 Petrography for Archaeological Geology

minerals, while never going completely to perpendicular light beams reflect off the speci-
extinction, will lighten and darken four times as men to create a lateral wave shift/displacement in
the stage is rotated through 360 . regions where surface relief exists. If the surface
The author has used reflected light microscopy is completely flat, nothing is observed. However,
extensively in the study of archaeomaterials— if there is, for example, a small difference
lithics, ceramics, metals, and slags, along with between the two wave fronts, one of the beams
fossiliferous materials as well. A transmitted must travel a path that is longer and is assigned
light microscope will typically be of little use to this path difference. Upon recombination, an
anyone wanting to examine the structure of metal- “interference (phase) contrast” is observed, giv-
lic samples or the surface of ceramics. Most ing, in many instances, a “three-dimensional
research microscopes can be fitted with both trans- effect” to the surface. It has to be kept in mind
mitted and reflected systems. In a reflected sys- that this is an optical effect and the relief does not
tem, light must be directed onto the surface and necessarily resemble the actual condition of a
eventually returned to the microscope objective surface. Petrographic usage of reflected light
by either specular or diffused reflection. Light microscopy rarely requires the resolution that
reflected from the surface of the specimen Nomarski optics provide, but it is there if needed.
re-enters the objective and passes into the binocu- Lithic and ceramic studies routinely use
lar head where it is directed either to the eyepieces reflected light microscopy as well as transmitted
or to a port for photomicrography. microscopy. The latter cannot examine unpre-
One of the most powerful techniques for pared surfaces of an artifact or item directly as
introducing contrast into reflected light micros- can reflected microscopes. So-called stereomi-
copy is differential interference phase contrast croscopes are most commonly used in petro-
optics. These special optics allow the visualiza- graphic studies. These microscopes do not have
tion of minute elevation differences on surfaces. the magnification capabilities of the petrographic
To produce interference, a birefringent prism microscope, but, in most cases, this magnifica-
(also known as a Wollaston or Nomarski prism, tion is not needed for evaluation of surfaces on
depending upon design) is introduced into the tools or fabric or temper in pottery. Pioneering
light path just above the objective, and a studies by Semenov (1964), Tringham
polarizer is installed in the vertical illuminator et al. (1974), and Odell (1975, 2001) in micro-
(similar to polarized light). The prism splits the scopic use were relied exclusively on reflected
polarized light into two orthogonal polarized light microscopy.
beams on their way to the specimen. These
Clay Minerals and Ceramics

7.1 Introduction contrast to earlier archaeological theories linking

pottery to plant and animal domestication or “the
The study of ceramics and the clays used to Neolithic Revolution” (Childe 1948), indicate
produce ancient earthenwares fill detailed and that pottery was a hunter-gatherer innovation
discursive publications in archaeology. that first emerged in East Asia between 20,000
Geoarchaeology chiefly concerns itself with and 12,000 years before the present, toward the
those aspects of ceramics and clays wherein it end of the Late Pleistocene epoch (Craig
can make the most pertinent contributions in et al. 2013).
geology, petrology, and mineralogy, to name Pottery represents a fundamental shift in
the most obvious areas. As an aluminosilicate human dietary history, conclusively
mineral, clay is practically ubiquitous. As a demonstrating that hunter-gatherers, not early
marine sediment, it floors the deepest portions farmers, in East Asia used pottery for some
of the world’s oceans (Thomsen 1877). Because 10,000 years before they became sedentary and
of its geological commonplace, clay is the began cultivating plants. The age for pottery
potter’s equivalent of ore for the metallurgist. production at Xianrendong of ~20,000 years
There is no one explanation for the first trans- ago coincides with the Last Glacial Maximum,
formation of clay to ceramic. Suffice to say, the MIS 2, when there was a decrease in the produc-
association of fire and clay, in some happen- tivity of regional food resources. Used for
stance setting—a hearth fragment, the accidently cooking, pottery allowed the preparation of
fired clay figurine, etc.—inspired early humans energy-rich starch foods and meat. Scorch and
to systemize the manufactured fired clay objects soot marks on shard exterior surfaces indicate
and, ultimately, pottery (Vandiver et al. 1989). that Xianrendong pottery was used in cooking,
Pottery’s antiquity has doubled in the past decade but complete physiochemical studies have not
with discoveries in China pushing back pottery’s been reported yet. However, lipid residues on
invention to the late Pleistocene ~18–17,000 cal- 15,000–11,800 BP pots (the Incipient Jōmon
endar years before the present (BP) first, at period) are unequivocally derived from
Yuchanyan Cave (Hunan, China) (Boaretto processing and cooking freshwater and marine
et al. 2009) then over 20,000 years ago with organisms (Brown 2000; Evershed 2008; Craig
discovery of a pottery at Xianrendong Cave et al. 2013).
with round bottoms and stripe-marked designs, Rice, writing in 1999, states “that objects of
tempered with course quartzite grains (Keally unfired and low-fired clay were created as part of
et al. 2004; Wu et al. 2012). These findings, in early “prestige technologies” of material

# Springer-Verlag Berlin Heidelberg 2016 179

E. Garrison, Techniques in Archaeological Geology, Natural Science in Archaeology,
DOI 10.1007/978-3-319-30232-4_7
180 7 Clay Minerals and Ceramics

representations beginning in the Upper Paleo- interest so that most archaeologists and potters
lithic and are part of an early “software horizon.” have a different idea of what constitutes clay than
Clay began to be more widely manipulated by do geologists, pedologists, or agronomists. To
non-sedentary, complex hunter-gatherers in the the potter, clay is any earthy aggregate that
very Late Pleistocene and early Holocene in (a) when pulverized and mixed with water
areas of resource abundance, especially in tropi- exhibits plasticity, (b) becomes rigid on drying,
cal/subtropical coastal/riverine zones. . .” The and (c) develops hardness and strength when
reliance on archaeological discovery of ancient heated to a sufficiently high temperature.
pottery, its age determination, and chemical anal- For the geologist, clay is a fine-grained
ysis seems to leave little room for weathering product of silicate minerals. The
geoarchaeology if it were not for the familiar structural breakdown of minerals to form clay is
names associated with these discoveries such as caused by natural geochemical reactions induced
Bar-Josef and Goldberg (Xiaohong, et al. 2012). by atmospheric and meteorological agents
In the geoarchaeological analyses at Yuchanyan including water, vapor, and gasses rich in oxy-
and Xianrendong, mineralogical and micromor- gen, carbon dioxide (CO2), and sulfur oxide
phological analyses of the sediments both indi- forms (SOx), as well as by various biological
cate that ash calcite was a major component of agents, such as lichens and anaerobic bacteria.
almost all samples, implying that they were pro- Depending on the parent rock composition, the
duced mainly during periods of human end result of weathering can be a variable mix-
occupations. Another unusual anthropogenic ture of clay, quartz, hydrous ferric and aluminum
activity is indicated by the clay-rich sediment oxides, resistant detrital minerals, and micaceous
formed into lenticular bands that must have material such as vermiculite. Texturally, clay is
been brought into the cave by humans and func- characterized by extremely fine particle size,
tioned as prepared surfaces (supra). Understand- below 2 μm in diameter, and with highly variable
ing the geologic—sedimentologic and chemical and physical properties.
mineralogical—context is crucial for a compre- As noted, clays can form either in place as
hensive understanding of clays and pottery. “primary” or residual weathering products gen-
erally overlying their source rocks, or as “sec-
ondary,” sedimentary clays that formed
7.2 Clays elsewhere and were transported by water from
their source, or as clay-size particles formed
Clays as noted are geologically ubiquitous from sediments, transported and refined by
minerals. This one fact may be, in large part, a water, and redeposited as sediment. In the state
major factor in the almost universal and indepen- of Georgia, one of the world’s principal kaolin
dent occurrence of pottery in the world. In a resources, the kaolin deposits of Cretaceous age
geologic sense, all clays are “secondary” are primary and higher grade; the tertiary
minerals. They are formed by the breakdown of deposits are secondary with more impurities.
other minerals, such as feldspars or the diage- The secondary kaolin formed by weathering of
netic solution of carbonate rocks such as crystalline rocks on an upland surface then was
limestones containing clayey impurities, which, transported away from their source; refined to
being insoluble, are deposited as clay. Clays are finer-grain sizes and picked-up impurities, espe-
formed by the disintegration and solution of cially organic matter; and redeposited where they
shale as well. In this sense, clays are all second- are found today. Other clays such as montmoril-
ary, but their deposits are termed “primary” lonite and bentonite are also found as primary or
(authigenic) or “secondary” (allogenic). Thus, secondary deposits. Since secondary clays are
“clay” is a term applied to a variety of earth carried by water, they tend to pick up many
materials of different origins and composition. impurities which affect the manner in which the
The actual definition depends on the field of clay may be fashioned, tempered, and fired.
7.2 Clays 181

Because of abrasion and alteration during trans- Table 7.1 Climate and weathering of silicate precursor
port, the particle sizes tend to be finer than in minerals
primary clays. Their fine-grain size and decayed Climate Mineral Clay(s)
organic matter tend to make them generally more Arid Biotite Smectite
plastic than primary clays. Plagioclase Kaolinite
Clay has the general formula of an Temperate Biotite Vermiculite
“alumina–silicate–hydroxyl” mineral. Clays are Plagioclase Smectite,
fine particulates that are >2 μm (microns
Tropical Biotite Kaolinite
(1 μm ¼ 1/1,000th mm) or less of triclinic K-spar
crystals (ex. kaolin). Clays are sheetlike Plagioclase
structures. Clay minerals are characterized by Humid Biotite Kaolinite
substitutional cations such as Si4+, Al3+, Fe3+, tropical K-spar, Kaolinite,
Mg2+, K+, and Na+. plagioclase gibbsite
Cations are common in the interlayers and not
in the silica-alumina structural units. The “2:1; The structure of kaolinites is 1:1; the structure
2:1:2 clay minerals” are more cationic than the of illites and smectites is 2:1.
1:1 clay minerals, e.g., kaolin. Kaolinites have
few if any cations in the interlayer. Structurally,
Fe3+, Al3+ cannot form tetrahedra, while Mg2+
can form octohedra along with Fe, Al. Illites
have one-sixth of Si4+ replaced with Al3+ with a
charge deficiency of +1 usually balanced by
potassium K+. Montmorillonites have many
cations in the interlayer, whereas Na+, Ca2+ are
common in smectites.
Rocks with high glass weather more rapidly
than those with lower glass (e.g., tuffs). Fine-
grained rocks weather more slowly than coarse-
grained rocks. Biotite-bearing rocks (granites)
weather more rapidly than those without (basalt). The interlayer, in illites, is filled with cations
Sedimentary rocks weather according to grain and, therefore, makes illites more rigid than
size and to some degree, clay content. Fine- montmorillonites. Kaolinites are the most rigid
grained sedimentary rocks weather more slowly of all clays. The property of rigidity is related to
as a rule. Cements in sedimentary rocks impact plasticity, an aspect very important to the potter.
weathering—silica cements are typically stron-
ger than carbonate (micrite/sparite) cements.
A “rock stability scale” is as follows: quartz- 7.2.1 Clays in Paleosols
ite, chert–granite, basalt–sandstone/siltstone–-
dolomite/limestone (bioclastic, silica-cemented While the foregoing suggests that most primary
“limestone” is an exception). Climate/ clays originate from silicate rocks, a large num-
weathering of the same lithology can produce ber of clay-rich or “terra rossa” soils do not. In
differing clays (Table 7.1; Fig. 7.1). many instances, clay-rich soil such as ultisols
Clay “sheets” are composed of silica sheets— (USDA) forms on carbonate bedrock (Driese
silica units are tetrahedral (four-sided—shown et al. 2011). Terra rossa soils across the globe
below). are characterized by red and yellow colors on the
Alumina sheets—alumina units are octahedral Munsell scale (cf. Chap. 3). Likewise, they have
(eight sided). clay-rich B horizons (supra). On paleosurfaces
182 7 Clay Minerals and Ceramics

Fig. 7.1 Weathering of silicate minerals (After Horgen 2000). Granites gneisses of the Appalachians produce kaolinite
with gabbros yielding smectite and olivine-rich rocks yielding montmorillonite via the mechanisms illustrated below
7.2 Clays 183

where clay-rich horizons indicate a paleosol or Studies of Southeastern United States

