Article 4 Proceeding2019 Crop

Results of
Crop Improvement
&
Management Research
2018
Edited by
Taye Tadesse
Berhane Lakew
Fasil Tadesse
የኢትዮጵያ የግብርና ምርምር ኢንስቲትዩት

Ethiopian Institute of Agricultural Research
Results of
Crop Improvement and
Management Research
2018
Proceedings of the National Conference on

Crop Improvement and Management Research
10-12 October 2019
Addis Ababa, Ethiopia
Copy Editing and Design:

Elizabeth Baslyos
[i]
©EIAR, 2020
Website: http://www.eiar.gov.et
Tel: +251-11-6462633
+251-11-6-454416
P.O.Box: 2003
Addis Ababa, Ethiopia
ISBN: 978-99944-66-55-9
Citation
Taye Tadesse, Berhane Lakew and Fasil Tadesse (eds.). 2019. Results of Crop Improvement and
Management Research 2018. Proceedings of the National Conference on Crop Improvement
and Management Research. October 10-12, Addis Ababa, Ethiopia
[ii]
Contents
Contents i
Foreword vi
Preface vii
Field Crops
Breeding and Genetics
Evaluation of Food Barley Genotypes for Grain Yield and Agronomic
Traits in the Central Highlands of Ethiopia
Tigist Shiferaw, Thomas Tsige, Adane Chofere , Kefyalew Taye, Wami Hailu,
Berhane Lakew, and Anberber Haile 1
Evaluation of Malt Barley Genotypes for Grain Yield and Malting Quality
Traits in the Central Highlands of Ethiopia
Thomas Tsige, Tigist Shiferaw, Berhane Lakew, Shimles Gezahegn, Kefyalew Taye and Seid Ahmed 9
Stability and Performance Evaluation of Food Barley Varieties for Moisture Stress Environments
Shimeles Gezahegn, Adane Chofere, Workeneh Mekasa,Tesfahun Alemu, and Amberber Haile 21
Stability Analysis of Malt Barley Varieties in the Highlands of Arsi and West
Arsi Zones of Ethiopia
Adane Chofere, Shimeles Gezahegn, Workeneh Mekasa, Amberber Haile, and Tesfahun Alemu 33
Mean Performance and Stability Analysis of Yield and Fiber Quality Traits
of Early Maturing Cotton (Gossypium hirsutum L.) Genotypes in Ethiopia
Merdasa Balcha, Donis Gurmessa, Bedane Gudeta, Samuel Damtew, Arkebe G/Egziabher,
Alehegn Workie and Mekashaw Arega 49
Evaluation and Stability Analysis of Cotton (Gossypium hirsutum L.)

Genotypes for Yield and Fiber Quality Traits in Ethiopia
Merdasa Balcha, Bedane Gudeta, Arkebe G/Egziabher, Donis Gurmessa, Samuel Damtew,
Alehegn Workie and Mekashaw Arega 57
Genetic Diversity Study of G. barbadense L. and Interspecific

(G. hirsutum x G. barbadense) Hybrid Cotton Germplasm using SSR Markers
Donis Gurmessa, Kassahun Bantte and Kefyalew Negisho 67
Evaluation of Yield and Fiber Qualities Performance of Long-staple Cotton

(Gossypium barbadense L.) Genotypes Tested on Multi-locations of Potential
Cotton- Growing Areas of Ethiopia under Irrigation Conditions
Bedane Gudeta, Samuel Damtew, Arkebe G. Egziabher, Donis Gurmessa,
Merdasa Balcha, Alehegn workie and MekashawArega 79
Yield Performance of Faba Bean (Vicia faba L) Genotypes for High

Potential Production Area
Musa J., Asnakech T., Gemechu K., Sisay A, Mesfin T, Nigat T. Wondafrash M., Deressa T, Kedir Y.,
Gezachew Y.,Behaylu A., Jemal E, Danele A., Tadele T., and Nano A 91
Yield Performance of Large-Seeded Faba Bean Genotypes for Potential

Production Areas in Ethiopia
Asnakech T., Musa J., Gemechu K., Sisay A, Mesfin T, Nigat T., Aliy R.,
Kedir Y. Mesfen H., Abebe H., and Fantanesh S. 97
[i]
Performance Evaluations of Field Pea (Pisum sativum L.) (Kiki type) Genotypes
for High Potential Production Area
Asnakech T., Musa J., Gemechu K, Sisay A., Mesfin T, Nigat T., Wondafrash M., Gezachew Y.,
Deressa T, Kedir Y., Aliy R, Jemal E., and Tadele T. 103
Evaluation of Field Pea (Pisum sativum L) Kiki and Shero Yype Genotypes for
High Potential Production Area
Asnakech T., Musa J., Gemechu K., Nigat T., Sisay A., Mesfin T, Wondafrash M, Aliy R.,
Kedir Y, Jemal E, Nano A, and Tadele T. 111
Evaluation of Large Red Bean Genotypes for Lowland Areas of Ethiopia

Abel Moges Firew, Berhanu Amsalu Fenta, Dagmawit Tsegaye, Tigist Shiferaw, Kassaye Negash, Kidane Tumssa,
Behailu Tesfaye, Semir Hashim, Girum Kefle, Yasin Goa, Zeru Yimer, and Clare Mugisha Mukankusi 117
Development of Biofortified Large Red Common Beans for Major Growing Areas of Ethiopia
Berhanu Amsalu Fenta, Dagmawit Tsegaye, Abel Mogos, Tigist Shiferaw, Behailu Tesfaye,
Tarekegn Argaw, Semir Hashim, Girum Kifle, Abush Tesfaye, Ibsa Aleyi, and Solomon Bekele 129
Evaluation of Cowpea Genotypes for Grain Yield Using Additive Main Effects
and Multiplicative Interaction (AMMI) Model
Berhanu Amsalu Fenta, Dagmawit Tsegaye, Tesfaye Walle, Abel Moges Firew, Firew Mekbib,
Tigist Shiferaw, Behailu Tesfaye, Ibsa Aleyi, and Tekle Yoseph 139
Evaluation of Mung Bean Genotypes for Lowland Areas of Ethiopia

Dagmawit Tsegaye, Berhanu Amsalu Fenta, Abel Mogos, Tigist Shiferaw, Kassaye Negash,
Kidane Tumssa, Behailu Tesfaye, Semir Hashim, Girum Kifle, Ziru Yimer, and Ibsa Aliyi 147
Introduction and Evaluation of Quality Protein Maize Hybrid Varieties for

Low Moisture Stress Agro-ecology
Alemeshet Lemma, Lealem Tilahun, Dereje Ayalneh, Bulo Dhabesa,
Talef Wendosen, and Estefanos Habtemariam 161
Participatory Demonstration and Evaluation of Improved Maize Technologies in

Selected districts of West Shewa, East Wollega and Ilu-Ababora Zones of Western Oromia
Effa Wolteji, Bayisa Gedafa and Dubiso Gacheno 173
Green Super Rice (Oryza sativa L.) Variety Development for the Lowland Rice
Ecosystem in Ethiopia
Abebaw Dessie, Taddess Lakew, Zelalem Zewdu, Mulugeta Atnaf, Assaye Berie, Haile Gebriel kinfie,
Mulugeta Bitew, and Hailemariam Solomon 187
Evaluation of Lowland Rice (Oryza sativa L.) Varieties in the Rainfed-Lowland

Rice Producing Areas of Ethiopia
Zelalem Zewdu, Abebaw Dessie, Mulugeta Atinaf, Assaye Berie, Mulugeta Bitew, and Geleta Abdi 199
Statistical Analysis of Multi-Environment Trial Data to Enhance

Genetic Gain in Ethiopian Sorghum Breeding Program
Driba Tadese, Amare Seyoum, Taye Tadesse, Yohannes Fikadu, Amare Nega, Adane Gebreyohannes ,
Alemu Tirfessa, Kidanemaryam Wagaw, Sewmehone Siraw belay, Daniel Nadew,
Hailemariam Solomon, Rebuma Mereraa, and Tamirat Bejiga 207
Tef (Eragrostis tef) Variety Development for Moisture Deficit Areas of Ethiopia
Worku Kebede, Yazachew Genet, Tsion Fikre, Kidist Tolosa, Solomon Chanyalew, Mengistu Demissie,
Kebebew Assefa, Kidu Gebremeskel, Atinkut Fentahun and Zerihun Tadele 221
[ii]
Semi-dwarf Tef (Eragrostis tef) Variety Development for High Potential Areas of Ethiopia
Worku Kebede, Tsion Fikre, Kidist Tolosa,Yazachew Genet, Solomon Chanyalew, Mengistu Demissie,
Kebebew Assefa, Habte Jifar, Nigus Belay, Kidu Gebremeskel,Girma Chemeda, Molalign Assefa,
Atinkut Fentahu and Zerihun Tadele 233
Yield Potential and Stability of Irrigated Spring Bread Wheat Genotypes in

Central Rift Valley of Ethiopia
Mihratu Amanuel Kitil, Tadiyos Bayisa Sarbessa, Hailu Mengistu Biru, Desta Gebre Banje, and Ambesu Tiliye 243
Spatial META Analysis Methods Improve Heritability and Genetic Gain of Early Maturing
Sorghum [Sorghum bicolor (L.) Moench] Genotypes in Dry Lowland Areas of Ethiopia
Amare Seyoum, Amare Nega, Taye Tadesse, Diriba Tadesse, David Jordan, Yilma Kebede,
Adane Gebereyhones, Kedanemaryam Wagaw, Hailemariam Solomon, Alemayheu Assefa, Tamirat Bejiga,
Daniel Nadew, Tokuma Legesse, Alemu Tirfessa, Emma Mace, and Sewmehone Siraw Belay 251
Agronomy and Physiology

Effect of Seed Rate and Row Spacing on Yield and Yield
Components of Rainfed Lowland Rice (Oryza sativa L.) variety
Zelalem Tadesse, Tilahun Tadesse and HabtamuAssega` 265
Determination of NP Fertilizer Requirement for Newly Released

Medium Maturing Maize Varieties at Jimma Zone, Southwestern Ethiopia
Muhidin Biya, Sisay Gurmu, and Eshetu Yadete 279
Effect of NP Fertilizer Rates and Plant Population Density on

Late Maturing Maize Variety
Sisay Gurmu, Muhidin Biya, and Eshetu Yadete 289
Evaluation of QPM Maize Variety under Integrated Uses of Different

Fertilizer Sources in Jimma, South-Western Ethiopia
Sisay Gurmu, Muhidin Biya and Eshetu Yadete 305
Determination of Nitrogen and Phosphorous Fertilizer Rates on Lowland Rice Production

Tilahun Tadesse, Zelalem Tadesse, Habtamu Assega and Desta Abaychew 319
Optimum Nitrogen and Phosphorus Fertilizer Rates for Upland Rice Production in
North-Western Ethiopia
Tilahun Tadesse, Zelalem Tadesse, Habtamu Asega, and Christian Tafere 333
Response of Cotton, (Gossypium hirsutum L.) Varieties to Different Stages of

Topping Under Irrigated Condition, Ethiopia
Yonas Bekele, Arkebe G/egziabher and Bedane Gudeta 347
Coffee, Horticulture, Spices Medicinal and Aromatic Plants

Coffee
Assessment of the Effectiveness of Coffee
De-mucilager and Driers for Physical and Sensorial Coffee Quality
Mikru Tesfa and Nigussie Mekonen 363
[iii]
Vegetable
Performance of Garlic Varieties under Rain-fed Cultivation Practices in
Libokemkem and Fogera Areas, South Gondar Zone, Ethiopia
Dessie Getahun and Mulat Getaneh 371
Influence of Plant Population on Bulb Yield and Yield Components of

Some True-Seed Propagated Shallots
Getachew Tabor, Fasil Tadesse and Yenenesh Asfaw 381
Evaluation of Lettuce (Lactuca sativa) Genotypes for Yield and Quality in Ethiopia
Zebenay Dagne, Samuel Tilahun, Ousman Yimer, Gete Fekadu, Fekadu Gebretensay,
Fasil Tadesse Tewolde, Getachew Tabor and Gizachew Atinafu 389
Determination of Optimum Plant Population for Onion (Allium cepa var. cepa L.)
Seed Yield and Quality in Central Rift Valley of Ethiopia
Jibicho Geleto, Tesfa Binalfew, Fekadu Gebretensay Mengistu, Nigissie Chewaka, Shimeles Aklilu, Yosef Alemu,
Selamawit Ketema, Melkamu Hinsermu, Gebeyehu Wondimu, and Birhan Abera 397
Medicinal and Aromatic Plants

Effect of Plant Spacing and Harvesting Age on Growth, Yield and Yield Components
of Lavender (Lavandula angustifolia L.) under Supplemental Irrigation at
Wondo Genet, Southern Ethiopia
Basazinew Degu and Sulti Amano 415
Effect of Harvesting Age and Plant Spacing on Growth, Yield and Yield Component of
Lavender (Lavandula angustifolia L.) under Rainfed Condition at Hawassa, Southern Ethiopia
Basazinew Degu and Sulti Amano 427
Determination of Optimum Spacing and Harvesting Age of Citron Scented Gum (Eucalyptus
citriodora H.) for Maximum Biomass and Essential Oil Yield at Wondo Genet, South Ethiopia
Desta Fikadu Egata and Wondimkun Dikir Desta 439
Enhanced Income Generation through Sweet Basil (Ocimum basilicum L.)

and Maize (Zea mays L.) Intercropping at Wondo Genet, South Ethiopia
Nebret Tadesse, Midhekisa Chala and Basazinew Degu 449
Potato
Bridging Attainable Yield Potentials of Potato through Nutrient
Management in Northwestern Ethiopia
Tadele Amare, Zerfu Bazie, Erkihun Alemu, Amlaku Alemayehu, AbereTenagne, Bitewlgn Kerebih,
Yasin Taye, Assefa Derebe and Tesfaye Feyissa 459
Evaluation Performance of Potato Clones in the Mid-highland Areas of Ethiopia

Shamil Alo, Engida Tsegaye, Abirham Negara and Alemu Worku 469
Tuber Yield Performance and Stability of Potato Genotypes in the Highlands of Ethiopia
Baye Berihun, Animut Tarik, Alemu Worku, Manamino Workayehu, Melkamu Eneyew, Abebe Chindi,
Tesfaye Abebe, Zerihun Kebede, Fekadu G/Tensay, Wubet Awoke and Gebre Alamir 475
Adaptation of Korean Potato (Solanum tuberosum L.) Varieties under Irrigation

Abebe Chindi, Kasaye Negash, Egata Shunka, Gebremedhin W/Giorgis, Tesfaye Abebe,
Fikadu Gebretinsay, Niguse Abebe, Wasu Mohammed and Zerihun Kebede 485
Collection and Characterization of Local Potato Cultivars from Eastern and Western Ethiopia
Semagn Asredie, Alemu Worku, Manamino Workayehu, and Egeta Shunka 495
[iv]
Root and Tuber Crops
Performance and Stability Study of Newly Bred Orange Fleshed Sweetpotato
Genotypes in Ethiopia
Fekadu Gurmu, Bililign Mekonen, Yitages Kuma 503
Verification of Introduced Early Maturing Orange Fleshed Sweetpotato Varieties for

Registration in Ethiopia
Fekadu Gurmu, Bililign Mekonen, Yitages Kuma, Shiferaw Mekonen and Zeritu Shashego 515
Participatory Orange-fleshed Sweetpotato Variety Selection in Ethiopia:

Integrating Farmers’ and Researchers’ Criteria for Variety Selection
Fekadu Gurmu, Abiyot Aragaw, Roland Brouwer, Bililign Mekonen and Zeritu Shashego 523
Effects of Soil Amendments on Growth and Biomass Yield of Early Generation

Seeds of Sweetpotato (Ipomoea batatas (L.) Lam) Grown in Net Tunnels
Daniel Markos and Fekadu Gurmu 533
Evaluation of Cassava (Manihot esculenta Crantz) Genotypes for Total

Cyanide Content, Storage Tuber and Starch Yield in South Western Ethiopia
Tewodros Mulualem, Neim Semman and Getachew Etana 551
Spices
Chemical Composition of Essential Oils from Released Coriander Variety
(Coriandrum sativum L.) Grown In Ethiopia
Biruk Hirko, Sileshi Abera, Haimanot, Mitiku and Belay Gezahagn 561
Evaluation of Improved Turmeric Boiler at Different Boiling Durations

Biruk Hirko, Haimanot Mitiku, Sileshi Abera, Abukiya Getu and Belay Gezahagn 569
Evaluation of the Effects of Different Curing and Drying Methods on

Quality of Black Pepper (Piper nigruml.) in Southwestern Ethiopia
Biruk Hirko, Haimanot Mitiku, Sileshi Abera and Abukiya Getu 577
Fruit Crops
Performance of Commercial Avocado Varieties in Raya Azebo,
Southern Zone of Tigray Region, Ethiopia
Haile Abebe, Wakuma Biratu, Kidane Tesfay, Molla Berhe, and Haileselassie Gebremeske1 585
Evaluation of Table Grape Varieties for Yield and Quality Traits

Muluken Demelie, Fasil Tadesse Tewolde, Gadisa Lellisa and Etsegent Keflu 597
[v]
Foreword
Agriculture plays a significant role in the Ethiopian economy. The sector employs the majority of
the workforce, supplies food for the ever-increasing human population and feed for livestock
production. At macro level the largest share of national GDP is sourced from Agriculture. The
country is endowed with huge genetic recourses of crops, livestock, microbial and other forms of
fauna and flora that would serve as building blocks for propelling genetic improvement and
advancement of the agriculture sector in general. Despite this enormous potential coupled with
long tradition of crop production and animal husbandry and bounty of ecological diversity,
Ethiopia is still grappling with the challenges of alleviating vulnerability of the agricultural, agro-
pastoral and pastoral communities. It is evident that considerable numbers of the populations are
exposed to food and nutritional insecurity and with eminent degradation of the natural resources.
Ethiopian agriculture is epitomized as a predominantly low-input low-output system dominated by
smallholder producers generating agricultural products that are far less surplus, competitive,
diverse and sustainable than would be required for the sector and the country thereof to appease
relentlessly spiking food and feed gaps. Contrary to this, Ethiopia is now on the verge of
completing the second phase of its five-year growth and transformation plan (GTP-II) which was
launched in 2015/16. The core determination of the plan is to perpetuate the growth and
development trajectory attained during the previous phase and thereby ensuring: food security and
sovereignty, sustainable supply of raw materials for agro-industries and import substation,
expanding the base for foreign earnings from agricultural exports, and increasing livelihood
resilience and environmental sustainability.
The strategic direction envisaged for attaining the above objectives is by increasing agricultural
production, productivity, quality and market linkages through enhanced adoption and scaling up of
agricultural technologies, implementation of cluster-based agriculture, input supply and
participation of the private sector. Central to the success of the plan is the view that productivity of
crops would be increased from the base year’s 19.5 q/ha to 30.9 for cereals, 16.4 q/ha to 24.0 for
pulses, from 9.0 q/ha to 12.7 for oilseeds, and from 7.0 q/ha to 11.1 for coffee. Accordingly, the
Ethiopian Institute of Agricultural Research (EIAR), guided by its vision and mission, is striving to
contribute to the fulfillment of these fundamental national objectives by undertaking research and
generating appropriate agricultural technologies, information and knowledge on a host of crop
species comprising: field crops, horticultural crops, coffee, tea, root and tuber crops, spices and
medicinal and aromatic crops.
Quite a number of results and outputs are generated every year by the crops research programs
organized under the Crops Research Directorate of EIAR. However, before the research results that
are derived from completed projects and experiments reach the end users, they need to be
reviewed, evaluated and synthesized for their merits of relevance to bring about positive changes in
crop productivity. This volume of publication, therefore, is dedicated to present research results
obtained from experiments carried out during the past three years under the different national
programs of the directorate. It is my firm belief that the materials contained in the proceedings will
serve a broad spectrum of users engaged in developing crop production packages, guidelines and
manuals as well as those engaged in research and academia. Finally, I would like to thank the
authors, editors and the publication staff of the institute for their commendable contributions in
producing the proceedings.
Diriba Geleti (PhD)

Deputy Director General for Research, EIAR
[vi]
Preface
Agriculture is the backbone of Ethiopian economy contributing to food self-sufficiency, foreign
currency earning, supplying raw materials for the emerging agro-processing industries while
conserving the ecosystem for sustainability. In spite of the increasing production in the past two
decades, the gap between demand and supply remains to be a challenge for the sector to feed the
increasing human population and contribution to economic growth. The crop research directorate
undertaking research projects designed to generate technologies and information contributing to
increased production and productivity. This proceeding is the result of the past three years (2017 to
2019 crop seasons) research undertakings of the national programs. The research activities were
mainly supported by the government and AGP II project which has supported agronomic research
and research activities reached the final stage of evaluation. In the execution of the research
activities, the federal and regional research centers and some of the universities have been
participated.
The papers are the outcome of the completed review forum which was held from Oct 10 to 12,
2019 at EIAR-HQ. In this workshop, senior scientists representing the different programs, sectors
and partner institutions were in attendance. Two stages review, at the program level and by
assigning senior researchers from the institutions, have been conducted prior to presenting the
papers in the workshop. The papers included in this proceeding were selected based on the
contributions to generate appropriate technologies for users, scientific merits and contribution to
the advancement of scientific research in the country. It is believed that the document will give an
opportunity for both technology users and researchers to get information on the research activities
and major achievements for the future research and development endeavors.
The proceeding is mainly consisted of papers on plant breeding, agronomy, on-farm demonstration
of improved technologies and post-harvest related issues. The papers organized into different
sections as field crops, horticulture, root and tubers, spices and coffee and tea crops. The
proceeding contained 55 reviewed articles from the different disciplines. The respective program
coordinators and the team members have exerted maximum efforts in the execution of the research
activities to the quality of scientific standards and were participated in writing and edition the
papers within programs. The authors are recognized for their commitment to writing the papers as
per the standard set initially and incorporation of the comments given by the reviewers for the
betterment of the quality of the papers. The contribution of the senior researchers across canters in
reviewing the papers was immense and the editors would like to thank those who have been
involved in the review process. The final edition was done by W/o Elizabeth Baslyos and would
like to appreciate for her support given to make the proceeding to the standard. We believe that the
papers contained in this document would provide useful information for the scientific community
and other end users.
The Editors
Taye Tadesse (PhD)
Berhane Lakew (PhD)
Fasil Tadesse (PhD)
[vii]
Field Crops
Breeding and Genetics
Evaluation of Food Barley Genotypes for Grain

Yield and Agronomic Traits in the Central
Highlands of Ethiopia
1Tigist Shiferaw,1 Thomas Tsige, 2Adane Chofere , 3Kefyalew Taye,
1Wami Hailu, 1Berhane Lakew, and 2Anberber Haile
1
Holetta Agricultural Research Center, P.O. Box 31, Holetta, Ethiopia; 2Kulumsa Agricultural Research
Center, P.O. Box 489, Assela, Ethiopia; 3Debreberhan Agricultural Research Center, P.O. Box 112,
Debre Berhan, Ethiopia; Corresponding author: thomas.tsige@yahoo.com
Abstract
The present study was undertaken to evaluate the performance of promising food
barley genotypes for grain yield and yield-related traits. The trial was conducted in
the 2017 and 2018 main cropping season using a randomized complete block design
with three replications. Variance analysis and GGE biplot were used to understand
the nature of genotype × environment interaction (G × E) in a grain yield data
collected from eighteen barley genotypes grown in eight environments (Location and
year combinations). The combined analysis of variance (ANOVA) showed
significantly higher genotype, environment and genotype by environment interaction
effects for all the traits studied. Accordingly, genotypes EH1493 X HB1307 (G10) and
HB 1307 X ND25160 (G2) showed the highest mean grain yield of 4558 kg ha -1 and
4499 kg ha-1, respectively. GGE biplot showed that G10 was the winner genotype at
BK18, DB18 and HO18 environments and it has good grain yield stability across the
testing environments. Therefore, G10 is a potential candidate variety to be included in
the variety verification trial for possible release.
Keywords: ANOVA, GGE biplot, Stability, Grain yield
Introduction
Barley (Hordeum vulgare L) is an important staple food crop and has high
potential in narrowing food deficit and enables to achieve food security in
Ethiopia. It sustains the livelihood of millions of people residing on the highlands
and fetches a substantial income for farmers. It is widely grown in diverse rain-
fed agro-ecological zones of Ethiopia at an altitude of 1400 to over 3600 m.a.s.l.
The crop is more diversified and prominent in areas between 2300 to 3400 m.a.s.l
(Zemede, 2002). Food barley is commonly cultivated in stressed areas where soil
[1]
erosion, soil acidity, occasional drought or frost limits the choice of other crops. It
is cultivated in different production systems, namely; early, late, Belg and residual
moisture (Chilot et al., 1998).
Barley is the fifth important cereal crops after maize, tef, wheat and sorghum. It is
produced on about 1 million hectares of land from which 2.0 million tons of grain
are produced annually (CSA, 2018). The productivity of barley in Ethiopia is low
(2.1 t/ha) as compared to the world average of 3.1 t/ha. The low productivity is
mainly attributed to abiotic stresses (soil acidity, frost, waterlogging, low moisture
and low soil fertility) and biotic stresses (diseases, pests and weeds).
In general, barley remains an important crop in the Ethiopian agriculture because

of its role in providing food for the rapidly growing population (3.0 % per year).
Therefore, there is a need to focus on barley improvement and developing
alternative varieties to the different production systems and agro-ecologies.
Variety development effort is a dynamic process as one breeding program is
required to provide a large option of varieties for diverse environmental
conditions. Moreover, available varieties become obsolete due to the break down
of disease resistance and poor performance. Hence, it is crucial to provide new
improved varieties that could go with the improved production packages identified
continuously. Increasing the productivity of food barley is very important for the
resource-poor smallholder farmers in Ethiopia to improve the output and income
to satisfy the local food demand of the rapidly growing population.
The National Barley Research program at Holetta and Kulumsa in collaboration

with D/Berhane regional research center has conducted a multi-location variety
trial, including lines developed from local collections, local crosses and
introductions with the main objective of identifying stable and superior varieties to
be released for production and use in the future breeding.
Materials and Methods

A total of eighteen food barley genotypes including three check varieties were
evaluated using a completely randomized block design with three replications.
The experiment was executed at Holetta, Jeldu, Debrebirhan, Bekoji and Kofole
during the main cropping season in 2017 and 2018 under rain-fed conditions.
Descriptions of the testing sites are presented in Table 1. The materials were
evaluated for eight quantitative and qualitative traits such as DF, DHE, DMA,
PLH, SC, NB, TKW, HLW, GYLD at eight locations across the two years (Table
2). The traits studied were days to heading (days), days to maturity (days), plant
height (cm), thousand kernel weight (gm), hectoliter weight (Kg hl-1), grain yield
[2]
(Kg ha-1), scald and net blotch disease severity (%). A plot size of 1.2 m by 2.5 m
was used to layout the experiment and the spacing between blocks and plots
within blocks were 1.5 m and 0.4m, respectively. Analysis of variance (ANOVA)
was carried out using R-software to determine the effect of genotype, environment
and their interaction on various traits of the food barley genotypes. Linear mixed-
effect model (lmer) of package lme4 was used for data analysis and the
environments were considered as random and genotypes as fixed effects (R Core
Team, 2017). The following ANOVA models have been used to test the
performance of genotypes at each and combined locations, respectively (Singh
and Ceccarelli 1996).
Yij = µ + Gi +Bj + eij and Yijk = µ + Gi +Ej + GEij +Bk (j) + eijk.
Where, Yij = observed value of genotype i in block j, Yijk = observed value of
genotype i in block k of environment j, µ = grand mean of the experiment, Gi =
the effect of genotype i, Bj = the effect of block j, Bk(j) = the effect of block k in
environment j, eij = error effect of genotype i in block j , Ej = environment effect,
GEij = the interaction effect of genotype i with environment j, eijk = error
(residual) effect of genotype i in block k of environment j. Also, GGE bi-plots
were performed to determine the stability of the tested food barley genotypes
using GGEBiplot GUI package of R-software (R Core Team, 2017).
Table 1. Description of testing sites
No. Testing sites Altitude (m) Rainfall (mm) Longitude Latitude

1 Holetta 2400 1100 38°38’E 9°00’N
2 Jeldu 2800 1200 38°03’E 9°17’N
3 D/Berhane 2830 932 39°32’E 9°41’N
4 Bekoji 2810 1082 39°15'E 7°15'N
5 Kofole 2700 1232 38° 45' E 7° 00 N
Table 2. Lists of genotypes and environments used for the experiment

Entry Genotype Location Year Environment code
G1 EH 1493 x PPB 25 Bekoji 2017 BK17
G2 HB 1307 x ND 25160 Bekoji 2018 BK18
G3 HB 1493 x ND 25160 Kofele 2017 KF17
G4 EH 1493 x HB 42 Kofele 2018 KF18
G5 Cross # 41/98 x HB 42 Holetta 2017 HO17
G6 EH 1493 x Cross 41/98 Holetta 2018 HO18
G7 ARDU 12-60B x HB 1307 Jeldu 2018 JL18
G8 HB 1307 x PPB 25 Debreberhane 2018 DB18
G9 HB 1307 x Cross # 41/98
G10 EH 1493 x HB 1307
G11 Estayesh x Lab # 87
G12 Cross # 41/98 X HB 1307
G13 Tolese x HB 1307
G14 HB 1307 X PPB 53
G15 Shege X Cross #41/98
G16 HB 1965
G17 HB 1966
G18 Local check
[3]
Results and Discussion
Combined analysis of variance showed that genotypes, environment and genotype

x environment interaction had a highly significant effect for all the traits studied
(Table 3). The significant genotypes effect showed the existence of significant
variability among the tested genotypes. Highly significant genotype x environment
interaction effect observed in this trial showed the tested genotypes performed
differently across the testing environments for the traits. Therefore, researches
involved in the barley improvement program should have to see the adaptability of
genotypes to different test environments. Similarly, significant differences among
the testing environments also observed. In the present study all traits showed a
higher contribution of the environmental component to the total sums of squares
and similar results were reported by Abtew et al., (2015).
Table 3. Mean square values for combined analysis of 18 food barley genotypes
DF DHE DMA PLH SC (DF)§ NB (DF)§ TKW HLW GYLD
Gen 17 89.6** 96.9** 174** 602(17)** 131(17)** 151.5** 26.5** 1448234**
Env 7 2206** 8145** 5102** 22733(5)** 31876(5)** 2104** 1219** 23019573**
Gen:env 119 11.5 ** 22.0 ** 51 ** 250(85) ** 89(85)** 19.2 ** 7.3* 1443851**
Env:rep 16 8.3 13.7 155 383(12)** 154(12)** 10.4 7.9 2059673**
Residuals 272 7.5 8.8 33 138(204) 56(204) 9.9 5.5 657602
CV 3.3 2.1 5.1 36.1 21.1 7.5 3.7 19.4
Mean 82.7 140.8 113.5 32.5 35.6 42.1 64.1 4190.8
DF=degree of freedom, DHE=days to heading, DMA= days to maturity, PLH=plant height, SC=scaled,
NB=net blotch, TKW= thousand kernel weight, HLW= hectoliter weight, GYLD= grain yield, **, *
Significant at 5% and 1% probability level, ns=non significant, §these traits were

not recorded at DB18 and JL18 mean squares under those traits are angular
transformed values Based on the average data over eight environments, G10 (4558
kg ha-1) and G2 (4499 kg ha-1) scored the highest mean grain yield, though not
significantly different from the recently released standard check varieties (HB1965
and HB 1966). Accordingly, G10 and G2 had 9.75 and 8.33% grain yield advantage
over the highest yielding check variety (HB 1966) (Table 4). The highest mean
hectoliter weight (HLW) and thousand kernel weight (TKW) values were scored by
G7. The two high grain yielding genotypes (G10 and G2) also had higher HLW and
moderate TKW values. Regarding disease-related traits, most test genotypes
showed moderate tolerance to scald and net blotch disease. The check variety HB
1965 and 1966 scored 21.2 and 26.7% severity for scald and 33.3% and 32.8%
severity for net blotch. Similarly, the candidate genotypes G10 and G2 also showed
moderate scald (36.2, 29.0%) and net blotch (39.0, 37.2%) resistance, respectively.
The ranges of plant height values of 106-115 cm were recorded with the lower end
corresponding check variety HB 1965 while the upper end corresponds to the local
check variety, this longer plant height is in agreement with the inherent nature of
[4]
most local varieties. Furthermore, the phenological traits showed lower variability
with, five and seven days differences in mean days to heading and maturity
respectively (Table 4). Based on agronomic and grain physical quality traits (HLW
and TKW), G10 is the top-ranking genotype. Thus, G10 and the other two
promising varieties, namely G2 and G7 can be used for future breeding programs as
donor parents.
Table 4. Means from the combined analysis of variance on 8 traits of 18 food barley genotypes tested at 8 environments.
Trt# Genotype DHE DMA PLH SC§ NB§ TKW HLW GYLD
G1 EH 1493 x PPB 25 80.9de 144.3a 109.7c-f 30.0cde 35.5b 45.6ab 64.6bc 4037abc
G2 HB 1307 x ND 25160 82.4a-e 142.0abc 109.8c-f 29.0cde 37.2ab 43.0b-e 64.7abc 4499a
G3 HB 1493 x ND 25160 84.8 abc 143.7 a 116.1a 26.2 de 36.8 ab 45.6 ab 63.8 bcd 4103abc
G4 EH 1493 x HB 42 80.7e 141.2a-d 111.3a-e 31.2b-e 32.8b 44.0a-d 64.5bc 4331abc
G5 Cross # 41/98 x HB 42 85.2ab 144.8a 113.6abc 28.4cde 35.5b 43.9a-d 64.4bc 4270abc
G6 EH 1493 x Cross 41/98 82.8a-e 141.5a-d 113.2abc 35.6a-d 36.2b 43.7bcd 64.3bc 4262abc
G7 ARDU 12-60B x HB 1307 80.2e 142.4ab 113.1a-d 35.7a-d 35.6b 47.9a 66.7a 3693bc
G8 HB 1307 x PPB 25 82.1 b-e 143.1 ab 110.9b-e 32.9 bcd 34.5 b 43.8 bcd 64.5 bc 4326abc
G9 HB 1307 x Cross # 41/98 83.9 a-d 142.3 ab 108.3def 29.5 cde 34.5 b 42.7 b-e 64.9 abc 4219abc
G10 EH 1493 x HB 1307 82.1cde 141.9a-d 109.9c-f 36.2a-d 39.0ab 42.4b-e 65.5ab 4558a
G11 Estayesh x Lab # 87 85.5 a 143.6 a 113.2a-d 44.5 a 32.2 b 40.0 de 64.4 bc 4408ab
G12 Cross # 41/98 X HB 1307 80.6e 139.4bcd 110.4b-f 36.7a-d 34.5b 41.4cde 65.0abc 4279abc
G13 Tolese x HB 1307 82.2b-e 138.cd 111.2a-e 31.2b-e 38.4ab 39.4e 64.6bc 4101abc
G14 HB 1307 X PPB 53 83.3a-e 141.4a-d 107.2ef 38.4abc 34.5b 41.5cde 65.3ab 4153abc
G15 Shege X Cross #41/98 84.3abc 143.3ab 111.2a-e 41.7ab 34.4b 40.9de 65.0abc 4357abc
G16 HB 1965 80.2 e 138.0 d 105.6f 21.2 e 33.3 b 40.1 de 62.0 d 4035abc
G17 HB 1966 82.1b-e 143.9a 109.2c-f 26.7de 32.8b 44.1abcd 64.0bcd 4153abc
G18 Local check 85.0abc 142.7ab 114.9ab 30.0cde 43.4a 45.1abc 63.1cd 3659c
DHE=days to heading, DMA= days to maturity, PLH=plant height (cm), SC=scaled (%), NB=net blotch (%), TKW=
thousand kernel weight (g), HLW= hectoliter weight (Kg hl-1), GYLD= grain yield (Kg ha-1), §these traits were not recorded
at DB18 and JL18
Separate analysis for each of the test environments showed consistently large
variation among test varieties for grain yield (Table 5). G10 and G2 gave
consistently the highest yield in all testing environments exceeding the check
varieties. These genotypes scored mean grain yield ranged from 3628-5440 kg ha-1
and 3577-5389 kg ha-1, respectively, and this showed both genotypes could have
the potential to be selected for variety verification trial (Table 5). The two
standard check genotypes HB 1965 and HB1966 had relatively lower grain yield
than the candidate genotypes across test environments. Besides G7 and local
check G18 recorded lower mean values in all locations.
[5]
Table 5. Mean performance of grain yield (Kg ha-1) of food barley genotypes tested at eight environments
Trt# Genotypes BK17 HO17 KF17 BK18 DB18 HO18 JL18 KF18 Mean
G1 EH 1493 x PPB 25 4906 3300 3094 4437 3906 4611 4438 3634 4037
G2 HB 1307 x ND 25160 5389 3783 3577 4920 4389 5094 4921 4117 4499
G3 HB 1493 x ND 25160 4957 3350 3145 4488 3956 4661 4488 3685 4103
G4 EH 1493 x HB 42 5214 3608 3403 4745 4214 4919 4746 3943 4331
G5 Cross # 41/98 x HB 42 5133 3526 3321 4664 4132 4837 4665 3861 4270
G6 EH 1493 x Cross 41/98 5096 3489 3284 4627 4095 4800 4628 3824 4262
G7 ARDU 12-60B x HB 1307 4540 2934 2728 4071 3539 4244 4072 3268 3693
G8 HB 1307 x PPB 25 5197 3590 3385 4728 4196 4901 4729 3925 4326
G9 HB 1307 x Cross # 41/98 5069 3462 3257 4600 4068 4773 4601 3797 4219
G10 EH 1493 x HB 1307 5440 3833 3628 4971 4439 5144 4972 4168 4558
G11 Estayesh x Lab # 87 5250 3644 3439 4782 4250 4955 4782 3979 4408
G12 Cross # 41/98 X HB 1307 5142 3536 3331 4673 4142 4847 4674 3871 4279
G13 Tolese x HB 1307 4965 3359 3154 4496 3965 4670 4497 3694 4101
G14 HB 1307 X PPB 53 5005 3399 3193 4536 4004 4709 4537 3733 4153
G15 Shege X Cross #41/98 5195 3589 3383 4726 4194 4899 4727 3923 4357
G16 HB 1965 4908 3301 3096 4439 3907 4612 4440 3636 4035
G17 HB 1966 5026 3419 3214 4557 4025 4730 4558 3754 4153
G18 Local check 4463 2857 2651 3994 3463 4168 3995 3192 3659
Stability of the test genotypes across environments

The discussion in the previous section based mostly on genotype main effect (G)
and genotype-by-environment interactions (GE) are treated as noise or a
confounding factor (Yan, 2006). But recently various methods developed to study
GxE interaction and determine whether or not a genotype is stable in performance
over a range of environments. The GGE biplot method consists of a set of biplot
interpretation methods to evaluate genotype and test-environment (Yan, 2007).
The “Mean vs stability” view of GGE biplot helps to evaluate genotypes based on
their mean performance and stability (Figure 1). The graph consists of the x-axis
passes through the biplot origin and the marker of the average environment, which
is defined by the average PC1 and PC2 scores of overall environments. The
stability and average performance of the genotypes were measured by their
projection to the Y- axis and X-axis, respectively (Yan, 2001). Therefore, G10 had
both the highest average yield and stability. Then G-2 scored the second-highest
mean grain yield performance but relatively it lacks stability and more adapted to
specific environments such as at HO18. Similarly, G11 (Estayesh x Lab # 87) had
better mean grain yield and more adapted to environment JL18. Among the check
varieties, G17 (HB 1966) revealed moderate mean grain yield and good stability.
The local check variety G18 had the lowest average yield and the least stable
variety (Figure 1).
The “which-won-where” pattern of a genotype by environment data set is an

important feature of GGE biplot. The Biplot contains a polygon drawn on
genotypes that are furthest from the biplot origin so that all other genotypes are
[6]
contained within the polygon and these genotypes located on the vertices of the
polygon performed either the best or the poorest in one or more environments
(Yan 2006). The “which-won-where” view of the GGE biplot is an effective
visual tool in the mega-environment analysis (Yan et al., 2007). G10 was the
winner in environments DB18, BK18 and HO17. G2 was more adapted to the
testing environment HO18. G11 was the highest yielding genotype in
environments JL18. The check genotype G17 (HB1966) was a vertex genotypes
with no testing location specifically adapted to it. The local check variety (G18)
was a relatively low yielding genotype in all environments (Figure 2).
Figure 1. Mean grain yield performance and stability of genotypes based on the G × E data
Figure 2. The which-won-where view of the GGE biplot of grain yield of Food barley genotypes based on the
G × E data
[7]
Conclusion
In the present study, the eighteen food barley genotypes showed significant
genetic, environmental and genotype by environment interaction effects for all
traits considered in the experiment. Genotypes, G10 (EH 1493 x HB 1307) and G2
(HB 1307 x ND 25160) showed the highest mean grain yield potential at all
environments. These genotypes showed moderately resistant to scald and net
blotch, and acceptable TKW and HLW values. Based on the “Mean vs stability”
view of GGE biplot, G10 was the highest yielding and stable genotype. While G2
was the second-highest yielding genotypes but it is relatively unstable. Similarly
“Which won where” pattern of GGE biplot confirmed that G10 gave consistently
highest mean grain yield across the test environments and G2 specifically adapted
to the HO18 environment. Therefore, based on mean performance and stability,
G10 is the best potential variety identified for possible variety verification trials
for the mega and other similar environments. Besides, G10 and the other two
Genotypes G2 and G7 will be included in the food barley crossing program as
potential parents for their good yield potential and physical grain quality traits.
References
Abtew, Wosene G, Berhane Lakew, Bettina IG, Haussmann, Karl J Schmid. 2015. Ethiopian
Barley landraces show higher yield stability and comparable yield to improved varieties in multi-
environment field trials. Journal of Plant Breeding and Crop Science. 7(8):275-29.
Chilot Yirga, Fekadu Alemayehu and Woldeyesus Sinebo (Eds). 1998. Barley-based Farming
Systems in the Highlands of Ethiopia. Ethiopian Agricultural Research Organization. Addis
Ababa, Ethiopia.
CSA. 2018. Area and Production of Major Crops, Agricultural Sample Survey, 2017/18 (2010
E.C.). Central Statistics Agency, Statistical Bulletin 586. Addis Abeba, Ethiopia
Singh, RK and Chaudhury, BD. 1996. Biometrical Method in Quantitative Genetic Analysis.
Kalyani, Ludhiana.
R Core Team. 2017. R: A language and environment for statistical computing. R Foundation for
Statistical Computing, Vienna, Austria.URL http://www.R-project.org/.
Yan, W. 2001. GGEbiplot—A Windows Application for Graphical Analysis of Multi-Environment
Trial Data and Other Types of Two-Way Data. Agronomy journal, 93:1111-1118.
Yan, W and NA Tinker. 2006. Biplot analysis of multi-environment trial data: Principles and
Applications. Canadian Journal of plant science, 86: 623-645.
Yan W, Kang MS, Ma B, Woods S and Cornelius PL. 2007. GGE biplot vs. AMMI analysis of
genotype-by-environment data. Crop Science, 47: 643-653.
Zemede, A. 2002.The Barley of Ethiopia. In: Stephen, B.B (ed.). Genes in the Field. On-farm
Conservation of Crop Diversity. Lewis Publisher, Boca Raton, Pp 77-107.
[8]
Evaluation of Malt Barley Genotypes for Grain
Yield and Malting Quality Traits in
the Central Highlands of Ethiopia
1Thomas Tsige, 1Tigist Shiferaw, 1Berhane Lakew 1, 2
Shimles Gezahegn, 3Kefyalew Taye and 1Seid Ahmed
1
Holetta Agricultural Research Center, P.O. Box 31, Holetta, Ethiopia
2
Kulumsa Agricultural Research Center, P.O. Box 489, Assela, Ethiopia
3
Debreberhan Agricultural Research Center, P.O. Box 112, Debreberhan, Ethiopia
Corresponding author: thomas.tsige@yahoo.com
Abstract
A multi-location variety trial was conducted to identify suitable malt barley varieties
that satisfy the malt and brewing industry quality requirements and reduce the cost for
importing malt barley. The trial was conducted to evaluate the performance of 28
promising malt barley varieties in seven environments (site-season combinations)
during the 2017 and 2018 cropping seasons. The phenological and agronomic data
collected were subjected to analysis of variance (ANOVA) using ‘R’ software. All the
traits analysis showed significant genotype, environment and genotype by environment
interaction effects. Genotype, HB 52 X Bahati (G 5) exhibited high mean grain yield
(5128 kg ha-1) and significantly different for grain yield from one of the improved checks
(Traveller). The other promising genotype, Bekoji-1 X Grace (G-8) showed acceptable
malt quality results for extract (81.8 %), protein (10.0 %), friability (85 %) and lower
values for beta-glucann (369 mg/l) with comparable grain yield to the improved checks
(Traveller, HB1963). Similarly, genotype Sabin X Beka (G7) showed acceptable malt
quality results with a lower level of beta-glucan (287mg/l). The “which-won-where”
and “Mean vs. Stability” view of GGE biplot showed that G5 exhibited high mean grain
yield and moderate grain yield stability and is the winner genotype in all test
environments. Overall, G8 is identified as potential malting barley candidate varieties
to be tested for more industrial malt quality traits before variety verification trial and
G7 and G5 can be considered as a potential parent in the malt barley crossing program
for their good malt quality traits and yield potential
Keywords: ANOVA, GGE biplot, Malt barley, Grain yield, Malt quality, and
Stability
Introduction
Barley is one of the most important cereal crops widely grown in the highlands of
Ethiopia with an annual production of about 2.0 million tons cultivated on an area
of about 1 million hectares of land with an average national yield of 2.16 tons/ha
(CSA, 2018). In the barley-based farming systems of the central highlands,
smallholder farmers have very few alternative crops. One source of income could
[9]
be growing malting barley, which has dependable local buyers in the country
(Bayeh and Berhane, 2011). Both food and malt barley are grown side by side
sharing similar agro-ecologies. However, the share of malt barley is roughly 15-20
% of the total barley production, which is the major input for beer production
(Berhane et al., 2016). Traditionally, both six-row and two-row barley types are
cultivated in the country, but the best malt quality beer is produced from two-row
varieties.
Malt barley is a high-opportunity cash crop, with great room for profitable
expansion, particularly when connected to growing breweries. However, there is a
shortage of quality malt barley varieties to meet the demand of the local breweries
that forced the malt factories to import a large quantity of malt barley from
abroad. The gap between domestic supply and demand indicates an opportunity to
enhance local production and substitute imports through huge untapped malt
barley potential in the country. Malt imports have grown tremendously reaching
over 75 thousand tons in 2017 covering about 70% of total annual demand and
costing the country about 41.5 million USD (ERCA, 2017).
Malt barley production has not expanded as expected, despite the potential of the
country to grow malting barley both in quality and quantity. There is a relatively
huge domestic market for malt of reasonable quality, where a large number of
farmers in the highlands of Ethiopia can commit part of their barley area to malt
barley production. Even though barley grows in many highland regions of the
country, the adoption of the malt barley varieties is limited to the Arsi highlands
and a lesser extent in Bale where farmers can sale their produce to the Asela malt
factory and the emerging brewery companies such as Heineken, Diageo, and
Dashen. The lack of adoption of malt barley to other highland barley producing
areas is due to limited extension activities by the ministry of agriculture and
relevant organizations. Currently, there is an attempt to promote malt barley
production in the central highlands of Oromia and Amhara region to provide malt
to breweries through contractual production. Therefore, improving the knowledge
and skill of farmers through demonstrating new malt barley varieties would be
vital to increase production and productivity to fill the existing supply gap in the
country. In addition, the quality demand from the breweries and the malt factory is
a benchmark for the malt barley breeding program. The National Agricultural
Research System (NARS), has developed more than 15 malt barley varieties over
the last three decades with the collaboration of the national research system and
international organizations. However, most of the nationally released malt barley
varieties did not satisfy all the malt quality requirements of the industry that
showed the importance of a targeted breeding program to develop and release malt
barley varieties that satisfy the quality parameters of the breweries and the malt
factories.
[10]
This paper presents the results of the malt barley multi-location variety trials
conducted in seven environments. The target was to select high yielding,
agronomically superior varieties with good malting quality traits, disease and pest
resistance and promote best performing varieties for release or incorporate in the
breeding pipeline development as recipients/donors of useful genes.

Twenty-eight malt barley genotypes were evaluated in two sets of experiments in
2017 and 2018 cropping seasons and from these genotypes, twelve of them were
evaluated repeatedly in both years. However, the remaining sixteen genotypes
appeared once in each year (Table 1). Therefore, twenty materials were tested
each year using RCBD design with three replications. The experiments were
carried out at Holetta (9°00'N, 38°38'E, elevation 2400m), Bekoji (7° 15'N,
39°15'E, elevation 2830m), Kofele (7°00' N, 38°45' E, elevation 2700m) and
Debrebrhane (9°41’N, 39°32’E, elevation 2800m) experimental sites in a non-
orthogonal set of six environments (site-season combinations). Twenty-two of the
experimental materials were selected from 2016 and 2017 malt barley preliminary
variety trials and the other six genotypes were included as checks (Table 1).
Data were recorded on the following phenological and agronomic traits: days to
50% heading, days to 50% maturity, plant height (cm), hectoliter weight (Kg hl -1),
thousand kernel weight (gm) and grain yield (Kg ha-1) from four central rows. Plot
yields were adjusted to 12.5% moisture content and converted to kilograms per
hectare. Disease data were recorded on scald and net blotch on 0-9 scale and
changed to percentage data, where 0=0%, 1=3%,2=12%,3=25%,4=42%,
5=58%,6=75%,7=88%,8=97%,9=100% before transformed using angular
transformation for statistical analysis. These traits were subjected to analysis of
variance using R- software (R Core Team, 2017). In the analysis, the
environments were considered as random and genotypes as fixed effects, and a
mixed effect model ANOVA was used for statistical analysis. The individual and
combine environment analysis of variance of the experiment was conducted as the
model as suggested by Singh & Ceccarelli (1996).
Yij = µ + Gi +Bj + eij and Yijk = µ + Gi +Ej + GEij +Bk(j) + eijk.

Where, Yij = observed value of genotype i in block j, Yijk = observed value of
genotype i in block k of environment j, µ = grand mean of the experiment, Gi =
the effect of genotype i, Bj = the effect of block j, Bk(j) = the effect of block k in
environment j, eij = error effect of genotype i in block j , Ej = environment effect,
GEij = the interaction effect of genotype i with environment j, eijk = error
(residual) effect of genotype i in block k of environment j. GGE bi-plots were
[11]
performed on grain yield to determine the stability of the genotypes using
GGEBiplotGUI packages of R- software (R Core Team, 2017).
The malt quality traits, namely extract content [% DM], protein content [% DM],
friability [%] and ß-glucan content [mg/L] of selected genotypes were analyzed
using the wet chemistry method in Germany malt quality laboratory “Versuchs-
und Lehranstalt für Brauerei in Berlin” on malted grain following the appropriate
procedure. Malt extract content was determined according to a small-scale version
of the European Brewery Convention (EBC) Methods Manual, Section 4.9.1
(European Brewery Convention, 1998). Additionally, grain samples of all
genotypes included in this study were analyzed at Holetta quality laboratory
following Near-infrared spectroscopy (NIRs) technique using the Bruker Tango
instrument.
Table 1. Lists of malt barley genotypes and environments used for the trials
Trt Genotype Year Trt Genotype Year Loc Year Env

G1 Grace x HB 1307 17-18 G15 KWS_Grinada 2017 Holetta 2017 HO17
G2 Bekoji I xBahati 17-18 G16 KWS-Hobbs 2017 Bekoji 2017 BK17
G3 HB 1307 x Su-Lilly 17-18 G17 KWS-Sassy 2017 Bekoji 2018 BK18
G4 Belgium 2 17-18 G18 KWS_ Canton 2017 Kofele 2017 KF17
G5 HB 52 x Bahati 17-18 G19 KWS-Solicit 2017 Kofele 2018 KF18
G6 IBON 174/03 x Traveller 17-18 G20 IBON-HI13/14-41 2018 D/berhane 2018 DB18
G7 Sabini x Beka 17-18 G21 IBON-HI14/15-45 2018
G8 Bekoji-1 x Grace 17-18 G22 IBON-HI14/15-56 2018
G9 IBON 174/03 * 17-18 G23 IBON-HI14/15-96 2018
G10 Holker 17-18 G24 IBON-HI14/15-102 2018
G11 HB 1963 * 17-18 G25 IBON-HI14/15-147 2018
G12 HB 1964 17-18 G26 MBHIBYT-23 2018
G13 KWS-Dante 2017 G27 Explorer 2017
G14 KWS-Eileen 2017 G28 Traveller ** 2018
*improved Check, locally developed; ** = Improved check, Introduced, G1-G12 evaluated for two years, G13-G28
evaluated for one year

The combined analysis of variance, across seven environments, showed highly
significant variations among the malt barley genotypes tested for all quantitative
traits considered. The differences among the environments were statistically
significant for all traits. Similarly, genotype by environment interaction was
significant for all traits (Table 2). It showed that all genotypes were not affected
by the environment equally. This G x E interaction effect made the selection
process difficult and requires further stability analysis of the genotype to select the
one relatively stable across environments and provide a high yield.
[12]
Table 2. Mean squares of traits of 28 malt barley genotypes grown at six environments
SV DF DHE DMA PLH SC(DF)§ NB(DF)§ TKW HLW GYLD

Gen 27 709** 461** 4370** 1719(27)** 572(27)** 295** 35** 11779569**
Env 5 2599** 4063** 4501** 9367(4)** 12017(4)** 1726** 826** 61902099**
Gen:env 87 26 ** 53 ** 131 ** 214(68) ** 254(68) ** 17 ** 7 ** 932567**
Env:rep 12 21** 16ns 107** 188(10)** 61(10)ns 9ns 1ns 2828458**
Residuals 228 9 16 40 74(190) 67(190) 8 4 441982
CV 3.56 2.78 6.61 16.35 39.31 5.82 2.94 16.92
Mean 84.6 143.4 96.0 52.5 20.9 47.5 67.6 3929.7
DF=degree of freedom, DHE=days to heading, DMA= days to maturity, PLH=plant height (cm), SC=scald (%), NB=net
blotch (%), TKW= thousand kernel weight (g), HLW= hectoliter weight (Kghl-1), GYLD= grain yield(kg ha-1), **, * significant at
5% and 1% probability level, ns=non significant, §these traits were not recorded at DB18 and mean squares under those
traits are angular transformed values
Among the tested genotypes even though not significantly different from some of
the newly tested genotypes and check varieties (IBON 174/03, HB 1963, HB
1964), G5 (HB 52 x Bahati) showed the highest mean grain yield. However, it had
significantly highest mean grain yield than the registered European varieties
(Explorer and Traveller). Then, most genotypes (G21-G26) that were substituted
during the 2018 cropping season had higher mean grain yield comparable to G5,
but it should be noticed that these genotypes were evaluated only at three
environments (BK18, KF18 and DB18). Similarly, among the test genotypes
evaluated at all environments, G6 (IBON 174/03 x Traveller) scored better mean
grain yield. Whereas, the introduced malt barley (G13-G19) varieties scored the
lowest mean grain yield values. The maximum hectoliter weight (HLW) was
recorded for the check variety G11 (HB1963), genotype G8 (Bekoji-1 x Grace)
and G26 (MBHIBYT-23). Significantly higher mean TKW value was recorded on
HB1964 (56.6g) and KWS-Eileen (52.6g). On the other hand, most introduced
malt barley genotypes (viz. G13, G14, G15 G16, G17 and G19) had higher mean
scald value of 67, 65, 69, 63, 67 and 72%, respectively. This is because of
materials are developed and evaluated under different environmental conditions
than Ethiopia and as a result, they are highly susceptible to Ethiopian scald races
(Table 3). However, materials derived from crossing programs had relatively
better tolerance to scald. Accordingly, G1, G3, G4, G5, G6, G7 and G8 scored
mean scald values of 53, 46, 45, 38, 51, 48 and 54%. Regarding net blotch most
of the tested materials showed moderate resistance. In contrast, these genotypes
which had higher scald values showed lower net blotch scores. As an example
G15, G22 and G27 scored 69, 80 and 74% for scald and 9, 23 and 10% for net
blotch, respectively. This may be due to the confounding effect of scald on net
blotch. Plant height showed consistently large variation among the malt barley
varieties. Generally, most foreign materials have short plant height and in contrast
the local varieties such as G7 showed a high mean plant height value of 116 cm
followed by G8 (115 cm). G6 (IBON 174/03 x Traveller) and G9 (IBON 174/03)
were relatively early whereas G-17 (KWS-Sassy), G14 (KWS-Eileen), G-18
[13]
(KWS-Canton) and other European introduced materials were late in days to
maturity. Generally, among the malt barley genotypes tested in all environments,
HB 52 x Bahati and IBON 174/03 x Traveller showed grain yield advantage as
compared to the recently released check varieties (HB1963 and HB 1964) and
better disease resistance (Table 3). Similarly, Bekoji-1 x Grace had comparable
mean grain yield value as the standard checks (HB 1963, Traveller) and high
values in mean grain physical quality parameters (TKW and HLW). Besides, the
newly inserted genotypes (G21-G26) showed similar grain yield with recent check
varieties and to confirm their performance over the years, these genotypes are
being evaluated in the 2019/20 cropping season.
Table 3. Overall mean for eight traits of 28 malt barley genotypes tested during the 2017 and 2018 main cropping season
Trt# Genotype DHE DMA PLH SC§ NB§ TKW HLW GYLD
G1 Grace x HB 1307 84c-g 143d-j 114ab 53d-i 35abc 49.5b-g 69.8ab 3995cd
G2 Bekoji I xBahati 79i-l 137jk 97fgh 71abc 25b-g 44.1hi 68.3a-e 3972cd
G3 HB 1307 x Su-Lilly 84c-g 139h-k 110a-d 46g-j 30b-e 48.9b-g 68.3a-e 3920cde
G4 Belgium 2 84d-h 138ijk 111abc 45hij 31bcd 47.7d-g 68.9a-d 4049bcd
G5 HB 52 x Bahati 81g-j 141f-k 106b-e 38ij 27b-f 46.6ghi 64.0g 5128a
G6 IBON 174/03 x Traveller 73l 136k 96gh 51e-j 23c-g 51.2bc 67.6b-e 4470a-d
G7 Sabini x Beka 82e-i 140f-k 116a 48f-j 31b-e 50.7b-e 68.6a-d 3931cde
G8 Bekoji-1 x Grace 81g-j 140f-k 115ab 54c-i 29b-e 51.0b-d 70.1ab 4027bcd
G9 IBON 174/03* 75 kl 135 k 88h 54 c-i 28 b-e 48.8b-g 67.0 cde 4487a-d
G10 Holker 86cde 141f-k 106b-f 48f-j 41ab 47.5e-h 69.6abc 4043bcd
G11 HB 1963* 87cd 145c-f 104c-g 57b-h 28b-f 50.4b-e 70.5a 4785ab
G12 HB 1964 81g-j 140f-k 106b-e 53d-i 30b-e 56.6a 67.1cde 4409a-d
G13 KWS-Dante 96b 149bcd 55ij 67a-f 9g 38.4k 63.8fg 1936h
G14 KWS-Eileen 95b 152b 60ij 65a-g 17d-g 52.6ab 66.9b-g 1905h
G15 KWS_Grinada 94b 146b-g 52j 69a-e 9g 39.0jk 68.6a-e 2193gh
G16 KWS-Hobbs 97ab 150bc 60ij 63a-h 14efg 38.5k 65.5efg 2870fgh
G17 KWS-Sassy 102 a 164a 63ij 67a-f 20c-g 43.1ij 66.0d-g 2691gh
G18 KWS_ Canton 99 ab 152 b 66 i 49e-j 21 c-g 43.1 ij 66.7 c-g 3050efg
G19 KWS-Solicit 96 b 148b-e 63ij 72a-d 14efg 37.3k 66.2d-g 2430gh
G20 IBON-HI13/14-41 77 jkl 140 f-k 88h 66 a-g 21 c-g 46.8 e-i 68.1a-e 3799def
G21 IBON-HI14/15-45 77jkl 137ijk 100c-g 53b-j 23b-g 50.6b-f 69.4a-d 4300a-d
G22 IBON-HI14/15-56 82e-j 136jk 95gh 80a 23b-g 48.3b-g 66.1d-g 4791abc
G23 IBON-HI14/15-96 81f-j 138f-k 101c-g 31j 23b-g 47.5c-i 69.0a-e 4392a-d
G24 IBON-HI14/15-102 78i-l 138g-k 100d-g 48e-j 23b-g 46.4f-i 67.4a-f 4700a-d
G25 IBON-HI14/15-147 77 i-l 139 f-k 96e-h 78 a 23 b-g 47.0 e-i 69.4a-d 4370a-d
G26 MBHIBYT-23 79h-k 141e-k 102c-g 69a-e 24b-g 47.7c-i 70.0abc 4382a-d
G27 Explorer 87c-f 144c-i 59ij 74ab 10fg 39.5jk 66.0d-g 2775gh
G28 Traveller* 89c 145b-h 88h 48e-j 51a 46.0ghi 66.9b-g 4215bcd
DHE=days to heading, DMA= days to maturity, PLH=plant height (cm), SC=scald (%), NB=net blotch (%), TKW=
thousand kernel weight (g), HLW= hectoliter weight (hl g-1), GYLD= grain yield (kg ha-1), §these traits were not recorded
at DB18, * = Improved checks
Individual environment mean grain yield and malt quality parameters value of
tested malt barley genotypes have been presented in Table 4. Genotype, HB 52 x
Bahati (G5) showed higher mean grain yield values across most environments,
ranged from 3614 - 7026 kg ha-1. Besides the grain yield the genotype had better
[14]
malt quality with a value of 80.7 %, 8.3 %, 72 %, 699 mg/l for extract, protein,
friability and beta-glucan respectively. Similarly, the check variety (HB1963),
exhibited the next high mean grain yield in most of test environments and it also
showed better values in malt quality traits. This confirmed that the variety is an
alternative malting barley genotype for the malting industry. The other genotype,
IBON 174/03 x Traveller (G6) had higher mean grain yield values across test
environments, but the genotype exhibited relatively poor malting quality. On the
other hand, Bekoji-1 x Grace (G8) and Sabini x Beka (G7) showed premium malt
qualities of 81.8 % and 81.5 % for extract, 10.0 % and 9.8 % for protein, 85 % and
78 % for friability and 369 mg/l and 287 mg/l for beta-glucan, respectively. These
malt quality traits are the most important and relevant for the malt factories and
breweries. Accordingly, the G8 and G7 scored mean grain yield value ranged from
2915-5090 kg ha-1 and 2804-6248 kg ha-1, respectively across the test environments
with acceptable malt quality standards (Table 4). Moreover, genotype HB1307 x
Sully (G3) showed better values in all-malt quality parameters. However, unlike
other genotypes, these values were recorded using NIRs techniques. So, the values
should be further validated with malt samples using the wet chemistry method.
Consequently, we can recommend genotype Bekoji-1 x Grace for its premium malt
barley quality and acceptable grain yield potential.
GGEBiplot
Significant mean squares for G X E indicated inconsistency of mean grain yield
performance of genotypes in different environments. Studying the causes of G X E
interaction helps to establish breeding objectives and identify areas of optimal
cultivar adaptation (Yan and Hunt, 2001). The “which-won-where” view of the
GGE biplot is an important feature for mega environment identification. Therefore
in this study based on the mean yield performance of malt genotypes, all
environments fall into a single sector. The six environments (KF17, KF18, BK17,
BK18, DB18 and HO17) grouped as one mega environment and G5, G6, G9 and
G11 performed well in all environments included in mega environments (Figure 1).
In addition, G5 (HB 52 x Bahati) is the vertex genotype, which showed the
superiority of the genotype in mean grain yield performance than the other
genotypes included in the mega environment (Yan and Tinker, 2006). Moreover, no
environments fell into sectors that contained the remaining genotypes, which
indicates that they were the poorest genotypes in all test environments (Yan, 2001).
[15]
Table 4. Individual location average grain yield and malt quality traits performances of the 12 malt barley genotypes
Extract Protein Beta glucan

Trt# Genotype BK17 BK18 DB18 HO17 KF17 KF18 Mean Friability (%)
% % (mg/l)
G1 Grace x HB 1307 3370 4820 3013 3611 4159 4872 3995 80.4 7.6 70 462
G2 Bekoji I xBahati 3651 5813 2640 3101 3947 4822 3972 81.1* 13.5* 68* 333*
G3 HB 1307 x Su-Lilly 4236 6396 3634 2572 4015 2629 3920 84.0* 9.5* 86* 246*
G4 Belgium 2 3873 5926 2972 3325 3784 4426 4049 76.8 13.3 29 1000
G5 HB 52 x Bahati 5304 7026 3614 4293 4675 6046 5128 80.7 8.3 72 699
G6 IBON 174/03 x Traveller 3595 6265 3154 4019 4715 5075 4470 76.7 12.2 37 1000
G7 Sabini x Beka 4127 6248 2804 3441 3633 3287 3931 81.5 9.8 78 287
G8 Bekoji-1 x Grace 3880 5090 2915 4081 4000 3968 4027 81.8 10 85 369
G9 IBON 174/03 3874 6667 2348 4236 4855 5144 4487 79.4 10.6 50 1000
G10 Holker 3477 6373 3035 3658 3418 4179 4043 77.6 13 61 420
G11 HB 1963 4005 5818 3686 5355 5198 4252 4785 81.3 8.9 78 510
G12 HB 1964 3645 5350 2927 4277 4584 5625 4409 78.6 11.6 44 1000
*These data were recorded from Near infrared spectroscopy (NIRs) using Bruker Tango instrument
NB= malt quality standards: Extract, >78 %, Friability, >77 %, Beta glucan, <400 mg/l, Protein, 9-11.5
[16]
Figure 1. The which-won-where view of the GGE biplot of grain yield of Food barley genotypes based on the G × E data
The GGE biplot explained 72% of the grain yield variation due to GGE (Figure 2).
“Mean vs. Stability” view of GGE biplot is an efficient tool to compare genotype
based on mean performance and stability across environments within a mega-
environment (Yan et al., 2007). Mean vs. stability view of GGE biplot presented
in Figure 2. G5 showed higher mean grain yield value than the other test
genotypes and had moderate stability. The check varieties G9 (IBON 174/03),
G11 (HB 1963) and the other test genotype G6 (IBON 174/03 x Traveller) had the
next highest mean grain yield values and these genotypes revealed good stability.
On the other hand, among high malt quality yielding genotypes, G8 (Bekoji-1 x
Grace) showed relatively better stability (Figure 2).
[17]
Figure 2. Mean grain yield performance and stability of genotypes based on the G × E data
Conclusions
The major finding from this study was that G8 (Bekoji-1 x Grace) had shown a
premium malt and physical grain quality result meeting the standard of the malt
industry, which in most cases is missing in our elite varieties. The other genotype
G5 (HB 52 x Bahati) showed significantly higher mean grain yield and good malt
quality traits, except for friability. The GGE biplot, G5 was the winning genotype
in all test environments. Moreover, the “mean vs stability” view of GGE biplot
indicated that G5 recorded the highest mean grain yield. In terms of stability, the
high yielding genotype (G5) and the high-quality genotype (G8) showed moderate
stability across test environments. Overall, genotypes G8 is identified as potential
malt barley candidate variety for further malt quality test before a variety
verification trial. Genotype G5 is included in the crossing block as a potential
donor parent for its high yielding performance across the test environments. In
addition, G8 and G7 (Sabini x Beka) are recommended for the malt barley
crossing program as a potential parent for their good malt quality traits (high malt
extract and low beta-glucan).
[18]
References
Bayeh, M and Berhane L. 2011. Barley research and development in Ethiopia – an
overview. 1n: Mulatu, B. and Grando, S. (eds). 2011. Barley Research and
Development in Ethiopia. Proceedings of the 2nd National Barley Research and
Development Review Workshop. 28-30 November 2006, HARC, Holetta, Ethiopia.
ICARDA, PO Box 5466, Aleppo, Syria. pp xiv + 391
CSA. 2018. Area and Production of Major Crops, Agricultural Sample Survey, 2017/18
(2010E.C.). Central Statistics Agency, Statistical Bulletin 586. Addis Abeba,
Ethiopia
Berhane Lakew, Chilot Yirga and Wondimu Fikadu. 2016. Malt Barley Research and
Development in Ethiopia: Challenges and Opportunities. In: Dawit Alemu, Eshetu
Derso, Getnet Assefa and Abebe Kirub (eds.). Agricultural Research for Ethiopian
Renascence: Challenges and Opportunities. Proceedings of the national conference
on Agricultural Research for Ethiopian Renascence. 26-17 January 2016, UNECA,
Addis Abeba.
Ethiopian Revenues and Customs Authority – ERCA, 2017: http://www.erca.gov.et.
European Brewery Convention. 1998. Analytica EBC. 5th ed. Verlag Hans Carl,
Nurnberg,
Germany.
R Core Team. 2017. R: A language and environment for statistical computing. R
Foundation for Statistical Computing, Vienna, Austria. URL http://www.R-
project.org/.
Singh, M. & S. Ceccarelli. 1996. Estimation of Heritability of Crop Traits from Variety
Trial
Data. Technical Manual No. 21. International Center for Agricultural Research in the Dry
Areas, Aleppo, Syria.
Yan, W. 2001. GGEbiplot - A Windows Application for Graphical Analysis of Multi
environment Trial Data and Other Types of Two-Way Data. agronomy journal,
93:1111- 1118.
Yan, W and NA Tinker. 2006. Biplot analysis of multi-environment trial data: Principles
and applications. Canadian Journal of plant science, 86: 623-645.
Yan, W, MS Kang, Ma B, Woods S and Cornelius PL. 2007. GGE biplot vs. AMMI
analysis of genotype-by-environment data. Crop Science, 47: 643-653.
[19]
[20]
Stability and Performance Evaluation of Food
Barley Varieties for Moisture Stress Environments
Shimeles Gezahegn1, Adane Chofere1, Workeneh Mekasa3,
Tesfahun Alemu2, and Amberber Haile1
1
Kulumsa Agricultural Research center, P.O.box 489, Assela, Ethiopia; 2Holetta Agriculutral Research
center, P.O.Box 31, Holetta, Ethiopia; 3Ambo Agricultural Research center, P.o.box 37, Ambo Ethiopia.
Crossponding author: setotaw2006@gmail.com
Abstract
Two sets of experiments were conducted at Asasa, Dhera, and Arsi Negele from
2015/16 to 2018/19 cropping season under the national variety trial. The experiment
was laid using RCBD with three replications. The analysis of variance showed a
significant difference among the tested genotypes in the two sets of the experiment.
The mean separation test showed that in Set I genotype LMON IBON 12-15 has
similar yield as the standard check Diribe. In the second set LMON IBYT-MRA 12-11
showed maximum average yield than the standard check Gobe and Diribe. Similarly,
the same varieties showed good stability and yield performance with the GGEbiplot
analysis that showed those genotypes have good yield and better stability across
locations and year. Therefore, candidate varieties LMON IBON 12-15 and LMON
IBYT-MRA 12-11 will verify in 2019/20 cropping season for the low moisture areas of
the country.
Keywords: Stability analysis, GGEbiplot, AMMI analysis, moisture stress
Introduction
Barley is an important cereal crop in Ethiopia and grown in a wide range of
environments ranging from lowland (1500 m.a.sl) to extreme highlands (above
3500 m.a.sl.). The crop is also one of the major crops grown in the lowland areas
that suffer from drought due to moisture and heat stress. Drought is one of the
major production constraints in agriculture especially in Ethiopia that highly
suffered from climate change and global warming. This event showed the
importance of developing cultivars that can provide some harvest under the
moisture stress condition. With this in mind, the national barley research program
started a breeding program focusing to identify genotypes tolerant to the moisture
stress areas of the country. The predictions on the impact of climate change in the
world showed sever water and heat stress could be the most limiting factor for
crop production (Vorosmarty et al., 2000; Rizza et al.,2004). Barley has shown
water stress resistance due to its extensive root system (Fischerand Maurer, 1978).
That showed its adaptation to moisture stress environments. Even though the crop
[21]
has tolerance to drought, it’s yielding ability severely hindered by water deficit
during the growing season (Araya and Stroosnijder, 2010).
In Ethiopia barley annually covered about 951,993.15 ha and ranked the 5th
important crop contributing to food security. The average productivity of barley is
21.57 Qt/ha (CSA, 2018). This is very low when compared to the yield potential
obtained in other countries. The low productivity became very serious in the
lowland areas due to moisture stress which resulted in productivity below 10
qt/ha. Even if barley is predominantly a highland crop, about 200000 ha of land
was cultivated every year with it (CSA, 2018). In addition, due to its elasticity, the
barley is more suitable for moisture stress environment. A small amount of
increment in productivity can help to alleviate the food and nutritional insecurity
in the low moisture areas of the country where frequently affected by drought. The
barley breeding program of the Ethiopian Institute of Agricultural Research started
a research program to develop drought-tolerant barley varieties for mid to low
altitude areas of the country. With this initiative, the national program imported
genotypes from ICARDA and other sources developed for drought tolerant to be
evaluated under the Ethiopian condition. Till now four barley varieties were
released for the low moisture areas by the national program that showed the
importance of strengthening the program to develop more varieties.
The impact of climate change and its impact on the suitability map of barley
showed that more land will be pushed to early and moisture stress environment
that required further research work to develop varieties suitable to these parts of
the region (Niguse et al., 2019). To mitigate this situation, the most convenient
method is to develop early maturing and drought tolerant varieties for the low
moisture areas of the country. So, this research activity was conducted to identify
the food barley varieties tolerant or resistant to drought stress environments of the
country.
Material and Methods

The trail comprises of 30 genotypes grouped into two sets where the first set
comprises 13 candidate varieties and two checks (Diribe and Aruso) and the
second set organized with 11 candidate varieties and four checks (Aruso, Diribe,
Bentu and Gobe) (Table 1). The experiment laid in Randomized complete block
design with three replications with a plot size of 1.2m x 2.5 with between row
0.2m and among plants 0.15m spacing. The trials were planted at Dhera, Asasa,
and Arsi Negele from 2015-2018 cropping seasons. Dhera located in the rift valley
that experienced high moisture and heat stress whereas Arsi Negele and Asasa
experienced the terminal stress that caused yield penalty. All the cultural practices
[22]
such as weeding, fertilizer recommendation was done according to the
recommendation for each site.
Table 1: The list of genotypes and its description evaluated in the National variety set I and Set II
Entry Set I Entry Set II

Code Genotype Origin Code Genotype Origin
G15 Aruso Landrace G15 Aruso Local landrace
G14 Diribe Released variety G14 Bentu Rleased variety
G5 LMON IBON 12-10 ICARDA G13 Diribe Released variety
G6 LMON IBON 12-11 ICARDA G12 Gobe Released variety
G7 LMON IBON 12-12 ICARDA G1 LMON IBON 12-1 ICARDA material
G12 LMON IBON-MRA 12-7 ICARDA G6 LMON IBYT-MRA 12-11 ICARDA material
G13 LMON IBON-MRA 12-8 ICARDA G7 LMON IBYT-MRA 12-13 ICARDA material
G1 LMON IEBON 12-1 ICARDA G8 LMON IBYT-MRA 12-15 ICARDA material
The agronomic data Days to heading, days to maturity, plant height, diseases
(scald, net bloth and leaf rust), 1000-kernel weight, hectoliter weight, lodging, and
grain yield/ha were recorded from each experimental unit (plot). The collected
data was analyzed according to the model for RCBD over location as follows.
Yijly = µ + Yy + Ll + (R/YL)jyl + Gi + GLil + GYiy + GLYily + eijrly
Where, µ is the grand mean, Yy the year effect, Ll the location effect, (R/YL)jyl the
replication within year and location effect, Gi the genotype effect, GLil the
genotype x location interaction effect, GYiy the genotype x year interaction effect,
GLYily is the GenotypexLocationxYear interaction effect and eijrly is the error
term. The analysis of variance and adjusted mean was estimated using R software
v. 5.6 (R Core Team, 2019). GGEBiplot analysis was done to understand the
stability and the performance of the tested entries over year and locations using
the GGEBiplot model (Yan et al. 2000, 2007; Yan and Hunt 2001; Yan and Kang
2002) was employed using the GGEBiplotGUI package of R (Frutos et al. 2014).
Based on this analysis the high performing and stable genotypes were selected for
the variety verification trial in the low moisture areas of the country.

The analysis of variance for the measured traits for Set I (Table 2) and Set II
(Table3) trials showed a highly significant difference among the tested genotypes.
[23]
In set I, the standard check deribe gave the maximum over location yield (3513.6
kg/ha) followed by the candidate variety LMON IBON 12-15 (3345.9 kg/ha)
(Table 3). In a similar experiment entry, G11 recorded the highest TKW followed
by local check (Aruso). Since this trait is an important one for marketing the
variety can be used as a parent in the future crossing program. The test entry
LMON IBON 12-15 is the better performing entry over all the standard checks
except Diribe.
The mean separation analysis among the tested barley genotypes in Experiment II
(Table 5) showed a significant difference among the tested genotypes for all
measured traits (Table 4). In this experiment, the test entry LMON IBYT-MRA
12-11 (G6) outperforms the best standard check Gobe (Table 4). In addition, this
variety gave 9% yield advantage over the best check (Gobe). Similarly, entry
LMON IBYT-MRA 12-11 (G6) also produced the superior performance on TKW
over the standard checks (Dribie and Bentu) (Table 5). This trait is the most
important characteristic in barley since it showed the plumpness of the seed and
the flour production. Besides the trait also increase the marketability of a variety.
The performance evaluation of the two experiments showed the possibility of
identifying more competent varieties for the low moisture areas of the country. In
addition, the 9% yield advantage obtained from the second set of the experiment
showed the prospect to increase productivity in the farmer’s field in the low
moisture areas of the country.
Table 2. The ANOVA table of the evaluated traits among the tested genotypes Asasa, Dhera, and Arsi Negele in 2015 to
2017 cropping season (Set I).
SV DF DHE DMA TKW PLH DF HLW DF GYLD

geno 14 828.1*** 1475.2*** 321.2*** 65554.0*** 14 176.2*** 14 1576387.8***
env 7 2527.5*** 16968.4*** 653.9*** 3136154.1*** 6 268.8*** 7 55151610.2***
env:Rep 16 18.7*** 82.8* 10.9* 20954.2ns 14 7.4ns 16 455169.1ns
geno:env 98 43.5*** 84.2*** 21.5* 42894.1*** 77 11.2*** 96 949535.3***
Residuals 224 12.0 48.7 6.6 17524.4 162 5.1 201 405930.9
Note: DHE – Days to heading, DMA – Days to maturity, PLH – plant height(cm), TKW – 1000 kernel weight (gm), HLW-
hectoliter weight (kg/hl) and GYLD – grain yield (kg/ha)
[24]
Table 3: The mean table of the tested genotypes for various traits over year and location at Asasa, Dhera and Arsi Negel
from 2015 to 2018 (Set I)
Entry Genotype DHE DMA PLH TKW HLW GYLD
G15 Aruso 55.9abc 89.6a 90.3c 44.7c 58.2def 2973.5
G14 Diribe 53.3a 95.6abcde 70.5a 38.2b 60.4f 3513.6
G5 LMON IBON 12-10 64.8de 101bcdef 70a 32.5a 51.4a 2635.5
G6 LMON IBON 12-11 67.3de 106.4f 83.6c 39.2b 54.4abcd 2846.5
G7 LMON IBON 12-12 68.3e 106.9f 82bc 38.9b 53.8abc 2760.3
G8 LMON IBON 12-14 62.5cde 103.4def 67.3a 37.2b 55.6bcde 3031.8
G9 LMON IBON 12-15 64.2de 105.3ef 68.5a 37.7b 56.3cde 3345.9
G10 LMON IBON 12-16 61.2bcd 103cdef 69.6a 36.8ab 55.4bcd 2932.1
G11 LMON IBON 12-17 57.1abc 93.5abc 70.3a 46.8c 61.4f 2988.5
G12 LMON IBON-MRA 12-7 62.4cde 104def 67.1a 37.5b 55.9bcde 3094.4
G13 LMON IBON-MRA 12-8 68e 108.6f 83c 38.7b 53.4abc 2950.3
G1 LMON IEBON 12-1 69e 106.6f 83.4c 37.6b 52.2ab 2524.4
G2 LMON IEBON 12-3 55.5ab 91.8ab 69.2a 45.2c 59.5ef 3004.9
G3 LMON IEBON 12-5 67.8de 107.9f 83.6c 39.4b 54.2abc 2699.5
G4 LMON IEBON 12-6 51.8a 94.7abcd 73.1ab 39b 55.4bcd 3047.2
Table 4: The ANOVA table for the evaluated traits at Asasa, Dhera and Arsi Negele from 2016 to 2017 cropping season
(set II)
Mean Mean
Mean square square square Mean square
SV DF DHE DMA PLH DF TKW DF HLW DF GYLD
geno 14 188.7*** 237*** 1088*** 14 106*** 14 148*** 14 2476678.6***
env 7 3433.8*** 7978*** 16210*** 6 2290*** 7 4210*** 7 76108655.4***
env:Rep 16 19.1*** 24ns 141*** 14 5ns 16 11*** 16 1144203.9***
geno:env 98 42.7*** 42*** 78*** 84 21*** 97 24*** 98 712911.1***
Residuals 224 12.6 22 43 185 6 201 5 217 339071.6
Table 5: The mean of the evaluated traits for the tested genotypes in the National variety trial from 2016 to 2018 cropping
season (Set II).
Entry geno DHE DMA PLH HLW TKW GYLD
G15 Aruso 57.2bcd 91.5abc 88.6e 54.9e 48.1c 2754.9ab
G14 Bentu 50.8ab 88a 65.8abc 50.3abcde 43.2abc 3167.5abc
G13 Diribe 54.8abcd 95.7bcd 69.2abcd 51.7abcde 43.2abc 3270.6abc
G12 Gobe 57.3bcd 96.2cd 70.8bcd 53.4cde 47c 3432.6bc
G1 LMON IBON 12-1 60.5d 97.5cd 72.3cd 48.4abc 43.5abc 2952.6abc
G5 LMON IBON 12-10 56.4abcd 92.8abcd 61.2a 49.6abcd 43.6abc 2539.7a
G2 LMON IBON 12-5 55.8abcd 95.3bcd 65.2abc 52.3abcde 46.7bc 2814.9ab
G3 LMON IBON 12-6 59cd 96.2cd 68.3abcd 48.8abc 43.8abc 2900.1abc
G4 LMON IBON 12-7 55.4abcd 97cd 63.5ab 47.8ab 41.9ab 2419.1a
G6 LMON IBYT-MRA 12-11 56.6abcd 96.8cd 71bcd 49.6abcd 45.3abc 3737.4c
G7 LMON IBYT-MRA 12-13 58.4cd 98.6d 75.3d 54.7de 47c 2723.4ab
G8 LMON IBYT-MRA 12-15 55.2abcd 91abc 63.9abc 54.4de 47.9c 2815.5ab
G9 LMON IBYT-MRA 12-16 52.8abc 93.9abcd 64.2abc 47.9a 40.7a 3020.7abc
G10 LMON IBYT-MRA 12-17 50.3a 89.4ab 63.9abc 48.2ab 41.9ab 2871.6ab
G11 LMON IBYT-MRA 12-18 55abcd 93.2abcd 67.7abcd 52.9bcde 46.9bc 2966.1abc
[25]
GGEBIplot analysis
The GGEBiplot analysis was performed for the tested genotypes in experiment I
and Experiment II separately by using the GGEBiplotGUI package of R. The
GGEBiplot is the graphical representation of the environment and genotypes in
the two-dimensional graph at the same time help to investigate the genotypes and
to establish the mega environment formation (Yan and Hunt (2001), Yan and
Tinker (2006)). The mean vs stability analysis in the GGEBiplot presented in the
biplot by plotting the mean environment axis and another perpendicular line pass
through the origin where the mean-environment axis showed the approximate
mean of the genotypes whereas the another perpendicular line indicates the
stability of the genotypes. In Figure 1 the mean vs stability of the Experiment I
presented in this figure it is possible to see the candidate variety G9 has better
performance and relatively good stability since the distance from the mean-
environment axis too short (Yan and Hunt 2001, Yan 2001, and Yan and Tinker
(2006)). The highly stable genotypes are G14 and G7 even if the yield
performance is below the G9 and G13. The break lines drown from the mean-
environment axis showed the stability of the genotypes. From the graph, the ideal
genotype is G7 with high stability and above-average performance. Besides
according to Yan and Hunt (2001), the GGEBiplot explains a small portion of the
total variation that can contribute to this discrepancy.
Figure 1: GGEBiplot graph of mean vs stability for the tested genotypes in Experiment I at Asasa, Dhera, and A.Negele in
2015 to 2018 cropping seasons
[26]
The ranking of genotypes analysis based on the mean performance and stability of
the genotypes (Figure 2) showed that the genotypes G9 is more proximate to the
ideal genotype than the other testing genotypes. According to Yan (2001) the
genotypes near to the center of the concentric circle are the more approximate and
the ideal genotype (that is high yielder and stable). Similarly, the candidate
genotypes showed good yield performance and stability and are potential
candidate varieties for possible release.
Figure 2: The ranking of the genotypes based on the ideal genotypes using GGEBiplot for Experiment I.
The which-won-where view of the GGEBiplot analysis showed that the candidate
variety G9 best perfume in Dhera and A.Negele areas where the standard check
G14 wins at Dhera and Asasa (Figure 3). Besides the high fluctuation in the
performances of the genotypes, the significant interaction effect was observed
[27]
over year by location. Yan and Tinker (2006) indicated the genotypes in the corner
of the polygon are the best performing one in each set of environments within the
angle of the polygon formed by the broken lines. In this assumption, the G9, G13,
G8, G4, and G2 performed best in their respective environments.
Figure 3: The which-win-where view of the GGEBiplot analysis of the Experiment I
GGEBiplot analysis for Environment II

The GGEBiplot analysis is a graphical tool that helps to visualize the
performance, stability and mega environment formation in the biplot (Yan and
Tinker, 2006). The analysis of GGEBiplot of Experiment II showed the genotype
G6 is the high yielding genotypes than the standard check G14 (Diribe) that
showed the potential of this variety for release for the low moisture areas of the
country (Figure 4). In addition the same genotype showed relatively good stability
[28]
as indicated by the broken lines in Figure 4. Even if the standard check Deribe
(G14) showed better stability but its yielding ability is lower than the candidate
variety G6.
Figure 4: The GGEBiplot view of mean vs stability of the tested genotypes in Experiment II evaluated at Asasa, Dhera,
and Arsi Negele in 2016 to 2018 cropping season.
The ranking of genotypes analysis using the GGEBiplot method showed the G6 is
more close to the ideal genotype indicated in the center of the concentric circle
(Figure 5). The genotypes laid near to the concentric circle in the figure are
considered as the most ideal genotype among the tested one for the environments
(Yan 2001, Yan and Hunt (2001) and Yan and Tinker (2006). The standard
checks and the local checks used in this analysis are located far from the center of
the concentric circle that showed the candidate genotypes G9 is better than the
[29]
checks and have the potential to be released as a commercial variety in the coming
cropping season.
The Which-win-where graph of the GGEBiplot also showed that the G6 was best
performing genotypes at Asasa, Dhera and Arsi Negele in 2016, 2017 and 2018
cropping seasons (Figure 6). That showed how the variety is good enough for the
low moisture areas of the country. The same genotype also has good grain size
that is acceptable by farmers. The biplot here showed as the genotype G6 (LMON
IBYT-MRA 12-11) is performing well in most of the environments that showed
the stability and high yielding capacity of the genotype. This also supported with
the mean vs stability analysis of the GGEBiplot Figure 4.
Figure 5. The ranking of genotypes in relation of the ideal genotypes as presented by GGEBiplot analysis for
experiment II
[30]
Figure 6: the Which-Win-Where/what GGEBiplot figure produced for tested gentotypes at Asasa, Dhera and
Arsi Negele in 2016 to 2018 cropping season (Experimnt II).
Conclusion
This report presented the findings of two sets of experiments of food barley for the
low moisture areas of the country from 2015 to 2018 cropping seasons. The report
showed that in Experiment I the genotype LMON IBON 12-15 (G9) performed
well across locations even if its yield performance not superior to the standard
check deribe. So to increase the choice of the varieties for low moisture areas and
considering the comparable yield advantage with the local check we
recommended the genotypes for variety verification trial in 2019 cropping season.
[31]
In Experiment II the tested genotype LMON IBYT-MRA 12-11 (G6) is the highly
stable and best-performing genotype across locations. Similarly, this variety
performed better in almost all environments as shown in GGEBiplot graphs. In
addition, this variety considered as the ideal genotypes than other genotypes in
Experiment II located near the center of a concentric circle. Therefore, genotype
LMON IBYT-MRA 12-11 (G6) will be presented for a variety verification trial in
the 2019/20 cropping season for the low moisture areas of the country.
Generally from the two experiments LMON IBON 12-15 from experiment I, and
LOMN IBYT-MRA 12-11 from experiment II will be submitted for variety
verification trial in 2019 cropping season.
Reference
Araya, A, Stroosnijder L. 2010. Effects of tied ridges and mulch on barley (Hordeum
vulgare) rain water use efficiency. Water Manage.J.97,841–847.
CSA (central statistical agency). 2018. Agricultural sample survey: area and production of
major crops, Meher season. Vol. I. Addis Ababa, Ethiopia.
Nigussie, D, Mulugeta W, Molla A, Bishaw Z and Biradar C. 2019. Land Suitability
Mapping for Production of Chickpea, Faba Bean and Malt Barley Varieties in
Ethiopia. Beirut, Lebanon: International Center for Agricultural Research in the Dry
Areas (ICARDA).
Fischer, RA, Maurer, R. 1978. Drought resistance in spring wheat cultivars .I. Grain yield
response. Aust.J. Agric. Res.29,897–912.
Frutos, E, Galindo MP, Leiva V. 2014. An interactive biplot implementation in R for
modeling genotype-by-environment interaction. Stoch. Environ. Res. Risk Assess.
28:1629-1641.
Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-
project.org/.
Rizza, F, Badeck, FW, Cattivelli L, Lidestri O, Fonzo NDi, Stanca AM. 2004. Use of a
water stress index to identify barley genotypes adapted to rain fed and irrigated
conditions.CropSci.44,2127–2137.
Vorosmarty, CJ, Green P, Salisbury J, Lammert RB. 2000. Globalwater resources:
vulnerability from climate change and population growth. Science 289, 284–288.
Yan W, Hunt LA, Sheng Q, Szlavnics Z. 2000. Cultivar evaluation and mega
environment investigation based on GGE biplot. Crop Sci 40:597–605
Yan W, Hunt LA. 2001. Interpretation of genotype by environment interaction for winter
wheat yield in Ontario. Crop Sci 41:19–25.
Yan W, Kang MS. 2002. GGE biplot analysis: a graphical tool for breeders, geneticists,
and agronomists. Boca Raton, FL, CRC Press
[32]
Stability Analysis of Malt Barley Varieties in the
Highlands of Arsi and West Arsi Zones of Ethiopia
Adane Chofere1, Shimeles Gezahegn1, Workeneh Mekasa3,

Amberber Haile1, and Tesfahun Alemu2
1
Kulumsa Agricultural Research center, P.O.box 489, Assela, Ethiopia; 2Holetta Agricultural Research
center, P.O.Box 31, Holetta, Ethiopia; 3Ambo Agricultural Research center, P.o.box 37, Ambo Ethiopia.
Abstract
Two sets of experiments were conducted in the 2015/16 to 2018/19 cropping season at
Kofele and Bekoji. The experiments were laid out using a randomized complete block
design with three replications. The ANOVA showed significant differences among
genotypes and high genotype x environment interaction for most traits in the two
experiments. The mean separation test showed that in the set I experiment, Beka x HB
52 cross had the highest grain yield than the standard check, Bekoji I. In set t II
experiment genotype MN BRITE provides the maximum grain yield as compared to
the two checks and the other candidate genotypes. Besides, the two candidates
showed a good level of extract and protein content with the range of the malt factory
requirements. Therefore, these two genotypes will be verified in the 2019/20 cropping
season in the variety verification trial.
Keywords: Stability analysis, GGEbiplot, Malt barley and malt quality
Introduction
In Ethiopia barley is the fifth important cereal crop that covers about 9% of the
total area cultivated (CSA, 2018). The total area covered by barley was around 1
million hectares and the average national productivity reached about 2.2 t/ha
(CSA, 2018). Barley is grown from the low land to the highland areas of the
countries with various agro-ecological conditions that make Ethiopia to be one of
the centers of diversity for the crop.
Malt barley is one of the important cash crops in the highlands of Ethiopia. The
total household actively engaged in barley production during the 2017/18 cropping
season was more than 3.5 million (CSA 2018). But out of the total production
about 85- 90% is food barley and only 10-15 % is malting barley (Alemu et al.,
2014). On the contrary, the country allocated foreign currency for importing
jumped from 124 thousand USD to 40 million from 1997 to 2014 (ERCA, 2015).
With this trend, the projection jumped to 420 million by 2025 (Tarekegn, 2015).
This increase associated with the expansion of the capacity of the already existed
[33]
breweries and the establishment of new breweries like DIAGO and Hinken and
maltsters like Sufflet Malt and BortMalt.
Barley is mainly cultivated in the high potential areas of the country ranging from
2300-3000 m.a.sl with evenly distributed rainfall conditions (500 -800mm) with a
pH of 5.5 to 7.3. Even if the country has the best environmental condition for malt
barley production still now the country fulfills the demand for malt barley by
importing from abroad. With the increasing demand for malt barley in the country
and the expansion of breweries and maltsters the demand for raw malt barley is
increasing every year. On the other hand, as malt barley is the industrial crop it
needs to satisfy the quality standards required by the malt factories and the
breweries. That requires a well-designed breeding program to develop malt barley
varieties that satisfy the demand form the breweries and malt factories. Thus, the
national barley breeding program every year conducts a yield trial to identify malt
barley varieties that meet the quality standards of the malt industry. The national
barley improvement program works to develop malt barley varieties that satisfy
the quality requirements by the breweries and the malt factories. For malt barley
plumb seed, high extract (>80%), low protein (8-11.5%), high friability (>70%)
are the major criteria (AMF, 2016). One of the important criteria besides the
quality traits in malt barley breeding is the yield potential, disease resistance and
stability across the environments. The stability and yielding potential across
locations can be studied using the GGE Biplot analysis (Yan, 2001, Yan and Hunt,
2001 and Yan and Tinker, 2006). The GGE Biplot technique is the visual
presentation of the multi environments trial (MET) experiments to investigate the
stability and productivity of the tested genotypes (Yan, 2001).
The GGEBiplot analysis help to understand the stability and the yielding ability of
the genotypes using the biplot based on the G + GE model. The analysis helps to
visualize the performance of the genotypes and understand which variety performs
well where that helps to identify the best genotype for a specific mega
environment (Yan and Hunt (2001).
Therefore, these experiments were conducted to identify high yielding, disease

resistance (scald and net blotch) and stable across high potential areas of barley
growing environments with the desired malt quality parameters.

The trail comprises 24 genotypes grouped into two sets where each set with 10
candidate varieties and two checks (Bahati and Bekoji I). The experiment laid in
Randomized complete block design with three replications with a plot size of
[34]
1.2m x 2.5 with between row spacing of 0.2m. The set I trial w conducted at
Bekoji and Kofele from 2015-2017 cropping seasons whereas set II was
conducted from 2016 to 2018 cropping season. Both locations are the potential
malt barley producing areas in the region. All the cultural practices such as
weeding, fertilizer applications were done according to the recommendation for
each site. The agronomic data Days to heading, days to maturity, plant height,
diseases (scald, net blotch and leaf rust), 1000-kernel weight, hectoliter weight,
lodging, and grain yield/ha were recorded from each experimental unit (plot).
From the same plot the malt quality traits grain protein, friability, and malt extract
were estimated using the Tango NIR infrared analyzer at malt barley laboratory,
Holetta Agricultural Research Center. The collected data was analyzed according
to the model for RCBD over locations and years as follows.
Yijly = µ + Yy + Ll + (R/YL)jyl + Gi + GLil + GYiy + GLYily + eijrly
Where, µ is the grand mean, Yy the year effect, Ll the location effect, (R/YL)jyl the
replication within year and location effect, Gi the genotype effect, GLil the
genotype x location interaction effect, GYiy the genotype x year interaction effect,
GLYily is the GenotypexLocationxYear interaction effect and eijrly is the error term.
The analysis of variance and the adjusted mean was estimated using R software v.
5.6 (R Core Team, 2019).
GGEBiplot analysis: To understand the stability and the performance of the tested
entries over the years and locations the GGEBiplot model (Yan et al. 2000, 2007;
Yan andHunt2001; Yan and Kang 2003) was employed using the GGEBiplotGUI
package of R (Frutos et al. 2014). In the analysis, the G + GE model was
employed to partition the genotype and genotype x environment interaction effect.
Based on this analysis the high performing and stable malt barley genotypes were
selected for the variety verification trial for the highly potential malt barley
growing areas in the country.
Results and Discussions

The individual location analysis of variance for grain yield in each year revealed
that there was significant variation among the tested genotypes for both sets of
experiments except for Kofele in 2017 in the set-II experiment (Tables 1 and 2).
The maximum yields were consistently obtained from Beka x HB 52 and MN
BRITE in the two locations and all the years from set-I and set-II experiments,
respectively. The low yielding genotypes vary in set-I experiment across locations
and years but genotype, FEG192-69 showed the very minimum yield in the two
locations in all the years in set-II experiment. The Combined analysis of variance
[35]
for the national trial set I (2015 to 2017) showed a significant difference among
the tested genotypes for all traits except for malt extract and friability (Tables 3
and 5). The candidate variety Beka x HB 52 gave acceptable extract value (80 %)
and meets the malt quality standard of the malt industry and breweries (Table 5).
The same candidate variety provides a 14% and 12% grain yield advantage over
the standard checks Bekoji I and Bahati respectively. Similarly, the candidate
variety showed acceptable grain protein content 11% as compared to the standard
check Bekoji I. The candidate variety (Beka x HB 52) gave an average yield of
4490.2 Kg/ha as compared to the standard check Bekoji I (3939.3 kg/ha), and
comparable 1000 seed weight with the standard check.
The Combined analysis of variance for the second set of malt barley variety trial
(Set II), also showed a highly significant difference among the tested genotypes
for all traits except malt extract and friability (Tables 4 and 6). Among the
genotypes in set II, the genotype MN BRITE gave a 10.5 and 17.4% yield
advantages over the standard checks Bekoji I and Bahati, respectively. The MN
BRITE (the candidate variety) gave a grain yield of 5712.8kg/ha whereas the best
standard check (Bekoji-I recorded 5158.6 kg/ha). Similarly, the same genotype
has acceptable malt extract value required by the malt industry, and comparable to
the standard check (Bekoji I). The extract value is within the range of the Assela
Malt factory. Similarly, the protein content is within the acceptable range that is
from 9.5 to 11.5%. There was also a significant genotype by environment
interaction in both sets of experiments indicating the importance of identifying a
stable variety across the test environments or a specific genotype for a specific
environment (Tables 2 and 3).
[36]
Table 1. Mean seed yield (kg/ha) of 12 malt barley genotypes tested under National Variety Trial set-I in the years 2015, 2016
and 2017
2015 2016 2017
Genotype Bekoji Kofele Bekoji Kofele Bekoji Kofele Mean
Bahati 5291.5 3922.6 4950.3 4830.9 4535.2 3860.3 4008.1
Beka X HB-52 5773.6 4404.7 5432.5 5313 5017.3 4342.5 4490.2
Bekoji I 5222.6 3853.8 4881.5 4762.1 4466.4 3791.5 3939.3
Kulumsa-ON/# 16 5308.9 3940.1 4967.8 4848.4 4552.6 3877.8 4025.6
Kulumsa-ON/# 17 5426.8 4058 5085.7 4966.3 4670.5 3995.7 4143.5
Kulumsa-ON/# 19 5208.6 3839.8 4867.5 4748.1 4452.4 3777.5 3925.3
Kulumsa-ON/P# 15 4920.4 3551.5 4579.3 4459.8 4164.1 3489.3 3637
Miscal 21 X CDC Select(bulk) 5032.8 3663.9 4691.7 4572.3 4276.5 3601.7 3749.4
Miscal 21 X HB 120(BULK) 5268.9 3900.1 4927.8 4808.4 4512.7 3837.8 3985.6
Miscal 21 X sabini 5226.9 3858.1 4885.8 4766.4 4470.7 3795.8 3943.6
Miscal x Bahati 5421.7 4052.9 5080.6 4961.2 4665.5 3990.6 4138.4
Miscal x karne 5540.7 4171.8 5199.6 5080.1 4784.4 4109.6 4257.3
Mean 5303.6 3934.8 4962.5 4843.1 4547.4 3872.5 4020.3
Minimum 4920.4 3551.5 4579.3 4459.8 4164.1 3489.3 3637
Maximum 5773.6 4404.7 5432.5 5313 5017.3 4342.5 4490.2
LSD 0.05% 921.4 1238 1172.9 1214.5 945.3 1177.4 435.7
CV % 11.5 21.6 15.7 16.7 14 21 16.4
Table 2. Mean seed yield (kg/ha) of 12 malt barley genotypes tested under National Variety Trial set-II in the years 2016,
2017 and 2018
2016 2017 2018

Genotype Mean
Bekoji Kofele Bekoji Kofele Bekoji Kofele
Bahati 4842.6 5057.5 5161.5 4219.6 5747.3 4939.7 4994.7
Bekoji I 5006.5 5221.4 5325.4 4383.4 5911.1 5103.6 5158.6
Burton 4874.9 5089.8 5193.8 4251.9 5779.6 4972 5027
FEG192-69 4048.2 4263.1 4367.1 3425.2 4952.9 4145.3 4200.3
FEG 192-42 4786.4 5001.3 5105.2 4163.3 5691 4883.5 4938.4
FEG192-16 4604.7 4819.6 4923.6 3981.7 5509.4 4701.8 4756.8
GEN2 -036 4902.1 5117 5221 4279.1 5806.8 4999.2 5054.2
Hol 211 4797.7 5012.6 5116.6 4174.7 5702.4 4894.8 4949.8
M-135 5365.9 5580.8 5684.7 4742.8 6270.5 5462.9 5517.9
M-145 5040 5254.9 5358.9 4416.9 5944.6 5137.1 5192.1
MN BRITE 5560.8 5775.6 5879.6 4937.7 6465.4 5657.8 5712.8
STARSO 620B 4493.8 4708.7 4812.7 3870.8 5398.5 4590.9 4645.9
Mean 4860.3 5075.2 5179.2 4237.2 5764.9 4957.4 5012.4
Minimum 4048.2 4263.1 4367.1 3425.2 4952.9 4145.3 4200.3
Maximum 5560.8 5775.6 5879.6 4937.7 6465.4 5657.8 5712.8
LSD 0.05% 1101 871.3 892.5 NS 1437 1429.5 447.6
CV % 15.4 11.6 11.6 16.8 16.6 19.5 15.5
[37]
Table 3. Combined analysis of variance of the evaluated traits among the tested genotypes (Set I).
Source DF DHE DMA SC NB PLH TKW HLW GYLD

Geno 11 111.91*** 89.6713*** 15.06*** 0.47*** 1916.62*** 19.06*** 167.54*** 898780.83*
Env 5 1066.8*** 2159.749*** 33.93*** 12.82*** 1732.93*** 858.97*** 226.92*** 11930716***
Env:Rep 12 8.13 11.62037 3.26 0.10 99.18 1.90 11.40 1231609.2
Geno:Env 55 10.24** 14.41776*** 1.47** 0.50*** 46.54*** 6.17*** 11.73*** 681432.77*
Residuals 132 6.09 7.292088 0.81 0.11 30.44 3.24 5.69 436645.98
NB: Geno.-Genotypes, Env.- Environment, Rep.-Replication, DHE – Days to heading, DMA – Days to maturity, PLH-
plant height, HLW – hectoliter weight, SC- scald, NB – net blotch, TKW – 1000 kernel weight, GYLD – Grain yield kg/ha,
Table 4. Combined analysis of variance of the evaluated traits among the tested genotypes (SetII).
Source DF DHE DMA SC NB PLH TKW HLW GYLD

Geno 11 104.5*** 59.30*** 3.55*** 7.56*** 1400.51*** 79.96*** 167.55*** 1372976***
Env 4 959.4*** 984.64*** 23.27*** 15.66*** 1007.72*** 2063.28*** 4277.2*** 11668155***
Env:Rep 10 8.5 5.47 2.20 0.58 76.26 7.70 16.95 1262714
Geno:Env 44 24.0*** 13.52*** 1.13** 1.59*** 76.61*** 14.11*** 31.29*** 749975.1**
Residuals 110 5.5 6.70 0.54 0.32 16.39 4.40 6.03 410412.4
NB Abbreviations are as given in Table 3.
[38]
Table 5. The mean performance of the tested genotypes evaluated at Bekoji and Kofele from 2015 to 2017 cropping season
Entry Genotypes DHE DMA PLH HLW SC NB TKW GYLD Extract Friability Protein
G11 Bahati 86.2abc 148.7abc 96.8bc 64.3abc 4.7bcd 1.1a 52.3bc 4008.1a 81.2a 72a 10a
G1 Beka X HB-52 91ef 148.1abc 113.2e 64.8bc 4.2abc 1.3a 49.4b 4490.2a 80a 52a 11ab
G12 Bekoji I 89.8def 147.9ab 107.4de 65c 3.7ab 1.4a 50.5b 3939.3a 78.1a 60a 12ab
G3 Kulumsa-ON/# 16 83.2a 146.4a 87.2a 61.7a 5.9d 1.3a 51.2b 4025.6a 80.4a 68a 11ab
G4 Kulumsa-ON/# 17 87.6bcde 147.3ab 91.9ab 62.9abc 5.3cd 1.5a 49.4b 4143.5a 80.2a 64a 10ab
G5 Kulumsa-ON/# 19 88.7cdef 149abc 87.8a 63.6abc 5.3cd 1.5a 50b 3925.3a 80.2a 60a 10ab
G2 Kulumsa-ON/P# 15 87.2bcd 147.1a 91.9ab 62.1ab 5.2cd 1a 49.7b 3637a 79.1a 70a 11ab
G7 Miscal 21 X CDC Select(bulk) 86.9bcd 151.6bcd 101.8cd 63.3abc 5.4cd 1.2a 45.3a 3749.4a 81.1a 62a 11ab
G8 Miscal 21 X HB 120(BULK) 88.2cdef 150.6abcd 113.9e 64.7bc 3.6ab 1.2a 51.6bc 3985.6a 81.4a 71a 12ab
G10 Miscal 21 X sabini 84.1ab 149.7abcd 108.6de 63abc 5.3cd 1.3a 53.1bc 3943.6a 78.6a 51a 13b
G6 Miscal x Bahati 91.5f 153.6d 112.5e 64.2abc 3a 1.1a 55.3c4 4138.4a 79.3a 57a 12ab
G9 Miscal x karne 88.6cdef 152.3cd 110.3e 63.4abc 4.2abc 1.3a 57.14 4257.3a 78.4a 48a 11ab
NB: NB Abbreviations are as given in Table 3.
[39]
Table 6. The mean performance of the genotype evaluated at Bekoji and Kofele over 2016 to 2018 cropping seasons
Entry Genotype DHE DMA PLH HLW SC NB TKW GYLD Extract Friability Protein
G11 Bahati 86ab 148.1bc 106.1bcde 64.2b 2.8ab 1.6a 57.1c 4994.7abc 79.7ab 62a 11a
G12 Bekoji I 89.7b 145.9abc 111.4def 63.1b 2.2a 2.2a 54bc 5158.6abc 80ab 75a 11a
G1 Burton 83.7ab 143.2a 86.5a 62.9b 4.1b 1.2a 48.7ab 5027abc 82b 76a 11a
G4 FEG 192 -69 89.7b 150.1c 122.4f 62.8b 2.3a 2a 50.8abc 4200.3d 82.2b 65a 11a
G3 FEG 192-42 84.1ab 147.8abc 113.9ef 56.5a 2.3a 2a 45.3a 4938.4abcd 76.2a 66a 12a
G2 FEG192-16 88.7b 145ab 115.1ef 60.2ab 2.7ab 4.3b 55bc 4756.8bcd 79.4ab 50a 12a
G5 GEN2 -036 88.4b 146abc 111.1cde 64.2b 2.7ab 2.1a 54.3bc 5054.2abc 80.5b 71a 12a
G6 Hol 211 85.7ab 145.8abc 108cde 63.3b 2.7ab 2.4a 55.7bc 4949.8abcd 80ab 78a 12a
G7 M 135 85.5ab 147.9bc 100.3bc 60.5ab 1.8a 1.5a 52.2abc 5517.9ab 81.3b 67a 11a
G8 M 145 85.7ab 144.1ab 101.5bcd 61.8b 2.9ab 1.7a 52.4bc 5192.1abc 79.6ab 73a 11a
G9 MN BRITE 81.5a 144.7ab 95.8ab 63b 2.5a 1.6a 55.8bc 5712.8a 79.1ab 70a 9a
G10 STARSO 620B 89b 147.7abc 110.8cde 64.2b 2.8ab 2.2a 54.5bc 4645.9cd 80.7b 75a 11a
NB Abbreviations are as given in Table 3.
[40]
GGEBiplot and stability analysis
The GGE Biplot analysis of the national variety trial Set I showed the
performance of each tested candidate varieties to the respective environments
(Figure 1). The candidate variety G1 seems to perform better at Kofele and Bekoji
in 2015 and 2017 cropping seasons. The mean and stability analysis of the GGE
Biplot can be used to identify the high yielding and stable genotypes. In the figure
from the mean performance line indicated by the arrow, it is possible to identify
genotypes with high yield (Yan 2001). The variety on the longer side of the arrow
is G1 (Beka x HB52) that showed this variety provides high yield than the other
candidate varieties (Figure 2). To identify the stable genotype, look much the
genotype lies away from the genotype mean axis (Yan and Tinker, 2006).
Similarly in our study, the standard check G12 (Bekoji I) is the most stable variety
whereas G5 is unstable genotype. The candidate variety G1 (Beka x HB52) seems
relatively stable. Generally, the candidate variety G1 (Beka x HB 52) is the high
yielder and relatively stable among the genotypes tested at Bekoji and Kofele from
2015 to 2017 cropping seasons (Figure 2). The standard variety G11 (Bahati) is an
average yielder and unstable as compared to other candidate varieties. The other
standard check Bekoji (G12) is highly stable but yielded below the grand mean
(Figure 2). This showed the candidate variety Beka x HB52 is high yielder and
stable that can be verified through the variety verification trial in 2019/20
cropping season.
Figure 1: The GGEBiplot analysis showed the performance of the tested genotypes under set I at Bekoji and Kofele from
2015 to 2017 cropping season.
[41]
Figure 2: The GGEBiplots graph showing the mean performance and stability of the tested genotypes evaluated at Bekoji
and kofele from 2015 to 2017 cropping season.
The ranking of genotypes relative to the ideal genotypes showed that the candidate
variety located near the center of the concentric circle is the stable and high
yielder genotype and considered as one approach for the ideal genotype
identification (Yan and Tinker, 2006). The Biplot analysis showed that G1 (Beka
x HB52) is the genotype that approximates the ideal genotype in this study (Figure
3) whereas all the other genotypes located outside the concentric circle. This
indicates candidate variety G1 (Beka x HB 52) is high yielding and a more stable
genotype as compared to the rest of the genotypes included in this experiment.
The Biplot analysis of the tested genotypes which won where/what (Figure 4)
showed the candidate variety G1 (Beka x HB52) perform best at Kofele in 2015
cropping season. The same variety also performed well in the 2015 and 2017
cropping season at Bekoji. The standard check G12 is located in the center
indicating its average performance in most of the environments.
[42]
Figure 3: The biplot ranking of the genotypes according to the ideal genotypes
Figure 4: The GGEBiplot analysis figure showing which win where/what for the tested genotypes at Bekoji and kofele from
2015 to 2017.
[43]
Generally from the malt barley national variety trial set I conducted from 2015 to
2017 cropping season at Bekoji and Kofele, the candidate variety G1 (Beka x
HB52) showed the best performance with an average yield of 4400 kg/ha with a
yield advantage of 14% from the best standard check Bekoji I. The candidate
variety also showed relatively good yield stability and satisfies the malt quality
traits required by the maltsters and brewing industries. Therefore, the candidate
variety Beka x HB52 will be promoted to the variety verification trial in the
2019/20 cropping season.
The GGEBiplot analysis of the second experiment (Set II) evaluated from 2016 to
2018 cropping season at Bekoji and Kofele was presented in Figure 5. The
analysis showed that genotype G9 performed best at Bekoji in 2018, whereas the
standard check Bekoji I (G12) performed better at kofele in the 2016 and 2018
cropping season. The candidate varieties G6, G8, G10 did not perform best in any
of the environments (Figure 5). The mean and stability analysis of the GGE
Biplots (Figure 6) showed that candidate variety G9 (MN BBRITE) recorded the
highest average mean grain yield than the other candidate varieties and the
standard checks. According to Yan and Tinker (2006), the line with arrow shows
the performance of the genotypes whereas the second perpendicular line indicates
the stability of the genotypes across environments. From this aspect, G7 and G5
are the most stable genotypes whereas G4 and G8 are the most unstable genotypes
from this experiment (Figure 6). The standard check Bekoji I (G12) is more stable
than MN BRITE but it is poorly performed in its yielding ability. Therefore, we
can deduce from figure 6 that candidate variety G9 (MN BRITE) is a high yielder
and relatively stable genotypes in this experiment.
The ranking of the tested genotypes with the ideal genotype showed candidate
variety G9 (MN BBRITE) is more close to the ideal genotype (Figure 7) whereas
candidate variety G10 is far from the concentric circle. In the graph, the genotype
lied in the center of the concentric circle considered the most ideal genotype
whereas a genotype located far from the center considered as poor performing and
not satisfy the characteristics of an ideal genotype (Yan 2001; Yan and Hut 2001
and Yan and Tinker 2006)). In this graph, the standard check is far from the
center of the concentric circle that indicates its poor performance as compared to
the best candidate variety G9 (MN BRITE).
According to Yan (2001) the GGE Biplot figure of which won where/what
showed which genotypes best to which environment. Based on this assumption G9
(MN BRITE) perform best at Bekoji, and Kofele in 2017 and 2018 cropping
season and the same genotype considered as the best performer in these locations
than other genotypes. The standard check variety (G12) performed better at Kofele
[44]
in the 2016 cropping season. But it is not the best performing one than others in
this group since G8 is the best in this condition.
Generally, the result of the combined mean separation and GGE Biplot analysis
showed the candidate variety G9 (MN BRITE) is the high yielder among the
tested genotypes and relatively stable. The average yield (57100 kg/ha) was
recorded which is a 10.7% yield advantage over the standard check (Bekoji I).
Therefore, the candidate variety (MN BRITE) has a yield advantage and stable
performance across the test environments will be verified in the 2019/20 cropping
season.
Figure 5: The GGEBiplot analysis of the tested genotypes in Set II evaluated at Bekoji and kofele from 2016
to 2018 cropping season.
[45]
Figure 6. The GGEBiplot for mean and stability of the genotypes the line with arrow showed the yield capacity
of the genotypes and the second line showed the stability of the genotypes for the Set II evaluated
at Bekoji and Kofele from 2016 to 2018 cropping season.
Figure 7. The GEEBIplot figure showing the ranking of the genotype in relation to the ideal genotypes for the genotypes
tested at Bekoji and Kofele from 2016 to 2018 cropping season
[46]
Figure 8. The GGEBiplot showing which genotypes win where/what for Set II evaluated at Bekoji and Kofele from 2016 to
2018 cropping season
Conclusion
In this experiment two sets of malt barley varieties evaluated from 2015 to 2018
cropping season at Bekoji and Kofele under the national variety trial. The result
showed a highly significant difference among the genotypes for the recorded traits
except for extract and friability for both sets of experiments. In addition, a highly
significant genotype by environment interaction was observed for grain yield
among genotypes. The mean performance for grain yield in the two sets of
experiments showed that genotypes (MN Brite and Beka x HB52) were better than
standard checks (Bekoji I and Bahati) in the tested locations. These genotypes
showed good stability across the environments and years. Similarly, the two
genotypes showed the required malt quality standards with an average mean
extract of 80% and acceptable protein content (9.5-11.5 %) set by the malt factory.
Therefore, these genotypes were proposed for a variety verification trial in the
2019/20 cropping season.
[47]
References
Alemu, D, Kelemu K and Lakew B. 2014. Trends and prospects of malting barley value
chains in Ethiopia. Addis Ababa, Ethiopia.
Assela Malt Factory (AMF). 2016. 2015/2016 Disclosure Journal. December 2016.
Assela, Ethiopia.
major crops, Meher season. Vol. I. Addis Ababa, Ethiopia
major crops, Meher season. Vol. I. Addis Ababa, Ethiopia.
Ethiopian Revenues and Customs Authority – ERCA, 2015
Frutos et al. 2014 Frutos E, Galindo MP, Leiva V. 2014. An interactive biplot
implementation in R for modeling genotype-by-environment interaction. Stoch.
Environ. Res. Risk Assess. 28:1629-1641
Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-
project.org/.
Tarekegn Garomas. 2015. The Role of Smallholder Farmers in the Import Substitution
and Industrialization of Ethiopia: The Case of Malt Barley Producers in Arsi and
Bale Areas, Ethiopia. (MSC Thesis). Indira Gandhi National Open University
(IGNOU) St. Mary's University (SMU), Addis Ababa, Ethiopia
Yan, 2001, Yan W. 2001. GGEbiplot—a Windows application for graphical analysis of
multi-environment trial data and other types of two-way data. Agron J 93:1111–
1118
Yan, W et Tinker NA. 2006. Biplot analysis of multi-environment trial data: Principles
and applications. Can. J. Plant Sci. 86: 623–645.
Yan W, Hunt LA, Sheng Q, Szlavnics Z. 2000. Cultivar evaluation and mega-
environment investigation based on GGE biplot. Crop Sci 40:597–605
Yan W, Kang MS, Ma B-L, Woods S, Cornelius PL. 2007. GGE biplot vs. AMMI
analysis of genotype-by-environment data. Crop Sci 47:643–653.
Yan and Hunt. 2001. Biplot Analysis of Multi-environment Trial Data 289
Yan W, Kang MS. 2003. GGE biplot analysis: a graphical tool for breeders, geneticists,
and agronomists. CRC Press, Boca Raton, FL.
[48]
Mean Performance and Stability Analysis of Yield
and Fiber Quality Traits of Early Maturing Cotton
(Gossypium hirsutum L.) Genotypes in Ethiopia
Merdasa Balcha, Donis Gurmessa, Bedane Gudeta, Samuel Damtew,
Arkebe G/Egziabher, Alehegn Workie and Mekashaw Arega
Ethiopian Institute of Agricultural Research, Werer Agricultural Research Center, P. O. Box 2003, Addis
Ababa, Ethiopia
Abstract
This study aimed to evaluate the mean performance and stability of genotypes across the
environment for the development of high yielding and good fiber quality cultivars.
Fifteen genotypes were tested from the 2016-2018 cropping season at three locations
using a randomized complete block design. Studying the effect of genotype, environment,
and genotype-by-environment interaction and determining representative testing
environments are important for releasing new varieties. Hence, among tested genotypes,
genotypes namely, Chamo Farm no. 1A1-1 DP-90 F1#307 were best performed overall
location for important traits such as seed cotton yield, lint yield, ginning outturn, fiber
length, and fiber strength. Based on the polygon view of the GGE bi-plot, two mega-
environments were detected with different winning genotypes. Considering the
simultaneous selection of mean lint yield stability and fiber quality performances of the
genotypes, among tested genotypes, genotype G8 (Chamo Farm no 1A1-1 DP-90
F1#307) was selected and promoted to variety verification trials to be tested in the
coming cropping season for its relatively wide adaptable performance.
Keywords: GGE bi-plot, mega-environments, adaptable
Introduction
Cotton production attained from upland cotton (Gossypium hirsutum L.) is more
than 90% in the world with high yield and generally good fiber quality while
remaining cotton supply resulting from the closely related Gossypium barbadense
L., Gossypium arboreum L., and Gossypium herbaceum L. species (Zhang et al.,
2008). Cotton is a great source of textile fiber, food for animal and food-grade
vegetable oil (Wakelynet et al., 2010). It is also used as fuel (Singh et al., 2018).
Because of its economic importance, Gossypium hirsutum has attracted
considerable scientific interest, not only among plant breeders and agricultural
scientists, but also among taxonomists, geneticists, and evolutionary biologists
(Wendel and Cronn, 2003).
Classical cotton breeding methods most important ways of improving quality traits
through the tedious process of selection based on measuring the fiber traits in
[49]
advanced generations of the selected lines (Green and Gulp, 1990), but the genetic
improvement has been hampered by the association of poor fiber quality traits
with high yields in cotton. Another barrier to fiber quality improvement through
conventional breeding in cotton is the length of time required (Kohel, 1999).
Hence, the development of new variety with higher yield and fiber quality are the
primary objectives of cotton breeding programs through hybridization followed by
pedigree methods. However, cotton breeders seriously face the problem of
selecting suitable parents and promising crosses to breed high yielding and high-
quality cotton varieties due to the negative association between seed cotton yield
and fiber quality in efforts to improve fiber quality through conventional breeding
methods. However, the major objectives of the cotton breeding programs of
Ethiopia are the improvement of cotton seed cotton yield, fiber quality and other
agronomic traits. However, Ethiopia is endowed with 18 main and 32 sub-agro-
ecologies in which wide agro-ecological variability is the major challenge for field
crops which resulted in high genotypes by environment interaction (GEI) effect.
This GEI effect is a function of inconsistent responses of varieties due to genetic
vs. location effects beyond breeding methodology practice. This entails the need
to initiate such a study to close the information on the performance of Cotton
genotypes and to identify adapted and stable high yielding varieties for the
Environment.

The present study was carried out at different cotton production areas of Ethiopia
at the location of Werer and Gewane at Afar Region and Nasa Farm in the
Southern part of Ethiopia. The experimental material was grown during the years
of 2016-18 at Werer and Nasa Farm and, 2018 at Gewane. Experimental material
was selected from diverse selection material and tested for two years at a
preliminary variety trial at Werer and proceed to national variety trail which of
comprising of 15genotypes (13 lines + 2 checks) as presented in Table 1. A total
of 15 genotypes were attempted in a randomized complete block design (RCBD)
with three replications. Each entry i.e. genotypes were sown in a five-row of 5m
length adopting a spacing of 90 cm between rows and 20 cm between the plants.
All the recommended practices were adopted for all the entries from sowing till
Picking. For recording observations, 15 competitive plants were randomly
selected and tagged from each treatment in each replication and the average value
per plant was computed for agronomic and fiber quality traits viz. plant height,
number of seeds per boll, average boll weight (gm), seed cotton yield (t/ha),
ginning outturn (%), lint yield (t/ha), ginning outturn (%), fiber length (mm), fiber
strength (gm/tex), micronaire. The mean data of 15 plants over three replications
were used for statistical analysis. The data were analyzed to find out the extent of
[50]
mean performance and stability analysis using SAS software and R-package
software, respectively.
Table 1. Cotton genotypes evaluated in the seven environments (for 3years at three locations)
Environment
Code Genotypes Location Year
Code
G1 Etna F4# 1-11 Werer 2016 E1
G2 Europa X Gedera 236 – 02-4 2017 E2
G3 Europa X Gedera 236 – 02-5 2018 E3
G4 Stam 59A X Europa – 01-33 Nasa farm 2016 E4
G5 Stam 59A X Europa – 5-2-1 2017 E5
G6 Melkawerer Farm 2 Farm no 51 DP-90 F1#111 2018 E6
G7 Arbaminch Farm no 3-2 DP-90 F1#286 Gewane 2018 E7
G8 Chamo Farm no 1A1-1 DP-90 F1#307
G9 Farm no Ago1 DP-90 F1#334
G10 Farm no Ago1 DP-90 F1#337
G11 Weyto Farm no M1 DP-90 F1#375
G12 Weyto Farm no M1 DP-90 F1#376
G13 Deltapine -90
G14 Stam 59A
G15 Ionia
Result and Discussion

Results from the mean squares are presented in Tables 2 for agronomic and
quality traits. Data analysis indicated that significant differences among genotypes
for all tested i.e. plant height, average boll weight, seed cotton, ginning outturn,
lint yield, micronaire, fiber length, and fiber strength traits except the number of
boll per plant. The genotype and environment main effects were significant
differences for plant height, seed cotton yield, lint yield and fiber length while
non-significance of genotypes by environment interaction for the number of boll
per plant, average boll weight, ginning outturn, micronaire and fiber strength.
Table 2. Mean squares of cotton genotypes for agronomic and quality traits at Werer, Gewane and Nasa farm
Source Variation df PH NBPP ABW SCY GOT LY M FL FS

Genotypes 14 1554.91** 22.96 ns 1.42** 2.10** 61.35** 0.72** 0.62** 21.99** 22.06**
Geno x year 28 165.69 ns 15.89 ns 0.36 ns 0.87 ns 3.04** 0.11 ns 0.16** 2.27** 4.88**
Geno x Loc 28 128.59 ns 14.62 ns 0.27 ns 1.50** 1.50 ns 0.22** 0.07 ns 0.85ns 2.16ns
Geno x yrs x Loc 28 351.45** 16.37 ns 0.37 ns 1.25** 2.57 ns 0.17** 0.09 ns 0.99* 2.58ns
R2 (%) 76.16 77.57 47.29 76.65 87.32 79.02 81.41 84.17 86.65
ns, * and **, non-significant, significant at P<0.05 and P<0.01, respectively, LSD (5%) = least significant difference at P < 0.05,
Geno. = Genotypes, Loc=Location,yrs= year, CV (%) = coefficient of variation in percent, PH=Plant Height, NBPP= Number boll
Per plant, ABW= Average boll weight, SCY (t/ha)= seed cotton yield, GP(%)= ginning percentage and LY(t/ha)= lint yield,
M=micronaire, FL(mm)= Fiber Length and FS(g/tex)= Fiber Strength
[51]
Table 3. Seed Cotton Yield and Lint Yield Performance of Cotton Genotypes across test Location from 2016-2018 at
National Variety Trial
Seed Cotton Yield(t/ha) Lint Yield(t/ha)

Geno. Werer Nasa Farm Gewane Werer Nasa Farm Gewane
1 4.88bcde 4.03bcd 2.89abc 1.84c 1.48bcde 1.05bc
2 4.39e 3.48d 3.06abc 1.59c 1.28e 1.11bc
3 4.86bcde 3.97bcd 3.65ab 1.72c 1.419cde 1.21ab
4 4.58de 4.94ab 1.94c 1.59c 1.69abcd 0.64c
5 4.81bcde 3.83cd 3.55ab 1.72c 1.37de 1.25ab
6 5.48abc 4.57abc 3.15abc 2.22ab 1.79abc 1.25ab
7 5.83a 3.93bcd 3.62ab 2.29ab 1.50bcde 1.37ab
8 6.23a 4.32abcd 3.85ab 2.53a 1.73abcd 1.51ab
9 5.45abcd 4.39abcd 4.41a 2.19b 1.73abcd 1.76a
10 5.68ab 4.75abc 3.25abc 2.24ab 1.82ab 1.26ab
11 5.68ab 4.4abcd 3.15abc 2.23ab 1.71abcd 1.25ab
12 5.76a 4.32abcd 3.73ab 2.24ab 1.67abcd 1.43ab
13 4.84bcde 4.48abc 4.43a 1.87c 1.65abcd 1.62ab
14 4.61cde 4.02bcd 4.31ab 1.83c 1.53bcde 1.63ab
15 4.71cde 5.15a 2.76bc 1.81c 1.95a 1.04bc
Mean 5.19 4.31 3.45 1.99 1.62 1.29
LSD (0.05) 0.76 0.84 1.37 0.29 0.32 0.51
CV (%) 15.69 20.79 23.68 15.97 20.76 23.49
The genotypes variation for seed cotton yield and lint yield were presented in
Table 3. Mean performance among 15 cotton genotypes at Werer, Nasa farm, and
Gewane reported a wide range of seed cotton yield from 1.94 to 6.23 t ha-1 with a
mean of 4.09t ha-1and the maximum seed cotton yield obtained was 6.23t ha-1,
5.15t ha-1, and 4.43t ha-1, while, lint yield range 0.64-2.53 t ha-1 with a mean of
1.59t ha-1andmaximum lint yield of 2.53tha-1, 1.95tha-1 and 1.76t ha- 11,
respectively. Among locations, Werer exhibited the highest mean performance
genotypes namely, Chamo Farm no. 1A1-1 DP-90 F1#307(6.23t ha-1), Arbaminch
Farm no. 3-2 DP-90 F1#286(5.83t ha-1) and Weyto Farm no M1 DP-90
F1#376(5.76t ha-1) for seed cotton yield and lint yield. Seven, 9,8 genotypes
registered high mean values than the grand mean for seed cotton yield at Werer,
Nasafarm and Gewane, while 7, 9 and 6 for lint yield, respectively. Only one
genotypes namely, Chamo Farm no 1A1-1 DP-90 F1#307were performed better
than best-performed check variety for seed cotton yield at Werer while at Nasa
Farm and Gewane candidate variety were beaten by check variety. In addition,
Chamo Farm no. 1A1-1 DP-90 F1#307 was recorded highest values of seed cotton
yield and lint at Werer.
[52]
Table 4. Mean performances of fifteen Cotton genotypes for nine agronomic and quality traits across seven environments
Geno PH NBPP ABW SCY GOT LY M FL FS

1 114.71bc 18.76ab 4.85ef 4.23cd 36.98ef 1.57def 4.39cde 29.66c 29.4abc
2 90.99f 19.67ab 4.56f 3.81d 36.37f 1.39f 4.29de 28.29gf 26.66e
3 97.38ef 18.89ab 4.97de 4.31bcd 35.28g 1.52ef 4.26e 28.89de 26.6e
4 94.41f 17.55b 5.16bcde 4.36bc 34.32h 1.49ef 4.44cde 27.36i 26.92e
5 121.26 ab 17.43 b 5.47abc 4.21 cd 35.57 g 1.51 ef 4.31 de 30.29 b 29.99a
6 103.84de 19.88ab 5.32abcd 4.76abc 40.12a 1.89ab 4.7a 28.79de 28.77bcd
7 106.23cde 19.67ab 5.18bcde 4.71abc 38.60cd 1.83bc 4.62ab 29.18d 28.82bcd
8 105.98cde 20.76a 5.16bcde 5.07a 39.99a 2.04a 4.76a 27.93gh 28.53cd
9 105.54cde 18.89ab 5.0de 4.85ab 39.66ab 1.93ab 4.44cd 29.09de 28.75bcd
10 105.27 cde 19.2ab 5.26bcde 4.94 a 39.02 bcd 1.92 ab 4.69 a 28.56ef 28.41cd
11 108.16cd 19.27ab 5.33abcd 4.77abc 39.2bc 1.87abc 4.68a 27.78hi 28.18d
12 111.27cd 19.89ab 5.06cde 4.86ab 38.66cd 1.88ab 4.69a 28.91de 27.89d
13 105.14cde 19.04ab 5.27bcde 4.63abc 37.52e 1.74bcd 4.62ab 29.74c 28.12d
14 125.45a 17.23b 5.49ab 4.32bcd 38.85cd 1.67cde 4.42cde 31.01a 29.76ab
15 110.0cd 17.46b 5.69a 4.62abc 38.25d 1.76bcd 4.49bc 30.54ab 29.69ab
Mean 107.04 18.91 5.19 4.56 37.89 1.73 4.52 29.07 28.44
LSD (%) 8.31 2.3 0.36 0.48 0.7 0.18 0.16 0.47 0.93
CV (%) 12.76 19.99 11.43 17.22 3.05 17.32 5.81 2.68 5.39
Mean values followed by similar letter(s) in each column are not significant different each other. LSD (5%) = least
significant difference at P < 0.05, Geno. = Genotypes, CV (%) = coefficient of variation in percent, PH= Plant height,
NBPP= Number of boll per plant, ABW= Average Boll Weight, SCY (t/ha)= seed cotton yield, GP(%)= ginning percentage
and LY(t/ha)= lint yield, M=Micronaire, FL(mm)= Fiber Length and FS(g/tex)= Fiber Strength
Mean performance of most important traits in cotton that of seed cotton yield and
lint yield recorded 5.07 t ha-1 and 2.04 t ha-1, respectively. The mean values of 15
cotton genotypes ranged from 90.99-107.04 cm, 17.23-20.76, 4.56-5.69 gm for
plant height, number boll per plant and average boll weight, respectively (Table
4). Genotypes namely, Chamo Farm no 1A1-1 DP-90 F1#307, Farm no. Ago1
DP-90 F1#337 and Weyto Farm no. M1 DP-90 F1#376 had exceeded best-
performed check variety (Deltapine-90) with a percentage of 9,6 and 5%,
respectively for seed cotton yield. The maximum number of boll per plant, seed
cotton yield, lint yield and micronaire were recorded from Chamo Farm no 1A1-1
DP-90 F1#307 with good ginning outturn and fiber quality traits, however, the
high value of micronaire is undesirable traits in cotton breeding. From overall
mean values, 7, 5 and 7 genotypes were performed better than best-performed
check variety for seed cotton yield, ginning outturn and lint yield, respectively,
while four genotypes less performed than least performed check variety for
micronaire. On the other hand, no genotypes exceed the best check (Stam-59A)
for fiber length. The result showed that a wide range of variation in all agronomic
traits and quality traits.
The polygon view of the GGE-biplot analysis helps to detect cross-over and non-
crossover genotype-by-environment interaction and possible mega environments
in multi-location yield traits (Yan et al. 2007). Genotypes5, G2, G4, G15, G8 and
G10 were vertex genotypes (Figure 1) in which G5, G2, G4 and G15 were not a
[53]
winner in any environment. They are best in the environment lying within their
respective sector in the polygon view of the GGE-biplot (Yan and Tinker 2006);
thus these genotypes are considered specifically adapted. Genotypes close to the
origin of the axis have wider adaptation (Abay and Bjornstand, 2009). The
environments fall into two quadrants while the genotypes into six quadrants.
Genotypes 8 performed well in E7 (Gewane, 2018) and E1 (Werer, 2016) and
were moderately adapted to E6 (Nasafarm, 2018), whileG10 performed well in
E5(Nasa Farm,2017) and E4(Nasa Farm, 2016). Four vertex genotypes, G5, G2,
G4, and G15 had the highest yield in none of the environments.
Which Won Where/What
E1
0.5
E7
G5
G13 G7 G8
G9
G2 E6
G14 G11
AXIS2 19.96 %
G1
0.0
G3
G12 E3
G4 E2
G6
-0.5
G10
G15 E5
E4
-0.5 0.0 0.5 1.0
AXIS1 51.22 %
Figure 1. GGE bi-plot on genotypes using symmetrical scaling for mean lint yield and stability of 15cotton genotypes
tested across seven environments
The ideal genotype is located in the first concentric circle in the bi-plot. Desirable
genotypes are those located close to the center of the concentric circles (ideal
genotypes). Thus, starting from the middle concentric circle pointed with arrow
concentric circles were drawn to help visualize the distance between genotypes
and the ideal genotype (Yan and Tinker, 2006). An ideal genotype has the highest
mean yield and stable across environments (Farshadfar et al. 2012). The ideal
genotype can be used as a benchmark for selection. Genotypes that are far away
from the ideal genotype can be rejected in early breeding cycles while genotypes
that are close to it can be considered in further tests (Yan et al., 2007). Genotypes
near to the concentric circle namely, G12, G8, G9, G7, G6, G10 and G11 can be
[54]
desirable and benchmarks for evaluation of Cotton genotypes. Undesirable
genotypes were those distant from the first concentric circle, namely, G5, G2, G3
and G4 (Figure2).
Ranking Genotypes
E1
0.5
E7
G5
G13 G7 G8
G9
G2 E6
G14 G11
AXIS2 19.96 %
G1
0.0
G3
G12 E3
G4 E2
G6
-0.5
G10
G15 E5
E4
-0.5 0.0 0.5 1.0
AXIS1 51.22 %
Figure 2. The average- environment view to rank genotypes to ideal genotypes
Conclusion
The analysis of variance for genotype and genotype-by-environment interaction
effect was significant for cotton genotypes studied at Afar and the southern part of
Ethiopia. The interaction effect of genotypes and environment contributed most to
the variability most important traits of cotton that of seed cotton yield and lint
yield. Among tested genotypes, Chamo Farm no 1A1-1 DP-90
F1#307simultaneously best mean performance for seed cotton yield and lint yield
with good ginning outturn and quality traits, as well genotypes were close to the
ideal genotype and it was promoted to variety verification trial in 2019 cropping
season and submitted to variety releasing committee as candidate variety for
possible release for commercial cotton production.
[55]
Reference
Abay F, and Bjørnstad A. 2009. Specific adaptation of barley varieties in different
locations in Ethiopia. Euphytica, 167: 181–195.
Farshadfar E, Mohammadi R, Aghaee M, Vaisi Z. 2012. GGE biplot analysis of genotype
x environment interaction in wheat-barley disomic addition lines. Australian Journal
of Crop Science 6, 1074–1079.
Green, CC and Gulp TW. 1990. Simultaneous improvement of yield, fiber quality, and
strength in Upland cotton. Crop Sci. 30:66-69.
Jonathan, F, Wendel and Cronn R. 2003. Polyploidy and the Evolutionary History of
Cotton. Advances in Agronomy, 78, 139-186.
Kohel, RJ. 1999. Cotton germplasm resources and the potential for improved fiber
production and quality. p. 167-182. In A.S. Basra (ed.) Cotton Fibers. The Haworth
Press, inc., Binghamton, NY.
Singh, K, Wijewardana C, Gajanayake B, Lokhande S, Wallace T, Jones D and Reddy
KR. 2018. Genotypic variability among cotton cultivars for heat and drought
tolerance using reproductive and physiological traits. Euphytica. 214(3): 57.
Wakelyn, PJ, Bertoniere NR, French AD, Thibodeaux DP, Triplett BA, Rousselle MA,
Goynes WR, Edwards JV, Hunter L and McAlister DD. 2010. Cotton Fiber
Chemistry and Technology. CRC Press, New York, NY.
Yan, W and Tinker NA. 2006. Biplot analysis of multi-environment trial data: Principles
and applications. Canadian Journal of Plant Science, 86: 623–645.
Yan, W, Kang MS, Ma B, Woods S, Cornelius PL. 2007. GGE biplot vs. AMMI analysis
of genotype-by-environment data. Crop Science 47, 643-655.
Zhang, H-B, LiY, Wang B and Chee PW. 2008. Recent advances in cotton genomics.
International Journal of Plant Genome. 2008:1-20.
[56]
Evaluation and Stability Analysis of Cotton
(Gossypium hirsutum L.) Genotypes for Yield and
Fiber Quality Traits in Ethiopia
Merdasa Balcha, Bedane Gudeta, Arkebe G/Egziabher, Donis Gurmessa,
Samuel Damtew, Alehegn Workie and Mekashaw Arega
Ethiopian Institute of Agricultural Research, Werer Agricultural Research Center,
P. O. Box 2003, Addis Ababa, Ethiopia
Abstract
Cotton is an important natural fiber crop and grown mainly on the arid to semi-arid
regions of the tropic and sub-tropic areas of Ethiopia. Cotton traits are affected by
genotype, environment and genotype x environment interaction (GEI). The objectives
of this study were to study mean performance and stability analysis for agronomic and
quality traits in which the repeatability and the effect of the size of testing
environments in detecting cultivar genetic differences. Thus, fifteen cotton genotypes
were evaluated at two locations for three consecutive years for yield and quality-
related traits. The experiment was laid out in randomized complete block design with
three replications. Combined analysis for seed cotton yield, lint yield and some quality
traits revealed significant (P<0.01) difference among genotypes and genotype x
environment interaction. Considering the simultaneous selection of best mean
performance of agronomic traits and fiber quality traits with stability genotypes
Delcero x Deltapine-90 #F5-5-4-2-2 and Delcero x Deltapine-90 #F5-5-4-2-3 were
selected as adapted to a wide range of environments.
Keywords: Agronomic traits, GEI, lint yield, seed cotton yield
Introduction
Cotton (Gossypium hirsutum L.) is a world greatest in importance for fiber and
natural crop extending one of the worlds’ biggest textile industries having a yearly
economic impact (Ashraf et al., 2018). China, India, and the United States are the
top most leading countries in the world, whereas Mali and Burkina Faso are
among the top in Africa. Cotton is grown under peasant smallholdings (rain-fed
area) and large-scale commercial farms (irrigated area) in Ethiopia in which
production methods and problems vary considerably for both production areas.
There is vast potential to grow cotton both under rain-fed and irrigated conditions,
as Ethiopia has about 3,000,810 ha of land potentially available for cotton
production (EIA, 2017).
Genotype x environment interaction is of major importance to the plant breeder in

developing improved genotypes. When genotypes are compared over a series of
environments, the relative rankings usually differ. The strong GEI causes
[57]
difficulties in the selection of widely adapted and stable genotypes under diverse
environments. This causes difficulty in demonstrating the significant superiority of
any genotype. This interaction is usually present whether the varieties are pure
lines, single cross, double-cross hybrids, top crosses, S1 lines or any other material
with which the breeder may be working (Chinchane et al.,2018). Estimation of
phenotypic stability has a valuable tool in the assessment of varietals adaptability
and provides some measures of uniformity despite environmental fluctuations.
More knowledge about causes of GxE interaction is needed and would be useful
for establishing breeding objectives, identifying the best test condition, and
finding areas of optimal cultivar adaptation (Pinki et al, 2018). Among Stability
analysis, GGE bi-plot analysis increasingly being used in GxE interaction data
analysis in agriculture and the most effective tool for mega-environment analysis,
genotype evaluation and test-environmental evaluation (Kaya et al. 2006). The
newly developed GGE bi-plot methodology is a superior approach to the graphical
analysis of research data and may revolutionize the way researchers analyze data
and graphical tools for Breeders, Geneticists, and Agronomists. In addition, GGE
bi-plot enables the simplistic graphical visualization of complex G x E interactions
(Yan et al., 2000). Its analysis allows visual examination of the relationships
among the test environments, genotypes and the G x E interactions. Moreover,
GGE bi-plot is an effective tool used for mega-environment analysis (‘which-won-
where’ pattern), where specific genotypes can be recommended to specific mega-
environments (Yan and Tinker, 2006).
Many traits of economic worth are quantitatively inherited and manifest high
genotype-environment effects, therefore, the present study was carried out to
determine the mean average and stability of promising genotypes for identifying
the most stable genotypes for Variety Verification Trail program.

The experimental material comprised of twelve new upland cotton advanced lines
and three checks as listed in Table 1. All most all genotypes opted from the
breeding procedure of hybridization-based on pedigree selection. The experiments
were carried out at two different locations i.e. Werer at Werer Agricultural
research Center and Nasa-farm at the Southern part of Ethiopia for three
consecutive years viz. 2016, 2017, and 2018 of the main cropping season as a total
of six environments. The Experiment was laid out in Randomized Complete Block
Design (RCBD) with three replications. The size of the experimental plot was 5 m
long by 0.9 m wide with five rows which are 20 cm apart from each other as all
cultural practices were done manually throughout the entire growing season as
required. Data collected on a plot basis were seed cotton yield, ginning
[58]
percentage, lint yield, fiber length, fiber strength, and micronaire. Combined
analysis of variance of data collected for agronomic and quality traits was
subjected to the general analysis of variance using PROC GLM procedure in SAS
(SAS, 2014). GGE bi-plot pattern explorer performed using R statistical software.
The most recent method, the GGE (genotype main effect (G) plus G x E
interaction bi-plot model, provides breeders a more complete and visual
evaluation of all aspects of the data by creating a bi-plot that simultaneously
represents mean performance and stability, as well as identifying mega-
environments. The GGE bi-plot can be useful to display the which-won-where
pattern of the data that may lead to identifying high-yielding and stable genotypes
and discriminating and representative test environments. The presence of
significant genotype by environment interaction GEI (P<0.01) justified further
testing for GEI components using the GGE bi-plot analysis to enhance the
understanding of the effect and components.
Table 1. Description of Cotton genotypes used in the multi-environmental trials

Code Genotypes Location Year Environment Code
G1 Delcero x Deltapine90 #F5-5-2-2-1 Werer 2016 E1
G2 Delcero x Deltapine90 #F5-5-4-2-2 2017 E2
G3 Delcero x Deltapine90 #F5-5-4-2-3 2018 E3
G4 Delcero x Deltapine90 #F5-5-4-3-2 Nasa-farm 2016 E4
G5 Delcero x Krishna 16-2 2017 E5
G6 Delcero x Coker 315-3 2018 E6
G7 Delcero x Ionia-4
G8 Urania x Krishina 16-25
G9 HSC-99
G10 Europa x Stam 59A – 04-5
G11 Stam 59A x Gedera – 236-1
G12 Arbaminch Farm no 3-2 DP-90 F1#288
G13 Deltapine -90…Check
G14 Stam 59A……. Check
G15 Ionia…………..Check

The individual and Combined analysis of variance indicated that the variation in
all tested quantitative traits among genotypes were significant (Table 2). In
addition, the mean squares due to genotypes by location interaction for combined
analysis of variance revealed significant differences for three agronomic traits i.e.
seed cotton yield, ginning outturn and lint yield and one quality traits i.e.
micronaire, which illustrated the performance of the genotypes were not consistent
across the tested environment. On the other hand, mean squares due to genotypes
by year interaction for combined analysis of variance showed significant
differences for all traits except fiber strength, while seed cotton yield, lint yield,
and micronaire traits showed significant differences in all tested environment
[59]
(genotypes by years and locations). The significant effect of genotypes,
environment, and GEI suggested that genotypes exhibited different performances
in different environments i.e., the mean performance of the genotypes differed
from environment to environment and the genotypes responded differently to a
change in environment. The difference among the tested genotypes across the test
locations occurs due to their differences in genetic makeup or the variation due to
the environmental factors.
As yield traits are interacting with numerous external factors during the life span
of the plants in that of ranking genotypes based on seed cotton yield may be
considered as a reliable measure with the advantage of fiber quality traits. At
Werer, Delcero xDeltapine-90 #F5-5-4-2-3 had the highest mean seed cotton yield
of 6.19t/ha followed by Stam-59A x Gedera-236-1(6.16 t/ha) and
Delcero xDeltapine-90 #F5-5-4-2-2(6.12 t/ha) even Stam -59A x Gedera – 236-
1candidate variety had low ginning percentage. In addition, Delcero x Deltapine-
90 #F5-5-4-2-3(2.4 t/ha)showed better lint yield performance with good standard
fiber quality i.e. micronaire, fiber length and fiber strength, while, Delcero x
Ionia-4had showed high fiber length(32.17mm) and fiber strength (34.32g/tex) but
lowest in all agronomic traits i.e. seed cotton yield, ginning outturn and lint yield.
At Werer, 53.3% genotypes were exceeded total mean and only 33.3% of
genotypes beat best-performed check variety that of Deltapine-90 for seed cotton
yield.
Table 2. Mean Squares from the combined analysis of variance for six traits of 15 Cotton genotypes
Source Variation df SCY GOT LY FL FS M
Genotypes 14 5.47** 119.08** 1.2** 31.87** 61.76** 1.13**
Geno x year 28 0.76* 9.81** 0.13* 2.3* 3.12ns 0.17**
Geno x Loc 14 0.89* 7.26** 0.15* 1.83ns 2.06ns 0.18**
Geno x yrs x Loc 28 0.8* 5.27ns 0.13* 1.49ns 2.89ns 0.10*
R2 (%) 80.5 79.8 83.2 75.5 84.6 87.1
ns, * and **, non-significant, significant at P<0.05 and P<0.01, respectively, LSD (5%) = least significant difference, Geno.
= Genotypes, Loc=Location,yrs= year, CV (%) = coefficient of variation in percent, SCY (t/ha)= seed cotton yield, GP(%)=
ginning percentage and LY(t/ha)= lint yield, M=Micronaire, FL(mm)= Fiber Length and FS(g/tex)= Fiber Strength
At Nasa farm, Stam 59A x Gedera-236-1 (5.63t/ha), Delcero x Deltapine-90 #F5-

5-4-2-2 (5.11 t/ha) were among top genotypes for seed cotton yield in which only
genotypes exceed best-performed check variety (Deltapine-90) while, Stam-59A x
Gedera-236-1 (2.04t/ha), Arbaminch Farm no 3-2 DP-90 F1#288(2.02t/ha),
Europa x Stam-59A-04-5(1. 99 t/ha) were among high performed genotypes for
lint yield in which two former genotypes beat best-performed check variety that of
Deltapine-90. In quality traits, Delcero x Ionia-4(31.21gm/tex), Urania x Krishina
16-25(31.46mm) and Stam 59A x Gedera- 236-1(4.37%) best-performed
genotypes for fiber strength, fiber length and micronaire, respectively. In addition,
among 15 genotypes 53.3%, 40%, 60%, 46.67%, 46.67% and 46.67% of
[60]
genotypes had exceeded grand mean for seed cotton yield, ginning outturn, lint
yield, micronaire, fiber length and fiber strength, respectively.
Generally, the significant interaction effect suggests that seed cotton yield of
varieties varied across the tested environments. From tested genotypes, Delcero x
Deltapine-90 #F5-5-4-2-2 and Delcero x Deltapine-90 #F5-5-4-2-3 are best-
performed genotypes in seed cotton yield and lint with most important traits. On
the other hand, Europa x Stam 59A-04-5 had high ginning outturn and good lint
yield, but low fiber quality traits.
GGE bi-plot ‘Which-Won-Where’ Patterns Stability model for lint yield

One of the smartest features of a GGE biplot is its ability to show the which-won-
where pattern of a genotype by environment dataset and multivariate analytical
technique that graphically displays a two-way table and allows visualizing the
relation among genotypes, environments and their interactions (Kassaye et al.,
2017). The polygon view of the GGE bi-plot indicated the presence of a crossover
G x E interaction as the environments fell in different sectors of the polygon view
and had different high yielding genotypes (Yan and Kang, 2002). GGE bi-plot
graph the genotypes located at the vertex of the polygon in each sector were the
winner genotype at those environments whose markers fall into the respective
sector. Genotypes close to the origin of the axis have wider adaptation (Abay and
Bjornstand, 2009).
The GGE bi-plot has been used in crop genotypes trials to effectively identify the
best-performing genotype(s) across environments, identify the best genotypes for
specific environments delineation, whereby specific genotypes can be
recommended to specific environments and evaluate the yield and stability of
genotypes (Yan and Tinker, 2006). Environments within the same sector share the
same winning genotypes and environments in different sectors have different
winning genotypes while, the genotypes within the polygon and nearer to origin
were less responsive than vertex genotypes (Yan and Tinker, 2006). The “which-
won-where”(which is best for what) function of a GGE bi-plot is an extended use
of the pair-wise comparison function. In the present investigation, the GGE bi-plot
graphic analysis of fifteen Cotton genotypes revealed that on figure1with two
mega environments. Accordingly, the polygon drawn on five genotypes were G9,
G7, G6, G10 and G12in which furthest from the bi-plot origin in all other
genotypes are retained within the polygon. Hence, genotype 10 and 11 were
uniquely adapted in all most all environments, viz. E3, E2, E5, E6 and E1,
whereas genotype 12 won on E4 environment. The vertex genotypes of G6, G7
and G9 were not the best in any of the test environments due to there is not any
environment fell inside the sectors.
[61]
Table 3. Mean Performance of six traits of 15 Cotton genotypes tested at Nasa Farm and Werer in 2016-2018 cropping season
Werer Nasa Farm

Genotypes SCY GOT LY M FL FS SCY GOT LY M FL FS
1 5.38defg 37.63gh 2.02cde 4.84abcd 28.82f 29.71cde 4.1de 36.06cde 1.49bcd 3.99bcd 29.18c 27.16de
2 6.12abc 38.97cd 2.39a 4.74bcd 29.45def 29.11cde 5.11abc 37.83bcd 1.93a 3.96bcd 28.64cd 26.14ef
3 6.19a 38.77de 2.4a 4.89abc 29.92cdef 29.26cde 4.54cde 37.53bcd 1.71abc 4.19ab 29cd 27.61cde
4 6.04abcd 38.69def 2.33ab 4.77bcd 30.06bcde 28.78de 4.72bcd 37.85bcd 1.78ab 4.36a 29.12c 26.86de
5 5.27 efg 35.74 j 1.88 de 4.72bcd 28.68 f 28.11 ef 4.08 de 35.06 def 1.42 cd 3.96 bcd 27.85 de 25.38f
6 4.89fg 36.96hi 1.81e 4.04e 30.52bcd 31.14bc 3.92de 32.99f 1.28d 3.69de 30.39ab 27.81cd
7 4.19h 34.10k 1.43f 4.76bcd 32.17a 34.32a 4.05de 33.26ef 1.35d 3.84cde 31.19a 31.21a
8 4.74gh 39.61c 1.88de 4.09e 31bc 27.89ef 3.71e 36.5cd 1.34d 3.59e 31.46a 26.24ef
9 4.72gh 38.10efg 1.8e 4.95ab 30.67bc 30.98bc 3.96de 37.16bcd 1.49bcd 4.29ab 31.42a 28.16bcd
10 5.22 efg 43.72 a 2.29 ab 4.61d 26.36 g 26.64 f 4.36 cde 45.59 a 1.99 a 4.26 ab 26.99 e 25.24f
11 6.16ab 36.31ij 2.23abc 5.05a 28.75f 32.76ab 5.63a 35.99de 2.04a 4.37a 28.87cd 29.5b
12 5.73abcd 40.72b 2.33ab 4.86abc 29.17ef 30.6cd 5.04abc 39.87b 2.02a 4.09abc 29.75bc 28.9bc
13 5.49b-f 38.05efg 2.09bcd 4.82abcd 29.11ef 28ef 5.37ab 36.99bcd 1.99a 4.35a 28.7cd 25.29f
14 4.77gh 40.32b 1.93de 4.72bcd 31.23ab 31.17bc 4.5cde 39.17bc 1.77ab 4.25ab 31.16a 29.53b
15 5.47 cdef 37.95 fg 2.07 bcde 4.68cd 30.09 bcde 29.87 cde 4.74 bcd 36.32 cd 1.73 abc 3.74 de 30.37 ab 27.12de
Mean 5.36 38.38 2.06 4.7 29.73 29.89 4.52 37.21 1.69 4.06 29.61 27.48
LSD(%) 0.61 0.69 2.39 0.21 1.08 1.82 0.73 2.67 2.87 0.3 1.1 1.32
CV (%) 12.11 1.93 12.38 4.69 3.86 6.48 17.26 7.67 18.13 7.93 3.82 5.11
Mean values followed by similar letter(s) in each column are not significant different each other. LSD (5%) = least significant difference at P < 0.05, CV (%) = coefficient of variation in
percent, SCY (t/ha)= seed cotton yield, GP(%)= ginning percentage and LY(t/ha)= lint yield, M=Micronaire, FL(mm)= Fiber Length and FS(g/tex)= Fiber Strength
[62]
Thus, varieties G4, G3, G11, and G10 were widely adopted. Genotype G10 was
the vertex genotypes in environments of E3, E5, E2, E6 and E1, as they had wide
adaptation while G12 to E4 with specific adaptation. All most all environments
i.e. E3, E5, E2, E6 and E1 share the same sector with an indication of the same
winning genotypes, while environments in different sectors have different winning
genotypes. The other seven genotypes were fallen in sectors with no environment
markers.
Which Won Where/What

10
E4
G9
5
AXIS2 12.62 %
G14 G12
G7 G13G2
G5 G15
0
E3
G4G11
G3 E2 E6
G8 G1 E5
G10
E1
G6
-5
-5 0 5 10 15
AXIS1 68.96 %
Figure 1. Polygon view of GGE (genotype plus genotype by environment interaction) bi-plot for the “which-
won where” pattern of genotypes and environments.
Genotypes located on the vertices of the polygon performed either the best or the
poorest in one or more locations since they had the longest distance from the
origin of bi-plot (Yan and Rajcan, 2002). Therefore, among the vertex varieties,
G10 and G12 were identified as the high yielding varieties, while G6, G7 and G9
were considered as the low yielding varieties across the testing environments and
highly specific adapted genotypes, as they had the longest distance from the origin
(Figure 1). Genotypes G15, G2, G3, G4 and G13 were located within polygons
[63]
which were less vulnerable to environmental changes across the test
environments. Genotypes 15, G5 and G14 were located near to the origin which
indicated that these genotypes were stable or wider adapted across the test
locations.
Conclusion
The present study was carried out in 15 genotypes with over three locations to
study stability in the performance of different environmental conditions for
various economic traits. The mean squares due to genotype were significant for all
agronomic and quality traits, while the genotype × environment mean squares
were significant for seed cotton yield, lint yield and micronaire indicating different
responses of the genotypes in different environments. Mean performance of
genotypes Delcero x Deltapine90 #F5-5-4-2-2(G2) and Delcero x Deltapine90
#F5-5-4-2-3(G3) were best-performed genotypes in seed cotton yield, lint yield,
and other most important fiber quality traits. GGE Biplot analysis makes this
useful technology accessible on a wider scale to plant breeders, geneticists,
agronomists, ecologists and other related research areas using which-won-where
model under this experiment. Hence, seven genotypes namely G15, G14, G2, G3,
G4 and G13 were found more stable for lint yield over the years and location. As
far as G3 and G2 had a good mean average in all traits and somewhat stable over
the different environmental conditions. Therefore, these genotypes were promoted
to variety verification trials in 2019 cropping season and submitted to variety
releasing committee as candidate varieties to be released for commercial cotton
production purposes.
Reference
Abay F and Bjornstad A. 2009. Specific adaptation of barley varieties in different locations in
Ethiopia. Euphytica, 167: 181–195.
Ashraf, J, ZuoD, Wang Q, MalikW, ZhangY, Abid MA and Song G. 2018. Recent insights into
cotton functional genomics: progress and future perspectives. Plant Biotechnology. 16(3),
699.
AYA, YM and Taner S.2006: GGE-bi-plot analysis of multi-environment yield trials in Bread
Wheat. Turkish Journal of Agriculture & Forestry 30: 325–337. Introduction to Biplots for
GxE Tables. University of Queensland, Brisbane
Blanche, SB, Myers GO and Kang MS. 2007. GGE Bi-plots and Traditional Stability Measures for
Interpreting Genotype by Environment Interactions. Journal of Crop Improvement, 20: 123-
135.
Chinchane, VN, Deosarkar and Kalpande HV. 2018. Stability Analysis for Seed Cotton Yield and
its Component Traits in Hybrids of Desi Cotton (Gossypiumarboretum L.). International
Journal of Current Microbiology and Applied .7(9): 1000-1012.
Ethiopia Investment Agency. 2017. Cotton, Textile and Apparel Sector Investment Profile
Summary Ethiopia.
Pinki, SS, Siwach RS, Sangwan, Sombir Singh, Mor VS, Shiwani Mandhania, Sunayana and Neha.
2018. Stability Analysis for Seed Cotton Yield and Its Components in Upland Cotton
[64]
(Gossypium hirsutum L.), International Journal of Current Microbiology and Applied
Sciences, 7(4): 1630-1638.
Yan W and Tinker NA. 2006. Biplot analysis of multi-environment trial data, Principles and
applications. Canadian Journal of Plant Science, 86: 623-645.
Yan, W, Hunt LA, Sheng Q, and Szlavnics Z. 2000. Cultivar evaluation and mega-environment
investigation based on the GGE bi-plot. Crop Science, 40:597-605.
[65]
[66]
Genetic Diversity Study of G. barbadense L. and
Interspecific (G. hirsutum x G. barbadense)
Hybrid Cotton Germplasm using SSR Markers
Donis Gurmessa1, Kassahun Bantte2 and Kefyalew Negisho1
1
Ethiopian Institute of Agricultural Research. P.OBBox: 2003, Addis Ababa, Ethiopia
2
Jimma University College of Agriculture and Veterinary Medicine. Post office box: 307, Jimma University,
Ethiopia
Abstract
This study was initiated to assess the genetic diversity among and within G.
barbadense L. and interspecific hybrid cotton germplasm, using SSR markers. A
total of 25germplasm were genotyped using 26 SSR markers. The pair-wise genetic
distance among the germplasm was calculated using Dice genetic distance measure
by DARwin 6.0.18 software and the distances were used to construct Neighbor-
Joining tree using MEGA 7 software. A total of 166 alleles, with a range of 3 to 13
alleles and a mean of 6.38 per markers were detected. Among the complete panel of
cotton germplasm used in this study, the pair-wise genetic distance ranged from 0.08
to 0.94, with an overall mean of 0.62. The result indicates a narrow to a broad range
of genetic distances and overall, a reasonable genetic diversity among and within
germplasm which could be used to further broaden the genetic base and to enlarge
the number of available cotton germplasm.
Introduction
Cotton (Gossypium spp.) belongs to the Malvaceae family and tribe Gossypieae in
the genus Gossypium L.(Fryxell, 1979). It is widely considered, the top most basis
of natural fiber and the chief cash crop and lifeline of the textile industry globally
(Riaz et al., 2013). The place of origin of the genus Gossypium is enigmatic,
however, the primary centers of diversity for the genus are west-central and
southern Mexico, north-east Africa and Arabia and Australia. DNA sequence data
from the existing Gossypium species suggests that the genus arose about 10-20
million years ago (Wendel and Albert, 1992; Seelanan et al., 1997). Among the
species so far, four are cultivated species that include the New World
allopolyploids G. hirsutumL. and G. barbadense L. and Old World diploids G.
herbaceum L. and G. arboreum L. (Brubaker et al., 1999).
The practice of plant improvement is harnessing and manipulating heritable

variation to develop improved genotype or population of genotypes that make up a
cultivated variety. This practice is bounded by a range of available genetic base.
The wider genetic diversity of crop species provides an opportunity to improve
complex traits of interest by combining or pyramiding genetic variations within a
population, so that particular research problems may be addressed (Van Esbroeck
et al., 1998, 1999). Agronomic challenges and priorities change, ecosystems now
[67]
supposed safe may need rehabilitation in the future. Plant material not deemed
particularly valuable today may turn out to be crucial tomorrow so that the more
genetic diversity in plant germplasm provides insurance and the better the chances
of fulfilling future needs (Engels et al., 1995).
Understanding the range of genetic diversity in plants is critical for the effective
conservation, management, and effective utilization of plant germplasm. Some
germplasm accessions available for cultivated plants are duplicates while others
include rare alleles or very unusual allele combinations. In particular, knowledge
of existing genetic diversity, where the plant population is found and how to best
utilize it, is of fundamental importance for basic science and applied aspects
(Mondini and Pagnotta, 2009; Brumlop and Finckh, 2010). The choice of a parent
is the first basic step in a meaningful breeding program. Thus genetic diversity
study identifies parental combinations exploitable to generate segregating
progenies with maximum genetic potential for further selection, as proven by
Akoroda 1987; Aremu et al., 2007; Rahim et al., 2010). Dje et al., 2000,
suggested that the higher the genetic distance between parents, the higher the
heterosis in the developed progenies. Hence the heterotic progenies can be further
hybridized and selections based on transgressive segregation. Overall elucidating
the details of genetic diversity in the available plant germplasm is one of the first
steps to understanding the preservation and the efficient utilization of plant
germplasm for improvement purposes (Abdurakhmonov et al., 2012; Ulloa et al.,
2013). Hence this study was initiated to analyze the extent of genetic diversity
among and between G. barbadense L. and Interspecific (G. hirsutum x G.
barbadense) hybrid cotton germplasm using SSR markers.

In this study, a total of 12 G. barbadense L. and 13 interspecific (G. hirsutum x G.
barbadense) hybrid cotton germplasm were used. The list of germplasm used is
given in Table 1.
[68]
Table 1. List of cotton germplasm used in the study.
S. no Germplasm name Designated code

1 GL-7 G1
2 HS-46 G2
3 HS-46#078 G3
4 HTO#052 G4
5 HTO#052 X LS-90 24-11 G5
6 HTO#052 X Pima S3 22-4 G6
7 LS-90 G7
8 LS-90 X G-45 6-3 G8
9 LS-90 X Pima S3 5-7 G9
10 Pima S-3 G10
11 Pima S-5 G11
12 HTO#52 X LG-450 23-7-2-2-1 G12
13 Cucurova1518 X LG-450 35-4 G81
14 Delcero x GL-7 8-2 G82
15 G-45 X Cucurova1518 10-9 G83
16 GL-7 x Stam 59 A 27 -9 G84
17 HS-46 X Deltapine-90 G85
18 HS-46 x Stonevile 453 19-2 G86
19 HS-46 x Stonevile 453 19-8 G87
20 HTO#052 X DP-90 21-7 G88
21 HTO#052 X Turkey I G89
22 HTO#052 X Stam 59A 19-6 G90
23 Pima S3 X Stam 59A 17-4 G91
24 Cucurova1518 x Pima 34-5-5-2-1 G92
25 G-45 x DP-90# F5-11-1-3-4-1 G94
SSR Markers
Initially, based on chromosome coverage, 46 SSR markers were selected from the
Cotton Marker Database of CottonGen (http://www.cottongen.org). The annealing
temperatures of primers were optimized by performing gradient PCR (Altay
Scientific S.P.A, Italy). Finally, 26 polymorphic SSR markers were selected
excluding 17 markers that failed to amplify properly and five monomorphic
markers. Details of the SSR markers used for this study are shown in Table 2.
[69]
Table 2. SSR markers used in the study
Name
S. No Location Motif (s) Primers (5' - 3') AT (oC)
1 BNL2960 AD_Chr.10 GA F- TAAGCTCTGGAGGCCAAAAA 52
R- CCATTTCAATTTCAAGCATACG
2 BNL0946 AD_Chr.10, GA F- GCTGTTGCTCCACATCTCCT 58
AD_Chr.20 R- GGGCAAACAGATAGGCAGAA
3 JESPR0274 AD_Chr.9, CA F- GCCCACTCTTTCTTCAACAC 57
AD_Chr.23, R-TGATGTCATGTGCCTTGC
AD_Chr.26
4 BNL3545 AD_Chr.02, CA F- AGTCAGTTTTTTGTTAGCAATATGC 52
AD_Chr.14 R- AACCATTAATTCCCTATTTAACCG
5 DPL0196 AD_Chr.04, GA F- CACATTTGGTGGGTATTGAGAAG 56
AD_Chr.22 R- ACCATACACGTGCTAATGTCAAAG
6 BNL3594 AD_Chr.06, TC F- AGGGATTTTGATTGTTGTGC 51
AD_Chr.25 R-TGAATTCAAAACAAATGTTAGCC
7 NAU1119 AD_Chr.26 GCA F- CCCCAACAAACTGAAAAATC 52
R- AGTTCTGTTGCCTGGGTTAG
8 NAU3401 AD_Chr.12 ATC F- ATGCCGACGCTTTAAGTAAC 52
R- CGATATGGGCATGTTTGATA
9 NAU5189 AD_Chr.23 TTC F- TGTCCCCCAATCATATTTTC 52
R- CAACTTCCCAAGCTCGTATT
10 BNL4071 AD_Chr.05, GT, F- CATTTCAGAAGTTGACATTTTCG 52
AD_Chr.19 GA R- CACTGCCCCTAAGAAGTTGC
11 BNL0569 AD_Chr.13, AG F- TTGAGAAGTACTACCATTAATTATCCA 52
AD_Chr.18 R- GACTGATGCCAGTTGACCCT
12 NAU1070 AD_Chr.03, AGG F- CCCTCCATAACCAAAAGTTG 54
AD_Chr.14 R- ACCAACAATGGTGACCTCTT
13 BNL3065 AD_Chr.16 AG F- CAAACGGGAGACCAAAAAAA 54
R- CGAACTGGCGAGTTAGTGCT
F- TTCAATTCTGGGTTCGAGCT 55
14 BNL3985 AD_Chr.23 TC
R-CACCCATCAACCCAAATTTC
15 BNL1417 AD_Chr.25 AG F- TTATTCTAACCACCGCCTCC 58
R- TGAGTGGATATGCTTGGCCT
16 BNL1721 AD_Chr.18 AG
F- TGTCGGAATCTTAAGACCGG 57
R- GCGCAGATCCTCTTACCAAA
17 BNL3371 AD_Chr.17 AG F- CAATCCTTTACGTGGCCTGT 58
R- AAAGACAGGCAATCCCCTTT
18 BNL3644 AD_Chr.14 TC F- GTGCTGTTTGGGCCTTACAT 58
R- TAAGCGCATTGACACACACA
19 DPL0717 AD_Chr.11, TA, F- CTCCATGATTTCTGAAACACAGGT 57
AD_Chr.21 TTG R- TTGTACTAACATACCTCCGGGTCT
20 JESPR0220 AD_Chr.4 GA F- CGAGGAAGAAATGAGGTTGG 56
AD_Chr.22 R- CTAAGAACCAACATGTGAGACC
21 BNL3992 AD_Chr.05, TC, F- CAGAAGAGGAGGAGGTGGAG 55
AD_Chr.19 GA R- TGCCAATGATGGAAAACTCA
22 BNL1672 AD_Chr.09, AG F- TGGATTTGTCCCTCTGTGTG 57
AD_Chr.23 R- AACCAACTTTTCCAACACCG
23 BNL2634 AD_Chr.07, AG F- AACAACATTGAAAGTCGGGG 56
AD_Chr.16 R- CCCAGCTGCTTATTGGTTTC
24 NAU3665 AD_Chr.10, AAT F- CAGCATGGAAATCCTAATCC 54
AD_Chr.20 R- TGAACTAGCTTGGCTGAATG
25 NAU1297 AD_Chr.20 TTC F- CCGCTGCAAAATTCTCTTAC 55
R- CCCACTGGACATTTCTATCC
26 NAU3589 AD_Chr.18 AT F- TGCAAATATTTCATCCCAAG 50
R- GGAGCCACCAAAGAAACTTA
[70]
DNA Extraction
Seeds of the varieties were sown in plastic pots in the greenhouse. Leaves were
harvested from 12 days old seedlings and stored in a freezer (-80OC) to avoid
degradation until DNA was extracted following, Xin et al. (2003). A small piece
of leaf was collected in a 96 well PCR plate and 50µl of 0.1M NaOH and 2%
Tween 20 (v/v) was added and incubated for 10 minutes at 95OC using PCR
machine. Subsequently, 50µl of 0.1M Tris-HCl and 2mM EDTA was added and
gently mixed at a moderate speed of 550 revolutions per minute on vortex. A
0.5µl extract was used as a source of template DNA.
Polymerase Chain Reaction

The PCR was also performed according to Xin et al. (2003) with some
modifications. The total volume was 10µl mixture, composed of 0.1% Bovine
serum albumin (w/v), 0.25 µl each primer pairs [10µM], 0.5 µl crude DNA
templates, 1% Polyvinylpyrrolidone-40 (w/v), 2.9 µl Nuclease free water and 5 µl
One Taq® 2X Master Mix with standard buffer (New England BioLabs ®Inc.). The
thermal cycling conditions of the PCR was 2 minutes of initial DNA denaturation
at 940C followed by 30 cycles consisted of 15 seconds of DNA denaturation at
940C 15 seconds of primer annealing at the indicated temperature on Table 2 and
30 seconds of primer extension at 680C. The PCR cycling condition was
terminated with a final extension time of 5 minutes at 680C and the amplified
product was hold at 40C.
Gel Electrophoresis
The amplified DNA was mixed with 6X loading dye with gel red in the ratio of
3:2 (v/v) respectively. With DNA ladder on the adjacent lane and the amplified
products were loaded on 3.5% Hi-Res agarose gel in 0.5X TBE buffer
(Abdurakhmonov, 2007, Xin et al., 2003; Wangari et al., 2013) and gels were
run for 8 hours at 65V. Finally, the amplified bands were documented under
ultraviolet light (Bio-Rad, USA).
Data Scoring and Analysis

SSR profile within the expected product size range was manually scored
according to estimated molecular size compared with the DNA ladder and in the
form of binary data (1 for presence and 0 for absent allele). Missed value was
scored as -9. The allele number, allele frequency and polymorphic information
content (PIC) of the SSR markers were calculated using the Power Marker
software program, version 3.25 (Liu and Muse, 2005). The PIC of SSR markers
was calculated as per Botstein et al. (1980) using the following formula,
Where, is the frequency of ith allele

[71]
Binary data was used to calculate the pairwise genetic dissimilarity among the
varieties based on Dice dissimilarity (Dice, 1945) using DARwin software,
version 6.0.18 (Perrier and Jacquemoud-Collet, 2006).
Dij=
Where,
Dij = Dice dissimilarity indices
a= number of alleles present in both i and j germplasm
b= number of alleles present in i and absent in j
c= number of alleles absent in i and present in j
The pair-wise Dice dissimilarity matrix generated were used to cluster germplasm
by Neighbor-joining algorithm (Saitou and Nei, 1987) using MEGA software,
version 7.0.26 (Kumar et al., 2016).

Diversity among SSR Markers
A total of 166 SSR alleles were produced across 26 loci with a mean number of
alleles 6.38 (Table 3). The detected number of alleles ranged from 3 to 13 and the
lowest number of alleles produced belongs to each of BNL1721 and BNL 2960
locus while the highest number of alleles was detected in BNL1672 locus. The
mean frequency of the major allele of markers was 0.39 and within the range of
0.16 to 0.76. Other studies reported a lower mean number of alleles per locus. For
instance, Zhang et al. (2011) reported 5.08 alleles per locus (2 to 14) in 59 cotton
cultivars of China using 40 SSR markers. Lacape et al. (2007) reported a mean of
5.6 alleles per locus (2 to 17) evaluating 47 accessions of the Gossypium genus
through 320 SSR markers. Overall, the author suggested the number of alleles
depends on markers to be used, a platform used for resolution of amplified
products and plant material to be genotyped. The relatively higher result of this
study is probably due to the germplasm of different species used. The
polymorphic information content (PIC), of the markers, varied from 0.38 to 0.88
with a mean value of 0.68. The highest PIC value of 0.88 belongs to BNL1672
followed by 0.82 in BNL3401 and 0.81 in BNL3545 while the lowest of 0.38 was
recorded in BNL1417. In other studies, more or less similar results of mean PIC
value, 0.62 (Guang and Xiong-Ming, 2006) and 0.65 (Moiana et al., 2012) were
reported. In other studies higher mean PIC of 0.80 reported by Zhang et al. (2011).
[72]
Table 3. SSR marker indices used in the study.
Major allele Number of Polymorphism

S. no Marker name frequency allele information content
1 BNL2960 0.37 3 0.58
2 BNL946 0.36 9 0.76
3 JESPR274 0.33 7 0.76
4 BNL3545 0.26 10 0.81
5 DPL0196 0.32 9 0.78
6 BNL3594 0.35 4 0.64
7 NAU1119 0.53 5 0.52
8 NAU3401 0.25 10 0.82
9 NAU5189 0.50 4 0.60
10 BNL4071 0.28 7 0.74
11 BNL569 0.34 5 0.69
12 NAU1070 0.38 5 0.66
13 BNL3065 0.60 5 0.55
14 BNL3985 0.33 6 0.74
15 BNL1417 0.76 4 0.38
16 BNL1721 0.64 3 0.44
17 BNL3371 0.37 8 0.74
18 BNL3644 0.35 5 0.68
19 DPL0717 0.40 6 0.71
20 JESPR220 0.28 7 0.79
21 BNL3992 0.27 8 0.80
22 BNL1672 0.16 13 0.88
23 BNL2634 0.40 7 0.65
24 NAU3665 0.59 5 0.58
25 NAU1297 0.25 6 0.75
26 NAU3589 0.36 5 0.70
Mean 0.39 6.38 0.68
Genetic relationships among G. barbadense L. germplasm

Pair-wise genetic distance among G. barbadense L. germplasm used in this study
is presented in Table 4. The pair-wise genetic distance ranged from 0.08 to 0.74
and the overall mean genetic distance was 0.43. Among the germplasms HTO
#052 (G4) and HTO#052 X LS-90 # 24-11 (G5), had the maximum genetic
distance whereas LS-90 X G-45 #6-3 (G8) &LS-90 X Pima S3 #5-7(G9), and
Pima S-5 (G11) &HTO#52 X LG-450 #23-7-2-1 (G12), had the minimum genetic
distance between them. The result of this study indicates low to high pair-wise
genetic divergence among the germplasm and overall a fairly moderate genetic
diversity within the germplasm. Dje et al. (2000), suggested that the higher the
genetic distance between parents, the higher the heterosis in their hybrids. Hence,
G5, G1, G3, and G2 could produce the more diverse progeny. Moreover, this
germplasm could be potential plant materials to solve the scarcity of G.
barbadense L. germplasm. On the other hand, some of the germplasm exhibiting
relatively, lower value of genetic distance (G8 to G12) exhibited moderately high
genetic distance, particularly with G5. Consequently, emphasis should also be
given to the conservation of this germplasm for future use.
[73]
Wu et al. (2010) studied 20 germplasm of G. barbadense L. species using 108
SSR markers and reported an overall mean genetic distance of 0.19. de Almeida et
al. (2009) reported a higher overall mean genetic distance of 0.39 among 26
germplasm of G. barbadense L. by using SSR markers. Comparatively higher
overall mean genetic distance among the germplasm used in this study could be
because of the germplasms used were from different sources.
Table 4. Dice dissimilarity matrix among G. barbadense L. species germplasm.
GBSG G1 G2 G3 G4 G5 G6 G7 G8 G9 G10 G11

G2 0.64
G3 0.62 0.15
G4 0.38 0.55 0.55
G5 0.72 0.54 0.56 0.74
G6 0.53 0.58 0.62 0.44 0.62
G7 0.57 0.56 0.61 0.4 0.62 0.17
G8 0.39 0.52 0.53 0.29 0.65 0.32 0.28
G9 0.41 0.54 0.56 0.25 0.72 0.31 0.29 0.08
G10 0.49 0.5 0.42 0.4 0.56 0.44 0.43 0.31 0.22
G11 0.41 0.51 0.51 0.29 0.65 0.3 0.29 0.18 0.13 0.22
G12 0.42 0.44 0.45 0.29 0.66 0.33 0.35 0.21 0.12 0.17 0.08
GBSG = G. barbadense L. species germplasm
Genetic relationships among interspecific

(G. hirsutum L. x G. barbadense L.) hybrids
The observed pair-wise genetic distance exhibited among interspecific hybrids
(IH) ranged from 0.12 to 0.68 (Table 5), and the overall mean genetic distance
was 0.45. The highest pair-wise genetic distance observed was between each of
Delcero x GL-7 #8-2 (G82) and HTO#052 X Stam 59A #19-6 (G90) and Pima S3
X Stam 59A 17-4 (G91). The least genetic distance was observed between HS-46
X Deltapine-90 (G85) and HS-46 x Stonevile 453 #19-8 (G87) and they were the
crossing results of germplasm introduced from the USA. The result indicates a
low to moderately high available genetic base and overall, a modest genetic
variation among hybrids. Probably selection work targeting a specific and or few
breeding objectives and a limited number of parental lines they are derived from
contributed to modest genetic diversity. Overall, G90, G91, G89 and G94 could be
the best interspecific hybrid germplasm for further use.
[74]
Table 5. Dice genetic dissimilarity matrix among interspecific (G. hirsutum L. x G. barbadense L.) hybrids.
IH G81 G82 G83 G84 G85 G86 G87 G88 G89 G90 G91 G92
G82 0.32
G83 0.34 0.4
G84 0.43 0.51 0.36
G85 0.35 0.38 0.31 0.31
G86 0.37 0.43 0.38 0.35 0.24
G87 0.32 0.38 0.29 0.37 0.12 0.27
G88 0.27 0.48 0.32 0.38 0.24 0.28 0.24
G89 0.57 0.57 0.55 0.46 0.48 0.47 0.51 0.52
G90 0.56 0.68 0.51 0.54 0.52 0.59 0.52 0.48 0.66
G91 0.52 0.68 0.66 0.6 0.59 0.59 0.56 0.54 0.63 0.41
G92 0.42 0.45 0.49 0.43 0.35 0.33 0.35 0.33 0.39 0.57 0.47
G94 0.48 0.51 0.49 0.58 0.42 0.53 0.39 0.37 0.51 0.66 0.56 0.33
The Dice dissimilarity matrix among G. barbadense L. and interspecific hybrid

germplasm is given in Table 6. The pair-wise genetic distances were within 0.68
and 0.94 range indicating a broad genetic divergence. The highest pair-wise
genetic distance observed among G. barbadense L. and interspecific hybrids were
between HS-46 (G2) and Pima S3 X Stam 59A #17-4 (G91) and the least
between, GL-7 (G1) and HTO#052 X Turkey I (G89). The result indicates the
availability of broad genetic diversity among G. barbadense L. and interspecific
hybrid germplasm. Furthermore, G91, G94 and G89 of interspecific hybrids and
G6, G5 and G7 of G. barbadense L. germplasm could be the best germplasm.
Table 6. Dice genetic dissimilarity matrix between G. barbadense L. interspecific

(G. hirsutum L. x G. barbadense L.) hybrids
Germplasm G1 G2 G3 G4 G5 G6 G7 G8 G9 G10 G11 G12

G81 0.83 0.77 0.79 0.75 0.81 0.80 0.74 0.73 0.71 0.72 0.75 0.80
G82 0.80 0.71 0.70 0.78 0.70 0.85 0.82 0.73 0.76 0.69 0.72 0.77
G83 0.78 0.69 0.70 0.80 0.81 0.83 0.79 0.78 0.77 0.72 0.71 0.74
G84 0.81 0.75 0.77 0.76 0.91 0.88 0.82 0.79 0.77 0.79 0.73 0.81
G85 0.83 0.71 0.72 0.78 0.86 0.88 0.82 0.80 0.79 0.79 0.78 0.80
G86 0.83 0.75 0.77 0.75 0.87 0.83 0.77 0.75 0.74 0.76 0.72 0.78
G87 0.80 0.74 0.73 0.77 0.85 0.87 0.81 0.77 0.75 0.74 0.74 0.76
G88 0.78 0.76 0.76 0.79 0.86 0.84 0.78 0.75 0.73 0.74 0.78 0.80
G89 0.68 0.78 0.78 0.81 0.88 0.92 0.86 0.74 0.77 0.77 0.78 0.78
G90 0.71 0.86 0.88 0.82 0.83 0.82 0.73 0.79 0.78 0.83 0.83 0.82
G91 0.79 0.94 0.90 0.81 0.85 0.89 0.83 0.79 0.83 0.90 0.88 0.87
G92 0.77 0.76 0.73 0.77 0.85 0.84 0.81 0.69 0.73 0.78 0.77 0.79
G94 0.77 0.77 0.73 0.81 0.83 0.92 0.83 0.74 0.81 0.77 0.86 0.84
The overall mean genetic distance among cotton germplasm used in this study was
0.62. As shown in Table 4-6, the pair-wise genetic distance ranged from 0.08 to
0.94. The result indicates a narrow to a broad range of genetic distances and
overall, a moderately broad genetic diversity is available among the germplasm.
Overall, the result showed a wide genetic distance among the germplasm used in
this study and this is mainly between G. barbadense L. species and or interspecific
hybrid cotton germplasm. The result of this study is more or less similar to that of
[75]
Sapkal et al. (2011) who reported 0.57 mean genetic distances among 91 cotton
accessions of G. hirsutum L., 50 maintainer and 41 restorer lines using 26 SSR
markers.
The Neighbor-joining cluster analysis showed the germplasm could be assigned to

three clusters. G. barbadense L. germplasm was assigned to two clusters (Cluster I
to Cluster II) and interspecific hybrid germplasm makes one cluster (Cluster III).
NJ dendrogram illustrating genetic relationship within G. barbadense L. and
interspecific hybrid cotton germplasm is given in Figure 1.
G1
G4
G8
G9
G10 Cluster I
G11
G12
G6
G7
G2
G3 Cluster II
G5
G81
G82
G83
G84
G86
G85
G87 Cluster III
G88
G91
G92
G94
G89
G90
0.10
Figure 1. NJ dendrogram showing genetic relationship within G. barbadense L. and interspecific hybrid
germplasm.
[76]
The intra and inter-cluster mean genetic distance among cotton germplasm is
depicted in Table 7. The mean intra genetic distance within varieties of each
cluster was 0.34 (Cluster I), 0.40 (Cluster II) and 0.41 (Cluster III). The lowest
mean inter-cluster genetic distance of 0.61was between germplasm of cluster I and
cluster II, followed by 0.79 (cluster I and III) and 0.80 (cluster II and III).
Table 7. Intra and inter-cluster genetic distance among cotton varieties.
Cluster I II III
I 0.34
II 0.61 0.40
III 0.79 0.80 0.41
In conclusion, The SSR markers discriminated against all the germplasm as there
was no 100% genetic similarities. Overall there is a reasonable genetic diversity
among and within germplasm which could be used to further broaden the genetic
base and enlarge the number of available cotton germplasm.
References
Abdurakhmonov, IY. 2007. Exploiting genetic diversity. pp. 2153. Proceedings of World Cotton
Research Conference-4. Lubbock, Texas, USA, 10-14 September 2007.
Abdurakhmonov, IY, Buriev ZT, Shermatov ShE, Abdullaev AA, Urmonov K, Kushanov F,
Egamberdiev ShS, Shapulatov U, Abdukarimov A, Saha S, Jenkins JN, Kohel RJ, Yu JZ
Pepper AE, Kumpatla SP and Ulloa M. 2012. Genetic Diversity in Gossypium genus. Chapter
16. In: M.Caliskan (ed.). Genetic Diversity in Plants. INTECH. Internet document available at,
https://cdn.intechopen.com/pdfs-wm/31484.pdf,
Akoroda, MO. 1987. Principal component analysis metroglyph of variation among Nigerian
yellow yam. Euphytica, 32:565-573.
Aremu, CO, Adebayo TA, Oyekunle M and Ariyo OJ. 2007. The relative discriminatory abilities
of techniques measuring Genotype x Environmental interaction in soybean (Glycine Max (L)
Merr.) in semi-arid and rain forest environments of Nigeria. Agricultural Journal,2(2):210-
215.
Botstein, RL, White MS and Davis RW. 1980.Construction of a genetic linkage map in man using
restriction fragment length polymorphisms. American Journal of Human Genetics, 32
(3):314–331.
Brubaker, CL, Bourland FM and Wendel JE. 1999. The origin and domestication of cotton. In:
Smith, W. C., and Cothren, J. T. (ed) Cotton: Origin, History, Technology and Production,
John Wiley and Sons, Inc., New York, pp. 3-31.
Brumlop S and Finckh MR. 2010. Applications and potentials of marker-assisted selection (MAS)
in plant breeding.http://bfn.de/fileadmin/MDB/documents/service/Skript_298.pdf. de
Almeida, VC, Hoffman LV, Yokomizo GKI, daCosta JN, Giband M and Barroso PAV. 2009.
In situ genetic characterization of Gossypium barbadense populations from the states of Para
and Amapa, Brazil. Pesquisa Agropecuaria Brasileira, 44(7):719-725.
Dje, Y, Heuertz M, Lefebvre C and Vekemans X. 2000. Assessment of genetic diversity within
and among germplasm accessions in cultivated sorghum using microsatellite markers. TAG
Theoretical and Applied Genetics, 100(6):918-925.
[77]
Engels, JMM, Arora RK and Guarino L. 1995. An introduction to plant germplasm exploration and
collecting: planning, methods and procedures, follow-up. Collecting plant genetic diversity.
Technical guidelines. CAB International, Wallingford, United Kingdom. pp. 31-63.
Fryxell, PA. 1979. The natural history of the cotton tribe. Texas A & M University Press, College
Station, Texas, USA.
Guang, CHEN and Xiong-Ming DU. 2006. Genetic diversity of source germplasm of Upland
cotton in China as determined by SSR marker analysis. Acta Genetica Sinica, 33(8):733-745.
Kumar S, Stecher G and Tamura K. 2016. MEGA7: Molecular Evolutionary Genetics Analysis
version 7.0 for bigger datasets. Molecular Biology and Evolution,33:1870-1874.
Lacape, JM, Dessauw D, Rajab M, Noyer JL and Hau B. 2007. Microsatellite diversity in
tetraploid Gossypium germplasm: Assembling a highly informative genotyping set of cotton
SSRs. Molecular Breeding, 19(1):45-58.
Moiana, LD, VidigalFilho PS, Gonçalves-Vidigal MC, Lacanallo GF, Carvalho LP, Maleia MP,
Pacheco CM, Ribeiro T, Neto HZ and Coimbra GK. 2012. Genetic diversity and population
structure of cotton (Gossypium hirsutum L. race latifolium H.) using microsatellite markers.
African Journal of Biotechnology, 11:11640-11647.
Mondini, L, Noorani A and Pagnotta MA. 2009. Assessing plant genetic diversity by molecular
tools. Diversity,1(1):19-35.
Perrier, X and Jacquemoud-Collet, JP. 2006. DARwin software. http://darwin.cirad.fr/.
Rahim, MA, Mia AA, Mahmud F, Zeba N and Afrin KS. 2010. Genetic variability, character
association and genetic divergence in mung bean ('Vigna radiata'L. Wilczek). Plant
Omics,3(1):1.
Riaz, M, Naveed M, Farooq J, Farooq A, Mahmood A, Rafiq ChM, Nadeem M, Sadiq A. 2013.
AMMI analysis for stability, adaptability and GE interaction studies in cotton (Gossypium
hirsutum L.). The Journal of Animal & Plant Sciences, 23(3): 865-871.
Sapkal, DR, Sutar SR, Thakre PB, Patil BR, Paterson AH and Waghmare VN. 2011. Genetic
diversity analysis of maintainer and restorer accessions in Upland cotton (Gossypium hirsutum
L.). Journal of Plant Biochemistry and Biotechnology, 20(1):20-28.
Seelanan, T, Schnabel A and Wendel JF. 1997. Congruence and consensus in the cotton tribe.
Systemic Botany, 22: 259-290.
Ulloa, M, Stewart JM, Garcia, CEA, Goday AS, Gaytan MA and Acosta NS. 2006. Cotton genetic
resources in the western states of Mexico: in situ conservation status and germplasm collection
for ex situ preservation. Genetic Resources and Crop Evolution, 53(4):653-668
Van Esbroeck, GA, Bowman DT, Calhoun DS and May OL. 1998. Changes in the genetic
diversity of cotton in the USA from 1970 to 1995. Crop Science.38:33-37.
Van Esbroeck GA, Bowman DT, May OL and Calhoun DS. 1999. Genetic similarity indices for
ancestral cotton cultivars and their impact on genetic diversity estimates of modern cultivars.
Crop Science.39:323-328.
Wangari, NP, Gacheri KM, Theophilus MM and Lucas N. 2013. Use of SSR markers for genetic
diversity studies in mulberry accessions grown in Kenya. International Journal of
Biotechnology and Molecular Biology Research,4(3):38-44.
Wendel, JF and Albert VA. 1992. Phylogenetics of the cotton genus (Gossypium hirsutum L.):
Character-state weighted parsimony analysis of chloroplast DNA restriction site data and its
systematic and biogeographic implications. Systemic Botany, 17: 115-143.
Xin, Z, Velten JP, Oliver MJ and Burke JJ. 2003. High-throughput DNA extraction method
suitable for PCR. Biotechniques, 34(4):820-827.
Zhang, Y, Wang XF, Li ZK, Zhang GY and Ma ZY. 2011. Assessing genetic diversity of cotton
cultivars using genomic and newly developed expressed sequence tag-derived microsatellite
markers. Genetics and Molecular Research, 10(3):1462-1470.
[78]
Evaluation of Yield and Fiber Qualities
Performance of Long-staple Cotton (Gossypium
barbadense L.) Genotypes Tested on Multi-
locations of Potential Cotton- Growing Areas of
Ethiopia under Irrigation Conditions
Bedane Gudeta, Samuel Damtew, Arkebe G. Egziabher, Donis Gurmessa,
Merdasa Balcha, Alehegn workie and MekashawArega
EIAR, Werer Agricultural Research Center, P. O. Box 2003, Addis Ababa, Ethiopia
Abstract
Long-staple cotton is known by its fiber qualities rather than fiber yield. In Ethiopia,
almost all produced cotton is medium staple cotton and long-staple variety is not
released yet. Besides the stabilized yields and farm income needs, currently, there is
an increased emphasis on producing cotton with fiber quality that meets or exceeds
the minimum standard needs of the domestic industry and/or the export market. The
objective of this study was to evaluate the yield and fiber qualities performance of
long-staple cotton genotypes to recommend for producers as an option for cotton
fiber quality production. Nine cotton genotypes included three medium staple cotton
fiber varieties as standard checks were evaluated at five locations of a total of six
environments during the 2016 and 2017 cropping season by using randomized
complete block design (RCBD) with three replications. Combined analysis of variance
(ANOVA) over locations showed significant differences among cotton genotypes for
yield and fiber qualities traits. Standard checks recorded higher fiber yield than long-
staple type of cotton genotypes. But, even if standard checks were higher in yield, the
long-staple type exceeded in fiber qualities. Among tested barbadense lines (pima
cotton), genotypes HTO#052 X LS-90 24-7and HTO#052 X LS-90 24-14 were
recorded high fiber qualities with relatively better fiber yield of 10.27 and 10.04 qt ha-
1
, respectively with their respective upper half fiber mean length of 33.94 and
32.92mm and with their respective mean fiber strength of 38.69 and 38.38 g/tex and
with their premium range of micronaire values of 3.68 and 3.53 respectively. These
genotypes were much greater than the standard checks in all three main important
fiber quality parameters. These genotypes were promoted to variety verification trial
in the 2018 cropping season and submitted to variety releasing committee as
candidate varieties and they have been evaluated by the technical committee. Based
on their fiber quality performances, the national variety release standing committee
approved the release of these two varieties for commercial production purposes.
Keywords: long-staple, fiber qualities, micronaire, premium
Introduction
Cotton is the most important cash crop in Ethiopia (EIA, 2012) and plays a great
role in the agricultural and industrial development of the country’s economy.
Cotton provides basic raw materials (cotton fiber) to the textile industry for the
[79]
domestic market or export (EIA, 2014). Cotton is an important foreign currency
earner and as an important substitute crop, because it is both an important export
and import substitute commodity crop, it also serves as an important item to save
the country’s earning and thus the contribution of the cotton crop is immense to
the overall growth of the national economy.
The cotton industry evolved from home-yard cultivation into a modern state and
private farms, and from cottage weaving into modern spinning and weaving mills
in Ethiopia. To satisfy the growing demands of cotton from time to time, research
efforts made so far to develop high yielding varieties with better fiber qualities
through the utilization of hybridization methods and selection techniques from
existing stocks. The early years of cotton research focused mainly on improving
the yield of cotton as the needs of the expanding textile mills were for quantity
rather than quality.
In recent years, ginners and spinners become quality conscious they demanded
better quality. Hence, increased pressure is being placed on cotton breeders to
develop cultivars that meet the requirements of growers for high yield potential
and the demands of the textile industry for improved fiber quality. So far, while
the means to increase yield have been actively pursued by research, little efforts
have been made to evaluate fiber quality because of lack of research facilities, the
limited availability of cotton germplasm materials, and partly due to the negative
correlation between yield and quality of cotton. Because of this, Ethiopia is
importing long-staple cotton fiber.
In Ethiopia, research on G. barbadense has been neglected as compared to G.

hirsutum species. Therefore no sufficient progress has been achieved in improving
the yielding ability of G. barbadense cottons. Therefore it is necessary to give
much attention to improve the yielding ability of the high priced G. barbadense
cotton looking to its superior fiber quality. Based on changing consumption
patterns, competition from synthetic fibers, importing of extra-long and long-
staple cotton fibers, Ethiopia needs to increase cotton production in various quality
groups. So, WARC/EIAR has been making a maximum effort to develop long-
staple cotton varieties with better fiber yield from the existing stocks. The main
objective of this study was to evaluate the yield and fiber qualities performance of
long-staple cotton genotypes to recommend for producers as an option for cotton
fiber quality production.