Paleudult, for example, that may have formed Paleudults in the eastern Tennessee Valley-and-
from limestone or dolomite bedrock and are Ridge Physiographic Province suggest rather
indicative of a temperate to tropical complex origins for these ancient soils (Driese
paleoclimate. Additionally, these ultisols tend to et al. supra) where the bedrock is undisputedly a
indicate geologically old land surfaces but the carbonate lithology. The clay mineralogic
kandic (argillic) B horizons can form in as little analyses support the formation of pedogenic kao-
as 4,000 years due to the high rate of clay forma- linite, halloysite, and hydroxyl-interlayer
tion and illuviation (Layzell et al. 2012). minerals along with illites (see above) derived
In mixed (ionic) charge sediments, constant- exclusively from the carbonate bedrock. With
charge clay minerals (aluminosilicates with a these ancient soils, a hypothesis of the so-called
permanent structural charge) interact with oxide top-down pedogenesis due to in situ weathering
minerals of Fe and Al, which are negatively and was rejected for a more polygenetic origin that
positively charged, respectively, at the normal involved alluviation and colluviation as
pH of groundwater. Iron oxides or contributing influxes of detrital minerals such as
oxyhydroxides are known to form close quartz and, of course, feldspars.
associations with clay minerals, forming discon-
tinuous layers or small aggregates on the nega-
tively charged surfaces of clays (Bertsch and
7.2.2 Clay Deposits
Seaman 1999; Bunn 2002; Hong and Xiao-Nian
1992; Seaman et al. 1997). Some studies have
Most clay deposits are found with an admixture
suggested that the “binding agents” providing
of silt, sand, and other impurities including
physical stability in soils are mainly poorly crys-
organic material, micas, and iron hydroxides.
talline or amorphous phases of Fe, Si, or Al
The impurities may constitute over 50 %, so
(Arias et al. 1995; Swartz et al. 1997) and that
such deposits must be refined before they can
crystalline phases are more likely to form dis-
be used for pottery making. The differentiation
crete aggregates.
among clay, silt, and sand is by particle size, but
184 7 Clay Minerals and Ceramics

unfortunately, there is no agreement on where clay. The structural properties of these groups
the size boundaries should be drawn. In most govern many technological choices by potters.
widely accepted granulometry scales used by
agronomists, engineers, and geologists, clay is 1. As indicated above, the structural unit for kao-
smaller than either 0.002 or 0.004 (Wentworth linite is a l:1 layer, consisting of one tetrahe-
Scale) mm in diameter (2–4 μm), silt up to dral sheet with one octahedral sheet (Fig. 7.2).
0.05 mm, sand 0.05–2 mm, and gravel is greater The space between two successive 1: 1 layers
than 2 mm. The name assigned to mixtures of the represents an interlayer which will be devoid
first three depends on the relative amount of each of any chemical elements if the layers are
component present. Soil scientists call clay any electrostatically neutral, i.e., with no excess
mixture with more than about 40 % clay, and valences. Since this is true in most kaolins,
mixtures with much silt and sand are called the interlayers remain relatively free of cations
loam. Potters don’t classify clays on mineralogi- or water. The simple rigid structure of kaolin
cal basis. Rather the terminology one sees used does not allow for much substitution by exotic
among ceramicists is: elements (alkali elements lower melting tem-
perature). In addition, it has the highest Al:Si
China clay: hydrated kaolin ratio of the clay minerals (higher Al means
Pipe clay: kaolin with large % SiO2 higher melting temperature) and is least able
Potter’s clay: smectites, illites, bentonite, e.g., to absorb water (lowest plasticity). The use of
montmorillonites—not kaolin group pure kaolin clays in ceramics is thus restricted
Brick clay: clay with sand and Fe in large % to the most advanced ceramic technologies.
Fire clay: clay with no lime, iron, or alkaline 2. The smectite structure unit is a 2:1 layer in
earths; very refractory, resists fusion which two tetrahedral sheets sandwich an
Ball clays: white or gray with large alumina— octahedral sheet. The position of the two tet-
silicate %; e.g., kaolinite principal clay min- rahedral sheets is inverted so that all apical
eral (35 lb. “balls,” British term) oxygens point to the octahedral sheet where
they are shared by both sheets. Most 2:1 clay
Virtually any clay can be used for making minerals have an excess layer negative charge
pottery, depending upon the skill of the potter. produced by unbalanced ionic substitutions,
The main clay mineral groups include (1) kaolin- such as Mg2+ for Al3+ or OH for O. This is
ite, which has a relatively fixed structure, (2) a neutralized in the interlayer by various
group of expandable clays called smectites, cations, mainly Na and Mg, hydrated cations,
whose most common member is montmorillon- hydroxyl groups, and water. Abundant substi-
ite, and (3) the illite group with a structure simi- tution for Al and the presence of interlayer
lar to smectite but is not expandable. The sum of cations makes smectites less refractory than
a layer plus an interlayer is a structure unit in other clays. They are also expandable because

Fig. 7.2 The silica and

alumina tetrahedral/
octahedral units
7.3 Ceramics and “Ceramics” 185

their charge deficiency is in the interior alu- of different clay and detrital minerals can make
mina layer. At lower firing temperatures than drying and firing behavior hard to predict. This
other clays, about 600  C major explains why the choice of clay and the prepara-
dehydroxylation occurs which causes irre- tion recipe can be the most conservative aspects
versible structural changes. Smectites also of ceramic production by traditional potters.
have higher shrinkage in drying and firing Because chemical compositions of clays vary
than other clays. widely in both elements present, especially trace
3. The illite group of clays have structures simi- elements at the ppm level, and water content, a
lar to micas as well as to smectites. In the structural determination by XRD analysis is
illites, about one-sixth of the Si4+ is replaced more important than a chemical analysis to iden-
by Al3+ leading to a charge deficiency that is tify individual clay minerals. This is especially
balanced chiefly by K+ but also by Ca2+/Mg2+ important for characterization of provenance and
and H+. The charge deficiency in the illites is technological studies.
primarily in the outer silica layers and there-
fore closer to the surface than in the smectites.
The excess negative charges on the illite 7.3 Ceramics and “Ceramics”
interlayers are satisfied by cations which
accounts for a more rigid structure. Because In the field of archaeological ceramics, there is
of their fine particle size and low shrinkage extensive variety in the nature of the specific
which helps produce a natural luster through wares or types of pottery manufactured at any
minimal mechanical treatment, illite clays are one time in any one place throughout antiquity.
desirable as pottery slips. To the earth scientist, not versed in the complex-
4. Other less important clay minerals are vermic- ity of archaeological ceramics, or the variety of
ulite and the chlorite group. Vermiculite, way archaeologists look at ceramics, confusion is
so-named because it looks like small worms not a surprising first reaction. In large part,
(Latin: vermiculus) when heated, is related to archaeologists tend to rely more on stylistic/mor-
the smectites. Vermiculites also have an phological variation in ceramic wares than on
expanding lattice but not as great as in the petrographic parameters. A key concept for
smectites. The chlorite group are light green, archaeological studies, not just for ceramics,
mixed layer minerals common in low-grade although the concept was first used in ceramic
metamorphic rocks and sediments. They can analysis is that of “type.” A type is a category of
form in a variety of ways: as alteration of archaeological artifacts defined by a consistent
ferromagnesian silicate minerals or clustering of attributes such that an item, such as
authegenically in sediments. They are com- a pottery vessel can be considered a member of a
monly mixed in with other clay minerals and class/category (Clarke 1968; Thomas 1998a, b).
are thus hard to identify. Thus, the Variation or consistency in types is used to eval-
two-1ayered kaolinite and the three-1ayered uate the geographical extent and chronological
smectite clays, including montmorillonite, are persistence of a particular type, which in turn, is
expandable, but the three-layered illites often used as a proxy for cultural variability or
are not. persistence. The type-variety of ceramic typol-
ogy uses the type concept for vessels and groups
The exact chemical compositions of most clay successively finer categories into “varieties”
deposits are quite variable, especially between based on some combination of variables, such
beds; their physical properties vary as well, cre- as paste, decorative treatment, and surface finish.
ating serious problems for the potter. The most common form of type is that of the
Substitutions of water and different cations in “descriptive type.” It inheres no chronological
clay structures, as well as indeterminate mixtures nor functional dimension although from the
186 7 Clay Minerals and Ceramics

adoption of ceramic typology, these categories and, at the second, over 20,000 shards were
have been so used. For instance, if a type persists found (O’Keene 2004; Keene and Garrison
over a relatively long period of time the term 2013). No archaeologist, nor geoarchaeologist,
“ceramic tradition” is often applied (Willey and examines this amount of pottery as a rule with
Phillips 1958). Likewise the term “ceramic hori- rare exceptions. Bishop (1994) reports on a mas-
zon” is used to describe the “rapid” appearance sive neutron activation analysis (NAA) study
of a type, in a short time span, across a relatively done on over 12,000 Mayan pottery samples.
large area or region (supra). This system of clas- More and more pottery analysis projects are to
sification is more commonly found in the study be mounted on a similar scale.
of prehistoric earthenwares, whereas the term At the Dove Creek site (Colorado), for exam-
“ware” is more commonly used in discussions ple, an Ancestral Puebloan/Anasazi Period site
of historic ceramics (Majewski and O’Brein (ca. AD 900–1,050), Till (2006) in the wake of a
1987). topographic and geophysical mapping study,
Rarely does the field archaeologist look much evaluated 8,976 shards. Till points out the use
beyond the form/shape, color, or decoration of a of 4 “working types”—categories used for anal-
ceramic artifact except, perhaps, to remark on the ysis but not formally recognized as “types” per
type of temper used (in earthenwares, mostly). se—out of a suite of 31 types. Ancestral Pueblo
No archaeologist, working in Neolithic-Post- farmers settled the central Mesa Verde Region of
Neolithic or equivalent periods (Woodland; southwestern Colorado in the seventh century
Pre-Classic, etc.), can ignore the ceramic artifact AD, the Basketmaker III Period (Diederichs
as (a) most of these archaeological cultures made et al. 2013). The Basketmaker III Period
and used pottery and (b) by definition, the “Neo- (AD 600–750) was part of an important cultural
lithic” is often characterized by the presence of shift in the history of the American Southwest,
pottery. As we noted in the introduction of this when a new technological package—including
chapter, early pottery appears in the Mesolithic “the maize suite”—enables direct precipitation
Period in Asia, specifically in the early Jōmon or “dry” farming, due in part to the recognition
Period of Japan, ca 10,500 BC (Ikawa-Smith of the deeper soils’ potential (Fadem et al. 2014).
1976; Rice 1999). The recent discoveries in These innovations—including pottery—enabled
Jiangxi Province, China, Xianrendong Cave, the rapid expansion of Puebloan populations,
have pushed back its invention into the late thereby representing a material culture
Paleolithic ca. 20,000 years ago (Wu et al. 2012). (“ceramic”) horizon in the archaeological record
It was independently invented and used in all (Lipe 2015; Diedreichs, supra). This change is an
parts of the world with the exceptions of American version of the “Neolithic Revolution”
Australia, the far north of North America, and (Lubbock 1865; Childe 1948), which
Eurasia. It is as ubiquitous an artifact in late emphasized the introduction and use of pottery.
Holocene archaeological cultures as those made The ceramic “sample,” nonetheless, in many
of lithic materials. archaeological instances can, therefore, be a real-
The ubiquity of pottery as an artifact can be as ity that challenges analyses. Another aspect,
much difficulty as it is an advantage for archae- beyond sheer numbers, is the salient difference,
ology. The sheer amount of pottery, mostly often overlooked, that exists between pottery and
recovered as shards, must be seen to be believed. other artifactual remains such as lithics. That
Pottery is recovered in metric ton amounts on difference lies in the phenomenological nature
many prehistoric sites. The author participated of pottery itself. Ceramic assemblages, unlike
on two excavations, one a Late Bronze Age vil- lithic materials, are pieces of the “whole cloth”
lage site in western Switzerland and another, or the vessel of which they were a part. Lithic
much smaller fourteenth- to fifteenth-century residue often represents all or portions of the
habitation site in coastal Georgia, where, in the chaine d’operatoire as originally defined by the
first case, over 20 tons of pottery was recovered French Paleolithic archaeologist, Leroi-Gourhan
7.3 Ceramics and “Ceramics” 187