The materials were originally developed through hybridization followed by the
pedigree selection method at Werer agricultural research center, and then the
outstanding lines have been evaluated and advanced to F6 stage at WARC by the
[80]
national cotton breeding program. The promising lines at the F6 stage were
selected and advanced to a preliminary variety trial and tested for 2 years at
Werer, then the best and superior genotypes were selected and advanced to low
land irrigated areas of cotton national variety trial at five locations and 6
environments. The parental lines were developed for improved fiber quality
purpose to full fill the demand of the domestic textile industry and international
export standards.
Table 1: Description of the test environments of the field experiment
Location Year Altitude (m.a.s.l) Lat/long.

Werer 2016, 2017 740 90 16'N 4009'E
Weyto 2016 556 50 23'N 36002'E
Gewane 2017 569 100 00'N 40031' E
Upper Awash 2017 1160 80 37' N 390 44' E
Omorate 2017 369 40 50'N 360 06'E
Experimental materials
The experiment consisted of 9 genotypes that included three standard checks
which are known by their fiber yield rather than fiber qualities and already
adapted to low land irrigated cotton production potential areas of Ethiopia.
Table 2: Lists of genotypes and their pedigrees used for the study
Entry Pedigree Selection number

1 LS-90 X HTO#052 4-8
2 LS-90 X Pima S3 5-7
3 HTO#052 X Pima S3 22-4
4 HTO#052 X LS-90 24-7
5 HTO#052 X LS-90 24-11
6 HTO#052 X LS-90 24-14
7 Deltapine-90(standard check)
8 Stam 59A(standard check)
9 Ionia(standard check)
Design of the experiment

The experiment was conducted in a randomized complete block design (RCBD)
with three replications. Each entry was planted in a plot having 5m length and 5
rows. The spacing between rows and plants were 0.90m and 0.20m, respectively.
The total plot and net plot area (harvestable plot area) were 5 rows x 5m x 0.9m
=22.5m2 and 3 rows x 5m x 0.9m =13.5m2, respectively. Analysis of variance
(ANOVA) was carried out by using SAS software to evaluate the performances of
the cotton genotypes. LSD was used for mean separation. The locations were
considered as random and genotypes as fixed effects, and a mixed effect model
ANOVA was used for statistical analysis. In the ANOVA for each location,
sources of variation are blocks, treatments and error terms.
[81]
Analysis of variances indicated that there were significant differences (p < 0.0001)
among the mean performance of the tested genotypes at all environments and
overall combined for average boll weight in gram (Table 3). Standard checks
recorded higher mean boll weight than all tested long-staple cotton genotypes.
Because standard checks are Gossypium hirsutum type which is a high yield
cotton species that exhibits only medium performance in fiber qualities while G.
barbadense, an important cultivated extra-long or long-staple cotton species. For
boll weight (g), ANOVA across environments indicated significant variation
among genotypes and environments but, non-significance differences for the G×E
interaction. This was also because of all genotypes similarly affected by the
environment. Among all barbadense lines, HTO # 052 X LS-90 24-14 exhibited
heaviest bolls (3.29 g) followed by the barbadense lines HTO # 052 X LS-90 24-7
and LS-90 X Pima S3 5-7 recorded (3.22 g) and (3.19 g) respectively, while the
line HTO # 052 X Pima S3 22-4 had minimum mean boll weight 2.94g followed
by line LS-90 X HTO#052 4-8 (3.09 g). In agreement with this study Yanal
Alkuddsi et al., 2013 in their studies, reported that among 52 evaluated
barbadense lines, line DB 533×DB 534 F5 IPS 30 exhibited heaviest bolls (3.23g)
followed by the barbadense lines DB 533×DB 534 F5 IPS 132 (3.17g), DB 533 ×
DB 534 F5 IPS 37 (2.93g) and DB 533 × DB 534 F5 IPS 18 (2.93g), while the
line DB 533 × DB 534 F5 IPS 33 had minimum mean boll weight (1.67 g)
followed by the line DB 533 × DB 534 F5 IPS 38 (1.73 g).
Table 3.Mean performance of boll weight (g) of tested genotypes
Gewane Weyto Omorate Upper wash Werer Overall

Treatment
2017 2016 2017 2017 2016 2017
LS-90 X HTO#052 4-8 2.41 3.09 2.26 3.74 3.79 3.24 3.09
LS-90 X Pima S3 5-7 3.29 2.96 2.22 3.87 3.32 3.46 3.19
HTO#052 X Pima S3 22-4 2.67 2.80 2.41 3.42 3.46 2.90 2.94
HTO#052 X LS-90 24-7 3.02 3.81 2.10 3.86 3.33 3.20 3.22
HTO#052 X LS-90 24-11 3.04 3.29 2.06 3.60 3.40 3.40 3.13
HTO#052 X LS-90 24-14 2.88 3.05 2.41 4.29 3.91 3.21 3.29
Deltapine-90(standard check) 4.41 4.58 3.54 4.56 5.28 4.50 4.48
Stam 59A(standard check) 4.54 5.62 3.46 4.78 5.34 4.72 4.74
Ionia(standard check) 4.64 4.88 3.79 5.41 5.27 5.00 4.83
Mean 3.43 3.79 2.69 4.17 4.12 3.74 3.66
Cv 16.64 16.47 14.05 4.16 11.23 7.21 12.10
Significance level
GENO *** *** *** *** *** *** ***
ENV ***
ENV * GENO NS
LSD 0.99 1.08 0.66 0.30 0.80 0.47 0.33
Analysis of variances indicated that there were significant differences (p < 0.05)
among the mean performance of the tested genotypes at all locations except at
[82]
Werer 2017 cropping season for seed cotton yield (Table 4). There were non-
significance differences among tested genotypes at Werer 2017 cropping season for
seed cotton yield. That means barbadense lines have competed with hirsutum type
varieties. So, the season was favorable for the tested barbadense lines. Standard
checks recorded higher mean seed cotton yield than all tested long-staple cotton
genotypes in other environments. Because standard checks are Gossypium hirsutum
types which are a high yield cotton species that exhibits only medium performance
in fiber qualities while G. barbadense, an important cultivated extra-long or long-
staple cotton species. Among the six G. barbadense genotypes, HTO # 052 X LS-
90 24-7 (27.19 qt/ha) recorded higher mean seed cotton yield of across
environments followed by HTO # 052 X LS-90 24-14 (26.35 qt/ha) (Table 4).
Similar findings were reported by Yanal Alkuddsi et al., 2013 in G. barbadense
genotypes, they recorded the highest seed cotton yield in lines DB 533 x DB 534
F5 IPS 132 (21.76 qt/ha) and DB 533×DB 534 F5 IPS 18 (21.51 qt/ha), whereas
Dhamayanthi and Subashree (2016) recorded higher seed cotton yield in lines
CCB-29 (14.89 qt/ha) and CCB-36 (12,31qt/ha) in G. barbadense genotypes which
were much less than our tested barbadense lines potential.
Table 4.Mean performance of Seed Cotton Yield (qt/ha) of tested genotypes
Upper Werer
Gewane Weyto Omorate
Treatment Awash Overall
2017 2016 2017 2017 2016 2017
LS-90 X HTO#052 4-8 24.54 18.22 12.85 24.46 33.48 31.02 24.09
LS-90 X Pima S3 5-7 20.55 14.71 13.09 29.17 31.89 32.94 23.73
HTO#052 X Pima S3 22-4 19.33 29.41 17.57 16.56 31.89 25.55 23.39
HTO#052 X LS-90 24-7 32.09 19.51 12.07 29.80 35.52 34.12 27.19
HTO#052 X LS-90 24-11 26.59 10.07 11.08 26.37 23.86 30.00 21.33
HTO#052 X LS-90 24-14 21.87 25.07 12.88 28.05 39.14 31.10 26.35
Mean 29.41 31.41 17.03 27.25 36.61 31.46 28.86
Cv 21.43 21.61 18.61 17.74 15.10 14.94 18.54
Significance level
GENO *** *** *** * *** Ns ***
ENV ***
ENV * GENO ***
LSD 10.91 11.75 5.49 8.37 9.57 9.50
There were also significant differences (p < 0.05) among the mean performances
of the tested genotypes at all locations for ginning outturn (%) (Table 5). Overall
ANOVA also showed that there were significant differences among genotypes,
environments and for the G×E. Variety Stam 59A which was standard check
recorded high ginning outturn (39.97%). Among barbadense lines, lines
HTO#052 X LS-90 24-11 (38.50%), LS-90 X HTO#052 4-8 (38.43%), HTO#052
X LS-90 24-14 (38.08%) and HTO#052 X LS-90 24-7 (37.57%) possessed higher
ginning out turn than standard checks Deltapine-90 (36.74%) and Ionia (37.24%).
[83]
Table 5.Mean performance of ginning out turn (GOT) (%) of tested genotypes
Upper Werer
2017 2016 2017 2017 2016 2017
LS-90 X HTO#052 4-8 40.17 36.05 37.99 40.05 38.36 37.96 38.43
LS-90 X Pima S3 5-7 37.54 35.93 36.07 38.53 36.72 35.58 36.73
HTO#052 X Pima S3 22-4 35.28 33.08 35.88 35.62 33.91 33.49 34.54
HTO#052 X LS-90 24-7 39.01 35.29 37.29 38.21 37.72 37.90 37.57
HTO#052 X LS-90 24-11 39.10 37.24 39.35 38.69 38.20 38.39 38.50
HTO#052 X LS-90 24-14 39.15 36.89 38.52 38.44 38.40 37.08 38.08
Mean 37.88 36.19 37.52 39.04 37.49 37.09 37.53
Cv 4.51 2.49 3.76 1.67 2.39 2.46 3.04
Significance level
GENO * *** ** *** *** *** ***
ENV ***
ENV * GENO **
LSD 2.96 1.56 2.44 1.13 1.55 1.58 1.08
Analysis of variances for lint yield indicated that there were significant differences
(p < 0.05) among the mean performance of the tested genotypes at all locations
except at Werer 2017 cropping season (Table 6). There were non-significance
differences among tested genotypes at Werer 2017 cropping season for lint yield.
That means barbadense lines were comparable with standard checks (G. hirsutum)
varieties. So, the season was favorable for the tested barbadense lines. Overall
ANOVA also showed that there were significant differences among genotypes,
environments and for the G×E interaction. The standard checks recorded higher
mean lint yield than all tested long-staple cotton genotypes in other environments.
Among barbadense lines, lines HTO#052 X LS-90 24-7 (10.27qt/ha) and
HTO#052 X LS-90 24-14 (10.04qt/ha) have recorded high lint yield.
[84]
Table 6. Mean performance of Lint yield (qt/ha) of tested genotypes
Upper Werer
2017 2016 2017 2017 2016 2017
LS-90 X HTO#052 4-8 9.88 6.57 4.88 9.82 12.89 11.78 9.30
LS-90 X Pima S3 5-7 7.78 5.29 4.72 11.25 11.72 11.72 8.75
HTO#052 X Pima S3 22-4 6.93 9.75 6.30 5.91 10.80 8.58 8.04
HTO#052 X LS-90 24-7 12.49 6.89 4.51 11.39 13.41 12.91 10.27
HTO#052 X LS-90 24-11 10.39 3.77 4.35 10.19 9.11 11.53 8.22
HTO#052 X LS-90 24-14 8.60 9.28 5.02 10.78 15.03 11.54 10.04
Mean 11.15 11.43 6.38 10.70 13.74 11.67 10.84
Cv 22.81 22.61 18.89 18.36 16.60 13.94 19.31
Significance level
GENO ** *** *** * *** Ns ***
ENV ***
ENV * GENO ***
LSD 4.40 4.47 2.09 3.40 3.95 3.57
There was also significant variability (p < 0.05) in micronaire among tested
genotypes at all locations except locations Weyto and Werer 2016 cropping
season at which were non- significant differences among genotypes (Table7).
Overall ANOVA also showed that there were significant differences among
genotypes, environments and for the G×E. The standard checks recorded higher
mean micronaire values than all tested long-staple cotton genotypes. But
micronaire values have an international standard. Micronaire reading from 3.7 –
4.2 is considered as premium; from 4.3- 4.9 as base range and greater than 5 is
discount range. Cotton lint with micronaire below 3.5 is usually considered
immature and weak (Chaudhry and Cuitchounts, 2003) whereas micronaire values
higher than 4.9 are less desirable as the fiber becomes too coarse for spinning.
The study revealed all mean micronaire values of barbadense lines were in the
premium range across environments whereas standard checks were found in the
base range across environments. Combined analysis of variance showed that
genotypes, environment and the GxE interaction had a very highly significant
effect (P<0.0001) on micronaire (Table 7). These indicated that micronaire is
affected by environments. This result is in agreement with the result of Campbell
& Jones (2005) and Pretorius et al. (2015) where they have got ANOVA across
environments indicating significant variation among genotypes, environments and
for the G×E for micronaire values. Micronaire was influenced more by
environments (Pretorius et al., 2015). But in our results, all micronaire values of
genotypes in this study were in the acceptable range according to Chaudhry and
Cuitchounts (2003).
[85]
Table 7. Mean performance of Micronaireof tested genotypes
Upper Werer
2017 2016 2017 2017 2016 2017
LS-90 X HTO#052 4-8 3.68 3.78 2.90 3.55 3.94 3.46 3.55
LS-90 X Pima S3 5-7 3.71 3.63 2.85 3.78 3.78 3.60 3.56
HTO#052 X Pima S3 22-4 3.75 3.63 2.94 3.46 3.11 3.84 3.45
HTO#052 X LS-90 24-7 3.99 3.80 3.02 4.03 3.30 3.93 3.68
HTO#052 X LS-90 24-11 4.18 3.65 2.86 3.81 3.49 3.76 3.63
HTO#052 X LS-90 24-14 3.58 3.78 2.84 3.75 3.72 3.52 3.53
Mean 4.15 3.74 3.22 3.79 3.79 4.02 3.78
Cv 7.08 13.54 5.54 4.76 17.22 4.61 10.05
Significance level
GENO *** Ns *** * Ns *** ***
ENV ***
ENV * GENO *
LSD 0.51 0.31 0.31 0.32 0.32
In terms of cotton fiber length, there were significant differences (p < 0.05) among
genotypes at all environments except at environments Weyto and Werer 2016
which were analysis of variances indicated non-significant differences among
genotypes (Table 8). But even if statistically indicated non-significant differences
among genotypes at environments Weyto 2016 and Werer 2016 the values of fiber
length (mm) of tested genotypes found in different ranges according to the
international standard of fiber length; fiber length less than 27mm is considered as
short, from 27.9 to 31.5mm considered as medium length, from 31.8 to 35.3mm is
long and above 35.3mm is extra-long. All barbadense lines recorded long staple
length whereas standard checks recorded medium length. Combined analysis of
variance showed that genotypes, environment and the genotypes x environment
interaction had a very highly significant effect (P<0.0001) on length (Table 8).
These indicated that fiber length is affected by environments. In agreement with
Fibre length and micronaire are significantly affected by agronomic and climate
effects mentioned in the Australian Cotton Production Manual 2019
(https://cottoninfo.com.au/publications/australian-cotton-production-manual). It
also affected by extreme temperatures, water stress, or nutrient deficiencies
(www.cottoninc.com/cotton-production/quality/classification-of-
cotton/classification-of-upland-cotton/).
[86]
Table 8.Mean performance of upper half mean length (mm) of tested genotypes
Upper Werer
Treatment Awash
Overall
2017 2016 2017 2017 2016 2017
LS-90 X HTO#052 4-8 35.67 32.45 31.26 32.79 31.94 35.03 33.19
LS-90 X Pima S3 5-7 34.54 30.66 31.02 32.96 31.78 36.04 32.83
HTO#052 X Pima S3 22-4 33.92 33.11 31.85 34.20 35.30 36.25 34.11
HTO#052 X LS-90 24-7 35.59 31.26 31.15 34.04 34.91 36.03 33.83
HTO#052 X LS-90 24-11 31.54 31.57 29.91 33.31 32.97 34.49 32.30
HTO#052 X LS-90 24-14 33.45 31.33 31.41 33.22 32.30 35.38 32.85
Mean 32.38 31.33 30.13 32.20 32.35 33.67 32.01
Cv 5.28 5.37 2.29 2.94 8.11 1.94 4.84
Significance level
GENO *** Ns *** *** Ns *** ***
ENV ***
ENV * GENO *
LSD 2.96 1.19 1.64 1.13 1.40
In terms of cotton fiber strength, there were significant differences (p < 0.05)
among genotypes at all environments except at environments Weyto 2016 and
Werer 2016 which were analysis of variances indicated non-significant differences
among genotypes (Table 9). But even if statistically indicated non-significant
differences among genotypes at environments Weyto and Werer 2016 the values
of fiber strength (g/tex) of tested genotypes found in different ranges according to
the international standard of fiber; fiber strength values of less than 23 is weak,
from 24 to 25 is intermediate, 26 to 28 is average, 29 to 30 is strong and greater
than 31 is very strong. The overall mean values of fiber strength of tested
genotypes were found in the very strong range except standard check Deltapine-90
variety. But, the mean values of the fiber strength of tested barbadense lines were
significantly very far from standard checks. ANOVA across environments
indicated significant variation among genotypes and environments but, non-
significance differences for the G×E interaction. This was also because of all
genotypes similarly affected by the environment. This result showed that there
were variations among genotypes and environments and all genotypes were
equally affected by environments since there were no interactions. This also
agrees with the finding of (Brooks, 2005) he found that a very highly significant
effect (<0.0001) for both among genotypes and environments and non-significant
effect (0.0907) for GxE interaction. It also mentioned in the Australian Cotton
Production Manual 2019 (https://cottoninfo.com.au/publications/australian-cotton-
production-manual) as fiber strength is more influenced by variety choice rather
than environments.
[87]
Table 9.Mean performance of Fiber strength (g/tex) of tested genotypes
Upper Werer
2017 2016 2017 2017 2016 2017
LS-90 X HTO#052 4-8 41.23 34.00 40.27 38.37 32.17 40.60 37.77
LS-90 X Pima S3 5-7 40.17 34.47 38.57 37.07 34.57 42.07 37.82
HTO#052 X Pima S3 22-4 39.30 36.50 39.47 38.03 39.63 43.00 39.32
HTO#052 X LS-90 24-7 39.33 34.93 39.27 40.07 37.90 43.73 39.21
HTO#052 X LS-90 24-11 37.57 32.03 39.97 39.50 35.73 42.60 37.90
HTO#052 X LS-90 24-14 40.57 35.03 41.75 40.43 34.20 44.43 39.40
Mean 36.97 33.57 37.49 35.79 33.74 39.33 36.15
Cv 8.59 12.53 4.12 4.90 14.71 4.55 8.83
Significance level
GENO ** Ns *** *** NS *** ***
ENV ***
ENV * GENO Ns
LSD 5.49 2.67 3.03 3.10 2.36
Conclusion and Recommendation
The experiment consisted of 6 barbadense lines and three standard checks which
are known by their fiber yield rather than fiber qualities and already adapted to
low land irrigated cotton production potential areas of Ethiopia. A total of 9
genotypes were evaluated five locations using randomized complete block design
(RCBD) with three replications during the 2016 and 2017 consecutive cropping
season. The experiments were carried out to evaluate the yield and fiber qualities
performance of long-staple cotton genotypes to recommend for producers as an
option for cotton fiber quality production. The result of the study showed that
there was a significant difference among genotypes, locations and GxE
interactions were also significant on yields. Combined analysis of variance over
locations indicated that genotypes HTO#052 X LS-90 24-7 and HTO#052 X LS-
90 24-14were the top two high yielding cotton genotypes among barbadense lines
with their respective mean lint yield of 10.27 and 10.04 qt ha-1, respectively. For
cotton fiber upper half mean length, genotypes HTO#052 X LS-90 24-7 and
HTO#052 X LS-90 24-14 were recorded 33.83 and 32.85mm respectively, which
were greater than standard checks Deltapine-90, Stam-59A and Ionia with their
respective mean fiber upper half mean length of 28.83, 30.55 and 29.61mm
respectively. In terms of cotton fiber strength, genotypes HTO#052 X LS-90 24-7
and HTO#052 X LS-90 24-14 were recorded 39.21and 39.40g/tex respectively,
which were much greater than standard checks Deltapine-90, Stam-59A and Ionia
with their respective mean fiber strength of 29.71, 32.51and 31.71 g/tex
respectively.
[88]
These two barbadense genotypes (pima cotton) were recorded high fiber qualities
with relatively better fiber yield. These genotypes were developed by crossing and
through selection at Werer Agricultural Research Center and the objectives, these
crosses were attained. So, these valuable genotypes were promoted to a variety
verification trial in the 2018 cropping season. These outstanding cotton genotypes
were submitted as candidate genotypes for variety verification trail and evaluated
by the national variety release technical committee. Based on their fiber quality
performances, the national variety release standing committee released these two
potential varieties for commercial production purposes in the 2019 crops season.
These genotypes can also be used as a source of valuable genetic materials that
could be used for future cotton breeding program to improve cotton fiber qualities.
References
Brooks SB. 2005. New methods to assess cotton varietal stability and identify discriminating
environments. Ph.D Thesis.Louisiana State University and Agricultural and Mechanical
College.16 p.
Campbell, BT and Jones MA. 2005. Assessment of Genotype × Environment Interactions for Yield
and Fiber Quality in Cotton Performance Trials. Euphytica 144: 69–78.
Chaudhry, MR and Guitchounts A. 2003. Cotton Facts. Common Fund for Commodities,
International Cotton Advisory Committee, Washington.
EIA (Ethiopian Investment Agency). 2012. Cotton Production and Ginning in Ethiopia.
EIA (Ethiopian investment agency). 2014. Investment Opportunity Profile for Cotton Spinning
Mill in Ethiopia.
https://cottoninfo.com.au/publications/australian-cotton-production-manual
Dhamayanthi, KPM and Subashree K. 2016. Assessment of yield and yield-related traits to
determine earliness in Egyptian cotton (Gossypiumbarbadense, L). Electronic Journal of
Plant Breeding, 7(3): 771-777 DOI: 10.5958/0975-928X.2016.00102.2
Pretorius, MMJ, Allemann and MF Smith. 2015. Use of the AMMI model to analyze cultivar-
environment interaction in cotton under irrigation in South Africa. 2(2): 76–80.
www.cottoninc.com/cotton-production/quality/classification-of-cotton/classification-of-
upland-cotton/
YanalAlkuddsi, Patil SS, Manjula SM, Patil BC, Nadaf HL and Nandihali BS. 2013. Identifying
of Extra Long Staple Suitable Lines (GossypiumbarbadenseL) with Improved Fiber Qualities
to Release New Lines as an Alternative for Suvin Variety of Barbadense, Cotton Genomics
and Genetics, Vol.4, No.1 1-12 (doi: 10.5376/cgg.2013.04.0001)
[89]
[90]
Yield Performance of Faba Bean (Vicia faba L)
Genotypes for High Potential Production Area
Musa J1., Asnakech T1., Gemechu K1., Sisay A1, Mesfin T1, Nigat T1.,
Wondafrash M.1, Deressa T. 2,, Kedir Y.2, Gezachew Y2., Behaylu A.3, Jemal E.4,
Danele A.5, Tadele T.6, and Nano A.7
1
Ethiopian Institute of Agricultural Research, Holetta Research Centre, P.O.Box 36, Ethiopia; 2Ethiopian
Institute of Agricultural Research, KulumsaResearch Centre, Ethiopia; 3Ethiopian Institute of Agricultural
Research, Jimma Research Centre, Ethiopia; 4Amhara Regional Agricultural Research Institute, Adet
Research Centre; 5Amhara Regional Agricultural Research Institute, Gonder Research Centre 6Oromiya
Regional Agricultural Research Institute, Sinana Research Centre, Ethiopia; 7Haramaya University, Ethiopia.
Abstract
Faba bean (Viciafaba L.) has high utility as a food and soil fertility improving crop.
Despite its multiples of importance, faba bean production and productivity have
been low due to several constraints including inherent low-yielding potential of the
local cultivars, biotic and abiotic stresses that require research priority. Although
significant research efforts were made to improve faba bean productivity, more
research, is still required to increase the productivity of the crop and to meet the
national demand. Therefore, the present study was initiated with the main objectives
to develop stable, high yielding and stress-tolerant faba bean, variety for different
agro-ecology in Ethiopia. A total of thirteen faba bean selected materials from
previous preliminary variety trials were evaluated using RCBD with three
replications as a national variety trial over nine locations for two years in 2016 and
2017 main cropping seasons. Significant differences (p<0.001) were observed
among the genotypes for most of the yield and yield component traits. Two genotypes
EH 09031-4 and EH08035-1 showed comparable yield and thousand seed weight as
the two standard checks Tumsa and Gora. Considering their high yield and large
seed size these two genotypes were recommended to be used as a parent material in
the faba bean improvement program.
Keywords: Evaluation; faba bean; yield, thousand seed weight,
Introduction
Ethiopia is a secondary origin and one of the secondary centers of Vavilovian
diversity for faba bean (Vicia faba L.). Faba is the major pulse crops grown in the
highlands and -mid-altitude areas (1800 - 3000 masl.) of Ethiopia. The crop is
grown for different economic and ecological purposes as sources of food, feed,
income, and foreign currency earnings and soil fertility restoration. Faba bean is a
crop of manifold merits in the economic lives of the farming communities in the
highlands of Ethiopia. It serves as sources of food with a valuable and cheap
protein compared to cereals. Faba bean plays significant roles in soil fertility
restoration as suitable rotation crops that fix atmospheric nitrogen, thereby
resulting in savings for smallholder farmers from less fertilizer use. Nitrogen
[91]
fixation by faba bean has a significant spillover effect to subsequently grown
wheat in Ethiopia. In order to keep the farming system healthy and sustainable,
legumes in general and faba bean, in particular, plays a significant role as
components of multiple cropping and in crop rotation with cereals. In addition, it
improves soil fertility, and also serves as a break crop to minimize disease-causing
pathogens of cereals.
Despite its multiples of importance, faba bean production and productivity have
been low due to several constraints including biotic and abiotic stresses, socio-
economic reasons and competition with other crops. Productions have been
constrained by several yield-limiting and reducing factors. The inherent low-
yielding potential of the indigenous cultivars is among the most important
production constraints. Moreover, foliar and root diseases, weed, and abiotic
stresses like drought, soil acidity, and waterlogging are among important
production constraints that deserved priority as research objectives. Thus, the
productivity of the crop is 2 t/ha far below the potential. Although efforts made for
the improvement of productivity of faba bean greater level of research is required
to meet the national demand to increase production and productivity of the crop
and to meet the national developmental desires, to supply sustainable raw material
for agro-industry and production of the crop in high standard quality and quantity
for export. Host plant resistance/tolerance development is most appropriate safe
and cost-effective which can be implemented as the core of an integrated strategy
for managing different biotic and abiotic production constraints of faba bean
production (Bouhassan et al., 2004). Consequently, improvement activities of faba
bean using a different breeding strategy is proceeding in faba bean breeding
program nationally and these must continue to sustain the release of high yielding
and stress-resistant / tolerant faba bean varieties. Therefore, the main objectives of
this trial were to develop stable, high yielding and disease resistant faba bean,
variety for faba bean growers in different agro-ecology of Ethiopia.

The study was conducted in nine locations, namely Gonder/Debark, Adet,
Holetta, Kulumsa, Bekoji, Kofele, Haramaya, Jimma, and Sinana with an altitude
range of 2000– 2800. All the locations are among the principal faba bean testing
sites in Ethiopia representing the faba bean belt of the country. A total of 13 faba
bean selected materials from previous preliminary variety trials were evaluated as
national variety trial for two years from 2016 to 2017 main cropping season. The
design was RCBD with three replications. Each germplasm planted in 4 rows with
0.4 m between rows, 0.1 m between plants in 4 m lengths of 0.6 m between
adjacent plots. DAP at the rate of 100 kg h-1 was applied at planting. Weeding and
other management practices were done as per recommendation. Data for the
following agronomic traits were recorded on a plot basis at all locations; days to
[92]
50% flowering (DTF) and days to physiological maturity (DTM) for the entire
plot, number of pods per plant (PPP), number of seeds per pod (SPP), and plant
height (PH). Disease scoring made from the entire plot using 1 to 9 scoring
method. Data thousand seed weight (TSW) from randomly selected 1000 seeds
from each plot and grain yield (GY) was taken from the entire plot. Grain yield
adjustments were made by weighing the oven drying and adjusting to a constant
moisture level of 10%. The grain yield recorded on a plot basis was converted to
kg ha-1 for statistical analysis. The analysis of variance (ANOVA) was performed
using the PROC GLM in SAS (SAS Institute, 2012).

The combined analysis of variance (ANOVA) revealed that significant (p < 0.001)
variation among the thirteen genotypes for DTF, DTM, PH, SPP, PPP, TSW and
grain yield (GY) across locations and year (Table 1 and 2). Out of the 13 faba
bean materials, the genotype EH 09031-4 (Moti XILB4726) showed 9.5 to 19%
thousand seed weight advantage over the two standard checks Gora and Tumsa
with comparable yield and genotype EH08035-1 (EH99019-5 X ILB1563)
showed 5% TSW advantage over the checks with comparable yield. These
genotypes were also performed better in other agronomic traits across locations
and over two years (Table 2). In addition, these genotypes had the merit of
chocolate spot tolerance as they are developed from the parent material of
chocolate spot resistant materials ILB4726 and ILB1563 from ICARDA. Thus,
the two genotypes will be used as a parent material for seed size in faba bean
breeding program.
[93]
Table 1. Mean grain yield (Kg/ha) performance of faba bean varieties tested in NVT at nine environments (year by location) from 2016 to 2017.
Geno Adt 16 Adt 17 Bekj 16 Bekj 17 DebT17 Hol 16 Hol 17 HU 16 Jim 16 Jim 17 Kof 16 Kof 17 Kul 16 Kul 17 Sin 16 Mean
Gora 2242 1887 4328 3825 3197 3667 3026 3706 6660 6261 4036 4473 4348 3417 2891 3869
EH 010010-1 2738 1407 3821 3583 3009 3426 4100 3427 4719 6360 4377 5003 3370 2814 1873 3598
EK05037-4 1683 1434 4969 3313 2632 3200 3369 4101 6061 6042 4010 4133 3665 2597 2048 3558
EH07023-3 1583 1849 4268 3273 3098 3083 3368 3181 4569 6165 3364 4235 4341 2732 2246 3421
EH09002-1 1801 1712 4923 3339 2276 3402 2756 3364 5516 6047 3637 4705 4710 2689 2376 3558
EH09004-2 1902 1258 4296 3003 2714 3480 4463 4036 5579 5730 3324 4365 4289 3352 1676 3562
EH09007-4 1777 1426 4772 3685 2497 3489 4006 3589 6472 6422 4040 4852 4496 3057 2392 3796
EH06007-4 2236 2263 5315 3444 3337 4041 4318 3623 6108 5493 3093 4458 4552 3130 2059 3828
EH09031-4 1494 1805 4687 2117 2769 4258 4730 3804 6258 6465 4578 5178 4516 2733 1577 3802
EH08031-2 1460 1379 4687 3289 2935 3085 4164 4039 5948 6165 3858 5475 4208 2999 1996 3720
EH08035-1 1952 2388 4804 3618 3395 3697 3923 4564 5960 5642 3714 5513 4769 3378 1696 3934
EH08035-3 2049 2344 4321 3272 2796 3999 4395 4577 6224 6146 3528 5404 4551 2864 2087 3906
Tumsa 1707 2199 4748 3019 3083 4178 3885 4081 6472 5973 4655 4787 4731 3629 2227 3958
Mean 1894 1760 4607 3284 2919 3606 3885 3846 5893 6056 3858 4834 4341 3033 2076 3735
CV (%) 14.4 12.7 13.6 7.8 8.4 7.8 6.5 6.4 7.6 4.2 8.3 4.9 8.4 8.3 11.5 8.9
LSD (5%) 392.4 322.7 NS 370.9 351 404.6 363.9 354.5 640.2 365.6 459.5 343.7 526.9 636 342.3 462.9
Where: Adt: Adet Agricultural Research Center, Bekj: Bekoji research station; BebT: Debretabor research station; Hol: Holetta Agricultural Research Center; HU: Haramaya University;
Jim: Jimma Agricultural Research Center; Kof: Kofele research station; Kul: Kulumsa Agricultural Research Center; Sin: Sinana Agricultural Research Center. (16=2016; 17=2017)
[94]
Table 2. Mean agronomic performance of faba bean varieties tested in NVT at nine environments (year by location) from 2016 to 2017
No Geno DTF DTM DTGF FPH PHT PPP SPP CS Rust TSW Gykgha
1 Gora 53 146 93 53 139 15 3 4 3 705 3869
2 EH 010010-1 55 148 93 51 132 15 3 2 3 650 3598
3 EK05037-4 53 146 93 47 148 16 3 4 3 623 3558
4 EH07023-3 52 168 117 47 138 15 3 4 3 717 3421
5 EH09002-1 54 147 92 44 130 13 3 3 3 700 3558
6 EH09004-2 51 143 92 40 130 15 3 4 4 634 3562
7 EH09007-4 53 146 92 52 138 14 3 3 3 717 3796
8 EH06007-4 56 146 90 50 134 15 3 3 3 719 3828
9 EH09031-4 54 147 93 51 135 14 3 3 3 772 3802
10 EH08031-2 52 145 94 48 135 14 3 4 4 689 3720
11 EH08035-1 51 144 93 44 150 15 3 4 4 676 3934
12 EH08035-3 53 144 91 49 136 16 3 4 3 660 3906
13 Tumsa 54 147 93 48 141 16 3 3 3 646 3958
Mean 53 146 92.7 48 136 15 2.7 3.4 3.2 685 3735
CV (%) 5.4 1.8 4 18 10 28 13 24 22 11.9 8.9
LSD (5%) 1 0.9 1.3 4.8 4.9 1.5 0.1 0.3 0.3 30.3 462.9
Where: DTF: Days to flowering; DTM: Days to maturity; DTFG: Days to grain filling; FPH: First pod hight from the ground; PHT: plant height; PPP:
Pod per plant; SPP: Seed per pod; CS: chocolate spot disease score; TSW: Thousand seed weight; Gykgha: Grain yield kilogram per hectare.
[95]
The evaluation made from the breeding material is either directly used for the
development of new varieties or utilized for incorporating their variable genetic
attributes in the locally adapted faba bean varieties through cross-breeding
program. From this trial the two genotypes EH09031-4 and EH08035-1 showed
comparable yield and better thousand seed weight compared to the two standard
checks. Therefore, these materials can be recommended to be used as a parent
material for seed size in faba bean breeding program.
Acknowledgments
This work was part of the national highland pulse breeding program and was
supported by the Ethiopian Institute of Agricultural Research. The authors
acknowledge staff members of the breeding and Genetics Research department of
Holetta Agricultural Research Center, Kulumsa ARC, Gonder ARC, Adet ARC,
Haramaya University, Jimma ARC and Sinana Agricultural Research Centers for
their collaboration and efforts in trial management and data collection.
Reference
Asnakech, TB John D and Julia S. 2015. Gene action determining grain yield and chocolate spot
(Botrytis fabae) resistance in faba bean. Euphytica 207:2
Asnakech TB, John D and Julia S. 2018. Genetic variability of faba bean genotypes for chocolate
spot (Botrytis fabae) resistance and yield. Euphytica 214:8.
Bouhassan, A, Sadiki M, and Tivoli B. 2004. Evaluation of a collection of faba bean (Vicia faba
L.) genotypes originating from the Maghreb for resistance to chocolate spot (Botrytis fabae)
by assessment in the field and laboratory. Euphytica 135:55–62.
CSA. 2018. Report on area and production of crops. Central Statistics Agency agricultural sample
survey for 2017 / 2018 Addis Ababa, Ethiopia.
SAS Institute. 2012. SAS proprietary software. Release 9.3 SAS Inst., Cary, NC, USA.
[96]
Yield Performance of Large-Seeded Faba Bean
Genotypes for Potential Production
Areas in Ethiopia
Asnakech T1., Musa J1., Gemechu K1., Sisay A1, Mesfin T1, Nigat T1., Aliy R. 2,
Kedir Y.2, Mesfen H.3, Abebe H.4, and Fantanesh S4.
1
EIAR, Holetta Research Centre, P.O.Box36, Ethiopia2 ; Ethiopian Institute of Agricultural Research,
KulumsaResearch Centre, Ethiopia; ; 3Ethiopian Institute of Agricultural Research, JimmaResearch Centre,
Ethiopia;4Amhara Regional Agricultural Research Institute, Gonder Research Centre
Abstract
Faba bean (Viciafaba L.) has high utility as a food and soil fertility improving crop.
Despite its multiples of importance, faba bean production and productivity has been
low due to several constraints including inherent low-yielding potential of the local
cultivars, biotic and abiotic stresses that require research priority. Although
research efforts made for the improvement of productivity of faba bean more is
required to increase the productivity of the crop and to meet the national demand.
Therefore, the present study was initiated with the main objectives of to develop
stable, high yielding and stress-tolerant faba bean, variety for different agro-ecology
in Ethiopia. A total of thirteen faba bean selected materials from previous
preliminary variety trials were evaluated using RCBD with three replications as a
national variety trial over nine locations for two years in 2017 to 2018 main
cropping seasons. Significant differences (p < 0.001) were observed among the
genotypes for most of the yield and yield component traits. Two genotypes EH
010058-1 and EH 010058-2 resulted in comparable yield and better in thousand
seed weight of the two standard checks Tumsa and Gora. Considering their high
yield and large seed size these two genotypes were recommended to be used as a
parent material in the faba bean improvement program.
Keywords: Evaluation; faba bean; yield, thousand seed weight,
Introduction
Faba bean (Viciafaba L.), production ranks fourth in the world as an important
food legume (Torres et al., 2006) is the major pulse crops grown in the highlands
and mid-altitude areas of Ethiopia. The crop is grown for different economic and
ecological purposes as sources of food, feed, income, and foreign currency
earnings and soil fertility restoration. Faba bean is a crop of manifold merits in the
economic lives of the farming communities in the highlands of Ethiopia. It serve
as sources of food with a valuable and cheap protein compared to cereals. Apart
from being a potential food and cash crop for resource-poor farmers, faba bean has
been used as a major break for cereal mono-cropping system which can increase
soil fertility by atmospheric nitrogen fixation through symbiosis (López-Bellido et
al., 2006). Its role as soil fertility restorers will remain an integral part of future
[97]
farming systems because of the rising price of nitrogen fertilizers, increasing
population pressure, and the negative side effects of commercial fertilizers. Like
any other legumes, faba bean enhances nutrient availability for associated or
subsequently grown cereals by producing exudates like cerate and malate that
solubilize soil minerals including a fixed form of phosphorus (Lindemann and
Glover 2003).
Although faba bean (Vicia faba L.) is the most important food legume cultivated
in Ethiopia, production and productivity has been low (2.1 t ha-1) compared to the
yield potential of 8 t ha-1(CSA, 2018) constrained by several yield-limiting and
reducing factors including biotic and abiotic stresses. Of which biotic stresses like
foliar and root diseases, weed, and abiotic stresses like drought, soil acidity, and
waterlogging are among important production constraints that merited importance
as research objectives. Efforts made for the improvement of productivity of faba
bean to meet the national demand and so far some improved varieties were
developed as Host plant resistance/tolerance development is most appropriate safe
and cost-effective for managing different biotic and abiotic production constraints
of faba bean production (Bouhassan et al., 2004). Therefore, the main objectives
of this trial were to develop stable, high yielding and disease resistant faba bean,
variety for faba bean growers in different agro-ecology of Ethiopia.

A total of 12 faba bean selected materials from previous preliminary variety trials
were evaluated as a national variety trial over nine locations (Holetta, Kulumsa,
Worabe, Areka, Adet, Gonder, Haromaya, Jimma, and Adet) for two years from
2017 to 2018 main cropping season. The design was RCBD with three
replications. Each germplasm planted in 4 rows with 0.4 m between rows, 0.1 m
between plants in 4 m lengths of 0.6 m between adjacent plots. NPS at the rate of
121 kg h-1 was applied at planting. Weeding and other management practices were
done as per recommendation. Data for the following agronomic traits were
recorded on a plot basis at all locations; days to 50% flowering (DTF) and days to
physiological maturity (DTM) for the entire plot, number of pods per plant (PPP),
number of seeds per pod (SPP), and plant height (PH). Data thousand seed weight
(TSW) from randomly selected 1000 seeds from each plot and grain yield (GY)
was taken from the entire plot. Grain yield adjustments were made by weighing
the oven drying and adjusting to a constant moisture level of 10%. The grain yield
recorded on a plot basis was converted to kg ha-1 for statistical analysis. The
analysis of variance (ANOVA) was performed using the PROC GLM in SAS
(SAS Institute, 2012).
[98]
Before the release of a variety of genotypes are evaluated for performance
comparative to the standard checks under various environmental conditions over
seasons and years. This enables the prediction of future performance of the
materials in farmers’ fields (Yan et al., 2000). This trial was evaluated over nine
locations for two years. The data only from sex locations (Holetta, Kulumsa,
Bekoji, Kofeele, Jimma and Gonder) were used for the statistical analysis. The
combined analysis of variance (ANOVA) revealed that significant (p < 0.001)
variation among the twelve genotypes for DTF, DTM, PH, SPP, PPP, TSW and
grain yield (GY) across locations and year are summarized in separate and
combined across locations and year (Table 1 ). Higher seed size is the biggest
reason for incorporating ICARDA large seed size faba bean materials in faba bean
breeding program and most of the materials were a bi-parental crosses of faba
bean materials introduced for large seed size like Atoma and Syria varieties. In
this trial, the weight of thousand faba bean seed ranged from 702.5 gm for the
standard check verity Gora to 935 gm and largest in genotype EH 010058-1
followed by EH 09028-3 with thousand seed of 912.7 gm. Most of the genotypes
evaluated have got the highest grain yield not statistically different from the
standard checks where the highest grain yield of 38.8.4 kg/ha obtained from the
standard check Gora followed by genotype EH 010008-5 with grain yield of 38.5
kg /ha. Large seed size was also an attempt of this trial thus out of the 12 faba
bean materials the genotype EH010058-1 and EH 09028-3 showed better seed
size over the two standard checks Gora and Tumsa with comparable yield
consistently over years (Table 1 ).

Two genotypes EH010058-1 and EH09028-3 showed comparable yield and better
in thousand seed weight than the two standard checks Tumsa and Gora.
Considering their high yield and large seed size these two genotypes were
recommended to be used as a parent material in the faba bean improvement
program.
[99]
Table 1: Combined analysis of Faba bean large seed size NVT for high potential area evaluated in 2017 and 2018 over locations
2017 2018
TRT DTF DTM PH PPP TSW GY q/ha PH PPP TSW GY q/ha
Gora 56.9cd 148.3ab 136.5a 13.8ab 782.3g 36.1abcd 130.5a 15a 702.5d 38.81a
EH 010002-1-1 54.1g 146.2b 125.2cd 14.3a 817.8fg 33.2d 120.45cde 13.15abcd 771.1cd 36.78ab
EH 010008-5 55.6ef 151.6a 135.5a 12.5bcd 952.0abc 35.5bcd 126.65ab 13.75abc 819.3bc 38.46a
EH 010051-1 55.1fg 152.4a 134ab 10.9d 964.8ab 36.4abcd 125.05abc 13.25abcd 816.7bc 37.5ab
EH 010058-1 55.8def 153a 129.5bc 12.3bcd 993.9a 35.8abcd 121.85bcd 11.45d 935a 36.41abc
EH 010058-2 54.3g 150.6ab 118.2f 13.4abc 796.1g 39.1a 117.4de 13.95ab 728.3d 37.05ab
EH 09012-1 58.0c 150.5ab 118.5f 12.4bcd 856.2ef 34.6cd 117.3de 12.95bcd 773.6cd 32.62c
EH 09017-5 54.8fg 149.8ab 124.1de 12.2bcd 886.4de 37.5abc 120.6cde 12.9bcd 821.4bc 37.97ab
EH 09021-1 68.42a 152.2a 120.3def 12.1bcd 946.9abc 35.1bcd 118.85de 13.4abc 819.4bc 34.05bc
EH 09028-3 62.8b 151.1a 119.8ef 12.1cd 982.8a 36.4abcd 115.8e 11.9cd 912.7a 36.52abc
EH 09046-3 56.5de 150.9ab 124cde 12.8abc 906.5cde 38.5ab 121.4bcd 13.8ab 820.3bc 37.18ab
Tumsa 55.1fg 152.9a 137.9a 13.4abc 919.3bcd 38.6ab 125.95abc 13bcd 865.7ab 37.54ab
Cv 3.7085 5.6648 6.8054 23.5 10.6 17.2349 7.285421 22.82423 14.444 16.0353
Mean 57.274 150.77 127 12.698 900.4 36.4104 121.8167 13.2083 815.53 36.7433
LSD 1.2086 4.8601 4.9182 1.7006 54.129 3.5708 5.5385 1.8814 73.51 3.6769
Where: DTF: Days to flowering; DTM: Days to maturity; PH: plant height; PPP: Pod per plant; TSW: Thousand seed weight; GY q/ha: Grain yield quintal per hectare.
[100]
Acknowledgments
Holetta Agricultural Research Center, Kulumsa ARC, Gonder ARC, Adet ARC,
Haramaya University, Jimma ARC and Worabe Agricultural Research Centers for
their collaboration and efforts in trial management and data collection.
Reference
Bouhassan, A, Sadiki M and Tivoli B. 2004. Evaluation of a collection of faba bean (Vicia faba L.)
genotypes originating from the Maghreb for resistance to chocolate spot (Botrytis fabae) by
assessment in the field and laboratory. Euphytica 135:55–62.
survey for 2017 / 2018 Addis Ababa, Ethiopia
López-Bellido, L, López-Bellido R, Redondo R and Benítez J. 2006. Faba bean nitrogen fixation in
a wheat-based rotation under rainfed Mediterranean conditions: Effect of tillage system. Field
Crops Research 98:253-260.
Lindemann, WC and Glover CR. 2003. Nitrogen fixation by legumes. In: Cooperative Extension
Service Guide A-129, editor New Mexico State University and the USDA.
Torres, AM, Rom B, Avila CMm, Satovic Z, Rubiales D, Sillero JC. 2006. Faba bean breeding for
resistance against biotic stresses: Towards application of marker technology. Euphytica
147:67-80.
Yan W, Hunt LA, Sheng Q, and Szlavnics Z. 2000 . Cultivar evaluation and mega environment
investigation based on the GGE biplot,” Crop Science 40:597-605.
[101]
[102]
Performance Evaluations of Field Pea
(Pisumsativum L.) (Kiki type) Genotypes
for High Potential Production Area
Asnakech T1., Musa J1., Gemechu K1., Sisay A1., Mesfin T1, Nigat T1.,Wondafrash M.1,
Gezachew Y2., Deressa T. 2, Kedir Y.2, Aliy R2, Jemal E.3, and Tadele T.4.
1
EIAR, Holetta Research Centre, P.O.Box 36, Ethiopia; 2Ethiopian Institute of Agricultural Research,
KulumsaResearch Centre, Ethiopia; 3Amhara Regional Agricultural Research Institute, Adet Research
Centre;4Oromiya Regional Agricultural Research Institute, Sinana Research Centre, Ethiopia.
Abstract
Field pea (Pisumsativum L) has high utility as a food and soil fertility improving crop.
Despite its multiples of importance, field pea production and productivity has been
below its potential due to biotic and abiotic stresses that require research priority.
Research efforts have been made however more is required to increase productivity
and to meet the national demand for field pea. Therefore, the present studies were
initiated with the main objectives of to develop stable, high yielding and stress-tolerant
field pea, varieties for different agro-ecology in Ethiopia and to verify the performance
of field pea candidate varieties under on-station and farmer field condition for national
release. A total of thirteen field pea genotypes selected from previous preliminary
variety trials were evaluated using RCBD with three replications as a national variety
trial over seven locations for two years from 2015 to 2016 main cropping seasons.
There were highly significant differences (p<0.001) among the genotypes for most of
the yield and yield component traits. Two genotypes EH-07006-5 and EH07007-3 were
found better in yield and other agronomic characters over the standard checks Burkitu
and Adi. These genotypes were recommended to be advanced to a variety verification
trial for release. Consequently, EH-07006-5 and EH07007-3 including the two
standard checks Burkitu and Adi were evaluated in a single plot of 10 m x 10 m
nationally across two regions at Holetta, Bekoji and Adet on station and on-farm field
conditions in 2017/18 main cropping season. These two candidate varieties were
evaluated by the national variety releasing technical committee and variety EH-07006-
5 was accepted and released to be used by field pea growers in similar agroecology of
the testing environments.
Keywords: Field pea, kiki type, stable, verification, yield
Introduction
Highland pulses in general and field pea in particular play important economic and
ecological roles in human diets and for the sustainability of the farming system.
Because of their nutritive (2-3 times higher protein content than cereals) and
market values, they provide substantial health and income opportunities. Canada
and China are the largest producers of field pea in the world and Ethiopia is the
most important producer of field peas as dry seed type in Africa and (FAOSTAT,
2017). It is a crop with better economic and partiality for improving the
[103]
livelihoods of poor farmers in Ethiopia in terms of area coverage and volume of
annual national production (CSA 2017/18). Field pea has been used as a major
break for cereal mono-cropping system which can increase soil fertility by fixing
atmospheric nitrogen (López-Bellido et al., 2006). The role of pulses as soil
fertility restorers shall be maintainable in the farming systems as environmentally
friendly and cost-effective production because of the rising price of commercial
fertilizers.
low and not as its potential due to several constraints including biotic and abiotic
stresses, socio-economic reasons and competition with other crops. Productions
have been constrained by several yield-limiting and reducing factors such as the
inherent low-yielding potential of the indigenous cultivars, foliar diseases like
Ascochyta blight, and powdery mildew, weed, and abiotic stresses like drought
and soil acidity are among important production constraints that require priority as
research resolutions. Productivity of field pea is increased from about 1.0 t/ha in
2006 to 1.67 t/ha in 2017 (CSA 2017/18). However, there is a gap between the
national average productivity using the currently available improved technologies
and on-farm productivity. These needs intended improvements of the field pea
varieties resolving the challenges of existing biotic and abiotic stresses. Ethiopia
being the secondary center of genetic diversity for field pea, there is a wealth of
gene pool that needs to make efforts to develop high yielding and tolerant verities
for different traits of interest. Consequently, improvement of field pea using a
different breeding strategy is in progress in field pea breeding program nationally
and these must continue to sustain the release of high yielding and stress-resistant
/ tolerant field pea varieties. Therefore, the main objectives of these trials were to
develop stable, high yielding and disease resistant field pea variety for field pea
growers in Ethiopia and to verify the performance of field pea candidate variety
under on-station and farmers’ field condition for release Nationally.

Field pea (Kiki type) Variety Evaluation
A total of 13 field pea selected materials from previous preliminary variety trials
were evaluated as national variety trial over eleven locations for two years from
between plants in 4 m lengths. DAP at the rate of 100kgh-1 was applied at planting.
Weeding and other management practices were done as per recommendation. Data
for the following agronomic traits were recorded on a plot basis at all locations;
days to 50% flowering (DTF) and days to physiological maturity (DTM) for the
entire plot, number of pods per plant (PPP), number of seeds per pod (SPP) and
plant height (PH). Disease scoring made from the entire plot using 1 to 9 scoring
[104]
method Data thousand seed weight (TSW) from randomly selected 1000 seeds
kg ha-1 for statistical analysis. The analysis of variance (ANOVA) was performed
using the PROC GLM in SAS (SAS Institute, 2012). The data analyses for
interpreting genotype by environment interaction were done by GGE stability
analysis R software.
Field pea Verification:

The two genotypes EH-07006-5(G22763-2C X KFP-11) and EH07007-3
(G22762-2C X MG102521 and IFPI 5258) selected from previous national variety
trial including the two standard checks Burkitu and Adi were evaluated nationally
at three on-station across two regions Amhara and Oromiya Agricultural Research
Centers (Holetta, Bekoji, and Adet) and at six on-farm that is two per each on
station in 2017/18 main cropping season. The candidate varieties including the
standard checks were planted in a single plot of 10 m x 10 m and all recommended
agronomic practices were applied. All the data for the following agronomic traits
were recorded on a plot basis at all locations; days to 50% flowering (DTF) and
days to physiological maturity (DTM) for the entire plot, number of pods per plant
(PPP), number of seeds per pod (SPP) and plant height (PH). Powdery mildew and
Ascochyta blight disease severity from the entire plot was recorded using a 1-9
scale. Data on grain yield (GY) was taken from the entire plot. Grain yield
kg ha-1 for statistical analysis. The mean grain yield of the candidate varieties and
standard checks were calculated on station and on-farm bases.

Field pea (Kiki type) Variety Evaluation:
The data from seven locations were used for analysis. The combined analysis of
variance (ANOVA) revealed that highly significant (p < 0.001) variation among
the thirteen genotypes for DTF, DTM, PH, SPP, PPP, TSW, and grain yield (GY)
across locations and year (Table 1, & 2). Out of the 13 faba bean materials, the
genotype EH-07006-5 was found with 15.7 to 18.5 % yield advantage and 19.3 to
30.1% thousand seed weight advantage over the standard checks Burkitu and Adi.
The second best genotype identified was EH07007-3 with 12.6 to 15.8 % yield
advantage and 18.1 to 28.9% thousand seed weight advantage over the standard
checks Burkitu and Adi. These genotypes were also performed better in other
agronomic traits across locations and over two years (Table 2). Therefore, these
genotypes were recommended to be submitted for the national variety releasing
committee and evaluated nationally for release.
[105]
The ‘which –won-where’ pattern of GGE biplot
The G x E interaction was further explored through the genotype and genotype x
environment GGE biplot analysis. The GGE biplot constructed by plotting the first
two principal components (PC-1 and 2) to explore the genotype x environment
interaction. The PC1 explained 55.42% of total variation and PC2 explained 22%,
thus together they accounted for 77.42% of the variation for the genotype and
genotype x environment interaction for the grain yield of the field pea evaluated
over five locations. The GGE biplot compares the ‘ideal genotype’ with 13 field
pea genotypes shown in Figure 1. The small circle, which is located on the
average environment coordinate (AEC) abscissa and with an arrow pointing to it,
represents the ideal genotype. It has the highest yield of the entire dataset and it is
stable (Yan and Kang, 2002). In this case genotype G4 followed by G5 was
closest to the hypothetical ideal genotype and therefore, most desirable of all
tested genotypes including from the standard checks (Figure 1).
Verification trial of the candidate genotypes EH-07006-5 and EH07007-3 was

conducted on station and on-farmers fields for release. Both candidate varieties
were performed well during the field evaluation by the technical committee (Table
3, and Figure 1). As a result, the National Vatietey Releasing Commeete decided
the candidate variety EH-07006-5 to be released nationally for its best
performance over locations and years compared to the two standard checks.
However, the second candidate verity EH07007-3 was rejected by the National
Variety Releasing Committee. The released field pea variety EH-07006-5 can be
used by the field pea growers in potential highlands of Ethiopia.
[106]
Table 1. Mean grain yield (Kg ha-1) performance of field pea varieties tested in NVT at 11 environments 2016 and 2017
No. Genotype Environments (location by year) Mean

Adet Adet Asasa Asasa Bekoji Bekoji Holetta Holetta Jeldu Kofele Sinana
2016 2017 2016 2017 2016 2017 2016 2017 2016 2017 2016
1 EH 07002-1 2239 1357 3754 3689 4579 2934 3199 1850 4164 4280 639 2971
2 EH 07005-1 1904 1485 4054 4194 4088 3417 2542 1871 4135 2902 693 2844
3 EH 07006-5 2271 1796 3619 5037 5560 3604 3292 2377 6113 3450 721 3440
4 EH 07007-3 2345 1834 3718 4213 5213 3670 3498 2706 5820 3120 689 3348
5 EH 08003-1 2514 1531 3713 4768 4588 3152 2682 2653 6686 3612 679 3325
6 EH 08003-2 3076 1949 3551 4818 4847 3120 3251 2737 4960 4130 643 3371
7 EH 08041-1 2803 1524 3716 4198 4754 3377 2880 2065 4669 4578 532 3190
8 EH 08041-3 1824 1516 2244 4115 4404 2939 2805 2721 5989 2846 559 2906
9 EH 08042-2 2633 1665 3101 3885 3856 2961 2606 1918 5879 2953 640 2918
10 EK 08024-1 2295 1198 3142 3230 5358 3872 1792 896 4979 2778 765 2755
11 EH 05048-3 2822 1645 3976 4622 5298 3413 2472 2121 5295 3504 641 3255
12 Adi 2142 1516 3199 3873 5639 3732 2089 1894 5208 2866 559 2974
13 Burkitu 2460 1550 3758 3590 5076 3128 2157 1991 4464 3063 608 2895
Mean 2410 1582 3503 4172 4866 3332 2713 2138 5258 3391 644 3091.7
CV (%) 15.9 22.3 25.8 13.1 13.4 12.7 21.5 15.1 10.5 13.3 5.2 17.1
LSD (5%) 548.4 506 1295 785.1 936.2 608.6 834.7 463.8 789.2 645 48.1 203.6
[107]
Table 2. Combined analysis of field pea genotypes (Kiki type) tested in NVT over seven locations from 2016 to 2017
No.Genotype AB PM DTF DTM DTGF PHT PPP SPP TSW

1 EH 07002-1 3.4 2.4 73 142 69 157 12 5 198
2 EH 07005-1 3.5 2.7 72 143 71 161 12 4 224
3 EH 07006-5 3.6 2.5 71 139 67 143 12 4 210
4 EH 07007-3 3.5 2.5 73 137 64 145 13 4 208
5 EH 08003-1 3.9 2.4 72 140 68 150 12 4 209
6 EH 08003-2 3.7 2.9 75 141 66 151 12 5 192
7 EH 08041-1 3.6 2.5 74 140 65 156 14 4 171
8 EH 08041-3 3.6 2.7 74 141 67 156 12 5 187
9 EH 08042-2 3.5 2.7 72 139 68 158 12 5 198
10 EK 08024-1 3.8 2.1 70 136 66 145 11 5 197
11 EH 05048-3 3.6 2.6 72 140 68 153 12 5 196
12 Adi 3.5 2.3 72 138 66 155 12 5 161
13 Burkitu 3.7 2.6 71 138 67 144 12 4 176
Mean 3.6 2.5 73 140 67.1 152 12 4.5 194
CV (%) 15 23 7.3 4.9 8.2 9 31 18 7.2
LSD (5%) 0.2 0.2 2 2.67 2.12 5.2 1.4 0.3 5.4
DTF, DTM, DTGF = number of days to flowering, maturity and grain filling, respectively; PHT=plant height (cm); PPP and
SPP=number of pods per plant and seeds per pod; AB =Ascochyta blight (in 1-9 scale); PM=Powdery mildew (1-9 scale);
TSW=thousand seeds weight (g)
Ranking Genotypes
11
10
1 Jel
5
Bek 9
10
AXIS2 22 %
13 6
0
12 4
8 5
32
Adt
-5
7
-10
Asa
Hol
-15
-5 0 5 10 15 20 25
AXIS1 55.42 %
Figure 1. The Which won where or which is best at what’ for grain yield based on genotype x environment of 13 Field
pea genotypes evaluated in five environments.
[108]
Table 3: Yield performance of the two candidate field pea genotype and the standard checks under on-farm
and on-station condition 2017/18
Varieties Grain Yield in qha-1

Candidate genotype EH 07006-5 On-station 3.4 - 3.7
On-farm 2.5 - 2.8
Candidate genotype EH07007-3 On-station 2.2 - 3.9
On-farm 3.4
Burkitu (Standard check) On-station 2.5 - 3.5
On-farm 2.4 - 3.3
Adi (Standard check) On-station 2.2 - 3.2
On-farm 2.4 - 2.9

The two genotypes EH-07006-5 and genotype EH07007-3 were identified best for
their grain yield and other agronomic characteristics compared to the two standard
checks Burkitu and Adi. Therefore, these genotypes were recommended to be
submitted for the national variety releasing committee to be evaluated nationally
for release. Thus, based on the field verification trial executed over location on
farmer’s field and research station the field pea candidate variety EH-07006-5 was
released nationally and recommended for field pea growers in potential highland
areas of Ethiopia and other similar agroecology.
Acknowledgments
Holetta Agricultural Research Center, Kulumsa ARC, Adet ARC, and Worabe
Agricultural Research Centers for their collaboration and efforts in trial
management and data collection.
Reference
Abel T, Tomas B, Kifle D and Mulatu G. 2015 Assessment of genetic diversity in Ethiopian field
pea (Pisum sativum L.) accessions with newly developed EST-SSR markers. BMC
Genetics 16: 102.
survey for 2017 / 2018 Addis Ababa, Ethiopia
FAOSTAT. 2018. Data base. Available at: http://fao.org/faostat.
[109]
Girma M, Chemeda D, Abaya T, Dangachew L, Negash G. 2011. Genotype x Environment
Interaction for Yield in Field Pea (Pisum sativum L.). East Africa Journal of Scienc. 5: 1
(6-11)
López-Bellido, L, López-Bellido R, Redondo R and Benítez J. 2006. Faba bean nitrogen fixation in
a wheat-based rotation under rainfed Mediterranean conditions: Effect of tillage system.
Field Crops Research 98:253-260.
Yan, W. 2002. Singular-value partitioning in biplot analysis of multienvironmental trial data.
Agronomy Journal 94:990-996.
Yan W, Hunt LA, Sheng Q, and Szlavnics Z. 2000 . Cultivar evaluation and mega environment
investigation based on the GGE biplot. Crop Science 40:597-605.
[110]
Evaluation of Field Pea (Pisumsativum L) Kiki and
Shero Yype Genotypes for High Potential
Production Area
Asnakech T1., Musa J1., Gemechu K1., Nigat T1., Sisay A1., Mesfin T1, Wondafrash M.1,
Aliy R2. Kedir Y. 2, Jemal E.3 Nano A.4, and Tadele T.5.
1
Ethiopian Institute of Agricultural Research, Holetta Research Centre, P.O.Box 36, Ethiopia; 2Ethiopian
Institute of Agricultural Research, KulumsaResearch Centre, Ethiopia; 3Amhara Regional Agricultural
Research Institute, Adet Research Centre; 4Haramaya University, thiopia; 5Oromiya Regional Agricultural
Research Institute, Sinana Research Centre, Ethiopia
Abstract
Field pea (Pisum sativum L) has high utility as a food and soil fertility improving
crop. Despite its multiples of importance, field pea production and productivity has
been below its potential due to several constraints including inherent low-yielding
potential of the local cultivars, biotic and abiotic stresses that require research
priority. Research efforts have been made for the improvement of productivity of field
pea however more is required to increase the productivity of the crop and to meet the
national demand. Therefore, the present study was initiated with the main objectives
of to develop stable, high yielding and stress-tolerant field pea, varieties for different
agro-ecology in Ethiopia. A total of ten field pea genotypes selected from previous
preliminary variety trials were evaluated using RCBD with three replications as a
national variety trial over seven locations for two years from 2017 to 2018 main
cropping seasons. There were highly significant differences (p < 0.001) among the
genotypes for most of the yield and yield component traits. Field pea genotype
EH010011-3 was found better in yield and other agronomic characters over the
standard checks Burkitu, Adi, Bilalo and Bursa. This genotype was recommended to
be advanced and evaluated in a variety verification trial in the 209 cropping season
Keywords: Evaluation, Field pea, stable, yield, thousand seed weight
Introduction
Highland pulses in general and field pea in particular play important economic and
ecological roles in human diets and for the sustainability of the farming systems.
Because of their nutritive (2-3 times higher protein content than cereals) and
market values, they provide substantial health and income opportunities. Canada
and China are the largest producers of field pea in the world and Ethiopia is the
most important producer of field peas as dry seed type in Africa and (FAOSTAT,
2017). It is a crop with better economic and partiality for improving the
livelihoods of poor farmers in Ethiopia in terms of area coverage and volume of
annual national production (CSA 2017/18). Field pea has been used as a major
break for cereal mono-cropping system which can increase soil fertility by fixing
atmospheric nitrogen (López-Bellido et al., 2006). The role of pulses as soil
[111]
fertility restorers shall be maintainable in the farming systems as environmentally
friendly and cost-effective production because of the rising price of commercial
fertilizers.
low and not as its potential due to several constraints including biotic and abiotic
stresses, socio-economic reasons and competition with other crops. Productions
have been constrained by several yield-limiting and reducing factors such as the
inherent low-yielding potential of the indigenous cultivars, foliar diseases like
Ascochyta blight and powdery mildew, weed, and abiotic stresses like drought and
soil acidity are among important production constraints that require priority as
research resolutions. Productivity of field pea is increased from about 1.0 t/ha in
2006 to 1.67 t/ha in 2017 (CSA 2017/18). However, there is a gap between the
national average productivity using the currently available improved technologies
and on-farm productivity. These needs intended improvements of the field pea
varieties resolving the challenges of existing biotic and abiotic stresses. Ethiopia
being the secondary center of genetic diversity for field pea, there is a wealth of
gene pool needs to make efforts to develop high yielding and tolerant verities for
different traits of interest. Consequently, improvement of field pea using a
different breeding strategy is in progress in field pea breeding program nationally
and these must continue to sustain the release of high yielding and stress-resistant
/ tolerant field pea varieties. Therefore, the main objectives of these trials was to
develop stable, high yielding and disease resistant field pea variety for field pea
growers in Ethiopia and to verify the performance of field pea candidate variety
under on-station and farmers’ field condition for release Nationally.

Field pea (Kiki and Shero type) Variety Evaluation
A total of 10 field pea selected materials from previous preliminary variety trials
were evaluated as a national variety trial over nine locations (Holetta, Jeldu, Adet,
Kulumsa, Kofele, Bekoji, Haromaya, Sinana, and Mekele) for two years from
between plants in 4 m lengths. NPS at the rate of 121 kg-1 was applied at planting.
Weeding and other management practices were done as per recommendation. Data
for the following agronomic traits were recorded on a plot basis at all locations;
days to 50% flowering (DTF) and days to physiological maturity (DTM) for the
entire plot, number of pods per plant (PPP), number of seeds per pod (SPP), and
plant height (PH). Disease scoring made from the entire plot using 1 to 9 scoring
method Data thousand seed weight (TSW) from randomly selected 1000 seeds
[112]
kg ha-1 for statistical analysis. The analyses of variance (ANOVA) were
performed using the PROC GLM in SAS (SAS Institute, 2012). The data analyses
for interpreting genotype by environment interaction were done by GGE stability
analysis R software.

The data used for the ANOVA were only from six locations (Holetta, Jeldu, Adet,
Kulumsa, Kofele, and Bekoji). The combined analysis of variance (ANOVA)
revealed that highly significant (p < 0.001) variation among the ten genotypes for
DTF, DTM, PH, SPP, PPP, TSW and grain yield (Gy) across locations and year
(Table 1, & 2). Out of the 10 field pea materials the genotype EH010011-3 was
found with 20 to 23 % yield advantage and 13 to 23% thousand seed weight
advantage over the standard checks Burkitu and Adi. These genotypes were also
performed better in other agronomic traits across locations and over two years
(Table 3). Therefore, this genotype was recommended to be submitted for the
national variety releasing committee and evaluated nationally for release.
The G x E interaction was further explored through the genotype and genotype x
environment GGE biplot analysis. The GGE biplot was constructed by plotting the
first two principal components (PC 1 and 2) to explore the genotype x
environment interaction. The PC1 explained 56.6% of total variation and PC2
explained 22.3%, thus together they accounted for 78.9% of the variation for the
genotype and genotype x environment interaction for the grain yield of the field
pea evaluated over six locations. The polygon view of the GGE biplot indicated
which cultivar performed best in which environment (Figure 4.1 A). In this case,
five genotypes were identified as vertex genotypes. The vertex genotypes were
either the highest or lowest yielding in the locations which fell within the same
sectors with those genotypes indicating different genotypes won in different
environments (Yan et al., 2010). Based on this, Genotypes G5 and G1 were the
vertices genotype where Jeldu, Bekoji, Adet, and Kofele fell, thus were the highest
yielding in these environments. Genotypes fell within the polygon, indicating that
they were less responsive than the vertex cultivars for the interaction. The GGE
biplot compares the ‘ideal genotype’ with 10 field pea genotypes shown in Figure
4.1 B. The small circle, which is located on the average environment coordinate
(AEC) abscissa and with an arrow pointing to it, represents the ideal genotype. It
has the highest yield of the entire dataset and it is stable (Yan and Kang, 2002). In
this case genotype G5 was closest to the hypothetical ideal genotype and
therefore, most desirable of all tested genotypes including the standard checks.
Field pea genotype EH010011-3 was recommended to be advanced to a variety
verification trial and evaluated under on-farm and on-station field condition in
2019 cropping season.
[113]
Table 1: Combined analysis of Field pea NVT (Kiki and Shero type) evaluated over six locations from 2017/18 to 2018/19
Entry Type PM AB DF DTM PH PPP TSW Gyqha-1

Bilalo Kule-shro 3.0c 3.5abc 71.1bc 138.3ab 155.9a 9.6c 196.0a 31.3ab
Bursa Shro 4a 3.4bc 69.9d 137.8bc 154.3a 10.1abc 179.1d 28.8bc
EH010006-2 Kule-shro 3.5b 3.5abc 71.4b 137.4bcd 154.8a 10.7a 158.5f 27.7c
EH010009-1 Kiki 2.8c 3.5abc 70.3cd 136.6ed 156a 9.6c 191.8abc 27.9c
EH010011-3 Kule-kiki 3.1bc 3.6ab 70.9bc 136.8cd 151.3abc 10.6ab 194.0ab 32.5a
EH010009-2 Kule-Kiki 3.4b 3.8a 69.7d 135.7ef 146.3cd 9.8bc 190.6bc 26.7c
Adi kiki 2.9c 3.7ab 68.7e 135.4f 148.1bcd 10.8a 156.9f 26.3c
Burkitu kiki 3.1c 3.5abc 68.7e 135.6f 144.3d 10.6ab 171.2e 26.9c
EH010004-1 Shro 3.5b 3.6ab 71.7b 137.4bcd 152.0abc 10.7a 187.8c 29.2bc
EH010005-2 kiki 3c 3.3c 73.9a 138.7a 153.5ab 9.4c 176.4d 28.9bc
CV 21.490 16.241 2.988 1.725 9.609 22.105 7.108 25.148
Mean 3.259 3.481 70.673 137.017 151.696 10.220 180.248 28.657
LSD 0.345 0.279 0.848 0.949 5.851 0.907 5.144 2.893
Where: AB: Ascochyta blight severity, PM: Powdery mildew severity (1-9 scale); DTM: Days to maturity; DF: Days to
flowering PH: Plant height; PPP: number of pod per plant; TWS: Thousand seed weight in gram GY: Grain yield quintal
per hectare;
Table 4.2: Mean performance of grain yield and thousand seed weight of Field pea genotypes from NVT (Kiki and Shero
type) evaluated over six locations in 2017 and 2018
2017 2018
Entry Type TSW Gyqha-1 TSW Gy qha-1
Bilalo Kule-shro 190a 25.2a 201a 37.5ab
Bursa Shro 173b 23.1ab 184c 34.6bc
EH010006-2 Kule-shro 153c 22.9ab 163e 32.5c
EH010009-1 Kiki 187a 24.1a 196ab 31.7c
EH010011-3 Kule-kiki 188a 25.7a 199a 39.5a
EH010009-2 Kule-Kiki 183a 19.5cd 197ab 34.1bc
Adi kiki 155c 20.5bcd 158e 32.c
Burkitu kiki 167b 18.9d 174d 34.9abc
EH010004-1 Shro 183a 22.4abc 191b 35.9abc
EH010005-2 kiki 174b 22.8abc 177cd 35.0abc
CV 7.684 27.0673 6.72675 23.7503
Mean 175.746 22.5204 184.75 34.794
LSD 7.7024 3.4769 7.0887 4.7136
Where: TSW: thousand seed weight; Gy qha-1: Greain yield quntal per heactar
Table 4.3: Mean performance of other agronomic traits of Field pea genotypes from NVT (Kiki and Shero type) evaluated
over six locations in 2017 and 2018
2017 2018
Entry Type DTF DTM PH PPP DTF DTM PH PPP
Bilalo Kule-shro 74.7bc 142.1ab 153a 7.9b 67bc 134ab 158a 11.4abc
Bursa Shro 74.1bcd 141.3abc 149ab 8.6ab 65de 134abc 158a 11.7abc
EH010006-2 Kule-shro 75b 141.3abc 155a 9.1a 67b 133bcd 154ab 12.3ab
EH010009-1 Kiki 73.9bcde 140.9bc 158a 8.2ab 66cd 132def 153ab 11.1abc
EH010011-3 Kule-kiki 74.8bc 140.7cd 154a 9.0ab 67bc 133cde 148bc 12.2ab
EH010009-2 Kule-Kiki 73.4cde 139.4de 143b 8.7ab 66d 132ef 149bc 10.8bc
Adi Kiki 72.7de 139.3de 149ab 9.2a 64f 131f 146bc 12.4a
Burkitu Kiki 72.6e 139.2e 144b 8.7ab 64ef 131ef 144c 12.5a
EH010004-1 Shro 75.5b 141.5abc 154a 8.9ab 68b 133bcd 150bc 12.6a
EH010005-2 Kiki 77.5a 142.5a 154a 8.2ab 70a 135a 152abc 10.6c
CV 3.339 1.754 10.626 22.309 2.6117 1.7437 9.1169 22.6096
Mean 74.4 140.833 151.8 8.6483 66.9458 133.2 151.6 11.7917
LSD 1.4168 1.409 9.198 1.101 0.9973 1.3248 7.8855 1.5207
Where: DTF: days to flowering; DTM: days to maturity; PH: plant hight; PPP: pod per plant
[114]
Which Won Where/What Ranking Genotypes
A B
2 Kof 2 Kof
5
5
1 1
AXIS2 22.29 %
7 6
AXIS2 22.29 %
7 6
Adt
0
8 Adt
0
8
3 10 3
4 10
4
9 9
Hol Hol
Kul Kul
-5
-5
Bek Bek
Jel5 Jel5
-10
-10
-5 0 5 10 15 -5 0 5 10 15
AXIS1 58.57 % AXIS1 58.57 %
Figure 1. Polygon View of GGE biplot, showing best genotype (A) and Which won where or which is best at what’ for grain
yield (B) based on genotype x environment of 10 Field pea genotypes evaluated in nine environments.
Reference
CSA. 2018. Report on area and production of crops. Central Statistics Agency agricultural
sample survey for 2017 / 2018 Addis Ababa, Ethiopia
FAOSTAT. 20117. Database. Available at: http://faostat3.fao.org/faostat.
López-Bellido, L, López-Bellido R, Redondo R and Benítez J. 2006. Faba bean nitrogen
fixation in a wheat-based rotation under rainfed Mediterranean conditions: Effect of
tillage system. Field Crops Research 98:253-260.
Yan, W. 2002. Singular-value partitioning in biplot analysis of multi environmental trial
data. Agronomy Journal 94:990-996.
Yan, W, Frégeau-Reid J, Pageau D, Martin R, Mitchell-Fetch J, Etienne M. 2010.
‘Identifying essential test locations for oat breeding in Eastern Canada’. Crop
Science 50:504-515.
[115]
[116]
Evaluation of Large Red Bean Genotypes for
Lowland Areas of Ethiopia
1Abel Moges Firew, 1Berhanu Amsalu Fenta, 1Dagmawit Tsegaye, 1Tigist Shiferaw,
1Kassaye Negash, 1Kidane Tumssa, 1Behailu Tesfaye, 1Semir Hashim, 1Girum Kefle,
2Yasin Goa, 3Zeru Yimer, and 4Clare Mugisha Mukankusi
1
Ethiopian Institute of Agricultural Research, Melkassa Research Centre, P.O.Box 436, Ethiopia
2
Southern Agricultural Research Institute, Areka Research Center, P.O.Box, Areka , Ethiopia
3
Ethiopian Institute of Agricultural Research, Pawi Research Centre, P.O.Box, Pawi. Ethiopia
4
nternational Center for Tropical Agriculture (CIAT), P.O.Box 6247, Kampala, Uganda
Corresponding author: E-mail: abvan23@gmail.com
Abstract
Production of large red bean types in Ethiopia is low because of a limited number of
high yielding varieties and adverse environmental conditions. Therefore, the
objective of this study was to evaluate the performance of large red bean genotypes
across bean growing areas of the country. The experiment was executed at eleven
locations during the 2015-2017 main season using sixteen genotypes including two
checks. The experimental design was 4 × 4 triple lattice across environments. The
study result revealed that large red bean genotypes show tsubstantial genetic
variation for grain yield and yield component traits. The combined analysis of
variance over environments explained bean grain yield was significantly (p<0.001)
affected by environments (66.07%), genotypes (5.4%), and genotype by environment
interaction (14.96%). Additive main effects and multiplicative interaction (AMMI)
and genotype main effect and genotype by environment interaction (GGE) biplot
models were applied to analyze and visualize the pattern of the interaction
component. Genotypes G14 (DAB 481) and G15 (Melkadima) combined both high
mean yield and relatively stable genotypes across environments. In addition to yield
performance by considering measured traits and additional quality traits (seed coat
color bleaching behaviour, cooking and canning quality) best performing genotypes
could be y considered for variety verification or it can be used as parental material
to improve common bean breeding
Keywords: Large red bean, Ethiopia, AMMI, GGE biplot, stability, yield.
Introduction
Common beans (Phaseolus vulgaris L.) are the most important grain legumes for
direct human consumption and as a source of income dominantly in the lowland
and mid-altitude areas of Ethiopia. The current annual national production of
common bean in the country is estimated to be 290,202 hectares with a total
production of 483,923 tons and average productivity of 1.67 tons per hectare
(CSA, 2017). Common bean is preferred by the majority of Ethiopian farmers
because of its fast maturing characteristics that enable households to get cash
income required purchasing food and other household needs when other crops
have not yet matured (Legesse et al., 2006).
[117]
In Ethiopia, the common bean has diverse production and utilization in the major
production areas. For instance, the white pea bean types are largely produced in
the Ethiopian Great Rift Valley mainly for export markets (Yayis et al., 2011).
While, the southern bean producing zones that consist of Sidamo, Wolaita and
Gamo Gofa zones produce red and mottled bean types mainly for household
consumption. On the other hand, the large red bean production in many parts of
the country is low because of the limited number of high yielding varieties and
adverse environmental conditions. Yield is a complex character that depends on
several other attributable traits and is highly influenced by genetic and
environmental factors. Hence, the analysis of genotype by environment interaction
(GEI) is an important aspect to be considered in the breeding and variety
development programs. GEI refers to the differential response of different
genotypes across environments (Kang, 2004). Some genotypes are characterized
by stable phenotypic performance in a wide range of environments, while others
display considerable variation over environments. To this effect, the analysis of
GEI is found to be essentially important to identify superior and stable genotypes
in multi-environment trials.
There are arrays of statistical tools for analyzing GEI data. Additive main effects
and multiplicative interaction (AMMI) and genotype main effect and genotype by
environment interaction (GGE) biplot models are capable of summarizing the
interaction incomprehensible ways, which will enable breeders to visualize
phenotypic stability across environments from multi-locational trial data
(Thillainathan and Fernandez, 2001) and allows easy identification of superior
genotypes in each environment (which-won-where) (Yan et al., 2007). Therefore,
the objective of this study was to assess the effect of GEI and the stability of large
red bean genotypes in diverse bean growing areas of Ethiopia.

Sixteen promising large red bean genotypes advanced from preliminary national
variety trial (PNVT) of CIAT introduction materials including two check varieties
called Melkadima and Dark Red Kidney (DRK) were evaluated at six bean
growing locations of Ethiopia using 4 x 4 triple lattice design from 2015-2017.
The test locations include Melkassa, Meiso, Pawe, Arsinegelle, Goffa, and
Alemtena. Each year and location was treated as a separate environment, making
11 test environments of the study. The spacing between rows and plants was 40
and 10cm, respectively. Each genotype was planted in 6 rows in 4m plot length
where only the four central rows were used for data recording. Di-ammonium
phosphate (DAP) fertilizer was applied using a 100 kg/ha rate during planting.
Days to 50% flower, days to 90% physiological maturity, grain yield (kg ha-1),
common bacterial blight (1-9), halo blight (1-9), angular leaf spot and rust were
assessed based on plot bases, while pod per plant and seed per pod were recorded
[118]
on five random samples of plants selected from the central four rows of each plot.
Grain yield measured from the entire four rows of each plot was converted into kg
ha-1 at 12.5% moisture content.
Statistical Analysis
AMMI analysis uses ANOVA and PCA for estimating stability and GEI (Gauch
1992).
Where: = the mean yield of genotype g in environment e, = the grand mean,

= the deviation of the genotype mean from the grand mean, = the deviation of
the environment mean from the grand mean, = the singular value for the IPCA n,
= the number of PCA axis retained in the model, = the PCA score of a
genotype for PCA axis n, = the environmental PCA score for PCA axis n, =
the AMMI residual and = the residuals.
GGE-Biplot model based on singular value decomposition (SVD) of the first two
principal components was used to fit the GGE biplot model (Yan, 2002).
Where, is the trait mean of genotype i in environment j, is the grand mean,

is the main effect of environment j, being the mean yield across all
genotypes in environment j, and are the singular values (SV) for the first and
second principal components (PC1 and PC2), respectively, and are
eigenvectors of genotype i for PC1 and PC2, respectively, and are
eigenvectors of j for PC1 and PC2, respectively, is the residual associated with
genotype i in environment j.

GEI interaction and genotypic mean performance
Combined analysis of variance showed highly significant differences (p<0.01) for
environments, genotypes, and GEI (Table 1). In the study, the largest variance was
due to the environment (66.07%) followed by GEI (14.96%) and genotype effects
(5.4%). A significant effect of GEI demonstrates the differential performance of
genotypes in different environments (Table 1). Moreover, GEI exerted a three
times larger effect than the genotypes which complicates selection for superior
and adaptable genotypes. Therefore, simultaneous consideration of both higher
mean grain yield and varietal stability is essentially important in the process of
selecting superior genotypes.
[119]
Table 1. Combined analysis of variance for grain yield (kg/ha) of 16 large red common bean genotypes at 11
environments.
Source of Degrees of Sum of Mean Total Variation

G × E explained (%)
variation freedom Square square explained (%)
Total 527 391213302 742340
Treatments 175 338120300 1932116
Genotypes 15 21120119 1408008*** 5.40
Environments 10 258487351 25848735*** 66.07
Block 22 13326547 605752*** 3.41
Interactions 150 58512829 390086*** 14.96
IPCA 1 24 23654595 985608*** 40.43
IPCA 2 22 13648931 620406*** 23.33
IPCA 3 20 6660303 333015*** 11.38
IPCA 4 18 5105284 283627** 8.73
Pooled error 330 39766455 119997 10.16
*** And ** are significant at 0.001 and 0.01 probability levels respectively; IPCA = Interaction Principal Component
Analysis
The yield performances of 16 genotypes for each environment as well as their

average performance across eleven environments are presented in Table 2. The
mean environmental grain yield across genotypes ranged from the lowest of 906.9
kg/ha at Goffa in 2016 to the highest 3336 kg/ha at Arsinegelle in 2017. The
highest yield of 4222.4 and 4244.6 kg/ha was obtained from genotypes DAB 513
and DAB 496 at Melkassa in 2017, while one of the check cultivar, DRK, gave the
lowest grain yield about 580 kg/ha at Goffa site in 2016. Genotypes DAB 496
with grain yield of 2626, DAB 544 with grain yield of 2621.9, DAB 481 with
grain yield of 2576 and DAB 513, with grain yield of 2624 kg/ha, record the best
in terms of mean grain yield across environments (Table 2). However, yields of
four genotypes were not significantly different from the best check variety called
Melkadima which yielded 2649.4 kg/ha. Whereas, the other checks variety, DRK,
together with the other test genotype DAB 482 gave the lowest grain yield across
environments.
The mean performance of other agronomic traits across locations is presented in

Table 2. The flowering date of genotypes across environments ranged from 34 to
39 days. Among genotypes evaluated, DAB 544 was the earliest flowering (34)
genotype, while DAB 496 was found to be a late flowering genotype (38.7) (Table
2). Statistically, there was a significant difference for days to flowering which
ranged from 77 days for DAB 525 to 80.6 days for DAB 540, however, this
difference was narrow to classify genotypes according to their maturity group. The
highest mean number of pod per plants (14.4) was obtained from genotype DAB
523 and the lowest (9.1) was obtained from DAB 481. The mean number of seeds
per pod was ranged from 2.8 to 3.5.
[120]
Figure 1. Mean disease reaction (1-9 scoring scale) of large red bean genotypes from 2015 to 2017 cropping
season. Abbreviations: CBB: Common bacterial blight, HB: Halo blight, ALS: Angular leaf spot
Figure 1 showed the effects of major diseases; common bacterial blight

(Xanthomonas campestris), halo blight (Pseudomonas syringae), angular leaf spot
(Phaeoisariopsis griseola) and rust (Uromyces appendiculatus) on large red bean
genotypes across environments from 2015 to 2017 main cropping seasons. All
genotypes including check varieties showed a high level of resistance to rust,
while the reaction to angular leaf spot and halo blight were varied from resistance
to moderately resistance. However, reaction to common bacterial blight genotypes
showed moderate resistance. Two best performing genotypes, DAB 496 and
Melkadima were identified for a good level of multiple disease resistance except
common bacterial blight (Figure 1).
[121]
Table 2. Grain yield and other agronomic parameters /traits of 16 large red bean genotypes evaluated across 11 environments of Ethiopia.
Mean grain yield (kg/ha)

Code Genotype
AN15 AN17 AN18 AT16 AT17 GF16 ME16 MK16 MK17 MK18 PW16 Yield Mean Mean Mean Mean
GM SPP PPP FD MD
G1 DAB 497 2006.9 3352.3 3355 1922.7 2367 884.2 1904.2 1666.2 3420.5 2317.4 1143.7 2212.7def 3.0efg 10.2c-f 36.4a-d 79.6bc
G2 DAB 317 2781.4 3037 2064.4 2444.8 2078.3 939.4 2703.4 2766.7 2519.6 2521.7 1373.3 2293.6cde 3.2c-f 13.6ab 35.2bcd 77.4g
G3 DAB 482 1580.5 2563.5 3192.6 2083 1991.9 943.5 1969.1 2742 1946 2773.1 1227.7 2092.1f 2.9fg 11.2cde 36.9ab 79.8abc
G4 DAB 523 2679.8 3433 2612.2 2519.9 2314.1 831.6 3201.7 1920.3 2843.2 2860.8 1191 2400.7bc 2.9fg 14.4a 35.5bcd 80.3ab
G5 DAB 512 2044.9 2500.1 3228.3 2346.8 2064.3 607.9 2744.2 2181.5 3758.1 2663.1 1184.4 2302.1cde 3.1d-g 10.9c-f 36.3a-d 79.3cd
G6 DAB 513 2320.2 3575.7 2782.4 2389.3 2963.5 1106.3 2940.4 2599.1 4222.4 2631.3 1333.8 2624.0a 3.4abc 9.4efg 36.7abc 78.7de
G7 DAB 496 2519.9 3439.6 2801.9 2737.5 3243.3 1093.2 2837.7 1750.7 4244.6 2774 1443.6 2626.0a 3.0efg 11.3cd 38.7a 77.5fg
G8 DAB 478 2299.4 3415.7 3004.4 2473.2 2673.1 763.3 2116.3 2238.1 3212.1 2521.6 1002.4 2338.1cd 3.5a 9.6efg 36.2a-d 78.3ef
G9 DAB 544 2450.5 4073 2410.7 2793.9 3203 1027.3 3316.4 2392 3198.9 2666.9 1308.4 2621.9a 3.2b-f 10.3c-f 34.0d 77.3g
G10 DAB 525 2290 3586.3 2817 2879.4 2595.3 1143.6 2616.9 2749.7 3036.5 3099.2 1317.4 2557.4ab 3.5ab 10.4c-f 34.3cd 77.2g
G11 DAB 545 1987.2 2636.1 3607.6 2537.6 1618.1 901.3 1851.1 2078 2148.1 2668.8 1454.9 2135.4ef 2.8g 13.7ab 34.5bcd 77.8fg
G12 DAB 540 1974 3301.9 2917.5 1769.9 2488 1044.4 2449.9 1814.9 2766.1 2611.8 1146 2207.7def 3.4a-d 10.6c-f 35.4bcd 80.6a
G13 DAB 532 2453.7 3271.1 2875.4 2818.1 2365.1 805.5 2165.6 2378.5 3265 2630.7 936.9 2360.5cd 3.2b-e 10.5c-f 35.5bcd 77.9efg
G14 DAB 481 2776.2 3993.9 3177.3 2360.2 2384 999.7 3350.2 2024.8 3220.1 2827.2 1222.3 2576.0a 3.4abc 9.1efg 35.8bcd 79.8bc
G15 Melkadima 2674.7 3639.4 2861 2571.2 2859.5 839.7 3138.1 2759.9 3482.2 2803.4 1514.6 2649.4a 3.6a 12.2bc 36.0bcd 80.4ab
G16 Red kideny 2237.5 3557 2422.1 1561.2 2083.2 579.9 2198.2 1767.9 3246.4 1949.6 1056.2 2059.9f 3.2b-f 10.5c-f 35.5bcd 77.5fg
EM 2317.3 3336 2883.1 2388 2455.7 906.9 2594 2239.4 3158.1 2645 1241
CV 14.8 5 18.3 16.4 6.5 13.1 16.5 21.2 5.4 14.75 15.3
Abbreviations: AN15 = Arsinegelle in 2015; AN17 = Arsinegelle in 2017; AN18 = Arsinegelle in 2018; AT16 = Alemtena 2016; AT17 = Alemtena in 2017; GF16 = Goffa in 2016; ME16 =
Meisso in 2016; MK16 = Melkassa in 2016; MK17 = Melkassa in 2017; MK18 = Melkassa in 2018; PW16 = Pawe in 2016; GM = Grand mean; EM =Environmental mean; SPP = seed per
pod; PPP = pod per plant; FD = Flowering date; MD = Maturity date
[122]
Stability and mega environment analysis
The application of the AMMI model for partitioning of GEI (Table 1) revealed the
first four terms of AMMI were significant. In the trial, the first and second
multiplicative axes terms explained 40.43 and 23.33% of GEI sum of squares,
respectively. According to Zobel et al, (1988) AMMI with the first two
multiplicative terms was the best predictive model. Thus, AMMI2 biplot was
generated using genotypic and environmental scores of the two PC to cross-
validate the interaction pattern of the 16 red bean genotypes in 11 environments
(Figure 2). The choice of AMMI2 instead of AMM1was made, information from
AMMI2 biplot may be more accurate as this model contains IPCA1 and IPCA2
information and therefore contains a greater pattern portion compared to the
AMMI1 biplot, which considered only IPCA1. In the biplot, G2, G3, G11, G1,
G7, and G4 placed furthest away from the biplot origin and expressed a highly
interactive behavior (positive or negative). While, G12, G13, and G8 were the
most stable genotypes since they placed relatively close to the biplot origin. Best
performing genotypes G14 and G15 were moderately stable genotypes. Among
the test environments, PW16, GF16, and AT16 were the main contributors to the
phenotypic stability of the varieties (Figure 2). In contrast, MK17, MS16, and
AN18 environments made a significant contribution to the GEI, since they were
positioned away from the origin of the AMMI biplot.
Figure 3. shows the ranking of 16 large red bean genotypes based on their mean
yield and stability performance using GGE biplot analysis. The estimation of yield
and stability of genotypes was done by using the average coordinates of the
environment (AEC) methods (Yan 2001, 2002). The average environment is
defined by the average values of PC1 and PC2 for all environments and it is
presented with a circle. A single arrowed lines passing through the small circle
and the biplot origin is called average environment coordinate abscissa with its
arrow pointing towards the increased yield. The genotypes are ranked along the
AEC abscissa and their stability is projected as a vertical line from the AEC
abscissa. A stable genotype will have a short projection from the AEC abscissa.
Thus, based on the graphical interpretation, genotypes G14, G15 and G9 with high
mean yield and stability performance can be considered ideal/promising
genotypes. Genotypes with high yield but low levels of stability were G6 and G7.
[123]
Figure 2. AMMI2 biplot (PC1 vs PC 2) for grain yield of 16 genotypes and 11 environments. Abbreviations for
environments and genotypes were as indicated In Table 1.
Figure 3. Genotypes mean yield performance and stability of the GGE biplot across environments.
Abbreviations for environments and genotypes were as indicated In Table 1.
[124]
Visualization of the ‘which-won-where’ pattern in the polygon view is helpful to
estimate the possible existence of different mega environments in the target
environments (Yan et al., 2000; Yan and Tinker, 2006). Figure 4 presents a
polygon view of sixteen large red bean genotypes tested in eleven environments.
The polygon is formed by connecting the genotypes further away from the biplot
origin such that all other genotypes are contained within the polygon. Genotypes,
G9, G7, G1, G11, G3, and G2 were vertexes of the polygon (Figure 4). These
genotypes are performed either the best or poorest in one or more locations since
they had the longest distance the origin of biplot. To each side of the polygon,
perpendicular line starting from the origin are drawn and extended beyond the
polygon so that the biplot divided into several sectors, and the different
environments were separated into different sectors. There were six sectors, and the
environments fall into four of them. This pattern suggests that the target
environment may consist of four different mega environments. Melkassa in 2016
and 2018 (MK16 and MK18) fell in one sector (mega-environment) suggesting
the repeatable performance of genotypes in this location with a winning genotype
of G2. Sector 2 (mega environment 2) consists of AT16, MS16, AN15, GE16, and
PW16, or five environments have good carrying capacity for genotype G9. The
mega environments 3 represents AN17, AT17, and MK17 are environments that
are suitable for genotype G7. Genotype G1 was the best genotype for mega
environment 4 which consists of only AN18.
Figure 4. Polygon view of the GGE biplot to show which genotype wins where. Abbreviations for environments and
genotypes were as indicated In Table 1.
[125]
Conclusion
The study result revealed that large red bean genotypes show substantial genetic
variation for grain yield and yield component traits. Four candidate genotypes had
comparable average grain yield performance with the standard check variety. The
magnitude of GEI effect was about three times more than the genotypic effect.
Generally from the trial, AMMI and GGE biplot revealed that the highest yielding
genotypes G14 (DAB 481) and G15 (Melkadima) were top-ranked in most of the
environments and found relatively stable genotypes across environments;
however, there was no single genotype that showed superior performance across
all environments. Among tested locations Melkassa and Arsinegllein 2017 have
been the best locations for expressing the genetic potential of the tested genotypes,
where these locations gave average grain yield of 3158 kg ha-1 and 3336 kg ha-1,
respectively. Since there is no significant difference between best-performing
genotypes and standard check variety, in the future by looking other quality traits
(seed coat color bleaching behavior, cooking and canning quality) the best
performing genotypes may consider either for verify to users or may use in future
breeding as a parental line.
Acknowledgments
We are grateful for long-lasting productive and impact-oriented partnerships with
the International Centre of Tropical Agriculture (CIAT), and the Pan African Bean
Research Alliance (PABRA) for technical and financial supports. We are also
grateful to Bill and Melinda Gates Foundation (BMGF) for the financial support
through Tropical Legume III (TLIII) project.
References
Central Statistical Agency (CSA). 2017. The Federal Democratic Republic of Ethiopian
Agricultural Sample survey, 2015/2016. Report on area and production of major crops.
Statistical bulletin 578, volume v. Addis Ababa, Ethiopia.
Gauch HG. 1992. Statistical analysis of regional trials. AMMI analysis of factorial design. 1st edn.
Elsevier, New York
Kang MS. 2004. Breeding: Genotype-by-environment interaction. In R.M. Goodman (ed.)
Encyclopedia of plant and crop science. Marcel - Dekker, New York. pp. 218-221.
Legesse, D. G. Kumssa, T. Assefa, M. Taha, J. Gobena, T. Alemaw, A. Abebe, Y. Mohhamed H.
Terefe. 2006. Production and Marketing of White Pea Beans in the Rift Valley, Ethiopia. A
Sub-Sector Analysis. National Bean Research Program of the Ethiopian Institute of
Agricultural Research.
Thillainathan M, Fernandez GCJ. 2001. SAS application of Tai’s stability analysis and AMMI
model genotype x environmental interaction (GEI) effects. Journal of Heredity 92(4): 367-
371.
[126]
Yan W. 2001. GGE biplot – a windows application for graphical analysis of multi-environment
trial data and other types of two-way data. Agronomy Journal., 93: 1111-1118.
Yan W. 2002. Singular value partitioning in biplot analysis of multi-environment trial data.
Agronomy Journal 94, 990-996.
Yan W, Hunt LA, Sheng Q, Szlavnics Z. 2000. Cultivar evaluation and mega-environment
investigation based on the GGE biplot. Crop Sci. 40: 597-605.
Yan W, Kang MS, Ma BL, Woods S, Cornelius PL. 2007. GGE biplot vs. AMMI analysis of
genotype-by-environment data. Crop Science, 47: 643-653.
Yan W, Tinker NA. 2006. Biplot analysis of multi-environment trial data: Principles and
applications. Can. J. Plant Sci. 86(3): 623-645.
Yayis R., G. Setegn Z. Habtamu. 2011. Genetic variability for drought resistance in small red
seeded common bean genotypes. African Crop Science Journal 19(4):303 – 311
Zobel RW, Wright MJ, Gauch HG. 1988. Statistical Analysis of a Yield trial. Agronomy journal
80: 388-393
[127]
[128]
Development of Biofortified Large Red Common
Beans for Major Growing Areas of Ethiopia
Berhanu Amsalu Fenta1, Dagmawit Tsegaye1, Abel Mogos1, Tigist Shiferaw1,
Behailu Tesfaye1, Tarekegn Argaw, 2, 1Semir Hashim, 1Girum Kifle,
Abush Tesfaye3, Ibsa Aleyi4, and Solomon Bekele5,
1
Ethiopian Institute of Agricultural Research, Melkassa Agricultural Research Centre, P.O.Box 436,
Ethiopia,2 Ethiopian Institute of Agricultural Research, (EIAR), P.O.Box 2003, 3Ethiopia, Ethiopian Institute
of Agricultural Research, 3Jimma Research Centre, P.O.Box, 4Haramaya University,
5
Ethiopia, Oromia Agricultural Research Institute, Bako Research Centre
Corresponding author: berhanufenta@gmail.com
Abstract
Common bean is one of the potential crops which can play a role in food and
nutrition security. A study has been conducted to select large red biofortified bean
variety with high yielding and adaptable to the wider or specific localities of the
country. The trial was conducted for three years and data of ten locations was
analyzed using a special analysis and best linear unbiased prediction (BLUP) of the
tested genotypes to compare the performance of the genotypes. Accordingly, the
tested genotypes exhibited average predicted grain yield up to 2.7 t/ha. But, some of
the test genotypes gave grain yield of 3-4 t/ha for specific sites and years. This shows
the need for a deep analysis of the existing data based on the correlated sites to look
for specific recommendations. In the future, nutritional analysis (Fe and Zn content)
will be analyzed, and based on the analysis result as well as other measured traits
result selection decision will be made to verify for the user or to use the selected
genotype as a crossing parent.
Keywords: Biofort, red beans, grain yield, special analysis
Introduction
Common bean (Phaseolus vulgaris L.) is one of the most commonly consumed
grain legumes worldwide, and it is the most important legume produced for
consumption and its’ commercial importance in Ethiopia. Common bean is an
important source of amino acids like lysine and tryptophan; minerals (iron, zinc,
and calcium), beneficial phytochemicals and antioxidants. For this reason it is an
important source of dietary protein supplementing those amino acids lacking in
cereal-based diets especially in Africa (Barampama and Simard 1993; Beebe, et
al., 2000).
Common bean is widely produced in Ethiopia in two seasons (main & short
rainfall seasons). It is produced by more than 10 million smallholder farmers
annually in more than half a million ha with annual grain production of
764,597metric tons (CSA, 2016). Although all colored beans production statistics
[129]
being reported as red beans; the most dominant colored bean production in the
country is red beans. Accordingly, the recent production of red bean in the country
was 850,000ton in 673,847area of land (CSA 2016).
Production of red bean in Ethiopia has been largely meant for local consumption.
In recent years, however, it has been recognized as one of the potential exportable
commodities regardless of its grain sizes. For instance, in the 2017/18 production
season about 111,140MT of colored beans mainly red beans were exported and
gained a value of 57.8 million USD from the export market. From overall
common bean export of the same season red beans counts almost 55% (78,067
ton) in volume and 47.4 % (37,526 million USD) export value (EPOSPEA 2018).
This shows the potential of red beans as exportable commodity and source of
income for growers.
Apart from its importance as a commercial crop, common bean plays an important
role as local consumption supplementing important nutrients. One of the
solutions/approaches is the selection of genotypes that can accumulate nutrients
(i.e. naturally bio-fortified) in their edible parts, at levels that can positively
impact human nutrition (Sperotto et al., 2013) would be indispensable. Vacuolar
Iron Transport (VIT) genes are thought to be involved in the storage of Fe in seeds
(Kim et al., 2006). Nutrition profile of red beans revealed that ½ cup of cooked
red beans to have 11% Fe and 6-8% Zn (https://beaninstitute.com/beans-101-2/).
Thus, selecting the most efficient varieties which can accumulate high Fe and
other minerals like Zn would be pertinent to avail for the consumers. In the past
studies genetic variability for nutrient (mainly Fe and Zn) accumulation in
common bean has been confirmed by Blair et al (2016). Thus, performance
evaluation/studies of lines derived from rich Fe and Zn population development
programs for their wider adaptability would be essentially important.
Although, more than 50 improved common bean varieties were released for
production in Ethiopia, the share of large red beans in the variety development
program was very limited. Owing to the growing demand for local consumption as
nutritious food and for the export market, identification of high yielding and
nutritionally rich common bean cultivars is pertinent. Thus, the objective of the
current study is to select nutritionally enhanced large red bean varieties combining
high grain yield, disease resistance, and adaptable to major common bean growing
areas of the country. Accordingly, this paper summarizes the efforts/studies
undertaken towards the development of large red bean germplasm which is
derived from the population development of Fe and Zn (biofortified lines)
originated from CIAT.
[130]
About 100 large red bean lines were obtained from CIAT and evaluation was
started at observation nursery in a single row observation plot at Melkassa. At this
stage genotypes were mainly evaluated for their adaptability and resistance to
major diseases. The selected genotypes were promoted to preliminary national
variety trials (PNVT) and each genotype was planted in four rows in larger plots
(4 m row length) using (5x5/6x6) triple lattice design replicated how many times
at 3-4 locations for one year. At this stage the treatments were planted in a larger
plot size (4 rows x 4 m length x 0.4m between the rows). At PNVT genotypes
were evaluated for yield and yield attributable traits as well as for disease
resistance. Based on the evaluation of this trial, 16-25 genotypes were selected and
advanced to the National variety trial (NVT) to be evaluated for overall
performances using designed experiments for two seasons.
Statistical analysis
Multi environmental trial data across locations and over years were conducted by
converting the trial into row and column to analyze using R- software package.
The statistical model used to analyze the trial data was a linear mixed model
which is given by the following equation.
where y(n x 1) is the vector of responses measured on n observations, is a

vector fixed effects, X(n x t) is a design matrix assigning fixed effects to each
observation, is a vector of random effects, Z (n x b) is a design matrix
assigning random factors to each observation, and e (n x 1) is the vector of residual
errors.
Analysis of Single Site Trial Data

Consider a crop variety selection trial consisting of nj plots in a rectangular array
cj column by rj rows (nj=cjrj), in which m genotypes are grown in a single trial j is
made, and let yj be the nj x 1 vector of data observations ordered as a row within
columns. The linear mixed model for the data vector yj can be written as
where the vectors represent fixed effects, random variety effects, and
random non-genetic (or peripheral, i.e., design and additional) effects,
respectively. Here of length nj = rjc, where, j=1, 2…t). All random effects are
assumed to follow a Gaussian distribution, with mean zero and each of the three
[131]
random effect vectors are assumed pairwise independent. Variance models used
for the random and residual effects are given by
Current methods of spatial analysis, popularized by the approach of Gilmour et al.

(1997), focus on modeling spatial variation on three scales; global, extraneous,
and local trends. Global variation is a form of non-stationary trend (i.e., has a
change in mean over the spatial dimension of the field) and is typically modeled as
a low-order polynomial fixed effect. Extraneous variation is also often aligned
with the dimensions of the field trial (i.e., rows or columns) and is induced by
management operations such as planting, harvesting, plot trimming and irrigation
(Stefanova et al., 2009). These terms are modeled as random effects in the mixed
model, and the covariance structure is captured through the variance of the
random effects .
Local trend is a smooth stationary process resulting from local soil or fertility
fluctuations in the field. This type of trend is modeled as a covariance structure
between neighboring plots at the residual level through the variance matrix R. The
model for the local trend in spatial analysis is introduced to address the lack of
independence between neighboring plots in the field. It is assumed that the trial is
arranged as a rectangular array of plots in the field as for the row-column design.
The variance matrix of the residuals R can then be formed as the Kronecker
product of the variance structure in the row and column directions.
The assumption of separability underpins the most popular models for small-scale
spatial trend in two dimensions, and appears to provide a reasonable approach for
a wide range of applications, (Martin, 1990; Cullis & Gleeson, 1991). A general
spatial model for the analysis of field trials in now considered, together with a
specific form for the matrix of correlated residuals.
The most common spatial variance model for the smooth trend in field trials is a
separable autoregressive process in the field row and column directions, where
and column are the nrxnr and ncxnc correlation matrices in the row and column
directions respectively (Gilmour et al., 1997). The variance matrix for the
residuals is
Analysis of Multi Environmental Trial Data

Suppose we have trials in which perennial crop variety grown (not all
varieties need to be grown in all trials. The jth trial consists of nj plots in a
[132]
rectangular array consisting of cj columns by rj rows (nj = rj cj). Let denote the
number of harvests for the trial and let be the total number of trial by
harvest combination ( . Let be the vector of
observation (for example Arabica coffee bean yield) for trial j , ordered as a row
within columns within harvests. The data combined across trials is denoted by
; this is an vector with
A linear mixed model for the data can be written, based on the model [2] as
Where is a vector of fixed effects with design matrix , is the

vector of genotype effect for individual trials by harvest combination
with associated design matrix , is a vector of other random effects
with associated design matrix and e is the vector of residuals.
The random-effects from a linear mixed model (equation [5]) are assumed to
follow a Normal distribution with zero mean vector and variance-covariance
matrix
Generally, in the previous studies revealed that the BLUP based a variance
component mixed model provided both a high precision and a high efficiency in
the prediction of the location-specific genotype effects with random locations than
ANOVA (Xiyuan Hu, 2015).

Model summary for the special analysis of the genetic variance result revealed that
out of 10 trials; two of 2017 for Melkassa and Alemtena, exhibited higher genetic
variance for yield and number of pod per plant traits (Table 1). However, for the
other two phenological traits (days to flowering and maturity) higher genetic
variance was observed for Alemtena 2017 and Jimma 2016. Moreover, these two
trials also showed a a higher mean grain yield than the other trials (locations)
except Alemtena 2017 with a lower value of error variance. This shows that these
testing sites are potential sites for discriminating test genotypes. The probable
reason for this could be due to relatively a better rainfall amount and distribution
in 2017 than the previous year (2016). This also suggests the importance of
climatic data of a given cropping season in recommending the best performing
genotype and for wider use throughout the production of different agro-ecologies
in the country.
[133]
Table 1. Trial means and estimated variance parameters for genetic effects and errors of model summary for the MET
data analysis from the biofort large red beans trials
Mean Genetic Variance Error Variance

Trials
DF DM NPP Yield DF DM NPP Yield DF DM NPP Yield
MK17 31.93 74.17 15.1 3.28 0.1 2.23 11.52 0.48 0.59 4.27 4.43 0.07
AT17 31.59 73.35 10.87 1.95 3.79 0.41 12.11 0.47 0.77 1.63 4.94 0.03
AN17 35.23 89.99 7.53 2.07 22.23 8.91 5.24 0.35 2.45 3.89 16.21 0.04
HU16 50.15 9.25 2.52 15.98 2.51 0.22 1.67 4.21 0.08
JM16 46.24 100.03 12.13 2.33 9.71 2.7 6.53 0.14 0.89 2.09 3.56 0.15
SR18 41.53 82.36 10.05 1.94 17.89 4.43 2.55 0.11 5.18 4.42 1.17 0.05
BK18 31.11 82.56 9.33 1.88 4.35 6.07 2.95 0.1 0.54 1.04 4.20 0.06
MK18 36.36 74.05 10.78 2.23 0.4 0.21 5.8 0.08 0.4 0.14 9.39 0.12
AN18 47.09 92.58 9.47 1.89 1.66 3 4.74 0.07 1.01 5.22 3.85 0.09
HS18 39.02 87.03 7.51 1.46 1.27 2.74 4.3 0.06 0.89 2.38 1.80 0.15
Note: MK= Melkassa, AT= Alemtena, AN= Arsingele, HU= Haremaya University, JM= Jimma, SR= Sirinka, BK= Bako,
HS= Harosebu, t/ha= tons per hectar
Table 2 Average predicted value of grain (PLUPs) yield (t/ha) of biofortified large red beans
Trials
Genotype AN17 AN18 AT17 BK18 HS18 HU16 JM16 MK17 MK18 SR18 Grand Total
Zoasho 2.94 2.48 3.82 2.68 1.56 3.63 3.14 4.82 2.99 2.86 3.09
Roba-1 1.64 2.33 2.09 2.11 2.02 2.48 2.93 4.50 2.79 2.20 2.51
NUA 355 2.42 2.14 3.62 2.49 1.33 3.02 2.68 4.89 2.71 2.47 2.78
NUA 648 3.15 2.40 2.05 1.98 1.73 3.53 2.96 4.33 2.51 2.16 2.68
NUA 350 1.94 2.03 2.03 2.13 1.60 2.55 2.49 3.17 2.41 2.09 2.24
NUA 536 2.08 2.03 1.92 1.98 1.55 2.79 2.48 3.33 2.38 2.06 2.26
NUA 636 2.76 2.12 2.65 1.90 1.45 3.08 2.58 2.70 2.37 1.95 2.36
NUA 577 2.34 2.10 2.05 1.73 1.58 2.76 2.56 3.35 2.36 2.18 2.30
NUA 645 2.18 1.98 2.07 2.27 1.43 2.79 2.42 3.72 2.35 1.97 2.32
NABE 3 1.00 1.76 2.04 1.96 1.49 2.11 2.15 2.84 2.34 1.88 1.96
NUA 656 1.10 1.65 2.52 2.13 1.24 2.24 2.01 2.43 2.32 2.11 1.97
NUA 497 2.38 1.98 2.54 1.83 1.35 2.85 2.41 3.45 2.31 2.24 2.33
NUA 591 1.59 1.95 1.53 1.71 1.64 2.58 2.38 3.02 2.30 1.81 2.05
NUA 353 2.67 2.10 1.95 1.78 1.56 2.67 2.55 3.19 2.29 1.98 2.27
NUA 582 2.08 1.88 2.44 1.89 1.32 2.48 2.28 3.37 2.27 1.56 2.16
NUA 590 1.76 1.91 1.52 1.83 1.61 2.04 2.31 2.46 2.23 1.92 1.96
NUA 640 3.28 2.24 1.50 1.68 1.64 2.79 2.70 3.68 2.20 1.91 2.36
NUA 561 2.16 1.82 2.04 2.05 1.26 2.51 2.19 3.49 2.16 2.11 2.18
NUA 597 2.35 2.05 1.00 1.71 1.70 2.59 2.46 2.85 2.13 1.47 2.03
NUA 615 1.62 1.70 1.90 1.67 1.29 2.20 2.04 3.37 2.12 1.81 1.97
NUA 654 1.82 1.74 1.66 1.84 1.33 1.94 2.08 2.96 2.08 1.71 1.92
NUA 605 2.29 1.67 2.36 2.05 0.99 2.58 1.99 3.09 2.06 1.89 2.10
NUA 619 2.19 1.85 1.00 1.37 1.49 2.12 2.18 3.21 1.97 1.92 1.93
NUA 637 1.64 1.67 1.32 1.64 1.32 2.23 1.96 2.62 1.96 1.58 1.79
NUA 575 1.87 1.50 1.67 1.84 0.98 2.14 1.75 3.36 1.88 1.72 1.87
NUA 347 1.39 1.49 1.17 1.40 1.20 1.81 1.73 2.60 1.83 1.85 1.65
Grand
Total 2.10 1.95 2.02 1.91 1.45 2.56 2.36 3.34 2.28 1.98 2.19
Best linear unbiased prediction (BLUP) is a standard method for estimating

random effects of a mixed model. BLUP is the method where the information of
all relatives is used the best and the fixed effects are accounted for unbiased and
[134]
the method accounts for selection over time. Its advantage is to handle unbalanced
designs and uses information for all relatives measured to improve estimates.
Accordingly, the special analysis/BLUP of grain yield for 10 trials carried out at
eight locations for three years revealed that out of 26 tested genotypes, more than
50% of them exhibited average grain yield of greater than 2 t/ha. However, the
predicted mean grain yield revealed that three genotypes with a higher mean yield
greater than 2.7 t/ha; one being the recent check-, Zoasho, followed by the other
two comparable to Zoasho are NUA 335 and NUA 648 (Table 2). BLUP analysis
further confirmed that trials of Melkassa in 2017, and that of Haremaya in 2016
gave high grain yield, suggesting that these sites are the best testing locations to
discriminate among bean genotypes and representing the most suitable agro-
ecologies to bean crop production in general.
Investigate the genetic correlation pattern using a heatmap. Accordingly, several

pairs of trials have shown positively and highly correlated (r > 0.5), indicating that
the testing sites exhibited fairly similar growing conditions to discriminate against
the tested genotypes (Fig. 1). Likewise, this correlation was also ascertained using
Dendrogram analysis (Fig. 2). The dendrogram analysis revealed that the trials
were clustered into three main clusters (fig. 2). Such a cluster would be pertinent
for site-specific recommendations. For instance, if we look at the first cluster
(AN17, HU16, and MK 17, the two best performing genotypes consistently at
these three locations are genotypes NUA 648 and Zoasho, yielding from 3 to 4.8
t/ha.
Figure 1: Heatmap of biofort large red beans conducted from 2016 to 2018
[135]
Figure 2: Dendrogram of grain yield of biofort large red bean trial
Conclusion
According to the result of this study, generally, biofortified large red bean
genotypes have been shown substantial genetic variation for grain yield. Although
the recently released large red bean variety, Zoasho exhibited the highest grain
yield than the other large red bean varieties, potential genotypes which have
biofortified background have been identified. The whole trial samples have been
sent abroad to analyze Fe and Zn content and genotypes showing better national
density with reasonable grain yield will be verified for wider use in nutritional
insecure areas. Further, the selected high Fe and Zn containing varieties can also
be used as trait donors in the crossing programs in the national bean improvement.
Acknowledgments
We are grateful for financial and germplasm support of the International Centre of
Tropical Agricultural (CIAT)-Uganda, and / Pan Africa Bean Research Alliance
(PABRA) and Tropical legume III (TL III) project.
[136]
References
Barampama, Z and Simard RE. 1993. Nutrient composition, protein quality, and
antinutritional factors of some varieties of dry beans (Phaseolus vulgaris)
grown in Burundi. Food Chem. 1993, 47, 159–167
Beebe, S, Gonzalez AV, Rengifo J. 2000. Research on trace minerals in the
common bean. FoodNutr. Bull. 2000, 21, 387–39
Blair MW, Wu X, Bhandari D, and Astudillo C. 2016. Genetic Dissection of ICP-
Detected Nutrient Accumulation in the Whole Seed of Common Bean
(Phaseolus vulgaris L.). Front. Plant Sci. 7:219. doi: 10.3389/fpls.2016.00219
Cullis BR and AC Gleeson. 1991. Spatial Analysis of Field Experiments-An
Extension to Two Dimensions. Biometrics, Vol. 47, No. 4 (Dec., 1991), pp.
1449-1460
CSA. 2016. Central Statistics Authority Annual Report for the Year 2012, Addis
Ababa, Ethiopia.
Ethiopian Pulses Oilseeds and Spices Processors Exporters Association
(EPOSPEA). 2018. The 8th International Conference on Pulses, Oil Seeds and
Spices Workshop Bulletin, November 23-24, 2018 Sheraton Addis Hotel,
Addis Ababa, Ethiopia.
Gilmour A, Cullis BB and Verbyla AP. 1997. Accounting for Natural and
Extraneous Variation in the Analysis of Field Experiments. Journal of
Agricultural, Biological and Experimental Statistics: Vol 2, No:3 269-293
Kim SA, Punshon P, Lanzirotti A, Li L, José M, Alonso JR, Ecker J, Kaplan M,
Guerinot Lou. 2006. Localization of Iron in Arabidopsis Seed Requires the
Vacuolar Membrane Transporter VIT1. Science 24: 1295-1298.
Sperotto RA, Ricachenevsky FK, de V, Waldow A, Müller ALH and Dressler VL,
Fett JP. 2013. Rice grain Fe, Mn and Zn accumulation: How important are
flag leaves and seed number? Plant Soil Environ. Vol. 59: 6: 262–266.
Stefanova, KT, AB Smith and Cullis BR. 2009. Enhanced diagnostics for the
spatial analysis of field trials. JABES 14, 392
https://doi.org/10.1198/jabes.2009.07098
Xiyuan Hu. 2015. A comprehensive comparison between ANOVA and BLUP to
evaluate location-specific genotype effects for rape cultivar trials with random
locations. Field Crops Research 179:144-149.
[137]
[138]
Evaluation of Cowpea Genotypes for Grain Yield
Using Additive Main Effects and Multiplicative
Interaction (AMMI) Model
Berhanu Amsalu Fenta1, Dagmawit Tsegaye1, Tesfaye Walle2, Abel Moges Firew1,
Firew Mekbib3, Tigist Shiferaw1, Behailu Tesfaye1, Ibsa Aleyi3, and Tekle Yoseph4
1
EIAR, Melkassa Research Centre, P.O.Box 436, Ethiopia; 2Department of Plant Sciences, Wolkite
University, Wolkite, Ethiopia; 3School of Plant Sciences, Haramaya University, Dire Dawa, Ethiopia
4
Jinka Agricultural Research center, Jinka, Ethiopia; Corresponding author: berhanufenta@gmail.com
Abstract
Twenty three cowpea genotypes collected from different growing areas of Ethiopia
and two released varieties were cultivated in three locations in 2017 and five
locations in the 2018 cropping season. The experiment was conducted using 5 * 5
triple lattice across locations. Only 3 genotypes namely, NLLP-CPC-07-54 (G13),
NLLP-CPC-07-145-21 (G21), and NLLP-CPC-07-10 (G1) gives higher yield than
the best check variety Bole (G24). Data on grain yield was subjected to Additive
main effects and multiplicative interaction (AMMI) model and showed a highly
significant difference for Genotype x Environment interaction (GEI) and indicating
the possibility of selection for stable genotypes. AMMI analysis indicated seven
interaction principal components analysis (IPCAs) were significant (P<0.01).
Among them, the first and second multiplicative axis terms explained 33.22 % and
27.11 % of GEI sum of squares, respectively. According to AMMI biplot genotypes
G11 (NLLP-CPC-103-B), G13 (NLLP-CPC-07-145-21), G14 (NLLP-CPC-07-28)
and G21 (NLLP-CPC-07-54) were as the stable genotypes and can be used as
superior genotypes in cowpea breeding.
Keywords: Grain yield; Additive main effects and multiplicative interaction

(AMMI); Interaction Principal Components analysis (IPCA);
Genotype x Environment interaction (GEI)
Introduction
The cowpea (Vigna unguiculata L.) is an important annual herbaceous legume
crop widely grown in sub-Saharan Africa, South America and Asia. In Ethiopia,
cowpea is grown in southern and eastern Tigray, north and south Wollo, Hararghe,
Gambella, and some parts of the southern region of the country. Cowpea plays a
critical role in the lives of millions of people in the developing world, providing
them a major source of dietary protein that nutritionally complements low protein
cereal and tuber crop staples. The nutritional profile of cowpea grain is similar to
that of other pulses with relatively low-fat content and total protein content that is
two- to fourfold higher than cereal and tuber crops. According to Thulin (1989), in
Ethiopia, cowpea is cultivated primarily for its edible seeds and the leaves that are
sometimes used as human food in the form of cooked leafy vegetables.
[139]
Even though Ethiopia is considered a secondary center of diversity (Vavilove,
1951) for cowpea, production is limited only in some parts of the country and
productivity is very low because of a few numbers of improved varieties. In order
to develop and provide the best performing cultivars to grow in different
environments, variability among genotypes is required. Variability can be created
by the collection of all available germplasms from all growing areas of the crop.
Based on it the national lowland pulse improvement program made collection
from all cowpea growing areas of the country.
Cowpea yield is dependent on genetic, environmental, and interaction which

reduces the association between genotypic and phenotypic components. Genotype
x Environment interaction (GEI) can be defined as the differential response of
varying genotypes under change(s) in the environment (Mather and Caligari,
1976). Multi-environment yield trials are essential in the estimation of GE
interaction and identification of superior and stable genotypes in the final selection
cycles (Kaya et al., 2006; Mitrovic et al., 2012). Genotypes that can adjust its
phenotypic state in response to environmental fluctuations in such a way that it
gives maximum stable economic return can be termed as well buffered or stable
(Allard and Bardshow, 1964). Additive Main effect and Multiplicative Interaction
(AMMI) model is a hybrid analysis since it incorporates classical additive effect
and multiplicative components into integrated least square analysis and thus more
effective in the selection of stable genotypes (Crossa et al., 1991; Yan and Hunt,
2001). The objective of this study, therefore, is to conduct GEI and yield stability
analysis of cowpea collections using the AMMI model.

Cowpea genotypes used for this experiment were local collection from major
cowpea growing areas of Ethiopia. Best performing 25 cowpea genotypes
including 2 check varieties that were advanced from preliminary variety trials
were grown at three locations (Babile, Melkassa and Meiso) in 2017 and five
locations (Babile, Melkassa, Jinka, Kobo and Meiso) in 2018 during the main
season. Each year at each location was considered as a separate environment,
making eight test environments. The descriptions of 25 genotypes are given in
Tables 1.
[140]
Table 1. Code name and source of tested genotypes
Code Genotypes Source

G1 NLLP-CPC-07-10 Collection
G2 CP-EXTERETIS Collection
G4 Dass 001 Collection
G5 ACC-215-821 Collection
G11 NLLP-CPC-103-B Collection
G13 NLLP-CPC-07-145-21 Collection
G19 NLLP-CPC-07-03-B Collection
G24 Bole Melkassa research center
G25 Kenketi Melkassa research center
The treatments were laid out in 5 x 5 triple lattice across environments. Each plot
consists of four rows and 4 m long with a spacing of 0.4 m between rows. For data
analysis, grain yield measured from a net plot size of 2.4 m2 was converted into kg
ha-1 at 12.5 % grain moisture content.
Data analysis
The data were subject to statistical analysis using Genstat 17 computer software.
A combined analysis of variance was done using the AMMI model (Gauch, 1992).
Where: = the mean yield of genotype g in environment e, = the grand mean,

= the deviation of the genotype mean from the grand mean, = the deviation of
the environment mean from the grand mean, = the singular value for the IPCA n,
= the number of PCA axis retained in the model, = the PCA score of a
genotype for PCA axis n, = the environmental PCA score for PCA axis n, =
the AMMI residual and = the residuals.
[141]
Mean grain yield
The genotypes grain yield across environments ranged from 861 kg/ha for NLLP-
CPC-07-143 to 1970 kg/ha for NLLP-CPC-07-145-21 (Table 2). Among tested
genotypes 3 genotypes namely, NLLP-CPC-07-54 (G13), NLLP-CPC-07-145-21
(G21), and NLLP-CPC-07-10 (G1) gives higher yield than the best check variety
Bole (G24). G9 (NLLP-CPC-07-143) was a poor-performing genotype with a
mean yield of 861 kg/ha. Across the specific locations, no genotype showed a
consistent yield advantage. Among the testing environments, the general
performances of the genotypes for grain yield were high at Meiso 2018 (MS18) as
compared to the other seven environments with a mean yield of 2956 kg/ha and
followed by Melkassa 2018 (MK18) (1877 /ha) and Meiso 2017 (MS17) (1665
kg/ha) (Table 2).
Table 2. Mean grain yield (kg/ha) of 25 cowpea genotypes tested at eight environments
Environments
Genotype
BA17 BA18 JN18 KB18 MK17 MK18 MS17 MS18 G.M
NLLP-CPC-07-10 970 784 1595 1895 1295 2140 1622 3890 1774
CP-EXTERETIS 241 696 688 1540 649 2028 1379 1951 1147
NLLP-CPC-07-169 816 748 1863 1624 1051 2107 1883 3776 1734
Dass 001 559 1436 1724 1628 968 2194 1397 3039 1618
ACC-215-821 705 1222 1712 1834 967 1722 2000 2645 1601
ACC-216-747 572 1266 1694 896 703 1702 1322 2433 1324
NLLP-CPC-07-139 542 1075 1547 2064 992 2370 780 3332 1588
NLLP-CPC-07-140 1025 1149 2734 506 1616 1422 1344 2772 1571
NLLP-CPC-07-143 499 553 758 979 408 1136 925 1626 861
ACC-215-762 681 322 647 1774 1164 1796 1990 2195 1321
NLLP-CPC-103-B 959 1301 1254 1708 1037 2318 1619 3276 1684
NLLP-CPC-07-156 247 1266 1563 1247 410 1745 168 2798 1181
NLLP-CPC-07-145-21 1173 1657 1890 2063 1542 2203 2264 2968 1970
NLLP-CPC-07-28 483 1219 1356 1427 1214 1754 1896 3187 1567
NLLP-CPC-07-36 407 683 1471 1240 1314 1777 1671 2629 1399
NLLP-CPC-07-166 365 1026 917 1556 1242 1656 1673 2626 1382
NLLP-CPC-07-166 831 858 1578 931 1040 614 1172 4359 1423
NLLP-CPC-07-167 952 1641 1267 1186 908 2083 1270 3148 1557
NLLP-CPC-07-03-B 678 278 835 2063 737 1676 1632 3043 1368
NLLP-CPC-07-157 975 694 1012 1534 1163 1984 2403 4011 1722
NLLP-CPC-07-54 1075 1309 1750 1941 1647 1862 2321 2883 1848
NLLP-CPC-07-57 367 676 1508 1557 1041 2601 1671 2259 1460
ACC-211-490 973 1534 1304 726 689 1507 2178 2700 1451
Bole 588 752 788 1064 1225 2664 3099 3725 1738
Kenketi 583 769 1126 1650 1118 1871 1946 2631 1462
EM 691 997 1383 1465 1046 1877 1665 2956
CV 9.22 15.6 7.27 7.18 11.28 9.12 7.39 5.19
LSD at 5% 242.2 257.6 166.6 124.4 210.0 283.9 203.9 254.8
Abbreviations: BA17 = Babile in 2017; BA18 = Babile in 2018; JN18 = Jinka in 2018; KB18 = Kobo in 2018; MK17 =
Melkassa in 2017; MK18 = Melkassa in 2018; MS17 = Meiso in 2017; MS18 = Meiso in 2018; GM = Grand mean; EM
=Environmental mean;
[142]
AMMI analysis
AMMI analysis of variance (ANOVA) for grain yield indicated that environment,
genotype and GEI effects accounted for 66.37 %, 8.91 % and 23.16 %
respectively, of the total variation. Only 1.48 % of the total variation was
remained unexplained (as error) (Table 3). A large yield variation explained by the
environment indicates that the locations were diverse, with a large difference
among environmental means causing most of the variation in grain yield. Hence,
GEI exerted about three times larger effect than the genotype which complicates
the selection of superior and adaptable genotype.
The presence of GEI demonstrated by the AMMI biplot, when the interaction
partitioned into seven Interaction Principal Component Axes (IPCA) as they were
significant in postdictive analysis. In the study, the first and second multiplicative
axis terms explained 33.22 % and 27.11 % of GEI sum of squares, respectively
(Table 2), which recommended that the most accurate model for AMMI can be
predicted using the first two IPCAs (Gauch and Zobel, 1996), Zobel, 1996).
Table 3. Combined analysis of variance for grain yield (kg/ha) of 25 cowpea genotypes at 8 environments.
Total
G×E
Source of Degrees of Sum of Variation
Mean square explained
variation freedom Square explained
(%)
(%)
Total 599 384841137 642473
Treatments 199 378834350 1903690
Genotypes 24 34305174 1429382*** 8.91
Environments 7 255414478 36487783*** 66.37
Block 16 326479 20405*** 0.08
Interactions 168 89114699 530445*** 23.16
IPCA 1 30 29601103 986703*** 33.22
IPCA 2 28 24157434 862766*** 27.11
IPCA 3 26 16670077 641157*** 18.71
IPCA 4 24 8305075 346045*** 9.32
IPCA 5 22 6309918 286814*** 7.08
IPCA 6 20 2123068 106153*** 2.38
IPCA 7 18 1948023 108224*** 2.19
Error 384 5680307 14792 1.48
In the AMMI1 biplot model, the abscissa represents the main effects and its
ordinate represents IPC1 scores. Genotypes and environments on the right side of
the midpoint (abscissa) of the perpendicular line have higher yields than those on
the left side. As a result, 13 genotypes (G18, G14, G8, G7, G5, G4, G11, G20, G3,
G24, G1, G21, and G13) were generally high yielding. In contrast, the remaining
12 genotypes (G9, G2, G12, G10, G6, G19, G16, G15, G17, G23, G22 and G25)
were generally low yielding genotypes. Environments MS17, MK18, and MS18
were on the right-hand side of the midpoint of the main effect axis, seemed to be
potential environments, while KB18, MK17, BA17, BA18, and JN18 were poor
environments. Genotypes with IPC1 scores close to zero expressed general
adaptation whereas the larger scores depict more specific adaptation to
[143]
environments with IPC1 scores of the same sign (Ebdon and Gauch, 2002).
Accordingly, genotypes G14, G5, G11, G21, and G13 with mean yields greater
than the overall mean and low IPC1 scores had a combination of high yield and
stability performance. Among environments, the poor performing environments
MK17 and BA17 was the most stable environment due to low IPC1 score, while
the remaining six environments had the highest interaction with genotypes
because they had higher IPCA scores.
Figure 1. AMMI1 biplot showing the mean (main effect) vs. stability (IPCA1) view of both genotype and environment on
grain yield.
AMMI2 biplot was generated using genotypic and environmental scores of the
first two IPCs to cross-validate the interaction pattern of the 25 cowpea genotypes
within eight environments (Figure 2). The distance from the biplot origin (0, 0) is
indicative of the amount of interaction that was exhibited by genotypes over
environments or environments over genotypes (Voltas et al., 2002). Genotypes or
environments located near the biplot origin were less responsive than vertex
genotypes or environments. Thus, G11, G13, G15 G14 and G21 were
demonstrated low interactive effect over the environment, thus these genotypes
show lower fluctuation to the changes in the growing environment. G17, G20,
G24, G10, G22, G12, and G8 scattered away from the origin in the biplot
indicating that these genotypes were more sensitive to environmental effects.
Based on their proximity to the origin, MK17 and BA17 relatively exhibited lesser
genotype discriminating ability and proved to be more representative of the
[144]
average environment than the remaining environments. While the remaining six
environments (MK18, KB18, BA18, JN18, MS17, and MS18) indicated the
longest distance between their markers and biplot origin, demonstrated higher
genotype discriminating ability and found to be less representative of the average
environment (Figure 2).
Figure 2. AMMI2 biplot of the first interaction principal component axis (IPCA1) vs. the second principal component axis
(IPCA2) for grain yield.
Conclusion
The present study revealed environment was the largest contributor to the
observed variation and followed by GEI and genotype. AMMI model is an
effective tool to understand complex GEI in a multi-environment trial of cowpea.
The results obtained from AMMI analysis indicated that genotypes G11 (NLLP-
CPC-103-B), G13 (NLLP-CPC-07-145-21), G14 (NLLP-CPC-07-28) and G21
(NLLP-CPC-07-54) were as the stable genotypes and has been selected for
verification trial. Accordingly, in the future, the released varieties will be
multiplied and disseminated to the end-users.
[145]
References
Crossa J, Fox PN, Pfeffer WH, Rajaran S, Gauch HG. 1991. AMMI adjustment for statistical
analysis of an international wheat trial. Theor Appl Genet 81:27-37.
Ebdon JS, Gauch HG. 2002. Additive main effect and multiplicative interaction analysis of
national turfgrass performance trials: I. Interpretation of genotype × environment
interaction. Crop Sci. 42:489–496.
Gauch HG. 1992. Statistical analysis of regional trials. AMMI analysis of factorial design. 1 st eds.
Elsevier, New York
Gauch HG and Zobel RW. 1996. Optimal replication in selection experiments. Crop Sci. 36: 838–
843.
Kaya Y, Akcura M and Taner S. 2006. GGE-biplot analysis of multi-environmentt yield trials in
bread wheat. Turk. J. Agric. For. 30:325337.
Mather, K and Caligari PDS. 1976. Genotype x environment interactions IV. The effect of the
background genotype. Heredity 36(1), 41-48.
Mitrovic B, Stanisavljevi D, Treski S, Stojakovic M, Ivanovic M, Bekavac G, Rajkovic M. 2012.
Evaluation of experimental Maize hybrids tested in Multi-location trials using AMMI and
GGE biplot analysis. Turk. J. Field Crops 17(1):35-40.
Thulin M. 1989. Fabaceae (Leguminosae). In: Hedberg I. and Edwards S. (Eds). Flora of Ethiopia.
Pittosporaceae to Araliaceae. Addis Ababa and Asmara, Ethiopia, Uppsala, Sweden 3:49-
174.
Vavilove, NI. 1951. The origin, variation Immunity, and breeding of cultivated plants chronica
bot.13 (1-6):1-36.
Voltas J, Van EF, Igartua E, García del Moral LF, Molina-Cano JL, Romagosa I. 2002. Genotype
by environment interaction and adaptation in barley breeding: Basic concepts and methods
of analysis. In Slafer GA, J L. Molina-Cano, R. Savin, J L. Araus, Romagosa (eds.) Barley
Science: Recent Advances from Molecular Biology to Agronomy of Yield and Quality. The
Harworth Press Inc., New York, pp. 205241.
Yan W, Hunt LA. 2001. Interpretation of genotype x environment interaction for winter wheat
yield in Ontario. Crop Sci 41:19-25.
[146]
Evaluation of Mung Bean Genotypes
for Lowland Areas of Ethiopia
Dagmawit Tsegaye1, Berhanu Amsalu Fenta1, Abel Mogos1, Tigist Shiferaw1, Kassaye
Negash1, Kidane Tumssa1, Behailu Tesfaye1, Semir Hashim1, Girum Kifle1,
Ziru Yimer1, and Ibsa Aliyi
1
EIAR, Melkassa Research Centre, P.O.Box 436, Ethiopia; 1EARI, Pawi Research Centre, Pawe, Ethiopia;
2
Haremaya University, Haremaya, Ethiopia; Corresponding author: damy.we@gmail.com
Abstract
This study was carried out to determine the response of mung bean (Vigina radiata
L.) genotypes to different environments. Nine mung bean genotypes were used as
plant material. This research was conducted in Babile, Melkassa, Miesso, Pawe,
Sekota and Srilanka . In all locations, the experimental design used was a 3*3 lattice
with 3 replications. The GGE biplot analysis used for the analysis of environmental
correlation, stable genotype across locations and unique environments for identifying
better-adapted genotypes in the low land areas. The genotype MB 6173-B-33 Black
Bean had the highest stability level, whereas the check MH-97-6 also successfully
grown. The genotype Black Bean and MB 6173-B-33 showed the highest yield
potential. To conclude, the ideal yield would be obtained if the requirements of the
genotypes are fulfilled by desired environmental conditions.
Keywords: Mung bean; Genotype; Environment; Location; Stability; Yield
Introduction
Mung bean [Vigna radiate (L.) Wilczek] is a warm-season annual grain legume.
The optimum temperature range for good production is 27-30°C (Imrie, 1998).
Mung bean is a quick crop, requiring 75–90 days to mature. It is a useful crop in
drier areas and has a good potential for crop rotation and relay cropping with
cereals using residual moisture. Smallholder farmers in drier marginal
environments in Ethiopia grow mung bean. As compared to other pulse crops, its
production in Ethiopia is very negligible. However, for resource-poor farmers in
drier marginal environments, it has been an important grain legume.
Mung bean has an important place as it contains more digestible protein than the
other pulses along with a reasonably good yield. Commonly, the dry seed of mung
bean can be prepared in different forms like, “nifro” (boiled beans), mixed with
sorghum or maize, flour/ split-grain can be used to prepare stew (wat), whole seed
can be used to prepare “sambusa” or soup (Berhanu et, al 2016). Mung bean is
produced in both Meher and Belg seasons mainly produced in the Belg season.
Recently mung bean has been included by CSA sample study. Accordingly, based
on the reports of the 2017/18 cropping season the production of mung bean covers
41,633.20 hectares with 514,227.41 quintals were produced. The national average
[147]
productivity is 12.35 qt/ha (CSA 2017/18). Being a source of cash for smallholder
farmers, mung bean is one of the main export commodities for the country among
the lowland pulses (Berhanu et, al 2016).
In a plant Breeding program, potential genotypes are usually evaluated in different

environments to select desirable ones. For stabilizing yield, it is necessary to
identify the stable genotypes suitable for a wide range of environments. To
identify such genotypes, genotype by environment interactions is of major concern
for a breeder because such interactions confound the selection of superior cultivars
by altering their relative productiveness in different environments (Eagles and
Frey, 1977). Varietal stability in grain yield for a wide range of environments is
one of the most desired properties of a genotype to fit the crop under available
cropping patterns. So adaptability and stability are the main considerations in
formulating an efficient breeding program. The objectives of this study were to
evaluate and to quantify the magnitude of the GEI and describe the which-won-
where pattern using GGE biplot. However, due to the growing demand of mung
bean variety for local consumption and export market, this study is important to
generating alterative and also high yielding and drought tolerant mung bean for
lowland areas of Ethiopia

Mung bean genotypes were introduced from India. The selection and evaluation
were done starting from the observation nursery stage for their adaptiveness and
other merits. Superior genotypes were selected and advanced to Pre National
Variety Trial (PNVT) in a plot size of four-meter lengths of four rows in designed
experiments with three replications at three locations. Yield, phonological data
and disease and insect pest incidence were recorded. Lattice design with three
replications was used with a plot size of 4 rows x 4 m length x 0.4m between rows
for mung bean Yield, yield components, disease severity and all other relevant
agronomic characteristics were critically evaluated. At National Variety Trial
(NVT) stage nine mung bean genotypes due to the availability of information we
were obliged to experiment for additional years. A simple lattice design with three
replications was used for both NVT trials. Except for the number of rows are
increased from 4 rows to 6 rows all measurable traits are similar to PNVT stage.
Varieties with outstanding performance identified from the multi-environment
trials, in terms of yield and quality compared to the standard checks will be
verified both on-station and on-farm on a larger plot (10 m x 10 m) along with the
recently released standard check. The candidate varieties were thoroughly
assessed by farmers for their suitability to socio-economic circumstances and
contribution to their livelihoods. The experiment was conducted at Babile,
Melkassa, Mieos, Pawe, Sekota and Sirinka.
[148]
Genetic Materials
The nine experimental genotypes were evaluated and advanced for high Yield and
early maturing. These mung bean genotypes were selected during the early stages
of breeding based on there highly heritable traits. The two genotypes (Showarobit
and MH-97-6) are the Local check and standard check commercial varieties
respectively.
Table 2. List of the nine mung bean genotypes evaluated across years and environments
Genotype code Genotype name

1 MB 6148-05-12
2 Mong Hole
3 Black Bean
4 Asha
5 MH-85-1
6 MH-BR-1
7 MB 6173-B-33
8 Showarobit
9 MH-97-6
Statistical analysis Methods

GE analysis of variance (ANOVA) was done on grain yield and related agronomic
traits using Genstat software version13 (Genstat, 2010). For individual location
results, the ANOVA is done using SAS software version SAS 9.2. The suitable F-
test for a mixed model that involves fixed genotypes and years, and location
comparison biplot for classifying the most discriminating and representative
locations were produced using the suitable SVP methods (Yan and Tinker, 2006).
The statement used for the combined trials is that the effect of random interactions
sums to zero across each level of a fixed factor (Moore and Dixon 2014). The
mean squares for genotypes, environments, and GE were tested compared to the
pooled error mean square, while locations, years, and locations x years were tested
compared to the mean square of replications within locations and years (McIntosh,
1983).

Based on the analysis of multi-location data, environments SK18 followed by
SK16, MS18, MS16 and MS14 and SR18 were the highest yielding locations.
Genotypes significantly varied for all the parameters. The study was shown that
two genotypes (3 and 7) were found to have higher mean grain yield standard
local check MH-97-6 and local check Shewa robit in combined environments but
the genotype 3 late in flowering and maturity date these types of genotypes are not
important for drought-stressed environments. The performance on genotype 7 for
listed traits (Yield, days to flowering and days to maturity indicate a positive
relationship between grain yields with days to 50% flowering and 90 % of
maturity. These types of genotype are very important in moisture stressed mung
[149]
bean growing areas. Hence, there is an opportunity to find genotypes, among the
tested entries, that perform better than the existing varieties in moisture stressed
areas and/or to use it as parents for hybridization programs.
The GE variation in this study was due to expectable factors (location) and
changeable factors (years). In this study, the GE could be attributed to diverse
factors like soil types, rainfall patterns, and temperatures. The economical option
is to develop mung bean varieties adapted to the target environments. However,
the locations have no clearly defined boundaries and farmers have of influencing
each other in the choice of variety that is grown, the development of varieties with
broad adaptation is strongly supported, rather than location-specific varieties
(Chhetri et al., 2012). Based on general performances, field observation and
farmers’ preferences one promising variety, MB 6173-B-33 were selected for
variety verification trial and to be evaluated by the National Variety Release
Committee (NVRC) to satisfy the emerging international and local demand of
mung bean and to have an alternative variety for previously released variety.
[150]
Table 3. Mean yield performance (qt ha-1) of nine mung bean lines evaluated at 10 environments
Environments
Genotype Mean
BA18 MK14 MK18 MS14 MS16 MS18 PW14 PW16 SK16 SK18 SR18
1 4.07 3.77 7.66 11.26 10.93 12.11 8.37 10.51 13.96 18.92 8.59 10.01
2 5.17 6.89 7.33 10.05 10.05 14.48 4.87 4.87 17.92 16.4 8.57 9.69
3 6.05 15.3 7.36 13.57 14.23 11.9 9.25 9.87 16.25 9.3 18.93 12.0
4 4.86 6.73 6.92 10.68 10.38 13.38 8.8 8.78 12.5 16.89 9.82 9.98
5 4.83 2.94 5.04 10.35 10.35 13.24 3.88 10.68 14.68 13.14 7.89 8.82
6 5.88 3.39 3.72 9.73 9.09 10.84 10.68 7.87 9.84 20.37 9.99 9.22
7 3.42 6.05 7.86 11.48 12.66 15.66 11.23 11.17 15.42 18.74 13.36 11.55
8 4.71 5.09 6.57 11.76 11.76 15.43 9.57 9.57 9.77 20.99 7.53 10.25
9 5.61 8.95 6.87 10.71 12.02 14.95 7.88 3.88 19.38 20.58 10.94 11.07
Mean 4.96 6.57 6.59 11.07 11.28 13.56 8.28 8.58 14.4 17.26 10.63 1029
CV 11.5 10.89 17.5 10.16 5.13 5.22 8.84 9.77 10.57 4.26 10.11
LSD 104.3 130.21 209.8 204.5 105.4 128.76 133.31 152 277.3 133.85 195.6
Signif ** ** * * ** ** ** ** ** ** **
Key: BA18= Babile2018; MK14=Melkassa 2014; MK18= Melkassa 2018; MS14=Miesso 2014; MS16=Miesso 2016; MS18=Miesso 2018; PW14=Pawe 2014; PW16=Pawe 2016;
SK16=Sekota16; SK18=Sekota 2018; SR18 Srinka 2018.
[151]
Table 4. Mean days to Flowering (DF) and days to Maturity (DM) of nine mung bean lines evaluated across environments
DF DM
PW14
PW16
PW14
PW16
MK14
MK18
MS14
MS16
MS18
MK14
MK18
MS14
MS16
MS18
SR18
SR18
Mean
BA18
SK16
SK18
BA18
SK16
SK18
Mean
Genotype
1 54 38 48 42 41 43 39 39 53 52 30 43 71 6477 69 68 71 77 77 69 85 53 71
2 52 38 45 42 41 41 42 42 51 52 28 43 73 6679 70 69 69 79 79 73 85 52 72
3 55 42 49 44 43 44 44 44 50 51 53 47 74 7378 77 70 75 88 88 81 92 73 79
4 54 35 48 41 40 42 39 39 50 51 27 42 76 6677 71 70 70 75 75 70 84 52 72
5 55 37 47 42 41 42 40 39 53 51 28 43 72 6577 71 70 71 80 73 73 83 51 71
6 55 38 47 42 41 42 39 41 49 51 32 43 76 5577 71 70 70 73 77 69 85 53 71
7 54 35 47 42 41 42 40 40 53 52 27 43 75 6571 70 69 71 73 73 69 86 53 70
8 54 37 47 41 40 42 42 42 55 52 28 44 74 6670 70 69 71 78 78 69 85 52 71
9 55 36 48 42 41 41 41 40 59 54 28 44 75 6877 70 69 70 77 80 71 87 52 72
Mean 54 37 47 42 41 42 40 40 53 52 31 74 6576 71 69 71 78 78 72 86 55
CV 2.8 4.2 1.5 2.05 2.1 3.06 2.95 2.95 5.28 3.1 22.78 3.4 10 1 6.3 36 1.4 3.1 3.1 3.1 1.6 11
LSD 1.26 2.34 2.17 2.17 5.06 2.97 12.97 1.5 1.8 4.4 4.4 4 2.4 11
Signif ns Ns ** ns ns ns ** * ** ns * Ns ns ** ns ns ** *** ** ** ** *
Key: BA18= Babile 2018; MK14=Melkassa 2014; MK18= Melkassa 2018; MS14=Miesso 2014; MS16=Miesso 2016; MS18=Miesso 2018; PW14=Pawe 2014; PW16=Pawe 2016;
SK16=Sekota16; SK18=Sekota 2018; SR18 Srinka 20
[152]
AMMI analysis for grain yield
When ranking of the genotypes changes across environments, the GEI occurs in
various forms. It has the most extreme form it consists of crossover interactions.
(Rodrigues et al., 2014). The Additive Main effect and Multiplicative Interaction
(AMMI) model is one of the most widely statistical models it can be used to
understand and structure interaction between genotypes and
environments(Gauch,1992). Combined analysis of variance (ANOVA): The
ANOVA for grain yield data using the Additive Main effect and Multiplicative
Interaction (AMMI) method is presented in Table 5. The results of combined-
ANOVA indicated that the differences among all sources of variation were highly
significant (P<0.001). The environment (E), genotype (G), and G × E interaction
effects were accounted for 65.53 %, 5.10%, and 26.08% of the total sum of
squares (TSS) in that order. This result revealed that there was a differential yield
performance among the genotypes across the testing environments due to the
presence of G × E interaction. In the AMMI analysis, the first and second IPCA
scores explained 35.32% and 26.31%, respectively, of the G×E interaction effect.
Table 5: AMMI ANOVA of grain yield for 9 mung bean genotypes across environments.
Source Degree of Sum of Mean Total Variation GxE Explained
Freedom square Square explained (%) (%)
Treatments 98 56124687 572701***
Genotypes 8 2957286 369661*** 5.10
Environments 10 38032210 3803221*** 65.53
Block 22 374965 17044*** 0.65
Interactions 80 15135190 189190*** 26.08
IPCA 1 17 7939249 467015*** 52.46
IPCA 2 15 3674960 244997*** 24.28
IPCA 3 13 2053872 157990*** 13.57
IPCA 4 11 623553 56687*** 4.12
Pooled error 176 1535799 8726 2.65
Total 296 58035450 196066
Test environments evaluation based on GGE biplots

Although the AMMI model used for genotype evaluation the GGE biplot also
used for environmental evaluation. The line passing through the biplot origin is
called the average environment coordinate (AEC) (Figure 2), in which any test
environments are closed to AEC that are more representative environments (Yan
and Kang, 2006). MK18, MS14, and MS16 had very short vectors, small angles,
and closed to AEC which means those environments are the most representative
environment. The environment SK16 and SK18 had the longest vectors from the
biplot origin and they were also a more discriminating and less representative
environment but SR16 and MK14 are closest to the ideal environment.
[153]
Fig 1.The environment-Vector view of the GGE biplot to display test environments.
Correlation of tested environment

The concentric circles on the biplot help to visualize the length of the environment
vectors, which is relative to the standard deviation within the respective
environments and is a measure of the discriminating ability of the environments.
Therefore, among the 11 environments, MS16 and SR18 were most discriminating
(informative). The comparison between two environments is determined by the
length of their vectors and the angle between them (Figure 2). Environment SK16
and PW14, PW16 as shown by a long environmental vector and had almost a right
angle between them that showed there is not any correlation. The unrelated
correlation between two environments SK16 and PW14, PW16 indicated a
crossover GE interaction; thus, the changes in ranking order of genotypes form
one year to another in the same location, However, MK18 and MS18 had short
environmental vector. This means if the study is carried out for several seasons
and the same sites continue to be non-discriminating; it means the locations can be
throw down and not to be used as test locations.
Environments are fewer valuable because they bring little discriminating

information about the genotypes as reported by Yan and Tinker (2006).
[154]
Fig 2.The discrimination view of GGE biplot.
The ‘which – won – where’ pattern and mega environments

The polygon view of a biplot provides the best way to imagine the interaction
patterns between genotypes and environments and to effectively understand a
biplot. It is drawn by connecting the genotype markers located. Furthest from the
biplot origin using straight convex hull to form a polygon such that all other
genotype markers are contained within the polygon. Environments that fall in
different sectors have different best genotypes. The which-won-where biplot
showed different winning genotypes in different environments (Gasura et al.,
2015).
Mega-environment idea needs multi-year data, in this study four mega

environments were formed (Figure 3). Thus environments, PW14, PW16, SR18,
MS14, MS18, MK14, and MK18 formed one mega environment, while SK16,
MS18, and SK18 formed a single split mega-environment respectively. The
engaging genotypes for each segment are those located at the vertex. The vertex
genotypes in this investigation were 3, 2, and 6. Genotypes 1, 6, and 8 were the
attractive genotypes for the SK18 formed one single split environment and 3was
also the winning genotype for the SR18 mega environment. Genotypes 5, was the
winning genotypes for the MS18 formed one single split environment, while 7and
3 was the winning genotype for the PW14, PW16, SR18, MS14, MS18, MK14
and MK18 formed one mega environment.
[155]
The convex hull is fairness lines between adjacent genotypes on the polygon,
which facilitate graphic comparison of the genotypes. The equality line between 3
and 2 shows that the 3 was better than 2 in all environments. Genotype 9 is located
on the convex hull that connects 3 and 2. This means that the three can be ranked
3, 2, and 9 in all the environments (Figure 4). This pattern recommends that the
mark environment may contain four different mega-environments and that
dissimilar cultivars should be selected and organized for each. This allows the
researcher to have a specific and effective explanation to recommend genotypes
that are best for that particular environment (Gasura et al., 2015).
Fig 3.GGE biplot analysis showing the mega-environments and their respective high yielding genotypes.
Mean Performance and Stability of the Genotypes

Figure 4 illustrates the graphic evaluation of the comparative performance of
tested genotypes in environment MS16 which produced high and stable yield
among the 11 environments. A line was drawn that passed through the biplot’s
origin and the MS16 marker to make an average environment axis (AEA), and
then another line drawn vertical from each genotype toward the AEA. The
genotypes were ranked based on their projections onto the AEA, with rank
increasing in the direction toward the positive end (Yan and Tinker, 2006).
Genotypes should be evaluated on both mean performance and stability across
[156]
environments. The average environment coordination (AEC) view of the GGE
biplot origin (Figure 4). Lines vertical to average environment axis (AEA)
measures the stability of genotypes in either direction. Genotypes with the
smallest vertical line and close to AEC are called stable genotype. Lowest mean
yield. Stability and high performance make a candidate the best genotype (Mare et
al., 2017). In this biplot, 3, and 7 were most stable with the higher yield
performances than the overall mean, whereas 4, 5and and 1 were also most stable
with the lower mean yield performances than the overall mean, as they were
located almost onto the AEA, indicating that their ranks were highly consistent
across environments. Genotypes were the best or the poorest genotypes in some or
all of the environments since they had the greatest distance from the origin as
reported by Yan and Kang (2006). Based on their stability and high yield, two
genotypes (3 and 7) were outstanding in the common of the environments. Such
genotypes were showed stability and best yield performance in high yield
environments, which was as reported by Finlay and Wilkinson (1963). However,
2, 9, 8, and 6 are the most unstable genotypes.
Fig 4. GGE for ranking of all genotypes relative to the test environments.
[157]
Conclusion
The significant GEI for DTF, DTM, and GY observed from analysis of variance in
this study shows that early mung bean genotypes respond differently when grown
in different environmental conditions. Environmental effects, as well as GEI, had
a strong effect on yield of early maturing mung bean genotypes; it means a
breeder faces the challenge of selection genotypes for improvement and or release,
thus additional testing for genotypes with wider and specific adaptation and
locations with good discriminating ability and representativeness were done. The
results from this study indicated that 3, and 7 were stable and high yielding and
best genotype across environments whereas environment SK16, SK18, and MS18
had excellent potential for grain yield in areas. The genotypes 9, 6, 2, and 8 were
poor performance for grain yield to be found outside limits of any environments.
The performance of early maturing mung bean is attributed to both genetic make-
up and environment.
References
Berhanu Amsalu., Kidane Tumsa, Kassaye Negash, Getachew Ayana, Amare Fufa, Mulatwa
Wondemu, Mulugeta Teamir, J.C. Rubyogo. 2016. Lowland pulses research in Ethiopia:
achievement, challenges and future prospect. pp 44-60. In: Dawit Alemu, Eshetu Derso,
Getnet Assefa and Abebe Kirub (eds). Agricultural Research for Ethiopian Renaissance.
Proceedings of the National Conference on Agricultural Research for Ethiopian Renaissance
held on January 26-27, 2016, in UNECA, Addis Ababa to mark the 50th Anniversary of the
establishment of the Ethiopian Institute of Agricultural Research (EIAR).
Central Statistical Agency. 2017/18. Ethiopian agricultural sample enumeration: statistical reports
on report crop production. Central Statistical Agency, Addis Ababa.
Chhetri N, Chaudhary P, Tiwari PR, Yadaw RB. 2012. Institutional and technological innovation:
Understanding agricultural adaptation to climate change in Nepal. Applied Geography, 33,
142-150.
Eagles HA and Frey KT. 1977. Repeatability of the stability variance parameter in oats. Crop Sci.,
17: 253-256.
Gauch HG. 1992. Statistical analysis of regional yield trials: AMMI analysis of factorial designs.
Elsevier, Amsterdam.
Gauch HG. 1988. Model selection and validation for yield trial with interaction. Biometrics
44(3):705-715.
Gasura E, Setimela PS, Souta CM. 2015. Evaluation of the performance of sorghum genotypes
using GGE biplot. Canadian J. Plant Sci. 95: 1205-1214.
Imrie, B. 1998. The New Rural Industries: AHand Book for Farmers and Investors: Mung
Bean.(http://www.rirdc.gov.au/pub/handbook/mun gbean.html) (Accessed on November 10,
2009).
Mare M, Manjeru P, Ncub B, Sisito G. 2017. GGE biplot analysis of genotypes by environment
interaction on Sorghum bicolor L. (Moench) in Zimbabwe. African J. Plant Sci.11: 308-319.
McIntosh MS. 1983. Analysis of Combined Experiments 1. Agronomy Journal 75(1):153-5.
Moore IS, Jones A, Dixon S. 2014. The pursuit of improved running performance: Can changes in
cushioning and somatosensory feedback influence running economy and injury risk?.
Footwear Sci. 6(1):11.
[158]
Rodrigues PC, Malosetti M, Gauch HG, Van Eeuwijk FA. 2014. A weighted AMMI algorithm to
study genotype-by-environment interaction and QTLby-environment interaction. Crop
Science.
GOMASHE SS, PATIL JV, DESHMUKH SB, SARODE SB. 2008. Asian J. of Bio Sci. (April,
2008) Stability for seed yield and its components in mungbean [Vigna radiata (L.) Wilczek]
Vol. 3 No. 1 : (111-114).
Yan W and Tinker NA. 2006. Biplot analysis of multi environment trial data: Principles and
applications. Canadian J. Plant Sci. 86: 623–645.
[159]
[160]
Introduction and Evaluation of Quality Protein
Maize Hybrid Varieties for Low Moisture
Stress Agro-ecology
Alemeshet Lemma1, Lealem Tilahun1, Dereje Ayalneh1, Bulo Dhabesa1,
Talef Wendosen1, and Estefanos Habtemariam1
1
Melkassa Agricultural Research Center, ,E-mail: melkasaagriculturalresearch@eiar.gov.et,
PO Box 436, Adama, Ethiopia: Corresponding author:; E-mail:alelgc98@gmail.com
Abstract
Maize has been instrumental for food security for Ethiopia through nutrition is far
from answered. The study was proposed to avail varieties with improved protein
content over conventional maize. Fifty QPM hybrids were introduced and evaluated
in 2014 and then the top ten and eight of them were selected and evaluated as
national variety trials in 2016 and 2018, respectively. Alpha lattice and RCBD
designs were used for the experiment in 2014 and 2016+2018 respectively. Analysis
was done using field book for across locations in 2014 and 2016 experiments while
Individual location analysis in 2018. Across year ANOVA was not conducted due to
variations in the number of genotypes and their kind. NOVA for the 2014 trial
showed there was a significant difference among the genotypes for Grain Yield (GY)
and related traits. The GY ranged from 1.84tonha-1(ZS621) to 5.8tonha-1(CZH084)
with a mean of 2.8ton ha-1. However, in the 2016 trial, it showed that there was no
significant difference among the genotypes for grain yield. In the 2018 trial, it
showed that there was a significant difference among the genotypes at Ziway and
Miesso for GY. The best-performed genotype in terms of GY was CZH132041Q with
a yield of 6.91tonha-1, 2.58tonha-1 and 1.60 ton ha-1 at Melkassa, Ziway and Miesso
respectively. The result of this experiment showed that CZH132041Q was an
outstanding hybrid that out-yielded both the conventional and QPM checks across
the tree year experiment across all the locations which showed that the variety will
be promoted to variety verification trials for possible release.
Keyword: Quality Protein Maize (QPM), National Variety Trial (NVT)
Introduction
Maize grows in 13 agro-ecological zones of Ethiopia and adapted to 90% of arable
land (Dawit A. et al., 2014). The national average yield increased over the years
significantly from 1.74 t/ha in 1993 to 3.73 t/ha in 2017 so as the land size from
0.84 million hectares to 2.14 million hectares in these years (FAO, 2017). The
yield increment recorded was majorly attributed to other factors like the use of
improved agricultural inputs like varieties and mineral fertilizers than an increase
in the production area (Tsedeke A. et al., 2015).
[161]
Maize research in Ethiopia was started in the 1950’s targeting mainly smallholder
farmers (Kebede Mulatu et al., 1992). Research for drought-tolerant maize was
initiated by the then Awassa Junior College of Agriculture in 1976 and released
the first synthetic variety called Alamura yellow in 1984 (Hussein M. and Kebede
M., 1992). The low Moisture and heat stress maize research program has been
focused on variety development adapted to the moisture stress agro-ecologies of
Ethiopia since its establishment in 1993. The subprogram released about eight
OPVs and three hybrids during these times. Introduction and evaluation for
identification has been playing a significant role in fast-tracking and releasing
useful varieties (Kebede Mulatu et al., 1992). So far all the eleven varieties
released from Melkassa for the low moisture stress agro-ecologies are from
introductions which depicted its importance in product identifications. CIMMYT
Kenya, CIMMYT Zimbabwe and IITA are our major sources of germplasms for
introduction.
The research for Quality Protein Maize (QPM) breeding was started in 1994 by
introducing and evaluating open-pollinated varieties and pools from CIMMYT.
The research was strengthened through Quality Protein Maize Development
(QPMD) nd Nutritious Maize for Ethiopia (NuME) projects from 2003 to 2018.
(Teklewold A., et al., 2015). During these periods varieties like Melkassa 6Q;
Intermediate maturing QPM OP variety was released in 2008 and has been
majorly utilized in seed production and distribution followed by MH138Q a
hybrid variety released in 2012 was significantly produced and demonstrated in
the central rift valley area. In 2013 Melkassa 1Q; the QPM version of Melkassa
one was released and has been under production since then. Therefore the study
was proposed to avail varieties with improved protein content over conventional
maize for those users whose diet was dominated by maize and who cannot afford
animal protein.

Genotypes
Forty-nine three-way cross QPM hybrids were introduced from CIMMYT
Zimbabwe in 2014 and were evaluated at Melkassa and Dhera along with one
local check; the two quarantine sites where the subprogram testes introduced
materials. Seventeen genotypes showed better performance over the local check.
In 2015 the selected materials were planted in four different locations but data was
not collected from any of them due to extreme drought. In 2016 ten genotypes
were selected again and planted across four locations namely Melkassa, Dhera,
Ziway, and Miesso and again the Data from Ziway was not included due to its
poor quality. In 2018 eight genotypes were evaluated across four different
locations described above.
[162]
Table 1. List of QPM three-way cross hybrids introduced and evaluated at Melkassa and Dhera in 2014.
Genotype Pedigree
CZH084 [CML141/[CML141/CML395]F2-1sx]-4-2-1-B*4-1-BB/CML511//CML181
CZH132011Q [CML141/[CML141/CML395]F2-1sx]-4-2-1-B*4-1-BB/CZL083//[[CML202/CML144]F2-1-1-3-B-1-
B*6/CZL066]-B*5-3-B
CZH132017Q CML144/CZL083//([CML443/CZL066]-2/CML443)-B*4-2-B
CZH132018Q CML491/CML511//[CML507/CZL066]-2-B-1-BBB-2-B
CZH132032Q [GQL5/[GQL5/[MSRXPOOL9]C1F2-205-1(OSU23i)-5-3-X-X-1-BB]F2-4sx]-11-3-1-1-
B*5/[CML141/[CML141/CML395]F2-1sx]-4-2-1-
B*5//(CML312IR/[(CLQRCWQ50/CML312SR//CML312SR)-3-1-B/(CLQRCWQ50/CML312SR)-2-
2-1-BB]-BB)F2-B-21-2-1-B-B
CZH132010Q [CML141/[CML141/CML395]F2-1sx]-4-2-1-B*4-1-
BB/CML511//([CML507/[CML182/[CML182/CML395]F2-3sx]-4-1-B]-1/CML507)-BBB-2-B
CZH132019Q CML491/CZL083//[(CLQRCWQ50/CML312SR//CML312SR)-3-1-B/CML312SR]-B*5-1-B-B
CZH132015Q [CML141/[CML141/CML395]F2-1sx]-4-2-1-B*4-1-BB/CML511//[CML507/CZL066]-2-B-1-BBB-2-B-
B
CZH132020Q CML491/CZL083//[Syn01E2/CML511]-16-B-2-BBB-11-B-B
B*5/[CML141/[CML141/CML395]F2-1sx]-4-2-1-B*5//([CML144/[CML144/CML395]F2-8sx]-1-2-3-2-
B*4-1-B/[INTA-2-1-3/INTA-60-1-2]-X-11-6-3-BB)F2-B-16-6-3-B-B
CZH132039Q CZL083/[GQL5/[GQL5/[MSRXPOOL9]C1F2-205-1(OSU23i)-5-3-X-X-1-BB]F2-4sx]-11-3-1-1-B*5-
3-B//(CML197-B/(CLQRCWQ50/CML312SR)-2-2-1-BBB)F2-B-13-3-2-B-B
CZH132014Q [CML141/[CML141/CML395]F2-1sx]-4-2-1-B*4-1-BB/CZL083//ObatanpaSR-278-B*4-2-B-B
3-B//(CML312IR/[(CLQRCWQ50/CML312SR//CML312SR)-3-1-B/(CLQRCWQ50/CML312SR)-2-
2-1-BB]-BB)F2-B-4-2-3-B-B
CZH132013Q [CML141/[CML141/CML395]F2-1sx]-4-2-1-B*4-1-BB/CZL083//[Syn01E2/CML511]-16-B-1-BB-1-
B-B
VH05291 CML181/WWO1408-1-1-2-B//CML511-B
CZH132007Q [(CLQRCWQ50/CML312SR)-2-2-1-BB/[CML390//[CML390/GQL5]-B-10sx]-B-6-1-BBB]-B*5-1-
B/[[CZL083/CML509]F2-1/CZL066]-9-B-1-BBB-6-B//CML511
B*5/[CML141/[CML141/CML395]F2-1sx]-4-2-1-B*5//(CML197-B/(CLQRCWQ50/CML312SR)-2-2-
1-BBB)F2-B-13-3-2-B-B
MH138Q CML144/CML159//Pool15QPMSR538
B*5/[CML141/[CML141/CML395]F2-1sx]-4-2-1-
B*5//(CML312IR/[(CLQRCWQ50/CML312SR//CML312SR)-3-1-B/(CLQRCWQ50/CML312SR)-2-
2-1-BB]-BB)F2-B-4-2-3-B-B
CZH132012Q [CML141/[CML141/CML395]F2-1sx]-4-2-1-B*4-1-BB/CZL083//[[CML506/[CML205/CML176]-B-2-
1-1-B]F2-1/[CML144/[CML144/CML395]F2-8sx]-1-2-3-2-B*4]-24-B-2-BBB-3-B-B
B/[[[CML150/CML373]-B-2-2-B/CML509]F2-2/CML511]-5-B-1-BBB-1-B//CML491
BB/[CML205/CML176]-B-2-1-1-2-B//(CML197-B/(CLQRCWQ50/CML312SR)-2-2-1-BBB)F2-B-13-
3-2-B-B
PAN 53 PAN 53
B/[[CML202/CML144]F2-1-1-3-B-1-B*6/CZL066]-B*5-1-B//CML491
CZH132003Q [CML141/[CML141/CML395]F2-1sx]-4-2-1-B*4-1-BB/CML491//[CZL066/CML511]-1-B-1-BBB-1-B-
B
CZH132009Q [[[CML150/CML373]-B-2-2-B/CML509]F2-2/CML511]-5-B-1-BBB-1-B/[VP047/CML511]-25-B-2-
BBB-6-B//CML491
CZH132044Q CZL1210/CZL1227//CML511
CZH132022Q [[CML202/CML144]F2-1-1-3-B-1-B*6/CZL066]-B*5-1-B/[[[CML159/[CML159/[MSRXPOOL9]C1F2-
205-1(OSU23i)-5-3-X-X-1-BB]F2-3sx]-8-1-BB/CML509]F2-1/CZL066]-9-B-1-BBB-6-B//CML181
[163]
CZH132043Q CZL1210/CZL1227//CML181
CZLH132006Q [(CLQRCWQ50/CML312SR)-2-2-1-BB/[CML390//[CML390/GQL5]-B-10sx]-B-6-1-BBB]-B*5-1-
B/[[CZL083/CML509]F2-1/CZL066]-9-B-1-BBB-6-B//CML181
CZH132016Q [CML141/[CML141/CML395]F2-1sx]-4-2-1-B*4-1-BB/CZL083//CompMSRXG9Q-B*5-1-B-B
3-B//(CML312IR/[(CLQRCWQ50/CML312SR//CML312SR)-3-1-B/(CLQRCWQ50/CML312SR)-2-
2-1-BB]-BB)F2-B-21-2-1-B-B
CZH132028Q [CML144/SNSYNF2[N3/TUX-A-90]-102-1-2-2-BSR-B*4]-B-4-3-B*4-1-
B/[GQL5/[GQL5/[MSRXPOOL9]C1F2-205-1(OSU23i)-5-3-X-X-1-BB]F2-4sx]-11-3-1-1-B*5-3-
B//(CML312IR/[(CLQRCWQ50/CML312SR//CML312SR)-3-1-B/(CLQRCWQ50/CML312SR)-2-2-
1-BB]-BB)F2-B-21-2-1-B-B
B//(CML312IR/[(CLQRCWQ50/CML312SR//CML312SR)-3-1-B/(CLQRCWQ50/CML312SR)-2-2-
1-BB]-BB)F2-B-4-2-3-B-B
B//(CML197-B/(CLQRCWQ50/CML312SR)-2-2-1-BBB)F2-B-13-3-2-B-B
CZH132019Q CML491/CZL083//[CZL066/CML511]-1-B-1-BBB-1-B
B/CompMSRXG9Q-B*5-1-B//CML491
CZH132024Q [CML141/[CML141/CML395]F2-1sx]-4-2-1-B*4-1-
B/CML491//[[[CML159/[CML159/[MSRXPOOL9]C1F2-205-1(OSU23i)-5-3-X-X-1-BB]F2-3sx]-8-1-
BB/CML509]F2-2/CZL057]-5-B-1-BBB-2-B-B-B
3-B//([CML144/[CML144/CML395]F2-8sx]-1-2-3-2-B*4-1-B/[INTA-2-1-3/INTA-60-1-2]-X-11-6-3-
BB)F2-B-16-6-3-B-B
CZH132021Q CML491/CZL083//CompMSRXG9Q-B*5-1-B-B
BB/[CML205/CML176]-B-2-1-1-2-B//(CML312IR/[(CLQRCWQ50/CML312SR//CML312SR)-3-1-
B/(CLQRCWQ50/CML312SR)-2-2-1-BB]-BB)F2-B-4-2-3-B-B
B//(CML197-B/(CLQRCWQ50/CML312SR)-2-2-1-BBB)F2-B-49-1-2-B-B
CZH132041Q CZL1210/CZL1227//CML491
B/(CLQRCWQ01/CML312SR)-4-2-1-B//[[GQL5/[GQL5/[MSRXPOOL9]C1F2-205-1(OSU23i)-5-3-
X-X-1-BB]F2-4sx]-11-3-1-1-B*4/CML511]-1-B-1-BBB-1-B-B
CZH132023Q [[CML202/CML144]F2-1-1-3-B-1-B*6/CZL066]-B*5-1-B/[[[CML159/[CML159/[MSRXPOOL9]C1F2-
205-1(OSU23i)-5-3-X-X-1-BB]F2-3sx]-8-1-BB/CML509]F2-1/CZL066]-9-B-1-BBB-6-B//CML491
CZH132042Q CZL1210/CZL1227//[CML141/[CML141/CML395]F2-1sx]-4-2-1-B*4-1-BB-B
CZH132002Q [CML141/[CML141/CML395]F2-1sx]-4-2-1-B*4-1-BB/CML511//[[CML202/CML144]F2-1-1-3-B-1-
B*6/CZL066]-B*5-3-B
BB/[CML205/CML176]-B-2-1-1-2-B//([CML144/[CML144/CML395]F2-8sx]-1-2-3-2-B*4-1-B/[INTA-
2-1-3/INTA-60-1-2]-X-11-6-3-BB)F2-B-16-6-3-B-B
CZH132037Q CML491/CML511//[[[CML159/[CML159/[MSRXPOOL9]C1F2-205-1(OSU23i)-5-3-X-X-1-BB]F2-
3sx]-8-1-BB/CML509]F2-1/CZL066]-9-B-1-BBB-1-B-B-B
ZS621 ZS261
[164]
Description of the study area
Table 2. Description of the Study areas
Location Annual Rain fall (mm) Altitude (masl) Latitude Longitude Soil Type Reference
Dhera 520 1680 8o20’N 39o23’E Sandy-Clay Gelana et al. 2001.
Melkasa 710 1550 8o24’N 39o21’E Sandy- Gelana et al. 2001.
Clay_loam
Miesso 560 1470 9o12’N 40o52’E Clay Loam Gelana et al. 2001.
Ziway 640 1637 8oN 38o35’E Silt Gelana et al. 2001.
Experimental Design and Data Management

In 2014 the experiment was laid in 5x10 α-lattice design (Patterson and Williams,
1976) with two replications. Design and randomization of the trials were
generated using CIMMYT’s computer software known as Fieldbook. Spacing was
75 cm between rows and 25 cm between plants within the row. Each entry
consisted of two rows of 4m long. In 2016 and 2018 the experiment was laid out
in RCBD and randomization of the trials was generated using CIMMYT’s
Fieldbook. Spacing was 75 cm between rows and 25 cm between plants within the
row. Each entry consisted of two rows of 4m long. Two seeds per hill were
planted to ensure the establishment of uniform and enough stand. At 2-4 leaf
stage, they were thinned to one plant per hill for both crops. All other agronomic
(fertilizer application, weeding, and pest control) practices were done based on
research recommendations. Data were collected on different agronomic traits
including grain yield, plant and ear height, ear per plant, male and female
flowering dates, and other agronomic traits on a plot basis. Data Collected on a
plot basis was analyzed using CIMMYT field book for individual and across
location performances analysis.