(1993). This is to say a reduction flake is just that firing, and (4) other esthetic properties, including
while the biface from which it was removed is color, etc.
the finished artifact temporally and spatially sep-
arate from that fragment of its manufacture. The Plasticity
shard is basically a smaller “vessel” in and of The addition of water to raw clay will produce a
itself. It is not a fragment of a manufacturing plastic mixture which can readily be shaped
process. It can yield information of that process without rupture. Clay exhibits plastic properties
but it is the product. This phenomenological for primary reasons: the clay particles themselves
reality gives greater credence to descriptive anal- are platelike, a reflection of their crystalline
ysis of these singular fragments of the whole. structure; the small size of particles, clay sized,
They are, in a real sense, the “whole.” This is below 0.002 mm in diameter; and the presence
gives the shard an analytical standing that belies of water, called “water of formation” between
its humble nature. A ceramic analysis of a single the clay particles. Water acts as a lubricant and
shard can be much like a mineralogical study of allows the clay particles to slide across each
fragment of a stone outcrop or facies in that the other rather than rupture when a shear force is
same relationships here, for instance, use of the applied.
Delasse Method of calculating volume The development of plasticity in clay is criti-
percentages of minerals in a specimen (cf. cal to the manufacture of a ceramic, and
Chap. 5). What is generally technologically true depending on the composition of the clay, plas-
for the shard is true for the pot as a whole. ticity will develop over a narrow range of water
Archaeological analytical protocols for pre- content. Dry clay is nonplastic, and as water is
historic ceramics have led to many important added, the clay will become plastic, and finally,
key insights. One of the first and most famous with too much water, it will become too fluid to
uses of a systematic examination of ceramic retain its shape when formed (Table 7.2). Kao-
types was that of Sir Flinders Petrie (1904) in linite has a non-expanding two-layer structure
his use and expansion of ceramic seriation/typo- and forms relatively large particles, 0.3–2 μm in
logical sequence analysis for chronological diameter; smectites have an expandable three-
purposes. For our purposes, the Earth science layer structure with a finer-grain size. Thus, kao-
study of ceramics, specifically petrography and linite has much lower plasticity and undergoes
geochemical, provides additional ways, in the much less shrinkage than does smectite upon
rubric of epistemology, giving independent evi- drying. If a potter has access to clays which are
dence, for suppositions and hypotheses about the highly plastic, this property can be modified by
manufacture, technological evolution, and use of the addition of nonplastic materials (soil, less
ceramics in the past (Hayden 1995, 1998; Tite plastic clay, crushed rock or pottery, or organic
1999; Vandiver 2001). material) as temper.
Of the many factors that control the plasticity
of clays, the most important is the shape and size
of the clay particles themselves. Clay is very fine
7.3.1 Ceramic Properties of Clay
grained and crystallographically platelike with a
thickness to diameter ratio of about 1:12. When
In the making of ceramics, it is crucial that the
potter know how the clay body will behave dur-
Table 7.2 Consistency water %
ing the different stages of fabrication: aging the
Dry 5–18
clay, forming the matrix, finishing the ceramic,
Stiff 10–20
drying, and finally firing. The properties that
Stiff-plastic 12–25
determine the usefulness and problems for each
Plastic 15–30
clay raw material are (l) plasticity in forming, Oversoft 15–35
(2) shrinkage by drying, (3) thermal during Liquid 20–50
188 7 Clay Minerals and Ceramics

water is added to clay, it acts as a lubricant, Paste and Temper A ceramic vessel or
surrounding the clay platelets and allowing them object—figurine, statue, bead, etc.—is composed
to glide over each other. Surface tension forces of clay and its impurities, generally accidental or
acting on the water and the clay bind them weakly accessory minerals and, in the case of low-fired
together and allow a molded object to retain its earthenwares, some form of tempering agent
shape. When excess water is added, more than the which is often mineralogical in nature. There
minimum needed to maintain surface tension, the are a few commonly used terms to denote paste
water and the mass will start to flow. and temper—matrix, fabric, body—are fre-
Other important factors which affect plasticity quently encountered. Stoltman (1991) considers
are the nature of adsorbed ions and clay particle paste to be “The aggregate of natural materials,
size. Large monovalent ions, including Na+, K+, ie, clays and larger mineral inclusions, to which
and ammonia, NH, diminish plasticity; smaller temper is later added....”. Kamilli and
di- and trivalent ions including Mg2+, Fe2+ and Steinberg’s definition was paste “consists of a
Fe3+, and Ca2+ enhance plasticity. The large coarse mineral fraction, a fine-grained matrix,
monovalent ions, too large to fit into the clay and perhaps some plant material” (Kamilli and
lattice, tend to disrupt the nonliquid water struc- Sternberg 1985). Body, as defined by Stoltman
ture promoting disorder. The smaller, more (supra), is the “bulk composition of a ceramic
highly charged ions easily fit into the space avail- vessel including clays, larger natural mineral
able in the lattice and thus do not exercise any inclusion in the silt, sand and gravel size ranges,
disordering effect. and temper (emphasis added).” Body is equiva-
lent to the term fabric (Shepard 1936, 1954). It is
temper that differentiates body/fabric from paste. Shrinkage Temper, by extension, is that which is not natu-
After the ceramic has been formed, the wet clay rally occurring within the paste of a ceramic—
must be dried. Water will be lost in two phases. e.g., it has been added by the potter.
During the first phase, the water of formation will Temper varies widely, in early ceramics,
evaporate freely and the ceramic will contract. imaging from organic materials such as plant
As the water is removed from around the clay material, bone, and shell; mineral temper such
particles, the particles will draw closer together as rock of many varieties; sand or volcanic ash or
until they eventually touch each other and start to tephra; to grog or ceramic fragments reused to
adhere. The ceramic will contract no farther at temper subsequent vessels. Because temper,
this stage. The next water loss takes place more other than that of organic origin, is similar to
slowly, and it will come from the pore spaces naturally occurring minerals and rock, some
between the clay particles. authors simply use size as a discriminating factor
Internal stresses are set up in the ceramic if it (Lombard 1987; Ferring and Perttula 1987). The
loses water too fast at this first stage. identification of temper, because of its similarity
The rate of evaporation is critical to prevent to the materials of the paste, is oftentimes diffi-
the clay from cracking or breaking up. Again, cult (Stoltman 1991), but the petrographer must
different clays vary greatly in their drying shrink- make the attempt wherever possible. Where the
age, that is, the reduction in size upon evaporation distinction is straightforward, the temper
of water. Clays also vary in their dry strength, that becomes a useful tool in distinguishing ceramic
is, the resistance to breaking when a force is types. Even in cases, such as were there the
applied after the water of formation has been coarser added material may be differentiated
removed (Table 7.2). This property is important from a fine-grained matrix by textural analyses.
to potters who scrape and smooth vessel walls at Potters use temper to help the clay dry and fire
this stage, perhaps prior to applying a slip. uniformly:
7.3 Ceramics and “Ceramics” 189

• Grog temper (basically ground up pottery) Temper morphology and size were illustrated
• Shell (after it has been fired) by (1984) Moundville (Alabama) prehistoric
• Sand/stone—like marble ceramics of the Mississippian Period (ca. AD
• Tuff 1,000–1,500). Temper gradually changed from
• Organic (Spanish moss, grasses, other fibrous coarse quartz then to fine quartz, grog, and finally
plants) coarse shell. Shell-temper became one principal
characteristic of evolved Mississippian ceramics.
Studies of temper and actualistic studies have Steponaitis showed that the shell resisted thermal
yielded significant insights into the function of this expansion better than other tempers and required
amendment to clays in terms of “performance lower firing temperatures (~700 oC.) or the shell
characteristics” which the ability of the vessels to would decompose to lime. Additionally, rehydra-
retain their contents and to survive impact without tion of the decomposed shell would cause failure
cracking; heating effectiveness and the ability to in the vessels as well (supra).
survive rapid changes in temperature without
cracking, in the case of cooking vessels; and Firing and Thermal Properties
cooling effectiveness, in the case of water storage Before firing any ceramic, the first stage—dry-
vessels (Tite 1999; Schiffer 1988, 1990; Skibo ing—is considered “air drying” and is generally
et al. 1989; Skibo and Schiffer 1995; Kilikoglou carried out <150  C. During drying, the mechan-
et al. 1995, 1998.) One area, other than temper per ically admixed water, which is present as films
se, was the importance of surface finish in surrounding clay particles, is lost. The clay mix-
preventing water loss. Resin, as we shall see with ture shrinks and loses its plasticity. As the clay
a Navaho example, applied to the surface, notably particles come closer together, they may come
the interior, minimizes water loss (Schiffer 1990). into contact with each other. Stresses set up at
Interestingly, simply finger-smoothing of the inte- this stage may later lead to structural damage to
rior surface accomplishes the same effect (supra). the body. During this period of drying, the body
But it is the role of temper, type density, and size will undergo its maximum shrinkage and exhibit
that most determines the strength and “toughness” its maximum water loss.
of the vessel (Tite 1999. Water first evaporates from the surface of the
Ceramic “toughness” is difficult to define. body. Then by capillary action, water rises from
Resistance to fracture is primarily the best the interior to the surface and in turn evaporates.
definition. To ease the process of evaporation, nonplastic
Repeated exposure to heat and subsequent cool fillers, called temper, earlier added to the clay
downs test the ability of a pottery vessel to resist paste, help provide pathways for water to escape.
fracturing. Prolonged exposure to cooking fires Other materials or characteristics may be present
increased with the specific foods such as the exam- or added to help bind the clay together and
ple of beans in the Basketmaker III Period. strengthen it during drying and shrinkage. This
Kilikoglou et al. (1995, 1998) showed ceramic includes fine rather than coarse particle size clay,
toughness increased within an increasing number the presence of Na1+, and added organic
of quartz inclusions as well as an increase in their materials such as flour or gums of various trees.
particle size. Feathers (1989) and West (1992) All these will help bind the clay particles
further showed that the type and morphology of together and help prevent uneven drying which
the temper is particularly relevant. Platy or fibrous may cause cracking or warping.
temper, such as mica or shell, greatly increases Much early pottery was fired in open fires, but
toughness when compared to angular temper such with the passage of time, more and more firing
as limestone, marble, quartz, or grog (supra). has been in periodic kilns. The process of firing
This morphological difference results in both takes place in distinct phases: setting up the kiln,
resistance to crack propagation and thermal firing the air-dried clay bodies, cooling down the
shock fracture. kiln, and finally, withdrawing the pottery. The
190 7 Clay Minerals and Ceramics

Fig. 7.3 Closed kiln

design of such kilns varies widely through time different results in the production of
and place. Most are so-called “open-flame” kilns earthenwares (Gosselain 1992; Kingery and
where the heat is used directly. The fire box is in Vandiver 1988). In pit kilns, maximum
the fire pit directly below the ware and temperatures (~600–800 oC) are reached within
surrounded by flames (Fig. 7.3). The combustion 20–30 min. Because of the rapid heating rate,
gases generated in the fire box move directly coarse pottery fared better.
upward into the firing chamber by convection. Thin-walled fine ware survived equally as
Temperature control, both in absolute degrees well because of its ability to allow steam to
Celsius and in the uniformity of heating through- escape the vessel (Tite 1999). Closed kilns
out the firing chamber of open flame kilns, is very greatly enhanced the control and quality of
poor. The temperature throughout the kiln will earthenwares (supra).
vary widely, with both vertical and horizontal Firing the clay body to the hardness of a
gradients. Vertical gradients are controlled by ceramic body is the next critical stage in manu-
the distance to the bottom of the kiln where it is facture. The ceramic then acquires its hardness,
the hottest and horizontal gradients develop by toughness, and durability. Original mineral and
the flow of combustion gasses toward the vent. clay structures are transformed, and chemical
Some areas within the kiln will have a reducing compositions change as the body loses its basi-
atmosphere, especially nearer the bottom where cally sedimentary condition becomes its meta-
highly reducing combustion gasses enter the fir- morphic equivalent. The principal changes that
ing chamber, and other areas nearer the vent will take place are by (a) oxidation or reduction,
be oxidizing, where a strong draft exists, oxygen (b) loss of volatiles and dehydration, principally
is readily available, and the combustion gasses loss of hydroxyl ion (OH) from the clay
themselves have become oxidized, e.g., CO2 has structures, and (c) creation of higher temperature
become CO2—indeed, temperature gradients phases and finally, at very high temperatures of
exist within single vessels, varying according to firing, vitrification.
size and wall thickness. As we will detail in the following section, clay
Time of firing and peak temperature are two minerals undergo important structural and chem-
principal factors, beyond temper, in determining ical changes, due primarily to dehydroxylation,
the strength of pottery vessels. Pit kiln firings when they are fired at temperatures over 600  C,
were shown by Shepard (1956), and kiln firings and particularly at higher temperatures of
have been demonstrated to produce very ~1,200  C. These changes are important to
7.3 Ceramics and “Ceramics” 191