Analysis of Variance showed that Genotypes had quite sufficient variability to
select among them in different agronomic traits. The minimum yield recorded was
1.89 ton ha-1 while the maximum was 5.89 ton ha-1 with a mean of 3.80 ton-1.
Seventeen genotypes showed a yield advantage ranging from 1% to 43.9% over
the local check (MH138Q) among which only one was non QPM hybrid
(CZH084). VH05291 was the only commercial check performed better than the
local check having a yield advantage of 5.7%. These genotypes had Days to
Anthesis (DA) in the range between 81.6 to 91.8 days. The result showed that
there were significant delays in DA from the normal flowering time in the
moisture stress agroecology which is about 70 days for intermediate maturity
groups. The reseason for delayed flowering was due to terminal drought tress and
was significantly reflected on the yield performance of the genotypes as it was
reported before (Campos et al. 2006; Qingfeng Meng et al. 2016; Abdullah et al.
2017). EPP was considered as a secondary trait especially in drought stress
[165]
breeding (Campos et al. 2006). The genotypes chowed good variability varied
from 0.76 to 1.95 with a mean of 1.13 depicted some materials showed a good
level of resistance to the terminal stress that happened during the experiment
(Campos et al. 2006).
Tabel 3. Across location performance of genotypes across Melkassa and Dhera in 2014 for Grain yield and related traits.
Entry Name RelGY RANK GY DA EPP RL HC PA TLB EH ASI NP

10 CZH084 1.44 1 5.98 83.4 1.25 5.5 6 3 2 131.6 -0 45
11 CZH132011Q 1.18 2 4.89 88.2 0.97 15.3 1.7 2.8 2 115.9 0.3 36.5
17 CZH132017Q 1.16 3 4.83 84.6 0.89 18.3 2 2.5 2.4 95.4 3.6 39
18 CZH132018Q 1.15 4 4.78 84.5 1.2 21.8 4 2.8 2 99.1 2.7 32.2
37 CZH132032Q 1.15 5 4.77 89.3 1.6 1.3 6.5 1.8 1.5 107.9 -1 19.7
7 CZH132010Q 1.13 6 4.68 83.3 1.11 2.6 9.3 2.8 1.7 116.4 1.8 32.5
19 CZH132019Q 1.12 7 4.67 85.1 1.11 9.6 5.5 2.8 2.5 115.5 3.3 34
15 CZH132015Q 1.12 8 4.65 85.1 0.84 27 0.9 2.6 2 114.9 -0 44.2
21 CZH132020Q 1.11 9 4.61 85 1.26 5.3 11 2.8 2 88.2 -0 32.7
35 CZH132030Q 1.09 10 4.54 84.7 0.94 37.5 4.5 2.6 2.3 120.5 -1 29.7
45 CZH132039Q 1.09 11 4.52 85.6 0.96 12 4.5 2.9 2.5 105 -1 34.6
14 CZH132014Q 1.08 12 4.47 81.6 0.96 12.5 13 2.3 2 105.1 1.8 27.9
47 CZH132041Q 1.07 13 4.45 86.5 0.91 41.4 -1.5 2.5 1.8 107.3 0.2 31.2
13 CZH132013Q 1.06 14 4.39 82 1.12 4.5 5.8 2.6 2.2 107.6 2.1 35
42 VH05291 1.06 15 4.39 83.1 0.99 4.5 5.2 2.4 2.3 98.8 2 33.2
4 CZH132007Q 1.01 16 4.19 87 1.09 9.2 -1 2.9 2 114.8 1.8 38.6
36 CZH132031Q 1.01 17 4.19 85.3 1.01 10.6 10 2.5 2.3 109.7 0.6 30.9
50 MH138Q 1 18 4.15 84.3 1.28 10.7 0.7 2.9 2.2 113.9 3.4 29.4
38 CZH132033Q 0.99 19 4.1 88.4 1.3 7.4 5 2.5 2 88.7 -4 20.3
12 CZH132012Q 0.97 20 4.01 86.3 1.03 24.5 3.5 2.5 1.7 110.3 3.6 34.3
1 CZH132004Q 0.96 21 4 89.2 0.96 4.3 7.2 2.6 2 124.5 0.1 37.4
40 CZH132035Q 0.94 22 3.9 89.4 1.36 10.7 -0.8 2.6 2 130.1 2.9 26.4
49 PAN 53 0.94 23 3.89 87.8 0.76 8.3 2.7 2.8 2 96.8 1.5 31.5
2 CZH132005Q 0.93 24 3.87 88 1.1 12.2 3.5 2.8 2.2 119.3 2.6 36.9
8 CZH132003Q 0.92 25 3.83 88.8 1.25 6 5.3 2.5 1.8 109.7 1.2 29.9
6 CZH132009Q 0.92 26 3.82 87.8 0.99 7.6 2.7 2.7 2 103.6 1 42.4
26 CZH132044Q 0.91 27 3.77 88.4 1 6.8 14 2.7 2 121.2 -1 33.8
27 CZH132022Q 0.89 28 3.69 87.3 1.04 1.4 12 2.5 2.5 100.7 3.4 32
25 CZH132043Q 0.88 29 3.64 87.3 1.02 3.3 9.5 2.9 2 101.9 1.6 27.1
3 CZLH132006Q 0.87 30 3.63 85.1 0.95 5.6 5.1 2.8 2.2 113.5 0.4 32.7
16 CZH132016Q 0.87 31 3.59 85.6 0.93 8.3 15 2.8 2.3 130.9 0.9 31.6
46 CZH132040Q 0.85 32 3.54 87.8 1.54 1.8 3.2 2.5 1.7 107.2 0.3 17.7
33 CZH132028Q 0.84 33 3.48 90 1.31 17.2 0.8 2.1 1.7 126.2 -1 18.1
34 CZH132029Q 0.83 34 3.43 87 1.63 8 17 2.2 1.8 107.2 -3 16.8
31 CZH132026Q 0.81 35 3.37 90.5 0.96 14.9 -0.3 3.2 2.2 116.8 -0 27.9
20 CZH132019Q 0.81 36 3.36 87 0.93 3.3 5.3 2.9 2 103.5 -1 33.8
5 CZH132008Q 0.8 37 3.32 89.8 1.32 12.5 0 2.3 1.7 93.8 0 25.9
29 CZH132024Q 0.8 38 3.31 86.6 0.94 15 5.1 2.9 2 116.2 3.4 32.4
44 CZH132038Q 0.78 39 3.25 86.4 0.82 27.1 3.4 2.5 2.5 114.4 0.1 38.9
22 CZH132021Q 0.74 40 3.09 88.7 1.12 22 11 2.7 2 111.4 0 27.2
41 CZH132036Q 0.73 41 3.05 88.6 1.95 20.2 0.5 2.4 1.8 105.3 -2 10.1
32 CZH132027Q 0.73 42 3.01 88.7 1.14 18.6 -0.8 2.5 2.2 108.5 -1 22.5
23 CZH132041Q 0.71 43 2.93 88 0.93 20.8 3 3.1 2.3 100.9 1.7 38.3
30 CZH132025Q 0.7 44 2.89 91 1.04 4.6 1.5 2.8 2 98.8 1.4 27.8
28 CZH132023Q 0.69 45 2.87 90.2 1.37 22.8 5.3 2.3 2.2 102.4 1 18.4
24 CZH132042Q 0.68 46 2.8 86.8 1.16 4.3 6.2 2.8 2 114.7 1.1 25.6
9 CZH132002Q 0.66 47 2.72 91.8 1.13 11.7 21 2.9 2 103.6 1.7 19.4
39 CZH132034Q 0.56 48 2.31 88 1.33 22.2 1.6 2.6 1.8 99.6 0.4 14
[166]
43 CZH132037Q 0.55 49 2.29 91.3 1.23 -4.4 16 2.7 1.8 130.1 1.1 13.8
48 ZS621 0.46 50 1.89 88.1 1.22 -4.5 29 2.7 1.5 105.9 -0 8.6
Mean 100 26 3.8 87.1 1.13 11.9 6 2.6 2 109.7 0.7 29.2
LSD_0.05 21 15 1.6 5.7 0.44 20.3 14 0.6 0.6 26.5 3.2 10.5
MSe 0.63 8.1 0.05 102 49 0.1 0.1 174.4 2.5 27.5
Min 50 1 1.89 81.6 0.76 -4.5 -1.5 1.8 1.5 88.2 -4 8.6
Max 157 50 5.98 91.8 1.95 41.4 29 3.2 2.5 131.6 3.6 45
RelGY= Relative Grain Yield, GY= Grain Yield (ton/ha), DA= Days to Anthesis (days), EPP= Number of Ears Per Plant,
RL= Root Lodging (%), HC= Bad Husk Cover (%), PA= Plant Aspect (1-5 Scale, 1= Excellent, 5= bad), TLB= Turcicum
Leaf Blight (1-5 Scale, 1= Resistant, 5= Susceptible), EH= Ear Heigh (cm), ASI= Anthesis Silking Interval, NP= Number of
Plants per plot.
Based on Seed availability CZH132018Q, CZH132019Q, CZH132041Q,

CZH132030Q, CZH132032Q, CZH132020Q, CZH132039Q were selected for
2016 national variety trial as 2015 was so dry, most of the trials did not produce
quality data for analysis and decision making.
The best Performing Genotype (CZH132041Q) Showed a 42 % Yield advantage

over the best non QPM check (MH140) while it showed a 46% yield advantage
over the QPM check (MH138Q). Whereas three varieties CZH132018Q,
CZH132019Q, and CZH132020Q took the next ranks and performed more or less
similarly having yield advantage ranging from 25-30 % over MH138Q and 21-
26% yield advantage over MH140 (Table 4). CZH132041Q Was one of the top
performed variety back in 2014 across Melkassa and Dhera (Table 3) which
showed that this variety had a potential for the low moisture stress agro-ecologies
of Ethiopia if it was approved for release.
Table 4. Performance of Selected genotypes from 2014 introductions evaluated across three locations (Melkassa, Dhera
and Miesso) in 2016.
Yield Adv Yield Adv
over Non over QPM
Entry Genotypes QPM Check Check Rank GY DA EPP EA PA PH EH
1 CZH132018Q 1.2423 1.2786 4 4.27 69.4 1.13 2.6 2.9 223.9 59.4
2 CZH132041Q 1.4226 1.4642 2 4.89 67.7 1.10 1.7 2.9 243.7 72.2
3 CZH132030Q 0.9544 0.9824 8 3.28 69.9 1.03 2.9 2.7 216.0 71.1
4 CZH132039Q 1.2646 1.3016 6 4.34 68.7 1.01 2.8 2.7 221.7 88.4
5 CZH132019Q 1.2648 1.3018 4 4.35 68.4 0.92 2.2 2.4 244.0 74.2
6 CZH132020Q 1.2153 1.2508 4 4.18 66.6 1.03 2.1 2.3 231.2 59.9
7 CZH132032Q 1.2009 1.2360 6 4.13 69.3 1.14 2.3 2.4 208.4 73.3
8 MH138Q 0.9716 1.0000 8 3.34 68.2 1.07 3.0 3.0 237.6 62.0
9 MH140 1.0000 1.0293 8 3.44 68.3 0.83 2.6 3.2 231.5 79.3
10 Melkassa2 0.9296 0.9568 9 3.19 67.6 1.04 2.6 2.8 227.8 72.6
Mean 100 6 3.94 68.4 1.03 2.5 2.7 228.6 71.2
LSD (0.05) 14 2 0.97 2.3 1.34 0.1 1.0 0.9 18.9
MSe 0.64 1.2 0.01 0.3 0.3 121.8 82.4
Min 81 2 3.19 66.6 0.83 1.7 2.3 208.4 59.4
Max 125 9 4.89 69.9 1.14 3.0 3.2 244.0 88.4
Yield Adv Over Non QPM check= Yield Advantage over Non QPM Check, Yield Adv over QPM Check= Yield Advantage over the the
QPM Check calculated as a coefficient between (GY of the Genotype/GY of the check) GY= Grain Yield (ton/ha), DA= Days to Anthesis
(days), EPP= Number of Ears Per Plant, PA= Plant Aspect (1-5 Scale, 1= Excellent, 5= bad), EA= Ear Aspect (1-5 Scale, 1= Excellent,
5= bad), PH= Plant Height (cm), EH= Ear Height (cm).
[167]
In 2018 five hybrids and three checks were evaluated across four locations, but
Dhera was excluded from the analysis due to poor data quality. The result showed
that the variety CZH132041Q was found to be best performed across the three
locations. The result showed that once again CZH132041Q showed its superior
performance over the normal check, the QPM check over locations and years
(Tables 3, 4, 5). CZH132041Q showed 23% yield advantage over the non QPM
check at Melkassa and a 48% Yield advantage over MH140 at Miesso and 6%
yield advantage over MH141 at Ziway (Table 5). The result showed that this
variety can be promoted to the level of variety verification trial as it was showing
consistently better performance over both QPM and non QPM checks. This
variety had the smallest ear Aspect which showed that the variety is very good in
terms of yield potential, flintiness, or very good kernel modification and no rotting
or bare-tips or bird attack which all are very good indicators for varietal selection
in our breeding project.
[168]
Table 5. Performance of Selected genotypes Under National Variety Trial Evaluated at Three Different locations in 2018.
Melkassa Ziway Mieso

Entry Name GW Rank DA % GY MH140 GW Rank AD %GY MH140 GW Rank AD % GY MH140
1 CZH132039Q 6.62 2 72.3 1.18 1.94 4 58.6 0.80 1.18 5 73.4 1.09
2 CZH132041Q 6.91 1 70.2 1.23 2.58 1 59.2 1.06 1.60 3 72.2 1.48
3 CZH132018Q 5.85 4 74.7 1.04 0.89 8 61.3 0.37 0.49 8 71.5 0.45
4 CZH132032Q 5.60 6 75.1 1.00 1.87 5 60.5 0.77 2.25 1 74.2 2.08
5 CZH132019Q 6.08 3 71.3 1.08 1.81 6 59.6 0.74 1.28 4 73.0 1.19
6 MH138Q 5.52 7 73.3 0.98 1.57 7 60.5 0.65 0.83 7 72.0 0.77
7 MH140 5.61 5 70.5 1.00 2.43 2 57.8 1.00 1.08 6 72.2 1.00
8 Melkassa2 5.422 8.000 66.966 0.97 2.405 3.000 52.375 0.99 1.784 2.000 68.116 1.65
Mean 5.95 5 71.8 1.94 5 58.7 1.31 5 72.1
LSD (0.05) 2.45 2 1.2 0.96 2 0.61 2 1.5
MSe 3 0 0 3 0 1
CV 27 1 28 3 47 1
p ns *** * ns *** ***
Min 5.42 1 67.0 0.89 1 52.4 0.49 1 68.1
Max 6.91 8 75.1 2.58 8 61.3 2.25 8 74.2
FW= Grain Weight calculated from Field Weight (tonha-1)
[169]
Recommendation
Combined analysis of variance was not performed for locations and years as the
types and number of genotypes tested were different in each of the years.
CZH132041Q was found to be a high yielder over the normal and QPM checks
across most of the test locations and over the years. The variety showed its
worthiness and superior performance over our local checks as well as commercial
checks. By Requesting its parental lines from CIMMYT Zimbabwe or by looking
the pedigrees and find out if the parental lines can be found in either of the three
breeding programs within Ethiopia, the variety should be promoted to variety
verification trial stages and it should be considered in the new project which will
run from 2020-2022.
Acknowledgment
We would like to acknowledge Melkassa Agricultural Research Center,
CIMMYT, and AGP-II projects for the support we get to conduct the trials. The
field assistants of Low Moisture and heat stress maize research subprogram are
duly acknowledged
References
Abdullah Izzam, Hidayatur Rehman, Amir Sohail, Shahzad Ali, Manzoor and Quaid Hussain.
2017. Genetic variability and correlation studies for morphological and yield traits in maize
(Zea mays L.). Pure and Applied Biology. http://dx.doi.org/10.19045/bspab.2017.600131.
Adefris Teklewold, Dagne Wegary, Abraham Tadesse, Birhanu Tadesse, Kassahun Bantte, Dennis
Friesen and B.M. Prasanna. 2015. Quality Protein Maize (QPM): A Guide to the Technology
and Its Promotion in Ethiopia. CIMMYT: Addis Ababa, Ethiopia.
Dawit Alemu, Chilot Yirga, Adam Bekele, Agajie Tesfaye. 2014. Situation and Outlook of Maize
in Ethiopia. Ethiopian Institute of Agricultural Research. ISBN: 978-99944-53-97-9.
FAO. 2017. FAOSTAT. Food and Agriculture Organization of the United Nations, Rome, Italy.
Gelana Seboksa, Mandefro Nigussie and Gezahegne Bogale. 2001. Stability of Drought Tolerant
Maize Genotypes in The Drought Stressed Areas Of Ethiopia. Pp 301-304. In: Friesen, D.K.
and A.F.E. Palmer (Eds.). 2004. Integrated Approaches to Higher Maize Productivity in the
New Millennium: Proceedings of the Seventh Eastern and Southern Africa Regional Maize
Conference, 5-11February, 2002, Nairobi, Kenya: CIMMYT (International Maize and Wheat
Improvement Center) and KARI (Kenya Agricultural Research Institute).
Campos H, Cooper M, Edmeades GO, Löffler C, Schussler JR, Ibañez M. 2006. Changes in
Drought Tolerance in Maize Associated with Fifty Years of Breeding for Yield in the U.S.
Corn Belt. Maydica 51 (2006): 369-381.
Hussein M and Kebede Mulat. 1992. Maize Germplasm Development for Moisture Stress Areas.
Pp25-29. In: Proceedings of the First National Maize Workshop of Ethiopia. 5-7 May 1992,
Addis Abeba, Ethiopia.
Kebede Mulatu, Gezahegn Bogale and Benit Tolesa. 1992. Maize production trend in Ethiopia.
Pp4-12 In:Benti Tolesa and J. K. Ranson (Eds). Proceedings of the First National Maize
Workshop of Ethiopia. -7 May 1992, Addis Abeba, Ethiopia.
[170]
Qingfeng Meng, Xinping Chen, David B. Lobell, Zhenling Cui, Yi Zhang, HaishunYang
and Fusuo Zhang. 2016. Growing sensitivity of maize to water scarcity under climate change.
Nature. Scientific RepoRts | 6:19605 | DOI: 10.1038/srep19605.
Patterson, HD and Williams ER. 1976. A new class of resolvable incomplete block designs.
Biometerica. 63:83-89.
Tsedeke Abate, Bekele Shiferaw, Abebe Menkir, Dagne Wegary, Yilma Kebede, Kindie Tesfaye,
Menale Kassie, Gezahegn Bogale, Berhanu Tadesse, Tolera Keno. 2015. Factors that
transformed maize productivity in Ethiopia. Food Sec. (2015) 7:965–981.
[171]
[172]
Participatory Demonstration and Evaluation of
Improved Maize Technologies in Selected districts
of West Shewa, East Wollega and Ilu-Ababora
Zones of Western Oromia
Effa Wolteji*, Bayisa Gedafa and Dubiso Gacheno
*Corresponding Author, effaw03@gmail.com
Abstract
This activity was conducted in Ilu-Gelan, Bako Tibe, Gobu Sayo, Sibu Sire, Bonaya
Boshe and Ilu-Harar districts of Western Oromia to demonstrate the recently released
maize varieties, BH-546 and BH-547 along with Limu as a standard check to the
farming community in these districts. These districts were purposively selected based
on potentiality for maize production; and one potential PA from each district were
selected based on accessibility and potentiality. Three varieties of maize, BH-546 and
BH-547as a recently released variety along with Limu; as standard check were
planted on 20m*10m adjacent plots on 24 farmers’ fields. All recommended
agronomic practices were equally applied to all the plots and the fields were closely
supervised and were managed well. At maturity, the varieties were jointly evaluated
with a team composed of researchers, Farmers and DAs. Despite the slight variability
in criteria set by farmers at the respective locations, yield, disease tolerance, grain
color, pest resistance, cob length, number of cob per plant, grain size, lodging
resistant, early maturity, bare tip, taste shelf-life quality and time for the backing of
the four were the common selection criteria across all locations. In almost all the
locations BH-546 beat both HB-547 and Limu varieties in yield and the criteria set
for evaluation; except longer time taking to be baked and late-maturing nature, and
impressing the farmers; especially the number of cob per plant, cob length, number of
seeds per plant, taste and retaining shelf-life quality for a long time (not less than ten
days). With regard to yield and farmers’ preference ranking and selection, 9.11t/ha,
8.92t/ha and 8.68 t/ha were obtained from BH-546, BH-547 and Limu varieties;
respectively putting BH-546 on the first rank, then BH-547 second and finally Limu.
Besides; in terms of yield advantage BH-546and BH-547 had 4.94 % and 2.72 %
yield advantage over standard check Limu and this implies that both demonstrated
BH-546and BH-547 had higher yield advantage than Limu. Further; statistically
mean yield comparison (t-test) results obtained from this study revealed that as there
were no significant differences at (p˂0.05) on mean yield performance among the
varieties demonstrated. Furthermore; in terms of profitability, the financial analysis
result of the study also showed that using BH-546 and BH-547 varieties can make it
more profitable than Limu. Cognizant that though there were no statistically
significant differences among the varieties mean yield based on farmers’ preference
ranking, yield advantage, profitability (cost-benefit analysis) gained from the study
the more demanded variety BH-546 will be further taken to the next step. Therefore;
as the variety has met the intended criteria of the farmers and suits farmers’
evaluation criteria in almost all the districts where this activity undertaken the pre-
scaling up activity shall follow next season.
Keywords: Maize; Participatory demonstration, evaluation and selection; Yield

Advantage, BH-546 and BH-547; Limu
[173]
Introduction
Maize was originated in Central America and introduced to West Africa in the
early 16th century (FAO 1992) and to Ethiopia between the 16th and the 17th
century (McCann 2005). Africa produces 6% of the total world maize production,
most of which is used for human consumption (Reynolds 1999). In Ethiopia,
maize grain is mainly used for human consumption as a variety of food products
and also for local beverages. The use of maize for industrial processing is low but
increasing. It is predominantly grown for cash in western Oromia and food in most
other parts of the country.
Maize is one of the three leading global cereals that feed the world (Shiferaw et
al., 2011). Maize, together with rice and wheat, dominate human diets (Ignaciuk,
2014) and provide at least 30% of the food calories of more than 4.5 billion people
in 94 developing countries. Maize alone contributes over 20% of total calories in
human diets in 21 low-income countries, and over 30% in 12 countries that are
home to a total of more than 310 million people. Of the 22 countries in the world
where maize forms the highest percentage of calorie intake in the national diet, 16
are in Africa (Nuss and Tanumihardjo, 2011). Maize’s central role as a staple food
in Africa and Central America is comparable to that of rice or wheat in Asia, with
consumption rates being the highest in eastern and southern Africa (ESA). Maize
accounts for almost half of the calories and protein consumed in ESA, and one-
fifth of the calories and protein consumed in West Africa.
Maize (Zea mays L.) is one of the most important staple food crops in sub-Saharan
Africa (SSA), predominantly produced and consumed directly by the smallholder
farmers (Shiferaw et al., 2011). In Ethiopia, maize is one of the principal cereal
crops ranking first in total production and productivity, and second to tef
(Eragrostis tef) in area coverage. A total of 7.8 million tons of maize (31% of the
total cereal) was produced on 2.1 million hectares (21% of the total area planted
cereals) of land by nearly 11 million small households (31% of the total cereal) in
2016 (FAO, 2017). Maize is everything for the Ethiopian maize farmers. Three
fourth of the maize produced is consumed at the household level by the small-
scale producers themselves (CSA, 2017). The grain is consumed in different forms
of food; the stover is used as feed, fuel and construction material. Besides, it
serves as a major source of income and means of employment for tens of millions
of farming and business communities. Due to its wider significance in the country,
maize is one of the strategic field crops targeted to ensure food security in
Ethiopia.
Regardless of the importance of maize as a principal food crop, its average

productivity of both at National and Regional level is very low (3.6 t/ha) as
contrasted with the world average yield of 5.6 t/ha (CSA, 2016 and FAO, 2017). A
[174]
significant portion of this yield gap (very low productivity) is attributable to
various factors such as biotic and abiotic stresses. Mainly these factors include;
such as shortage of disease-resistant, adaptable and high yielding varieties, the
slow turnover of varieties tolerant or resistant to these stresses and appropriate
crop management practices, low-level use of improved management and other
inputs and post-harvest management, (Worku et al., 2012; Abate et al., 2017).
Some of the main abiotic factors affecting maize production and productivity are
drought, heat, soil acidity, frost and poor soil fertility mainly in N and P. Biotic
stresses hindering maize production in Ethiopia are diseases (e.g., Grey Leaf Spot,
Turcicum Leaf Blight, Common Leaf Rust, Maize Streak Virus, Maize Lethal
Necrosis), parasitic weeds (mainly Strigahermontica), and insect pests (such as the
maize stem borer, maize weevils and the newly emerged fall armyworm). In order
to increase the productivity of this crop, National Agricultural Research System
(NARS) has been making great efforts over the last twenty years to develop and
release large numbers of maize crop varieties and associated production
technologies for diversified agroecology of Ethiopia.
Maize is the second most widely cultivated crop in Ethiopia and is grown under
diverse agro-ecologies and socioeconomic conditions typically under rain-fed
production. It is extensively grown in most parts of the country, with the major
production areas concentrated at an altitude of 1000 to 2200 m.s.l. (MoARD,
2015). Further; during 2015/16 cropping season 2,111,518.23 ha of land was
covered by maize and about 75,508,354.11 quintals were harvested with the
average yield of 33.87 quintals per hectare at a national level (CSA, 2016).
Similarly, according to CSA (2016), the land covered by maize production in
Oromia in general was 1, 125,747. 85 hectares and about 39,538,593.74 quintals
were harvested with an average yield of 35.12 quintals per hectare at a regional
level. Moreover; maize area covered in the 2015/16 Maher production season
particularly in West Shewa, East Wollega and Ilu-Ababora Zones were
102,865.15, 126, 372.40 and 99,625.89 hectares; respectively (CSA, 2016).
Further, maize harvested in West Shewa, East Wollega and Ilu-Ababora Zones
were about 3, 579,519.87, 5,560,184.82 and 3,633,541.06 quintals; respectively
with an average yield of 34.80, 44.00, 36.47 quintals per hectare in West Shewa,
East Wollega and Ilu-Ababora Zones; respectively. Even though, most agro-
ecologies of West Shewa, East Wollega and Ilu-Ababora Zones are the potential
areas for maize production, the yield obtained by farming communities was below
the potential.
Despite the availability of several improved maize technologies generated by the

research system in Ethiopia over the last one decade, most of the farmers in the
Oromia region, West Shewa, East Wollega and Ilu-Ababora Zones specially Bako
Tibe, Ilu-Gelan, Gobu Sayo, Sibu Sire, Bonaya Boshe, Guto Gida and Ilu-Harar
mainly depend on the varieties that incur them higher costs than their profit and
[175]
return they earn from the output they produced. This is mainly due to shortage of
disease-resistant, adaptable and high yielding varieties, the slow turnover of
varieties tolerant or resistant to these stresses. Therefore, this activity was initiated
with objectives of demonstrating improved maize technologies (BH-546 and BH-
547) to verify the aforementioned problem and to evaluate, select and familiarize
the farming communities with low-cost maize varieties, high yielder and that can
help them to earn better profit and investment returns which in turn will facilitate
the adoption process and bridge the productivity gap.
Objectives
 To demonstrate and evaluate the performance of maize varieties with full
package under farmers’ condition;
 To evaluate the productivity and profitability of the technology under
farmers’ condition;
 To create awareness on the importance and availability of the technology
 To enhance linkage among the different stakeholders;
 To collect feedback from the participants for further research design and
the way;

Site and Farmers selection
This activity was conducted in some selected districts of East and West Shewa
zones. Selections of the districts were based on potentiality for maize production
and accessibility for supervision. Accordingly, Bako Tibe Gobu Sayo, Bonaya
Boshe Sibu Sire and Ilu-Gelan were selected based on the aforementioned criteria.
One potential PA from each district was selected based on accessibility and
potentiality for maize production. In each district, 4 experimental farmers were
selected. Development Agents and district experts were collaborating on-site and
farmer selection. The farmers were selected based on; willingness to be held as a
member, accessibility for the supervision of activities, good history of
compatibility with group dynamics, willingness to share innovations to other
farmers, availability and accessibility of sufficient land to accommodate the trials,
vicinity to roads to facilitate the chance of being visited by many farmers, good
history of handling experimental plots in the past or loyalty to entrust trials and
genuineness and transparency to explain the technology to others.
After sites and farmers’ were selected both theoretical and practical training were
given to farmers, Development agents and district experts. Training provided on
the following areas; such as, maize technology transfer approaches, production
management, breeding aspect, post-harvesting (seed quality). The training aimed
to create awareness of farmers’, Development agent and district experts on maize
technology.
[176]
The plots were properly plowed and made ready for planting ahead of the planting
date. Then after all the necessary inputs (seed, fertilizers) were delivered to the
farmers and planting was made on the farmers’ field by researchers, TAs as well
as farmers. Accordingly; three maize varieties BH-547, BH-546 and Limu as a
standard check were planted on adjacent plots of 10m*20m each. All the
necessary recommended agronomic practices were equally applied for all of the
plots and every field was supervised to check the status and to identify gaps. At
the maturity stage, a participatory variety evaluation platform was arranged that
attended by the experimenting farmers, neighboring farmers, researchers from
BARC and other stakeholders.
Both qualitative and quantitative data were collected. Data collected includes;
yield data, total number of farmers and other stakeholders’ participated in training
and field days, farmers’ perception and preferences and costs and income gained.
The data was analyzed using descriptive statistics such as mean, frequency
distribution, tables and percentages. Also, quantitative data collected were
subjected to SPSS software to analyze mean, standard deviation, t-test and
ANOVA table. Besides; farmers’ preferences were collected and analyzed using
score ranking techniques. And also this technique was employed to evaluate and
select best bet variety/ies and /or technology/gies and to rank their criteria and
parameters according to the real situation of the area. Further, gross margin
analysis was used as it is very useful in a situation where fixed capital forms a
negligible portion of the production. Thus; it is the difference between gross
income and the total variable cost (Mohammed et.al., 2016). Furthermore;
according to (Sumai et.al., 2000), the technology gap and technology index was
also calculated using the following formula.
Technology gap = Potential yield qt/ha – demonstration yield
Technology index = Potential yield - demonstration yield * 100
Potential yield
Result and Discussions

Participatory Variety Evaluation and Selection
At maturity, the varieties were then be evaluated based on the farmers’ selection
criteria. At this juncture, the farmers were assisted to record their evaluation
criteria, which then be ordered using pair-wise ranking technique. Each variety
was then be evaluated against the criteria ordered based on the weight attached to
each parameter. At the end of the evaluation process, the result of the evaluation
was displayed to the evaluators, and discussion was made on the way ahead. To
this end; FRG farmers scored each variety for individual traits considered
important by them and ranking of varieties was done on a scale of 1-5, 1 being
very poor and 5 being the highest score representing superiority. Maize yield,
early maturity, lodging tolerant, number of cob per plant, disease tolerant and
[177]
other traits were considered as the most selection criteria for each maize variety.
Based on the overall mean score the best-preferred variety/ies was/ were evaluated
and ranked. The variety/ies selected, accordingly, will be proposed for further
scaling up.
Following the procedure, the farmers set their criteria to evaluate, rank and select
the best variety/ies under their growing environment. Accordingly; against each
of the criteria and weight attached the farmers’ evaluated, ranked and selected the
best bet varieties from the demonstrated maize varieties. Therefore; BH-546 was
ranked and selected first by all the traits including yield then followed by BH-547
and finally Limmu was ranked thirdly against the criteria they already set. Hence;
based on the overall mean score the best preferred and selected variety/ies by
farmers’ at large; BH-546 and BH-547, will be proposed for further scaling up.
Below (Table 1) summarizes farmers score rank of the varieties across the
districts.
According to (table 1) above ranking and scoring of maize varieties across the
districts was done; accordingly, the highest score was recorded for BH-546
variety; 4.39, then by BH-547 variety; 3.6, and the lowest score was recorded for
standard check; Limu; 3.36. BH-546 variety was ranked as best followed by BH-
547 variety and finally, the least ranked variety was the standard check Limu.
Besides; except for its late-maturing nature of the variety and longer time for
baking, farmers selected BH-546 variety as a best because of its ability to tolerate
disease and pest, large cob size, number of cobs per plants, number of seeds per
cobs, grain size, grain color, taste, long shelf life quality of dough and
marketability. Furthermore, secondly selected BH-547 variety was almost similar
in the entire trait acquired with BH-546 variety except low yielder as compared
with BH-546. The least ranked and selected variety; the standard check (Limu),
was mainly subjected to open its cob (bare tip) after maturity which in turn highly
expose the variety to bird attack and rotting problem that may not a problem with
the others two varieties as they turn their cob down. Generally; comparable yield
could be obtained from all the three varieties but the differences mainly lie on the
other related traits listed for evaluation and selection of the best variety that suits
the need of the farmers and most preferred by the farmers at large.
[178]
Table 1: Score ranking of Maize varieties by farmers across the districts.
Variety Bako Tibe Gobu Sayo Bonaya Boshe Ilu Gelan Overall
Total Mean Rank Total Mean Rank Total Mean Rank Total Mean Rank Rank
Score Score Score Score Score Score Score Score
BH-546 44 3.67 1st 54 4.56 1st 57 4.89 1st 53 4.44 1st 1st
BH-547 42 3.50 2 nd 42 3.67 3rd 43 3.44 2 nd 46 3.78 2nd 2nd
Limu 34 2.83 3rd 46 3.83 2 nd 40 3.22 3 rd 43 3.56 3 rd 3rd
NB: selection criteria; 1= Lodging tolerant, 2= Disease tolerant 3=early maturity, 4=cob length, 5= number of cob/plant, 6= grain color, 7=grain size, 8= Bare tip;
9= high yielder; 10= Taste; 11= Less baking time and 12= Long shelf life quality
[179]
On-farm performance of the varieties
Despite the inevitable variability in performance between and even within
locations, yield performance of the varieties was still promising. The variability in
yield performance might have stemmed from differences in the status of soil
fertility, the difference in management (usage of recommended cultural practices
and others. Despite this fact, a yield of 8.92 tones/ha for BH-547, 9.11 tones/ha for
BH-546 and 8.68 tones/ha for Limmu, respectively was obtained. In line with this
below (Table 2) summarize the mean yield gained across the districts.
Table 2. Mean yield performances of Maize varieties in tones/ha
Name N Mean SD Min Max

Limmu 10 8.68 14.31 7.35 10.20
BH-546 10 9.11 17.61 7.37 10.56
BH-547 10 8.92 14.73 8.13 9.64
Source: Own data 2018
Besides; on-farm mean yield performances of maize varieties for each of the
locations (districts) were summarized and presented below (Table 3).
Accordingly; when compared yield wise except at Bako Tibe where Limmu
variety beat both BH-546 and BH-547, in all the other locations (districts) BH-546
variety gave the highest yield followed by BH-547 and finally Limmu variety (
Table 3).
Table 3: Mean yield of maize varieties across the districts
Districts Mean yield the varieties in tones/ha

BH-546 BH-547 Limmu
Bako Tibe 8.73 8.35 10.2
Boneya Boshe 7.14 8.13 7.35
Gobu Sayo 10.56 9.14 8.77
Ilu-Galan 10.51 9.64 9.85
Sibu Sire 9.41 9.43 8.67
Mean Yield 9.11 8.92 8.68
Source: Own data 2018
Result of the above (Table 3) also indicated that BH-546 and BH-547 performed
well and excel Limmu variety; respectively in all the locations except at Bako
Tibe where the only location Limmu gave better yield than the two varieties.
Table 4. Mean yield comparisons (t-test) of maize varieties at (P˂0.05) across the districts
Group 1 Group 2 Critical P -Value Significant

BH-546 Limmu 0.017 0.59 No
BH-547 Limmu 0.025 0.72 No
BH-546 BH-547 0.05 0.79 No
[180]
Statistically, the ANOVA table result summarized above (table 4) showed that as
there is no significant difference in mean yield performances among the varieties
at (P˂0.05). Therefore; statistically though there is no significant difference
among the varieties at (P˂0.05) based on yield advantage and farmers' preferences
BH-546 and BH-547 will be scaled up on large area and more farmers for the
coming years.
Yield Advantage
Calculating yield advantage of the varieties helps to show the extra benefit in
percentage that the farmers’ obtained from producing improved variety. Besides;
it helps to recommend based on the relative yield advantage over other varieties.
Accordingly; as indicated in (table 5) below BH-546 had higher yield advantage
over other varieties which is 4.94 % then by BH-547 2.72 % and could be
calculated using the underlying formula.
Yield advantage %= Yield of new variety - Yield of standard check X 100

Yield of standard check
Table 5. Yield advantage of recently released maize varieties over the standard check
Yield advantage over the

standard check
Demonstrated Varieties Yield obtained (tones/ha) (Limmu)
BH-546 9.11 4.94 %
BH-547 8.92 2.72 %
Limmu 8.68 -
Cost-Benefit Analysis
In terms of profitability and returns that could be gained from each of the
varieties, the economic analysis result of the study was summarized and presented
below (Table 6). Accordingly; the highest profit and returns were gained from
using BH-546 then followed by BH-547 and finally from Limu variety. This
means that a total of 50,105 ETB (Fifty thousand and one hundred five Birr) profit
can be obtained from a hectare of land investment on BH-546 variety. Further,
from the study result the highest returns to investment were gained from BH-546
variety; 2.7 Birr, next from BH-547 variety; 2.63 Birr and finally from Limu
variety; 2.04 Birr. Therefore; from this result, it can be concluded that BH-546
variety is high yielder than BH-547 and Limmu varieties using improved
variety/ies seeds of BH-546 and/or BH-547 were economically profitable than
Limmu. This is mainly because a cost that the farmers’ incur for purchase of BH-
546 and /or BH-547 is very low as compared to costs for seed procurement of
Limmu variety that may expose the farmers’ to the extraordinary cost of up to
two-fold for purchase of the seed. But also though statistically there is no
significant difference on mean yield gained among the varieties there is a reality
on the ground that using either of the varieties and/or both BH-546 and BH-547
[181]
will be encouraged or recommended as they suit farmers preference (need), have
better yield advantage, better profit and returns as well as incur low cost for seed
than Limmu. Eventually; the ultimate goal that we intended for is that
recommending any technologies/ varieties/ practices that can reduce the cost of
production that the farmers incur and maximizing the profit that the farmers could
fetch from the technologies/ varieties/ practices they employed. The cost-benefit
analysis result of the study for each of maize varieties is summarized and
presented in underlying (Table 6).
Table 6. Cost-Benefit analysis of maize varieties demonstration across the districts
Yield Fertilizer Seed Labor

(tones Price Cost Cost Cost TR Return to
Variety /ha) (Birr/Qt) Birr/Q) (Birr/Q) in Birr TVC (P*Q) GM Investment
BH-546 91.1 550 3150 2850 7500 13500 50105 36605 2.71
BH-547 89.2 550 3150 2850 7500 13500 49060 35560 2.63
Limmu 86.8 550 3150 5040 7500 15690 47740 32050 2.04
Injera was also prepared from the three varieties; namely, BH-546, BH-547 and
Limu to evaluate, rank and select the best variety that meets their criteria.
Accordingly; the farmers then evaluate, rank the varieties following the criteria
they set. To this end, the criteria they set were baking time required, shelf-life
quality, sweetness, softness and “injera” color (attractiveness). After they set the
criteria they were invited to taste the “injera” prepared from the three varieties
then by to rank from the varieties. Hence, Limu variety takes a short time for
baking followed by BH-547 and finally BH-546. This is to mean that BH-546
requires a long time for baking to be “injera” and also as the farmers replied
taking a long time for baking indicates that this variety had long shelf quality than
Limu that require a short time for baking. And also Limu variety cannot maintain
and retain its quality for a long period as the other varieties. Moreover, as to the
sweetness of the varieties, BH-546 was ranked first then by BH-547 and Limu
lastly. Furthermore, the farmers also ranked softness and attractiveness of the
varieties with this BH-546 was ranked first against these criteria then by BH-547
and lastly Limu. Lastly; the farmers concluded that in all its traits BH-546 was the
best and preferred variety to prepared as an “injera” from it so this necessitates for
wider scaling up/out of the variety in a wider area and on a large number of
farmers’ fields for the coming years.
Table 7: Maize food processing and utilization of the varieties against their score rank
Criteria Maize varieties Score Rank (scale 1-5)

BH-547 Limu BH-546
Minimum Baking Time 4 5 3
Long shelf life quality 4 3 5
Sweetness 4 3 5
Softness 4 3 5
Attractiveness of the “Injera” 4 3 5
[182]
The technology gap indicates that the gap in the demonstration yield over potential
yield. The observed technology gap is attributed to dissimilarities in soil fertility,
acidity, rainfall and other natural calamities (Dhaka et. al., 2010). According to
Dhaka et.al., 2010 its contribution is to narrow down the gap between the yields of
different varieties and to provide location-specific recommendations. The yield
gaps can be further categorized into the technology index which is used to show
the feasibility of the variety at the farmer’s field. The lower the values of
technology index the more the feasibility of the varieties. To this end, the
technology gap and index of demonstrated varieties (BH-546, BH-547 and
Limmu) were calculated using the underlying formulas and presented in the below
table.
Technology gap = Potential yield qt/ha – demonstration yield
Technology index = Potential yield- demonstration yield * 100

Potential yield
Accordingly; from the result gained and summarized below (Table 8) on yield gap
observed of the demonstrated maize varieties in all the varieties there was no yield
reduction observed from the potential the varieties inherited. Accordingly; all the
varieties retained and expressed the potential they acquired while demonstrated on
the farmers’ field. Moreover; similarly the values of technology index observed
for the varieties demonstrated were very low; even positive, that these varieties are
very feasible for the farmers to be used
Table 8. Technology gap and technology index for demonstrated maize varieties
Maize Varieties
Parameter
BH-546 BH-547 Limmu
Yield gap (qt/ha) (-) 1.11 (-) 0.92 (-) 0.8
Technology index (%) (-) 13.86 (-)11.46 (-) 2.12
During participatory variety evaluation and selection, farmers’ set their criteria
under their real situation existing. Accordingly; farmers evaluated, ranked and
selected BH-546 first then BH-547 second and Limu thirdly against the criteria
they set and weight attached for each of the criteria listed. Hence; farmers
preferred BH-546 because it is a high yielder, disease tolerant, grain color, number
of cob per plant, cob length, taste, long shelf life quality of dough and grain
weight as compared with others. Besides, secondly selected BH-547 except late-
maturing nature of the variety, comparable yield with BH-546 variety can be
obtained. Even the standard check; Limu the major problem associated with this
variety mainly was the cost of production the farmers’ incurred, very short shelf
life quality of dough (maximum of 3 days) and opening the cob (bare tip) after
maturity that in turn expose to bird attack and rotting problem due to rain whereas
this was not a problem on BH-546 and BH-547 as they turn down their cob after
maturity; otherwise, comparable yield can be gained from Limu variety.
[183]
Therefore; since the farmers’ evaluated, ranked and selected the varieties one after
the other and accordingly BH-546 was ranked and selected firstly, and the other
two varieties (as an option) will be disseminated on a large area and for more
number of farmers’ where the activity was carried out and others similar agro-
ecologies.
Conclusion and Recommendations

The activity was carried out in selected zones of West Shewa, East Wollega and
Bunno Bedelle zones of Western Oromia. These districts were Bako Tibe, Ilu-
Gelan, Gobu Sayo, Bonaya Boshe and Sibu Sire. In each district, 4 experimental
farmers were selected and participated in the activity implementation. Three
varieties of maize BH-547, BH-546 and Limu as a standard check were used along
with all recommended full packages. Besides; a plot of land 10*20 m2 was used
for each of the variety under consideration. At the end, each of the variety was
evaluated and selected by participant farmers as well as surrounding farmers who
were not part of the activity. Accordingly, farmers score and ranked different
traits/criteria that were used to evaluate, rank and select the best bet variety for
their area. To this end in all other traits except yield, late maturing and lodging
farmers ranked BH-547 then by Limu and BH-546. But from yield and economic
point of view BH-546 was ranked first followed by BH-547. And also from profit
and investment return the highest profit and investment was from BH-546 then
followed by BH-547 and the least economic profit and investment return was from
Limu as compared to the other two varieties. Therefore, BH-546 and BH-547 will
be scaled up/out based on farmers’ preference, yield advantage, highest revenue
and high investment return from the varieties.
References
Abate T, Fisher M, Abdoulaye T, Kassie GT, Lunduka R, Marenya P, Asnake W. 2017.
Characteristics of maize cultivars in Africa: How modern are they and how many do
smallholder farmers grow? Agriculture & Food Security 6(1):30.
Central Statistical Agency. CSA. 2016. The Federal Democratic Republic of Ethiopia. Central
Statistical Agency Agricultural Sample Survey 2015/2016 (2008 E.C.): Report on Area and
Production of Major Crops (Private Peasant Holdings, Meher Season), Volume I. Addis
Ababa, Ethiopia.
Central Statistical Agency. CSA. 2017. Centeral Statistical Agency Agricultural Sample Survey,
Report on area and production of major crops. www.csa.gov.et. Accessed 20 December 2017.
Food and Agriculture Organization. FAO. 1992. Maize in human nutrition. Food and Agriculture
Organization of the United Nations. Rome, 192 pp
<http://www.fao.org/docrep/TO395E/TO395E00.htm>.
Food and Agriculture Organization. FAO. 2004. Rice and narrowing the yield gap. Retrieved from
http://www.fao.org/rice2004/en/f-sheet/factsheet5.pdf
Food and Agriculture Organization Corporate Statistical Database (FAOSTAT). 2017. Statistical
databases and data-sets of the Food and Agriculture Organization of the United Nations.
http://faostat.fao.org/default.aspx. Accessed 18 Dec 2017.
[184]
Ignaciuk, A., and Mason-D'Croz, D. 2014. “Modelling Adaptation to Climate Change in
Agriculture”, OECD Food, Agriculture and Fisheries Papers, No. 70, OECD Publishing.
McCann, JC. 2005. Maize and grace: Africa’s encounter with a new world crop, 1500–2000.
Cambridge: Harvard University Press.
Ministry of Agriculture and Natural Resource (MoARD). 2015. Variety Registration Book. Addis
Ababa, Ethiopia.
Nuss E. T., and Tanumihardjo, SA. 2011. “Quality Protein Maize for Africa: Closing the Protein
Inadequacy Gap in Vulnerable Populations,” Advances in Nutrition 2: 217-222.
Reynolds L. 1999a. Maize for food, feed and fertiliser. A literature review on multiple uses of
maize in mixed farming systems in east and southern Africa. Consultant’s Report to ILRI and
SLP (System-wide Livestock Programme). ILRI (International Livestock Research Institute),
Nairobi, Kenya. 41 pp.
Shiferaw B, Prasanna BM, Hellin J, Bänziger M. 2011. Crops that feed the world 6. Past successes
and future challenges to the role played by maize in global food security. Food Security
3(3):307-327.
Worku M, Legesse W, Berhanu T, Girma D, Girum A,Wende A, Tolera K, Gezahegn B, Dagne W,
Solomon A. 2012. Status and future direction of maize research and production in Ethiopia. In
M. Worku, Twumasi-Afriyie, S., Wolde, L., Tadesse, B., Demisie G., Bogale, G., Wegary, D.
and Prasanna, B.M. (Ed.), Meeting the Challenges of Global Climate Change and Food
Security through Innovative Maize Research, Addis Ababa pp. 17.
[185]
[186]
Green Super Rice (Oryza sativa L.) Variety
Development for the Lowland Rice
Ecosystem in Ethiopia
Abebaw Dessie1, Taddess Lakew1, Zelalem Zewdu1, Mulugeta Atnaf1, Assaye Berie1,
Haile Gebriel kinfie2, Mulugeta Bitew3, and Hailemariam Solomon4
1
Fogera National Rice Research and Training Center/EIAR, Ethiopia; 2Shire Agricultural Research
Center/Tigrai Institute of Agricultural Research Center, Ethiopia: 3Pawe Agricultural Research Center/EIAR,
Ethiopia; 4Assosa Agricultural Research Center/EIAR, Ethiopia
Corresponding Author: Abebaw Dessie
Abstract
Two different sets of field experiments were conducted from 2016 to 2018 main
cropping seasons. In experiment I, 15 GSR genotypes including two checks were
evaluated in Fogera and Shire-Maitsebri to select high yielding, cold tolerant, disease
resistance and early maturing GSR rice varieties for the high altitude of the lowland
production system. In experiment II, a total of 20 GSR genotypes including one check
were evaluated in Pawe and Assosa to select high yielding and disease resistance in
the lowland ecosystems of Ethiopia. In both sets of experiments, the trials were laid
out in randomized complete block design with three replications. The combined
analysis of variance in experiment I revealed a significant difference (P≤0.01). in
grain yield, days to maturity, days to heading, panicle length, filled grain per panicle,
plant height and thousand grains weight Three genotypes (G2, G6, and G4) showed a
significant difference than the standard checks on grain yield and gave grain yield
advantage of 32.6 %, 27.9 % and 22.3 %, respectively. GGE-biplot analysis revealed
that G2 and G6 are high yielding and most stable among tested genotypes in both
environments and these genotypes were proposed for national variety release. In
Experiment II, the mean grain yield of the 20 lowland green super rice genotypes
ranged from 2730.30kgha-1(G7) to 3683.40 kg ha-1 (G13). The combined analysis
over the environments revealed that no genotype significantly gave a higher yield
over the standard check. There was no genotype for wider adaptability. However, the
separate location analysis revealed that at Assosa, one genotype (G13) and at Pawe
two genotypes (G14 and G1) gave significantly higher grain yield than the standard
check with a yield advantage of 26.7%, 21.7% and 20.6%, respectively.
Keywords: GSR, cold, stability, GGE bi-plot, rice
Introduction
Rice (Oryza sativa L.) which was originated from the tropics and subtropics is
widely cultivated in diverse environments. The tremendous growth of the human
population worldwide has increased the demand for rice production (Khush,
2001), requiring an improvement of 50% by the year 2025 (Liang et al., 2010).
Due to its origin in tropical and subtropical regions, rice is more sensitive to cold
stress than other cereals crops such as Wheat and Barley. Therefore, the
[187]
production of rice is severely limited by cold stress in temperate areas (Africa
Rice Center, 2018). Cold stress is the major factor affecting rice growth,
productivity, its distribution worldwide (Moraes de Freitasetal., 2016). Production
of rice is affected primarily due to its vulnerability to cold stress at the seedling
stage, as well as the reproductive stage leading to spikelet sterility.
In Africa, rice also constantly increasing as a staple food and there has been
increasing demand in Africa in the past three decades from 1999-2018 (Zhang et
al., 2014); however, these demands have not been commensurate with the total
production and most of the African countries are a net importer of milled rice,
which costs 6.4 billion USD annually (Zhang et al., 2014). Although rice is
recently introduced to Ethiopia, it is recognizing its importance as a food security
crop and a source of income and employment opportunity
(http://www.africarice.org/publications/ricetoday/Ethiopia%E2%80%99s%20mill
ennium%20crop.pdf). The production, productivity, and consumption of rice in
Ethiopia is constantly increasing in the country (CSA, 2017). Ethiopia’s
geography is noticeable by immense depressions and mountains. Consequently,
vast arable lands are located at high altitudes more than 2000 meters above sea
level. Rice can grow in wide agro-climatic zones; however, low-temperature
stresses are serious challenges for rice farmers at high elevations in the tropics.
Lack of cold-tolerant rice varieties in the high lands of Ethiopia is the main
constraint for the promotion of rice. Therefore, the main objective of this study
was to evaluate the performance and stability of introduced lowland green super
rice genotypes for their wider or specific adaption in the North-West Ethiopia and
similar agro-ecologies.

Two sets of low land Green Super Rice (GSR) experiments were conducted.
Experiment I
A total of 16 genotypes were introduced from China/CAASin 2014 with cold
tolerant and high grain yield genotypes (Table 1), Laboratory and field
quarantine/observation for both sets of experiment I and II were conducted in
Holeta Agricultural Research Center and Andasa, respectively in 2014 main
cropping season. Both sets of preliminary variety trials were conducted in 2015 in
woreta on station. The multi-environment trials for sets of experiments were
conducted for 3 years from 2016 to 2018. Fogera, Shire- Maitsebri,
Gondar/Demba and Jima were the testing locations for the first experiment.
However, the complete data were generated from Fogera and Shire-Maitsebri for
the first experiment. The locations where the trials conducted differ in soil types,
annual rainfall, altitude and annual temperature (Tables 3 and 4). For multi-
[188]
environment trials, a total of 15 genotypes including two checks were used (Table
5).
Experiment II
A total of 40 GSR genotypes were introduced from IRRI in 2014 for the second
experiment (Table 2). Thirty-six (36) genotypes including one check were planted
as preliminary variety trial in a place where temperature and humidity are high
(Pawe). The plot size and type of design for observation and preliminary variety
trial were not standard and vary based on the amount of seed and experimental
area. For multi-environment trials, Pawe, Assosa, Tepi, and Gambela were used
for the second experiment and 20 genotypes including one check were used for the
second experiment (Table 6). However, data were generated only from Pawe and
Assosa for the second experiment.
Design and trial management

The trials were laid out in a randomized complete block design with three
replications for both sets of experiments in all locations. Each plot had a size of
7.5 m2 (Six rows with 5 m long x0.25 m row spacing). A seed rate of 60 kg ha-1
was used with direct seeding methods in a row was applied. Fertilizer (UREA and
DAP) were applied based on each location recommendation. All DAP was applied
at the time of sowing. For UREA, split application was applied; 1/3 at sowing, 1/3
at active tillering and the remaining 1/3 during panicle initiations.
Phenological and agronomic data were collected including days to heading, days
to maturity, plant height (cm), Panicle length (cm), number of effective tillers per
plant, number of filled grain per panicle, thousand-grain weight (g), phenotypic
acceptability, grain yield per plot (g). The data were subjected to the GLM
procedure for analysis of variance using SAS software V.9.0. And Genotype x
environment and stability analysis were done by using Genstat 18th edition
software.
Table 1: List of genotypes introduced from China/CAAS for experiment I
No. Genotype Seed Source No. Genotype Seed Source

1 Yungeng 44 CASS/ China 10 P-28 CASS/ China
5 Fengdao 23 CASS/ China 14 Songgeng15 CASS/ China
6 Songgeng20 CASS/ China 15 Li Jing 9 China
7 KB-2 CASS/ China 16 Li jing 11 China
8 Songgeng3 CASS/ China 17 Ediget (check 1) Fogera/FNRRTC
9 Songgeng9 CASS/ China 18 X-Jigina (Ceck 2) Fogera/FNRRTC
[189]
Table 2: List of genotypes introduced from IRRI and used for observation and preliminary variety trial used for experiment
II
No Genotype Seed Source No Genotype Seed Source
1 GSR IR1-5-D1-D1 IRRI 22 GSR IR1-8-Y7-D2-S1 IRRI
2 GSR IR1-5-D7-Y3-S1 IRRI 23 GSR IR1-9-D12-D1-SU1 IRRI
3 GSR IR1-5-D20-D2-D1 IRRI 24 GSR IR1-11-D7-S1-S1 IRRI
4 GSR IR1-5-D20-D3-Y2 IRRI 25 GSR IR1-11-Y10-D3-Y3 IRRI
5 GSR IR1-5-S8-D2-S1 IRRI 26 GSR IR1-12-D10-S1-D1 IRRI
6 GSR IR1-5-S10-D1-D1 IRRI 27 GSR IR1-12-S2-Y3-Y1 IRRI
7 GSR IR1-5-S10-D3-Y2 IRRI 28 GSR IR1-12-S2-Y3-Y2 IRRI
8 GSR IR1-5-S12-D3-Y2 IRRI 29 GSR IR1-12-S8-Y1-S1 IRRI
9 GSR IR1-5-S14-S2-Y1 IRRI 30 GSR IR1-12-S8-Y1-Y2 IRRI
10 GSR IR1-5-S14-S2-Y2 IRRI 31 GSR IR1-12-Y4-D1-Y1 IRRI
11 GSR IR1-5-Y3-S2-SU1 IRRI 32 GSR IR1-12-Y4-D1-Y2 IRRI
12 GSR IR1-5-Y3-Y1-D1 IRRI 33 GSR IR1-12-Y4-D1-Y3 IRRI
13 GSR IR1-5-Y4-S1-Y1 IRRI 34 GSR IR1-12-Y4-Y1-D1 IRRI
14 GSR IR1-5-Y7-Y2-SU1 IRRI 35 GSR IR1-15-D4-D1-Y1 IRRI
15 GSR IR1-8-S6-S3-S1 IRRI 36 GSR IR1-15-D7-S4-S1 IRRI
16 GSR IR1-8-S6-S3-Y1 IRRI 37 GSR IR1-17-D6-Y1-D1 IRRI
17 GSR IR1-8-S6-S3-Y2 IRRI 38 GSR IR1-17-Y16-Y3-S1 IRRI
18 GSR IR1-8-S9-D2-Y2 IRRI 39 GSR IR1-17-Y16-Y3-Y1 IRRI
19 GSR IR1-8-S12-Y2-D1 IRRI 40 GSR IR1-17-Y16-Y3-Y2 IRRI
20 GSR IR1-8-S14-S1-SU1 IRRI 41 Ediget (Check) Fogera/NRRTC
21 GSR IR1-8-S14-S3-Y2 IRRI 42 Gumera (Check) Fogera/NRRTC
Table 3: Description of study environment for Lowland GSR experiment for cold-tolerant for experiment I
Altitude Annual rainfall Temperature 0C (Mean)

Location (m) Latitude Longitude (mm) Max Min
Fogera/Woreta 1810 11058’N 37041’ E 1300 27.9 11.5
Shire/Mai-tsebri 1350 13005’ N 38008’ E 1296 36.0 15.0
Gondar/Dembiya NA NA NA NA NA NA
Jimma NA NA NA NA NA NA
Table 4: Description of study environment for Lowland GSR experiment for experiment II
Altitude Annual rainfall Temperature 0C (Mean)

Location (m) Latitude Longitude (mm) Max Min
Pawe 1050 1109’ N 3603’ E 1457 32.8 17.2
Assosa 1590 10003’N 34059’E 1120 28.0 14.5
Tepi NA NA NA NA NA NA
Gambella NA NA NA NA NA NA
Table 5: List of genotypes used for national variety trial for cold in experiment I
No. Genotype Seed Source No. Genotype Seed Source

4 Yungeng 38 CASS/ China 12 Li Jing 9 CASS/ China
5 Fengdao 23 CASS/ China 13 Li jing 11 CASS/ China
6 KB-2 CASS/ China 14 Check1(Ediget) Fogera/NRRTC
7 Songgeng9 CASS/ China 15 Check (KOMBOKA) Fogera/NRRTC
8 P-28 CASS/ China
[190]
Table 6: List of genotypes used for national variety trial in experient II
No Genotype Seed Source No Genotype Seed Source
1 GSR IR1-17-Y16-Y3-Y2 IRRI 11 GSR IR1-5-Y3-S2-SU1 IRRI
2 GSR IR1-15-D4-D1-Y1 IRRI 12 GSR IR1-11-Y10-D3-Y3 IRRI
3 GSR IR1-5-D1-D1 IRRI 13 GSR IR1-12-D10-S1-D1 IRRI
4 GSR IR1-12-Y4-Y1-D1 IRRI 14 GSR IR1-12-Y4-D1-Y2 IRRI
5 GSR IR1-8-S9-D2-Y2 IRRI 15 GSR IR1-12-S8-Y1-Y2 IRRI
6 GSR IR1-12-S2-Y3-Y2 IRRI 16 GSR IR1-5-S10-D1-D1 IRRI
7 GSR IR1-5-D20-D2-D1 IRRI 17 GSR IR1-8-S6-S3-S1 IRRI
8 GSR IR1-5-S10-D3-Y2 IRRI 18 GSR IR1-5-S12-D3-Y2 IRRI
9 GSR IR1-12-S8-Y1-S1 IRRI 19 GSR IR1-5-S8-D2-S1 IRRI
10 GSR IR1-8-S14-S1-SU1 IRRI 20 KOMBOKA (Check)- Fogera/NRTTC

Observation and preliminary variety trial
All introduced genotypes for both sets of experiments were free from any
quarantine pests and disease during quarantine and observation stages. The
preliminary variety trial of the first experiment showed good performance and
significant variation among tested genotypes (Figure 1). Whereas the genotypes in
the second experiment revealed stunted growth and failed to seed set at Wereta
however those genotypes showed a significant difference in grain yield (Figure 2)
and other traits like the number of tillers and plant height in Pawe on station in
2015/16. Following this result, the national variety trials were designed and
conducted.
Figure 1 Grain yield performance of 18 lowland GSR cold-tolerant genotypes at Wereta in 2015/16
Figure 2. Grain yield performance of 36 lowland GSR genotypes at Pawe in 2015/16
[191]
Experiment I
The combined analysis of variance for grain yield, days to maturity, days to
heading, panicle length, and filled grain per panicle, plant height and thousand
grains weight showed significant difference (P≤0.01). The analysis of the
environment effect also revealed a significant difference (P≤0.01) for grain yield
and other agronomic characters. The genotype x environment interaction effect
was significant for all traits (P≤0.01). The three-way interaction analysis for
genotypes x location x years were also showed significant variation (P≤0.01) for
grain yield and other agronomic characters (Table 7). The study revealed that
genotypes responded differently for grain yield and other agronomic characters to
different environments. This pointed out the advantage of executing multi-location
trials to explore the response of genotypes for their specific or wider adaptability.
The significant interaction difference of the three-way interaction of genotype x

location x years revealed that the possibility of getting genotypes that can be
adapted widely/or specifically. The mean grain yield of the 15lowland green
super rice genotypes ranged from 3108.4 kg ha-1 (G13) to 4840.3 kg ha-1 (G2).
Compared to the standard checks (G14 and G15), ten genotypes (G2, G6, G4, G1,
G12, G5, G8, G9, G3 and G10) gave higher yield than the checks. However only
the three genotypes viz. G2, G6and G4 showed a significant difference from the
standard checks fo grain yield with grain yield advantage of 32.6 %, 27.9 %, and
22.3 %, respectively, and other agronomic traits. Moreover, GGE-biplot analysis
revealed that G2 and G6 are high yielding and most stable among tested genotypes
in both environments (Fig. 3). There were no grain sterility problems observed in
both genotypes (G2 &G6) and have better panicle exertion which can fit the cold-
tolerant characteristics. Following this result, these two genotypes (G2 and G6)
were proposed for national variety release.
[192]
Table 7. Mean grain yield and other yield-related parameters of lowland green super rice genotypes for cold-tolerant at
Fogera and Shire-Mai-Tsebrifrom2016 to 2018.
Genotype Code DTH DTM PL PH FTP FGP PHA GY

Yungeng 44 G1 95.3 132.7 17.6 81.4 10.5 114.2 1.2 4233.1
Yungeng 31 G2 91.3 127.4 19.2 87.4 10.6 123.4 1.0 4840.3
Yungeng 45 G3 94.3 134.3 18.5 78.7 11.5 116.4 1.3 3830.6
Yungeng 38 G4 91.8 127.9 19.6 85.5 10.2 124.2 1.2 4464.7
Fengdao 23 G5 92.4 151.1 16.7 74.5 16.6 91.7 1.7 3981.7
KB-2 G6 90.4 127.2 17.4 78.2 11.0 104.3 1.2 4667.8
Songgeng9 G7 84.7 123.8 17.7 75.7 11.7 95.6 2.2 3272.8
P-28 G8 89.4 127.9 16.7 74.6 11.8 99.1 1.5 3898.7
P-37 G9 87.4 122.7 16.7 73.6 12.8 101.7 1.0 3863.8
P-38 G10 100.4 137.4 19.0 80.9 12.4 101.7 2.0 3734.3
P-39 G11 88.1 121.3 17.2 77.4 13.0 96.3 1.7 3309.2
Li Jing 9 G12 86.1 122.8 19.7 86.6 10.8 119.0 1.0 4079.5
Li jing 11 G13 103.6 138.8 16.9 63.0 12.7 89.3 3.0 3108.4
Check -Ediget G14 89.1 120.2 18.8 85.1 10.9 96.1 1.3 3649.5
Check - Komboka G15 104.6 116.6 16.8 60.4 13.4 104.3 1.7 3373.2
Mean 92.6 128.9 17.9 77.5 12.0 105.1 1.5 3883.0
CV (%) 5.7 14.3 7.2 5.9 22.9 10.8 34 23.1
Gen. (G) *** *** *** *** *** *** *** ***
Loc. (L) *** *** *** *** *** *** - ***
Year (Y) *** NS *** *** *** *** - ***
G*L *** *** *** *** *** NS - ***
G*L*Y *** *** *** *** *** *** - ***
Note: *, **, and *** refers to significant at 5%, 1% and 0.1% level, NS=non -significant, CV= coefficient of variation, G*L=
genotype by location, G*L*Y= genotype by location by year, DH= days to 50% heading, DM= days to 85% maturity, FGP=
filled grains/panicle, PH= plant height (cm), PL= panicle length (cm), PHA=Phenotypic acceptability and Gy= grain yield
(kg/ha).
Figure 3.The average-environment coordination (AEC) view of ranking rice genotypes relative to an ideal genotype
(center of the concentric circle).
[193]
The candidate genotypes G2 and G6 which are closer to the ideal genotype
(designated by the arrow) are desirable (stable and high yielder).
Experiment 2
The combined analysis of variance for grain yield, days to maturity, days to
heading, plant height and thousand grains weight showed a highly significant
difference (P≤0.01) and for panicle length significantly different (P≤0.05) among
the genotypes. There was no significant difference observed among tested
genotypes for filled grain per panicle. The combined analysis of the environment
effect also revealed a significant difference (P≤0.01) for grain yield and other
agronomic characters. Except for panicle length and filled grain per panicle, the
three-way interaction of genotypes x location x year were showed significant
variation (P≤0.01) for grain yield and other agronomic characters (Table 8). The
significant interaction difference of the three-way interaction of genotype x
location x years revealed that the possibility of getting genotypes which can be
adapted widely/or specifically to all or specific environment. The mean grain yield
for the 20 lowland green super rice genotypes ranged from 2730.30 kg ha-1 (G7) to
3683.40 kg ha-1 (G13) with the mean grain yield of 3310.04 kg ha-1. Compared to
the standard check (G20), seven genotypes (G13, G14, G1, G12, G15, G4, G8,)
gave a higher yield than the check. However, there is no genotype showed a
significant difference from the standard check on grain yield. This revealed that
the tested genotypes failed to give significantly higher grain yield than the
standard check (KOMBOKA) and there is no genotype for wider adaptability.
However, the separate location combined analysis revealed that there is a
significant difference in grain yield and other agronomic characteristics than the
standard check.
At Assosa, the three years combined analysis for grain yield ranged from1872.5
kg ha-1 to 3809.7 kg ha-1 and mean grain yield of 2836.23 kg ha-1. Seven
genotypes (G13, G12, G6, G4, G3, G15 and G8) gave higher grain yield than the
standard check (Table 9). However, only one genotype (G13) gave statistically
significantly higher grain yield than the check and with grain yield advantage of
26.7% as compared to the check.
[194]
Table 8. Mean grain yield and other yield-related parameters of 20 lowland green super rice genotypes at Assosa and Pawefrom2016 to 2018.
Genotype Code DTH DTM PL PH FTP FGP GY
GSR IR1-17-Y16-Y3-Y2 1 104.11 139.22 19.62 77.75 6.40 87.71 3653.70
GSR IR1-15-D4-D1-Y1 2 104.89 140.22 19.93 78.21 6.08 96.84 2929.70
GSR IR1-5-D1-D1 3 105.61 142.72 20.53 78.36 6.57 96.01 3347.10
GSR IR1-12-Y4-Y1-D1 4 105.56 140.17 19.81 78.48 6.37 91.57 3436.40
GSR IR1-8-S9-D2-Y2 5 107.94 142.17 20.41 79.26 6.40 90.35 3307.00
GSR IR1-12-S2-Y3-Y2 6 106.50 142.50 21.23 79.39 5.91 91.98 3351.80
GSR IR1-5-D20-D2-D1 7 107.11 142.00 19.47 79.43 6.71 95.68 2730.30
GSR IR1-5-S10-D3-Y2 8 110.78 144.78 19.95 80.51 5.98 93.27 3409.20
GSR IR1-12-S8-Y1-S1 9 104.17 137.72 20.02 80.87 6.47 86.69 3089.50
GSR IR1-8-S14-S1-SU1 10 104.56 140.00 21.70 81.31 5.68 100.6 3113.80
GSR IR1-5-Y3-S2-SU1 11 106.28 141.67 21.51 81.43 5.92 89.61 3136.40
GSR IR1-11-Y10-D3-Y3 12 103.78 138.61 20.48 81.72 6.12 91.29 3603.30
GSR IR1-12-D10-S1-D1 13 108.44 142.22 19.89 81.91 5.72 90.73 3683.40
GSR IR1-12-Y4-D1-Y2 14 105.72 139.61 19.59 82.22 6.41 91.97 3644.50
GSR IR1-12-S8-Y1-Y2 15 106.89 142.33 19.79 83.28 5.71 92.67 3578.90
GSR IR1-5-S10-D1-D1 16 108.17 142.50 20.06 83.44 5.78 100.8 3211.20
GSR IR1-8-S6-S3-S1 17 105.44 141.00 19.38 83.57 6.19 87.48 3113.90
GSR IR1-5-S12-D3-Y2 18 105.39 141.06 19.69 83.89 6.99 101.9 3284.80
GSR IR1-5-S8-D2-S1 19 104.22 140.28 19.82 85.62 5.42 98.91 3205.00
KOMBOKA (Check) 20 103.89 139.00 20.90 86.09 5.97 97.80 3371.00
Mean 106.00 140.99 20.19 81.34 6.14 93.70 3310.04
CV (%) 3.30 2.50 10.70 7.10 25.60 14.60 18.90
LSD (5 %) 2.32 2.36 1.42 3.81 1.03 8.99 411.09
Gen. (G) *** *** * *** NS *** ***
Loc. (L) *** *** *** *** *** *** ***
Year (Y) *** *** NS *** *** *** ***
G*L *** *** ** *** NS * ***
G*L*Y *** *** NS *** *** NS ***
Note: *, **, and *** refers to significant at 5%, 1% and 0.1% level, NS=non -significant, CV= coefficient of variation, LSD= list significance
different, G*L= genotype by environment, G*L*Y= genotype by locationby year, DH= days to 50% heading, DM= days to 85% maturity,
FGP= filled grains/panicle, PH= plant height (cm), PL= panicle length (cm) and GY= grain yield (kg/ha)
Table 9. Mean grain yield and other yield-related parameters of 20 lowland green super rice genotypes at Assosafrom 2016 to 2018.
Genotype Code DTH DTM PL PH FTP FGP TGW GY
GSR IR1-17-Y16-Y3-Y2 1 115.9 153 19.6 83.94 4.16 74.22 22.48 2804
GSR IR1-15-D4-D1-Y1 2 115.1 152.6 19.9 85.91 4.44 86.07 23.82 2195
GSR IR1-5-D1-D1 3 114.7 155.7 21 91.96 4.73 84.2 24.42 3266
GSR IR1-12-Y4-Y1-D1 4 116.8 153.6 20 89.81 5 82.56 23.74 3240
GSR IR1-8-S9-D2-Y2 5 115.2 153.1 22.1 91.27 4.56 82.87 24.82 2803
GSR IR1-12-S2-Y3-Y2 6 115.7 153.2 23.5 85.39 4.78 81.13 23.37 3287
GSR IR1-5-D20-D2-D1 7 119.1 155.6 20.7 90.46 4.44 85.29 52.21 1873
GSR IR1-5-S10-D3-Y2 8 119.6 156.2 20.8 93.56 4.38 81.91 52 3143
GSR IR1-12-S8-Y1-S1 9 108.2 143.7 22.2 93.49 5.16 72.2 26.53 2775
GSR IR1-8-S14-S1-SU1 10 113.1 150.3 23.6 100.7 4.91 94.71 27.38 2698
GSR IR1-5-Y3-S2-SU1 11 116.2 154.8 22.3 93.67 4.02 76.29 25.03 2575
GSR IR1-11-Y10-D3-Y3 12 118 154 20.4 100.3 4.18 83.73 22.93 3440
GSR IR1-12-D10-S1-D1 13 119.9 154.8 21 91.03 4.51 82.09 25.41 3810
GSR IR1-12-Y4-D1-Y2 14 114 150.3 19.6 87.72 4.82 76.04 23.5 2743
GSR IR1-12-S8-Y1-Y2 15 118.1 155.6 20.8 91.23 4.22 88.36 23.86 3171
GSR IR1-5-S10-D1-D1 16 120 156 20.6 91.96 4.6 95.33 24.02 2449
GSR IR1-8-S6-S3-S1 17 114.6 152.2 20.7 97.84 4.4 68.4 25.19 2403
GSR IR1-5-S12-D3-Y2 18 115.1 152.9 20.8 94.83 4.62 103 24.47 2569
GSR IR1-5-S8-D2-S1 19 113.6 151.6 21 97.16 4.2 87.33 25.77 2475
KOMBOKA (Check) 20 111.4 148.6 21.6 96.29 4.33 85.6 22.58 3008
Mean 115.7 152.9 21.1 92.43 4.52 83.57 27.18 2836
CV (%) 2.8 2.3 12.9 7.5 18.5 18.5 93.8 20.8
LSD (5 %) 3.05 3.21 2.54 6.44 0.78 14.42 23.79 552
Genotype (G) *** *** NS *** NS *** NS ***
Year (Y) NS *** ** *** *** *** NS ***
G*Y NS NS NS NS NS NS NS NS
Note: *, **, and *** refers to significant at 5%, 1% and 0.1% level, NS=non -significant, CV= coefficient of variation, LSD= list significance
different, G*E= genotype by environment, DH= days to 50% heading, DM= days to 85% maturity, FGP= filled grains/panicle, PH= plant
height (cm), PL= panicle length (cm), TGW=thousand grain weight and GY= grain yield (kg/ha)
[195]
At Pawe, the combined analysis of variance showed that significant difference
(P≤0.01)for grain yield, days to maturity, days to heading and plant height; and
significant difference (P≤0.05) for thousand-grain weight among the
genotypes(Table 10). There was no significant difference for fertile tillers per
plant and fertile grain per panicle among tested genotypes. The genotype by year
interaction showed a significant difference for plant height (P≤0.01) and grain
yield (P≤0.05). The mean grain yield ranges from 3416.7 kg ha-1to 4545.8 kg ha-1.
Two genotypes (G14 and G1) showed significantly different than the check and
gave yield advantage of 21.7% and 20.6%, respectively.
Table 10. Mean grain yield and other yield-related parameters of 20 lowland green super rice genotypes at Pawefrom
2016 to 2018.
Genotype Code DTH DTM PL PH FTP FGP TGW GY

GSR IR1-17-Y16-Y3-Y2 1 92 125 19.62 77.07 8.64 101.2 21.17 4503.2
GSR IR1-15-D4-D1-Y1 2 95 128 19.96 72.6 7.71 107.62 25.94 3664.8
GSR IR1-5-D1-D1 3 97 130 20.09 74.93 8.4 107.82 21.5 3428.4
GSR IR1-12-Y4-Y1-D1 4 94 127 19.6 73.04 7.73 100.58 22.38 3633.1
GSR IR1-8-S9-D2-Y2 5 101 131 18.73 67.51 8.24 97.83 20.06 3811.3
GSR IR1-12-S2-Y3-Y2 6 97 132 18.96 70.11 7.04 102.82 20.89 3416.7
GSR IR1-5-D20-D2-D1 7 95 128 18.24 66.27 8.98 106.07 20.06 3588.2
GSR IR1-5-S10-D3-Y2 8 102 133 19.07 69.89 7.58 104.62 20.22 3675.2
GSR IR1-12-S8-Y1-S1 9 100 132 17.84 70.33 7.78 101.18 19.23 3403.9
GSR IR1-8-S14-S1-SU1 10 96 130 19.84 71.51 6.44 106.64 21.17 3529.7
GSR IR1-5-Y3-S2-SU1 11 96 129 20.69 70.78 7.82 102.93 22.89 3698.2
GSR IR1-11-Y10-D3-Y3 12 90 123 20.53 70.91 8.07 98.84 21.73 3766.1
GSR IR1-12-D10-S1-D1 13 97 130 18.76 65.38 6.93 99.38 19.94 3557.1
GSR IR1-12-Y4-D1-Y2 14 97 129 19.6 69.24 8 107.89 21.96 4545.8
GSR IR1-12-S8-Y1-Y2 15 96 129 18.78 67.62 7.2 96.98 20.73 3987
GSR IR1-5-S10-D1-D1 16 96 129 19.51 69.78 6.96 106.36 21.28 3973.3
GSR IR1-8-S6-S3-S1 17 96 130 18.07 69.93 7.98 106.56 20.22 3825
GSR IR1-5-S12-D3-Y2 18 96 129 18.6 72.31 9.36 100.87 20.56 4000.6
GSR IR1-5-S8-D2-S1 19 95 129 18.6 65.47 6.64 110.49 20.06 3934.9
KOMBOKA (Check) 20 96 129 20.18 70.27 7.6 110 19.83 3734.2
Mean 96 129 19.26 70.25 7.76 103.83 21.09 3783.84
CV (%) 3.9 2.9 7.1 6.3 26.7 10.8 14.9 17.2
LSD (5 %) 3.46 3.48 1.28 4.16 1.94 10.51 2.94 608.89
Gen. (G) *** *** *** *** NS NS * ***
Year (Y) *** *** *** *** *** *** *** ***
G*Y NS NS NS *** NS NS NS *
Note: *, **, and *** refers to significant at 5%, 1% and 0.1% level, NS=non -significant, CV= coefficient of variation, LSD=
list significance different, G*E= genotype by environment, DH= days to 50% heading, DM= days to 85% maturity, FGP=
filled grains/panicle, PH= plant height (cm), PL= panicle length (cm), TGW=thousand grain weight and GY= grain yield
(kg/ha)
Summary and Recommendation
The present study in the first experiment revealed that significant differences
among genotypes and environments for grain yield and related agronomic traits
[196]
suggesting a differential response of genotypes to varied environments. In the first
set of experiments, G2 (Yungeng 31) and G6 (KB-2) are proposed for national
variety release. For the second experiment, the combined analysis over years and
locations revealed that there was no genotype statistically significantly higher than
the standard check for grain yield and related agronomic traits. However,site-
specific analysis at Paweshowed that two genotypes, G14 (GSR IR1-12-Y4-D1-
Y2and G1(GSR IR1-17-Y16-Y3-Y2) gave significantly higher grain yield than
the check with a yield advantage of 21.7% and 20.6%, respectively. Similarly,
atAssosaG13 (GSR IR1-12-D10-S1-D1) gave significantly higher grain yield than
the check with a yield advantage of 26.7%. Therefore, further investigation of
regional variety trial is necessary for Pawe and Assosa areas to recommend for
their specific locations. The first experiment revealed the importance of cold
tolerance, high yield, and disease resistance in the evaluation of genotypes. Cold
tolerance varieties allow rice producers to use high elevation areas for rice
production. The proposed two varieties from the first experiment are crucial to
boost production and productivity in the high elevation of rainfed lowland rice
production system of Ethiopia.
References
Khush, GS. 2001. Green revolution: the way forward. Nat. Rev. 2: 815–22.
Liang, J, Y Lu, P Xiao, M Sun, H Corke, and J Bao. 2010. Genetic diversity and population
structure of a diverse set of rice germplasm for association mapping. Theor. Appl. Genet.
121:475–87.
Africa Rice Center (Africa Rice). 2018. Continental Investment Plan for accelerating Rice Self-
Sufficiency in Africa (CIPRiSSA). Abidjan, Cote dIvoire: 204pp.
Moraes de Freitas, GP, Basu S, Ramegowda V, Braga EB and Pereira A. 2016. Comparative
analysis of gene expression in response to cold stress in diverse rice genotypes, Biochemical
and Biophysical Research Communications,http://dx.doi.org/10.1016/j.bbrc.2016.02.004.
Zhang Q, Chen Q, Wang S, Hong Y, and Wang Z. 2014. Rice and cold stress: methods for its
evaluation and summary of cold tolerance-related quantitative trait loci. Rice, 7(1): 24.
Rice.Ethiopia’s millennium crops.
http://www.africarice.org/publications/ricetoday/Ethiopia%E2%80%99s%20millennium%20c
rop.pdf
CSA. 2017. Central Statistical Agency.Agricultural Sample Survey, Report on Area and
Production of Major Crops (Private Peasant Holdings, Meher Season), Statistical Bulletin,
No.?? Addis Ababa, Ethiopia.
[197]
[198]
Evaluation of Lowland Rice (Oryza sativa L.) Varieties in
the Rainfed-Lowland Rice Producing Areas of Ethiopia
Zelalem Zewdu1*, Abebaw Dessie1, Mulugeta Atinaf1
Assaye Berie1, Mulugeta Bitew2, and Geleta Abdi3
1Fogera National Rice Research and Training Centre,2Pawe Agricultural Research Center,3Bako Agricultural Research
Center; *Corresponding Author:zelalemsafe@gmail.com;
Abstract
A field experiment was conducted to evaluate the performance of rice varieties for
grain yield and other agronomic traits in the rainfed lowland rice-producing areas of
Ethiopia; Assosa, Bako, Pawe and Tepi during 2017 and 2018 cropping season. Six
lowland rice varieties were included for evaluation and laid out in Randomized
Complete Block Design with four replications. The combined analysis across years
and locations revealed significant variations (p≤0.001) among the tested varieties for
days to heading, days to maturity, plant height, grain yield and thousand seed weight
while non-significant difference was observed for the number of filled grains per
panicle and panicle length. Both location and year effects were highly significant for
all measured traits. Variety by location and variety by year was highly significant
(p≤0.001) for days to heading, days to maturity, plant height, number of filled grains
per panicle, and grain yield though variety by year was not significant for thousand
seed weight. Variety by location and variety by year were non-significant for panicle
length. Based on the performance of the variety in each test location and year, variety
Ediget is recommended for Assosa (29.30 qt/ha) and Bako (33.90qt/ha) areas with
days to heading and maturity (85 and 145 days) among the tested varieties and
variety Fogera-2 (51.88 qt/ha) for Pawe areas with maturity date(121 days) while in
Tepi area variety Hibir followed by Gumara are recommended for better
productivity(48.30 qt/ha and 46.80 qt/ha) respectively. However, based on average
grain yield over four locations and two years variety Ediget (32.22 qt/ha) followed by
Fogera-2 (30.73qt/ha) performed better than others and they are recommended for
production.
Keywords: Lowland rice, Adaptation, Variety, Yield
Introduction
Rice production is a recent phenomenon in Ethiopia, as compared to other cereal
crops. It had been introduced in the 1960s at Fogera and Gambela areas through
North Korean and Dutch people respectively (CSA, 2011; Gebey et al., 2012).
The country has a huge potential of land for rice production in three ecosystems,
upland (5.6 million ha), irrigated (3.7 million ha) and rainfed lowland. Now, a
days the production and productivity of rice have been expanding throughout the
country (MoARD, 2010). The production has been increased from 112,443
quintals in 2007 to 1,268,064 quintals in 2016. Even though there is an increment
in production, the imports are also increasing from year to year. From 2008 to
2016 the imported rice had increased from 225,000 quintals to 3,118,270 quintals
[199]
and causes more than 170 US dollars. This indicates the increasing demand for
rice from year to year in the country. Rice is not only food securing crops but also
a source of employment and income source for youths in most rice-producing
areas of the country (Dawit et al., 2018).
Even though there is huge potential and increasing demand for the crop, lack of
high yielding varieties, terminal moisture stress, low soil fertility, disease, and
cold effect is the constraints that hinder the expansion and productivity of the crop
(Abebaw, 2018). One of the major constraints in rice-producing areas is the lack
of adaptable and high yielding improved rice varieties. So far eleven improved
lowland rice varieties had been released in Ethiopia. However, the cultivation of
improved lowland varieties in most rice-growing is limited, due to the
unavailability of adaptable rice varieties. Farmers grow old rice varieties
venerable to disease and end with low productivity. Therefore, this study was
designed to select adapted and high yielding lowland rice varieties in the testing
rice producing areas.

The experiment was conducted at Assosa, Pawe, Bako, and TepiResearch Centers
during the 2017 and 2018 cropping seasons in the rainfed lowland condition. The
description of the trail site is indicated in Table 1. Five lowland rice varieties
released from national and regional research centers and one widely produced rice
varieties, but not yet registered (Table 2) were evaluated in a randomized
complete block design replicated four times, on a plot size of 5 long ×1.5 width. A
spacing of 25 cm, 50 cm and 1.5 m were used between rows, between plots and
between blocks, respectively.
Data were recorded on 5 randomly selected plants from the middle four rows for
panicle length, plant height, number of filled grains per panicle, and on plot bases
for days for heading, days for maturity. Grain yield and thousand seed weight
were taken on a plot basis from the four harvestable rows. Fertilizer application
and weeding were carried out following the recommended packages. Data were
subjected to analysis of variance (ANOVA) using GenStat version 12 edition
software package (Payne et al., 2009).
Table 1. Description of experimental locations
Elevation Annual rain Mean temperature(oC)

Location (m) Latitude Longitude fall (mm) Min Max Soil type
Assosa 1580 10° 03’ N 34° 59’ E 1,275 14°C 39°C Nitosols
Bako 1590 9° 06’ N 37° 09’ E 1,245 9°C 34.4°C Nitosol
Pawe 1120 11°19’N 36°24'E 1587 16.3°C 32.6°C Vertisols
Tepi 1200 7° 3' N 35°18’ E 1522.1 15°C 30°C Nitosols
Sources (Tefera et, al, 2017, Hailemariam et al.,2017)
[200]
Table 2. Description of experimental materials used for the study
Released Year of Altitude Productivity(kg ha-1)

Varieties
Center released (m.a.s.l) Rain fall On-farm On-station
Ediget Adet RC 2011 1150-1850 800-1400 32 52
Hiber Adet RC 2013 1150-1850 1000-1400 26-36 34-47
Gumera Adet RC 1999 1740-1900 1520 30 38
Fogera-2 FNRRTC 2016 500-1590 1000-1450 37-49 42-61
Demoze*
X-Jigna*
*Demoze and X-jigna have not yet registered, but they are under production in lowland rice producing areas of Ethiopia

The analysis of variance over years at Assosa showed a highly significant
difference (p≤0.01) among varieties for days to maturity, plant height, number of
filled grains per panicle, grain yield, and thousand seed weight but not on panicle
length (Table 3). The effect of year was highly significant (p≤0.01) on days to
maturity, number of filled grains per panicle and thousand seed weight, and
significantly for panicle length at (p≤0.01), and it was non-significant on plant
height and grain. The interaction effect of the year by variety showed a highly
significant difference (p≤0.01) for the number of filled grains per panicle, grain
yield, and thousand seed weight and days to maturity (p≤0.05); while it was non-
significant for panicle length and plant height. A similar finding was also reported
by Hailemariam et al. 2017.
The average days to maturity ranged from 93 to 115 days. The variety Ediget
matured earlier than the others while Fogera-2 was matured later than the others.
The varieties had minimum mean differences for panicle length. The average plant
height was ranged from 80.6 cm to 98.6 cm; Fogera-2 has the shortest plant height
and Demoze has the longest height among the tested varieties. A maximum filled
grain was recorded on X-jigna (105.6), and a small amount of filled grain was
obtained on Gumara (80.9). The average grain yield was ranged from 2933 kg ha-1
for Ediget to 2065 kg ha-1 for X-Jigna. The highest thousand seed weight was
obtained for Hibir (38.7 g) followed by Gumara (37.0 g). The results revealed that
Ediget (2933 kg ha-1) and Demoze (2933 kg ha-1) are the high yielders and well-
adapted varieties in the rainfed lowland rice-producing areas of Assosa. The
variety Ediget was the highest yielder followed by Demoze for Assosa areas and
could be recommended for production. Hailemariam et al. 2017 had also reported
the same result which confirmed that variety Ediget is less responsive across
different years and stable genotype in the area.
[201]
Table 3. Mean performance of grain yield and yield-related traits of six lowland rice varieties
evaluated in 2017 & 2018 cropping season at Assosa
Variety DM PL PH FGPP GY TSW

Ediget 93 20.0 95.1 89.9 2933.0 36.6
Hiber 95 19.7 89.5 82.7 2124.0 38.7
Demoze 97 20.0 98.6 101.3 2457.0 31.2
Gumera 98 21.2 96.0 80.9 2162.0 37.0
X-Jigna 99 20.6 95.4 105.6 2065.0 31.2
Fogera-2 115 19.9 80.6 101.3 2069.0 23.9
Mean 100 20.2 92.5 93.6 2301.6 33.1
CV (%) 2.2 9.4 9.3 17.1 21.8 7.1
LSD (5 %) 3.163 2.8 12.3 23.1 722.4 3.4
Variety (V) *** NS *** *** *** ***
Year (Y) *** ** NS *** NS ***
VxY * NS NS *** *** ***
DM=Days to maturity, PL=Panicle length (cm), PH=Plant height (cm) FGPP= Number of filled grains per panicle, GY=,
Grain yield (kg ha-1 ), TSW=Thousand seed weight (gm), * , **, ***, and NS denotes significance difference at p ≤0.05,
p≤0.01 and p ≤ 0.001 and non-significant difference respectively
The combined analysis of variance across years at Bako (Table 4) showed a highly
significant difference (p≤0.01) among the tested varieties for days to maturity,
days to heading, panicle length, plant height, number of filled grains per panicle,
grain yield, and thousand seed weight. The effect of year was highly significant
(p≤0.01) for days to maturity, days to heading, plant height, number of filled
grains per panicle, grain yield, while it was non-significant on panicle length and
thousand seed weight. The interaction effect of variety by year showed a highly
significant difference (p≤0.01) for all measured traits (days to maturity, days to
heading, panicle length, plant height, number of filled grains per panicle, grain
yield) except for thousand seed weight.
Days to heading ranged between 84 to 101 days; Edidetgave head earlier than the
others, whereas Fogera-1 was the late heading variety. Ediget was the earliest
maturing (145 days) and Fogera-2 (153 days) was the late-maturing among the
tested varieties. The average panicle length was ranged from 19.7 cm to 22.4 cm,
the shortest panicle length recorded for Demoze, and the longest was for Gumara.
The highest plant height was recorded for Demoze (127.9 cm) followed by X-
jigna (125.4) whereas the shortest plant height was obtained for Fogera-2 (104.2).
The number of filled grains per panicle ranged from 30.5 (Fogera-2) to 55.1
(Ediget). Thousand seed weight was highest for variety Ediget (30.4) and Hiber
(30.4), while the small grain weight was obtained for Fogera-2 (18.9). Regarding
the grain yield, Ediget was the higher yielder followed by Hiber and the two
varieties could be recommended for production in Bako areas.
[202]
Table 4. Mean performance of grain yield and yield related traits of six lowland rice varieties evaluated
in 2017 & 2018 cropping season at Bako
Variety DH DM PL PH FGPP GY TSW

Ediget 85 145 20.3 114.1 55.1 3398.0 30.4
Hiber 84 146 20.1 107.1 54.0 2820.0 30.3
Demoze 88 145 19.7 127.9 44.2 2637.0 26.3
Gumera 94 145 22.4 122.8 33.1 1847.0 26.9
X-Jigna 89 146 20.6 125.4 35.8 1440.0 26.6
Fogera-2 101 153 19.8 104.2 30.5 1419.0 18.9
Mean 90 147 20.5 116.9 42.0 2260.1 26.6
CV (%) 2.6 1.1 4.7 3 25.4 36.9 11.6
LSD (5 %) 2 2 1.0 3.6 10.9 847.4 3.1
Variety (V) *** *** *** *** *** *** ***
Year (Y) *** *** NS *** *** *** NS
VxY *** *** *** *** *** *** NS
DH= Days to heading, DM=Days to maturity, PL=Panicle length (cm), PH=Plant height (cm) FGPP= Number of filled
grains per panicle, GY=, Grain yield (kg ha-1), TSW=Thousand seed weight (gm), *** denotes significance difference at p
≤ 0.001, NS=Non-significant difference
The combined analysis of variance across years at Pawe (Table 5) showed highly
significant difference(p≤0.01) for days to heading, days to maturity, plant height,
number of filled grains per panicle, grain weight and thousand seed weight; and
significant difference (p≤0.05) was observed for panicle length among the tested
varieties. The effect of year was also highly significant on all measured traits
(p≤0.01) except panicle length. The interaction of variety and year revealed a
highly significant difference (p≤0.01) for days to heading, days to maturity, plant
height, whereas the number of filled grains per panicle, grain yield, and thousand
seed weight didn’t show a significant difference. There was a highly significant
difference among varieties. The non-significant effect of the year by variety
indicates that the varieties are stable across years for the testing locations.
Variety Demoze (70 days) followed by Ediget (72 days) were the early variety for
heading while Fogera-2 (121 days) was the late heading variety. Panicle length is
one of the major yield contributing traits. The longest panicle was recorded on
variety Gumara followed by Fogera-2 while the shortest panicle length was
recorded on Hibir. The average plant height ranged from 85.6 (Hibir) to 107.2
(Gumara). The tallest variety was Gumara followed by Demoze and x-jigna, while
the shortest height was recorded on Hibir and Fogera-2. Filled grain is one of the
yields contributing factor and the highest filled grain and grain yield were
recorded on the same variety Fogera-2 which were117.6 and 5188kg ha-1and
followed by Gumara 99.2 and 3619 kg ha-1 respectively. However, the small
number of filled grains was recorded on variety Hiber whereas Demoze was the
lowest yielder variety. Thousand seed weight was higher for variety Hiber
followed by Gumara while the smaller was recorded on Fogera-2.
[203]
Table 5. Mean performance of grain yield and yield-related traits of six lowland rice varieties evaluated in 2017 & 2018 cropping season
at Pawe
Ediget 72 106 18.2 89.2 95.6 2860.0 32.0
Hiber 68 102 17.7 85.6 81.7 2702.0 30.8
Demoze 70 106 18.1 95.4 83.0 2168.0 24.3
Gumera 79 115 19.8 107.2 99.2 3619.0 29.1
X-Jigna 69 104 18.8 95.8 89.2 2218.0 25.7
Fogera-2 85 121 19.7 87.1 117.6 5188.0 20.9
Mean 74 109 18.7 93.4 94.4 3125.6 27.1
CV (%) 2.5 2.1 6.8 7.1 10.8 14.1 3.9
LSD (5 %) 2 2 1.3 6.7 10.4 447.5 1.1
Variety (V) *** *** ** *** *** *** ***
Year (Y) *** *** NS *** *** *** ***
VxY *** *** NS *** NS NS NS
DH= Days to heading, DM=Days to maturity, PL=Panicle length (cm), PH=Plant height (cm) FGPP= Number of filled grains per panicle,
GY=, Grain yield (kg ha-1 ), TSW=Thousand seed weight (gm), ** and *** denotes significance difference at p ≤ 0.01 and
p≤0.001respectively, NS=Non-significant difference
The analysis of variance at Tepi revealed that the varieties were highly significant
different (p≤0.01) for days to heading days to maturity, plant height, and number
of filled grains per panicle, grain yield, and thousand seed weight while panicle
length didn’t show significant difference in 2017 cropping season (Table 6). A
similar result was also previously reported by Tefera et. al 2017. The average days
to heading for the varieties in this cropping season was 92 days. The variety which
took long days to head was Fogera-2 (108 days), while variety Ediget and Demoze
(87 days) followed by Hiber (88 days) are earlier to head. Days to maturity ranged
from 118 to 137 days, the early maturing variety was Ediget while Fogera-2 (137
days) was the late maturing variety in the area. The longest panicle length and
tallest plant height were recorded on Gumara 19.7 cm and 111.1 cm respectively.
Gumara was also the best yielding variety (4680 kg ha-1) followed by Hibir (4830
kg ha-1) whereas Demoze (2398 kg ha-1) and X-jigna were the lower yielder
varieties in the area. The higher thousand seed weight were also obtained on those
high yielding varieties, Hibir (35.8 g) and Gumara 35.8 g). As a result, Hiber
followed by Gumara were the best variety in grain yield, filed grain, and also
thousand seed weight in the cropping season and area.
Table 6. Mean performance of grain yield and yield-related traits of six lowland rice varieties evaluated
in 2017 cropping season at Tepi
Ediget 87 118 18.6 108.3 77.1 4173.0 34.8
Hiber 88 128 18.4 99.5 78.1 4830.0 35.8
Demoze 87 129 18.8 99.6 55.7 2398.0 28.3
Gumera 93 130 19.7 111.1 75.4 4680.0 35.8
X-Jigna 91 131 19.1 101.9 67.4 2784.0 29.2
Fogera-2 108 137 19.1 79.6 66.3 4163.0 20.2
Mean 92 129 18.9 100.0 70.0 3837.9 30.6
CV (%) 4.3 1.8 5.9 6.7 14.5 22.8 12.2
LSD (5 %) 5.9 3.5 1.7 10.0 15.3 1316.9 5.6
Variety (V) *** *** NS *** *** *** ***
GY=, Grain yield (kg ha-1 ), TSW=Thousand seed weight (gm), *** and NS denotes significance difference at p≤0.001and non-significant
difference respectively
[204]
A combined analysis of variance was performed to assess the performance of
varieties with the test locations and years. Significant variation (p≤0.001) was
observed among the tested varieties for days to heading, days to maturity, panicle
length, plant height, grain yield and thousand seed weight while non-significant
difference was observed for the number of filled grains per panicle (Table 7). Both
location and year effects were highly significant for all measured traits indicating
the test locations and years differed. Variety by location and variety by year was
highly significant (p≤0.001) for days to heading, days to maturity, plant height,
number of filled grains per panicle, and grain yield indicating that varieties
performed differently in each location and year. Though variety by year was not
significantly different for thousand seed weight, variety by location and variety by
year were non-significant for panicle length and suggesting that the stability of the
varieties across locations and years. The interaction of variety by location by year
was highly significant for days of heading, days of maturity, plant height, grain
yield and thousand seed weight while panicle length and number of filled grains
per panicle were not significant for this interaction.
Grain yield is the most important economic trait in most crop improvement
programs. The highest grain yield was recorded for Ediget (3222 kg ha-1) and
Fogera-2(3073 kg ha-1) while the low grain yield was recorded on X-Jigna. This is
due to the long age of the variety under production. The average maturity days for
the variety were 88 days; Fogera-2 and Gumara were the late-maturing varieties
and Hiber and Ediget were early maturing. Variety Fogera-2 had a high number of
filled grains per panicle and grain yield. However, variety Ediget was the best
variety across locations and year in terms of the number of filled grains per
panicle and grain yield per ha (3222 qt/ha) with high thousand seed weight.
Table 7. Combined analysis of mean performance of grain yield and yield-related traits of six lowland rice varieties over locations and
years

Ediget 84 131 19.4 100.8 83.1 3222.0 33.2
Hiber 83 133 19.0 94.8 73.5 2875.0 33.7
Demoze 86 134 19.2 106.2 74.0 2417.0 27.4
Gumera 91 136 20.9 109.0 75.9 2848.0 31.7
X-Jigna 86 133 19.9 105.0 76.5 2033.0 28.0
Fogera-2 101 143 19.7 89.1 81.6 3073.0 21.1
Mean 88 135 19.7 100.8 77.4 2744.6 29.2
CV (%) 2.9 1.5 7.2 6.5 26.7 25.8 8.8
Variety (V) *** *** *** *** NS *** ***
Location *** *** *** *** *** *** ***
Year *** *** *** *** ** *** ***
VxL *** *** NS *** ** *** **
VxY *** *** NS *** ** *** NS
LxY *** *** * *** ** *** ***
VxLxY *** *** NS *** NS *** ***
GY=, Grain yield (kg ha-1 ), TSW=Thousand seed weight (gm), *,**, *** and NS denotes significance difference at p ≤ 0.05, p≤0.01,
p≤0.001 and non-significant difference respectively
[205]
The combined analysis of variance over years and locations showed that there
were high interactions of variety by location, variety by year and variety by
location by year. This indicated that varieties performance varies in each location
and year. Therefore, the variety Ediget had the best performance in Bako and
Assosa and it is recommended for productions in these areas, while Fogera-2 had
better performance in Pawe and is recommended for production in the area. In
Tepi area variety Hibir and Gumarafollowed byEdigethad better performance and
they are recommended for production. However, the combined analysis over the
years and locations indicated that variety Edigethad better performance and it is
recommended for early maturity and good yield followed by Fogera-2 for good
yield and late maturity.
References
Abebaw Dessie. 2018. Cereal Crops Research Achievements and Challenges in
EthiopiaInternational Journal of Research & Review (www.ijrrjournal.com)Vol.5; Issue:9;
September 2018
Tefera A, Alamerew S, Tulu L. 2017. Genetic Variability, Heritability and Genetic Advance for
Yield and its Related Traits in Rainfed Lowland Rice (Oryza sativa L.) Genotypes at Fogera
and Pawe, Ethiopia. Adv Crop Sci Tech 5: 272. doi: 10.4172/2329-8863.1000272
Central Statistical Agency (CSA, 2011). 2013. Report on Area and Production of Major Crops
(Private Peasant Holdings, Meher Season). ADDIS ABABA, Ethiopia.May 2013.
Dawit A, Agajie T, Abebaw A, Degu A, Tilahun T and John T. 2018. A historical analysis of rice
commercialization in Ethiopia: the case of the Fogera plain. A blog posted in future
agricultures consortium website. Posted on May 31, 2018. Link: https://www.future-
agricultures.org/blog/from-the-fogera-plain-how-one-man-helped-to-build-a-regional-
industry/
Gebey T, Berhe K, Hoekstra D and Bogale A. 2012. Rice Value Chain Development in Fogera
Woreda Based on the IPMS Experience. Nairobi, Kenya: ILRI. 23pp.
Hailemariam S., Taddesse L, Abebaw D. 2017. GGE Biplot Analysis of Yield Stability in Multi-
Season Trials of Early Matured Rice (O. Sativa) Genotypes in Rain Fed Lowland Ecosystem
Ministry of Agriculture and Rural Development (MoARD). 2010. National Rice Research and
Development Strategy of Ethiopia. Addis Ababa, Ethiopia, pp. 48.
Payne RW, Murray DA, Harding SA, Baird DB, Soutar DM. 2009. An introduction to
GenStat for Windows (12th edition). VSN International, 5. The Waterhouse, Waterhouse
Street, HemelHemstead, UK.
[206]
Statistical Analysis of Multi-Environment Trial
Data to Enhance Genetic Gain in Ethiopian
Sorghum Breeding Program
Driba Tadese, Amare Seyoum, Taye Tadesse, Yohannes Fikadu, Amare Nega, Adane
Gebreyohannes , Alemu Tirfessa, Kidanemaryam Wagaw, Sewmehone Siraw belay,
Daniel Nadew, Hailemariam Solomon, Rebuma Mereraa, and Tamirat Bejiga
EIAR, Crop Research Directorate, Melkassa Agricultural Research Centre,
National Sorghum Research program P. O. Box 436/2003, Adama, Ethiopia.
Abstract
Different statistical methods for the analysis of multi-environment trial (MET) data in
a plant breeding program have been advanced over the past decades. However, many
plant breeding programs have not yet taken advantage of these methods. The aim of
plant breeding is most often to select either high performing genotypes for target
environments or stable genotypes across a given set of environments. Hence, we
demonstrate the benefits of implementing these methods into the sorghum breeding
program in Ethiopia. Then, the specific objective of this study would be, adjusting for
the effect of field trend in each trial using spatial models based on the shared
environment of plots that are close together, exploiting the correlation between trials
using factor analytic models for genotype by environment (GxE) interaction and
exploiting the correlation between genotypes by incorporating pedigree information in
the analysis. These advanced methods are based on a linear mixed model that can
incorporate all of these correlation models. Therefore, the use of correlated data in
this study shows high improvements in enhancing genetic gain.
Keywords: Spatial analysis, Sorghum MET, Linear Mixed Model, Heritability.
Introduction
Sorghum (Sorghum bicolor (L.) Moench) is a diploid C4 cereal crop that was
domesticated in Africa. It has 2n = 20 chromosome and genome size of 750 Mb
(Paterson et al., 2009). Globally, Sorghum [Sorghum bicolor (L.) Moench] 2n =
20) is the fifth most important cereal crop and is the dietary staple of more than
500 million people in 30 countries (FAO,2011). It is grown on 40 million ha in
105 countries of Africa, Asia, Oceania and the Americas. Sorghum plays an
important role as a dietary staple for millions of people, especially in arid and
semi-arid countries of Africa and Asia. In Ethiopia, the crop stands third next to
tef and maize both in area and total production (CSA, 2018) and it is
predominantly cultivated in dryland areas where drought causes frequent
failures of other crops (Amare S et al., 2019). The current sorghum production in
Ethiopia is estimated to be 5.1 million tonnes on an area of 1.88 million ha of land
-1
giving the national average grain yield of 2.71 tones ha (CSA, 2018).
[207]
Developments of early maturing and drought tolerant varieties containing stay
green trait that assist the plant to delay leaf senescence are the major strategies in
the national sorghum breeding program of Ethiopia. Exploitation of genetic
variability is the most important tool in plant breeding and this has to be inferred
by phenotypic expression. The consequences of the phenotypic variation depend
largely on the environment. This variation is further complicated by the fact that
not all genotypes react in similar ways to the change in environment and no two
environments are the same. If the relative performance of genotypes grown in
different environments is different, then GxE interaction becomes a major
challenge to crop improvement.
Different statistical methods for the analysis of multi-environment trial (MET)

data in plant breeding programs have advanced over the past decades. However,
many breeding programs have not yet taken advantage of these methods. The aim
of plant breeding is most often to select either high performing genotypes for
target environments or stable genotypes across a given set of environments.
Breeders are also interested in stable genotypes, given that such genotypes are
among the winner genotypes. But the challenging issue could be getting high
performing stable genotypes, mainly due to interactions between genotypes and
the environment. Usually, breeders put their trials across several locations within a
given agro-ecology and for a number of years. Several statistical techniques have
been proposed to facilitate the estimation and interpretation of genotype by
environment interactions (Tadese et al., 2019).
The classical statistical techniques mostly used for analysis of multi-environment

trial (MET) is the two cross classifications (genotype by environment interaction)
of Analysis of Variance (ANOVA). However, while this technique can adequately
explain only the main effects, and identify GxE interactions as a source of
variation, it fails to decompose and analyse the inherent interaction effects. This is
due to the additive nature of the ordinary ANOVA model, which does not allow
for analysis of nonadditive interaction components. Therefore, other statistical
approaches were proposed to identify the pattern and complexity of interaction.
Oakey et al. (2007) compared the performance of the ANOVA method with the
regression method, and found that ANOVA fails to detect significant interaction
components; and the regression approach accounts for only a small portion of the
interaction sum of squares when the pattern fits specific regression model.
MET is usually analysed using a two-stage approach this approach is an

approximation of the combined analysis of the raw plot data from all trials Smith
and Cullis (2001) presented a weighted mixed model for the second-stage analysis
that aimed at accommodate sources of error variation, thereby reducing efficiency
losses in estimation. The superior approach, however, is the spatial MET analysis
[208]
of Kelly et al. (2007), in which individual plot data are analysed and a separate
spatial covariance structure and error variance allowed for each trial.
Spatial variation within trials and error variance heterogeneity between trials
occurs routinely, and if unaccounted for, will result in biased and inefficient
estimates of variety. In MET data analysis, usually, there is interest in the analysis
of both overall performance of each variety (across all environments) and whether
variety performance is affected by GxE interaction. The approach of
multiplicative models for the interaction effect allows for a separate genetic
variance for each trial and provides a parsimonious and interpretable model for the
genetic covariances between pairs of trials. Then, the genetic model can be
regarded as a random-effects analogue of the additive main effect and
multiplicative interaction (AMMI) model. Hence, this approach combines the
strength of AMMI with the advantages afforded by the mixed model framework.
Therefore, in this study, a linear mixed-effect model was used to analyse MET
data, with the specific objective of clustering sorghum testing environment based
on the correlation ( how the locations respond to each genotype) of the trials, to
show advantages of using advanced statistical methods for precise estimation and
improvement in heritability of sorghum breeding program in Ethiopia. So, to
achieve these objectives, the field trend at each trial was included in LMM to
capture spatial field trends then, the second stage analysis under LMM, combining
the trial through Factor Analytic (FA) model to capture correlated trial in MET
data.

A multi-environment trial (MET) data set was taken from the sorghum breeding
program of the Ethiopian Institutes of Agricultural Research (EIAR). The data set
contains 13 trials of five locations of four years. The number of genotypes were
685 while the number of genotypes included at each year was different (Table 1)
Table 1. Total number of genotypes on the diagonal and number of genotypes in common between trials on
the off-diagonal.
Trial series 2014 PYT 2015 PYT 2016 NVT 2017 NVT
2014 PYT 433
2015 PYT 77 300
2016 NVT 77 80 80
2017 NVT 33 20 19 60
[209]
Commonly, field experiment is laid down in a rectangular form like jth trials j= 1. .
. p, which consist Nj, plots with rj, rows and cj columns (Nj = rj x cj) (Smith and
Cullis 2001). For spatial analysis, the vector is sorted accordingly rows
within columns. Hence, the combined vector of data across all environments can
be modeled as,
= {yj), n = is given by
y = X + Zu+ e --------------------------------------------------------------(1)
where and are vectors of fixed and random factors respectively

while, and be associated design matrices for fixed and random
effects with the former assumed to be of full column ranks and e is a vector of the
residual. The joint distribution of ( is assumed to be,
Where and are vectors of variance parameters. Then, the distribution form of
the vector data y, is a Gaussian distribution with mean X and variance matrix H=
.
In addition, the error term contains a vector of sub error {ej}, where is a
th
vector of plot errors for a given j trial and decomposed into a spatially dependent
process while independent white noise process . The matrix of the error term
for trial j can be written as = , where is the matrix of
spatial correlation which is a function of with associated variance while
is the variance parameters of the white noise process. The spatial process
is assumed to be the second-order stationery given that the correlation between
plot depends on the lag distance. Furthermore, the columns and row dimensions
are separable. Hence, one can write , where and are the
two-dimensional correlations respectively. However, several research findings
(Oakey, et al., 2006) show that the first ordered autoregressive which is denoted
by AR1xAR1 most often provides an adequate variance structure for local spatial
trends (Tadese et al 2019).
The random effect u consists of sub vectors { }, where is the vector of

th
effect for the i random term, i=1. .q. the matrix Z is partitioned conformably as
[Z1. .Zq]. It assumed that the sub-vector of u are mutually independent. Variance
[210]
matrix Gi for the ith random term has many possible forms including the standard
variance component structure Gi = .
Let ug be the mpx1 vector of genetic effect for m varieties for each p environment
ordered as varieties within environments. It represents a two dimensional
(varieties by environment) array of effect, namely , where = . It
is assumed that the associated variance structure has separable form with
= , where and are the symmetric p x p and m x m component
matrices for environment and varieties, respectively. When = , just for
simplicity, therefore , and the matrix is the so-
called genetic variance matrix. The diagonal elements are genetic variance for
individual environments and the off-diagonal elements are genetic covariance
between pairs of environments.
The spatial mixed model for MET can then be written as
= , -----------------(2)
The fixed effect includes environmental main effects and trial-specific effects
for extraneous field variation (Oakey et al., 2006), is a variety effects at each
environment with associated design matrix and comprise an
additional random effect with a design matrix .
Therefore,
The random term consists sub vectors of { }, where is the vector of

th
effect for the i random term.
Using genetic relatedness in a pedigree model, we can partition genotype effects
into additive and non-additive effects. Most often, plant breeders are interested in
two things the merit of an individual (total genetic effect) and the value of an
individual as a parent of a cross (additive effect or breeding value). The breeding
value (additive effect) is the value of an individual, as judged by the mean value of
the individual's progeny. The expected performance of offspring can be derived
from parental performance. Including genetic relatedness in the MET, analysis
allows for estimation of breeding values and increases the accuracy of total
genetic effects. Furthermore, extending the standard statistical approach to include
[211]
pedigree information; genetic line effects g can be partitioned into a vector of
additive line effects and a vector of non-additive effects i, such that g = a + i
(Oakey et al., 2006). Generally, in non-additive components, dominance and
epistasis cannot be separated in this method; since in inbreeding the dominance
effect of a line is generally assumed to be zero. Hence, the non-additive effects
here are referred to as epistatic effects.
Therefore, the standard LMM can be extended as;
, --------------------------------------(3)
Where A(mxm) is a known additive relationship matrix.
In a breeding program, there are many possible forms of genetic variance matrix
structures. Under a linear mixed model, the standard structure is given by =
, where and are the variance components for variety main
effects and interaction effects respectively, where is a p x p matric of one. This
implies that all environments have constant genetic variance and all pair of
environment has the same genetic covariance. In this case, since inefficient
estimation, Smith and Cullis (2001) consider an alternative variance structure
model which is known as the Factor Analytic model which is analogous of the
AMMI model. In multi-environment trial data, this model captures the nature of
heterogeneous variance-covariance
The model for factor analytic is given as
=(
=( ,
where
, and
).
[212]
Therefore, in this study first field trend, global variability (variability from one
corner to the other in the experiment), extraneous variation and pedigree
information is considered for each specific trial. Later on, combined analysis
followed through the FA model considering all information at individual trials
through LMM.

A visual display of residual variation before a spatial adjustment and residual
variation after adjustment are presented in Figures 1a and b. In large field trials,
field variations are a substantial source of error, since neighboring plots shared
similar characteristics compared to those far apart. This spatial variability can bias
the estimated potential of a given genotype, when comparing genotypes. Hence,
neighbor lines of plots were used to model the systematic field trends to minimize
bias, and reduce error to increase precision in estimation. Figure1a indicates areas
of high and low yields in the field. It also shows that neighboring plots tended to
be more similar than those far apart. Thus, there is a need to include spatial
correlation in variance-covariance analysis, to handle field trends that extend from
one corner to another corner in the field.
Residual variation after spatial adjustment
[213]
b
Figure 1. Residual variation in field trial before spatial adjustments and after spatial adjustments.
When trials are correlated (similar response of genotypes in some environment)

best materials in a given environment is the same in another environment. Then,
information from multiple environments can be combined to improve the accuracy
of genetic gains of particular trials. In this case, MET analysis can also help to
understand the broad and specific adaptation of genotypes over a range of target
populations. In this study, the factor analytic model was used for MET data
analysis. The correlations between environments ranged from -1 and 1.
Correlations of -1 indicates that the performance of the environments fall in
opposite direction (the angle between the two environments is more than 90
degree), implying that the highest performing genotypes in one environment were
the lowest-performing genotypes in another environment.
Correlation of +1 is an indication of perfect similarity between two environments,

hence selecting superior genotypes based on one environment is the same as
selection for another environment. In Figure 2, for instance, ER14PYT had nearly
0 correlation with ER15PYT, ER17NVT, MI17NVT and SH16NVT2; while
KB14PYT, SH17NVT, MI14PYT, MI16NVT1, SR17NVT had nearly perfect
correlation. But sometimes, there may be an indication of none or even reverse
correlations between trials within the same locations, but for different years due to
seasonal effect or other factors; for example ER14PYT with ER17NVT (Fig. 2).
Generally, correlations ranged from negative to positive (Figs. 2 - 3) both sides.
For example, ER17NVT had a negative correlation with SH16NVT1; hence
selection of best genotypes based on one trial may be lower in the other trials.
MI17NVT did not correlate with ER14PYT. Therefore, it is important to classify
the trial for selection strategies into different clusters based on the correlation
structure in a trial.
[214]
Figure 2. Heatmap to show patterns of Genetic correlation in sorghum experiment of 13 trials in Ethiopia.
SH16NVT1
KB14PYT
Axis 1
MI14PYT
(30%) SR17NVT SH16NVT2
ER17NVT
Axis 2 (22%)
Figure 3. Biplot of GxE in 13 sorghum trials.
Graphical description of MET data is commonly used to explain genotype by

environment. In Figure 3, the concepts of genotypes won and where is illustrated.
Plots showed that the environment with the longest line from the center measures
the discriminativeness of that environment when compared with others. For
example, SH16NVT1 and SH17NVT were among the most discriminative
environments; this means that environments had considerable contributions in
discriminating genetic variations. On the other hand, environments with less
distances from the center were those stable environments, hence they explained
[215]
less genetic variations. In addition, when a specific genotype is close to a given
environment, it indicates that the genotype is the winner for that specific
environment. That means, that this genotype is the best performer for that trial.
Figure 3 is also about correlations between the trials. Environments with less
angle (less than 90 degrees) between their two lines are more correlated. Then, the
selection of the best genotypes based on one of the two environments does not
change the ranks of the genotypes in another environment. When an angle
between two lines is greater than 90 degrees, selection based on the two
environments can change the ranks of genotypes; for example. Heritability can be
estimated more precisely through appropriate statistical methods, without
additional costs. Figure 4 indicates an improvement in heritability generated from
different statistical models; including heritability in classical RCBD, in spatial
modeling and spatial plus correlated trials (MET). In Figure 4, the horizontal axis
is the list of sorghum trials; while the vertical line is the percent of heritability for
each trial in different statistical models. In most of the trials, there was an
improvement in heritability when spatial plus correlated trials were modeled
simultaneously. This means that if there is a spatial field trend in the trial,
heritability improves through modeling field trends. Also, in MET there is
improvement in heritability since it borrows information from different trials.
It is well known that the reliability of varietal selection from a single trial is
greatly improved with the use of sound design and analysis techniques.
Consideration of spatial trends in field experiments increased gains in accuracy
and precision, as this study indicated, for which also many authors recognized the
need to model spatial trends within a field trial (Cullis et al., 2006). If not
considered, spatial variability can bias the selection of genotypes; implying that
genotypes are selected only due to the plot where they are planted. For example,
Figure 1a indicates the potential existence of field trends, which is usually
expected in large trials. Hence, for selection purposes, adjustment for spatial
variability is important. The adjustment is simultaneously fitted under a linear
mixed model, with other sources of variability at a time.
The Factor Analytic (FA3) model was used to indicate the correlation between
trials. In Figure 2, there are environments with strongly positive correlations
KB14PYT, SH17NVT, MI14PYT, MI16NVT1, SR17NVT. Environments with
negative correlations include ER17NVT, SH16NVT1, MI15PYT, SH16NVT2,
ER15PYT.
This study showed which genotype was the winner and in which environments, as
well as the correlations between environments based on angles of the two
environmental axes (Figs. 3). Discriminative environments are also identified in
this study (Fig. 3); this means environments with high interaction with genotypes.
[216]
Environments with less distances lines from the center are indications of a stable
environment.
The performance of each genotype is determined by the closest environment, and

one genotype can be closer to one or more environments; that means this genotype
is superior in these environments; an observation that agrees with the ideas of Yan
and Tinker (2006).
Figure 4. Improvement in heritability through different advanced statistical methods.
The improvement in heritability under the different statistical methodologies is

illustrated in Figure 4; for example, heritability gained under classical RCBD,
heritability under spatial analysis and heritability under spatial plus MET through
use of correlated environments are compared. In general, there is high
improvements of heritability through modeling spatial field trend with MET under
linear mixed effect model.
[217]
In this study, the importance of spatial field trend, pedigree analysis and correlated
environment through the FA model is indicated. Estimations through
maximization of the linear mixed model is more appropriate and applicable for the
analysis of MET, since it can handle and fit simultaneously all source of variations
(extraneous variations, random variations, trend variations, and pedigree
information) at a time. In addition, to select the best variety, it is better to consider
the correlation between the environments. Smith and Cullis (2001) also strongly
suggested the need of considering field trends in experiments for precise
estimation.
Acknowledgment
In the course of modernizing and redesigning the national sorghum breeding
program, there have been various institutions and individuals who have made a
significant contribution to make sorghum one of the best breeding programs in
Africa. In this regard, we want to extend our heartfelt appreciation to the Bill &
Melinda Gates Foundation (BMGF) for funding of sorghum research projects and
the University of Queensland for technical support. A special thanks and
appreciation are due to Dr.Yilma Kebede, Professor David Jordan, and Emma
Mace and Alison from the University of Queensland for their unreserved
assistance for the introduction, implementation, and adoption of modernization of
the breeding program into the national sorghum breeding program of Ethiopian
Institute of Agricultural Research (EIAR) through iMashilla and MERCI projects.
.
References
Amare S, Gebreyohannes A, Nega A, Nida H, Tadesse T, Tirfessa A and Bejiga T. 2019.
Performance Evaluation of Sorghum (Sorghum bicolor (L.) Moench) Genotypes for Grain
Yield and Yield Related Traits in Drought Prone Areas of Ethiopia. Adv Crop Sci Tech 7:
439. doi:10.4172/2329-8863.1000439
Central Statistical Agency (CSA). 2018. Agricultural sample survey. Report on area, production of
major crops, (private peasant holding, Meher season), Addis Ababa
Cullis BR, Smith AB and Combes NE. 2006. On the design of early generation variety trials with
correlated data. American Statistical Association and the International Biometric Society.
Journal of Agricultural, Biological, and Environmental Statistics 11(4):381–393.
Tadese D, Taye G and Lakew B. 2019. Spatial analysis in multi-environment trials of malt barley
in Ethiopia. African Crop Science Journal, Vol. 27, No. 3, pp. 515 – 527. DOI:
https://dx.doi.org/10.4314/acsj.v27i3.13
Food and Agricultural Organization of the United. 2011. FAO STAT Prod STAT, Production
Crops. http://faostat.fao.org/site/567/default.aspx#ancor (accessed June 23, 2011)
Kelly A, Smith A, Eccleston J, and Cullis B. 2007. The accuracy of varietal selection using factor
analytic models for Multi-environment plant breeding trials. Crop Science 47:1063-1070.
[218]
Oakey H, Verbyla A, Cullis B, and Pitchford W. 2007. Joint modeling of additive and non-additive
(genetic line) effects in multi-environment trials. Theoretical and Applied Genetics
114:1319–1332.
Oakey H, Verbyla A, Pitchford W, Cullis B, Kuchel H. 2006. Joint modeling of additive and non-
additive genetic line effects in single field trials. Theoret Appl Genet 113:809–819.
Paterson, AH, Bowers JE, Bruggmann R, Dubchak I, Grimwood J, Gundlach H, Haberer G,
Hellsten U, Mitros T, Poliakov A and Schmutz J. 2009. The Sorghum bicolor genome and
the diversification of grasses’, Nature, 457(7229), pp. 551–556.
Smith, AB and Cullis BR. 2001. The analysis of crop variety evaluation data in Australia.
Australian and New Zealand Journal of Statistics 43:129-145.
R., Cullis B. 2010. A comparison of analysis methods for late-stage variety evaluation trials.
Australian and New Zealand Journal of Statistics 52:125–149.
Yan and Tinker. 2006. Eastern Cereal and Oilseed Research Centre, Agriculture and Agri-Food
Canada, 960 Carling Ave., Ottawa, Ontario, Canada K1A 0C6.
[219]
[220]
Tef (Eragrostis tef) Variety Development for
Moisture Deficit Areas of Ethiopia
Worku Kebede1, Yazachew Genet1, Tsion Fikre1, KidistTolosa1, Solomon Chanyalew1,
Mengistu Demissie1, Kebebew Assefa1, Kidu Gebremeskel2,
Atinkut Fentahun3 and Zerihun Tadele4
1EIAR, DebreZeit Research Center, P.O.Box 32; 2TARI, Axum Research Center, P.O.Box30, Axum, Ethiopia;
AARI, Adet Agricultural Research Center, P.O. Box 8, Bahir Dar, Ethiopia
4University of Bern, Institute of Plant Sciences, Altenbergrain 21, CH-3013 Bern, Switzerland
Abstract
Tef [Eragrostis tef (Zucc.) Trotter] is an extensively cultivated cereal crop in
Ethiopia where it is annually grown by about 6.5 million smallholder farmers on
about 30% of the total area allocated to cereal crops. However, the productivity of
tef is very low compared to other cereals mainly due to lack of adequate scientific
improvement done on the crop, and consequently the widespread use of local
varieties coupled with traditional cultural practices. The objective of this experiment
was to determine and understand the effect of genotype, environment, and their
interaction on grain yield of tef and to identify and release stable and high yielding
tef genotype for moisture deficit areas of the country. Fourteen selected tef genotypes
derived from two independent crosses and advanced through selection for at least
eight generations plus a local and standard check varieties were tested over three
years at seven tef growing sites in moisture stress areas in Ethiopia (viz. Debre Zeit,
Alemetena, Dhera, Axum, Sirinka, Minjar, and Mehoni) using randomized complete
block design with four replications. Among the test lines, (DZ-Cr-387 X 3774-
13(RIL120B)) was found superior in performance. In terms of pedigree, this variety
was obtained from a targeted cross between DZ-Cr-387(Quncho) selected as a
maternal parent for its high yielding ability and wide adaptability and 3774-13
selected as a pollen parent for its very white seed color and earliness. The latter
parent was obtained through TILLING (Target Induced local lesions IN genomes)
technique from the University of Bern. Eventually, the resulting variety christened
“(DZ-Cr-387 X 3774-13(RIL120B))” was approved for release by the National
Variety Release Committee in May 2019. The main advantages of DZ-Cr-453(RIL
120B) over other tested lines were its higher grain yield and low moisture stress
tolerance. (DZ-Cr-387 X 3774-13(RIL120B)) showed grain yield advantage of 8.5%
and 30.7% over the standard (Boset) and local check, respectively. In addition, its
main positive feature(s) of (DZ-Cr-387 X 3774-13(RIL120B)) include high
adaptability and stability in mega environments.
Keywords: Eragrostis tef, bora, early maturity, Variety Release
Introduction
Tef, Eragrostis tef (Zucc.)Trotter is a member of the Grass or Poaceae Family and
Chlorodoideae sub-family. Tef is an allotetraploid (2n=4x=40) with estimated
genome size ranging from 648 to 926Mbp (Mulu et al., 1996; Fufa et al., 2000),
which is approximately 60% larger than the rice genome. It is self-pollinated with
[221]
a very low degree of out-crossing that ranges from 0.2% to 1.0% (Seyfu, 1997).
Tef represents a unique biodiversity component in the agriculture and food
security systems of millions of poor farmers in Ethiopia. It is the most important
cereal crop of Ethiopia, where the last two decades (1995 to 2015) have shown
tremendous increases in both the total area devoted to the crop and its production
(CSA, 2015). Ethiopia is the center of both origin and diversity for tef (Vavilov,
1951).
Currently, the crop is increasingly receiving global attention for its nutritional
advantages because it is rich in nutrients and is gluten-free. Consumers prefer tef
not only because it makes good quality “injera”, a pancake-like soft bread, but it
also is nutritious due to its high protein and mineral contents (Geremew et al.,
2002). Furthermore, the absence of gluten in the grain (Spaenij-Dekking et al.,
2005) renders it an alternative food for people suffering from a gluten allergy or
intolerance and celiac disease. Tef is also known to be tolerant to extreme climatic
and soil conditions; hence, it is a favorite crop in the semi-arid areas with moisture
limitations (Zerihun and Kebebew, 2012).
Despite its numerous relative advantages and economic importance, the

productivity of tef in Ethiopia is low amounting to 1.66 tons ha -1 (CSA, 2017).
The major yield-limiting factors in tef is lack of cultivars tolerant to lodging and
drought (Kebebew et al., 2011), as well as small seed size. Yield losses are
estimated to reach up to 40% during severe moisture stress (Mulu, 1993). Further,
yield reduction of69% to 77% has been reported to have occurred as a result of
drought at the anthesis stage of tef (Abuhay et al., 2001).
In Ethiopia, to increase tef productivity, more than 42 improved varieties have

been developed and released by national and regional agricultural research of the
country mainly by Debre Zeit Agricultural Research Center (MoANR, 2017).
Among these, recently released varieties including Quncho (Kebebew et al.,
2011), Kora (Kebebew et al., 2017), Dagim (Solomon et al., 2017) and Tesfa
(Worku et al., 2018) showed significant yield benefits. Most of them are targeted
to favorable and wide environments but they are not suited for diverse and
challenging environments. In plant breeding programs many new genotypes are
usually evaluated in different environments (location and years) to identify and
advance desirable ones towards release. A genotype or cultivar stable if it has
adaptability for a trait of economic importance across diverse environments. The
environmental component (E) generally represents the largest component in
analyses of variance, but it is not relevant to cultivar selection; only G and GE are
relevant to meaningful cultivar evaluation and must be considered simultaneously
for making selection decisions(Yan and Kang 2003). In general genotype by
environment (GxE) interaction affects the efficiency of crop improvement
programs that may lead to complicates recommendation of varieties across diverse
[222]
environments. Information on the structure and nature of GxE interaction is
particularly useful to breeders (Yayis et al., 2014). Hence, this trial was initiated
with the following objectives:
 To determine and understand the effect of genotype, environment, and
their interaction on grain yield of tef.
 To identify and release stable and high yielding tef genotype for moisture
deficit areas of the country

The field experiment was carried out over three years (2014/15, 2015/16 &
2016/17) at seven tef growing sites in moisture stress areas of the country
including Debre Zeit, Alemetena, Dhera, Axum, Sirinka, Minjar and Mehoni. In
some of the locations such as Debre Zeit low moisture stress was simulated by late
sowing.
Targeted hybridization between DZ-Cr-387 X 3774-13 was made in 2010. The

purpose was to develop stable, high yielding, early maturing, white seeded and
farmers, and consumers preferred tef varieties for the moisture stress areas of the
country. In other words, it was targeted at developing varieties with high yield and
better quality than the improved variety Boset. The maternal parental line was the
popular tef variety called Quncho (DZ-Cr-387 RIL355) (Kebebew et al., 2011)
selected for its high yielding ability and wide adaptability. Line 3774-13 was
selected as the pollen parent for its very white seed color and earliness, and it was
obtained through TILLING (Target Induced local lesion IN genomes) technique
from the University of Bern, Switzerland. From the crosses, 500 F2-derived
genotypes generated and progressively advanced through eight selfed generations
of stringent selection focusing on standing ability and grain yield. Crossing and
early generation testing were done for all breeding populations at Debre Zeit
Agricultural Research Center from where the National Tef Breeding Program is
coordinated. The crossing combinations and names of recombinant inbred lines, as
well as control materials used in the current study, are shown in Table 1.
The performance of fourteen tef genotypes including 12 selected early maturing

inbred lines from the crosses as well as two controls (farmers’ variety or local and
improved standard check variety, Boset) were tested at Debre Zeit, Alemetena,
Dhera, Axum, Sirinka, Minjar and Mehoni using randomized complete block
design with four replications. The trial was conducted on the plot size of 2m*2m
with 10 rows per plot throughout all trial sites and 1.5m between replication, 1m
between plot, 20cm between rows. Agronomic and yield data were collected and
subjected to statistical analysis to identify the best genotypes of the evaluated
genotypes. A variety verification trial was conducted at Debre Zeit, Alemetena,
[223]
Axum, Sirinka and Minjar on the trial station and in ten farmers’ field during the
2018/19 main production season.
Data Collection and analysis

Data on grain yield and yield-related traits were collected on plot and plant basis
from each plot, respectively. Date of heading and maturity were taken when each
plot attained a 50% heading (panicle emergency), 90% physiological maturity, and
lodging index taken during 90% physiological maturity, respectively, and days
were calculated beginning from the date of sowing. Data for plant height (cm),
Panicle length (cm) were collected based on five sample plants which were
randomly taken from each plot and the average of five sample plants was used for
analysis. Thousand seed weight (g) and grain yield (g) of each plot were measured
on clean, dried seed and the measured grain yield value (g) has converted to
kilogram per hectare for analysis. All agronomic practices were done as per the
recommendation for tef.
Data from individual environments and combined over seven locations were
analyzed by using SAS (2009) software. The analysis of variance for grain yield
and yield-related traits for each environment and over seven environments was
analyzed by using a randomized complete block design. The combined analysis of
variance across the environment was done to determine the differences between
genotypes across environments, among environments and their interaction.
Bartlett’s test was used to assess the homogeneity of error variances before doing
combine analysis over environments. Mean comparison using Duncan's Multiple
Range Test (DMRT) was performed to explain the significant differences among
means of genotypes and environments.
Analysis of Variance (ANOVA) of data of individual location, Yij =µ + Gi +Bj +

eij model was used, where; Yij = observed value of genotype i in block j, µ =
Grand mean of the experiment, Gi = the effect of genotype i, Bj = the effect of
block j, eij =the error of genotype i in block j. A combined analysis of variance
over locations was carried out using the following statistical model. Yijk=µ + Gi
+Ej +GEij+ Bk(j)+ eijk, where; Yijk = observed value of genotype i in block k of
environment (location) j, µ = Grand mean of the experiment, Gi = the effect of
genotype i, Ej =the effect of the jth environment GEj = the interaction effect of
genotype i with environment j, Bk(j) =the effect of block k in location
(environment) j, eijk =the error effect of genotype i in block k of environment j.
R software (3.5 version) was used to visualize GEI patterns. Based on principles
of GGE biplot, for the yield characters; Environmental evaluation (The power of
environments to discriminate among genotypes), Genotype evaluation (the mean
performance and stability) and Mega-environment analysis (which-won-where
[224]
pattern), whereby specific genotypes can be recommended for specific mega-
environments (Yan and Tinker 2006).

Distinguishing features and performance of the new variety
The new tef variety (DZ-Cr-387 X 3774-13(RIL120B)) is the descendant product
of hybridization between DZ-Cr-387 and 3774-13. Quncho was selected as a
maternal parent for its high yielding ability and wide adaptability, while 3774-13
was selected as a pollen parent for its very white seed color and earliness. The
latter parent was obtained through TILLING (Target Induced local lesions IN
genomes) technique from the University of Bern. The major distinguishing
characters of (DZ-Cr-387 X 3774-13(RIL120B)) are yellowish lemma color,
yellowish anther color, and a very white seed color.
The summary of the agronomic performance results of the test genotypes is shown
in Table 1. Highly significant variations among the genotypes were observed in
shoot biomass and total grain yield in all study years and locations. The average
grain yield of (DZ-Cr-387 X 3774-13(RIL120B)) was 2.0 -2.5 t ha-1 at research
centers and 1.9-2.2 t ha-1 on farmers’ fields. On the average, (DZ-Cr-387 X 3774-
13(RIL120B)) reaches the panicle emergence stage in 37 days and physiological
maturity in 75 days after sowing. From the average plant height of 97.9cm the
panicle of (DZ-Cr-387 X 3774-13(RIL120B)) contributes to 40%. (DZ-Cr-387 X
3774-13(RIL120B)) is released for low rainfall and terminal low moisture stress
areas of the country. It performs very well in areas having an altitude 1200-1800m
above mean sea level.
(DZ-Cr-387 X 3774-13(RIL120B)) showed grain yield advantage of 8.5% and

30.7% over the standard (Boset) and local checks, respectively. Based on a three-
year multi-location trial, RIL120B was selected for its high grain and moisture
stress tolerance as well as other desirable traits. Hence, RIL 120B was given the
vernacular name (DZ-Cr-387 X 3774-13(RIL120B)) and put under a variety
verification trial for release as a new improved tef variety. Based on the
application, the National Variety Release Committee in Ethiopia evaluated the
three-year performance of RIL 120B and visited several locations where the new
variety was grown for evaluation. Consequently, the committee approved the
release of RIL 120B as a new variety with the name ‘DZ-Cr-453(RIL 120B)’.
[225]
Table1.Mean agronomic performance of 14tef genotypes evaluated in the national variety trial (Early set) over three years
(2014, 2015 and 2016).
Code Genotypes DPE DTM PH PL LI BY GY
1 DZ-Cr-409(Boset, standard check) 34.3 73.4 92 36 87 11250 1940
2 DZ-Cr-387 X 3774-13(RIL49B) 38.3 75.9 99.2 41.8 83.5 12314.3 1688.3
3 DZ-Cr-387 X 3774-13(RIL8C) 36.7 74.3 94.6 38.7 85.5 11895.8 1783
4 DZ-Cr-387 X 3774-13(RIL74C) 34.9 74.2 96.9 37.8 81.8 12506.9 1813.2
5 DZ-Cr-387 X 3774-13(RIL93) 33.9 74.3 98.6 39.6 87.1 11187.5 1711.2
6 DZ-Cr-387 X 3774-13(RIL110A) 37.3 75.2 100 39.9 82.3 12847.2 1762.9
7 DZ-Cr-387 X 3774-13(RIL120B) 37.3 75.3 97.9 41 87 13186 2105
8 DZ-Cr-387 X 3774-13(RIL55A) 37.8 75.1 96 40.5 86.4 12007.1 1738.5
9 DZ-Cr-387 X 3774-13(RIL17B) 35.1 74.4 95.7 42.2 88.6 11312.5 1727.9
10 DZ-Cr-387 X 3774-13(RIL4A) 37 76.1 98.7 39.5 85.7 11528.6 1709.6
11 DZ-Cr-387 X 3774-13(RIL56B) 35.3 74.7 98.4 40.5 87.2 12342.9 1882.6
12 DZ-Cr-387 X 3774-13(RIL77) 36 74.4 96.1 38.1 86 11611.1 1714.8
13 DZ-Cr-387 X 3774-13(RIL118) 37.9 75.5 99.6 40.3 79 12812.5 1691.5
14 Local check 34.3 72.9 92.3 36 91 11243 1611
Mean 36.1 74.7 96.9 39 86 12001 1777
Coefficient of variation (%) 15.5 11.7 21.5 18 18 24.99 29.7
LSD (0.05) 2.34 4.23 9.65 3.3 6.8 2789.4 598.3
DPE =days to panicle emergence, DTM = days to maturity, LI(%)= lodging index, PH(cm)= Plant height,
PAL(cm)=panicle length, BY(kg/ha)=biomass yield, and GY(kg/ha)= grain yield
GGE biplot analysis of tef genotypes

As shown on Figure 1, the genotypes code 7 (DZ-Cr-387 X 3774-13(RIL120B)),
code 1 (DZ-Cr-409), code 11 (DZ-Cr-387 X 3774-13(RIL56B)), code 4 (DZ-Cr-
387 X 3774-13(RIL74C)) and code 6 (DZ-Cr-387 X 3774-13(RIL110A)) showed
positive interaction with most environments
(Debre Zeit/Bishoftu, Alemtena, Dhera, Axum and Mehoni). But the remaining
genotypes did not show positive interaction to most of the environments.
Figure 1.GGE biplot analysis of 14 tef genotypes tested at seven locations
[226]
Identification of stable genotypes with the highest performance
The single-arrowed line is the AEC abscissa (or AEA) and points to higher mean
yield across environments (Figure 2). Thus, ‘the genotype code "7 (DZ-Cr-387 X
3774-13(RIL120B))"’ had the highest mean yield, followed by genotype code "1
(DZ-Cr-409)" and "6 (DZ-Cr-387 X 3774-13(RIL110A)", whereas genotype code
"14 (Local)" and "10 (DZ-Cr-387 X 3774-13(RIL4A))” had the lowest mean
yield. The AEC ordinate passes through the plot origin and is perpendicular to the
AEC abscissa and points to greater variability (poorer stability) in either direction.
Thus, genotype code "6 (DZ-Cr-387 X 3774-13(RIL110A))" was highly unstable,
whereas genotype code "7 (DZ-Cr-387 X 3774-13(RIL120B))" was highly stable.
Which genotypes (s) won where? .

The genotype code "7 (DZ-Cr-387 X 3774-13(RIL120B))", "6 (DZ-Cr-387 X
3774-13(RIL110A))", "14 (Local)", "10 (DZ-Cr-387 X 3774-13(RIL4A))" and "5
DZ-Cr-387 X 3774-13(RIL93)" were as the corner or vertex genotypes.
Environment Dhera, Axum, Minjar and Sirinka fell in the sector in which
‘genotype code "6 (DZ-Cr-387 X 3774-13(RIL110A))" was the vertex genotype
(Figure 3). This means that the genotype code "6(DZ-Cr-387 X 3774-
13(RIL110A))" was the best genotype for those environments. The five
environments fell in the sector in which ‘code 7(DZ-Cr-387 X 3774-
13(RIL120B))’ was the vertex cultivar, meaning that ’code 7 (DZ-Cr-387 X 3774-
13(RIL120B))’ was the best cultivar for these five environments.
Figure 2. Identification of stable genotypes with the highest performance
[227]
Figure 3. Which won where pattern of the GGE biplot 14 tef genotypes tested at seven locations
Ranking genotypes
The arrow is where an ideal genotype should be. Its projection on the AEA was
designed to be equal to the longest vector of all cultivars, and its projection on the
AEC was zero, meaning that it is stable. Therefore, genotypes located closer to the
ideal genotype are more desirable than others. Thus, genotype code "7 (DZ-Cr-
387 X 3774-13(RIL120B))" was more desirable than the genotype code "1(DZ-
Cr-409)" and "6 (DZ-Cr-387 X 3774-13(RIL110A))"(Figure 4).
[228]
Figure 4. Ranking genotypes
Brief description of the variety

A brief description of the new tef variety (DZ-Cr-387 X 3774-13(RIL120B))
including its names, distinguishing pheno-morphic features and yield is given in
Table 2.
Table 2.Summary of description of the new tef variety(DZ-Cr-387 X 3774-13(RIL120B))
No. Descriptor parameter Description

1 Breeder’s name DZ-Cr-453 RIL120B
2 Pedigree DZ-Cr-387(Quncho)x 3774-13-RIL120B
3 Vernacular name given (DZ-Cr-387 X 3774-13(RIL120B))
4 Days to panicle emergence 34-42
5 Days to maturity 70-80
4 Plant height (cm) 90-116
5 Panicle length (cm) 38-46
6 Lemma color Yellowish
7 Anther color Yellowish
8 Caryopsis color Very white
9 Grain yield – On-station (kg/ha) 2000-2500
10 Grain yield - On-farm (kg/ha) 1900-2100
11 Straw yield (kg/ha) 13100-14000

Genotype by environment interaction has a key effect on crop variety
development by complicating the release of varieties across challenging
environments. Analysis of variance for each seven locations and combined over
seven locations showed significant differences among genotypes, environments,
and genotypes x environments interaction (GEI) for grain yield and most of the
[229]
yield-related traits. The significant genotypes x environments interaction effects
indicated the inconsistent performance of genotypes across the tested
environments. Among the tested genotypes, the genotype code "7 (DZ-Cr-387 X
3774-13 (RIL120B))"’ had the highest mean yield, followed by genotype code "1
(DZ-Cr-409)" and "6 (DZ-Cr-387 X 3774-13(RIL110A)", whereas genotype code
"14 (Local)" and "10 (DZ-Cr-387 X 3774-13 (RIL4A))” had the lowest mean
yield. Considering the seven environments data and field performance evaluation
during the variety verification trial, the national variety releasing committee has
approved the official release of candidate genotype, genotype code "7(DZ-Cr-387
X 3774-13(RIL120B))", with the vernacular name of “Bora” for moisture deficit
areas of the country.
Acknowledgments
We are very grateful to Syngenta Foundation for Sustainable Agriculture
supported EIAR-University of Bern Collative Tef Research Project, DZARC, and
other federal and regional research centers for supporting the implantation of the
experiment, and also all researchers and technical assistants for participating in
data collection and analyses.
.
References
AbuhayTakele, HirutKebede and Belay Simane. 2001. Physiological research in tef. In: Narrowing
the Rift: Tef Research and Development (HailuTefera, Getachew Belay and Mark Sorrells,
eds.), Proceedings of the International Workshop on Tef Genetics and Improvement, 16-19
October 2000, Addis Ababa, Ethiopia, pp 177-189.
CSA. 2015. Central Statistics Agency, Agricultural Sample Survey. Report on Area, Production
and Yield of Major Crops (Private Peasant Holdings,Meher Season), Addis Ababa, Ethiopia
CSA. 2017. Central Statistical Agency, Agricultural Sample Survey 2016/2017 (2009 E.C).
Volume I. Report on Area and Production of Major Crops (Private Peasant Holdings, Meher
Season). Statistical Bulletin 584. Addis Ababa, Ethiopia.
Fufa Hundera, HailuTefera, KebebewAssefa, TesfyeTefera and TirunehKefyalew. 2000. Genetic
variability and correlation of morpho-agronomic characters in tef landraces. Tropical
Science, 39: 140-146
Geremew Bultosa, Hall AN and Taylor JRN. 2002.Physico-chemical characterization of grain tef
[Eragrostistef(Zucc.) Trotter] starch, Starch-Starke, 54: 461-468.
KebebewAssefa Yu JK, Zeid M, Getachew Belay, HailuTefera and Sorrells ME. 2011. Breeding
tef [Eragrostistef (Zucc.) trotter]: conventional and molecular approaches. Plant Breeding,
130: 1-9.
KebebewAssefa, Solomon Chanyalew, Yazachew Genet, MitikuAsfaw, TsionFikre, HabteJifar,
NigussuHussen, Mihireteab H/Sellassie, and MolalignAssefa. 2017. Tef (Eragrostistef)
variety “Kora”.Ethiop.J.Agri.Sci. 27(2):137-140.
SeyfuKetema. 1997. Tef: [Eragrostistef (Zucc.) Trotter]: Promoting the Conservation and Use of
Underutilized and Neglected Crops12 Institute of Plant Genetics and Crop Plant Research,
Gatersleben/International Plant Genetic Resources Institute, Rome, Italy.
Solomon Chanyalew, KebebewAssefa, Mitiku Asfaw, Yazachew Genet, Kidist Tolossa,Worku
Kebede, TsionFikre, Nigussu Hussen, Habite Jifar, Atinikut Fentahun, Kidu Gebremeskel,
[230]
Girma Chemeda and Tegegn Belete. 2017. Tef (Eragrostis tef) variety
“Dagim”.Ethiop.J.Agri.Sci. 27(2): 131-135.
Spaenij-Dekking L., Kooy-Winkelaar Y. and Koning F. 2005.The Ethiopian cereal tef in celiac
disease.The New England Journal of Medicine, 353: 1748-1749.
Vavilov, NI. 1951. The origin, variation, Immunity and Breeding of cultivated plants. Ronald
Press, New York. [Translated from Russian by Sarrchester, K].
Worku Kebede, KidistTolossa, Nigussu Hussein, TsionFikre, Yazachew Genet, Abate Bekele,
KiduGebremeskel, AtinkutFentahun, HabteJifar, Sonia Plaza-Wüthrich, RegulaBlösch,
Solomon Chanyalew, KebebewAssefa, and ZerihunTadele.(2018). Tef (Eragrostis tef)
Variety ‘(DZ-Cr-387 X 3774-13(RIL120B))’.Ethiop.J.Agri.Sci. 28(2): 107-112.
Yayis R, Fitsum Alemayehu, Gurmu F et al., “Registration of ‘AMBERICHO’ a newly released
tef (Pisum sativum L) variety for the Southern Highlands of Ethiopia,” Journal of Plant
Studies, vol. 4, no. 2, pp. 42-43, 2015.
Yan W and MS Kang. 2003. GGE biplot analysis: A geographical tool for breeders geneticists, and
agronomists. Boca Raton, FL; CRC Press.
Yan W and Tinker NA. 2006. Biplot analysis of multi-environment trial data: Principles and
applications. Can. J. Plant Sci.86:623-45.
ZerihunTadele and KebebewAssefa. 2012. Increasing food production in Africa by boosting the
productivity of understudied crops. Agronomy, 2: 240–28.
[231]
[232]
Semi-dwarf Tef (Eragrostis tef) Variety
Development for High Potential Areas of Ethiopia
Worku Kebede1, Tsion Fikre1, Kidist Tolosa1,Yazachew Genet1, Solomon Chanyalew1,
Mengistu Demissie1, Kebebew Assefa1, Habte Jifar2, Nigus Belay2, Kidu
Gebremeskel3,Girma Chemeda4, Molalign Assefa5, Atinkut Fentahu6
and Zerihun Tadele7
1
EIAR, DebreZeit Research Center, P.O.Box 32, Ethiopia; 2EIAR, Holetta Research center, P.O.Box 2003,
Addis Ababa, Ethiopia; 3TARI, Axum Research Center, P.O.Box30, Axum, Ethiopia
4
OARI, Bako Research Center, P.O.Box03, Bako Ethiopia; 5SARI, Worabe Research Center, P.O.Box021,
Worabe, Ethiopia; 6AARI, Adet Agricultural Research Center, P.O. Box 8, Bahir Dar, Ethiopia
7University of Bern, Institute of Plant Sciences, Altenbergrain 21, CH-3013 Bern, Switzerland
Abstract
Genetic improvement of native crops is a new and promising strategy to combat
hunger in the developing world. Tef is the major staple food crop for more than 70
million people in Ethiopia. As an indigenous cereal, it is well adapted to diverse
climatic and soil conditions; however, its productivity is extremely low mainly due to
susceptibility to lodging. Tef has a tall and weak stem liable to lodge, and this is
aggravated by wind, rain, or application of nitrogen fertilizer. The objective of this
experiment was to determine and understand the effect of genotype, environment, and
their interaction on grain yield of tef and to identify and release stable, semi-dwarf
and high yielding tef genotype for high potential areas of the country. Twelve tef
genotypes including two checks were laid out in randomized complete block design
using four replications for two years (2016 and 2017) at twelve locations of high
potential areas of the country. The combined data analysis over locations and years
indicated that the candidate variety DZ-Cr-458(RIL.18) performed better than the two
checks and the other test genotypes. The National Variety Release Committee in
Ethiopia investigated the two-year performance of DZ-Cr-458(RIL.18) and visited
several locations where the new candidate variety was grown for evaluation under a
variety verification trial. Based on the evaluations, the National Variety Release
Committee the candidate variety was approved for release in May 2019. The variety
christened “DZ-Cr-458(RIL.18)” was obtained from F7 recombinant inbred lines of
the targeted intra-specific cross Kaye Murri and 3774-13 (Kegne) made in 2011. The
female parent Kaye Murri was selected for its extra white seed color, thick culm, and
vigorous growth habit and the anther parent 3774-13 were identified for its short
height through TILLING. It gave 7.5% and 10.65% grain yield advantages over the
standard check (Quncho) and local check, respectively.
Keywords: Tef, Ebba, GGE biplot, Stability Variety Release
Introduction
Tef is one of the most important crops in which 70% of Ethiopian used it as stable
food. It is a resilient crop that performs better than other cereals under marginal
conditions including drought, waterlogging, and poor soil fertility. Since it
produces a reasonable yield when grown in areas that experience moisture
[233]
scarcity, it is considered as a low-risk crop (Seyfu, 1997). Tef is nutritious due to
its high protein and mineral content (Geremew et al., 2002; Abebe et al., 2007),
and the absence of gluten (Spaenij Dekking et al., 2005) makes it an alternative
food for people suffering from coeliac disease.
Tef has become globally known, and various products are available in Europe and
North America as health foods especially for persons with gluten intolerance
(Saturni et al., 2010). In South Africa, India, Pakistan, Uganda, Kenya, and
Mozambique tef is mainly grown as a forage or pasture crop (Kebebew et al.,
2011). Tef grows under a wide range of ecological conditions from sea level up to
3000 meters above sea level (m.a.s.l). It is annually cultivated on over three
million hectares of land, and as such, it accounts for about 30% of the total area
and 20% of the gross grain production of cereals grown in the country (CSA,
2017). However, the productivity of tef is relatively low compared to most other
crops. The national breeding program for tef has since nearly six decades been
striving in many ways to increase the production and productivity of tef by
eliminating and/or reducing the yield-limiting factors. Lodging is one of the most
yield-limiting factors, which occurs when plants grow long in height and fail to
hold the panicle. The effect of the length of basal internodes on lodging has been
demonstrated in rice (Hoshikawa, 1990). According to our knowledge, the effect
of the lengths of the lower basal culm internodes on the lodging tolerance
behavior of tef plants has not been addressed or reported so far. Reducing the
height of tef is one of the mechanisms presumed to solve the lodging problem.
In plant breeding programs many new genotypes are usually evaluated in different
environments (location and years) to identify and advance desirable ones towards
release. A genotype or cultivar is stable if it has adaptability for a trait of
economic importance across diverse environments. The environmental component
(E) generally represents the largest component in analyses of variance, but it is not
relevant to cultivar selection; only G and GE are relevant to meaningful cultivar
evaluation and must be considered simultaneously for making selection
decisions(Yan and Kang 2003). Genotype by environment (GxE) interaction
affects the efficiency of crop improvement programs that may lead to complicate
the recommendation of varieties across diverse environments. Therefore,
information on the structure and nature of GxE interaction is particularly useful to
breeders (Yayis et al., 2014). Consequently, the objective of the present work has
been to determine and understand the effect of genotype, environment, and their
interaction on grain yield of tef and to identify and release stable, semi-dwarf and
high yielding tef genotype, and farmers and consumers preferred tef varieties with
better grain quality for high potential areas of the country.
[234]
The field experiment was carried out over two years (2015/16 & 2016/17) at
twelve tef growing sites in high potential areas of the country (Debre Zeit, Minjar,
Holeta, Adet, Fenote Selam, Akaki, Assosa, Jimma, Ginchi, Hatsebo, Adadi
Mariam and Bichena). The genetic materials used in the study were obtained from
the intra-specifi cross made in 2011. The crossing method used to develop these
materials was standard surgical hand emasculation and pollination technique.
Female and male parents used for the crossing were Kaye Murri and 3774-
13(“Kegne”), respectively. Kaye Murri was selected as a maternal or ovule parent
for its extra-white seed color, thick culm, and vigorous growth habit, while 3774-
13 (“Kegne”) was identified at the University of Bern in Switzerland from
screening 5000 mutagenized tef populations (Jost et al., 2015) using the method of
TILLING (Targeted Induced Local Lesions IN Genomes). The candidate variety
with the pedigree DZ-Cr-458 (Kaye Murri X 3774-13 (RIL8)) was identified from
F7 homozygous recombinant inbred lines (RILs) advanced using the single seed
descent method (SSD) from F2-derived seeds. The crossing combinations and
names of recombinant inbred lines, as well as control materials used in the current
study, are shown in Table 1.
Genotypes, Testing Sites, and Experimental Design.

Crossing and early generation testing for all breeding populations were performed
at Debre Zeit Agricultural Research Center from where the National Tef Breeding
Program is coordinated. The performance of 12 tef genotypes which includes 10
inbred lines from the two independent crosses to semi-dwarf tef lines as well as
two controls (farmers’ check and improved Quncho variety were tested at several
locations (Debre Zeit, Minjar, Holeta, Adet, Fenote Selam, Akaki, Assosa, Jimma,
Ginchi, Hatsebo, Adadi Mariam and Bichena) using Randomized Complete Block
Design with four replications. The trial was conducted on the plot size of 2m*2m
with 10 rows per plot throughout all trial sites and 1.5m between replication, 1m
between plot, 20cm between rows. Agronomic and yield data were collected and
subjected to statistical analysis to identify the best genotypes of the evaluated
genotypes. A variety verification trial was conducted at (Debre Zeit, Minjar,
Holeta, Adet, and Bichena) on the trial station and in ten farmers’ field during the
2018/19 main production season.
Data on grain yield and yield-related traits were collected on plot and plant basis
from each plot, respectively. Date of heading and maturity were taken when each
plot attained 50% heading (panicle emergency), 90% physiological maturity, and
lodging index taken during 90% physiological maturity, respectively, and days
were calculated beginning from the date of sowing. Data for plant height (cm),
Panicle length (cm) were collected based on five sample plants which were
randomly taken from each plot and the average of five sample plants was used for
analysis. Thousand seed weight (g) and grain yield (g) of each plot were measured
[235]
on clean, dried seed and the measured grain yield value (g) has converted to
kilogram per hectare for analysis. All agronomic practices were done as per the
recommendation for tef.
Data from individual environments and combined over twelve locations were
analyzed by using SAS (2009) software. The analysis of variance for grain yield
and yield-related traits for each environment and over twelve environments was
analyzed by using a randomized complete block design. The combined analysis of
variance across the environment was done to determine the differences between
genotypes across environments, among environments and their interaction.
Bartlett’s test was used to assess the homogeneity of error variances before doing
combine analysis over environments. Mean comparison using Duncan's Multiple
Range Test (DMRT) was performed to explain the significant differences among
means of genotypes and environments.
Analysis of Variance (ANOVA) of data of individual location, Yij =µ + Gi +Bj +

eij model was used, where; Yij = observed value of genotype i in block j, µ =
Grand mean of the experiment, Gi = the effect of genotype i, Bj = the effect of
block j, eij =the error of genotype i in block j. A combined analysis of variance
over locations was carried out using the following statistical model. Yijk=µ + Gi
+Ej +GEij+ Bk(j)+ eijk, where; Yijk = observed value of genotype i in block k of
environment (location) j, µ = Grand mean of the experiment, Gi = the effect of
genotype i, Ej =the effect of the jth environment GEj = the interaction effect of
genotype i with environment j, Bk(j) =the effect of block k in location
(environment) j, eijk =the error effect of genotype i in block k of environment j. R
software (3.5 version) was used to visualize GEI patterns. Based on principles of
GGE biplot, for the yield characters; Environmental evaluation (The power of
environments to discriminate among genotypes), Genotype evaluation (the mean
performance and stability) and Mega-environment analysis (which-won-where
pattern), whereby specific genotypes can be recommended for specific mega-
environments (Yan and Tinker 2006).

Analysis of Variance
The combined analysis of variance for grain yield of the 12 tef genotypes across
12 testing environments revealed highly significant (P <0.01) effects due to
genotypes, environments, and genotype by environment interactions. The
significant variability among the tef varieties in the present study is in line with
the previous reports in tef (Kefyalew, 2001; Habte at el, 2019).
[236]
Distinguishing features and performance of the new variety
The new tef variety (Kaye Murri x 3774-13(RIL18)) is the descendant product of
hybridization between Kaye Murri and 3774-13. Kaye Murri was selected as a
maternal or ovule parent for its extra-white seed color, thick culm, and vigorous
growth habit, while 3774-13 was selected as a pollen parent for its very white seed
color and earliness. The latter parent was obtained through TILLING (Target
Induced local lesions IN genomes) technique from the University of Bern. The
major distinguishing characters of (Kaye Murri x 3774-13(RIL18)) are semi-
dwarf, yellowish lemma color, yellowish anther color, and a very white seed color.
Highly significant variations among the genotypes were observed in shoot

biomass and total grain yield in all study years and locations. The average grain
yield of (Kaye Murri x 3774-13(RIL18)) was 2.0-2.6 t ha-1 at research centers and
1.9-2.3 t ha-1 on farmers’ fields. On the average, (Kaye Murri x 3774-13(RIL18))
reaches the panicle emergence stage in 48 days and physiological maturity in 102
days after sowing. From the average plant height of 85-100 cm, the panicle of
(Kaye Murri x 3774-13(RIL18)) contributes to 36-44%. (Kaye Murri x 3774-
13(RIL18)) is released for mid to optimum rainfall and high potential areas of the
country. The candidate variety gave 7.5% and 10.65% grain yield advantages over
the standard check (Quncho) and local checks, respectively. The candidates has
also comparable shoot biomass yield. It has got an immense farmer’s attention due
to its high yielding potential, white seed color, and good straw yield. It is to be
noted that the straw yield is equally important as grain in tef during participatory
variety selection trials. Moreover, the candidate had a newly identified character
which is semi-dwarf in height, which might help in reducing yield loss due to
lodging.
Based on a two-year multi-location trial, Kaye Murri x 3774-13(RIL18) was

selected for its high grain and biomass yield as well as additional traits indicated
below. Hence, RIL-18 was selected and put under a variety verification trial for
release as a new variety. Based on the application for release, the National Variety
Release Committee in Ethiopia scrutinized the two-year performance of RIL 18
and visited several locations where the new variety was grown for evaluation.
Consequently, the committee approved the release of RIL18 as a new variety. The
new variety was christened as ‘Ebba’ as a standing witnessed for the earliest
known tef scientist, Dr. Tadesse Ebba.
[237]
Table 1.Mean performance of tef genotypes in the white-seeded semi-dwarf late set national tef variety trial over twelve locations
Gen. Grain yield Biomass yield Lodging Plant height (cm)

Code Genotypes (kg/ha) (kg/ha) index
1 DZ-Cr-387 2109 8645 65 103
2 GA-10-3 X Kaye Murri(RIL182) 2163 8141 65 92
3 DZ-Cr-387 X GA-10-3(RIL176) 2124 8323 65 104
4 GA-10-3 X Kaye Murri(RIL.184) 2046 6899 66 90
5 DZ-Cr-387 X GA-10-3(RIL182) 2187 7756 70 91
6 DZ-Cr-387 X GA-10-3(RIL178) 2190 8236 63 99
7 DZ-Cr-387 X GA-10-3(RIL210) 2112 8874 64 104
8 DZ-Cr-387 X GA-10-3(RIL157) 2221 8527 65 100
9 Kaye Murri x 3774-13(RIL18) 2265 7868 60 92
10 GA-10-3 X Kaye Murri(RIL281) 2086 7821 63 94
11 DZ-Cr-387 X GA-10-3 (RIL91) 2231 8674 65 101
12 Local check 2047 7569 72 92
Mean 2150 8111 65 97
Coefficient of variation (%) 16.72 25 10 8
LSD (0.05) 124.9 66.8 5.2 6.0
Coefficient of determination (R2) 90.0 75 73 75
Ranking of genotypes relative to ideal genotypes

The average environment coordination view of the GGE biplot shows the ranking
of genotypes based on the performance of an ideal genotype (Figure 1). The
relative adaptation of the ideal genotype is evaluated by drawing a line passing
through the biplot origin and the best genotype marker. This line is called a
genotype axis and is connected to the best genotype (Yan et al; 2000). Such
ranking of genotypes based on the performance of the ideal genotype revealed that
genotype code 8(DZ-Cr-387 X GA-10-3(RIL157)), 9(Kaye Murri x 3774-
13(RIL18)), 6(DZ-Cr-387 X GA-10-3(RIL178)) and 11(DZ-Cr-387 X GA-10-3
(RIL91) in that diminishing order were among the top-yielding genotypes.
Ranking Genotypes
1500
Minjar
1000
8
11
500
3
AXIS2 25.2 %
1 10 6
Holeta
Jimma
Assosa
DZBS
Ginchi
0
Bichena
F_Selam
12 Adet 9
Akaki Hatsebo
4
-500
Adadi
7
2
-1000
-1000 -500 0 500 1000 1500 2000
AXIS1 51.87 %
Figure 1; Ranking genotypes relative to ideal genotypes
[238]
M ean vs. Stability
1000 1500 Minjar
8
11
500
3
AXIS2 25.2 %
1 10 6
Holeta
Jimma
Assosa
DZBS
Ginchi
0
Bichena
F_Selam
12 Adet 9
Akaki Hatsebo
4
-1000 -500
Adadi
7
2
-1000 -500 0 500 1000 1500 2000
AXIS1 51.87 %
Figure 2.Ranking of genotypes based on mean performance and stability
Genotypes mean yield and stability

The average environment coordinate (AEC) is a line that passes through the origin
and points to the higher mean yield across environments, and it shows the increase
in rank of genotypes towards the positive end (Yan et al., 2000). This line was
reported to be useful to evaluate mean grain yield and stability of genotypes (Yan,
2000, 2001; Yan and Hunt, 2001). According to such reports, genotypes
considered to be stable are those that appear closer to the origin with the shortest
vector from the AEC. Accordingly, Figure 2 shows the mean performance and
stability of the genotypes in the present study. Based on this, the candidate variety
genotype code 9 (Kaye Murri x 3774-13 (RIL18)), genotype 6 (DZ-Cr-387 X GA-
10-3 (RIL178)), and genotype 8 (DZ-Cr-387 X GA-10-3 (RIL157), with the
shortest vector from the AEC axis, were identified as the most stable genotypes,
while genotype code 5 (DZ-Cr-387 X GA-10-3 (RIL182) and 2 (GA-10-3 X Kaye
Murri (RIL182)) with the longest vector from AEC were the most unstable
genotypes. On the other hand, genotype code 8 (DZ-Cr-387 X GA-10-
3(RIL157))and 9 (Kaye Murri x 3774-13 (RIL18)) followed by genotype code 6
(DZ-Cr-387 X GA-10-3(RIL178)) and 11 (DZ-Cr-387 X GA-10-3 (RIL91) scored
higher grain yield.
GGE biplot pattern of tef genotypes

On Figure 3, Genotypes code 9 (Kaye Murri x 3774-13(RIL18), 8 (DZ-Cr-387 X
GA-10-3(RIL157)), 5 (DZ-Cr-387 X GA-10-3(RIL182)), 6 (DZ-Cr-387 X GA-10-
3(RIL178), 11 (DZ-Cr-387 X GA-10-3 (RIL91)and 2 (GA-10-3 X Kaye
Murri(RIL182) showed positive interaction with most of the environments (viz.,
Minjar, Hatseb, Akaki, Adadi Mariam, Adet, Ginchi, Fenote Selam and Debre
[239]
Zeit). However, the remaining genotypes did not show positive interaction with
most of the environments.
GGE Biplot
500 1000 1500
Minjar
8
11
3
AXIS2 25.2 %
1 10 6
Holeta
Jimma
Assosa
DZBS
Ginchi
0
Bichena
F_Selam
12 Adet 9
Akaki Hatsebo
4
-1000 -500
Adadi
7
2
-1000 -500 0 500 1000 1500 2000
AXIS1 51.87 %
Figure 3.GGE biplot analysis of 12 tef genotypes tested at twelve locations
Brief description of the variety

A brief description of the new tef variety Ebba including its pedigree,
distinguishing agro-morphological features and grain yield data is given in Table
2.
Table 2.Summary of description of the new tef variety Ebba
No. Descriptor parameter Description

1 Breeder’s name DZ-Cr-458 RIL18
2 Pedigree Kaye Murri x 3774-13-RIL18
3 Vernacular name given Ebba
4 Days to panicle emergence 44-52
5 Days to maturity 98-110
4 Plant height (cm) 85-100
5 Panicle length (cm) 36-44
6 Lemma color Yellowish
7 Anther color Yellowish
8 Caryopsis color Very white
9 Grain yield – On-station (kg/ha) 20000-2600
10 Grain yield - On-farm (kg/ha) 1900-2300
11 Straw yield (kg/ha) 7868-11300

Genotype by environment interaction has a key effect on crop variety
development by complicating the release of varieties across challenging
environments. Analysis of variance for every seven locations and combined over
[240]
seven locations showed significant differences among genotypes, environments,
and genotypes x environments interaction (GEI) for grain yield and most of the
yield-related traits. The significant genotypes x environments interaction effects
indicated the inconsistent performance of genotypes across the tested
environments.
Among the tested genotypes, the genotype code 8 (DZ-Cr-387 X GA-10-3

(RIL157)) and 9 (Kaye Murri x 3774-13 (RIL18)) followed by genotype code 6
(DZ-Cr-387 X GA-10-3(RIL178)) and 11(DZ-Cr-387 X GA-10-3 (RIL91) scored
higher grain yield. Considering the seven environments data and field
performance evaluation during the variety verification trial, the national variety
releasing committee has approved the official release of candidate genotype code
"9(Kaye Murri x 3774-13(RIL18))", for high potential areas of the country with
the vernacular name of “Ebba” as a standing witnessed for the earliest known tef
scientist, Dr. Tadesse Ebba.
Acknowledgments
The authors are very grateful to EIAR, University of Bern, DZARC and other
federal and regional research centers for implementing the experiment and also to
all researchers and technician assistance for participating in data collection and
analyses. The financial support of the Syngenta Foundation for Sustainable
Agriculture through a collaborative research project on tef is highly appreciated.
References
Abebe Y, Bogale A, Hambidge KM, Stoecker BJ, Bailey K, Gibson
RS. 2007. Phytate, zinc, iron and calcium content of selected raw and prepared foods
consumed in rural Sidama, Southern Ethiopia, and implications for bioavailability. Journal of
Food Composition and Analysis20, 161–168.
CSA. 2017. Central Statistical Agency, Agricultural Sample Survey 2016/2017 (2009 E.C).
Volume I. Report on Area and Production of Major Crops (Private Peasant Holdings, Meher
Season). Statistical Bulletin 584. Addis Ababa, Ethiopia.
Geremew Bultosa, Hall AN and Taylor JRN. 2002.Physico-chemical characterization of grain tef
[Eragrostistef (Zucc.) Trotter] starch, Starch-Starke 54: 461-468.
HabteJifar, KebebewAssefa, KassahunTesfaye, Kifle Dagne and ZerihunTadele. 2019.Genotype-
by-environment interaction and stability analysis in grain yield of improved tef(Eragrostis tef)
varieties evaluated in Ethiopia.American Journal of Experimental Agriculture35(5): 1-13.
Hoshikawa KW. 1990. Studies on lodging in rice plants. Japan Journal of Crop Science 59(4):
809-814.
Jost M, Esfeld K, Burian A, Cannarozzi G, Solomon C, Kuhlemeier C, Kebebew A,
Zerihun T. 2015. Semi-dwarfism and lodging tolerance in tef (Eragrostis tef) is
linked to a mutation in the alpha-Tubulin 1 gene. Journal of Experimental
Botany 66: 933-944.
Kebebew Assefa, Yu JK, Zeid M, Getachew Belay, HailuTefera and Sorrells ME. 2011. Breeding
tef [Eragrostistef (Zucc.) trotter]: conventional and molecular approaches. Plant Breeding,
130: 1-9.
[241]
Kefyalew T. 200. Genotype x environment interaction in tef.In: Narrowing the Rift: Tef Research
and Development. (HailuTefera, Getachew Belay and Mark Sorrells, eds.). Proceedings of the
International Workshop on Tef Genetics and Improvement, 16-19 October 2000, Addis
Ababa/DebreZeit, Ethiopia
Seyfu Ketema. 1997. Tef: [Eragrostistef (Zucc.) Trotter]: Promoting the Conservation and Use of
Underutilized and Neglected crops. 12 Institute of Plant Genetics and Crop Plant Research,
Gatersleben/International Plant Genetic Resources Institute, Rome, Italy.
Spaenij-Dekking L, Kooy-Winkelaar Y and Koning F. 2005.The Ethiopian cereal tef in celiac
disease.The New England Journal of Medicine 353: 1748-1749.
Yan W and Hunt LA. 2000. Biplot analysis of diallel data. Crop Sci. 42:21–30.
Yan W. 2001.GGE biplot; A windows application for graphical analysis of multi-environment trial
data and other types of two -way data.Agron.J.93:1111-1118.
Yan W. 2002.Singular-value partitioning in biplot analysis of multi-environment trial data.
Agron.J.94:990-996.
Yan W and MS Kang. 2003. GGE biplot analysis: A geographical tool for breeders geneticists, and
agronomists. Boca Raton, FL; CRC Press.
Yan W and NA Tinker. 2006. Biplot analysis of multi-environment trial data: Principles and
applications. Can. J. Plant Sci.86:623-45.
[242]
Yield Potential and Stability of Irrigated Spring
Bread Wheat Genotypes in Central Rift Valley of
Ethiopia
Mihratu Amanuel Kitil1, Tadiyos Bayisa Sarbessa1, Hailu Mengistu Biru1,
Desta Gebre Banje2, and Ambesu Tiliye1
1
EIAR,Werer Agricultural Research Center, Irrigated Wheat Research,
2
Technology Multiplication and Seed Research Directorate; P. O. Box 2003 Addis Ababa, Ethiopia
Abstract
The available potential resources under arid and semi-arid lowland irrigated areas of
Ethiopia could be more exploited by irrigated wheat technology generation on progress and
production to fill the gap of national wheat demand. Bread wheat genotypes evaluated in
different locations of Amibara district in Afar Region and Fentale district of Oromia Region
for three consecutive years 2015/16-2017/18 were considered in the study. The 25 bread
wheat genotypes were evaluated in triple lattice with three replications on 9 m 2 plot area.
The study was targeted to identify and select high yielding potential and stable candidate
varieties for release and also further breeding purposes. The analysis of variances of bread
wheat genotypes evaluated revealed a highly significant difference (P≤ 0.01) among
genotypes for all traits and the genotypes by environment interactions. The overall mean
performance of genotypes evaluated across different environments the two genotypes
(HEILO//MILAN/MUNIA/3/KIRITATI/2*TRCH) 3877 kg/ha and (MUU/FRNC
LN//FRANCO LIN #1) with 3655 kg/ha were better out yielded with 12% and 5% than
check respectively. The third promising genotype (GLADIUS/2* BAVIS) is the most stable
and it is better in early maturing (12%), 2nd best plant height (13%) next to
MUU/FRNCLN// FRANCOLIN #1, and it has the maximum thousand kernel weight (8%)
and related quality traits comparing to check. The out yielded genotypes and performed well
in their important traits were selected as candidate varieties and submitted for variety
releasing committee waiting for the final decision. Therefore from the current results, it
has been observed better yield potential than the check variety and showing stability among
the studied genotypes and this could be exploited in future large scale seed production.
Keywords: Environment, Genotypes, Stability, Yield.
Introduction
Wheat is one of the major cereals grown for use as food and industrial raw
materials in Ethiopia. It is an important staple food in the diets of many
Ethiopians, providing an estimated 12% of the daily per capita caloric intake for
the country’s over 90 million population (FAO, 2017). It is the important staple
crop in Ethiopia and ranked 4th in the area (13.38%) and grain production
(15.17%) of total grain crops which has increased production mainly by
smallholder farmers using rain-fed based production system (CSA, 2018).
Ethiopia is still importing about 1.6 ml tons of wheat which estimated to 25% in
deficit to fulfill domestic wheat demand by foreign currency (USDA, 2018).
[243]
Several Bread wheat (Triticum aestivum L.) varieties have been released for
rainfed production targeted for different midland and highland area agro-
ecologies. Although there is a potential for irrigated wheat production in the
lowlands, the capacity to produce at a commercial scale is at its very initial stage
in the country. Main wheat production constraints are biotic and abiotic stresses
across rain-fed and irrigated environments which accentuated by the increasing
incidence of climate change heat and drought (Wuletaw et.al, 2018). So; the
climate change effect abiotic and biotic stress tolerant high yielding variety
development is vital in continuous as per Wuletaw et.al, 2018. Irrigated Lowland
area wheat production has limitations in the high yielding and stability of released
wheat varieties and their seed system. It is very important to increase the irrigated
lowland wheat varieties as a choice for variety development as well as for
production and also for irrigable areas with better seed supply. The Irrigated wheat
Research and development works done in the last decade achieved better in a
continuous variety of development and pre-scaling up which utilized by
policymakers and get attention for irrigated wheat production and diversification
in lowland areas of Ethiopia. The progress and promising irrigated wheat research
results to be sustainable in technology generation to supply the required varieties
and related packages to feed the irrigated wheat development works and
diversification throughout the Ethiopian lowland areas. The high yielding and
stable varieties are very crucial to attain the required production gain. Hence,
evaluation of the extent and performance of available high yielding genotypes
present under lowland areas are essential for effective crop production in an
extensive way for local consumption. Thus, this study aimed to identify high
yielding and stable genotypes in different environments for variety development
works then to release and access related merits for further breeding activities.
Experimental material, sites and Agronomic practice

The field experiments were conducted in Awash River basins at Amibara district
Werer Agricultural research center (WARC) under the Ethiopian Institute of
Agricultural Research (EIAR). The WARC is located at 740 m a.s.l (9o16’8”N,
40o09’41” É). In Oromia regional state the trials were done at Saru-weeba Fentale
district. The 25 bread wheat Genotypes in triple lattice with three replications
were planted under irrigated lowland area on 3m x 3m = 9m2 plot area. The trials
were planted at a relatively cool season of the middle awash from mid-October to
November per consecutive years.
The fertilizers (UREA=100N Kg/ha, DAP=50P2O5 kg/ha) were applied based on
previous practice in the irrigable areas. UREA Fertilizer application was on a split
basis; 1/2 at 25-30 days after planting and 1/2 at booting stage and DAP applied
all at planting. All experimental plots irrigated uniformly using furrow irrigation
methods in 10days interval until the wheat crop reached physiological maturity.
[244]
Other management practices performed as per previous recommendations. Data
were recorded for Days to 50 % heading, Days to 75% Maturity, Spike length,
spikelet number per spike, Plant height (cm), Number of kernel per spike,
Thousand kernel weight (g) and Grain yield (kg ha-1).
Statistical Analysis of Data

The recorded all yield components and average yield across locations data were
subjected to analysis of variance (ANOVA) and Varieties by environment
interaction (GGE) biplot analysis using appropriate software GenStat statistical
package18th Edition (VSN International 2015). A comparison of treatment means
was done using Fischer’s least significant difference (LSD) test at 5% probability
levels. The combined analysis of variance was carried out to estimate the effects
of environment (E), Genotypes (G) and Genotype by Environment interaction.

Analysis of Variance
The Analysis of variances across location from 25 bread wheat genotypes
evaluated revealed that highly significant (P≤ 0.01) difference among genotypes
for all traits and the genotypes by environment interaction was a highly significant
difference (P≤ 0.01) for all traits (Table 1). This indicated that the measured traits
of bread wheat genotypes were highly influenced by environmental factors.
Significant genotype x environment interaction found for all traits studied would
mean that evaluation of bread wheat genotype on several environments would
give a more accurate estimate of different traits. This results agreed with the study
indicated that bread wheat genotypes responded deferentially to environments
with significant genotype X environment interaction (Leta Tulu and Addishiwot
Wondimu, 2016)
Table 1 Analysis of Variance (ANOVA) of irrigated wheat genotypes
Source of Mean square of Traits

Variation
d.f. DH DAM PLH SL NSPS NKSP GY TKW
Trt 24 114** 72.4** 125** 8*** 16*** 174** 1551795** 133**
Trt.En 125 30** 107** 366** 3.1** 4.7** 83** 1258917** 27**
Trt.Rep 50 7.8 15 40 1.7 3.1 30 858775 15
Trt.Ib 68 12.5 17 29.6 1.5 2.4 92 482880 14
Residual 200 8 15 34 2.1 3.2 31 753892 15
** DH = Days to heading, DM = Days to maturity, PH = Plant height, SL = Spike length, NSPS = Number of Spikeletes per
Spike, NKPS = Number of Kernels per spike, GY = Grain yield and TKW = Thousand kernel weight.
Mean Performance of Genotypes

The mean performance of yield in specific environments (year and location) was
described in Table 2 that clearly showing the performance genotypes across
[245]
environments. The two candidates HEILO//MILAN/MUNIA/3/KIRITATI/
2*TRCH and MUU/FRNCLN//FRANCOLIN #1 were among the superior
genotypes almost in all environments and similarly the third candidate
GLADIUS/2*BAVIS is average yielder in addition to its extra earliness except for
the first environment (Table 2 and 3). The combined mean performance of the
evaluated genotypes different traits across different environment were well
illustrated in table 3. From all the genotypes evaluated in different environments
the two genotypes (HEILO//MILAN/MUNIA/3/KIRITATI/2*TRCH) with a mean
yield of 3877 kg/ha followed by (MUU/FRNCLN//FRANCOLIN #1) the second
most genotype with 3655 kg/ha mean performance out yielded by 12% and 5%
than the standard check respectively. The third promising genotype
(GLADIUS/2*BAVIS) is characterized by its extra early maturing (12%), it is
second-best in plant height (13%) next to MUU/FRNCLN//FRANCOLIN #1, the
best in its thousand kernel weight (8%) and related quality traits mean
performance than check (Table 3). For the yield-related traits, spike length and
Number of spikelets per spike BAJ #1/KIS KADEE#1 and
PFAU/MILAN/5/CHEN/AEGILOPSSQUARROSA(TAUS)//BCN/3/VEE#7/
BOW/ /PAST OR/6/2*BAVIS#1 genotypes have the best mean performance
followed by HEILO//MILAN/MU NIA/3/KIRITATI/2*TRCH (Table 3). The high
yielding genotypes and their promising yield-related traits help to use them in
large scale production and also further breeding similar to the (S. S. PUNIA et al.,
2011) findings. In addition to high yielding, genotypes that are characterized by an
early maturity could be promising because these adaptation mechanisms are
associated with an escape strategy for stress conditions. This study was very
important in arid areas with respect to the existing stresses for different purposes
(M. Afzal arain et al., 2011).
[246]
Table 2: Mean performance of yield (kg/ha) of wheat trial in lowland irrigated areas over different environments.
2015-16 2015-16 2015-16 2015/16 2015/16 2016-17 2017-18 Mean
S.N Genotypes
Werer Fentale WARC FWCF FHF WARC WARC Yield
1 KIRITATI//HUW234+LR34/PRINIA/3/BAJ #1 3161 2503 3797 3067 3718 2596 2549 3064
2 BAJ #1/KISKADEE #1 2893 2708 3713 2951 2787 3605 2926 3145
3 QUAIU//2*BRBT1*2/KIRITATI 2984 1991 3545 2642 3839 2584 3020 2929
4 ND643/2*WBLL1//2*KACHU 2896 2382 3908 2230 4542 3783 3155 3243
5 BAVIS//ATTILA*2/PBW65 3691 2045 3849 2107 3972 4312 3088 3331
6 BAJ #1/KISKADEE #1 2824 2646 3385 3804 4153 3640 3056 3287
7 GLADIUS/2*BAVIS 1424 3088 3272 3286 3527 3003 2813 2706
8 Gaambo 3318 2395 4143 3777 3161 4066 3117 3469
9 HEILO//MILAN/MUNIA/3/KIRITATI/2*TRCH 3958 3168 3812 4512 4089 4526 3297 3877
10 WHEAR/VIVITSI//WHEAR/3/FRNCLN 2844 3087 4328 3377 2788 3594 3020 3329
11 MUU/FRNCLN//FRANCOLIN #1 2716 3274 4574 4072 4062 3651 3217 3655
12 DANPHE #1//ND643/2*WBLL1/3/DANPHE 2146 2263 4082 2723 3334 3419 2517 2916
PFAU/MILAN/5/CHEN/AEGILOPS SQUARROSA
13 2735 3469 4408 3160 2133 2896 3325 3222
(TAUS)//BCN/3/VEE#7/BOW/4/PASTOR/6/2*BAVIS #1
14 BAVIS/4/MILAN/KAUZ//DHARWAR DRY/3/BAV92 1888 3300 4095 3112 4828 4113 3043 3455
15 MUTUS//ND643/2*WBLL1 1810 2736 4357 1871 3677 3129 3113 3039
16 KIRITATI//HUW234+LR34/PRINIA/3/FRANCOLIN #1 2516 1739 3611 2646 3589 3644 3540 2952
17 VEE/MJI//2*TUI/3/PASTOR/4/BERKUT/5/BAVIS 1692 2866 3318 3125 4369 4486 3209 3233
WHEAR/KUKUNA/3/C80.1/3*BATAVIA//2*WBLL1/5/PRL/2*PAST
18 2186 3194 2287 3377 1420 3946 2740 2876
OR/4/CHOIX/STAR/3/HE1/3*CNO79//2*SERI
19 DANPHE #1*2/CHYAK 2879 2973 3478 3684 2036 3945 3169 3280
20 MINO/898.97/4/PFAU/SERI.1B//AMAD/3/KRONSTAD F2004 2783 2716 2500 2912 2143 3547 2605 2838
WHEAR/KUKUNA/3/C80.1/3*BATAVIA//2*WBLL1*2/4/ND643/2*W
21 2995 2233 2818 2826 1626 3506 2707 2751
BLL1
22 TACUPETO F2001*2/BRAMBLING//KIRITATI/2*TRCH 2664 2771 3081 1838 2358 3541 2995 2849
23 PRL/2*PASTOR//WHEAR/SOKOLL 2275 1696 2571 1963 3461 3657 3045 2600
24 DANPHE #1*2/CHYAK 3223 2894 3114 2359 3722 3866 3286 3205
25 W15.92/4/PASTOR//HXL7573/2*BAU/3/WBLL1/5/MUU 3116 2434 3543 1704 4553 3798 3285 3181
Mean 2705 2656 3584 2748 3485 3634 3029 3137
CV% 30.6 27 26 29.25 25.06 17.3 19.4 32.7
LSD 1368 1471 1563 43.33 1922 1163 1191 673
*FWCF- Fentale Women Cooperative farm, FHF- Fentale Hindi Farm, WARC-Werer Agricultural Research center,
[247]
Table 3 Combined Analysis of Bread wheat Genotypes Mean Performance across different locations (Amibara district at Afar Region and Fentale district at Oromia region) 2015/16 -to- 2018/19.
Trt Genotypes DH DAM PLH SL NSPS NKPS Yld_kg/ha TKW