Fig. 7.4 Phase changes in


convert the clay into a useful ceramic. High 6. The atmosphere in the kiln, whether reducing
temperatures dislodge ions from original or oxidizing. Color is strongly dependent on
positions held in the clay structure into more this: ferrous iron can cause a green color,
favorable sites, thus converting the clay into ferric, and unoxidized organic material black.
chemically stable and structurally stronger mate-
rial. Chemical changes that take place during Tite et al. (1982) in a study of Greek Red and
firing of clay are shown in Fig. 7.4. The exact Black Figure pottery demonstrated the use of the
changes that take place for each stage of firing three-stage oxidizing-reducing-oxidizing firing
depend on many variables, including those that cycle to produce the reduced iron-oxide pigment
are material dependent: Studies of firing temperatures have proven to
be capable of ascertaining both the ranges of
l. The chemical, mineralogical, and structural temperatures and firing times (Tite and Maniatis
composition of the clay and the nature of 1975; Tite 1995; Wagner et al. 1986; 1998;
included materials. Kingery and Friermanm 1974; Kingery 1992a,
2. The wall width of the ceramic body. b).
3. The nonplastic temper added by the potter. At
450–500  C, water that is chemically com- Oxidation and Reduction
bined in the clay structure is lost. The clay If materials such as dung, straw, etc. are utilized
crystalline structure itself will break down in the making of the ceramic body, then, when
above 800  C and be largely converted to heated to 200–500  C in an oxidizing atmo-
new mineral phases. Above 1,800  C, the sphere, such organics will oxidize and be given
ceramic body will fuse, a temperature far off as CO2, leaving vacant spaces and a porous
above the possibilities of any ancient body. If, on the other hand, the ceramic body was
technologies and those that are dependent on heated under reducing conditions, in those parts
firing conditions. of the kiln where oxygen was not available, then
4. The rate of firing, allowing gases including the organic material is only charred and a residue
those formed by the breakdown of organic of elemental carbon is left. Under these
material and carbonates to escape. Too fast a conditions, a gray-black ware will result. Com-
rate will trap gases and result in higher porosity. monly, firing is timed to allow oxidation to go to
5. The duration of firing. The longer high completion so that no organic matter is trapped in
temperatures are held, the more complete the fired clay before the stages of dehydration of
will be mineral reactions forming glass and clay and the formation of new mineral phases
high temperature mineral phases, such as and glass begin. However, much archaeological
spinels and mullite which lead to a stronger pottery with a gray core shows that oxidation
ceramic body does not always go to completion.
192 7 Clay Minerals and Ceramics

As the temperature is raised, and a good draft temperatures are maintained for a long enough
is obtained in the kiln, the clays themselves will period, high-temperature aluminum silicates will
start to oxidize. Clear colors such as red, buff, or start to form. Depending on the (1) length of time
yellow suggest an oxidizing atmosphere at firing held, (2) the temperatures attained, as well as
temperatures of about 700–900  C. If the draft is (3) other minerals present, such as micas, that
cut off preventing outside oxygen from entering, may react with the clays, a variety of higher-
the entire kiln will be filled with reducing gases. temperature mineral phases may crystallize.
The most important oxidation-reduction reaction These include, from lower to higher temperature,
affecting color is the change from red ferric feldspars, spinels, and above about 1,050  C
oxide and hematite Fe203 to black ferrous oxide mullite (3Al203• 2Si02). Mullite is a synthetic
FeO and magnetite Fe304. aluminosilicate used in the manufacture of
high-temperature and high-shock ceramics, such Loss of Volatiles and Dehydration as automobile spark plugs. Identification of such
When temperatures of firing reach about high-temperature minerals can be important
200–300  C, all mechanically bound or pore clues to a vessel’s firing history.
water still present after ambient drying is Starting about 600  C, vitrification or
vaporized and driven off. As the kiln temperature “sintering” starts in the more fusible impurities.
rises, hydroxyl ions that are chemically bound Feldspars, among the most common impurities,
within the clay lattice start to be lost. The rate of and commonly added to the clay paste as a flux
loss depends on the clay species with halloysite by modem potters, will melt: pure Na-feldspars
and kaolinite abruptly losing most hydroxyls at 1,118  C and pure K-feldspars at 1,100  C;
about 4,000  C (Fig. 7.4). Other clays, such as solid solution mixtures will melt at lower
smectites, start to lose interlayer water (held temperatures. Many potters, both ancient and
between the silicate sheets) at lower modem, recognized a variety of other materials
temperatures. They then continue to lose water that can act as fluxes, including illitic clay, iron
but at a slower rate than kaolinite. oxides, etc. When enough vitrification has taken
In addition to water, as the temperature rises place to give a moderate degree of hardness, the
to over 200  C, any carbon or organic matter in clay is considered baked.
the clay will start to oxidize and escape as CO or Full firing, or the burning stage for most clay
CO2. The rate and completeness of oxidation bodies, begins at about 850  C. Two baked
depends on the rate of temperature rise, the vessels struck together will emit a dull sound,
length of time the maximum temperature is whereas when fully burned, that is for some
held, the wall width of the pot, the clay composi- mixtures fired as low as 850  C, for most about
tion, and the nature of the paste. On average, 1,000  C, the vessels will emit a ring. To sum-
organic carbon is not completely eliminated marize, the following are the expected changes in
until temperatures of about 500  C are reached. the clay paste during firing:

150–200  C: All mechanically combined High-Temperature Phases
adsorbed water is lost from the clay. Much
and Vitrification
interlayer water will also be lost from
Dehydroxylation or the loss of OH- from the clay
molecules causes major structural and chemical
200–400  C: Organic matter begins to oxidize
changes. Heating kaolinite above about 500  C
and migrate to the surface, blackening the
results in a disordered crystalline form which is
clay body before finally being driven off as
still reversible by rehydration back to its ordered
kaolinite structure. At 900  C and higher
450–550  C: Chemically bound water is lost
temperatures, all water is lost and the crystal
from clay lattice, dehydroxylation. The loss
lattice is completely destroyed. If high firing
7.3 Ceramics and “Ceramics” 193

takes place abruptly in kaolin in the structural production of earthenwares by indigenous

transformation to metakaolin with minor artisans. Some authors suggest that primary evi-
shrinkage and a great increase in porosity. In dence for pottery manufacture is rare (Sullivan
illite and smectite, the loss is more gradual. At 1988). In defense of Sullivan, most modern
470  C, a red glow can be seen in the dark; at indigenous pottery is not made for the same set
550  C, all organic matter is oxidized. of cultural uses as in the premodern era. Much of
550–750  C: Quartz inverts at 573  C from the the modern ceramic output of the American
alpha, low-temperature form, to beta, the Southwest is for consumption as art rather than
high-temperature form, accompanied by a as utilitarian or ceremonial use. The reader
volume expansion of 2 %. Beginning at should not view this statement as in any way
600  C, micas lose structural hydroxyls from critical of the modern indigenous potter. Rather,
their lattice. the appeal to collectors and museums of modern
750–850  C: Most organic water has been suc- production has allowed traditional designs to
cessfully burned out of the clay. Kaolin and survive into the modern era. Likewise, there are
smectite begin to lose their characteristic crys- continued cultural uses for modern-day vessels
talline structure. beyond their sale outside the native culture.
850–950  C: By 870  C, CaC03 decomposes to Indeed, the genesis of this study of Navaho
CaO and loses CO2 gas. Beta quartz inverts to ceramics with a similar pronouncement by a tribal
a higher-temperature form, tridymite, but the official in Arizona on the Tohono O’Odom reser-
reaction is very sluggish and will only take vation (Robert S. Hill, Personal Communication
place if the firing temperature is held for a 2011). The official, when queried about local
relatively long period of time. potters, opined that “no one made pottery any-
950–1,100  C: The structure of most kaolin and more.” Such was not the case, happily. A brief
smectite is irreversibly lost, but some illite drive around a neighborhood produced the names
may persist. With the breakdown in clays, and vessels of more than one local potter. Unfor-
high-temperature alteration and reaction tunately, the Tohono O’Odom artisans were less
products can form, accompanied by pro- able to share cultural knowledge than those of the
nounced shrinkage. Navaho; hence, this section’s subject is of the
Above 1,000  C: Wollastonite (CaSi03), calcium latter’s ceramics and not the former.
ferrisilicates, and, in a reducing atmosphere, Pottery has “meaning” in Native American
hedenbergite (CaFeSi20 6) with reduced Fe2+ cultures. It is more than just utilitarian in nature.
iron are typical reaction products. Other high- Pottery contains a large amount of cultural infor-
temperature phases expected to form are spi- mation as noted not only by Longacre and Kramer
nel (MgAl204) and mullite (3Al203.2Si02). (supra). This section is entitled “ethnotech-
>1,100  C: Feldspars melt dissolving silica and nological study” rather than “ethnoarch-
initiating a glassy phase of the ceramic. Poros- aeological” to stress the technological aspects of
ity diminishes rapidly. Navaho pottery manufacture over the specific cul-
tural context within which it resides. The study did
ascertain that, for the Navaho, pottery connotes:

7.3.2 Ceramic Ethnotechnological – Symbols of nationality/identity.

Study: Making Navaho/Diné – Spirituality/belief systems.
Pottery – Natural world in motifs of animals, nature.
– Vessel forms indicate use as well as aesthetic.
Kramer (1985) and Longacre (1991) have
described the importance of ethnography as an Among the Navajo (Diné), the Changing
important method in understanding the Woman story is not told until after the first
194 7 Clay Minerals and Ceramics

Fig. 7.5 Forming the

“wet” pot. Note appliqué

frost. Changing Woman is the first woman in the Several techniques can be used to form pots.
Navajo belief and mother to the Hero Twins, Three of the most used by the Navaho are:
Monster Slayer and Child-Born-of-Water, who
rid the world of monsters so humans could 1. Anvil—paddle
inhabit the earth. Navajo and native potters, in 2. Draped clay—on/in basket
general, do not refer to the vessel clay as 3. Pinch—coil
“paste”—but rather it is the “clay body.” Paste
is an archaeological term. In the 2011 study, the pinch-coil method was
Manure is used for fuel particularly in the primarily used although the anvil technique was
American Southwest, whereas wood was used demonstrated. The potter kneads the clay to
in the Southeast. There are more sheep and remove air, pinching the clay – “pull-and-in”-
cows on the Navajo Reservation than useable to shape the coils. Navajo potters start with the
trees. small “pinch pot.” Coils are laid, sequentially, on
In the present study, dung was the preferred top of the pinch pot base (Fig. 7.5). “Bé-ho” or
fuel source (Fig. 7.3). All the Navaho potters biyo—an appliqué of beaded clay can be used on
observed in the 2011–2012 study dug their own the rims. These help with handling of a hot pot.
clay. The location is closely held but shared The appliqué is never closed—it can overlap but
among fellow potters in many cases. Navaho is is never closed. This allows spirits to enter and
a huge place, being nearly as large as the whole leave the vessel. This spiritual aspect extends to
of New England. Clay sources abound. One pot- pottery manufacture. The Navajo only make pot-
ter specifically sought out “brown clays” to help tery for use in eating. Therefore, they do not
ensure a firing to a high-red color. This particular make ceramic dolls. The mud dolls are unfired.
potter applied an iron-bearing slip to the drying Small pots are manufactured as containers for the
vessels and burnished them. sands (and clays) used in Navajo sand painting
Temper was usually crushed pottery or grog, ceremonies.
ground up, and worked into the clay. Sand was When firing their pots, Navajo potters partic-
used as well. Some potters, working in more ularly appreciate the “fire clouds” that manure
traditional forms, shaped the vessel, first with a fires create (Fig. 7.6). These fire clouds add inter-
corncob then burnished using a smooth stone est to the pot’s surface. Pı̃non pine tar is used to
during the drying process (Hill 2012). As noted “glaze” pot exteriors. It is applied to hot pots
above, slips can be used besides appliqué and allowing a sheen to form on the pot’s surface
piñon pitch applied after firing. In comparison, (Fig. 7.7). Santa Clara potters use the so-called
Santa Clara pueblo potters use a1:3 ratio of tuff saggar firing instead of pit kilns (Fig. 7.3). They
temper to clay for their ceramics. can fire pots in trash cans—aka a “saggar.”
7.3 Ceramics and “Ceramics” 195

Fig. 7.6 Modern Navaho

pottery (Photograph by
Robert Hill)

Fig. 7.7 Two thin sections of late prehistoric illustrates more uniform firing conditions as well as a
earthenwares, Georgia, USA. In the left, firing more uniform grain-temper sizes. The temper in both is
differences—black indicating reducing conditions and primarily quartz-dominated medium sand. Both
red-brown, oxidizing. Such differences are not unusual photomicrographs were in PPL (Courtesy of Mark
for pit kiln-fired wares. The thin section on the right Williams)

7.3.3 Geoarchaeological Study (a) The nature and the source of the raw
of Ceramics materials (clays, temper), the “slip”
(applied surface pigment) and a reconstruc-
Pottery then is the general term used here for tion of the working methods of ancient
artifacts made entirely or largely of clay and hard- potters
ened by heat. Today, a distinction is sometimes (b) The physical properties of the raw
made between pottery, applied to lower-quality materials, from their preparation as a clay
ceramic wares, and higher-grade porcelain. No body, drying and firing, and through their
such distinction will be made here so the term transformations during manufacture into a
pottery alone will be used. Raw material that goes final ceramic product
into the making of a pot includes primarily the clay (c) The nature of the chemical and mineral
body, but also varying amounts of “temper,” added reactions that take place during firing as a
to make the material more manageable and to help clue to the technology available to the potter
preserve the worked shape of the pot during firing. (d) The uses, provenance, and trade of the
Of primary interest in ceramic studies are: wares produced
196 7 Clay Minerals and Ceramics

Neff (1993) defines “compositional” analyses preserved. Pottery is classified as a “ware”

as the characterization of the ceramic fabric, largely on the basis of its firing characteristics
either by chemical or mineralogical analysis. and surface treatment decoration (Table 7.3). A
Indeed, trace element studies are foremost distinction is made between vitrified ceramics,
among the methods to study earthenwares i.e., high enough firing temperatures to fuse the
(Vandiver 2001). Compositional approaches typ- clay into a glassy matrix versus lower-
ically seek to recognize cultural transmission by temperature unvitrified, porous material. The for-
use of material characterization to differentiate mer include stoneware and porcelain; the latter
local from nonlocal ceramics by their clays or terracotta and earthenware.
tempering agents. Efforts to recognize local ver- The earliest fired pottery worldwide are
sus nonlocal wares derives much of their impetus terracottas. They are fired at relatively low
from exchange studies, wherein the assumption temperatures, usually below 900  C. Terracotta
is made that local pottery is likely to pertain to a vessels, sculptures, and tiles are generally not
single (local) tradition within which pottery- glazed, although they may be surface treated by
making information was perpetuated over some roughening, to absorb heat and prevent slipping
period of time. Chemical data (e.g., Foust when wet, or covered with a slip. The slip is a
et al. 1989; Olin and Blackman 1989) or petro- suspension of fine-grained clay and water that
graphic data (Stoltman 1989, 1991, 2001) seek to may be added to enhance the beauty of the pot
refine these local or nonlocal ceramic typologies by coloring and smoothing, as well as to reduce
(Neff, supra). Not all archaeologists agree with a porosity and retard leaking.
strictly compositional approach. Sullivan Earthenware bodies are also porous and
(1984:77) questions the value of technological unvitrified but are fired at higher temperature
attributes (including mineralogy) in southwest- than terra-cottas, up to about 1,100–1,200  C.
ern ceramic taxonomy, arguing that decorative Because of the higher temperatures of firing,
attributes provide a superior foundation for material at the surface might fuse and form a
ceramic taxonomy because they may be shared natural glaze or they may also be unglazed.
across local technological bases, thus facilitating When fired at lower temperature, they grade
description and comparison of assemblages. into terra-cotta. The clays used in the manufac-
Much of the information needed to answer ture of earthenware are generally coarse and
compositional questions is, as Neff (supra) plastic when wet and many are red firing.
suggests, available through the use of standard Earthenwares include a wide variety of products,
geochemical and petrographic analysis of ranging from “coarse” such as bricks and tiles to
ceramic artifacts. Insight into the working “fine” wall tiles and majolica, tin enameled and
methods of ancient potters also has been obtained made from a good-quality white-burning clay.
through ethnographic studies of cultures where, Stonewares are fired at temperatures that are
either because of isolation or conservative sufficiently high enough, 1,200–1,350  C, to
traditions or both, ancient methods have been achieve partial fusion or vitrification of the clay

Table 7.3 Characteristics of ceramic bodies

Porosity Firing temp Applications/notes
Terracotta High:~30 % <1,000  C Roof tiles, bricks unglazed, most prehistoric coarse
Earthenware 0–25 % 900–1,200  C Coarse: tiles, bricks glazed or fine: wall and floor tile, unglazed
majolica, unvitrified
Stoneware 0.5–2.0 % 1,200–1,350  C Roof tiles, tableware, glazed or artware unglazed, vitrified
Chinaware <1 % 1,100–1,200  C Tableware, white, vitrified
Porcelain <1 % 1,300–1,400  C Fine tableware, hard body, artware, fine, “rings” when tapped
From Rice (1987)
7.3 Ceramics and “Ceramics” 197

body. The body itself is medium to coarse much sought after property, especially when
grained, opaque, and commonly light brown or form blue-and-white wares grew in popularity.
gray. Modern stoneware is usually made of sedi- Pottery stone could be fired at a lower tempera-
mentary clays that are transported rather than ture (1,250  C) than paste mixed with kaolin,
residual deposits, such as ball clays which are which required 1,350  C.
highly plastic and low in iron. Archaeological In Ming porcelains, cobalt was used for
stonewares may or may not be glazed. underglaze blue decoration. Prior to this, the
Porcelains were first produced in China during cobalt had been brilliant in color, but with a
the time of the T’ang dynasty in the ninth to tenth tendency to bleed in firing; by adding manga-
century AD. They are fired at 1,280–1,400  C or nese, the blue color was duller, but the line
even higher, required to vitrify the highly refrac- crisper. Xuande porcelain was the finest of all
tory white clay, kaolin, used in their manufac- Ming output. During a significant portion of the
ture. The clay is relatively pure and mixed with Ming Dynastic period, exports of Chinese
quartz and alkali feldspars which act as a flux, porcelains were banned. This allowed Vietnam-
allowing lower firing temperatures than needed ese porcelains and stonewares to become widely
to vitrify pure kaolin. At 1,150  C, K-feldspar traded. Evidence of this period of a Vietnamese
(microcline and orthoclase) melt, and at ceramic florescence comes from both early kiln
1,118  C, Na-rich plagioclase (albite) melts, giv- sites excavated in the 1990s as well as South
ing the porcelain a translucency, hardness, and China Sea shipwrecks of both Chinese and Viet-
ring when tapped. Porcelains today are made up namese origins: Chu lâu-My Xa kiln site and in
of about half kaolin plus roughly equal amounts the Cu Lao Chàm (Hôi An) cargo (~250,000
of feldspar and quartz. ceramic vessels). The latter shipwreck was
A major Chinese production center, 600 years excavated by a team led by Oxford University’s
ago, and today is Jingdezhen. AD 907–960, the Mensun Bound, in cooperation with the Viet-
“5 Dynasties Period.” Other major periods are as namese government. Elements of the Hôi An
follows: Shufu (12th c.), Qingbai/Yingqing cargo were sold at auction (Pope 2014).
(12th–13th c.), “Blue and White” (post-14th c.), The importance of Vietnamese ceramics pro-
Ming Period (1368–1644), and “Monochromes” duced during the fifteenth century has been
(post-18th c.). Today, Jingdezhen, produces established beyond any doubt. At this time, the
300 million pieces of porcelain a year. ban on Chinese ceramic exportation by the Ming
Vogt, in 1900, established the chemical and dynasty offered an open market to Vietnamese
mineralogical nature of Chinese porcelain. He products: blue-and-white porcelains, celadons
determined the major minerals present were kao- (jade-like monochrome) and tam thai (three
linite; potash mica; soda feldspar, albite; and colors, to be compared with the Chinese sam
quartz (SiO2). cai equivalent) overglazed wares. The Vietnam-
The kilns at Jingdezhen became the main pro- ese learned ceramic skills while under Chinese
duction center for large-scale porcelain exports domination up to the 11th c. AD and continued
to Europe starting with the reign of the Wanli their indigenous development thereafter. They
Emperor (1572–1620). were experts in both porcelains and elegant
Kaolin and pottery stone—pu-tan—were stonewares. The latter fall under the heading of
mixed in about equal proportions (d’Entrecolles celadons or jade-like, colored stonewares
1743) identified this as “Best Quality or equal (Colomban et al. 2004). Likewise, the Vietnam-
parts petuntse and kaolin” (Pollard and Wood ese used color to great advantage in their designs.
1986). Kaolin produced wares of great strength The constituent elements as well as the struc-
when added to the paste; it also enhanced the tural phases of the blue-and-white and tam thai
whiteness of the body—a trait that became a glazes of the fifteenth-century Vietnamese
198 7 Clay Minerals and Ceramics

ceramics have been determined by a combination northern China result from the divergent evolu-
of XRF, EDS, and nondestructive Raman spec- tion in distinct ceramic environments. Chaff tem-
troscopy. The transparent glaze is an aluminosil- per and sequential slab construction in
icate glass, with CaO as the chief fluxing agent. northwestern Iran were adaptations to highly
The firing temperature for this base glaze was as plastic montmorillonite clays, while coiling and
high as the sintering temperature of the body, that tamping with paddle and anvil were adaptations
is, more than 1,200  C. The jade-like color of the to the gritty loess-derived illitic clays of northern
celadon glazes was achieved by iron doping in China. In addition, the northern Chinese clays
the glazes, while the black colors or metallic- were more forgiving over a wide range of firing
black dendrites were obtained by the dispersion temperatures, permitting higher firing
of FeO and/or Fe3O4 under the transparent glaze temperatures in China. Coincidentally, the
(in a very strong reducing atmosphere). The blue uniform refractory properties of loess soils of
decoration under the transparent glaze was northern China favored high-temperature firing
achieved using Co-containing manganese ore. in subterranean kilns, whereas Near Eastern
In some cases, the addition of iron made a dark potters developed aboveground kilns, the
greenish-blue color. temperatures of which could usually be
Cyfflé’s Terre de Lorraine manufacture in maintained below the level at which
Lunéville (1766–1780) produced some of the montmorillonitic kiln walls would fail.
first European porcelains (Maggetti 2011). The
diversity of Cyfflé’s production is now better
recognized. His trial-and-error experiments 7.4 Provenance
made use of a remarkably wide range of paste
mixtures, with porcelain bodies in the French Determining the provenance of clay used in the
(soft-paste) and the German (hard-paste) tradi- manufacture of a pot is an important problem that
tion (supra). is traditionally attacked by geochemical
An example of soft-paste (artificial) porcelain techniques. Many studies have used trace ele-
body (10 % water adsorption), contained quartz, ment analysis to compare the possible sources
calcic plagioclase (An 88–95) in a glassy matrix. to a fired ceramic. However, conclusions of
Electron-backscattered diffraction analysis such studies must be carefully interpreted since
(EBDS) revealed a coronitic reaction rim visible ceramic pastes will not always be similar miner-
around the quartz. The K-rich and Na-poor frit is alogically or chemically to their sources, except
a mixture of potassium nitrate, alum, calcined in the cases of unrefined clay, typical of very
gypsum, sand, and moderate amounts of salt ancient potters and very low-temperature firing.
and soda. The water adsorption values indicated Most ceramic clay has been refined and mixed
a somewhat underfired condition. In other resulting in a great variation in particle size dis-
samples, Maggetti (supra) found evidence of tribution and mineralogical variations. Blackman
the high-temperature silicate mullite typical of (1992) found that levigation of the raw clay
hard-paste porcelains. resulted in major changes in mineralogy and
In another study employing a comparative chemistry if silt and finer sizes were removed.
approach to technological characteristics, If only sand size particles were removed, the
Vandiver (1988) reconstructed and compared changes were not as serious except for trace
the forming and firing practices of Neolithic elements that were concentrated in the accessory
potters in northwestern Iran and northern China. minerals.
She argues that the complex technological A study by Kilikoglou et al. (1988) found that
differences between pottery from Hajji Firuz, most trace elements showed great variations in
northwestern Iran, and Yangshao pottery from comparing the raw to the fired clay. Whether the
7.6 Optical Petrography 199