1 KIRITATI//HUW234+LR34/PRINIA/3/BAJ #1 53b-e 85bc 72abc 8fgh 14c-f 36d-g 3064b-e 35d-j
2 BAJ #1/KISKADEE #1 54cde 86c 75abc 10a 16ab 37d-g 3145b-e 38bcd
3 QUAIU//2*BRBT1*2/KIRITATI 53b-e 84bc 77bc 9a-d 15abc 43b-g 2929cde 35d-i
4 ND643/2*WBLL1//2*KACHU 53b-e 85bc 72abc 8.1d-g 15abc 37d-g 3243a-e 30m
5 BAVIS//ATTILA*2/PBW65 54cde 86c 73abc 8.1d-g 13.3fg 34fg 3331a-d 34f-k
6 BAJ #1/KISKADEE #1 54cde 86c 72abc 9.4ab 14c-f 34fg 3287a-d 31lm
7 GLADIUS/2*BAVIS 41a 76a 69ab 7h 12g 32g 2706de 42a
8 Gaambo 55de 86c 79c 9.3abc 16ab 38d-g 3469abc 39bc
9 HEILO//MILAN/MUNIA/3/KIRITATI/2*TRCH 54cde 86c 76abc 9.4ab 16ab 41c-g 3877a 40ab
10 WHEAR/VIVITSI//WHEAR/3/FRNCLN 54cde 87c 72abc 8.7a-f 15abc 39c-g 3329a-d 35d-i
11 MUU/FRNCLN//FRANCOLIN #1 53bcd 86c 68a 8.2d-g 14c-f 34fg 3655ab 37b-f
12 DANPHE #1//ND643/2*WBLL1/3/DANPHE 54cde 85bc 79c 8.4b-g 14.7b-e 38d-g 2916cde 35d-i
13 PFAU/MILAN/5/CHEN/AEGILOPS SQUARROSA
52bcd 87c 75abc 8.1d-g 16.2a 39c-g 3222a-e 36ch
(TAUS)//BCN/3/VEE#7/BOW/4/PASTOR/6/2*BAVIS #1
14 BAVIS/4/MILAN/KAUZ//DHARWAR DRY/3/BAV92 52bcd 85bc 75abc 8.8a-e 14c-f 35efg 3455abc 35d-i
15 MUTUS//ND643/2*WBLL1 52bcd 85bc 71abc 8.4b-g 14.8bcd 36d-g 3039b-e 34f-k
16 KIRITATI//HUW234+LR34/PRINIA/3/FRANCOLIN #1 54cde 85bc 70ab 8.3d-g 14c-f 37d-g 2952cde 33i-m
17 VEE/MJI//2*TUI/3/PASTOR/4/BERKUT/5/BAVIS 51bc 85bc 73abc 7.7fgh 13.7def 35efg 3233a-e 36b-g
18 WHEAR/KUKUNA/3/C80.1/3*BATAVIA//2*WBLL1/5/PRL/2*PA
52bcd 85bc 72abc 7.6gh 13.5ef 47a-d 2876cde 33i-m
STOR/4/CHOIX/STAR/3/HE1/3*CNO79//2*SERI
19 DANPHE #1*2/CHYAK 52bcd 84bc 72abc 8.09efg 14.8bcd 55ab 3280a-e 34e-j
20 MINO/898.97/4/PFAU/SERI.1B//AMAD/3/KRONSTAD F2004 53bcd 87c 73abc 7.6fgh 13.97c-f 51abc 2838cde 32j-m
21 WHEAR/KUKUNA/3/C80.1/3*BATAVIA//2*WBLL1*2/4/ND643/
51b 84bc 73abc 8.09efg 14.5b-f 46b-e 2751de 34g-l
2*WBLL1
22 TACUPETO F2001*2/BRAMBLING//KIRITATI/2*TRCH 56e 86c 73abc 8.2d-g 14.3c-f 44b-f 2849cde 37b-e
23 PRL/2*PASTOR//WHEAR/SOKOLL 52bcd 85bc 71abc 7.5gh 14.2c-f 50abc 2600e 31klm
24 DANPHE #1*2/CHYAK 51b 81ab 71abc 8.3c-g 16ab 59a 3205a-e 33h-m
25 W15.92/4/PASTOR//HXL7573/2*BAU/3/WBLL1/5/MUU 53 86bc 74abc 7.7fgh 14c-f 54ab 3181b-e 36d-i
Mean 53 85 73 8 15 41 3137 35
CV 7.3 8.3 17.5 19.3 13.7 44.6 32.7 13
LSD 2.8 5.1 8.4 1.04 1.3 12.0 673.3 3
** DH = Days to heading, DM = Days to maturity, PH = Plant height, SL = Spike length, NSPS = Number of Spikeletes per Spike, NKPS = Number of Kernels per spike, GY = Grain yield and TKW = Thousand
kernel weight.
[248]
GGbiplot analysis
The GGbiplot analysis showed that overall performance of the Genotypes across
different environments (locations/years) and cumulative mean performance as per
(Anputhas et al., 2011 and Yan, W. and Kang, M. S. 2003). Genotypes assigned
on treatment 9 (HEILO//MILAN/MUNIA/3/KIRIT ATI/2*TRCH) and 11
(MUU/FRNCLN//FRAN COLIN #1) out yielded in overall mean and more stable
on environments except environment-5. Early maturing Genotype assigned on
treatment 7 (GLADIUS/2*BAVIS) is the most stable genotype in its yield on
overall environments Fig 1.
Fig1 GGbiplot Analysis of the genotypes across different locations and years
Conclusion
The trial was evaluated mainly in middle Awash areas following the river basin
where irrigable land is available for the ease of irrigation water supply using
furrow irrigation method. The variability among genotypes for different traits was
well recognized and stability of promising genotypes in yield also well
discriminated. Based on the study results two candidate varieties for yield
potential and one candidate for different agronomic merits were identified.
[249]
Generally, the outperformed candidate varieties were potential yielders across all
the six environments exception to the fifth environment. The out yielded
genotypes were promising in most of the traits that could be utilized for further
selection as candidate variety for possible release and also for further breeding
purposes. More locations and area-based better package utilization could express
more the potential which requires attention in the future to broaden the study.
Acknowledgments
We sincerely acknowledge EIAR, WARC, National wheat Research Program for
technical and managerial supports, CIMMYT and AGP-II projects for their
Germplasm, Technical and financial supports.
Reference
Anputhas M, Samita S, Abeysiriwardena DS. 2011. Stability and adaptability analysis of rice
cultivars using environment-centered yield in two-way ANOVA model. Communications in
Biometry and Crop Science 6 (2), 80–86.
CSA (Central Statistical Agency). 2018. Report on Area and Production of Major Crops (Private
Peasant Holdings, Meher Season). Agricultural Sample Survey. Central Statistics Agency,
Food and Agricultural Organization of the United Nations (FAO). 2017. Crop Prospects and Food
Situation. Rome, Italy.
Muhammad Afzal Arain, Mahboob Ali Sial, M. Arif Rajput, and Ameer A. Mirbahar. 2011. Yield
stability in bread wheat genotypes. Pak. J. Bot., 43(4): 2071-2074.
Leta Tulu and Addishiwot Wondimu. 2016. Adaptability and yield stability of bread wheat
(Triticum aestivum) varieties studied using GGE-biplot analysis in the highland
environments of South-western Ethiopia African Journal of Plant Science Vol. 13(6), pp.
153-162, June 2019 DOI: 10.5897/AJPS2019.1785
Punia SS. Mansoor Shah A and Ram Ranwha B. 2011. Genetic analysis for high temperature
tolerance in bread wheat. African Crop Science Journal, Vol. 19, No. 3, pp. 149 – 163.
USDA (United States Department of Agriculture). 2018. Ethiopia Grain and Feed Annual Report.
Global Agricultural Information Network (GAIN) Report ET-1813. Foreign Agricultural
Service, USDA, Washington, DC. (http://www.fas.usda.gov/data/ethiopia-grain-andfeed-
annual)
VSN International. 2015. Genstat for Windows 18th Edition. VSN International, Hemel
Hempstead, UK. Web page: Genstat.co.uk
Wuletaw Tadesse, Zewdie Bishaw, Solomon Assefa. 2018. "Wheat production and breeding in
SubSaharan Africa: Challenges and opportunities in the face of climate change", International
Journal of Climate Change Strategies and Management, https://doi.org/10.1108/IJCCSM-02-
2018-0015
Yan W and Kang MS. 2003. GGE biplot analysis: A graphical tool for breeders, geneticists, and
agronomists. CRC Press, Boca Raton, FL.
[250]
Spatial META Analysis Methods Improve
Heritability and Genetic Gain of Early Maturing
Sorghum [Sorghum bicolor (L.) Moench]
Genotypes in Dry Lowland Areas of Ethiopia
Amare Seyoum1, Amare Nega1 Taye Tadesse2, Diriba Tadesse1, David Jordan3, Yilma
Kebede, Adane Gebereyhones1, Kedanemaryam Wagaw1, Hailemariam Solomon1,
Alemayheu Assefa2, Tamirat Bejiga1, Daniel Nadew1, Tokuma Legesse1, Alemu
Tirfessa1, Emma Mace3, and Sewmehone Siraw belay1.
1
Ethiopian Institute of Agricultural Research (EIAR), Melkassa Agriculture Research Center, P. O. Box 436,
Adama, Ethiopia, 2 Ethiopian Institute of Agricultural Research (EIAR), Head Quarter, Addis Ababa,
Ethiopia; 3Queensland Alliance for Agriculture and Food Innovation, Hermitage Research Facility 604
Yangan Rd Warwick QLD 4370, Australia
Abstract
In Ethiopia, drought is usually occurring due to delay in onset, dry spell after
sowing, drought during critical crop stage (flowering and grain filling stage) and too
early cessation of rainfall. These situations can be overcome by developing improved
sorghum varieties that are resistant to drought. Developments of sorghum varieties
resistant to drought and producing better grain yield while addressing the plant
biomass requirement is one of the strategies in the sorghum breeding program to the
dry lowland environment. A total of 90 early maturing sorghum genotypes were
evaluated along with two standard check varieties to estimate the grain yield, plant
height, days to flowering, days to maturity and overall agronomic aspects and
stability of performance across the test environments. The trial was conducted using
RCBD in row and column arrangement. The linear mixed model has been used to
predict and identify stable and superior varieties cross the test environment.
Correlations of the trials range from positive +1 to -1 where positive correlation is
an indication of similarity among the testing environments while negative
correlation is an indication of non-similarity among testing environments. Moreover,
using the biplot it was observed that the stability and correlation among the testing
site where the angle between the two lines measures the strength of the correlation.
Improvement in heritability has been obtained due to spatial variation using
advanced statistical analysis methods without any additional cost. Three genotypes
exhibited better yield advantage, higher plant biomass and overall plant aspect
including drought tolerance. In addition, these genotypes were preferred by farmers
in their overall agronomic desirability (drought tolerance, earliness, head exertion
and compactness, grain size and shape and threshability. Accordingly, , the three
candidate varieties were promoted to the variety verification trial l in 2018/19, and
finally variety 14MWLSDT7114 (2005MI5060/E-36-1) is released for commercial
production in dry lowland environment.
Keywords: Genotype, heritability, mixed model, Spatial analysis, GEI, correlated

environment
[251]
Introduction
Sorghum is the second most widely cultivated cereal in sub-Saharan Africa
following maize (FAO, 2012). Over 23 million hectares of land in the continent
are allocated for sorghum production annually with the total annual grain volume
of about 26 million tones. It is believed to have originated in Ethiopia as
evidenced by the early history of the domestication of the sorghum crop there
(Mann et al., 1983; Ayana and Bekele, 1999; Tesso et al., 2008; Adugna and
Bekele, 2012). Ethiopia is the third-largest producer of sorghum in Africa after
Nigeria and Sudan. It is also the most important cereal crop worldwide used for
food, feed, production of alcoholic beverages, and biofuels. In addition, the crop is
primarily grown as a food grain crop in Ethiopia and preferred next to tef for its
injera (leaven bread) making quality. Furthermore, the biomass produced from
sorghum is equally important for sorghum growing farmers in Ethiopia to address
the feed demand (Seyoum et al, 2019).
Regarding the production, area coverage and productivity of sorghum in Ethiopia,

it is the third most important cereal crop in area coverage which shares 18% of the
area covered by cereals and 14.6% of the area covered by grain crops. The total
production of sorghum is nearly 5.1 million tons produced from 1.9 million
hectares of land with the national average productivity of 2.71 tons per hectare
(CSA, 2018). Sorghum is the dominant crop particularly in the dry lowland agro-
ecology and staple food crop for one-third of the Ethiopian population. The
overall production and productivity of sorghum have shown an increasing trend
over the past decades due to an increase in productivity per unit area and to a
lesser extent an increase in area coverage. A meager improvement in productivity
of sorghum has the potential to transform rural livelihoods and also boost the
national economy as the area coverage of sorghum is huge. Most of the sorghum
acreages in Africa including Ethiopia are located in areas that are prone to high
temperatures and frequent drought stress. Drought stress caused by low and erratic
rainfall and exasperated by high temperatures common in most sorghum growing
regions of the world and it is the most important abiotic factor limiting sorghum
productivity (Boyer, 1982; Bohnet and Jensen, 1996).
In Ethiopia, drought is usually manifested by delay in onset, dry spell after

sowing, drought during critical crop stage (flowering and grain filling stage), and
too early cessation of rain. These situations can be overcome by developing
improved sorghum varieties that are resistant to drought. The stage of plant
development in which moisture stress is most severe determines the associated
yield reduction in the crop. Yield reduction is most severe when the stress occurs
when the plant sink capacity is being set, and pollination is disrupted and embryos
are aborted (Westgate and Boyer, 1985). Since the inception of sorghum
research in Ethiopia, concerted efforts have been made to realize a strong research
[252]
program that could be able to develop a variety with high, widely adapted, high
yielding, early maturing and striga resistance varieties with multiple resistance
traits to address major biotic and abiotic factors.
However, until recently the breeding program relied on exotic germplasms which
had a high harvest index compared to local landraces that led to a low adoption
rate because of several factors such as poor grain market, farmers' interest in
multi-purpose and high biomass cultivars. Hence, the notion of client-oriented
breeding to increase the adoption of improved technologies and enhancing genetic
gain through breeding are timely agenda for sorghum breeders (Mindaye et al., in
press). Taking these into account, modification of the breeding program has been
undergone to increase efficiency and genetic gain and bring a sustainable impact
in the research and development endeavors.
Furthermore, development and deployment of high yielding, early maturing,

drought-tolerant, and striga resistance varieties with improved nutrition have been
the major strategies of the national sorghum breeding program in addressing the
varietal demand of end-users. The redesigned breeding pipelines will produce
varieties that are more acceptable for farmers due to higher grain yield, good grain
quality, and acceptable biomass production while providing much greater stability
of performance than currently grown land-races and improved varieties.
Exploitation of genetic diversity for economic benefit is the most important

strategy in plant breeding, and this has to be inferred by field performance
expression of the phenotypes. The consequences of the phenotypic variation
depend largely on the environment. This variation is further complicated by the
fact that not all genotypes react similar ways to the change in the environment.
Genotype by environment interactions (GEI) happens when two or more
genotypes perform differently in more than two environments. This results in
differential responses of genotypes across the testing environment which is
considered as a hindrance in identifying, selecting and recommending crop
varieties (Mindaye et al., 2016). On the other hand, the use of appropriate design
and analysis models could be very vital either to identify high performing
genotypes for target environments or stable genotypes across a given set of
environments. Breeding for drought tolerance requires careful selection of the
target environment; the choice of selection environment is important to achieving
high genetic gain from selection (Cooper et al., 2006).
An important aim of the analysis of agricultural field trials is to obtain good

predictions for genotypic performance, by correcting for spatial effects. Different
statistical methods have been developed for the analysis of multi-environment trial
(MET) data in crop improvement programs (for example; Patterson et al., 1977;
Lin et al., 1986; Becker and Leon, 1988; Guach, 1992, Smith et al.2005). The aim
[253]
of any crop improvement is most often to select either high performing genotypes
for target environments or stable genotypes across a given set of environments.
MET is usually analyzed using a two-stage approach, in which variety means are
first estimated separately for each trial and then combined to form the data for an
overall analysis. The latter methods include mixed effect models (Talbot, 1984;
Smith et al., 2005) and the fixed effects AMMI (additive main effects and
multiplicative interactions) model (Welham et al., 2010). The two-stage approach
is an approximation of the combined analysis of the raw plot data from all trials. If
there is error variance heterogeneity between trials and spatial variation or unequal
replication within trials, the approximation may be poor in estimation by classical
ANOVA. Smith and Cullis (2001) presented a weighted mixed model for the
second-stage analysis that aimed at accommodating these sources of error
variation, thereby reducing efficiency losses. The superior approach, however, is
the spatial MET analysis of Kelly et al. (2007), in which individual plot data are
analyzed and a separate spatial covariance structure and error variance allowed for
each trial. Therefore, the objective of this study was to quantify the magnitude of
genotype by environment interaction and identify high yielding, early maturing,
high biomass and stable sorghum variety for commercial release using spatial and
MET analysis

Description of Experimental Site
The experiment was conducted at six locations which represent the dry lowland
sorghum growing agroecology during 2014, 2015, and 2016 main cropping
seasons. These locations are found in different parts of the country such as Kobo
in North Wello, Mieso in West Hararghe, Erer in East Hararghe, Sheraro and
Humera in Western Tigray and Mehoni in southern Tigray. The detailed
description of each testing location is presented in Table 1 including longitude and
latitude coordinates, altitude (m.a.sl), average annual rainfall amount (mm), soil
type and average maximum and minimum temperatures (oc).
Table 1: Description of testing locations
Sites Longitude Latitude Altitude Soil type Rainfall Minimum Maximum

(m.a.s.l) (mm) T° T°
Mieso 39°21'E 8°30'N 1470 Vertisol 571 16 31

Sheraro 39°9'E 14°6'N 1179 Vertisol 615 20.4 34
Kobo 39°38'E 12°09'N 1513 Vertisol 678 14.8 32
Mehoni 39°70'E 12°70'N 1578 Clay 539 12.81 23.24
(Fachagama)
Humera 4009'E 9016'N 750 Vertisol & flui soil 590 26.7 40.8
Erer 42015’E 9010’N 1297 Vertisol 778 17 37
Source: National Metrology data of 2016/17 cropping season, m.a.s.l=meters above sea level, T°=Temperature.
[254]
Descriptions of genotypes evaluated in this variety trial
Sorghum genotypes derived from pedigree selection were used in this study to
evaluate their performance across locations and over the years. These genotypes
were developed by Melkasa Agricultural research Center via the Pedigree
breeding method and they consisted of ninety (90) lines including two standard
checks Melkam and Dekeba (lists of genotypes and their pedigree description not
shown here for the sake of space). The genotypes involved in this variety trial
were developed by crossing and advanced through pedigree selection up to F5 and
F6 generations based on their response to yield component traits and plant height.
Multi-environment evaluation trials were conducted as PYT and NVT from 2014
to 2016 targeting the dry lowland environments. Based on grain yield
performance, plant height, and flowering time three candidate varieties were
proposed to be verified and later release for growers.
Table 2: List of key pedigree lines studied in Mini environment trial
Genotype Pedigree Background

14MWLSDT7362 2005MI5060/B-35 line developed at Melkassa
14MWLSDT7114 2005MI5060/E-36-1 lines developed at Melkassa
14MWLSDT7115 ICSR24010/B_35 lines developed at Melkassa
14MWLSDT7129 ICSR24010/E-36-1 lines developed at Melkassa
14MWLSDT7060 Macia/76T1#23 lines developed at Melkassa
14MWLSDT7234 Macia/E-36-1 lines developed at Melkassa
14MWLSDT7098 MR812/76T1#23 lines developed at Melkassa
14MWLSDT7425 MR812/B-35 lines developed at Melkassa
14MWLSDT7073 SDSL2690-2/76T1#23 lines developed at Melkassa
14MWLSDT7308 Teshale/B-35 lines developed at Melkassa
14MWLSDT7026 WSV387/76T1#23 lines developed at Melkassa
14MWLSDT7138 WSV387/E-36-1 lines developed at Melkassa
Dekeba ICSR24004 Popular variety check
Melkam WSV387 Popular variety check
Experimental design and field management

A randomized Complete Block Design (RCBD) with two replications in a row-
column arrangement was used l to account for the bidirectional variability
(trends), and to improve the estimation of heritability and genetic values. Each
plot area contained two rows of 5 m length separated by 0.75 m. Sowing date
varies with location but it was conducted between the last week of June to first
week of July when a sufficient amount of rainfall was available. Planting was
conducted manually by drilling along the furrow and the population was adjusted
by thinning considering 0.15 m spacing between plants. DAP fertilizer was
applied at the time of planting with the rate of 100 kg/ha and Urea was side-
dressed when the plant reached at knee height with a rate of 50 kg/ha basis. Data
were collected on the traits of days to 50% flowering, plant height (cm), grain
yield per plot (kg/plot), days to 90% physiological maturity, overall plant aspect
and the data were subjected to statistical analysis using linear mixed model (LMD)
to identify stable and superior varieties compared with the standard check variety.
[255]
Statistical analysis
Mixed-effect models have been developed over the past three decades, first with
applications to animal breeding (Henderson, 1984) and then to other disciplines.
The mixed-effects model contains experimental factors of both fixed and random-
effect types, with appropriately different interpretations and analyses for the two
types. The estimation of variance components in the mixed model assumes
Gaussian random terms by restricted maximum likelihood (REML); where REML
procedure maximizes the joint likelihood of all error contrasts rather than of all
contrasts as in ordinary maximum likelihood (Patterson and Thompson, 1971). In
the original description of REML, it suggests that the score equation for the
variance components may be solved iteratively using the Fisher scoring (FS)
algorithm. For many applications, this strategy presents computational difficulties
due to the large size of the matrices to be inverted and multiplied. Thompson
(1977) presented an overview of the methodology with particular reference to
animal breeding applications and showed how some of the computational burdens
of the FS algorithm may be overcome.
Spatial analysis of individual trials

Adjustment for spatial trends in plant breeding field trials is essential for efficient
evaluation and selection of genotypes. Current mixed model methods of spatial
analysis are based on a multi-step modeling process where global and local trends
are fitted after trying several candidate spatial models (Gilmour et al., 1997).
Before conducting a one-stage model fitting in the META analysis, different
models have been fitted to individual environment spatial plot errors to identify
the best model to be fitted in the final META analysis model.
Spatial mixed model for MET trials

The experiment was laid out in a rectangular array of jth trials j= 1. . . p, consists
of Nj, plots with rj, rows and cj columns (Nj = rj x cj) (Smith 2001). The vector of
data is ordered correspondingly as rows with in column. The model for
the combined vector of data across the environment (trials) = {yj), n =
is given by
y = X + Zu+ e --------------------------------------------------------------(1)
where and are vectors of fixed and random effects, respectively.

and are the associated design matrices with the former assumed to be
of full column ranks. The vector of residuals is given by e. Therefore, the
distribution of the vector of data y is Gaussian with mean X and variance matrix
H= .
[256]
Error e term also consists of a vector of sub error {ej}, where is vector of
th
plot errors for j trial and decomposed into a spatially dependent process and an
independent white noise process . The error variance matrix for trial j is given
by = , where is a spatial correlation matrices that is a
function of with associated variance . The parameter is variance of the
white noise process. The assumption is that spatial process for is second-order
stationary so that the correlation between plot depends only on the distance
between them. It also further assumed that the two-dimensional process is
separable so one can write , where and are the correlation
matrices for columns and rows respectively. However, many researchers
(Gilmour, Cullis, and Verbyla, 1997) show that separable autoregressive process
of order one which is denoted by AR1xAR1 most often provides an adequate
variance structure for local spatial trend. In addition, errors from different trials
are assumed to be independent.
The random effect u consists of sub vectors { } , where is the vector of

effect for the ith random term, i=1. . .q. the matrix Z is partitioned conformably as
[Z1 . . .Zq]. it assumed that the sub-vector of u are mutually independent.
Variance matrix Gi for the ith random term has many possible forms including the
standard variance component structure Gi = .
Let ug be the mpx1 vector of genetic effect for m varieties for each p
environments ordered as varieties within environments. It represents a two-
dimensional (varieties by environment) array of effect, namely , where =
. It is assumed that the associated variance structure has a separable form
with = , where and are the symmetric p x p and m x m
component matrices for environment and varieties, respectively. When = ,
just for simplicity, therefore , and the matrix is
the so-called genetic variance matrix. The diagonal elements are genetic variance
for individual environments and the off-diagonal elements are genetic covariance
between pairs of environments.
The spatial mixed model for the above model 1 of MET data can then be written
as
= , -----------------(2)
the fixed effect includes environmental main effects and trial-specific effects for
extraneous field variation (Gilmour et al., 1997), which is a variety effects at
[257]
each environment with associated design matrix and comprise and
additional random effect with the design matrix , and variance matrix .
In a breeding program, there are many possible forms of genetic variance matrix
structures. In the mixed-model of MET data, the standard structure is given by =
, where and are the variance components for variety main
effects and V x E interactions respectively, whereas is a p x p matric of one.
This implies that all environments have the same genetic variance and all pair of
environment has the same genetic covariance. Due to inefficient estimation or
unstable even for moderately large values of p. Smith (et. al, 2001) proposes an
alternative variance structure model which is called factor analysis that is
analogous of AMMI of Gauch, (1988, 1992).
This model captures the nature of heterogeneous variances and covariance that
occur in most MET data. The factor analytic (FA) model is a regression-type
model (y=ax+b), which can be fitted for an increasing number of dimensions k.
Therefore, the model with variety main effects and k-factor analytic model for
genotypes by environment interaction is a special case of (k+1) factor analytic
model for variety effects at each environment. The concurrence of genotypes and
populations between testing sites was used to allow the trial series to be analyzed
as a single MET as of each trial consisting of similar hybrids, which is the current
best practice method for analyzing field trials for plant breeding programs.

If trials are correlated then information from multiple trials can be combined to
improve the accuracy of the prediction of genetic merit of individuals in a
particular trial. Hence, a total of 90 early maturing sorghum genotypes were
evaluated along with two standard check varieties to estimate the grain yield, plant
height, days to 50% flowering, days to maturity, and overall agronomic aspects
and their stability of performance across eight environments. Spatial Linear mixed
model has been fitted to the data analysis to each of the individual trials and
estimates of the overall trial means, the variance parameters for genetic effect, and
the residuals showed that Kobo is the highest yielder for all genotypes and the
genetic variance is also highest among all environments. On the contrary, Erer is
the lowest yielder and had the lost genetic variance and error variance when
compared with all other environments in terms of yield trait.
In this study, three genotypes exhibited better yield advantage, higher plant
biomass, and overall plant aspects including drought-tolerant. Hence, variety
14MWLSDT7114 (2005MI5060/E-36-1) had 21.2%, variety 14MWLSDT7196
(WSV387/76T1#23) had 17.7% and 14MWLSDT7329 (SDSL2690-2/76T1#23)
[258]
variety had 27.2% superiority in plant height. Mean grain yield performance of
genotype 14MWLSDT7114 displayed 12.2%, genotype 14MWLSDT7196 and
genotype 14MWLSDT7329 varieties had also 13.7 and 20.2% grain yield
advantage compared to the standard check variety Dekeba and Melkam (Table 1).
In addition to the grain yield advantage in comparison to the recently nationally
released standard check variety these three genotypes were also showed stable
grain yield performance across years and environments (Figure 2). Moreover,
these genotypes were also highly preferred by farmers by their overall agronomic
desirability (drought tolerance, earliness, head exertion and compactness, grain
size and shape, threshability. The national variety releasing committee has been
evaluated the variety verification trial both on station and farmer field condition in
2018/19 and then the decision has been made by standing committee for the
release of a variety with a pedigree of 14MWLSDT7114 (2005MI5060/E-36-1)
for commercial production.
The analysis result displayed in Figure 1 indicates that the correlation between
environments how early maturing sorghum lines are correlated over the
environments. Correlation of the trials’ ranges from positive +1 to -1 where
positive correlation is an indication of similarity among the testing environments
while negative correlation is an indication of non-similarity among testing
environments. This implies the selection of the materials. When the environments
have a positive correlation, then selection best materials based on a given
environment is similar to selecting the best materials in the other environment; for
example, KB14PYTLSL with MS15SG2N02, MS16SG2N02, SH16SG2N02 has a
high correlation. Whereas, when environments are negatively correlated there is a
rank change among the genotype so that selecting the best material based on a
given environment can be less performing materials in the other environment like
SH16SG2N02, HM15SG2N02, BB15SG2N02.
[259]
Figure 1. Genetic correlation among early maturing sorghum breeding environments.
One can briefly, observe the stability and correlation among the testing site using
Figure 2 where the angle between the two lines measures the strength of the
correlation. Less angle between the two lines indicates the existence of
correlation most often if less than 90’ but if the angle is just around 90’ it indicates
the independency among the environments. Furthermore, when the angle between
the two lines is more than 90’, this indicates the negative correlation between the
two environments. The distance of the line from the center measures the stability
of the line. Less distance from the center indicates stability while the far distance
from the center indicates non-stability of the environment. Hence, from Figure 3 it
is possible to derive that the angle between the trials (environments)
MS14PYTLSL, MS15SG2N02, SH15SG2N02, MH16SG2N02, MS16SG2N02,
SH16SG2N02, ER15SG2N02 is acute (less than 90o) suggesting that these
environments are correlated and their genotypic response is similar. Whereas
KB14PYTLSL and HM15SG2N02 have an obtuse angle suggesting that this
environment is less correlated environments and the genotype discriminating
ability is different.
[260]
Figure 2. Correlation among testing environment using the angle between two linesThe comparison biplot
presented in Figure 4 can also be used to identify three superior and stable genotypes across the
testing environments.
Comparison biplot (Total - 75.71%)
3
SH15
39
18
5326 24 43 49
PC2 - 22.28%
2938 27
31
4820 51 4 36 MI16 14
5
34
52 712
2 KB14
10 54 6 4247
44
MI15
324641
28 582230 40 21 55 15
25 SH16
MI14
11 16
23
8 56
50 1
195735 13 9
17 33 45
37
PC1 - 53.44%
Genotype scores
Environment scores
AEC
Figure 3. Comparison of genotype grain yield performance
[261]
Increasing heritability and genetic gain is the objective of a crop breeding
program; where high heritability is good for selection. Experimental field
variation decreases heritability, and as a result grain yield is poorly estimated.
Blocking is typically used to remove spatial variation but this approach is limited
(a trend not captured by blocks). Spatial analysis makes use of the correlation
between plots located close to each other. Adjustment for variance due to spatial
trends increases heritability by reducing error and potentially allows reduced
replication which in turn permits increased population size and increased selection
intensity. In the recent study, significant improvement in heritability and or
genetic gain obtained. Figure 4 indicates improvements in heritability by these
advanced statistical methods without any additional cost.
Figure 4. Improvements in heritability through applications of different statistical models.
Conclusion
Many breeding programs evaluate their materials at several locations to exploit
GxE. These trials are often correlated in the way they rank varieties (both positive and
negative). MET analysis makes use of data from other sites where a genotype is grown to
increase the accuracy of estimates of genotype performance at particular sites (ie
increases heritability). MET analysis can be useful for determining the most efficient
distribution of trial sites to achieve good coverage of the target population of
environments. Increasing heritability and genetic gain is the objective of a crop breeding
program; where high heritability is a good parameter for selection. Field variation
decreases heritability. Blocking is typically used to remove spatial variation but this
approach is limited (trend not captured by blocks). Spatial analysis makes use of the
correlation between plots located close to each other. In a recent study, significant
[262]
improvement was observed in improving heritability and genetic gain. This shows that
mixed model statistical analysis has had a major impact on the modern plant breeding
program. In this multi-environment and Spatial Analysis of Early Maturing Sorghum field
trials a total of 90 early maturing sorghum genotypes were evaluated along with two
standard check varieties to estimate the grain yield for grain yield, plant height, days to
flowering, days to maturity and overall agronomic aspects and stability of performance
across the test environments. Linear mixed model has been used to predict and identify
stable and superior varieties across the test environment presented from Table 3-5 and
Figure 1-4. Three genotypes exhibited better yield advantage, higher plant biomass, and
overall plant aspect including drought-tolerant. Hence, variety 14MWLSDT7114
(2005MI5060/E-36-1) had 21.2%, variety 14MWLSDT7196 (WSV387/76T1#23) had
17.7% and 14MWLSDT7329 (SDSL2690-2/76T1#23) variety had 27.2% superiority in
plant height. Mean grain yield performance of genotype 14MWLSDT7114 displayed
12.2%, genotype 14MWLSDT7196, and genotype 14MWLSDT7329 had also showed
13.7 and 20.2% grain yield advantage compared to the standard check variety Dekeba and
Melkam.
In this study Figure 1 indicates how early maturing sorghum lines are correlated over the
environments. Correlation of the trial ranges from positive +1 to -1 where positive
correlation is an indication of similarity among the testing environments while negative
correlation is an indication of -non-similarity among testing environments. One can
briefly, observe the stability and correlation among the testing site using Figure 2 where
the angle between the two lines measures the strength of the correlation. Figure 4
indicates improvements in heritability by these advanced statistical methods without any
additional cost. Form this study three genotypes were showed stable grain yield
performance across years and environments. In addition, these genotypes were preferred
by farmers in their overall agronomic desirability (drought tolerance, earliness, head
exertion and compactness, grain size, shape, and threshability. The three candidate
genotypes were promoted to the variety verification trial in 2018/19, and finally the
national variety release standing committee decided the release of 14MWLSDT7114
(2005MI5060/E-36-1) variety for commercial production
Reference
Adugna A and Bekele E. 2012 Geographical distribution and phenotypic diversity of wild/weedy
sorghum [Sorghum bicolor (L.) Moench] in Ethiopia: implications for germplasm
conservation and crop–wild gene flow. Plant Genetic Resources Available on CJO
doi:10.1017/S1479262112000391.
Amare S, Gebreyohannes A, Nega A, Nida H, Tadesse T, Tirfessa A and Bejiga T. 2019.
Performance Evaluation of Sorghum (Sorghum bicolor (L.) Moench) Genotypes for Grain
Yield and Yield Related Traits in Drought Prone Areas of Ethiopia. Adv Crop Sci Tech 7:
439. doi:10.4172/2329-8863.1000439
Ayana A and Eshetu B. 1999. Multivariate analysis of morphological variation in sorghum
(Sorghum bicolor (L.) Moench) germplasm from Ethiopia and Eritrea. Genet. Resour. Crop
Evol. 46: 387–384
Becker H and Leon J. 1988. Stability analysis in plant breeding. Plant Breed. 101, 1- 23
Bohnet HJ and Jensen RG. 1996. Strategies for engineering water-stress tolerance in plants.
TIBTECH 14:89–97.
[263]
Boyer JS. 1982. Plant productivity and environment. Science 218:443–448
Cooper M, FS van Eeuwijk, Chapman SC, Oodlich DW and Loffler. 2006. Genotype by
environment interactions under water-limited conditions. In: Drought adaptation in cereals
(j. M. Ribaut ed.). The Waworth Press Inc, Binghampton N.Y. pp. 51-9
CSA. The Federal Democratic Republic of Ethiopia Central Statistical Agency (CSA) Agricultural
Sample Survey Report on Area and Production of Major Crops. I, (2018).
Gauch HG. 1988. Model selection and validation for yield trials with interaction. Biornetrics 44,
705-715.
Gauch HG, Jr. 1992. Statistical Analysis of Regional Yield Trials: AMMI Analysis of Factorial
Designs. Amster- dam: Elsevier.
Gauch HG and Zobel RW. 1988. Predictive and postdictive success of statistical analyses of yield
trials. Ther. Appl. Genet. 76: 1-10.
Gilmour AR, Cullis BR, and Verbyla A P. 1997. Accounting for natural and extraneous variation
in the analysis of field experiments. Journal of Agricultural, Bi- ological, and Environmental
Statistics 2, 269-273.
Gleeson AC and Cullis BR. 1987. Residual maximum likelihood (REML) estimation of a
neighbour model for field experiments. Biometrics 43, 277-288.
Henderson CR. 1984. Applications of linear models in animal breeding University of Guelph,
Guelph, ON.
Lin C, Binns M and Lefkovitch L. 1986. Stability analysis: where do we stand? Crop Sc. 26, 894–
900.
Liu C, Rubin DB and Wu YN. 1998. Parameter expansion to accelerate EM: The PX-EM
algorithm. Bio- metrika 85, 755-770.
Mann, JA, CT Kimber and Miller FR. 1983. The origin and early cultivation of sorghums in
Africa. Texas A & M University, Texas Agric. Exp. Stn. Bull. No. 1454.
Mindaye TT, Tesso T and Ejeta G. 2008. Combining ability of introduced sorghum parental lines
for major morpho-agronomic traits. J. SAT Agric. Res. 6 6, 2–7 (2008
Mindaye TT, Amare S, Tamirat B, Kidanemariam W, Adane G/Yohannes, Habte N, Tokuma L,
Sewmehon S and Alemu T. Major Achievements, Challenges and Future Prospects of
National Sorghum and Millet. in press
Patterson HD and Nabugoomu F. 1992. REML and the analysis of series of crop variety trials. In
Proceed- ings from the 25th International Biometric Conference, Hamilton, New Zealand,
77-93.
Patterson HD and Thompson R. 1971. Recovery of in- terblock information when block sizes are
unequal. Bio- metrika 31, 100-109.
Patterson HD, Silvey V, Talbot M and Weatherup STC. 1977. Variability of yields of cereal
varieties in U.K. trials. Journal of Agricultural Science, Cambridg
Smith AB and Cullis BR. 2001. The analysis of crop variety evaluation data in Australia.
Australian and New Zealand Journal of Statistics 43, 129-145.
Smith AB, Cullis BR, and Thompson R. 2005. The analysis of crop cultivar breeding and
evaluation trials: an overview of current mixed model approaches. J. Agric. Sci. 143, 449–
462
Talbot M. 1984. Yield variability of crop varieties in the U.K. Journal of Agricultural Science,
Cambridge 102, 315-321
Tesso Tefaye, Issoufou Kapran, Cécile Grenier, Allison Snow, Patricia Sweeney, Jeff Pedersen,
David Marx, Gurling Bothma, and Gebisa Ejeta. 2008. The Potential for Crop-to-Wild Gene
Flow in Sorghum in Ethiopia and Niger: A Geographic Survey. Crop Sci. 48:1425–1431.
Wilkinson GN, Eckert SR, Hancock TW, and Mayo 0. 1983. Nearest neighbour (NN) analysis of
field experiments (with discussion). Journal of the Royal Statis- tical Society, Series B 45,
151-211.
Westgate ME and Boyer JS. 1985. Carbohydrate reserves and reproductive development at low
leaf water potentials in maize. Crop Sci. 25:762-769
[264]
Agronomy and Physiology
Effect of Seed Rate and Row Spacing on Yield and
Yield Components of Rainfed Lowland Rice
(Oryza sativa L.) variety
Zelalem Tadesse, Tilahun Tadesse and HabtamuAssega
Ethiopian Institute of Agricultural Research (EIAR), Fogera National Rice Research and Training Center
(FNRRiTC), P.O.Box 1937 Fogera, Ethiopia; E-mail-zelalemtadesse46@gmail.com
Abstract
Rice (Oryza sativa L.), which is one of the most important food crops for more than
half of the global population. However, its productivity is very low because of poor
agronomic practices like inappropriate seeding rate, improper adjustment of row
spacing, and other factors. A field experiment was conducted at Fogera Plain during
the rainy seasons of 2017 and 2018 to determine the effects of seed rate and row
spacing on yield and yield component of rainfed lowland rice ecosystem. Four-row
spacing’s (15, 20, 25and 30 cm) and Seven seed rates (40, 60, 80, 100, 120,140
and160 kg/ha) were combined factorially and laid in Randomized Complete Block
Design with three replications. Data were collected on plant height, panicle length,
number of total and effective tillers per row meter length, grain yield, straw yield,
and harvest index. All collected data were subjected to analysis of variance.
Economic analysis was also done for grain yield. The results of the experiment
showed that the main effect of seed rate was significantly affecting, panicle length,
number of total and effective tillers per row meter length, grain yield, straw yield,
and harvest index whereas row spacing was significantly affected grain yield and
number of total and effective tillers per row meter length. The interaction effect of
seed rate and row spacing was significantly (p<0.01) affected total tillers, number of
effective tillers, and grain yield. The highest grain yield (3.25 t ha -1) was obtained at
a seed rate of 120 kg ha-1 and 30 cm row spacing. The economic analysis indicated
that a seed rate of 100 kg ha-1 and a row spacing of 20 cm is the most profitable
treatment with a mean net benefit of 50982.5.Birr ha -1. Therefore it can be concluded
that a seed rate of 100 kg/ha and row spacing of 20 cm is preferable and
recommended for rainfed lowland rice production ecosystem in Fogera plain.
Keywords: Low land rice, grain yield; row spacing; seeding rate.
Introduction
Rice (Oryza sativa L.), is one of the most important food crops and is considered
as a major source of calories for more than half of the global population (Carrijoet
al., 2017), covers 11% of total arable land (Khush, 2005). Rice is grown in
different ecosystems, mainly classified into three types according to the
environmental water conditions: irrigated, rain-fed lowland, and upland
[265]
ecosystems, which account for approximately 75%, 19%, and 4% of the global
rice production, respectively (IRRI, 2013). Rice has become a commodity of
strategic significance across many African countries (Hegde and Hegde, 2013). It
is also the most rapidly growing food sources across the continent due to the great
urbanization in Africa more than any other region in the world. In 2008, Africa
produced an estimated quantity of 23 million Mt of rice on 9.5 million hectares.
The western, northern, and eastern regions of Africa had the largest shares, with
10.2 million Mt, 7.3 million Mt, and 5 million Mt, respectively. These quantities
of rice were harvested on 5.8 million hectares, 0.8 million hectares and 2.4 million
hectares in West, North, and East Africa, respectively.
Rice consumption is increasing faster than that of any other food staple in Africa
at about 5.5% per year (2000–2010 average). This increase is driven by
urbanization and related changes in eating habits, and population growth (Seck et
al., 2012). Rice consumption was approximately 24 million tons per year in SSA
in 2012. With only about 60% of rice consumption being satisfied by domestic
production, rice imports stand at 10–12 Mt. This is equivalent to one-third of the
rice traded on the world market. Demand for milled rice in SSA is expected to
increase by 30 Mt by 2035, equivalent to an increase of 130% in rice consumption
(Seck et al., 2012).
Africa has sufficient land and water resource to produce enough rice to feed its
population and, in the long term, generate export revenues. Rice cultivars, rice-
based cropping systems and the rice itself will, however, have to undergo
adaptations and improvements to meet future demands for both food security of
the growing population and environmental conservation (Asch and Brueck, 2010).
Rice productivity in Africa is generally low about 1 t ha-1 in uplands, 1 to 2 t ha-1
in rainfed lowlands and 3 to 4 t ha-1 in the irrigated zones, and a range of factors
explain this low productivity (African Rice, 2010).
Ethiopia has a huge potential in both rain-fed and irrigated areas for rice
production, which is estimated about thirty million ha (MoARD, 2010; CSA,
2012) The cultivation of rice in Ethiopia is of more recent history than its
utilization as a food crop. Some evidences indicate that the cultivation of rice in
Ethiopia was first started at the Fogera and Gambella plains in the early 1970s.
Recently, the Fogera, Gambella, Metema, and Pawe plains located in the northern,
northwestern, and western regions Ethiopia have become the major rice-producing
areas in Ethiopia (MulugetaSeyoum, 1999, 2000). According to the National Rice
Research and Development Strategy of Ethiopia, the trend in the number of rice-
producing farmers, the area allocated and production showed a high increase
especially since 2006 (NRRDSE, 2009). Area rose from 6,000 hectares in 2005 to
nearly 222,000 hectares in 2010 and paddy production from 15,460 tons to
887,400 tons, at the same time, the number of rice farmers increased from 18,000
[266]
to more than 565,000 (MoARD, 2010). At the Fogera plain, rice plays an
important role in relaxing the problem of food insecurity in the farming
community. Besides, rice is among the target commodities that have received due
emphasis in the promotion of agricultural production, and it is considered as the
“millennium crop” expected to contribute to ensuring food security in Ethiopia.
Although rice is introduced to the country very recently, it has proven to be a crop
that can assure food security in Ethiopia, the second-most populous nation in SSA
(MoARD, 2010).
Dawit (2015) indicated that the production area of rice was lower than other
cereals (teff, sorghum, maize, and wheat). But the production status was increased
in the production area, yield in tons and producing farmers. Rice is the second
among cereals in terms of average national yield (2.79 t ha-1) next to maize (3.387
t ha-1) (CSA, 2016). The average rice yield ranges from 2.5 to 4.0 t ha-1 on-farmers
field conditions in different agro-ecologies of the country (MoARD, 2011). In
2014/15 and 2015/16 cropping season, rice was produced on 46,823.22 ha and
45454.18 of land in the country with a total production of 1,318,218.53 tons and
1268,064.47 tons, respectively (CSA, 2016).
Optimum seeding rate and proper adjustment of row spacing are the most
important production factors for higher grain yield. Spacing determines the
number of plants per unit area (Yoshida et al., 1981). The plant to plant and row to
row distance determines the plant population per unit area which has a direct
effect on yield. Closer spacing hampers intercultural operations, more competition
arises among the plants for nutrient, air, and light as a result plants become weaker
and thinner and consequently, yield is reduced (Alam et al, 2012). Wider spacing
also allows more competition among crop plants and weeds. As a result plant
growth slows down and their grain yield decreases. On the other hand, Closer
spacing increases competition among plants for nutrients, air, light, which results
in weaker plants, mutual shading thus favors more straw yield than grain yield
(Sultana et al., 2012). Lower and higher plant densities have a positive influence
on the yield of rice (Harris et al, 2015). High seeding rate leads to non-productive
tillers, more severe disease pressure and susceptible to lodging (Garba et al, 2013).
Optimum plant spacing ensures the plant to grow properly with their aerial and
underground parts by utilizing more solar radiation and nutrients (Mohaddesi et
al., 2011). There are several agronomic management constraints with this crop.
Rice is becoming a high potential crop and there is a lack of appropriate
agronomic management recommendations that could help to maximize the
productivity of the cultivation techniques in the study area. Among the rice
production constraints, seed rate and row spacing are an important factor Fogera
Plain. Area-specific recommendation of seed rate is vital to set optimum seed rate
and row spacing for rice production in the study area. Therefore the objective of
[267]
the present study is to determine the effect of row spacing and seed rate on yield
and yield components of rice in Fogera plain.

The experimental site is located at 11049’55 North latitude and 370 37′ 40 East
longitudes. The altitude of the experimental site is 1815 meters above sea level.
The average mean annual minimum and maximum temperature is 12.75 0C and
27.37 0C, respectively. The main rainfall is in July and August (Fig. 1). The
experiment was conducted from 2017 to 2018.
Figure 1.The average rainfall and temperature condition of Fogera plain for the period of 1981-2017.
Treatments, design and management

The treatments consist of a factorial combination of four levels of row spacings
(15, 20, 25and 30 cm) and seven levels of seed rates (40, 60, 80, 100, 120,140
and160 kg ha-1) in randomized complete block design with three replications. The
gross plot size was 3m*4 m. The net plot was made by excluding the left and right
two outer rows and a plot length of 0.5 m from the top and bottom sides of the plot
area. Thus the net plot size for the respective row spacings (15, 20, 25 and 30 cm)
were 3m*2.4m; 3m*2.2m, 3m*2m, and 3m*1.8m. The variety Edget was used for
this experiment. Recommended fertilizer rates of 69/23 kg N/P2O5 ha-1was used in
the experiment (Tilahun et al., 2007).
Data such as plant height (cm), Panicle length (cm), number of total tillers/row
meter length, number of effective tillers/row meter length,, thousand seed weight,
grain yield, straw yield, and harvest index were recorded from the net plot. All
collected data were subjected to analysis of variance (ANOVA) using SAS
[268]
software version 9.2 (SAS-Institute, 2008). Combined analysis over locations and
years was performed. Between treatments, comparisons of means were made
using the Least Significant Difference (LSD) test at 1 and 5% probability levels.
The mean grain and straw yield data was adjusted by 90% and economic analysis
was carried out by following CIMMYT (1988) procedures by taking all variable
costs. The prevailing cost of inputs and outputs in the year 2019 considered for the
analysis. The cost of rice grain and straw was Birr 13.5 and 1.2 /kg, respectively.
Total costs that varied ( seed and planting cost) for each treatment were calculated
and treatments were ranked in order of ascending total variable cost (TVC) and
dominance analysis was used to eliminate those treatments costing more but
producing a lower net benefit than the next lowest cost treatment. The prices of
the inputs that were prevailing at the time of their use were considered for working
out the cost of cultivation. Net returns per hectare were calculated by deducting
the cost of production per hectare from gross income per hectare. A treatment that
is non-dominated and having the highest net benefit is said to be economically
profitable (CIMMYT, 1988).

The analysis of variance exhibited that the seed rate had significantly (P< 0.05)
influenced panicle length. Whereas, row spacing and interaction effect did not
show significant effect on panicle length.(Table 1).The highest panicle length
(18.6) was observed at a seed rate of 40 kg ha-1 and the lowest panicle length
(17.1 and 17.2) recorded from 160 and 140 kg ha-1 seed rate respectively (Table
2). At the lower seed rate of 40 kg ha-1, the panicle length was higher compared to
a higher seed rate of 140 and 160 kg ha-1. This might be due to more free space
between plants at the lower seed rate and less intra-plant competition for available
resources that resulted in higher panicle length. This result is in line with the
finding of Zewdie Bishawet al. (2014) who reported that plant height and panicle
length are negatively interrelated on wheat. Shorter plants produce longer panicle
length and longer plant produce shorter panicle and higher biomass production.
Gafaar (2007) also stated that increasing sowing density from 200 up to 400 grains
m-2 significantly decreased spike length. Similarly, Seleimanet al. (2010) reported
that the longest spikes were obtained from 250 and 300 grains per m2 but without
significant differences between both of them.
The analysis of variance indicated that the main effects of seed rate and row
spacing highly significantly (p<0.01) affected both the number of total and
effective tillers per row length. Moreover, the interaction of seed rate and row
spacing also highly significant (P<0.01) (Table1) The maximum number of total
tillers per row meter length (66.5) and the minimum (56.3) were recorded from the
seed rate of 140 kg ha-1 and 60 kg ha-1, respectively (Table 2). Regarding the row
spacing, more number of total tillers per row meter length (65.5) was recorded
[269]
from (30 cm) row spacing. On the other hand, the smallest number of total tillers
per row meter length (54.8) was observed from 15 cm row spacing. Concerning
the interaction effect, the maximum number of total tillers (75.85) was obtained at
120 kg ha-1 seed rate and a row spacing of 30 cm (Table 3).
The analysis of variance indicated that the main effects of seed rate and row
spacing highly significantly (p<0.01) affected the number of effective tillers per
row meter length (Table 1). However, the interaction effect of seed rate and row
spacing also had a highly significant (p<0.01) effect on the number of effective
tillers per row meter length. The highest number of effective tillers per row meter
length (64.96) was observed at a seed rate of 140 kg ha-1 while the lowest number
of effective tillers per 1m row length (55.28 and 54.95) observed at a seeding rate
of 40 and 60 kg ha-1 respectively and statistically similar. In addition to this, the
highest number of effective tillers (63.79) per row meter length was recorded at a
row spacing of 30 cm while the minimum number of effective tillers (53.73) per
row meter length was recorded at a row spacing of 15 cm(Table 2).
Wider spacing produced higher tillers per hill than closer spacing. The production
of more tillers in widely spaced plants was probably due to the absorption of more
nutrients and moisture and also to the availability of more sunlight in comparison
to dense plants. Similar results were reported by (Haque, 2002). This result was
also in agreement with (Sewunet Ashebir, 2005) who reported the highest
effective tillers were recorded at medium spacing performed better as compared to
lower spacing. (Yamada, 1961) also showed that under dense planting, the growth
of each plant decreases and the size of the plants and productive tillers become
smaller. The more densely the rice plants, the fewer are the number of stems or
tillers and productive tillers per hill but their number increases per unit area.
Analysis of variance showed that the main effect of seed rate and row spacing had
a highly significant effect (P < 0.01) on grain yield. On the other hand seed rate
affected significantly (P < 0.01) straw yield (Table 1). However, the interaction
effect of seeding rate and row spacing showed a highly significant (p < 0.01)
effect on grain yield but not in straw yield (Table 1). The highest grain yield (3.45
t ha-1) was obtained at the seed rate of 100 kg ha-1 and the lowest grain yield (3.1
t/ha-1) was obtained at a seed rate of 160 kg/ha (Table 2). The row spacing of 25
cm produced maximum grain yield (3.31 t/ha-1) followed by 30 and 20 cm (3.2
and 3.1 t/ha1) respectively and the lowest grain yield (3.12 t/ha) was observed at a
row spacing of 15 cm. From the above result, it could be suggested that as the
seeding rate increased there was no significant and proportional yield increment
observed. Likewise, grain yield of rice was significantly affected at the optimum
seed rate but further increase in seed rate did not increase the yield. (Balock et al.,
2002) who reported that the increased plant spacing considerably resulted in
vigorous plant growth and caused a significant increase in the number of panicles
[270]
per hill, grain yield per hill, filled grain per panicle, and thousand-grain weights.
Furthermore, other workers (Mosalem et al., 2000) reported that increasing
seeding rates decreased the number of spikelets/panicle, spike length, number and
weight of grains/spike in wheat. Grain yield of rice increased with the increase in
the number of plants per unit area as long as there is space in the cultivated areas.
When planting density exceeds the optimum level, competition among plants for
light above ground and nutrients below ground becomes severe. Consequently,
plant growth slows and grain yield decreases. Zeng and Shannon (2000) reported
that the reduction in fertility at high density was one of the causes for the
reduction of seed yield per plant with the increase of seeding density.
The main effect of seed rate highly significantly (P < 0.01) affected straw yield whereas
the main effect of row spacing and its interaction effect didn’t show significant (p>0.05)
effect on straw yield (Table 1). The highest straw yield (6.51 and 5.96 t ha-1) was
observed at a seed rate of 100 kg ha-1 and 160 kg ha-1 which is statistically similar. While
the lowest straw yield (4.75 t ha-1) was found with a seed rate of 40 kg ha-1 (Table 2).
There was a linear increase in straw yield as the seeding rate was increased. However, it
was statistically in parity within straw yield obtained in response to the seeding rate of 60,
80, and 140 kg ha-1. The lowest spacing might have influenced vegetative growth in terms
of plant height and number of tillers per hill (effective and non-effective tillers) which
resulted in increased straw yield. Sultana et al. (2012) also reported similar results and the
higher straw yield was obtained in 20 cm row spacing in rice.
The harvest index represents the ratio of grain yield to the total above-ground dry matter
production (Marschener, 1995). The present results showed that the harvest index was
highly significantly (p<0.01) affected by seed rate but not by row spacing and its
interaction effect (Table 1). The highest harvest index (38.68%) was recorded at a seed
rate of 40 kg ha-1 while the lowest harvest index (35.1%) was recorded at the seeding rate
of 100 kg ha-1 respectively, (Table 2). The ability of a cultivar to convert the dry matter
into economic yield is indicated by its harvest index. The higher the harvest index value,
the greater the physiological potential of the crop for converting the dry matter to grain
yield. One of the ways to increase the yield is to increase harvest index. Sink formation
and ripening are the two physiological processes that explain the improvement in HI
(Akita, 1982). Based on the principles of economic analysis CIMMYT (1988), the
minimum acceptable marginal rate of return (MRR %) should be 100%. The economic
analysis was done based on the prevailing prices of variable costs using the Ethiopian
currency (Birr). Grain and straw yields adjustments, calculations of total variable costs
(TVC), gross benefits (GB), and net benefits (NB) were performed (Table 4). Dominance
analysis was performed after arranging the treatments in their order of TVC. Treatments
are considered as dominated if it has higher TVC but lower NB than a previous treatment
with lower TVC and higher NB (Table 5). Non dominated treatments were taken out and
the marginal rate of return (MRR) was computed (Table 6). Highest NB (Birr 50982.5 ha -
1
) with an acceptable level of MRR (913.01) was observed at a row spacing of 20 cm and
100 kg seed rate ha-1 (Table 6). The combined application of a seed rate of 100 kg ha-1 and
20 cm row spacing is the most profitable rate to be recommended for rice production in
Fogera plain.
[271]
Table 1. Analysis of variance (ANOVA) for Panicle length, Total tillers, Number of effective tillers, Grain Yield, Straw Yield and Harvest Index of seed rate and row spacing of
low land rice
Panicle Total Number of No. of fertile Grain

Plant height length (cm) tillers/m Effective tillers/ spikes/panicle Yield Straw
Sources of variation DF (cm) row m row (t/ha) 1000sw/g Yield (t/ha) HI (%)
Location 6 2199.5 NS 207.05 ** 11167.33** 12275.75 ** 249.98 ** 42.32 ** 923.71 ** 150.69** 1221.08 **
RS_cm 3 90 7.92 NS 2757.40** 2434.77 ** 21.42 NS 0.90 * 0.88 NS 3.04 NS 18.49 NS
Rep 2 109.8 NS 48.66 ** 172.84 NS 295.70 NS 37.20 NS 3.04 ** 15.95 * 13.31 ** 34.73 NS
SR_kg 6 383.7 NS 23.86 ** 1130.66 ** 1085.35 ** 16.33 NS 2.57 ** 0.50 NS 24.51 ** 104.1 **
Location*RS_cm 18 27.58 NS 2.05 NS 114.99 NS 124.20 NS 15.55 NS 0.52 * 4.87 NS 2.04 NS 43.40 NS
Location*Rep 12 198.12 NS 22.35 ** 85.56 NS 82.23 NS 20.20 NS 0.49 NS 4.53 NS 3.58 * 29.40 NS
Location*SR_kg 36 47.53 NS 3.01 NS 196.6 * 189.21 * 21.74 NS 0.49 * 9.03 ** 2.55 NS 36.08 NS
Rep*RS_cm 6 10.59 NS 2.06 NS 56.73 NS 27.45 NS 3.24 NS 0.31 NS 2.57 NS 0.67 NS 20.27 NS
RS_cm_*SR_kg 18 48.25 NS 3.39 NS 542.71 ** 519.49 ** 14.37 NS 0.82 ** 5.95 NS 2.56 NS 37.90 NS
Rep*SR_kg 12 5.67 NS 1.47 NS 179.86 NS 171.22 NS 29.09 NS 0.59 * 6.92 NS 2.06 NS 32.00 NS
Location*Rep*RS_cm 36 30.33 NS 2.09 NS 87.05 NS 96.08 NS 20.54 NS 0.36 ns 5.61 NS 2.18 NS 21.62 NS
Locatio*RS_cm_*SR_kg 108 19.69 NS 2.59 NS 129.19 NS 129.13 NS 14.18 NS 0.49 ** 5.37 NS 2.74 ** 33.91 NS
Location*Rep*SR_kg 72 20.31 NS 1.84 NS 83.06 NS 80.32 NS 15.68 NS 0.21ns 4.89NS 1.81 NS 31.13 NS
Rep*RS_cm_*SR_kg 36 37.19 NS` 4.15 NS 90.77 NS 99.06 NS 15.48 NS 0.50 * 4.45 NS 2.25 NS 24.97
Error 216 26.81 2.85 121.94 116.72 374.78 0.30 5.03 1.81 30.38
CV% 6.71 9.55 18.26 18.33 24.31 17.23 6.93 23.59 15.02
PL=Panicle length, TT/RML, Number of Total tillers/row meter length, NET/RML=number of effective tillers /row meter length, GY (ton) grain yield ton /ha, SY (ton) straw yield ton /ha and
HI (%) harvest index.
[272]
Treatments had a significant effect on yield and yield components of rice. Among
the treatments, drilling 100 kg ha-1 seed rate and a row spacing of 25 cm produced
maximum grain yield (3.31 t/ha-1) than other treatments. The partial budget
analysis also revealed that the highest NB (Birr 50982.5 ha-1) with an acceptable
level of MRR (913.01) was observed at a row spacing of 20 cm and 100 kg seed
rate ha-1. Therefore the combined application of a seed rate of 100 kg ha-1 and 20
cm row spacing is the most profitable rate to be recommended for rice production
in Fogera plain and other similar agro-ecologies.
Table 2. Main effects of seed rate and row spacing on yield and yield components of low land rice
Seed rate (kg ha-1) PL (cm) TT/RML NET/RML GY(ton ha-1) SY(ton ha-1 HI (%)
40 18.6195A 56.619D 54.952D 2.89050D 4.7512D 38.6831A
60 18.0133B 56.369D 55.286D 3.24097BC 5.4290C 36.8615AB
80 17.9026BC 58.571CD 56.845CD 3.25961BC 5.6222BC 37.0289AB
100 17.5579BCD 62.262B 60.762B 3.45280A 6.5158A 35.1723B
120 17.3793CD 62.607B 60.798B 3.31124B 5.8944B 35.9450B
140 17.1824D 66.548A 64.964A 3.25198BC 5.8179BC 37.0223AB
160 17.1250D 60.238BC 58.857BC 3.10017C 5.9638A 36.1051B
Row Spacing (cm)
15 17.5376 54.850C 53.735C 3.12488B 5.5423 36.8460
20 17.4366 60.231B 58.476B 3.22968AB 5.8631 36.2114
25 17.8337 61.197B 59.687B 3.31715A 5.8147 37.0476
30 17.9235 65.558A 63.796A 3.18960AB 5.6339 36.6483
PL=Panicle length, TT/RML= Number of Total tillers/row meter length, NET/RML=number of effective tillers /row
meter length, GY= grain yield ton ha-1, SY= straw yield ton ha-1 and HI (%) = harvest index.
[273]
Table 3.Interaction effect of seed rate by row spacing on yield and yield components of lowland rice at Fogera.
TRT Row Seed TTN/RML NNET/ RML GY (tha-1) SY (t ha-1)

space Rate
1 15 40 53.381HIJ 51.952HIJ 2.9043IJKL 4.7247
2 15 60 53.524HIJ 52.571GHIJ 3.0649FGHIJKL 5.3416
3 15 80 52.762J 51.571IJ 3.3315BCDEFGH 6.0545
4 15 100 52.381J 51.238J 3.2075CDEFGHIJKL 5.7321
5 15 120 55.476FGHIJ 54.333GHIJ 3.0654FGHIJKL 5.5386
6 15 140 56.762FGHIJ 56.048GHIJ 3.3482BCDEFG 5.4599
7 15 160 59.667EFGHIJ 58.429DEFGHIJ 2.9523GHIJKL 5.9445
8 20 40 53.619HIJ 51.667IJ 2.8268L 4.6429
9 20 60 53.333IJ 51.095J 3.1543CDEFGHIJKL 5.6952
10 20 80 57.048FGHIJ 55.476GHIJ 3.0582FGHIJKL 5.6579
11 20 100 70.000ABC 69.000ABC 3.7705L 7.1490
12 20 120 57.952EFGHIJ 55.714GHIJ 3.1371DEFGHIJKL 5.9444
13 20 140 70.952ABC 68.762ABC 3.2904CDEFGHIJ 6.1108
14 20 160 58.714EFGHIJ 57.619EFGHIJ 3.3703BCDEF 5.8417
15 25 40 62.000DEF 60.190DEFG 2.9374HIJKL 4.7961
16 25 60 59.000EFGHIJ 57.048FGHIJ 2.8782KL 4.9622
17 25 80 54.000GHIJ 54.333GHIJ 3.5205ABCD 5.7009
18 25 100 59.286EFGHIJ 57.857EFGHIJ 3.6355AB 6.7918
19 25 120 61.143DEFGH 59.333DEFGH 3.5464ABC 6.3497
20 25 140 65.524CDE 63.667CDEF 3.2242CDEFGHIJKL 5.7718
21 25 160 67.429BCD 65.381BCD 3.4780ABCDE 6.3302
22 30 40 65.286CDE 63.571CDEF 2.8935JKL 4.8410
23 30 60 60.619DEFGHI 59.095DEFGHI 3.3033BCDEFGHI 5.7171
24 30 80 61.667DEFG 59.762DEFG 3.1283DEFGHIJKL 5.8584
25 30 100 67.381BCD 64.952BCDE 3.1977CDEFGHIJKL 6.3903
26 30 120 75.857A 73.810A 3.2589BCDEFGHIJK 5.7451
27 30 140 72.952AB 71.381AB 3.1011EFGHIJKL 5.1464
28 30 160 55.143FGHIJ 54.000GHIJ 3.4444ABCDEF 5.7386
CV 18.26 18.33 17.23 -
[274]
Table 4.Economic analysis for grain and straw yield for seed rate by row spacing interaction of low land rice at Fogera.
Treat RS SR GY SY AGY ASY TVC GB NB

ments (t/ha) (t/ha) (t/ha) (t/ha) (Birr/ha) (Birr/ha)
1 15 40 2.9043 4.7247 2613.87 4252.23 2140 40389.921 38249.921
2 15 60 3.0649 5.3416 2758.41 4807.44 2410 43007.463 40597.463
3 15 80 3.3315 6.0545 2998.35 5449.05 2680 47016.585 44336.585
4 15 100 3.2075 5.7321 2886.75 5158.89 2950 45161.793 42211.793
5 15 120 3.0654 5.5386 2758.86 4984.74 3220 43226.298 40006.298
6 15 140 3.3482 5.4599 3013.38 4913.91 3490 46577.322 43087.322
7 15 160 2.9523 5.9445 2657.07 5350.05 3760 42290.505 38530.505
8 20 40 2.8268 4.6429 2544.12 4178.61 1740 39359.952 37619.952
9 20 60 3.1543 5.6952 2838.87 5125.68 2010 44475.561 42465.561
10 20 80 3.0582 5.6579 2752.38 5092.11 2280 43267.662 40987.662
11 20 100 3.7705 7.149 3393.45 6434.1 2550 53532.495 50982.495
12 20 120 3.1371 5.9444 2823.39 5349.96 2820 44535.717 41715.717
13 20 140 3.2904 6.1108 2961.36 5499.72 3090 46578.024 43488.024
14 20 160 3.3703 5.8417 3033.27 5257.53 3360 47258.181 43898.181
15 25 40 2.9374 4.7961 2643.66 4316.49 1540 40869.198 39329.198
16 25 60 2.8782 4.9622 2590.38 4465.98 1810 40329.306 38519.306
17 25 80 3.5205 5.7009 3168.45 5130.81 2080 48931.047 46851.047
18 25 100 3.6355 6.7918 3271.95 6112.62 2350 51506.469 49156.469
19 25 120 3.5464 6.3497 3191.76 5714.73 2620 49946.436 47326.436
20 25 140 3.2242 5.7718 2901.78 5194.62 2890 45407.574 42517.574
21 25 160 3.478 6.3302 3130.2 5697.18 3160 49094.316 45934.316
22 30 40 2.8935 4.841 2604.15 4356.9 1440 40384.305 38944.305
23 30 60 3.3033 5.7171 2972.97 5145.39 1710 46309.563 44599.563
24 30 80 3.1283 5.8584 2815.47 5272.56 1980 44335.917 42355.917
25 30 100 3.1977 6.3903 2877.93 5751.27 2250 45753.579 43503.579
26 30 120 3.2589 5.7451 2933.01 5170.59 2520 45800.343 43280.343
27 30 140 3.1011 5.1464 2790.99 4631.76 2790 43236.477 40446.477
28 30 160 3.4444 5.7386 3099.96 5164.74 3060 48047.148 44987.148
RS= Row spacing (cm); SR= seed rate (kg ha-1) ; TVC= Total variable cost (Birr ha-1) GY, Average grain yield (t ha-1)
AGY= Adjusted grain yield ( ton ha-1); SY=Average straw yield (ton ha-1) ASY= Adjusted straw yield ( ton ha-1); GB=
Gross benefit (Birr ha-1); NB = Net benefit (Birr ha-1)
[275]
Table 5. Dominance analysis for seed rate by row spacing interaction for grain yield of low land rice at Fogera
Treatments RS SR TVC(B ha-1) NB Dominance

22 30 40 1440 38944.3
15 25 40 1540 39329.2
23 30 60 1710 44599.6
8 20 40 1740 37620.0 D
16 25 60 1810 38519.3 D
24 30 80 1980 42355.9 D
9 20 60 2010 42465.6 D
17 25 80 2080 46851.0
1 15 40 2140 38249.9 D
25 30 100 2250 43503.6 D
10 20 80 2280 40987.7 D
18 25 100 2350 49156.5
2 15 60 2410 40597.5 D
26 30 120 2520 43280.3 D
11 20 100 2550 50982.5
19 25 120 2620 47326.4 D
3 15 80 2680 44336.6 D
27 30 140 2790 40446.5 D
12 20 120 2820 41715.7 D
20 25 140 2890 42517.6 D
4 15 100 2950 42211.8 D
28 30 160 3060 44987.1 D
13 20 140 3090 43488.0 D
21 25 160 3160 45934.3 D
5 15 120 3220 40006.3 D
14 20 160 3360 43898.2 D
6 15 140 3490 43087.3 D
7 15 160 3760 38530.5 D
RS= Row spacing (cm) ; SR= seed rate (kg ha-1); TVC= Total variable cost (Birr ha-
1); NB= Net benefit (Birr ha-1)
Table 6. MRR analysis for seed rate by row spacing interaction for grain yield of low land rice at Fogera
Treatments RS SR TVC (Birr ha-1) NB Birr ha-1 MRR (%)

22 30 40 1440 38944.3
15 25 40 1540 39329.2 384.893
23 30 60 1710 44599.6 3100.2147
17 25 80 2080 46851.0 608.50919
18 25 100 2350 49156.5 853.86
11 20 100 2550 50982.5 913.013
RS= Row spacing (cm); SR= Seed rate (kg ha-1); TVC= Total variable cost (Birr ha-1); NB= Net benefit (Birr ha-1);
MRR%= Marginal rate of return in percent
References
Africa Rice Center (Africa rice). 2010. Improving Access to Rice Seed and Building a Rice
Data System for Sub-Saharan Africa (Japan Emergency Project). Final Report of the
Japan Emergency Project, Cotonou, Benin.
Alen, MS, Baki MA, Sultana MS, Ali KJ and Islam MS. 2012. Effect of variety, spacing and
number of seedlings per hill on the yield potentials of transplant Aman rice.International
Journal of Agronomy and Agricultural Research, 2(12): 10-15.
[276]
Asch, F and Brueck H. 2010. Rice Crop Innovations and Natural—Resource Management: A
Glimpse into the Future Presented at 2nd Africa Rice Congress: Innovations and
Partnerships to Realize Africa Rice Potential, March 22-26. Mali: Bamako.
Balock, AW, Soomro AM, Javed MA, Ahimed M, Bughio HR, Bughio MS and Mastoi NN.
2002. Optimum plant density for high yield in rice (Oryza sativa L.).Asian Journal of
Plant Sciences. 1(1): 25-27.
Carrijo, DR, Lundy ME, Linquist BA. 2017. Rice yields and water use under alternate wetting
and drying irrigation: a meta-analysis. Field Crops Res. 203, 173–180.
CIMMYT. 1988. From Agronomic Data to Farmer Recommendations: An Economics Training
Manual. Completely revised edition, Mexico, DF.
CSA (Central Statistical Agency). 2012. Federal democratic republic of Ethiopia. Statistical
Abstracts, Addis Ababa: Ethiopia.
CSA (Central Statistical Agency). 2016. Agricultural sample survey report on Area and
Production of major crops. Central Statistical Agency of Ethiopia, Addis Ababa, Ethiopia,
1: Statistical bulletin; 584.
Dawit Alemu. 2015. Rice in Ethiopia: Progress in Production Increase and Success Factors.6th
CARD General Meeting; Ethiopian Institute of Agricultural Research, pp 1-22.
Gafaar, NA. 2007. Response of some bread wheat varieties grown under different levels of
planting density and nitrogen fertilizer.Minufiya Journal of Agriculture, 32: 165-183.
Haque D. 2002. Effect of Madagascar Technique of Younger Seedling and Wider Spacing on
Growth and Yield of Boro rice.MS Thesis, Dept. Agron., Bangladesh Agril. Univ.,
Mymensingh;
Hegde, S and Hegde V. 2013. Assessment of Global Rice Production and Export Opportunity
for Economic Development in Ethiopia. International Journal of Science and Research
(IJSR), 2(6): 2319-7064.
International Rice Research Institute. 2013. Rice Almanac, 4th Edition: Source Book for One
of the Most Important Economic Activities on Earth. International Rice Research Institute
(Los Banos) p.283.
Khush, GS. 2005. What it will take to feed 5.0 billion rice consumers in 2030. Plant Mol. Biol.
59, 1–6.
Marschner, H. 1995. Mineral nutrition of higher plants. 2 nd Ed. Academic Press, London. 783p.
MoA (Ministry of Agriculture). 2011. Crop Variety Register. Issue No. 14. Animal and Plant
Health Regulatory Directorate, Addis Ababa Ethiopia.
MoARD (Ministry of Agriculture and Rural Development). 2005. Crop Variety Register, ,
Crop Development Department, Addis Ababa, Ethiopia.
MOARD (Ministry of Agriculture and Rural Development). 2010. Ministry of Agriculture and
Rural Development, Crop Variety Register, Crop Development Department, Addis
Ababa, Ethiopia: 48.
Mohaddesi, A Abbasian A, Bakhshipour S and Aminpanah H. 2011. Effect of different levels
of nitrogen and plant spacing on yield, yield components and physiological indices in
high yield rice. Amer-Eur Journal of Agriculture and Environment, 10: 893–900.
Mosalem, ME, Zahran M, El-Menofi MM and Moussa AM. 2002.Effect of sowing methods
and seeding rates on growth and yield of some wheat cultivars, pp. 239-392. In:
Proceedings of The 5th International Triticale Symposium, June 30-July 5, Radizikow,
Poland.
Mulugeta Seyoum. 1999. Rice: Potential crop for food security in Amhara Region; AgriTopia:
The Quarterly Newsletter of Ethiopian Agricultural Research Organization; 14(3/4): 1-2.
MulugetaSeyoum. 2000. Uptake and Response of Rice (Oryza sativa L.) to Nitrogen and
Phosphorus Fertilization Under Rain Fed Lowland Condition of The Fogera Plain,
Northern Ethiopia; Masters Thesis; Alemaya University.
NRRDSE. 2009. National Rice Research and Development Strategy of Ethiopia; MinistryOf
Agriculture and Rural Development, Addis Ababa, Ethiopia.
[277]
SAS (Statistical Analysis Software) Institute Inc. 2009. SAS/GRAPH 9.2: Reference, Fourth
Edition. Cary, NC: SAS Institute Inc.
Seleiman, MF, Ibrahim ME, Abdel-Aal SM and Zahran GA. 2010. Effect of seeding rates on
productivity, technological and rheological characteristics of bread wheat
(triticumaestivumL.).International Journal of Current Research. 4 : 075-80??.
Sewunet Ashebir. 2005. Effect of Nitrogen and Seeding Rates on Grain Yield, Yield
Components and Nitrogen Uptake of Rainfed Rice (OryzasativaL) in Fogera, South
Gonder. M. Sc. Thesis, Alemaya University.
Sultana M, Rahaman M, and M. Rahman. 2012. Effect of row and hill spacing on the yield
performance of Boro Rice (cv. BRRI dhan45) under aerobic system of cultivation. Journal
of the Bangladesh Agricultural University,.10(1): 39-42.
Sultana, MR, Rahman MM and Rahman MH. 2012. Effect of row and hill spacing on the yield
performance of boro rice (cv. BRRI dhan45) under aerobic system of
cultivation.Bangladesh Journal of Bangladesh Agricultural University, 10(1): 39-42.
Tilahun Tadesse, Minale Liben, Alemayehu Assefa, Belesti Yeshalem and Tesfaye Wossen.
2007. Effect of nitrogen and phosphorus fertilizers on the yield of rice in Fogera and
Metema areas. In: Ermiase A., Akalu T., Alemayehu A., Melaku W., Tadesse D., and
Tilahun T. (eds.). Proceedings of the 1st Annual Regional Conference on Completed Crop
Research Activities, 14- 17 August 2006. Amhara Regional Agricultural Research
Institute, Bahir -Dar, Ethiopia.
Yamada, N. 1961. On the relationship between yield and spacing in rice.Agricultural and
Horticultural.36(1): 13-18.
Yoshida, S. 1981. Fundamentals of rice crop science.Institutional Rice Research Institute,
Manila, Philippines: 269pp.
Zeng, L and Shannon MC. 2000. Effects of Salinity on grain yield and yield components of rice
at different seeding densities.Agronomy Journal; 92: 418–423.
Zewdie Bishaw, Paul C Struik and Van Gastel AJG. 2014. Assessment of on-farm diversity of
wheat varieties and landraces: evidence from farmer’s field in Ethiopia. African Journal
of Agricultural Research, 9 (38): 2948-2963.
[278]
Determination of NP Fertilizer Requirement for
Newly Released Medium Maturing Maize Varieties
at Jimma Zone, Southwestern Ethiopia
Muhidin Biya*, Sisay Gurmu, and Eshetu Yadete
Jimma Agricultural Research Center, P. O. Box 192, Jimma, Ethiopia
E-mail of the corresponding author: muhibiya@gmail.com
Abstract
The field experiments were carried out in two sites of Omonada woreda, Jimma
Zone, for three consecutive years 2016-2018 main cropping seasons. The treatments
consisted of factorial combinations of four Nitrogen and Phosphorous fertilizer rates
(69/52, 92/69, 115/86 and 138/104) and three maize varieties (BH547, BH546 and
SPRH 1) laid down in a randomized complete block design (RCBD) with three
replications. The analysis of data showed that all parameters of maize were
significantly affected by NP fertilizer rates and maize varieties except harvest index,
plant height, and logging percentage were not affected by NP fertilizer rates.
Maximum grain yield (7.11 t ha-1), biomass yield (10.46 t ha-1) and harvest index
(0.724) were recorded from BH546 variety and grain yield (7.11 t ha-1), biomass
yield (9.54 t ha-1) and harvest index (0.742) were recorded in the highest 138/104
NP fertilizer rates. The grain and biomass yield were significantly increased from
6.21 to 7.11 t ha-1 and 8.39 to 9.54 t ha-1 which means increased by 12.32% and
11.71% over control respectively from the application of 92/69 NP kg ha -1. In
conclusion, partial budget analysis based on the field prices of inputs and maize
grain yield showed that the BH546 gave the highest net benefit (25,006 ETB) with
acceptable MRR (160%) followed by the application of 138/104 kg ha -1 NP with net
benefit (21,445 ETB) and acceptable MMR (816%). Therefore, from current on-farm
input availability and economic feasibility, a medium maturing maize BH546 variety
with application 138/104 NP fertilizer rates taken as optimal and recommended for
farmer’s profitability under rainfed condition of maize (BH546 variety) in Omonada,
Jimma.
Keywords: Maize; BH547; BH546; SPRH 1; Nitrogen and Phosphorous fertilizer

rates
Introduction
Maize is the primary staple crop in Ethiopia and plays an important role in the
livelihood of the people of Ethiopia. Its availability and abundance determine the
level of welfare and food security in the country. In Ethiopia, future increases in
maize production to meet domestic demand will have to rely on improvements in
yield per hectare rather than on the expansion of maize production area. Enhanced
maize productivity can be achieved by increased use of modern production
techniques such as the adoption of hybrid maize varieties in line with the use of
fertilizer application and appropriate crop management practices.
[279]
Some soil types in Ethiopia contain enough amounts of essential nutrients for the
plant's development. But the majority of soils in the country contain low to
medium total N and found inherently low in available P. This could be due to long
term cultivation practice that depletes plant nutrients and soil resource degradation
as a result of soil erosion and run-off and this becomes a threat to agricultural
productivity in Ethiopia. Thus to produce some crop yields the application of
mineral fertilizers has been mandatory that previous fertilizer recommendations
for most potential maize production areas were found to vary from 90 - 110 kg/ha
N and 20 - 35 kg/ha P ( Tolessa et al.,2002). However, current observation in
maize production field show multi-nutrient deficiencies even in plots where
farmer attempted to apply fertilizers at rates they claim optimal for expected yield
levels. Thus, NP fertilizers must be applied at rates that can provide better nutrient
use efficiencies of some maize varieties or hybrid can be coupled with optimum
plant population and fertilizer application methods and rates that coincide with
peak need by the crop. Therefore, it is an appropriate period to propose fertilizer
studies for lately released medium maturing maize varieties that have been
developed and ready for use in different agro-ecologies of Ethiopia. Hence, an
effort made to study the effect of different fertilizer rates on the growth and yield
of newly released medium maturity maize varieties.

The current field experiments were conducted at two sites for three consecutive
main cropping seasons of Omonada woreda, Jimma Zone Southwestern Ethiopia
at farmers’ fields. The sites were located on 7º46' N and 36º 00'E and laid at an
altitude of 1753 m.a.s.l. with soil type of the area is Upland: Chromic Nitosol and
Combisol. The average maximum and minimum temperatures are 9ºC and 28ºC
respectively and reliably receive good rains 1561 mm per annum cropping season.
The medium maturing maize varieties (BH547, BH546 and SPRH 1) which are
adapted to low-mid altitude (1000-1800 masl) areas and released recently. They
were white-colored and used for the experiment to maximize the yield potential of
smallholder farmers. The combination of four Nitrogen and Phosphorous fertilizer
rates (69/52, 92/69, 115/86 and 138/104) and three varieties (BH547, BH546, and
SPRH 1) twelve treatments were laid out in a randomized complete block design
(RCBD) with three replications. Blocks were separated from each other by 1.5 m
wide-open space, while experimental plots within replications were separated by 1
m apart from each other. The gross size of each plot was 4.0m length by 5.25m
width (21 m2) accommodating 8 rows. The inner 6 rows used for data collection.
Nitrogen and phosphorus fertilizers were applied, respectively per stand or hill
base. To increase the nitrogen use efficiency, it was split into two equal rates and
applied at planting time and knee height stages. The experimental field was
prepared following the conventional tillage practice and furrow opened by using
[280]
oxen. Two maize seeds were planted per hill and thinned after establishment to
maintain a single healthy plant per hill. All other agronomic practices like three
times hand weeding were applied uniformly to both experimental plots as per their
respective recommendations for maize in the study area. The data collected were
growth, yield, yield related and other agronomic data were collected.
Soil Sampling and Analysis

A composite surface soil (0-30 cm depth) sample was collected from both sites
with a gauge auger before planting; the experimental field was blocked into three
parts depending upon land uniformity. Plant residues on the sampling soil surface
were removed. Soil pH was determined in a 1:2.5 soil-water suspension using a
combination of the glass electrode. Organic carbon was estimated by the wet
digestion method (Okalebo et al., 2002) and organic matter was calculated by
multiplying the percent organic carbon (OC) by a factor of 1.724. Total soil N was
measured using the micro-Kjeldahl digestion, distillation and titration procedure
as described by AOAC (1994). After extraction of the soil sample by sodium
bicarbonate solution as per the procedure outlined by Olsen et al. (1954), available
P was determined by measuring its absorbance using a spectrophotometer. The
investigated soil properties are shown in Table 1.
Table 1. Physiochemical properties of the experimental soil (0-30 cm depth) before sowing.
Omonada Rating Source

pH 1:2.5 (H2O) 5.03 Strongly acid Batjes, (1995) Tekalign (1991)
Organic matter (%) 3.52 Medium Tekalign (1991)
Total N (%) 0.17 Medium Berhanu, (1980)
Organic carbon 2.04 medium Cottenie, 1980
Available P (mg kg-1) 9.45 Low
Source: Jimma agricultural research center soil laboratory.
Economic analysis
To assess the costs and benefits associated with different treatments (inter and
intra row spacing), the partial budget technique as described by CIMMYT (1988)
was applied. Economic analysis was done using the prevailing market prices for
inputs at planting and outputs, at the time the crop was harvested. All costs and
benefits were calculated on ha-1basis of Ethiopian Birr (ETB). The inputs and/or
concepts used in the partial budget analysis were the mean grain yield of each
treatment in both years, the field price of QPM Hybrid, BHQPY545 maize grain
(sale price grain yield minus the costs of fertilizer, planting, seed), the gross field
benefit (GFB) ha-1(the product of field price of the mean yield for each treatment),
the field price of Seed rate kg ha-1, fertilizer and wage rate, the total costs that
varied (TCV) which included the sum of field cost of seed, fertilizer and its wage
for planting and application. The net benefit (NB) was calculated as the difference
between the GFB and the TCV. The actual yield was adjusted downward by 10%
[281]
to reflect the difference between the experimental yield and the yield farmers
could expect from the same treatment. There were optimum plant population
density, timely labour availability, and better management (e.g. weed control,
rainfall) under the experimental conditions CIMMYT, (1988). The dominance
analysis procedure as detailed in CIMMYT (1998) was used to select potentially
profitable treatments from the range that was tested. The discarded and selected
treatments using this technique were referred to as dominated and undominated
treatments, respectively. The undominated treatments were ranked from the lowest
to the highest cost. For each pair of ranked treatments, the percent marginal rate of
return (MRR) was calculated. The MRR (%) between any pair of undominated
treatments was the return per unit of investment in fertilizer. To obtain an estimate
of these returns the MRR (%) was calculated as changes in NB divided by changes
in cost. Thus, the MRR of 100% was used indicating for every one EtB expended
there is a return of one ETB for a given variable input. Sensitivity analysis for
different interventions was also carried out to test the recommendation made for
its ability to withstand price changes. Sensitivity analysis simply implied redoing
marginal analysis with the alternative prices. Through sensitivity analysis, the
maximum acceptable field price of an input was calculated with the minimum rate
of return as described by Shah et al. (2009). Analysis of variance (ANOVA) for
all collected data was computed using SAS software version 9.3. The significance
of differences between samples was separated using the least significant difference
(LSD) at 5% level of significance.

The combined analysis of variance (ANOVA) for the effects of NP fertilizer rates
and maize varieties both location and three seasons were did not show significant
(P <0.05) interaction all parameters. Grain and biomass yield were significantly (P
<0.01) influenced by both NP fertilizer rates maize and varieties (Table 6) and
Plan height, logging percentage and harvest index were showed significantly (P
<0.05) only influenced by maize varieties (Table 1 and 2), While Plan height,
logging percentage and harvest index were not showed significantly (P <0.05)
influenced NP fertilizer rates (Table 2).
The plant height was significantly (P<0.05) influenced by the main effects maize
varieties at both locations and three seasons and not significantly (P<0.05)
influenced by NP fertilizer rates and interaction effect (Table 2). The SPRH 1
variety had the highest plant height of 282.06 cm, followed by BH546 variety with
a height of 27.85cm. Among varieties, BH547 variety had the shortest 250.48cm
height (Table 2). The result showed the plant height and growth were vary with
varieties even though they were a similar group of maturity and variety SPRH 1
had the highest plant height 282.06 cm than all tested varieties. Similarly, Tolera
A. et al., (1999) reported that breeders should select maize varieties that combine
[282]
high grain yield and desirable stover characteristics because of large differences
that exist between cultivars. The highest plant height 266.58 cm and followed by
265.56 cm were recorded from 138/104 and 115/86 NP fertilizer rates
respectively, while the shortest plant height 258.56 was recorded from the lowest
69/52 NP fertilizer rates. There was an increase in plant highest with an increase
in N rate. The current result was in agreement with Rashid et al. plant height was
linearly increased with increasing levels of N fertilization. Similarly, Harris P.
(1995) reported plant height, on each treatment with an increase of NKP rate and
animal manure application.
Table 2. Across season and location main effects of NP and medium maturing maize
varieties on yield and yield component parameters at Omonada
Varieties Plant Logging

height (cm) percentage
BH547 250.48 32.66
BH546 257.85 15.97
SPRH1 282.06 41.89
LSD (0.05) 10.81 6.04
NP (P2O5 kg/h-1) fertilizer rates
69/52 258.56 30.53
92/69 263.17 29.32
115/86 265.56 30.41
138/104 266.58 30.44
Mean 263.47 30.17
LSD (0.05) Ns Ns
CV (%) 12.48 20.88
Where, LSD (0.05) =Least Significant Difference at 5% level; CV=Coefficient of Variation.
The highest Logging percentage of 41.89 was recorded from the SPRH 1 variety
and followed by 32.66 from BH547 variety, while the lowest 15.97 was recorded
from the BH546 variety. It was significant differences (P<0.01) effect only due to
varieties but didn’t show the interaction effect and influenced by NP fertilizer
rates (Table 2). The varietal difference shows that the variety SPRH 1 gave the
highest plant height and logging percentage. So, these highest logging percentage
was may be resulted due to the highest plant height even though there were other
factors like wind. Similarly, Odeleye and Odeleye (2001) reported that maize
varieties differ in their growth characters, yield and its components.
The biomass yield was significant differences (P<0.01) effect due to NP fertilizer
rates and maize varieties (Table 3). The highest biomass yield 10.46 and 9.54 t ha-
1
were recorded from the BH546 and 138/104 respectively from the highest NP
rate. On the contrary, the lowest 7.02 and 8.39 t/ha from the BH547 and 69/52
respectively from the lowest NP rate (Table 2). The highest 138/104 NP rate of
biomass yield was significantly increased from 8.57 to 9.54 t ha-1 which means
increased by 11.71% over control from the application of 92/69 NP kg ha-1
[283]
(control). The result also showed that an increase in biomass yield with increasing
NP rate. These results in agreement with Tariku B. e al (2018) reported that the
application of higher nitrogen increased the dry matter of plants.
Table 3. Across season and location main effects of NP and medium maturing maize varieties on yield and yield
component parameters at Omonada
Grain yield Biomass
Varieties tone ha-1 yield t ha-1 Harvest Index
BH547 6.83 9.18 0.739
BH546 7.82 10.46 0.746
SPRH1 5.10 7.02 0.724
LSD (0.05) 0.48 0.59 0.126
NP (P2O5 kg/h-1) fertilizer rates
69/52 6.21 8.39 0.736
92/69 6.33 8.57 0.733
115/86 6.68 9.06 0.734
138/104 7.11 9.54 0.742
Mean 6.58 8.89 0.736
LSD (0.05) 0.56 0.69 Ns
CV (%) 20.45 20.35 5.19
Where, LSD (0.05) =Least Significant Difference at 5% level; CV=Coefficient of Variation.
Grain yield was significantly (P<0.01) affected by the effect of NP fertilizer rates
and maize varieties and doesn’t show the interaction effect (Table 3). The results
of ANOVA showed that the highest mean grain yield (7.82 t ha-1) was recorded
from BH546 variety and followed by (7.11 t ha-1) were recorded in the highest
138/104 NP fertilizer rates, while the lowest grain yield (6.21 t ha-1) and biomass
yield (7.02 t ha-1) were recorded in SPRH 1variety and the lowest 69/52 NP
fertilizer rates. The highest 138/104 NP rate of Grain yield was significantly
increased from 6.33 t ha-1 to 7.11 t ha-1 which means increased by 12.32% over
control from the application of 92/69 NP kg ha-1 (control). This is similar to the
findings of Ghimire S. et al. (2016) reported that grain yield was significantly
affected by crop varieties sown. Irrespective of fertilizer dose and irrigation
maximum grain yield ranging from (3.17 to 7.25 t ha-1) and (1.60 to 6.32 t/ha) was
produced by Rajkumar in improved practice and farmers practice of cultivation
respectively. Also supported by Obi (1999), Kim (1997) and Udoh (2005) who
reported that some hybrid maize varieties have a yield advantage over other maize
varieties because they possess such special qualities as high yield, disease
resistance, and early maturity uniformity in flowering and ear placement,
harvester.
Harvest index: The harvest index was significantly (P<0.001) influenced by

varieties and not significantly (P<0.05) influenced by the application of NP
fertilizer rates (Table 3). Significantly higher harvest index of 0.746 was obtained
from BH546 variety and while the lowest was 0.724 from SPRH 1 variety. The
second-highest harvest index 0.742 was obtained from the application of the
[284]
highest 138/104 NP fertilizer rate. The result showed that there was an increase in
harvest index with an increase in N rate. In line with this result, Lawrence (2008)
reported that the harvest index in maize increased when nitrogen rates increased.
Similalry, Orkaido (2004) reported that increasing N level from 0 to 120 kg N ha-1
increased the harvest index of maize.
Economic Viability of maize varieties and NP fertilizer rates

Analysis of variance (Table 1) showed that verities and NP had a significant (P =
0.001) effect on the grain yield of QPM Hybrid, BHQPY545 maize whereas
interaction was not significant. An economic analysis of the combined results
using the partial budget technique was thus appropriate (CIMMYT, 1988). The
result of the partial budget analysis and the data used in the development of the
partial budget is given in (Tables 4 and 5). It was performed by considering
fertilizer cost, application cost, and labour as the main input, mean grain yield
obtained across season and location. The total costs of fertilizers (NPS = 15.90
ETB/kg and urea = 12.65 ETB/kg and sale of grain maize at Omonada open
market average price (4.00 ETB/kg). Dominance analysis (Table 9) led to the
selection of treatments only BH547 and BH546 from varieties and 69/52, 115/86
and 138/104 NP kg ha-1 from NP rates were ranked in increasing order of total
costs that vary. The treatments having MRR below 100% was considered and
unacceptable to farmers; thus, SPRH 1 variety and 92/69 NP kg ha-1 were
eliminated (CIMMYT, 1988) (Table 4). This was because such a return would not
offset the cost of capital (interest) and other related deal costs while still giving an
attractive profit margin to serve as an incentive. partial budget analysis based on
the field prices of inputs and maize grain yield showed that The BH546 gave the
highest net benefit (25006 ETB ha-1) with acceptable MMR (160%) followed by
and application of 138/104 kg ha-1 NP net benefit (21445 ETB ha-1) and
acceptable MMR (816%) (Table 5).
Market prices are ever-changing and as such a recalculation of the partial budget
using a set of likely future prices i.e., sensitivity analysis, was essential to identify
treatments that may likely remain stable and sustain satisfactory returns for
farmers despite price fluctuations. The sensitivity analysis study indicates an
increase in the field price of the total variable costs, and a fall in the price of maize
grain, which represented a price variation of 15% (Table 6). The price changes are
sensitive under market conditions prevailing at Omonada which were below the
minimum acceptable MRR of 100% for BH546 and also both application 115/86
and 138/104 kg ha-1 N/P2O5 fertilizer. Therefore, this investigation remained with
changes to BH546 with application 138/104 NP kg ha-1 as promising new
practices for farmers at Omonada under the prevailing price structure since they
gave more than 100% MRR. These results agree with Bekele, H. (2000).
[285]
Table 4. Partial budget analysis for varieties and NP rates for at current prices.
Grain yield Adjusted Grain Yield Gross Field TCV Net Benefit Marginal Rate of
Varieties tone ha-1 tone ha-1 Benefit (ETB/ha) (ETB/ha Return (%)
BH547 6.83 6.15 24,588.00 1776 22,812 U
BH546 7.82 7.04 28,152.00 3146 25,006 U
SPRH 1 5.10 4.59 18,360.00 3851 14,509 D
NP fertilizer rates (kg ha-1) -
69/52 6.21 5.59 22,356.00 2076 20,280 U
92/69 6.33 5.70 22,788.00 2761 20,027 D
115/86 6.68 6.01 24,048.00 3446 20,602 U
138/104 7.11 6.40 25,596.00 4151 21,445 U
TCV= total cost that varied, Retail price of grain =Birr 4.00 per kg; EtB = Ethiopian Birr; Fertilizers urea = Cost of Birr
12.65, per kg; NPs =Cost Birr 15.90 per kg; MMR= Marginal Rate of Return; NB = Net benefit;
Table 5. Partial budget with estimated marginal rate of return (%) for varieties and NP rates at current prices.
Grain yield Adjusted Grain Yield Gross Field TCV Net Benefit Marginal Rate of
Varieties tone ha-1 tone ha-1 Benefit (ETB/ha) (ETB/ha Return (%)
BH547 6.83 6.15 24,588.00 1776 22,812 ----
BH546 7.82 7.04 28,152.00 3146 25,006 160
NP fertilizer rates(kg ha-1)
69/52 6.21 5.589 22,356.00 2076 20,280 ----
115/86 6.68 6.012 24,048.00 3446 20,602 23
18/104 7.11 6.399 25,596.00 4151 21,445 120
TCV= total cost that varied, Retail price of grain =Birr 4.00 per kg; EtB = Ethiopian Birr; Fertilizers urea = Cost of Birr
Table 6 Sensitivity analysis of maize production based on a 15% rise in total cost and maize
price of gross field benefit fall
NB Raised Raised
N/P2O5 Fertilizer TVC(ETB ha-1) (ETB ha-1) cost benefit MRR (%)
BH547 5138.31 19390.20
BH546 3617.90 21255.10 1520.41 1864.9 ----
-1
NP fertilizer rates(kg ha )
69/52 2387.40 17238.00
115/86 3962.90 17511.70 1575.5 273.7 17
138/104 4773.65 18228.25 810.75 716.55 88
* TCV= total cost that varied, Retail price of grain =Birr 4.00 per kg; EtB = Ethiopian Birr; Fertilizers urea = Cost of Birr
Summary and Conclusion

From all, the most important parameters were significant (P = 0.001) for the
effects of maize varieties and NP fertilizer rates. The highest biomass yield 10.46
and 9.54 t ha-1 was recorded from the BH546 and 138/104 respectively from the
highest NP rate. The highest mean grain yield (7.82 t ha-1) was recorded from
BH546 variety and followed by (7.11 t ha-1) were recorded in the highest 138/104
NP fertilizer rates, while the lowest grain yield (6.21 t ha-1) and biomass yield
[286]
(7.02 t ha-1) were recorded in SPRH 1variety and the lowest 69/52 NP fertilizer
rates. The highest 138/104 NP rate of Grain yield was significantly increased from
6.33 t ha-1 to 7.11 t ha-1 which means increased by 12.32% over control from the
application of 92/69 NP kg ha-1 (control). The partial budget analysis based on the
field prices of inputs and maize grain yield showed that The BH546 gave the
highest net benefit (25006 ETB ha-1) with acceptable MMR (160%) followed by
and application of 138/104 kg ha-1 NP net benefit (21445 ETB ha-1) and
acceptable MMR (816%). Therefore, from current on-farm input availability and
economic feasibility, a medium maturing BH546 variety with the application of
138/104 NP fertilizer rates taken as optimal and recommended for farmer’s
profitability under rainfed condition of maize (BH546 variety) in Omonada,
Jimma zone and other similar humid agro-ecologies of the west and southwest
Ethiopia.
References
AOAC (Association of Official Analytical Chemist). 1994. Official Method of Analysis, 12 Ed.
Washington,
Batjes, NH, 1995. A global data set of soil pH properties. Technical Paper 27, International Soil
Reference and Information Centre (ISRIC), Wageningen.
Berhanu, D. 1980. A survey of studies conducted about soil resources appraisal and evaluation for
rural development in Ethiopia, Addis Ababa.70p
Cottenie, A. 1980. Soil and plant testing as a basis of fertilizer recommendations. FAO soils
Bulletin 38/2, FAO, Rome
Bekele, H. 2000. Integrated nutrient management in irrigated wheat (Triticumaestivum L.). MSc
Thesis, University of Agricultural Sciences, Dharwad, India. Effects of Sowing Methods
and Seed Rates on Yield Components and Yield of Tef in SoroWoreda, Hadya Zone,
Southern Ethiopia
CIMMYT (International Maize and Wheat Improvement Center). 1988. Partial budget analysis for
mean grain yield.
CIMMYT. 1988. From Agronomic Data to Farmer Recommendations: An Economics Training
Manual. Completely revised edition. Mexico, D.F.79p.
Ghimire, S Sherchan DP, Andersen P, Pokhrel C, Ghimire S and Khanal D. 2016. Effect of Variety
and Practice of Cultivation on Yield of Spring Maize in Terai of Nepal.
Harris, PB. 1995. The effect of different rates and times of application of nitrogen on the yield and
quality of whole crop barley. Experimental Husbandry, 28: 1-6.
Kim, SK. 1997. Achievement, challenges and future direction of hybrid maize research and
product in B. BaduApraku, Akoroda M.O., Oedraw M. and Quin F.M(eds). Proceedings of
Required Maize Workshop May 99-June 2, 1995. IITA Cotonou, Benin Republic.
Obi, IU. 1999. Effect of nitrogen Rates And Intra –Row Spacing on Local maize (Zea mays) in
southern Guinea Savannah Zone of Nigeria. Journal of Sustainable Agriculture and
Environment, 5: 147-152.
Odeleye, FO and Odeleye, MO. 2001. Evaluation of morphological and agronomic characteristics
of two exotic and two adapted varieties of tomato (Lycopersicomesculentum) in southwest
Nigeria. Proceedings of the 19 Annual Conference of HORTSON, 1: 140-145.
Okalebo, JR Gathua KW and Woomer PL. 2002. Laboratory methods of soil and plant analysis: A
working manual, 2 nd ed. TSBF-CIAT and SACREDAfrica, Nairobi, Kenya.
Olsen, SR Cole C, Watanable FS and Dean LA. 1954. Estimation with sodium phosphorus in soil
by extraction with sodium bicarbonate, USDA Circular 939: 1-19.
[287]
Orkaido, O. 2004. Effects of nitrogen and phosphorus fertilizers on yield and yield components of
maize (Zea mays L.) on black soil of Regede, MSc Thesis. Haramaya University, Ethiopia.
SAS Institute Inc. Cray. 2008. Users Guide. Version 9.3. NC.USA.
Tariku Beyene, Tolera Abera and Ermiyias Habte. 2018. Effect of Integrated Nutrient Management
on Growth and Yield of Food Barley (Hordeumulgare) Variety in Toke Kutaye District,
West Showa Zone, Ethiopia
Tekalign Tadese. 1991. Soil, Plant, Water, Fertilizer, Animal Manure, and Compost Analysis.
Working Document No. 13. International Livestock Research Center for Africa (ILCA),
Tolera, A, Berg T and Sundstol F. 1999. The effect of variety on maize grain and crop residue
yield and nutritive value of the Stover. Journal of Animal feed Science and Technology 79:
165-177.
Tolessa D, T Bogale, W Negassa and T Workayahu, M Liben, T Mesfin, B Mekonen and W
Mazengia. 2002. A Review of fertilizer management on maize in Ethiopia. pp46-55. In:
Mandefri Nigussie, D. Tanner and S.Twumasi-Afriya (eds.). Enhancing the contribution of
maize to food security in Ethiopia: Proceedings of the second national maize workshop of
Ethiopia. 12-16 November 2001, Addis Ababa, Ethiopia. EARO/CIMMYT.
Udoh, J. 2005. Crop Production Techniques for the Tropics Concept Publications Limited,
Munshin, Lagos Nigeria, 101-106.
[288]
Effect of NP Fertilizer Rates and Plant Population
Density on Late Maturing Maize Variety
Sisay Gurmu*, Muhidin Biya, and Eshetu Yadete
E-mail of the corresponding author: sis.sis1835@gmail.com
Abstract
Maize is among the leading cereals in production globally and an important
potential food security crop in Ethiopia. However, its productivity is very low mainly
due to low soil fertility and plant stands per area. Therefore, a field experiment was
conducted to determine the effect of NP fertilizer and plant density at Jimma and
BunoBedele Zone, Southwestern Ethiopia during the 2016-2018 main cropping
seasons. The experiment had four nitrogen and phosphorus rates (P2O5); 69/52,
92/69, 115/86 and 138/104 kg ha-1; blended NPSZnB fertilizer (150 kg ha-1 +140 kg
ha-1 urea top-dressed) and four plant density; 44,444, 53,333, 66,666, and 88,888.
Each treatment was assigned to each plot in split-plot design in which fertilizer
assigned to subplot and plant density assigned to the main plot. Data on the yield
and yield components were subjected to ANOVA using SAS version 9.3. The highest
grain yield and above ground biomass 7566.2 kg ha-1 and 15.07 ton ha-1 respectively
were recorded from 138/104 N/P2O5. On the other hand, the highest grain yield 7182
kg ha-1 and above ground biomass19.94 ton ha -1 was recorded from 66,666 plant ha-
1
. The highest net benefit 35,050.18 and 37,220.6 ETB were obtained from 138/104
kg ha-1 N/P2O5 and plant density of 53,333 (75*25cm). Therefore, it is advisable for
farmers in the study area and adjacent woredas with similar agro-ecologies, a plant
density of 53,333 plant ha-1 (75 x 25cm a plant hill-1 or 75 x 50 cm two plants hill-1)
in complement with N/P2O5 fertilizer rate of 92/69 kg ha-1 can be recommended.
Keywords: Grain yield, organic fertilizer, inorganic fertilizer, and plant

population
Introduction
Like any other tropical country, most cropping systems in Ethiopia are traditional
and crop bases are complex, vary across agro-ecologies, and diverse according to
the cultural food needs of resource-poor farmers. Maize is commonly planted in
rows of varying spaces; less effort has been made to plant at optimum densities to
maximize its productivity in different agro-ecologies of Ethiopia. Summaries of
earlier results from different studies on maize plant population densities indicate
that better yields were obtained at planting density in the range of 4-7 plants m-2
(40,000-70,000 plants ha-1)(Tenaw et al., 1992). Later studies confirmed that at 5-
7 plants/m2 medium to late maize maturity groups gave maximum yields in humid
regions, while early maturity groups produced maximum yields at higher densities
in both humid and moisture stress areas (Tenawet al., 2002). It is being observed
that lately innovated medium and early maize varieties in humid lowlands and low
[289]
moisture stress area found to be varied in structure and leaf arrangements from
known normal maize varieties. These variations in morphology may lead to
different planting density to reach at their at maximum yield potentials.
Plant population density have a significant impact on growth and yield of crops,
including maize, a popular C4 cereal crop (Cox, 1996). Therefore, understanding
how plants regulate their growth in response to plant population densities has
problems, such as the determination of optimal sowing density (Cox, 1996).
Increased plant populations could lead to increased yields under optimal climatic
and management conditions due to a greater number of smaller cobs per unit area
(Bavec and Bavec, 2002).
Plant population is the prime factor for getting maximum yield which is decided
by inter and intra row spacing of crops. Decreasing the distance between neighbor
rows at any particular plant population has several potential advantages. First, it
reduces competition among plants within rows for light, water, and nutrients due
to a more equidistant plant arrangement (Porter et al., 1997). The more favorable
planting pattern provided by closer rows enhances maize growth rate early in the
season (Bullock et al., 1988), leading to a better interception of sunlight, a higher
radiation use efficiency, and a greater grain yield (Westgate et al., 1997).
Secondly, the maximization of light interception from early canopy closure also
reduces transmittance through the canopy (McLachlan et al 1993). The smaller
amount of sunlight striking the ground decreases the potential for weed
interference, especially for shade-intolerant species (Johnson et al.,1998). Thirdly
the quicker shading of soil water being lost by evaporation (Karlen and
Camp,1985).This is especially important under favorable soil surface moisture
conditions because it allows maize plants to maximize photosynthesis and the
proportion of water that is used growth processes rather than evaporated from the
soil (Lauer,1994). Furthermore, earlier crop cover provided by smaller row width
is instrumental to enhance soil protection, diminishing water runoff and soil
erosion (Sangoi et al., 1998). The nutrient use efficiency can be improved with the
use of optimum plant population (Srikanth et al., 2009).
However, according to Duncan (1984) plant population above critical density has
a negative effect on yield per plant due to the effects of interplant competition for
light, water, nutrient and other potential yields limiting environmental factors.
Most soils contain an abundance of elements essential for plant development, but
the majority of these elements are rarely available for plant use due to nutrient
loss. Nutrient depletion and soil degradation have become a serious threat to
agricultural productivity in Ethiopia. These soils suffered multi-nutrient
deficiencies so that, application of mineral fertilizers has become mandatory to
increase crop yields in such soils (Adeniyan and Ojeniyi, 2005).
[290]
According to Srikanth et al. (2009) among the plant nutrients, primary nutrients
such as nitrogen and phosphorus play a crucial role in determining the growth and
yield. The nutrient use efficiency can be proved with the use of hybrids, optimum
plant population, and application of chemical fertilizer coinciding with peak need
by the crop. Hence, the experiment was done to evaluate the response of maize
hybrid BH661 to different rates of NP fertilizer and plant population density; and
to identify the economic optimum of NP fertilizer rate for production of maize.