elements increased or decreased upon firing and manufacturing technology can be resolved by
purification depended on many variables includ- using the complementary techniques of petro-
ing their: graphic and geochemical analysis to determine
mineral and chemical compositions.
1. Physical nature—whether they were held in The hand lens, binocular microscope, optical
fine- or coarse-grained material, organic mat- (polarizing) microscope, scanning electron
ter, in the clay, or in detrital minerals microscope (SEM) equipped with analytical
2. Chemistry—whether they were present as capabilities (EDS), X-ray diffraction (XRD),
oxides, hydroxyls, carbonates, and on their and electron microprobe (EMPA-WDS/SXES)
position in the clay lattice, i.e., interlayer, are the most important of these tools. The initial
adsorbed on the surface, or within a sheet study of a potsherd should begin with a hand
3. Physical properties—most important, the tem- lens, followed by thin sections and a petrographic
perature at which they volatilize microscope. These studies are done at low mag-
nification, that is, <200 times magnification.
The conclusion of their study was that under Textures, including pore spaces, and mineral
no circumstances can the chemistry of raw clay inclusions or phases can be identified. At this
be compared to that of a finished pot. To demon- point, unknown fine-grained minerals and phase
strate the provenance of the clay in the pot, the reactions at mineral boundaries can be labelled
steps taken by the potter in making the pot must for detailed study by EMPA or SEM. A prelimi-
be followed: the clay must be purified and fired nary grouping of the potsherd samples can be
under analogous conditions. The analytical made, and, in conjunction with the archaeologist,
methods then used to compare the synthetic pot decisions made on what other types of analysis,
to the original should be sensitive to elements and perhaps field sampling, are needed to obtain
present in trace amounts and include optical further information.
emission spectroscopy, neutron activation,
X-ray fluorescence, SEM, or the electron micro-
probe. Then, so that meaningful statistical anal-
ysis can be carried out, (1) the number of 7.6 Optical Petrography
elements determined should be as large as possi-
ble, and (3) a large number of samples should be A pottery sherd may be regarded as a
analyzed. metamorphosed sedimentary rock and clearly
amendable to petrographic analysis. Quinn
(2013), Stoltman (1989, 1991) and others
7.5 Methods (Kamilli and Sternberg 1985; Kamilli and
Lamberg-Karlovsky 1979; Rice 1987; Whitbread
Geochemical and petrographic analyses are 1989; Williams 1983; Williams et al. 1955), in
widely used to characterize archaeological recent years, have reiterated the importance of
ceramics (Tite 1999; Vandiver 2001). As noted petrography to the study of archaeological
above, bulk trace element analyses with either ceramics championed, earlier, by Shepard
NAA or ICP-MS methods are foremost among (1936, 1942, 1954, 1966) as well as by Matson
studies of trade and provenance and determining (1960, 1966) and Peacock (1981, 1970). A rela-
manufacturing/geologic sources. After trace ele- tively steady volume of studies, of both Old and
ment analysis follows petrographic analysis, New World focus, have continued utilizing tried-
bulk chemical analysis, and microstructural and-true geological protocols such as point
studies (Vandiver, supra). Important problems counting (Chayes 1954, 1956; Galehouse 1971)
of sourcing raw materials and characterizing and optical/instrumental petrographic procedures
200 7 Clay Minerals and Ceramics

(Freestone and Middleton 1987; Garret 1986: From the perspective of optical petrography, the
Lombard 1987; Maggetti 1982; Vaughn 1990, procedure is that for the study of archaeological
1991). rocks and minerals:
There are many good arguments for beginning
a pottery study with petrographic analysis. Not 1. Prepare a thin section. Two examples are
only is this the cheapest kind of analysis, but shown in Fig. 7.7.
information can be obtained that cannot be satis- 2. Stain, etch, or otherwise pretreat the thin
factorily obtained by any other method. In petro- section.
graphic and textural studies of ceramics, four 3. Examine the thin section under the low-power
major elements can be readily evaluated: PPL/XPL microscope, determining minerals/
phases by optical properties and the propor-
1. Minerals—major and accessory minerals can tion of each observed.
be examined by both petrography and particle 4. Quantitatively evaluate the thin section using
size studies. standard point-counting procedures.
2. Clay minerals—well-fired pottery does not 5. Examine the thin section under high-power
contain true clay minerals since they have PPL/XPL for relief, color/birefringence, etc.
reacted and transformed to other minerals.
3. Non-mineral inclusions—both petrography An integrated petrographic and chemical
and textural analyses are productive methods. approach (NAA) to ceramics of all varieties as
commonplace has been employed in the prove-
The use of particle (grain) size analysis, com- nance studies of amphoras, Egyptian and Canaan-
bined with the low-power study of grain mor- ite jars, as wells a Minoan ceramics (Day
phology, has much to offer the ceramic et al. 2011; Hancock et al. 1986; Hancock and
petrographer. Betancourt 1987); American Southwest prehis-
Because clay mineral identification is not typ- toric ceramics (Carr 1984,1990); joining a host of
ically done with the petrographic microscope, we studies enumerated in monographs and books
will hold our discussion until the following chap- (e.g., Bronitsky 1989; Kingery 1984, 1990, 1993;
ter where more appropriate instrumental Nelson 1984; Plog 1980; Rice 1987, 1996a, b,
techniques are examined. It is important to note 1999, 2001; Sassaman 1993; Sinopoli 1991;
that response of clay minerals and inclusions to Skibo and Feinman 1999; Glasscock et al. 2007a,
firing temperature produces progradational min- b)—all studies of ceramic technology.
eralization, as in metamorphism, as well as melt- Josephs (2005) makes good case for a
ing and recrystallization of the clays and standardization of terminology in the use of
accessories (Herz and Garrison 1998). A com- micromorphological analysis of thin sections.
mon reaction seen in highly fired wares is the Thin sections of ceramics have much in common
formation of mullite (3A12O3 2SiO3) its anteced- with those of sediment and soil. Whereas
ent minerals such as kyanite and sillimanite protocols for the micromorphology description
(cf. preceding section). Other minerals such as of soils and sediments are well established
hedenbergite and fayalite (pyroxene and iron- (Goldberg 1995; Courty et al. 1990; Kooistra
rich olivine, respectively) form as well, given and Kooistra 2003), the same cannot be said for
the appropriate clay matrix. All of these crystal micromorphological studies of ceramics (supra;
minerals are readily identified by their optical Courty and Roux 1995).
properties—kyanite, in particular, with its rich In ceramic studies of archaeological interest,
blue birefringence as well as the bright greens rocks and minerals continue to play an often
of the pyroxene. decisive role in questions regarding ceramic
In certain ceramics such as faience and technology, provenance, trade, and linkages to
porcelains, the effects of firing are defining belief systems; gender relations and power; as
elements in the identification of these varieties. well as art and aesthetics in the past cultures. In
7.8 Chemical Analysis 201

the next chapter, we shall examine instrumental slab to slab, or pulled/pinched section to section.
procedures in greater depth, while we close this In coarse-tempered coiled vessels, clustering can
chapter with a continuation of our discussion of also result from the rolling out of their coils.
radiographic and chemical methods. Coarse particles are driven to the centers of
coils (Carr, supra). Braun (1982) is more cau-
tionary stating that tempering inclusions of
7.7 X-Ray Radiography crushed sherds or simply crushed fired clay, for
example, may be no more or less radiographi-
Hidden features that can be detected radiograph- cally dense than the surrounding matrix. Radio-
ically include coils, slabs, their size, morphology, graphic experimental trial was unable to detect
methods of joining; the material type, approxi- crushed shard tempering in Woodland Period
mate mineralogy, size, density, and orientation of pottery from the south-central Illinois. Nonethe-
aplastic inclusions or voids; fracture systems; less, where tempering raw materials do not
paste texture; and hidden vessel parts. Under impose limitations, the use of radiographic
certain conditions, these features can be used to methods for analyzing temper appears both fea-
sort shards by their vessels of origin, to identify sible and, for several large areas of research,
the primary and secondary methods of vessel highly desirable (Braun, supra).
manufacture (Carr 1990). As Carr notes (supra), Computed tomography (CT) and computed
X-radiography had been used occasionally by radiography (CR)—digital X-ray methods—
archaeologists to study ceramics since at least have rapidly joined the earlier conventional
the 1930s. Titterington (1935), Digby (1948), X-ray methods of years past (Kim and Liaw
and Shepard (1956) defined a limited set of 1998; Casali 2006). In a study of Late Bronze
tasks to which radiography is applicable. They Age cremation urns, Harvig et al. (2011) used
focused largely on documenting vessel both CT and CR. CT scanners are common in
manufacturing techniques. Radiographs can be today’s hospitals and are equipped with various
used to reveal the primary and/or secondary for- computer programs allowing rapid image build-
mation procedures by which a vessel was ing and three-dimensional visualization, but the
constructed: pinching, slab building, drawing, software has been developed for medical
coiling, throwing (wheel), and molding/anvil, purposes. Harvig et al. (2011) utilized a program
for example. This application was first suggested package (MIMICS®, from Materialise®,
by Shepard (1956) and then systematically Belgium ( developed
investigated by Rye (1977). at the Laboratory of Biological Anthropology in
Radiographs of a vessel can allow one to Copenhagen. After importing the CT scanner
determine the volumetric density, size distribu- data file, the single-slice images were accessible
tion, shape, material type, and general mineral- and the program allowed immediate multiplanar
ogy of temper inclusions. The material type or reformatting and editing of single-image
general mineralogy of temper inclusions can be elements.
inferred from their relative grey levels on a radio-
graph supplemented with size and shape infor-
mation. This temper mineralogy is usually
further deduced using other X-ray analytical 7.8 Chemical Analysis
methods such as XRF and SEM-EDS or WDS.
Braun (1983) and Carr (1990), for example, After petrographic and optical examination,
have shown that temper volumetric density of a chemical analysis is generally carried out.
vessel can also be estimated more accurately by Much chemical information could have been
radiography. Temper particles can have a clus- obtained using X-ray diffraction (XRD), the ana-
tered rather than a random distribution within the lytical scanning electron microscope
paste of a vessel as a result of poor kneading or (SEM-EDS), and the electron microprobe
pounding. Temper density can vary from coil to (EMPA), and, in many cases, could be enough
202 7 Clay Minerals and Ceramics

to answer all important questions of provenance groupings of archaeological ceramics. Non-clay

or technology of firing or both. If more detailed inclusions in pottery paste can be identified with
information is needed, especially quantitative petrographically distinctive geologic sources.
chemical analysis of major and trace elements When this is not the case, geochemical analysis
of the paste, temper, slip, and detrital minerals such a NAA can be employed to search for more
for provenance determination, or for technology subtle, chemical evidence of such groupings.
of firing, then a method of analysis can be Where the objective is to discern local or subre-
decided upon. Many methods of chemical analy- gional patterns of pottery manufacture and
sis are available, and some, such as neutron acti- exchange, this has proven to be informative
vation analysis (NAA), are very expensive and (Stewart et al. 1990; Speakman et al. 2011a, b).
may not produce the needed results. This is a
critical step in a ceramic study; before one kind
of analysis is decided upon, the following should 7.8.1 An Example of Chemical Analysis
be considered: Using X-Ray Diffraction (XRD)

1. Some methods can only analyze material This instrumental analysis foreshadows a broader
within a few mm of the surface, such as discussion of the use of these techniques across
SEM and electron microprobe, but give geoarchaeology. While Tite (1999) is on record
exceptional small-scale detail on chemistry that because of the breakdown of clay minerals
and phase changes. during firing, infrared spectroscopy nor X-ray
2. Each method involves a varying degree of diffraction are useful methods to determine the
destructive or nondestructive sample original clay minerals, McReynolds et al. (2008)
preparation. have used X-ray diffraction analysis to examine
3. If samples are to be compared against an local clays from the Piedmont and Coastal Plain
existing database, it is preferable to use a regions of North Carolina in a study exploring
like type of data - XRF, NAA, ICP - for the residential mobility and resource use (Fig. 7.8).
comparative database. The same elements Variability in mineral assemblages, across
measured by NAA and XRF, for example, 42 clay samples, were examined in the context
will show completely different abundances. of region (Piedmont/Sandhills) in order to iden-
4. Chemical analysis by itself generally does not tify potential sources for the prehistoric pottery
resolve what parameter is responsible for found in the Sandhills region (Fig. 7.9). While
which element. This is best done by other clay, as a pottery resource is relatively common
methods where optical control is maintained, across the Carolina landscape, compositional
such as SEM and the electron microprobe. analyses of prehistoric ceramics indicated that a
However, NAA has been used for a relatively relatively fewer number of specific were utilized.
long period, and NAA data bases have been Availability was not necessarily a criterion for
established for ceramic provenance studies. selectability.
ICPS (inductively coupled plasma emission As will be noted in the following chapter on
spectroscopy) is a relatively new technique instrumental techniques, XRD is uniquely capa-
used in archaeology with many advantages ble of detecting structural differences among
over other analytical systems. Suites of clay minerals of unfired samples. As we have
major elements as well as trace elements are seen in the foregoing discussion of effects of
routinely determined with a high degree of firing temperatures—decomposition of existing
precision and at relatively inexpensive cost. minerals and the formation of new phases—
Numerous example studies abound where unfired clays preserve the most mineralogical
examination of thin sections has proven to be a information. XRD retrieves this information in
successful method for discovering provenance the form of diffractograms produced by
7.8 Chemical Analysis 203

Fig. 7.8 Sandhills regions,

North Carolina. Clay
sample locations are
marked (Fig. 7.3, p. 34,
Herbert and McReynolds

monochromatic X-rays bombarding a crystalline 10, and 14 Å. Within the 7 Å group were the
mineral such as clay. The reference variable is kaolins; the 10 Å group, the illite/micas; and in
that of the so-called d-spacing or the interatomic the 14 Å group, the hydroxyl-interlayered
layers or crystallographic planes of the mineral vermiculites/chlorites/smectites. Of these
(cf. Chap. 8). Specific d-spacings identify spe- groups, the 14 Å group clay minerals dominated
cific minerals when compared to reference in the Sandhills samples with several being
materials (Stanjek and Häusler 2004). enriched in the 7 Å kaolin group minerals. The
The Sandhills clay study examined ten clay- Coastal Plain clays, by comparison, were rich in
like minerals grouped within three primary the 14 Å group minerals and poor in the 10 Å
groups based on general d-spacing intervals—7, illite/mica minerals (Fig. 7.10).
204 7 Clay Minerals and Ceramics

Fig. 7.9 Various pottery shards from the Sandhills JMH056, New River Fabric Impressed (flex warp); (g)
regions. Pottery samples from the Kolb site (38 Da75): JMH057,New River Cord Marked; (h) JMH058, Cape
(a) JMH051, Yadkin Fabric Impressed; (b) JMH052, Fear Fabric Impressed; (i) JMH059, Cape Fear Fabric
Yadkin/Hanover Fabric Impressed; (c) JMH053, Impressed; (j) JMH060, Hanover I Fabric Impressed
Yadkin/Hanover Cord Marked; (d) JMH054, New River (Fig. 7.8, p. 28, Herbert and McReynolds 2008)
Cord Marked; (e) JMH055, Yadkin Cord Marked; (f)

Plagioclase feldspar is generally considered to 7.8.2 Diagenetic Changes

be the “parent” mineral for most Piedmont clays
(McReynolds et al.), whereas the Coastal Plain During burial elements of the pot will be subject
clays are potassium feldspars (K-feldspar) derived to both infiltration of new material and dissolu-
primarily from the Cretaceous-aged Cape Fear tion of some original constituents by solutions
Formation (Sohl and Owens 1991). When passing through and organisms present in the
included in the potting clay, the feldspars, most environment of burial. Unfortunately, in many
notably the K-feldspars, gave greater overall cases, the nature of diagenetic changes is
strength and durability to the pottery produced destroyed by archaeological treatment. On most
from them. Likewise, it was observed the smectitic digs, pottery is cleaned with relatively strong
(14 Å) clays were generally more plastic than the acids which dissolve much of the products of
kaolin-rich Coastal Plain (7 Å) clays. The XRD diagenesis, including the carbonates and zeolites.
study of the suite of 42 clay samples was able to Since both gypsum and calcite will be leached by
clearly differentiate the mineralogy of the source this treatment, important information will be lost.
clays. This information, in turn, at least in the case Gypsum is a product of cementation during
of the feldspar minerals, can aid archaeologists in burial; the form of the calcite, either powdery
assessing the production and movement of pottery or crystalline, will reveal whether or not
within the Sandhills regions. temperatures of lime formation were reached
7.8 Chemical Analysis 205

Fig. 7.10 Ternary plots of

Sandhills’ clay groups
(Fig. 7.2, p. 117, Herbert
and McReynolds 2008)

and if soil processes caused the formation of of this type of study is Richard Evershed (Heron
secondary calcite. and Evershed 1993). An interesting application
Diagenetic changes during burial are an espe- of residue analysis involved the dating of prehis-
cially serious problem for low-temperature toric steatite vessels from the Piedmont region of
ceramics, that is, those fired about 500–700  C. the southern Appalachians wherein charred
Such pottery is generally very porous allowing residues adhering to the interior surface have
easy access for outside solutions. Dissolution and been used to determine the age of the vessel
neocrystallization will take place: deposition in (Truncer 2004).
pore spaces, replacement of primary minerals by Chemical analysis of food residues is
secondary carbonates, hydrates, oxides, associated with Late Pleistocene pottery, focusing
hydrosilicates (zeolites), and sulfates (gypsum). on one of the best-studied prehistoric ceramic
The initial porosity is due to the manufacturing sequences in the world, the Japanese Jōmon
and firing processes, the chemistry and mineral- (Craig et al. 2013. The results of residue analyses
ogy of the starting material, and the loss of other demonstrate that lipids can be recovered reliably
substances. from charred surface deposits adhering to pottery
dating from about 15,000 to 11,800 cal bp (the
Incipient Jōmon period), the oldest pottery so far
7.8.3 Analysis of Residues investigated, and, that in most cases, these organic
compounds are unequivocally derived from
Analysis of organic materials has been carried processing freshwater and marine organisms.
out to determine reactions that may have taken The most useful surviving organic compounds
place between the food or perfume and the pot in are lipids, which are the constituents of fats, oils,
which it had been stored or cooked. This type of waxes, and resins (Heron and Evershed 1993;
analysis is generally carried out by FTIR (Fourier Tite 1999). Because of hydrophobic properties,
Transform Infrared) Spectroscopy or by similar lipids survive on pottery. The lipids are extracted
specialized means—chromatography—that can in a solvent, and the individual lipids are
identify organic molecules. A leading proponent separated out, using gas chromatography. The
206 7 Clay Minerals and Ceramics

mass associated with each lipid can then be deter- provided a vegetable protein which compensated
mined by mass spectrometry. Identification of the a lack of essential proteins in a predominately
lipids is based on a combination of the retention (75 %) maize-only diet (Lipe, supra).
time in gas chromatography and their mass spec- Beans require(d) longer cooking times than
tra. Alternatively, the pattern of lipids separated does maize. Basketmaker III groups adopted
out from the archaeological pottery vessel by gas ceramic wares as a result. To chemically distin-
chromatography can be compared with the guish food like beans, one can use sterols present
patterns obtained from modern examples of the to distinguish animal products (cholesterol) from
different animal and plant species which the ves- vegetable products (campesterol or sitosterol)
sel could have contained (REF). (Evershed et al.,supra). This example of the
Fatty acids can provide a broad classification need for the adoption of pottery by a culture
into animal fats, dairy products, vegetable oils, or may be a further corroboration of Brown’s
fish oils, but it is not normally possible to identify (1989) contention that the adoption of pottery
the actual animal species from the fatty acids as a container was “a response to conditions in
alone. However, on the basis of the stable carbon which the rising demand for watertight, fire-
isotopic composition of the fatty acids extracted resistant containers is coupled with constraints
from pottery, it has been possible to distinguish in meeting this demand.” Thus, pottery was
between fats from ruminant (i.e., ovine and adopted when existing other forms of containers
bovine) and nonruminant (i.e., porcine) animals failed to meet the increasing demand brought
(Evershed et al. 1997). about by new types of food processing particu-
As mentioned earlier in this chapter, larly that of beans rather than corn/maize.
Diederichs et al. (2013) and Lipe (2015) assert In Fig. 7.11, Reber and Evershed (2004) illus-
the Basketmaker III Period was part of an impor- trate the chromatographic spectrum of lipid resi-
tant cultural shift in the history of the American due of maize from Mississippian ceramics. Much
Southwest, when a new technological package— like the example of maize and beans in the
including cooking with pottery—enabled the Colorado Basketmaker III discoveries, maize,
rapid expansion of Puebloan populations and beans, and squash were the requisite staples of
represents a material culture horizon in the the Mississippian cultural florescence in the cen-
archaeological record. The addition of beans to tral Mississippi Valley and across the late prehis-
the dietary repertoire of this farming society toric American Southeast.

Fig. 7.11 The total ion

chromatogram (TIC) of
maize lipids. This shows
primarily the plant and/or
fish fatty acid ratio (more
unsaturated C18:1 than
saturated C16:0), plant
biomarkers (sitosterol), and
Triacylglycerols are
referred to as TAGs,
diacylglycerols as DAGs,
and monoacylglycerols as
MAGs (Reproduced from
Reber and Evershed 2004,
Fig. 7.1)
7.8 Chemical Analysis 207

Reber and Evershed (supra) term this food classes of compound. Cholesterol is a lipid bio-
residue approach as “biomarker analysis.” It is marker for meat (Heron and Evershed 1993),
the second approach used to identify the original sitosterol and other plant sterols are lipid
contents of vessels through absorbed organic res- biomarkers for plants (Heron 1989; Evershed
idue analysis. Biomarkers are unique to a single et al. 1992; Evershed 1990). The presence of a
source or group of foods. Lipid biomarkers sur- lipid biomarker allows a more specific identifica-
vive well and are more easily analyzed than other tion of the foods once cooked in a sherd.
Instrumental Analytical Techniques
for Archaeological Geology 8

8.1 Introduction materials of art and archaeology, the methods

must ofttimes be nondestructive or, at the very
A simple search of either modern scholarly or least, destructive of only a small portion of the
popular scientific literature confirms the rele- material under examination. To Sir Humphrey
vance of chemical analyses to archaeology. Davy’s chagrin, his 1818 attempt to unroll
Chemical analyses are not only used to determine Pompeiian papyrus scrolls met with less than
the chemical makeup of an artifact but also to success when all of the eleven scrolls were effec-
reveal clues to age, diet, and health when applied tively destroyed before any decipherment could
to both organic and inorganic materials. Chemis- be done (Herz and Garrison 1998). In contrast,
try is equally important in the preservation and analytical chemistry’s exposure of science’s
restoration of archaeological materials. The first most famous fraud—Piltdown—using the quan-
use of chemistry in the study of artifacts can be titative determination of the elements fluorine,
traced to German chemist Martin Heinrich nitrogen, and uranium in the “fossil” hominid
Klaproth’s 1796 studies of Greek and Roman remains and their comparison to both modern
coins, other metal artifacts, and some Roman and prehistoric bones, in the mid-twentieth cen-
glass in which he pioneered the use of gravimetric tury, was the watershed event that brought ana-
analysis (Goffer 2006). Gravimetric analysis is lytical chemistry into the forefront of subsequent
the determination of an element through the mea- scientific studies of archaeological materials.
surement of the weight of the insoluble product of The chemical examination of the physical
a reaction with that element. Since Klaproth’s properties of artifacts and other materials of
first foray, there have been numerous examples archaeological interest generally requires instru-
of the application of analytical chemistry to mental assistance. In the last chapter, that instru-
archaeological materials of all sorts by historical ment was the petrographic microscope, which, as
figures in both chemistry and physics—Sir we have seen, is a very useful tool in the service
Humphrey Davy, Jon Davy (Sir Humphrey’s of archaeological geology. A mineral’s optical
brother), Jöns Jakob Berzelius, Jon Voelker, properties are attributes which aid us in identifi-
Michael Faraday, Marcelin Berthelot, and cation and, by extension, the characterization of
Friedrich August von Kekulè (various in Meschel the archaeological item. With this characteriza-
1980; Goffer supra; and Pollard and Heron 1996). tion, we can, as archaeologists, hope to use this
According to Robinson (1995), one important mineralogical information to pose pertinent
function of analytical chemistry is to provide inquiries as Freestone and Middleton (1987)
data to other scientists. When applied to have aptly suggested in the following:

# Springer-Verlag Berlin Heidelberg 2016 209

E. Garrison, Techniques in Archaeological Geology, Natural Science in Archaeology,
DOI 10.1007/978-3-319-30232-4_8
210 8 Instrumental Analytical Techniques for Archaeological Geology

1. The characterization of the material from analysis determines it is bronze, as to its appro-
which an object was made priate typology, then we must do a quantitative
2. The reconstruction of the technology involved assessment.
in its manufacture Tite (1999) observes that in the early days of
3. The inference of the place of manufacture or provenance studies, using chemical analysis, the
source of raw materials principal techniques were quite different from
4. The changes that have occurred in the object those in use today. For instance, optical emission
during burial or storage spectroscopy (OES) was superseded first by
atomic absorption analysis (AAS) then by induc-
The reader can certainly suggest other impor- tively coupled plasma mass spectroscopy (ICP
tant questions, but these are surely common MS). AAS, itself, has continued to be used in
to most inquiries. Nikischer (1999) has many labs but the rise of ICP MS, particularly;
reminded us that contrary to popular belief, laser ablation (LA) ICP MS has impacted its
there aren’t any foolproof single (author’s usage as well. Neutron activation (INAA), ICP
emphasis) technological innovations that can MS, and X-ray fluorescence (XRF) along with
produce an accurate identification every time newer techniques, such as proton-induced X-ray
for every mineral (or material for that matter). emission (PIXE), have found the widest
Frequently, two or more confirming techniques applications (supra).
are needed. “This sentiment is echoed by more
than one compendium on techniques used in the
chemical/mineralogical analysis of archaeolog- 8.2 Analytical Techniques
ical materials (Artioli and Angelini 2011; and Their Pluses (and Minuses)
Ciliberto and Spoto 2000; Garrison 2001; for Archaeology
Glasscock et al. 2007b; Henderson 2000; Tite
1999; Vandiver 2001). Techniques commonly encountered in the study
In the application of analytical techniques to of archaeological materials include:
archaeological materials, one can seek data at (1)
the molecular level and (2) the elemental/isoto- 1. X-ray diffraction spectroscopy (XRD)
pic level in archaeological geology. In archaeo- 2. X-ray fluorescence (XRF) spectroscopy
logical geology where this would be illustrated 3. Electron microprobe/analytical scanning elec-
by either examining a lithic (marble) sample as to tron microscopy (EMPA/SEM)
its mineralogical (molecular) form, such as cal- 4. Atomic absorption spectroscopy (AAS)
cite (CaCO3), or assessing the metallurgy of an 5. Inductively coupled plasma emission spec-
early Bronze Age axe wherein the presence or troscopy/mass spectroscopy (ICP-ES/MS)
absence of the alloying metallic element tin 6. Mass spectroscopy (MS)
(Sn) is critical to its classification, both typologi- 7. Neutron activation analysis (NAA/INAA)
cally and chronologically. Likewise, a second 8. Infrared and Raman spectroscopy (IR/FTIR/
important distinction can be made in the analysis Raman)
of materials—that of (a) qualitative or
(b) quantitative. The former simply answers the European laboratories commonly use emis-
question, posed in our axe example, “what sion spectroscopy, atomic absorption spectros-
kind?” A quantitative study asks a second level copy, X-ray fluorescence and diffraction
of question—“how much?” Returning to the axe, spectroscopy, and neutron activation (Collon
if all we want to know is whether the artifact is a and Wiescher 2012). Additionally, ion probes
copper axe or a later bronze type, then a qualita- (PIXE; SHRIMP) and types of mass spectros-
tive study is appropriate to the question. But, if copy are commonly found. North and, to a some-
we desire to classify the axe, assuming our what lesser degree, South American laboratories
8.2 Analytical Techniques and Their Pluses (and Minuses) for Archaeology 211

mimic that of the Old World in the diversity of sensitive method for the quantitative study of
analytical methods regularly used on archaeolog- metals with detection limits down to 1012 g. It
ical materials as do Asian labs, particularly in is not generally used in the detection of
Japan and China. The European Union nonmetals. Newer AAS instruments have had to
(EU) Geochemical Facility at Bristol University expand as to rapid multielement, solid sample
(UK) has a comprehensive suite of modern ana- analyses because of the introduction of ICP or
lytical instruments that are available to member inductively coupled plasma spectroscopy.
European Union states together with 15 non-EU In its original form or in newer hybrid
members in eastern Europe and the Near East. configurations with mass spectrometers and
The instrumentation includes electron micro- lasers, ICP has rapidly challenged AAS with
probe (EMPA), ICP-MS, laser ablation (LA) - regard to the analyzable range of elements as
ICP-MS, XRF, ICP-AES, XRD, Mössbauer well as to quantitative sensitivity. Unlike AAS,
spectroscopy, Raman spectroscopy, Fourier ICP is restricted to the elemental analysis of
transform infrared spectroscopy (FTIR), nuclear materials. By itself, ICP is good for the analysis,
magnetic imaging (NMR), and SEM (with ana- like AAS, for metals. Coupled with mass spec-
lytical EDS). Surface analysis can be done using troscopy, ICP is routinely used for analyses of
Auger electron, secondary ion mass, and X-ray both metals and nonmetals. It has a wide, linear
photoelectron spectrometers. sensitivity range from parts per billion (ppb) to
From the perspectives of (1) molecular versus percent levels.
elemental and (2) qualitative versus quantitative, Mass spectroscopy is used in both qualitative
each technique has strengths and weaknesses. and quantitative studies of molecular, elemental/
The X-ray techniques of XRD and XRF are ionic, and isotopic forms up to very high atomic
good examples of this, XRD being a powerful mass. At the quantitative level, the mass spec-
instrumental technique in the qualitative and trometer is a commonly used tool for the deter-
quantitative determination of the molecular mination of liquid and gas samples and, when
nature of a material. XRF, by contrast, is a rou- combined with lasers, solid samples as well.
tine method for the qualitative study of elements (Instrumental) Neutron activation analysis
with atomic numbers higher than Z ¼ 10, includ- (INAA/NAA) requires either nuclear reactors or
ing both metals and nonmetals. Additionally, it is accelerators. The former are common in most
a very sensitive and reliable method for the quan- developed countries, but the reactors must be
titative study of elements of high atomic weights. lower power, research types rather than the
The electron microprobe/scanning electron more frequently found power generation
microscopic instruments (EMPA/SEM), like reactors. NAA is an extremely sensitive analyti-
that of XRF, have the capabilities of performing cal method that can reach the part per billion
both molecular and elemental analyses as well range (ppb ¼ ng/g), in many materials, with
those that are qualitative and/or quantitative. In ease. It is used for both qualitative and quantita-
terms of surface analysis, the SEM techniques tive elemental analyses.
are highly sensitive to the number of atoms Infrared spectroscopy has been a staple chem-
detected but not their respective percentages ical technique for the qualitative and quantitative
(Robinson 1995). Also, electron spectroscopy is study of organic compounds. It was the first
used only in a semiquantitative way. Newer elec- technique that the author first became acquainted
tron systems such as “analytical SEM” models with in its use on the adsorption of contaminants
have built-in X-ray energy dispersive spectrom- on surfaces. The literature is vast on this particu-
eter (EDS) capabilities that allow them to per- lar technique but its use in archaeological
form qualitative analyses. analyses has not been as extensive as many of
AAS or atomic absorption spectroscopy is not the other techniques we will survey. Newer
used in qualitative analyses except in single ele- variations on IR spectrometers such as the
ment studies. It is, perhaps, the most accurate and Fourier transform (FTIR) designs hold
212 8 Instrumental Analytical Techniques for Archaeological Geology

significant promise for the study of archaeolog- 1. Nondestructiveness of the technique

ical materials such as clays and particularly in 2. Accuracy and reliability of the results
field settings. 3. Availability of the technique
Raman spectroscopy is, as promoted by 4. Cost
Killick (2014), a powerful and swift alternative
to powder XRD (Smith and Clark 2004). It is The first consideration—nondestructiveness,
nondestructive, except for some compounds like by the technique, of the archaeological sample
some silver minerals that are degraded by expo- of interest—is not a trivial one. Because of the
sure to laser beams. It can be used on polished rarity and uniqueness of many archaeological
blocks and thin sections of metallic ores as well objects, any damage to them, beyond that of a
as unprepared specimens. Raman and infrared natural, diagenetic origin while unexcavated, is
spectroscopies measure the same phenomenon, generally to be avoided. Archaeologists and
which is the momentary absorption and loss of museum curators are loath to expose unique
vibrational energy by molecular bonds. collections to scientific analyses that pose the
Both Raman and infrared peaks correspond to threat of damage of destruction to the artifacts.
molecular bonds rather than crystal plane As an archaeologist, the author has had to deal
d-spacings as in XRD and work well in with just this problem when presented with the
identifying cryptocrystalline minerals. With choice between preservation of an excavated
regard to Killick’s observation concerning the item and the determination of a much-needed
speed of analysis using Raman spectroscopy, it absolute date. In this specific case, the decision
can take 1–2 h to obtain a publication-quality to sacrifice the archaeological sample—a prehis-
spectrum using powder XRD, whereas a spec- toric corn cob, rare in the context in which it was
trum of comparable quality can be obtained by found—proved too compelling when its preser-
Raman spectroscopy in as little as 5 min or a vation was weighed against the need for a reli-
spectrum sufficient for chemical identification able age determination of an incidence of
in as little as 30 s making it, like SEM-EDS, an prehistoric farming in the southeastern United
excellent technique for use in the rapid screening States. This is not usually the case in other
of samples. situations. Preservation of the analyzed archaeo-
In addition to these methods, we shall briefly logical material is considered paramount.
examine other analytical procedures, such as The accuracy and reliability of the analytical
those listed for the EU facility, whose usage has results are key considerations in the choice of an
produced insights into archaeological questions. instrumental technique. All the techniques we
Electron spin resonance (ESR), cathodolumi- shall discuss have varying levels for the detection
nescence (CL), and magnetic susceptibility will of elemental or mineralogical parameters. Some
be mentioned relative to specific archaeological can regularly achieve accurate determination of
findings such as provenance studies of marble, trace levels (1 μg) of interest. Others can only
cave sediment, and climate. A major consider- achieve accurate results at the percentage level of
ation in discussing this particular suite of instru- analysis. It is perhaps important to digress here
mental techniques is their common usage in the and eliminate any confusion about what is meant
analysis of materials of archaeological interest. by “accuracy” and, its oft-confused adjunct,
All have long track records in both their theoreti- “precision.” Accuracy refers to the correct mea-
cal underpinnings and their practical usage. For sure of a particular physical property such as
the archaeological researcher, other temperature, mass, age, etc. The precision of
considerations enter into the selection of the that measurement refers to its replication with
instrumental technique or techniques for his/her reliability and, most importantly, repeatability.
particular needs. Key among these practical If the true mass of an artifact is 30 g, then a
considerations are: measurement of 20 g is obviously in error.
8.3 X-ray Diffraction (XRD) 213

Should an instrumental technique repeatedly 8.3 X-ray Diffraction (XRD)

measure the instrumental mass as 20 g to three
or four place precision, it is still an inaccurate Spectroscopy is the determination of the physical
measurement; however, it is a “precise” inaccu- properties of a material—rock, mineral, biologic,
rate measurement. solid, liquid, or gas phase—by the absorption
The mineralogists Artioli and Angelini (2011) and/or emission of energy. The spectra of absorp-
caution that present-day trends in archaeometry, tion and emission can be characteristic of a mate-
in the service of cultural heritage, are in danger rial thus allowing its identification and
of a proliferation of “approximate” protocols and measurement. The following sections discuss
nonstandard instrumentation that can rapidly the properties and characteristics of individual
produce data that are simply wrong and instrumental techniques beginning with the
misleading. Other workers have raised similar X-ray spectroscopic techniques of XRD
warnings (Shackley 2011b; Speakman and XRF.
et al. 2011a, b; Speakman and Hunt 2014). The A colleague has called XRD the “first line of
potential to mass produce data that should never defense for archaeological (analytical) analyses”
enter the scientific literature is very real. (Paul Schroeder, 1996, personal communica-
Reproducibility and reliability are, therefore, tion). XRD is a technique that yields a mineral-
measurement-bound in that one seeks to achieve ogical identification rather than an elemental
accurate evaluations with instrumental methods. characterization. XRD is a monoenergetic spec-
The eight listed techniques have established reli- troscopic technique based on Bragg’s Law:
ability to differing measurement levels. Of them,
neutron activation analysis (NAA), a gamma-ray n λ ¼ 2 d sin φ
spectroscopic technique, can lay claim to the
where (n) is the order of wavelengths, (λ) the
crown of the most accuracy and precision at
wavelength, (d ) the interatomic spacing in a
trace element levels of over 70 elements. ICP,
crystal, and (φ) the diffraction angle of X-ray
inductively coupled plasma spectroscopy, is rap-
leaving the mineral’s lattice (Robinson 1995).
idly approaching NAA’s analytical standard and
This is shown in Fig. 8.1. For XRD to work, the
will likely supplant the older technique by virtue
archaeological material must be crystalline. With
of greater availability and cost per analysis.
the exception of glassy materials like obsidian,
The last two parameters—availability and
most minerals are crystalline.
cost—are, likewise, important to the choice of
In XRD analyses, monoenergetic collimated
analytical techn