Description of the Study Area
The on-farm field experiment was conducted in two sites of Jimma Zone (Kersa
and Omonada woreda) and Buno-Bedele Zone (Banshure kebele), Southwestern
Ethiopia during the cropping season of 2016-2018. The Kersa site was located on
latitude 7º42' N and longitude 36º 59'E and found at an altitude of 1753 m.a.s.l.
The average minimum and maximum temperature is 6ºC and 25.5ºC respectively
and reliably receives good rains 1712 mm per annum during the cropping season.
Whereas, Omonada site was located on latitude 7º37' N and longitude 37º 14'E
and laid at an altitude of 1753 m.a.s.l. The average minimum and maximum
temperature is 6ºC and 25ºC respectively and reliably receives good rains 1446
mm per annum cropping season. The Bedele site was located on latitude 8º32' N
and longitude 36º 22'E and found at an altitude of 1753 m.a.s.l. The average
minimum and maximum temperature is 6ºC and 24.5ºC respectively and reliably
receives good rains 1712 mm per annum cropping season. The farming system of
the study site is coffee and cereal crops dominated with coffee, maize, teff, and
sorghum also has a warm and cold climate, also convenient topography which is
very suitable for all agricultural practices. It was situated in the tepid to cool
humid-mid highlands of southwestern Ethiopia. The soil type of the experimental
area was Eutric-nitisols (reddish-brown).
Soil Physico-chemical Properties

The soil of the experimental field was characterized for selected physico-chemical
properties before the application of the treatments (Table 1). The average soil pH
of the trial sites ranges from 4.96 to 5.00, which was strongly acidic (Batjes, 1995)
and ideal for the production of most field crops. The pH of the soil affects maize
growth by suppressing the root development and reducing the availability of
macronutrients to plants especially phosphorus (Brady and Weil, 2008). The soil
total N ranges from 0.19 to 0.20% and OC from 2.02 to 2.20% were found a
medium rate for crop growth and development for both nutrients (Berhanu, 1980).
The Bray II extractable available P for Kersa woreda was 3.72 mg kg-1, which is
below the critical level (8 mg kg-1) for most crop plants whereas, Omonada and
Bedele sites were ranges from 8.96 to 12.32 mg kg-1 which is medium to high in
content as described by Tekalign and Haque (1991).
[291]
Table 1 Selected Physico-chemical properties of the soil of the experimental sites before planting
Location
Soil characters Kersa Omonada Bedele
pH(1:2.5) 4.96 5.15 5.00
av P(mg kg-1) 3.72 12.32 8.96
TN (%) 0.20 0.19 0.20
OC (%) 2.12 2.02 2.20
OM (%) 3.65 3.47 3.78
C:N ratio 10.61 11.05 11.40
Where pH= hydrogen power, OC=organic carbon, TN=Total Nitrogen, Av.
P=Available phosphorous, OM=Organic matter. Values are the means of duplicated samples.
Description of the experimental materials

In the present study, maize varieties BH661 hybrid was used. It is the most
promising variety released by Bako Agricultural Research Centre and which
adapted well to the agro-ecologies of Jimma and Buno Bedele areas. The
experimental field was plowed and prepared following the conventional tillage
practice before planting at all experimental locations. The land was leveled using
manual power before the field layout was made. Two maize seeds were planted
per hill and then thinned to one plant per hill after good establishment of seedlings
to maintain a single healthy plant per hill. Four nitrogen and phosphorus rates
(P2O5); 69/52, 92/69, 115/86 and 138/104 kg ha-1; blended NPSZnB fertilizer (150
kg ha-1 + 140 kg ha-1urea top-dressed) and four plant population density; 44,444,
53,333, 66,666, and 88,888 plants per hectare were combined and randomly
assigned to each plot. The split-plot design with three replications and plot size
22.95m2 (4.5 m x 5.1 m) for each treatment was used. Each factor within a
replication was randomly arranged to plant population densities in the main plot
and fertilizer rates in a subplot. Thus, the number of treatments per replication was
20 and total plots per experiment were sixty (60).
Nitrogen fertilizer rates were applied in a split in such a way that half during
planting and half at the knee height growth stage to increase the nitrogen use
efficiency. All other agronomic practices were applied uniformly to all
experimental plots as per their respective recommendations for maize in the study
area.
The data collected were growth, yield and yield-related parameters. Analysis of
variance (ANOVA) for all collected data was computed using SAS version 9.3
statistical software. Whenever the ANOVA results showed the significant
differences between sources of variation, the means were separated using the least
significant difference. The homogeneity test was done as suggested by Gomez and
Gomez, (1984).
[292]
Partial budget analysis
Partial budget analysis was performed to investigate the economic feasibility of
the treatments and assess the costs and benefits associated with different
treatments of chemical fertilizers and plant population density of the seed rates.
The partial budget technique as described by CIMMYT (1988) was applied. The
partial budget analysis was done using the prevailing market prices for inputs at
planting and outputs, at the time the crop was harvested. All costs and benefits
were calculated on a hectare basis in Ethiopian Birr (ETB). The inputs and/or
concepts used in the partial budget analysis were the mean grain yield of each
treatment, the gross field benefit (GFB) ha-1 (the product of field price and the
mean yield for each treatment), the field price of chemical fertilizers and urea kg-1
(the nutrient cost plus the cost of transportation from the point of sale to the farm),
cost of labor spent on seed purchase and planting, the total costs that varied (TVC)
which included the sum of field costs of fertilizers and their application, and seed
purchase and planting.
The net benefit (NB) was calculated as the difference between the GFB and the
TVC. The marginal rate of return (MRR %) was also calculated. To obtain an
estimate of these returns, the MRR (%) was calculated as changes in NB divided
by changes in cost. Thus, a minimum acceptable MRR of 100% was used
indicating, for every one ETB expended there is a return of
one ETB for a given variable input (CIMMYT, 1988), which is suggested to be
realistic. This enables’ to make farmer recommendations from marginal analysis.
The dominance analysis procedure as detailed in CIMMYT (1988) was used to
select potentially profitable treatments from the range that was tested. Sensitivity
analysis for different interventions was also carried out to test the recommendation
made for its ability to withstand price changes. Sensitivity analysis simply implied
redoing marginal analysis with the alternative prices. Through sensitivity analysis,
the maximum acceptable field price of an input was calculated with the minimum
rate of return as described by Shah et al. (2009).

The effect of different rates of NP fertilizer and plant population density on yield
and yield-related parameters and cost-benefit analysis were presented and
discussed as follows. It has been observed that late maturing maize variety like
BH661 vary in stature and leaf arrangements from other medium maturing/early
maturing maize varieties. These variations in morphology may lead to different
planting density to reach their maximum yield potentials.
The combined analysis effects of NP fertilizer rates and plant population density
over locations and seasons didn’t show significant (P >0.05) interaction effect on
plant height, ears per plant, lodging percent, grain yield, aboveground biomass and
[293]
HI. Plant height was not significantly influenced by both NP fertilizer and plant
population density, whereas ears per plant, harvest index, grain yield, and above-
ground biomass were highly significantly (P <0.01) influenced by both NP
fertilizer rates and plant population density. However, lodging percentage was not
significantly (P >0.05) affected by NP fertilizer rates, but it was highly
significantly (P <0.01) affected by plant population density (Table 2).
The interaction effect of NP fertilizer and plant population density, and the main
effect of NP fertilizer and plant population density were not significantly (P
>0.05) affected plant height (Table 2). Over season and location mean indicated
that the maximum number of ears per plant (0.97) was recorded from 138/104 kg
ha-1 N/P2O5 fertilizer. But its effect was statistically at par with 69/52, 92/69, and
115/86 kg ha-1 N/P2O5 fertilizer, while the minimum number of ears per plant
(0.90) was recorded from 150 kg ha-1 blended NPSZnB +140 kg ha-1 urea top-
dressed (Table 2). Application of 138/104 kg ha-1 N/P2O5 fertilizer increased the
number of ears per plant by 5.6% over blended NPSZnB fertilizer. These results
were in line with the result of Singh and Nepalia (2009) who reported the
application of a higher dose of chemical fertilizer improved the number of cobs
plant-1 in QPM hybrid significantly over the lower nutrient application.
On the other hand, the maximum number of ear per plant (1.06) was recorded
from 44,444 plants ha-1 (75*30cm), while the minimum number of ear per plant
(0.81) was recorded from 88,888 plants ha-1 (75*15cm) (Table 2). The plant
population density at 44,444 plants ha-1 increased 30.9% number of ears per plant
over 88,888 plants ha-1. The results indicate as plant population density increased
the number of ear per plant decreased. At low plant population density, the
number of plants limited the yield, while at high plant population density number
of barren plants limited yield as well. This might be due to the efficient use of the
crop to the nutrient applied per plant stand, and this which in turn had increased
the nutrient availability for vigorous plant growth thus might have increased the
number of ears plant-1. These findings are in agreement with Hashemi-Dezfouli
and Herbert (1992) who reported a significantly higher number of ears per plant at
lower plant density as compared to higher plant density.
The maize variety BH-661 was sensitive to lodging, as its plant height goes higher
up to 291 cm as indicated in table 2 below. Stalk lodging represents one of the
most serious constraints to the use of high plant densities in maize BH-661
variety. Numerically, the highest lodging percent (57.6%) was recorded from the
higher plant population density of 88,888 plants ha-1 (75*15cm) while, the
minimum lodging percent (46.9%) was recorded from 44,444 plants ha-1
(75*30cm) (Table 2). The plant population density at 44,444 plants ha-1 decreased
18.6% lodging percent over 88,888 plants ha-1. The results indicate as the plant
population density increased the lodging percent also increased and vice versa. As
the plant density increases the internodes become thinner, making the plant more
[294]
prone to stalk lodging (Song et al., 2016). The stored carbohydrates in the maize
stalks were transported to grains and weakened the basal internodes, thus reducing
the bending quality and providing an ease of lodging (Xue et al., 2016); this is
because the basal internodes act as a lever for holding the plants upright (Yuan et
al., 2002). These results were in line with the result of Gou et al., (2010), who
reported more lodging observed at high plant population density as compared with
lower densities.
The highest grain yield 7566.2 kg ha-1 was recorded from 138/104 N/P2O5 which
was statically at par with 115/86N/P2O5 fertilizer rate. On the other hand, the
lowest grain yield of 6223.9 kg ha-1 was recorded from blended NPSZnB fertilizer
150 kg ha-1 +140 kg ha-1 urea top-dressed (Table 2). The grain yield advantage of
17.8% was obtained from 115/86N/P2O5 when compared to the blended NPSZnB
fertilizer. Reduction of grain yield might be due to nutritional imbalance and
deficiency of certain important plant growth elements at various important growth
stages and also due to reduced leaf area development resulting in lesser radiation
interception and, consequently, low efficiency in the conversion of solar radiation.
The higher doses of chemical fertilizers increased grain yield as nutrients are the
main driving force to produce a high yield of maize (Nivong et al., 2007).
On the other hand, the highest grain yield 7182 kg ha-1 was recorded from 75*20
cm (66,666 plant ha-1) which was not statistically significant different from 88,888
and 53,333 plant ha-1. While, the lowest grain yield 6542.4 kg ha-1 was recorded
from the lowest plant population density of 44,444 plants ha-1 (Table 2). By
planting 53,333 plants ha-1 there was a 9.1% grain yield increase over plant
population density of 44,444 plant ha-1. This shows that an increase in plant
population density increase grain yield because the plant per meter square area
increase, the number of cobs harvested increase. Thus, balanced growth and
development of plants need optimum plant population density because optimum
density enables plants efficient utilization of available nutrients, soil water and
better light interception coupled with other growth influencing factors. This
finding was in agreement with Farnham (2001) who reported that maize grain
yield increased as plant density increased from 59,000 to 89,000 plant ha-1.
[295]
Table 2. Over season and location main effect of fertilizer rate and plant population density on yield and yield components of maize at Jimma and Buno-Bedele zone during 2016-2018
cropping season
Over location and year

N/P2O5 Plant EPP Lodging % Grain yield (kg ha-1) AGB(ton ha-1) HI
(Kg ha-1) height(cm)
69/52 285.2 0.94a 53.8 6636.6c 13.70c 0.49b
92/69 286.7 0.93ab 52.7 7020.7b 14.22bc 0.50ab
115/86 287.3 0.95a 51.2 7331.3ab 14.80ab 0.50ab
138/104 290.9 0.97a 52.6 7566.2a 15.07a 0.51a
NPSZnB (150+140 kg ha-1 urea top
285.4 0.90b 52.8 6223.9d 14.75ab 0.42c
dressed )
LSD (0.05) 4.324 0.041 6.0201 345.84 0.61 0.015
CV (%) 5.42 17.20 28.40 18.91 17.02 12.20
F-test NS ** NS ** ** **
Population density
88888(75*15cm) 287.9 0.81d 57.6a 6961.6a 14.82a 0.47c
66666(75*20cm)
288.9 0.89c 53.3a 7182.2a 14.94a 0.49ab
53333(75*25cm)
286.7 0.99b 52.7a 7136.7a 14.54a 0.50a
44444 (75*30cm) 284.8 1.06a 46.9b 6542.4b 13.72b 0.48bc
LSD (0.05) 3.87 0.037 5.546 309.33 0.55 0.013
CV (%) 6.78 19.87 33.02 19.37 18.92 13.49
F-test NS ** ** ** ** **
LSD= Least significant difference; CV=Coefficient of variation; NS=Non significant; EPP=Ears per plant; HI= Harvest index; AGB= Above ground biomass; Values followed by the same
letter within a column are not significantly different at P< 0.05.
[296]
The highest above ground biomass yield 15.07 ton ha-1 was recorded from 138/104 N/P 2O5 kg ha-1
which was statically at par with 115/86N/P2O5 fertilizer rate. On the other hand, the lowest
above-ground biomass yield 14.70 ton ha-1 was obtained from the 69/52 N/P2O5 fertilizer (Table
2). The above-ground biomass yield advantage of 8.0% was obtained from 115/86 N/P2O5 when
compared to 69/52 N/P2O5 kg ha-1 fertilizer. The result showed that the above-ground
biomass was increased by increasing rates of NP fertilizer application due to the
higher number of ear per plant, plant height, and grain yield. An adequate supply
of nutrients to the crop helps in the synthesis of carbohydrates, which are required
for the formation of protoplasm, thus resulting in higher cell division and cell
elongation. Thus an increase in biomass yield might have been on account of
overall improvement in the vegetative growth of the plant due to the application of
NP fertilizer. Similar results were obtained by Makinde and Ayoola (2010) who
reported that conjunctive application of organic and inorganic fertilizers is
effective for the growth of maize and improving the yields.
On the other hand, the highest above ground biomass 19.94 ton ha-1was recorded
from 75*20 cm (66,666 plant ha-1) which was not statistically significant different
from 88,888 and 53,333 plant ha-1. While, the lowest above-ground biomass 13.72
ton ha-1was recorded from the lowest plant population density of 44,444 plants ha-
1
(Table 2). By planting 53,333 plants ha-1 there was 6% above ground biomass
increase over plant population density of 44,444 plant ha-1. This shows that an
increase in plant population density increase above ground biomass yield because
the plant per meter square area increase, the number of cobs harvested increase.
Biomass yield was decreased in wider spacing due to minimum plant height
occurred in this treatment and plants in the wider inter and intra row spacing
decreased the ability of plants for capturing resources which were reflected as
evident in their decreased biomass production. These results were in agreement
with Bullock et al. (1998) who reported that narrow row spacing made more
efficient use of available light and shaded the surface soil more completely during
the early part of the growing season while the soil is still moist and therefore,
narrow row spacing is more effective in producing biomass.
Numerically, the highest harvest index (0.51) was recorded from the application of
138/104 kg ha-1 N/P2O5 fertilizer. But its effect was not statistically significant
different from 92/69 and 115/86 kg ha-1 N/P2O5 fertilizer, while the lowest harvest
index (0.49) was recorded from lowest fertilizer rate (69/52 kg ha-1 N/P2O5
fertilizer ) (Table 2). As harvest index is the ratio of grain yield to total above-
ground biomass, the highest harvest index was recorded from a higher NP
fertilizer rate. Adequate supply of NP fertilizer is essential for optimizing
partitioning of dry matter between grain and other parts of the maize plant.
Optimum utilization of solar radiation, higher assimilate production, and its
conversion to starch results in higher biomass, grain yield leading to higher
harvest index. These findings were in line with results by Kumar and Puri (2001)
[297]
who observed that the highest harvest index was recorded with 90 kg N ha-1
compared to 45 kg N ha-1 and control.
On the other hand, the highest harvest index (0.50) was recorded from 53,333
plants ha-1 (75*25cm) which was not statistically significantly different from
66,666 plants ha-1 (75*20cm), while the lowest harvest index (0.47) was recorded
from 88,888 plants ha-1 (75*15cm) (Table 2). The harvest index obtained was in
the acceptable range of 0.4 - 0.6 for maize (Hue, 1995). The result showed that the
harvest index was decreased by increasing plant density due to the increased
competition in high densities for resources. The obtained results were in
agreement with the findings of Zamir et al. (2011) and Moraditochaee et al (2012)
who claimed that with increasing the plant population, the harvest index was
decreased.
The open market price (6 birr kg-1) for maize crop and the official prices of
chemical fertilizer (13.5 birr kg-1), urea (10 birr kg-1), and the cost of labor spent
on chemical fertilizer application and transport, seed planting and purchase were
used for analysis. The cost of application and transport for fertilizer was taken to
be 15 birr 100 kg-1. Grain yield was adjusted by 10% for management differences
to reflect the difference between the experimental yield and the yield that farmers
could expect from the same treatment (Getachew and Taye, 2005, CIMMYT,
1988).

select potentially profitable treatments. Dominance analysis led to the selection of
treatments ranked in increasing order of total variable costs (Table 3). For each
pair of ranked treatments, the percent marginal rate of return (MRR) was
calculated. The MRR (%) between any pair of un-dominated treatments was the
return per unit of investment in chemical fertilizer and plant population density. It
was calculated by dividing the change in net benefit to the change in variable
costs. 100% MRR means for every 1 birr invested in different cost of fertilizer and
maize seed, farmers can expect to recover 1 birr and obtain an additional1 birr
(CIMMYT, 1988).
[298]
Table 3. Partial budget analyses of NP fertilizer rates and plant population density on grain yield of late-maturing maize
variety at Jimma and Buno-bedele zone during 2016-2018 cropping season
N/P2O5 GY Adj.GY GFB TVC NB

(Kg ha-1) (kg ha-1) (kg ha-1) (ETB ha-1) (ETB ha-1) (ETB ha-1)
69/52 6636.6 5972.9 35837.64 2933.6 32904.04
92/69 7020.7 6318.6 37911.78 3885.3 34026.48
115/86 7331.3 6598.2 39589.02 4827.5 34761.52
138/104 7566.2 6809.6 40857.48 5807.3 35050.18
NPSZnB (150+140 kg ha-1
6223.9 5601.5 33609.06 3605 30004.06
urea top dressed )
Plant population density
88888(75*15cm) 6961.6 6265.4 37592.64 2216 35376.6
66666(75*20cm) 7182.2 6464.0 38783.88 1662 37121.9
53333(75*25cm) 7136.7 6423.0 38538.18 1317.6 37220.6
44444 (75*30cm) 6542.4 5888.2 35328.96 1108 34221.0
*GY= Grain yield; GFB = Gross field benefit; TCV = Total cost that varied; NB = Net benefit;
ETB = Ethiopian Birr; Price of chemical fertilizer = 13.5birr kg-1; Price of Urea = 10 birr kg-1; Wage rate = 30 Birr man-day-
1; Retail price of grain = 6 birr kg-1.
The highest net benefit (35,050.18ETB) was obtained from 138/104 kg ha-1
N/P2O5 fertilizer followed by a net benefit of (34,761.52 ETB) by application of
115/86 kg ha-1 N/P2O5 fertilizer with un acceptable MRR(%) less than 100%
(74.4% and 98.1% respectively). Whereas, the lowest net benefit (30,004.06 ETB)
was obtained from blended fertilizer application of NPSZnB 150 kg ha-1+140 kg
ha-1urea top dressed (Table 4). Due to application of 138/104 and 115/86 kg ha-1
N/P2O5 fertilizer, there was net benefit increase by 16.8% (5046.12 ETB) and 16%
(4757.46 ETB) respectively, when compared with blended fertilizer application of
NPSZnB 150 kg ha-1 +140 kg ha-1urea top dressed.
Table 4. Partial budget with dominance analysis and MRR (%) of NP fertilizer and plant population density of late-
maturing maize variety at Jimma and Buno-Bedele zone in 2016-2018 cropping season
N/P2O5 (Kg ha-1) GY Adj.GY GFB TVC NB Domi MRR (%)

(kg ha-1) (kg ha-1) (ETB ha-1) (ETB ha-1) (ETB ha-1) nance
69/52 6636.6 5972.9 35837.64 2933.6 32904.04 --- ---
NPSZnB (150 +140 kg
6223.9 5601.5 33609.06 3605.0 30004.06 D ---
ha-1urea top dressed )
92/69 7020.7 6318.6 37911.78 3885.3 34026.48 UD 117.9
115/86 7331.3 6598.2 39589.02 4827.5 34761.52 UD 98.1
138/104 7566.2 6809.6 40857.48 5807.3 35050.18 UD 74.7
Population Density
44444 (75*30cm) 6542.4 5888.2 35328.96 1108.0 34221.0 --- ---
53333(75*25cm) 7136.7 6423.0 38538.18 1317.6 37220.6 UD 1431.1
66666(75*20cm) 7182.2 6464.0 38783.88 1662.0 37121.9 D ---
88888(75*15cm ) 6961.6 6265.4 37592.64 2216.0 35376.6 D ---
D=Dominated treatment; UD= Un-dominated treatments; ETB = Ethiopian Birr; Price of chemical fertilizer = 13.5birr kg-1;
Price of Urea = 10 birr kg-1; Wage rate = 30 Birr man-day-1; Retail price of grain = 6 birr kg-1
[299]
TVC (ETB ha-1) NB (ETB ha-1) MRR (%)
40000 140
Net benifit (ETB ha-1)

35000 117.9 120
30000 98.1 100
MRR(%)
25000
20000 74.780
15000 60
10000 40
5000 20
0 0
Fertilizer rate

40000 1600
35000 1431.1 1400

30000 1200
MRR(%)
25000 1000
20000 800
15000 600
10000 400
5000 200
0 0
44444 53333 66666 88888
Plant population density
Figure 1. Total variable cost, net benefit and MRR % of fertilizer rate and plant population density rate of BH661 maize
varieties at Jimma and Buno-Bedele zone in 2016-2018 cropping season
On the other hand, the maximum net benefit (37,220.6 ETB) was obtained from a
plant population density of 53,333 plant ha-1 (75*25cm) with acceptable MRR of
1431.1% (Table 4). Whereas, the minimum net benefit (34,221.0 ETB) was
obtained from a plant population density of 44,444 plant ha-1 (75*30cm). Due to
sowing plant population of 53,333 plant ha-1 (75*25cm), there was net benefit
increase by 8.8% (2,999.6 ETB), when compared with the plant population
density of 44,444 plant ha-1 (75*30cm), the recommended before. The MRR%
was decreased as the NP fertilizer rate increased to 138/104 kg ha-1 N/P2O5 and
the highest rate of return was recorded from 53,333 plants ha-1 (75*25cm) (Figure
1).
Figure 1 Total variable cost, net benefit and MRR % of fertilizer rate and plant
population density rate of BH661 maize varieties at Jimma and Buno-Bedele zone
in the 2016-2018 cropping season.
[300]
grain, which represented a price variation of 15% (Table 5). The price changes are
sensitive under market conditions prevailing at Jimma and Buno-Bedele zones
which were below the minimum acceptable MRR of 100% (87.17%) for
application 92/69 kg ha-1 N/P2O5 fertilizer; whereas the price changes are realistic
which were above the minimum acceptable MRR (1057.78%) for a plant
population of 53,333 plant ha-1(75*25cm) (Table 5).
Table 5. Sensitivity analysis of maize production based on a 15% rise in total cost and maize
N/P2O5 Fertilizer TVC (ETB ha-1) NB (ETB ha-1) Raised cost Raised benefit MRR (%)
69/52 3373.64 27968.43 ---- ---- ----
92/69 4468.095 28922.51 1094.46 954.07 87.17
Population density
44444(75*30cm ) 1274.2 29087.8 ---- ---- ----
53333(75*25cm) 1515.24 31637.5 241.04 2549.68 1057.78
*TCV = Total cost that varied; NB = Net benefit; ETB = Ethiopian Birr; MRR= Marginal rate of
return; Price of chemical fertilizer = 13.5 birr kg-1; Price of Urea = 10 birr kg-1, Wage rate = 30 Birr man-day-1; Retail price
of grain = 6 birr kg-1.

Declining soil fertility aggravated the challenge of agriculture to meet the world’s
increasing demand for food sustainability and the variations in the morphology of
the maize crop lead to different planting density to reach maximum yield. Because
of this, the study was conducted to investigate the response of maize hybrid
BH661 to different rates of NP fertilizer and plant population density at Jimma
and Buno-Bedele zones, southwestern Ethiopia.
Accordingly, rigorous research efforts were made on farmer's fields of Jimma

zone(Kersa and Omonada woredas) and Buno-Bedele zone in the vicinity of the
Jimma research center for three cropping seasons (2016-2018). The results
revealed that individual chemical fertilizer and plant population density improved
grain yield and above ground biomass yield of the maize. The improvement was
mainly due to the availability of nutrients from the chemical fertilizer for plant
development up to cob formation. The application of chemical fertilizer increased
grain yield mainly due to better grain development.
Summary of results from across seasons and sites indicated different fertilizer
doses and plant population density significantly affect grain and above-ground
biomass yield of the maize. From this experimental result, it can be concluded that
the maximum grain yield and biomass yield per hectare was recorded with
[301]
138/104 kg ha-1 N/P2O5 fertilizer with the highest net benefit of 35,050.18 ETB
with MRR 74.7%. On the other hand, the highest grain yield and biomass yield
per hectare was recorded from plant population density 66,666 (75*20cm).
However, this treatment grain yield was statistically not different from plant
population density of 53,333 plant ha-1(75*25cm) which gave the highest net
benefit of 37,220.6 ETB and highest MRR 1431.1%. This result contradicted the
previous recommendation of a plant population density of 44,444 plant ha-1
(75*30cm). Therefore, it is advisable for farmers in the study area and adjacent
woredas with similar agro-ecologies, a plant population density of 53,333 plant ha-
1
(75 x 25cm a plant hill-1 or 75 x 50 cm two plants hill-1) in complement with
N/P2O5 fertilizer rate of 92/69 kg ha-1 can be recommended.
References
Adeniyan, ON and Ojeniyi SO. 2005. Effect of poultry manure, NPK 15-15-15 and combination of
their reduced levels on maize growth and soil chemical properties. Nigerian Journal of Soil
Science, 15: 34-41.
Batjes, NH. 1995. A global data set of soil pH properties. Technical Paper 27, International Soil
Bavec F. and Bavec M. 2002. Effect of plant population on leaf area index, cob characteristics and
grain yield of early maturing maize cultivars (FAO 100-400). European Journal of
Agronomy, 16:151-159.
Brady, N and Weil RR. 2008. Nature and Properties of Soils (14th ed.). Prentice Hall, Upper
Saddle River. 992p.
Bullock, DG, Nielsen RL and Nyquist WE. 1988. A growth analysis comparison of corn grown in
conventional and equidistant plant spacing. Crop Science, 28:254-258.
CIMMYT. 1988. From Agronomic Data to Farmer Recommendations: An Economics
Training Manual. Completely revised edition. Mexico, D.F.79p
Cox, WJ. 1996. Whole plant physiology and yield response of maize to plant density. Agronomy
Journal, 88:489-496.
Farnham, DE. 2001. Row spacing, plant density and hybrid effects on corn grain yield and
moisture. Agron. J., 93: 1049-1053.
Getachew, A and Taye B. 2005. On-farm integrated soil fertility management in wheat on
nitisols of central Ethiopian highlands. Ethiopian Journal of Natural Resources, 7: 141-155.
Gou, L, Huang JJ, Sun R, Ding ZS, Dong ZQ and Zhao M. 2010. Variation characteristic of stalk
penetration strength of maize with different density-tolerance varieties. Trans. Chinese Soc.
Agric. Eng. 26, 156–162.
Hashemi-Dezfouli, A and Herbert SJ. 1992. Intensifying plant density response of corn with
artificial shade. Agron. J., 84, 547-551.
Hue, NV. 1995. Sewage sludge. pp.199-247. In: Jack E. Rechcigl (eds). Soil amendments and
environmental quality. Lewis Publisher, London.
Karlen, DL and Camp CR. 1985. Row Spacing, Plant Population, and Water Management Effects
on Corn in the Atlantic Coastal Plain 1. Agronomy Journal, 77:393-398.
Kumar, P and Puri UK. 2001. Effect of nitrogen and farmyard manure application on maize
varieties. Indian Journal of Agronomy, 46: 255-259.
[302]
Makinde, EA and Ayoola OT. 2010. Growth, yield and NPK uptake by maize with complementary
organic and inorganic fertilizers, African Journal of Food, Agriculture, Nutrition and
Development, 10:2203-2217.
McLachlan, SM, Tollenaar M, Swanton CJ and SF Weise. 1993. Effect of corn-induced shading on
dry matter accumulation, distribution, and architecture of redroot pigweed
(Amaranthusretroflexus). Weed Science, 41:568-573.
Moraditochaee, M, Motamed MK, Azarpour E, Khosravi Danesh R. 2012. Effects of nitrogen
fertilizer and plant density management in corn farming. ARPN J of Agric and Biol Sci.7:
133-137.
Nivong, S, Tasnee A and Russell Y. 2007. Nitrogen fertilizer response of maize on some important
soils from DSSAT software prediction. Natural Science, 41: 21-27.
Porter, PM, Hicks DR, Lueschen WE, Ford JH, Warnes DD and Hoverstad TR. 1997. Corn
response to row width and plant population in the northern Corn Belt. Journal of Production
Agriculture, 10: 293-300.
Sangoi, L and Salvador RJ. 1998. Maize susceptibility to drought at flowering: a new approach to
overcome the problem. Ciência Rural, 28:699-706.
Shah, STH, Zamir MSI, Waseem M, Ali A, Tahir M and Khalid WB. 2009. Growth
and yield response of maize to organic and inorganic sources of nitrogen. Pak. J. life soc.sci.,
7:108-111.
Singh, D and Nepalia V. 2009. Influence of integrated nutrient management on QPM productivity
and soil of South Rajasthan. Indian J. of Agr. Sci., 79: 1020-22.
Song, YH, Yukui R, Guta B and Jincai L. 2016. Morphological characteristics of maize canopy
development as affected by increased plant density. PLoS One 11:e0154084.
Tekalign, M. and Haque I. 1991. Phosphorus status of some Ethiopian soils, II. Forms and
distribution of inorganic phosphates and their relation to available phosphorus. Trop Agric.,
68: 2-8.
Westgate, ME, Forcella F, Reicosky DD and Somsen J. 1997. Rapid canopy closure for maize
production in the northern US corn belt: radiation-use efficiency and grain yield. Field Crops
Res.Amsterdam, 49:249-258
Xue, J, Gou L, Zhao Y, Yao M, Yao H, Tian J. 2016. Effects of light intensity within the canopy
on maize lodging. Field Crop Res. 188, 133–114.
Yuan, Z, H Li, YD and Chen HS. 2002. Dynamic model and lodging resistance analysis of maize
stem. J. Maize Sci. 3, 74–75.
Zamir MSI, Ahmad AH, Javeed HMR and Latif T. 2011. Growth and yield behavior of two maize
hybrids (Zea mays L.) towards different plant spacing. Cercetări Agronomice în Moldova.
14: 33-40.
[303]
[304]
Evaluation of QPM Maize Variety under Integrated
Uses of Different Fertilizer Sources in Jimma,
South-Western Ethiopia
Sisay Gurmu*, Muhidin Biya and Eshetu Yadete
E-mail of the corresponding author: sis.sis1835@gmail.com
Abstract
Studies on the combined use of organic and inorganic fertilizers for maize are lacking
at Jimma conditions. Therefore, a field experiment was conducted to determine the
effect of integrated uses of organic with inorganic fertilizers sources on the
productivity of Quality protein maize (QPM) variety BHQPY545 at Kersa, Omonada
woreda and Jimma on station, Southwestern Ethiopia during 2016 and 2017 main
cropping seasons and one-year validation at Omonada woreda. The experiment
involved factorial combinations of inorganic and organic fertilizer with negative
control, 150 kg ha-1 NPSZnB+140 kg ha-1urea top-dressed, recommended NP2O5
(92/69 kg ha-1), 100% compost, 25% compost + 75% recommended NP, 50% compost
+ 50% recommended NP, 75% compost + 25% recommended NP and 100% compost
+ 100% recommended NP fertilizer which was laid out in RCBD design. The compost
was added based on N equivalency of the recommended fertilizer rate. Data on yield
and yield components of maize were subjected to ANOVA using SAS version 9.3. The
over location and year highest grain yield (8443.1 kg ha -1) and above-ground biomass
(21.52 ton ha-1) were obtained from 100% compost + 100% recommended NP
fertilizer. Partial budget analysis revealed 100% compost + 100% recommended NP
fertilizer realized the maximum net return (39,717.7 ETB) with marginal rate of
return (94%). From farmers' decision point and current on-farm input availability
and technical feasibility and economic analysis, 92/69 kg ha-1 N/P205 fertilizer
application taken as optimal fertilizer application and recommended for production of
BHQPY545 in Jimma area and other similar humid agro-ecologies of southwestern
Ethiopia.
Keywords: Grain yield, inorganic fertilizer, organic fertilizer, recommended

fertilizer rate, soil health.
Introduction
Maize is the primary staple crop in Ethiopia and plays an important role in the
livelihood of the people of Ethiopia. Its availability and abundance determines the
level of welfare and food security in the country. In Ethiopia, future increases in
maize production to meet domestic demand will have to rely on improvements in
yield per hectare rather than on the expansion of maize production area. Enhanced
maize productivity can be achieved by increased use of modern production
techniques such as the adoption of hybrid maize varieties, the use of chemicals,
and integrated nutrient management. Small-scale maize production plays a major
[305]
role in Ethiopia's maize economy and the adoption of hybrid technology by small-
scale farmers would have the potential to address sustainability and supply issues.
Most soils contain an abundance of elements essential for plant development, but
the majority of these elements are rarely available for plant use due to nutrient
loss. Nutrient depletion and soil degradation have become a serious threat to
agricultural productivity in Ethiopia. These soils suffered multi-nutrient
deficiencies, the application of mineral fertilizers has become mandatory to
increase crop yields in such soils (Adeniyan and Ojeniyi, 2005).
The dissemination and adoption of QPM is lagging behind normal maize in Sub-
Saharan Africa where it is needed more (Aman et al., 2016). Twumasi-Afriyieet
al. (2016) reported that an estimated area of one million ha of land in Sub Saharan
Africa (SSA) was under QPM production in 2015. Research on QPM is of recent
history in Ethiopia. The work was started by testing introduced CIMMYT
(International Maize and Wheat Improvement Center) QPM pools and populations
in 1980 (Leta et al., 2001). From experimental variety trials conducted at some
locations in Ethiopia, for instance, some QPM entries yielded 9.5 ton ha-1, which
is an advantage of 20% over the best local check (Gemechu et al., 2016).
Among plant nutrients, nitrogen is a vitally important, a major yield determining

nutrient and its availability in sufficient quantity throughout the growing season is
essential for optimum maize growth (Kogbe and Adediran, 2003). It is a
component of protein, nucleic acids and other compounds essential for plant
growth process (Onasanyaet al., 2009). Whereas phosphorus is the second most
important nutrient element (after nitrogen) limiting agricultural
production (Kogbe and Adediran, 2003). It is used for growth, utilization of sugar
and starch, photosynthesis, the metabolic process which leads to a higher yield of
the crop (Ayubet al., 2002).
Compost is one of the organic fertilizers and it is an alternative source of plant

nutrients (Vanlauweet al., 2012; Ngwiraet al., 2013). Application of compost
improves soil fertility parameters, such as alleviate acidification, benefits better
microbial activity, soil aeration, increases soil organic matter, increases CEC, P
availability and sustainable increase in crop yields (Diacono and Montemurro,
2010). The use of compost and sometimes in combination with inorganic
fertilizers gave maximum grain yields of QPM (Balaiet al. 2011).
Currently, the lack of a balanced and integrated application of nutrients reduced

the yield potential of QPM and other maize varieties in most maize producing
areas of Ethiopia. The productivity of maize is low as a result of continuous
cropping, inadequate use of fertilizer inputs, very low or lack of the use of organic
manure neither alone nor in combinations with mineral fertilizers. Since the use of
inorganic fertilizer in maize production is costly and its effect is short term, there
[306]
is another option to use organic fertilizer in an integrated form. For this, site-
specific testing of integrated nutrient management (INM) is becoming critical and
sorting out of optimum nutrients for maize production that benefit farmers to get
better nutrition and economic advantage, and sustain soil health. Therefore, site-
specific INM investigation based on growth, yield and yield-related parameters of
QPM variety under agro-climatic conditions of Jimma, Southwestern Ethiopia is
needed. Even though the cultivar was under production, the cultivar performance
under the use of NPS fertilizer and compost has not been tested. Hence, for better
dissemination and adoption of QPM hybrid, there is a need to understand it
performance with various agronomic management practices, of which nutrient
management is vital in influencing the growth and yield of the crop. Therefore, the
objective of this paper was to evaluate the integrated uses of organic with
inorganic fertilizer sources on the productivity of QPM maize variety and
economic feasibility.
Description of the Experimental Site

The experiment was conducted at Jimma zone (Kersa, Omonada woreda and
Jimma on station), in the Oromiya region at the southwestern part of Ethiopia,
during the main cropping season of 2016 and 2017. Although, at the 2018 main
cropping season a validation experiment was done at Omonada woreda. The
average maximum and minimum temperature is 9ºC and 28ºC respectively and
reliably receives good rains up to 1747 mm per annum. The farming system of the
study sites is coffee and cereal crops dominated with coffee, maize, teff and
sorghum also has a warm and cold climate, also convenient topography which is
very suitable for all agricultural practices. It was situated in the tepid to cool
humid-mid highlands of southwestern Ethiopia. The soil type of the experimental
area was Eutric nitisols (reddish-brown).
Soil and Compost Physico-chemical Properties

The soil of the experimental field and compost were characterized for selected
Physico-chemical properties before the application of the treatments (Table 1).
The average soil pH of the trial sites ranges from 4.97 to 5.12, which was strongly
acidic (Batjes, 1995) and ideal for the production of most field crops. The pH of
the soil affects maize growth by suppressing the root development and reducing
the availability of macronutrients to plants especially phosphorus (Brady and
Weil, 2008). The soil total N ranges from 0.14 to 0.22% and OC from 1.47 to
2.17% were found a medium rate for crop growth and development for both
nutrients (Berhanu, 1980). The Bray II extractable available P for Omonada
woreda was 4.38 mg kg-1, which is below the critical level (8 mg kg-1) for most
[307]
crop plants whereas, Kersa and Jimma on station sites ranged from 18.16 to 30.14
mg kg-1 which is high in content as described by Tekalign and Haque (1991).
Table 2. Selected physico-chemical properties of the soil of the experimental sites and compost before planting
Characters Soil Compost

Kersa Omonada Jimma on station
pH(1:2.5) 4.97 5.12 5.11 9.05
av P(ppm) 30.14 4.38 18.16 508.66
TN (%) 0.22 0.16 0.14 0.60
OC (%) 2.17 1.94 1.47 7.12
OM (%) 3.73 3.34 2.53 12.27
C:N ratio 9.92 12.09 10.33 11.80
Where pH= hydrogen power, OC=organic carbon, TN=Total Nitrogen, Av. P=Available phosphorous,
OM=Organic matter. Values are the means of duplicated samples.
The chemical compositions of the compost utilized as an organic source of soil

fertility amendment in this study are presented in Table 1. Accordingly, the mean
OC and total N contents of the compost was 7.12% and 0.60% respectively, with a
resultant narrow C: N ratio of 11.80. It indicates the prepared compost was well
decomposed to the level of average soil organic matter. The C: N ratio of about 30
is considered N neutral, lower ratios will release N and act as N fertilizers and
higher ratios will immobilize N as a microbial breakdown of the carbon
component. The C: N ratio of compost should drop below 20% before application
to the soil (Brady and Weil, 2002) to have an expected impact from the
application of compost. The pH of compost (9.05) was strongly alkaline and it is
capable of ameliorating the acidic content of the soil (Onwudiwe et al., 2014).
Treatments and Experimental Design

The experimental field was plowed and prepared following the conventional
tillage practice before planting at all experimental locations. Two maize seeds
were planted per hill and then thinning was done after the good establishment of
seedlings to maintain a single healthy plant per hill. Eight inorganic and organic
fertilizer with negative Control (no input), blended NPSZnB 150 kg ha-1 + 140 kg
ha-1urea top-dressed, recommended NP (92/69 kg ha-1) (as check), 100% Compost
based on N equivalency of recommended fertilizer rate (RFR), 25% compost
based on N equivalency of RFR + 75% Recommended NP, 50% compost based
on N equivalency of RFR + 50% Recommended NP, 75% compost based on N
equivalency of RFR + 25% Recommended NP and 100% compost based on N
equivalency of RFR + 100% recommended NP fertilizer.
Medium maturing maize variety BHQPY545 was used for the study. It was
released by Bako Agricultural Research Center through the National Maize
Research Program in 2008. It performs well in agro-ecology of 1000-2000 m.a.s.l
with rainfall of 1000-1200 mm. It can give 8.0-9.5 and 5.5-6.5 t ha-1 grain yields
under on-station and on-farm experiments, respectively. It was moderately tolerant
[308]
to rust, blight and gray leaf spot with a maturity date of 138 and 25 kg ha-1 seed
rate. The Randomized Complete Block Design (RCBD) with three replications
and plot size 22.95m2 (4.5 m x 5.1 m) for each treatment was used.
Nitrogen fertilizer rates were applied in a split in such a way that half during
planting and half at the knee height growth stage to increase the nitrogen use
efficiency. All other agronomic practices were applied uniformly to all
experimental plots as per their respective recommendations for maize in the study
area.
The data collected were yield, yield related and other agronomic data were
collected. Analysis of variance (ANOVA) for all collected data was computed
using SAS version 9.3 statistical software. Whenever the ANOVA results showed
the significant differences between sources of variation, the means were separated
using the least significant difference (LSD). The homogeneity test was done as
suggested by Gomez and Gomez, (1984).
Partial budget analysis

The partial budget analysis was performed to investigate the economic feasibility
of the treatments and assess the costs and benefits associated with different
chemical fertilizers and compost levels. The partial budget technique as described
by CIMMYT (1988) was applied. The partial budget analysis was done using the
prevailing market prices for inputs at planting and outputs, at the time the crop
was harvested. All costs and benefits were calculated on a hectare basis in
Ethiopian Birr (ETB). The inputs and/or concepts used in the partial budget
analysis were the mean grain yield of each treatment, the gross field benefit (GFB)
ha-1(the
product of field price and the mean yield for each treatment), the field price of NP
fertilizer (including blended fertilizer) and urea (the nutrient cost plus the cost of
transportation from the point of sale to the farm), cost of labor spent on compost
preparation, transportation and incorporation, the total costs that varied (TVC)
which included the sum of field costs of fertilizers and their application.
The net benefit (NB) was calculated as the difference between the GFB and the
TVC. The marginal rate of return (MRR %) was also calculated. To obtain an
estimate of these returns the MRR (%) was calculated as changes in NB divided
by changes in cost. Thus, a minimum acceptable MRR of 100% was used
indicating, for every one ETB expended there is a return of one ETB for a given
variable input (CIMMYT, 1988), which is suggested to be realistic. This
enables’ to make farmer recommendations from marginal analysis. The
dominance analysis procedure as detailed in CIMMYT (1988) was used to select
potentially profitable treatments from the range that was tested. Sensitivity
analysis for different interventions was also carried out to test the recommendation
made for its ability to withstand price changes. Sensitivity analysis simply implied
[309]
redoing marginal analysis with the alternative prices. Through sensitivity analysis,
the maximum acceptable field price of an input was calculated with the minimum
rate of return as described by Shah et al. (2009).

The combined analysis over locations and seasons showed fertilizer types and
rates were highly significant (P <0.01) influenced plant height, ears per plant,
lodging percentage, grain yield and above-ground biomass. But the harvest index
was not significantly (P >0.05) affected by the fertilizer types and rates (Table 2).
Numerically the longest plant height (268.8cm) was recorded from 100%
recommended NP+100% compost, while the shortest plant height (216.5cm) was
recorded from control (Table 2). The plant height was increased by 24.2% at full
recommended NP fertilizer and full compost as compared to control. The increase
in plant height with an increasing rate of NP fertilizer and compost could be due to
their synergistic effects. Nitrogen promotes the formation of chlorophyll which in
turn resulted in higher photosynthetic activity, vigorous vegetative growth, and
taller plants and P is required for shoot and root development where metabolism is
high and cell division is rapid (Rao et al., 2001). Also, the compost acted as the
storehouse of different plant nutrients, reduce P fixation, and improve CEC,
aeration, root penetration and water storage capacity of the soil (Rahman et al.,
2012). These results were in line with the findings of Adekayode and Ogunkoya
(2010) who explained that there was a very high significant difference in maize
plant height in plots treated with high fertilizers compared with nil application.
The maximum number of ears per plant (1.63) was recorded from 100%compost
+100% recommended NP fertilizer, which was statistically at par with all
treatments except the control and 100% compost only, while the minimum number
of ears per plant (1.02) was recorded from the control (Table 2). The 100%
compost +100% recommended NP fertilizer increased 59.8% number of ears per
plant over control. The number of ear per plant was determined by the prolific
ability of the BHQPY545 maize variety as described by Adefris et al. (2015) and
the growth behavior of the crop which is dependent upon management practices
and edaphic and climatic factors. The improvement of the soil conditions or
enrichment with nutrients and organic matter due to soil-added materials might be
responsible for better cob production under plots treated with NP fertilizer and
compost. This is considered as the main yield component as it defines the yield
potential of a crop. The improvement of the soil conditions or enrichment with
nutrients and organic matter due to soil-added materials might be responsible for
better cob production under plots treated with NP fertilizer and compost. These
results were in line with the findings of Dagne (2016) who indicated that the
application of blended fertilizer increased the number of cobs harvested compared
to the control plot. Also, Amanullah et al. (2015) found that the application of
[310]
compost was most beneficial in terms of higher yield and yield components of
maize over the control.
The highest lodging percent (21.84%) was recorded from the control which was
not statistically significantly different from 100% compost, while the lowest
lodging percent (7.21%) was recorded from blended NPSZnB 150 kg ha-1 + 140
kg ha-1 urea top-dressed fertilizer which was statistically at par with recommended
NP fertilizer, 25% compost + 75% recommended NP, 50% compost + 50%
recommended NP and 100% compost+100% recommended NP (Table 2). The
application of blended NPSZnB fertilizer decreased 202.9% lodging percent over
the control. Nevertheless, increasing NP and compost fertilizer linearly increased
resistance of plants for lodging. Moreover, the number of data points with higher
lodging percentage was more for treatments that received lower fertilizer rates
than treatments with relatively higher fertilizer rates. Brady and Weil (2000)
reported that plants deficient in nutrients develop thin and spindly stems. Such
stems could be susceptible to lodging by the wind. Moreover, N deficient plants
have poor development of the root system, which reduces their anchorage
capacity.
Numerically the highest grain yield 8443.1 kg ha-1 was recorded from 100%
compost + 100% recommended NP fertilizer at which it was statically at par with
recommended N/P2O5fertilizer (92/69 kg ha-1) and 25% compost + 75%
recommended NP fertilizer. On the other hand, the lowest grain yield 3435.8 kg
ha-1 was obtained from the control (Table 2). The grain yield advantage of 145.7%
and 118.6% were recorded from 100% compost + 100% recommended NP
fertilizer and recommended N/P2O5fertilizer (92/69 kg ha-1) respectively when
compared to the negative control. The reduction of grain yield in unfertilized plots
might be due to nutritional imbalance and deficiency of certain important plant
growth elements at various important growth stages. The increase in grain yield
could be attributed to the beneficial influence of yield contributing characters and
increase of the nutrients in the soil and modification of soil environments that
resulted in better vegetative growth which in turn enabled the crop to produce
greater photo-assimilate.
N has synergistic effects on growth and yield attributes resulting in greater

translocation of photosynthates from source to sink, beneficial effect on the
physiological process, plant metabolism and growth (Yayock et al., 1988) and
thereby leading to higher grain yield. The P supply is particularly important for
stimulating early root formation and growth, functions in plant macromolecular
structures as a component of nucleic acids and phospholipids (Marschner 2012).
Mugwe et al. (2007) reported the higher grain yield of maize was recorded in
treatments of compost either alone or in combination with mineral fertilizer when
compared to the control and compost is responsible in improving soil physical,
[311]
chemical and microbial conditions in addition to giving different macro and
micronutrient to the plant. These results were in line with the findings of
N’Dayegamiye et al. (2010) who reported that the application of compost with
120 kg N ha-1 led to higher maize grain yields. The highest grain yield was
recorded with the application of FYM 10 t ha-1 + 100% RDF in QPM maize
hybrid, which was significantly superior to the other treatments as revealed by
Ravi et al. (2012).
Table 2. Over seasons and locations main effect of organic and inorganic fertilizer on yield and yield components of QPM
maize variety during 2016 and 2017 cropping season
Over location and year

Fertilizer types and rates Plant EPP Lodging Grain yield AGB HI
height(cm) % (kg ha-1) (ton ha-1)
Negative Control 216.5d 1.02c 21.84a 3435.8d 9.53d 0.38
150 kg ha-1 NPSZnB + 140
244.7bc 1.57a 7.21c 7284.2b 17.86b 0.48
kg ha-1 Urea top dressed
Rec NP (92/69 kg ha ) -1 236.2c 1.55a 8.48bc 7509.2ab 18.82ab 0.51
100% Compost 237.61bc 1.25b 20.51a 4991.0c 13.77c 0.39
25% compost +75% Rec NP 250.8b 1.52a 9.19bc 7383.8ab 18.37ab 0.41
50% compost + 50% RecNP 250.8b 1.55a 11.32bc 7031.9b 19.47ab 0.38
75% compost +25% Rec NP 248.4bc 1.53a 13.20b 7057.2b 18.60ab 0.40
100%compost+100%Rec NP 268.8a 1.63a 9.49bc 8443.1a 21.52a 0.41
LSD (0.05) 13.33 0.159 4.978 1147.6 3.198 0.140
CV (%) 8.28 16.58 35.67 19.21 18.13 20.62
F-test ** ** ** ** ** NS
*LSD= Least significant difference; CV=Coefficient of variation; Rec=Recommended; EPP= Ears per plant; HI= Harvest
index; AGB= Above ground biomass; Values followed by the same letter within a column are not significantly different at
P< 0.05. Compost was applied based on N-equivalence of recommended fertilizer rate (92 N kg ha-1)
The highest above ground biomass 21.52 ton ha-1 was recorded from 100% compost +
100% recommended NP fertilizer at which it was statically at par with recommended
N/P2O5fertilizer (92/69 kg ha-1), 25% compost + 75% recommended NP fertilizer,
75% compost + 25% recommended NP and 50% compost + 50% recommended
NP fertilizer. On the other hand, the lowest above-ground biomass 9.53 ton ha-1 was obtained
from the control (Table 2). The above-ground biomass yield advantage of 125.8% and 97.5% were
recorded from 100% compost + 100% recommended NP fertilizer and recommended
NP fertilizer respectively when compared to the negative control.
The result showed that the above-ground biomass was increased by increasing
rates of NP fertilizer and compost application due to a higher number of ears per
plant, plant height and grain yield. An adequate supply of nutrients to the crop
helps in the synthesis of carbohydrates, which are required for the formation of
protoplasm, thus resulting in higher cell division and cell elongation. Thus an
increase in biomass yield might have been on account of overall improvement in
the vegetative growth of the plant due to the application of NP fertilizer and in
combination with compost. The above-ground biomass improvement with
compost than those obtained with the control was probably attributed to the
[312]
improvement of the physical conditions and biological activity of the soil (Chang
et al., 1990). Similar results were obtained by Makinde and Ayoola (2010) who
reported that conjunctive application of organic and inorganic fertilizers is
effective for the growth of maize and improving the yields. The harvest index was
not significantly (P >0.05) affected by fertilizer types and rates over locations and
years (Table 2).
The open market price (6 birr kg-1) for maize crop and the official prices of
chemical fertilizer (13.5 birr kg-1), urea (10 birr kg-1) and the cost of labor spent on
compost preparation, transport and incorporation were used for analysis. The cost
of application and transport for fertilizer was taken to be 15 birr 100 kg-1. Grain
yield was adjusted by 10% for management difference to reflect the difference
between the experimental yield and the yield that farmers could expect
from the same treatment (Getachew and Taye, 2005, CIMMYT, 1988).

select potentially profitable treatments. Dominance analysis led to the selection of
treatments ranked in increasing order of total variable costs (Table 3). For each
pair of ranked treatments, the percent marginal rate of return (MRR) was
calculated. The MRR (%) between any pair of un-dominated treatments was the
return per unit of investment in chemical fertilizer and plant population density. It
was calculated by dividing the change in net benefit to the change in variable
costs. 100% MRR means for every 1 birr invested in different cost of fertilizer and
maize seed, farmers can expect to recover 1 birr and obtain an additional 1 birr
(CIMMYT, 1988).
Table 3. Partial budgets with dominance analysis of organic and inorganic fertilizer maize grain yield at Jimma zone in
2016 and 2017 cropping season
GY Adj.GY GFB TVC NB

Fertilizer types and rates (kg ha-1) (kg ha-1) (ETB ha-1) (ETB ha-1) (ETB ha-1) Dominance
egative Control 3435.8 3092.2 18553.3 0 18553.3 ---
% Compost (9.2 t ha-1 ) 4991.0 4491.9 26951.4 2206.0 24745.4 UD
75% compost + 25% Rec NP 7057.2 6351.5 38108.9 2887.3 35221.6 UD
Rec NP (92/69 kg ha-1) 7509.2 6758.3 40549.7 3273.0 37276.7 UD
150 kg ha-1 NPSZnB +
7284.2 6555.8 39334.7 3425.0 35909.7 D
140 kg ha-1 Urea top dressed
50% compost + 50% Rec NP 7031.9 6328.7 37972.3 3447.5 34524.8 D
25% compost + 75% Rec NP 7383.8 6645.4 39872.5 4008.3 35864.3 D
100% compost+100% Rec NP 8443.1 7598.8 45592.7 5875.0 39717.7 UD
D=Dominated treatment; UD= Un-dominated treatments; ETB = Ethiopian Birr; Price of chemical fertilizer = 13.5 birr kg -1;
The highest net benefit (39,717.7 ETB ha-1) was obtained from 100% compost +
100% recommended NP fertilizer with unacceptable MRR 94% followed by a net
benefit of (37,276.7 ETB ha-1) with acceptable MRR 533% by application of
[313]
recommended NP fertilizer (92/69 kg ha-1) (Table 4). The highest MRR 1538%
was obtained from the application of 75% compost + 25% recommended NP
fertilizer and it dramatically decreased when higher nutrients applied from both
compost and chemical fertilizer (Figure 1). Thus, those treatments above the
minimum acceptable marginal rate of return (100%), could be recommended as
alternative sources for users (CIMMYT, 1988).On the other hand, the lowest net
benefit (18,553.3 ETB) was obtained from control (no input). Due to the
application of recommended NP fertilizer (92/69 kg ha-1), there was a net benefit
increase by 114.1% (21,164.4 ETB) when compared with a negative control.
Table 4. Partial budget with estimated marginal rate of return (%) of organic and inorganic fertilizer on maize grain yield at
Jimma zone during 2016 and 2017 cropping season
TVC NB Raised Raised

Treatments
(ETB ha-1) (ETB ha-1) cost benefit MRR (%)
Negative Control 0 18553.3 --- --- ---
100% Compost (9.2 t ha-1 ) 2206 24745.4 2206 6192.08 281
75% compost + 25% Rec NP 2887 35221.6 681.25 10476.23 1538
Rec NP (92/69 kg ha-1) 3273 37276.7 385.75 2055.1 533
100% compost + 100% Rec NP 5875 39717.7 2602 2441.06 94
*GY= Grain yield; GFB = Gross field benefit; TCV = Total cost that varied; NB = Net benefit;Rec=Recommended;
D=Dominated treatment; UD= Un-dominated treatments; ETB = Ethiopian Birr; Price of chemical fertilizer = 13.5 birr kg -1;
50000 2000
40000
MRR(%)
1538 1500
30000 1000
20000
10000 533 500
281 94 0
0
100% 75% Rec NP 100%
Compost compost (92/69 kg compost
+ 25% ha-1) + 100%
Rec NP Rec NP
Fertilizer rate and types
Figure 1. Total variable cost, net benefit and MRR % of fertilizer rate and types of QPM maize variety at Jimma zone in
2016 and 2017 cropping season
grain, which represented a price variation of 15%. The price changes are realistic
under market conditions prevailing at Jimma area which were above the minimum
[314]
acceptable MRR of 100% for application of 100% Compost (9.2 t ha-1), 75%
compost + 25% recommended NP fertilizer and recommended NP fertilizer (92/69
kg ha-1).
Table 5. Sensitivity analysis of maize production based on a 15% rise in total cost and maize
TVC NB
Treatments Raised cost Raised benefit MRR (%)
(ETB ha-1) (ETB ha-1)
Negative Control --- 15770.3 --- --- ---
100% Compost (9.2 t ha-1 ) 2536.9 21033.6 2536.9 5263.3 207.5
75% compost + 25% Rec NP 3320.3 29938.4 783.4 8904.8 1136.6
Rec NP (92/69 kg ha-1) 3764 31685.2 443.6 1746.8 393.8
D=Dominated treatment; UD= Un-dominated treatments; Rec=Recommended; ETB = Ethiopian Birr; Price of chemical
fertilizer = 13.5 birr kg-1; Price of Urea = 10 birr kg-1; Wage rate = 30 Birr man-day-1; Retail price of grain = 6 birr kg-1
Validation of QPM Maize variety under Integrated Uses of Different Fertilizer Types in Jimma Zone, southwestern Ethiopia
After 2 years experiment of QPM maize variety evaluation under integrated uses
of different fertilizer types, validation of the experiment was done at Jimma zone
Omonada woreda on four on-farm sites for one year at 2018 main cropping season
with objective to determine one optimum integrated use of fertilizer where
maximum yield potentials for QPM maize BHQPY545 which performed better
and gave significant highest grain yield with 4 fertilizer rates 92/69 kg ha-1
N/P2O5, 25% Compost+75% recommended NP fertilizer, 50% Compost+ 50%
recommended NP and 100% Compost+100% recommended NP fertilizer (Table
6).
Table 6. Validation of QPM maize varieties under integrated uses of different fertilizer types
on grain yield and yield components at Omonada woreda, Jimma zone during 2018 cropping season
Fertilizer types and rates Plant height Grain yield Above ground biomass
(cm) (Kg ha-1) (ton ha-1)
92/69 N/P205 254.75 7123.5b 12.35b
25% Compost+75% Rec NP 250.75 6789.5b 11.63b
50% Compost+50% Rec NP 252.75 6698.8b 11.80b
100% Compost+100% Rec NP 262.25 8181.3a 13.70a
LSD (0.05) NS 887.23 1.34
CV (%) 12.00 7.71 6.76
Values followed by the same letter within a column are not significantly different at P< 0.05;
Rec= Recommended; Compost was applied based on N-equivalence rate of recommended fertilizer rate (92 N kg ha-1)
The farmer’s perception was collected at the green ear stage and harvest period.
Seven maize stand evaluation criteria were set by farmers to decide optimal
fertilizer recommendation for BHQPY545 QPM maize variety. Accordingly,
maize growth rate, probability of lodging, number of ears/plant and yield potential
were found the most important for the maize stand evaluation criteria.
Furthermore, based on maize stand evaluation criteria that were set by farmers
[315]
(Table 7) 35% of them chosen for 100% Compost+ 100% recommended NP
fertilizer.
Table 7. Farmers perception on optimal integrated uses of different fertilizer types at Omonada woreda at 2018 cropping
season
Farmers Evaluation 92/69 N-P205 25% Compost 50% Compost+ 100% Compost+
Criteria +75% Rec NP 50% Rec NP 100% Rec NP
Weeding Frequency Low Low medium High

Growth rate Slow slow Fast Fast
Probability of lodging medium medium medium Medium
Number of ears/plant 2 2 2 2-3
Cob size Medium Medium Medium Bigger
Number of ear rotting medium medium low low
Yield potential Medium Medium Medium Higher
choice in Percentage 25% 22% 18% 35%
In general, from farmer’s decision point and current on-farm input availability and
economic analysis with highest MRR 414.79% (Table 8), 92/69 N-P205 fertilizer
application taken as optimal fertilizer application and recommended for the
production of BHQPY545 in the Woredas and other similar humid agro-ecologies
of the west and southwest Ethiopia.
Table 8. Partial budget with estimated MRR (%) for the effect of organic and inorganic fertilizer on grain yield of QPM
maize variety at Omonada woreda in 2018 cropping season
GY Adj.GY GFB TVC NB Domi MRR (%)

Fertilizer types and rates (kg ha-1) (kg ha-1) (ETB ha-1) (ETB ha-1) (ETB ha-1) nance
50% Compost + 50% Rec NP 6698.8 6028.92 36173.52 3067.5 33106.02 --- ---
92/69 kg ha-1 N/P205 7123.5 6411.15 38466.9 3513 34953.9 UD 414.79
25% Compost+ 75% Rec NP 6789.5 6110.55 36663.3 3576.75 33086.55 D ---
100% Compost+100% Rec NP 8181.3 7363.17 44179.02 5443 38736.02 UD 195.96
D=Dominated treatment; UD= Un-dominated treatments; Rec= Recommended; ETB = Ethiopian Birr; Price of chemical
fertilizer = 13.5 birr kg-1; Price of Urea = 10 birr kg-1; Wage rate = 30 Birr man-day-1; Retail price of grain = 6 birr kg-1

The results revealed that individual chemical fertilizer and compost improved
grain yield and above ground biomass yield of the QPM maize varieties. The
improvement was mainly due to the availability of nutrients from the chemical
fertilizer and compost for plant development up to cob formation. The application
of integrated organic and chemical fertilizer increased grain yield and above-
ground biomass of the crop mainly due to better grain development. Summary of
[316]
results from across seasons and sites indicated compost and chemical fertilizer
significantly affects the grain and above-ground biomass yields of the QPM maize
varieties. To increase certainty and further use of optimum fertilizer use evaluation
criteria were set by farmers and subjected to choose or reject fertilizer rate and
types not technically appropriate. From farmer's decision point and current on-
farm input availability and technical feasibility and economic analysis, 92/69 kg
ha-1 N/P205 fertilizer application taken as optimal fertilizer application and
recommended for the production of BHQPY545 in Jimma area and other similar
humid agro-ecologies of the west and southwest Ethiopia. This approved the
recommended fertilizer application of before. Also, the application of 75%
compost + 25% recommended NP fertilizer could be recommended as alternative
sources for users.
References
Adefris,T, Dagne W, Abraham T, BirhanuT, Kassahun B, Dennis F and Prasanna BM. 2015.
Quality Protein Maize (QPM): A Guide to the Technology and Its Promotion in Ethiopia.
CIMMYT: Addis Ababa, Ethiopia.34p
Adekayode, FO and Ogunkoya MO. 2010. Effect of quantity and placement distances of inorganic
fertilizer in improving soil fertility status and the performance and yield of maize in a tropical
rain forest zone of Nigeria. J. of Soil Sci. and Env. Management, 1: 155-163.
Amanullah, A. Iqbal A and Iqbal M. 2015. Impact of potassium rates and their application time on
dry matter partitioning, biomass and harvest index of maize with and without cattle dung
application. Emirates Journal of Food and Agriculture, 27: 447-453.
Ayub, M, Nadeem MA, Sharar MS and Mahmood N. 2002. Response of maize fodder
to different levels of nitrogen and phosphorus. Asian Journal of Plant Science, 1: 352-354.
Aman, J, Kassahun B, Sintayehu A and Tolera B. 2016. Evaluation of Quality Protein
Maize (Zea mays L) Hybrids at Jimma, Western-Ethiopia. J. Forensic Anthropol, 1: 101.
Balai, ML, Verma A, Nepalia V and Kanthaliya PC. 2011. Productivity and quality of
maize (Zea mays) as influenced by integrated nutrient management under continuous
cropping and fertilization. The Indian Journal of Agricultural Sciences, 81: 374-376.
Batjes, NH. 1995. A global data set of soil pH properties. Technical Paper 27, International Soil
Brady, N and Weil RR. 2002. The nature and properties of soils. 13 edition. Prentince Hall. New
Jersey, USA. 598p.
Brady, N and Weil RR. 2008. Nature and Properties of Soils (14th ed.). Prentice Hall, Upper
Saddle River. 992p.
Brady, NC, Weil RR and Brady NC. 2010. Elements of the nature and properties of soils (No.
631.4 B733E.). Upper Saddle River, NJ: Pearson educational international.
Chang, C, Sommerfeldt TG and Entz T. 1990. Rates of soil chemical changes with eleven annual
applications of cattle feedlot manure. Canadian J. of Soil Sci., 70: 673-681.
CIMMYT. 988. From Agronomic Data to Farmer Recommendations: An Economics
Training Manual. Completely revised edition. Mexico, D.F.79p
Dagne, C. 2016. Blended Fertilizers Effects on Maize Yield and Yield components of Western
Oromia, Ethiopia. Agriculture, Forestry and Fisheries, 5:151-162.
Diacono, M and Montemurro F. 2010. Long-term effects of organic amendments on soil
fertility: a review. Agron Sustain Dev., 30: 401-422.
[317]
Gemechu, N, Sentayehu A, Leta T. 2016. Review on Quality Protein Maize Breeding for
Ethiopia. Journal of Biology, Agriculture and Healthcare, 6: 84-96.
Getachew, A and Taye B. 2005. On-farm integrated soil fertility management in wheat on
nitisols of central Ethiopian highlands. Ethiopian Journal of Natural Resources, 7: 141-155.
Kogbe, JOS and Adediran JA. 2003. Influence of nitrogen, phosphorus and potassium
application on the yield of maize in the savanna zone of Nigeria. African Journal of
Biotechnology, 2: 345-349.
Leta, T, Mosisa W, Gelana S, Jemal A, Hadji T, Sewagegne T and Twumasi-Afriyie S.
2001. Enhancing the Contribution of Maize to Food Security in Ethiopia. pp.39-45.
Proceedings of the Second National Maize Workshop of Ethiopia 12-16 November 2001.
Makinde, EA and Ayoola OT. 2010. Growth, yield and NPK uptake by maize with complementary
organic and inorganic fertilizers, African Journal of Food, Agriculture, Nutrition and
Development, 10:2203-2217.
Marschner, P. 2012. Rhizosphere biology. pp. 369-388. In: Marschner's Mineral Nutrition of
Higher Plants (Third Edition).
Mugwe, J, Mugendi D, Kungu J and Mucheru-Muna M. 2007. Effects of plant biomass, manure
and inorganic fertilizer on maize yield in the Central Highlands of Kenya. African Crop
Science Journal, 15: 111-126.
N'Dayegamiye, A, Nyiraneza J, Chantigny H and Laverdiere R. 2010. Long-term manure
application and forages reduce nitrogen fertilizer requirements of silage corn-cereal cropping
systems. Agronomy Journal, 102: 1244-1251.
Ngwira, AR, Nyirenda M and Taylor D. 2013. Toward sustainable agriculture: an
evaluation of compost and inorganic fertilizer on soil nutrient status and productivity of three
maize varieties across multiple sites in Malawi. Agroecology and sustainable food systems,
37: 859-881.
Onasanya, RO, Aiyelari OP, Onasanya A, Oikeh S, Nwilene FE and Oyelakin OO.
2009. Growth and yield response of maize to different rates of nitrogen and phosphorus
fertilizers in southern Nigeria. World J. of Agr. Sciences, 5: 400-407.
Onwudiwe, N, Benedict OU, Ogbonna PE and Ejiofor EE. 2014. Municipal solid waste and NPK
fertilizer effects on soil physical properties and maize performance in Nsukka, Southeast
Nigeria. African Journal of Biotechnology, 13: 68-75.
Rahman, MA, Rahman MM, Begum MF and Alam MF. 2012. Effect of bio compost, cow dung
compost and NPK fertilizers on growth, yield and yield components of chili. International
Journal of Biosciences, 2: 51-55.
Rao, S, Singh KK, Ali M. 2001. Sulfur: A key nutrient for higher pulse production. Fert.News, 46:
37-50.
Ravi, N, Basavarajappa R, Chandrashekar CP, Harlapur SI, Hosamani MH and Manjunatha MV.
2012. Effect of integrated nutrient management on growth and yield of quality protein maize.
Journal of Agricultural Sciences, 25: 395-396.
Tekalign M and Haque I. 1991. Phosphorus status of some Ethiopian soils, II. Forms and
distribution of inorganic phosphates and their relation to available phosphorus. Trop Agric.,
68: 2-8.
Twumasi-Afriye S, Palacios Rojas N, Friesen D, Teklewold A, Gissa DW, De Groote
H and Prasanna BM. 2016. Guidelines for the quality control of Quality Protein Maize seed
and grain. 38p.
Vanlauwe, B. 2002. Integrated plant nutrient management in sub-Saharan Africa: from
concept to practice. CABI. 355p
Yayock, JY, Lombin G and Owonubi JJ. 1988. Crop science and production in warm climates.
Macmillan publishers, London. 314p
[318]
Determination of Nitrogen and Phosphorous
Fertilizer Rates on Lowland Rice Production
Tilahun Tadesse1*, Zelalem Tadesse1, Habtamu Assega1 and Desta Abaychew1

1
Fogera National Rice Research and Training Center, Woreta, Ethiopia
*Corresponding author: email-tilahuntadesse2000@gmail.com
Abstract
Nitrogen and phosphorous fertilizer application experiment was conducted on rainfed
lowland rice production at Fogera plain in two cropping seasons of the years 2016 and
2017. The treatments were comprised of factorial combinations of five nitrogen
(0,92,184, 276 and 368 kg/ha) and four phosphorous levels (0, 23, 46, and 69 P 2O5
kg/ha). Data were collected on plant height. panicle length, number of total tillers/m,
number of effective tillers/m, panicle length, number of fertile spikelets/panicle,
thousand seed weight, grain yield, straw yield and harvest index. All collected data
were subjected to analysis of variance. Economic analysis was also carried out by
following CIMMYT (1988) procedures. The results of the experiment indicated that the
main effect of nitrogen application was significantly affecting plant height, panicle
length, total tillers/m, effective tillers/m, filled spikelets/panicle, grain yield, straw
yield, thousand seeds weight and harvest index while phosphorous was affecting only
grain and straw yields. The interaction of nitrogen and phosphorous were affecting
grain yield, straw yield and harvest index. Concerning the interaction effect, the
highest grain yield (7.0 t ha-1) was obtained at 276-69 N- P2O5 kg ha-1. The economic
analysis has exhibited that the combined application of 184-46 N- P2O5 kg ha-1 is the
most profitable treatment. It is thus concluded that the application of nitrogen and
phosphorous fertilizers at rates of 184-46 N- P2O5 kg ha-1 is the best recommended for
rainfed lowland rice production in Fogera plain and other similar agro-ecologies.
Keywords: Lowland rice; Nitrogen; Phosphorous; profitable.
Introduction
Rice (Oryza sativa L.) is one of the most popular cereal crops in the world. It is
the principal food for one-third of the world’s population. It is the most important
food crop for the world’s population, especially in South Asia, Middle East, Latin
America and West India (Zhao et al., 2011). More than 90% of rice is produced
and consumed in Asia (Subediet al, 2019). It provides some 700 calories per
person, mostly residing in developing countries. In Ethiopia, rice production was
started three decades ago in the early 1970s and the country has reasonable
potential to grow various rice types mainly in rainfed lowland, upland and
irrigated ecosystems (Mulugeta and Heluf, 2014). Though rice is a recent
introduction to the country, its importance is well recognized as the production
area coverage of about 10,000 ha in 2006 has increased to over 63,000 ha in 2018
(CSA 2019). The area coverage in domestic rice production has increased
[319]
considerably linked with the expansion of production in the wetland and upland
areas with the introduction of suitable rice varieties for the different agro-
ecologies. In line with the area expansion, the production levels have been
increasing consistently over the years. Rice production increased from 71,316.07
tons in 2008 to 171,854.1 tons in 2018 (CSA, 2018). The number of farmers
engaged in rice production has also grown year after year. Rice production has
brought a significant change in the livelihood of farmers and created job
opportunities for many citizens in different areas of the country. Currently,
Amhara, Southern Nations, Nationalities and Peoples Region (SNNPR), Oromiya,
Somali, Gambella, Beni Shangul Gumuz, and Tigray regions are the rice-
producing areas in Ethiopia (MoARD, 2010). The Amhara region takes the lion’s
share of producing the crop and accounted for 65-81% of the area coverage and
78-85% of the production in the years 2016-2018 (CSA 2017, CSA 2018 and
2019). According to the report of MoARD (2010), the potential rice production
area in Ethiopia is estimated to be over 5,590,895 ha. Most of Ethiopia’s rice
production potential area lies in the western part of the country.
The national average yield of rice is about 2.8 t ha-1 (CSA, 2018) which is lower
compared to the world average productivity of 4.6 tones ha-1 (FAOSTAT, 2018).
Weeds, pests, soil nutrient deficiencies and terminal moisture stress are the major
causes of low rice productivity in Ethiopia (MoARD, 2010; Gebeyet al., 2012).
Poor soil fertility is among the major factors limiting rice production in Ethiopia.
Nitrogen, phosphorus, and potassium are applied as fertilizers in large quantities
to rice fields, and a deficiency of either of the nutrient leads to yield losses
(Subediet al, 2019). Nitrogen and phosphorus are often cited as the most limiting
nutrients in agricultural soils of Ethiopia (Molla and Sofonyas, 2018). Appropriate
fertilizer application is an important management practice to improve soil fertility
and the production of rice. The availability of plant nutrients, particularly nitrogen
at various plant growth stages is of crucial importance in rice production
(ShaRadaet al., 2018; Daquiado, 2019). Therefore, a fertilizer experiment was
conducted on the lowland rice production of Fogera Plain to recommend
appropriate levels of nitrogen and phosphorous rates.

Rainfed lowland nitrogen and phosphorous rates experiment was conducted at
Fogera plain in two cropping seasons of the years 2017 and 2018 cropping
seasons. The experimental site is located between Latitude 11°49’55 North and
Longitude 37° 37′ 40 East at an altitude of 1815 meters above sea level. The study
site receives averages mean annual rainfall, minimum and maximum temperature
of 1219 mm, 12.75°C and 27.37°C, respectively. The long-term rainfall data
(1986-2017) years indicated that much of the rainfall appear in July and August
(Figure 1).
[320]
60 150
Temperature (°C)
Mean Rainfall (mm)
40 100
20 50
0 0
SR18
BK18
HS18
HU16
AT17
AN17
JM16
AN18
MK18
31.93 74.17 15.1
Figure 1. The Rainfall and Temperature condition of Fogera Plain for the period 1981-2017
The experimental site's soil was found to be heavy clay with a pH range of 5.87-
6.63, which is slightly acidic and it is a preferred range for most crops (Table 1).
Total nitrogen content (%) was with a range of 0.09-0.16, which is within the
range of low levels (0.02-0.5%) for tropical soils. The organic matter content of
the soil was between 2.13-3.09%, which is within a range of medium (2-4%) for
Ethiopian soils as per criteria developed by Murphy (1968). The available P
content of the soil of the experimental site was 11.4-25.13ppm, which lies in a
range of deficiency (< 20-40mg/kg) for most crops (Landon, 1991).
Table 3. Relevant soil physicochemical properties of the experimental rice field before planting in Fogera Plain of Ethiopia
Soil properties Units Minimum Value Maximum value

Textural class Heavy clay Heavy clay
Chemical properties
pH (H2O) 1:2.5 g soil - 5.87 6.63
Total nitrogen (TN) % 0.09 0.16
Organic carbon (OC) % 1.24 1.93
Organic matter (OM) % 2.13 3.09
Available Phosphorus Ppm 11.4 25.13
The experimental treatments were comprised of factorial combinations of five

nitrogen (0, 92,184, 276 and 368 kg/ha) and four phosphorous levels (0, 23, 46
and 69 P2O5 kg/ha). The gross size of plots was 2m x 4m consisting of 10 rows
planted at a spacing of 20 cm apart with a seed rate of 100kg/ha. The net plot was
made by excluding the left and right outer rows and a plot length of 0.5 m from
the top and bottom sides of the plot. The final net plot size was thus 1.6m x 3m.
Data was collected from the net plot area on plant height, number of total tillers/m,
number of effective tillers/m, number of filled spikelets/panicle, thousand seed
weight, grain yield, straw yield and harvest index. The plant height was taken at
the physiological maturity of the crop by selecting five random tillers. The number
of tillers was counted just before harvesting by random sampling using rulers. The
[321]
total sundried biomass of the harvested rice was recorded before threshing. The
harvest index was calculated as the ratio of grain yield to biological yield
following the equation:
The rice grain yield and thousand seed weight were adjusted at 14% standard
moisture content. All collected data were subjected to analysis of variance
(ANOVA) using SAS software version 9.2 (SAS-Institute, 2008). Since the test of
homogeneity of variances for each parameter was non-significant, combined
analysis of variance was done over the years to determine the effects of
application methods and N rates by year interaction. Wherever treatment
differences are be found significant, mean separation of treatments would be
calculated based on results of F-test and probability levels of 0.01 and 0.05
depending on the results of the ANOVA.
Agronomic Efficiency (AE) was calculated to assess the efficiencies of the applied
N rates as follows:
AE= (Gf - Gu / Na) kg rice grain/ kg N fertilizer applied
Where Gf is the grain yield of the fertilized plot (kg), Gu is the grain yield of the
unfertilized plot (kg), and Na is the rate of applied N fertilizer (kg) (Liu et al,.
2019). Economic analysis was carried out by following CIMMYT (1988)
procedures by taking all variable costs. The prevailing cost of inputs and outputs
in the year 2019 considered for the analysis. The cost of Urea and TSP fertilizers
for the stated period at Fogera was Birr 13.1 and 14.3, respectively while the price
of rice grain and straw was Birr 13.5 and 1.2.

The analysis of variance indicated that the main effect of nitrogen was highly
significantly (P<0.001) affecting plant height, panicle length, total tillers/m,
effective tillers/m, filled spikelets/panicle, grain yield, straw yield, thousand seeds
weight and harvest index (Table 2). On the other hand, phosphorous was highly
significant (P<0.001) affecting grain and straw yields (Table 3). The interaction of
nitrogen and phosphorous applications was highly significant (P<0.001) affecting
grain and straw yields and significantly (P<0.005) affecting harvest index (Table
3).
The comparison to the nitrogen rates indicated that the highest values of plant
height (109.0 cm), panicle length (20.9 cm)number of total tillers/m (82.1),
number of effective tillers/m (80.7), number of filled spikelets per panicle (102.9)
(Table 3). In line with the present findings, Sah et al., (2019) had reported that
different levels of N caused a significant difference in plant height, the height of
[322]
the plant found to increase from 60 kg-1 N to 120 kg N ha-1. The findings of many
authors had confirmed the significant effect of nitrogen levels on panicle length
(Fageria and Baligar, 2001; Gewaily et al., 2018; Sah et al., 2019). Sahet al.,
(2019) recorded the highest panicle length with 180 kg N application while
Fageria and Baligar, (2001) stated nitrogen application of 210 kg ha -1 exhibited
larger panicle length. In conformity of the present experiment, Fageria and
Baligar, (2001), Dong et al., (2016) and Sah et al., (2019) similarly concluded that
nitrogen application significantly increases the total tillers number. Sahet al.,
(2019) obtained the highest total tillers for N at 120 kg ha-1 while Dong et al.,
(2016) gained the largest tillers number at 210 Kg ha-1 N. Similar to the total
number, the number of productive tillers depends on environmental conditions
especially nutrient applications (Fageria and Baligar, 2001). Gewaily et al., (2018)
reported that fertile tillers were increased significantly with increasing nitrogen
levels from 0 to 220 kg N ha-1 for all rice. Many authors reported for the
significant response of the number of fertile spikelets per panicle to nitrogen
application confirming the current observation Fageria and Baligar, (2001), Wang
et al., (2017), Gewailyet al., (2018) and Liu et al,. (2019). A maximum number of
fertile spikelets per panicles were observed at N rates of 210-220 Kg ha-1by
Fageria and Baligar, (2001) and Gewaily et al., (2018).
Concerning the yields, the highest straw yield (21.7 t ha-1) was obtained at the
highest rate of 368 kg N ha-1 while the highest grain yield(5.99 t ha-1) was
exhibited at the 276 kg N ha-1 rate, and the highest the thousand seeds weight
(29.0 g) was at 184 kg N ha-1(Table 3). Quite differently, the highest harvest index
(34.51%) was observed at the no (0kg N ha-1) N rate (Table 3). The lower values
for the respective parameters except for the harvest index were recorded at no (0kg
N ha-1) N application. In the case of the harvest index, the lowest score was at the
maximum (368 kg N ha-1) nitrogen rate (Table 3). Similar to the observations
made at the study, Sah et al., (2019) reported for a significant effect of N on straw
yield mentioning that it was highest at N application of 180 kg ha-1. Reporting
significant responses of grain yield to N application, some authors observed the
highest rice grain yields at rates nearer to the current higher-yielding rate.
Different authors reported that nitrogen application increase the grain yield and
largest values recorded at the nitrogen application treatment of 209- 220kg N ha-1
(Fageria and Baligar, 2001, Dong et al., 2016, Gewaily et al., 2018). A bit
differently, Liu et al,. (2019) reported the highest mean grain yield of 10.5 t ha-1at
300 kg ha-1N treatment elaborating that as the N rates increased to 360 kg ha-1, the
mean grain yield decreased to 9.4 t ha-1. Optimum fertilizer level plays an
important role in achieving crop potential yield. Among the fertilizer, N is most
important for the proper growth and development of rice (Sah et al., 2019). The
increase in grain yield might be due to nitrogen application enhancing the dry
matter production, improving rice growth rate, promoting elongation of
[323]
internodes, and activity of growth hormones like gibberellins (Gewaily et al.,
2018).
The only growth parameters which were significantly responding to the main
effects of phosphorous application are grain and straw yields (Table 2). The
highest grain (5.20 t ha-1) and straw (15.2 t ha-1) yields were obtained at the
highest phosphorous (69 kg P2O5 ha-1) rate (Table 4). The non-phosphorous
application (0 kg P2O5 ha-1) gave the lowest grain (4.00 t ha-1) and straw(12.58 t
ha-1) yields.The increase in P levels resulted in higher rice productivity (90 > 60 >
30 > 0 kg P ha-1) (Amanullah et al., 2016). The higher grain yield may be
attributed due to better growth with higher nutrient availability and higher
photosynthetic rate of the plants and more photosynthate partitioning into the
reproductive parts (Amanullah et al., 2016). The response curve revealed that the
grain yield of rice showed a declining trend after the application of 276 kg N ha-1
while the yield had a linear response to the phosphorous application and the
maximum yield occurred at the maximum rate of 69 kg P2O5 ha-1 (Figure 2). This
indicates that future rice fertilizer experiments in Fogera plain shall consider
higher rates of phosphorous.
As revealed in the analysis of variance (Table 5), the grain yield, straw yield, and
harvest index responded significantly to the interaction effects of nitrogen and
phosphorous (Table 2). The comparison to the grain yield indicated that the
highest (7.00 t ha-1) was gained at the combination of 276-69 N- P2O5 kg ha-1
which is statistically at par with the yields recorded at 368-69, 368-46, and 184-46
N- P2O5 kg ha-1 applications (Table 5). The lowest grain yield was at nill (0-0 N-
P2O5 kg ha-1) which is statistically equivalent with that of 0-23, 0-46, 0-69 N-
P2O5 kg ha-1 applications (Table 5). Regarding the straw yield, the highest (23.16 t
ha-1) was gained at the combination of 276-69 N- P2O5 kg ha-1 which is
statistically at par with the straw yields shown at 368-69, 368-46, 184-46 and 368-
23 N- P2O5 kg ha-1 applications (Table 5). The lowest straw yield of 4.11 t ha-1was
exhibited at no (0-0 N- P2O5 kg ha-1) application which is statistically at par with
that of 0-23, 0-46, 0-69 N- P2O5 kg ha-1 applications (Table 5). Management of
soil fertility largely determines the availability of N and P for crop plants (Iqbal et
al., 2017). Mineral nutrition in rice requires 16 essential elements of which
nitrogen (N), phosphorus (P), and potassium (K) are applied to rice fields as
chemical fertilizers in large quantities (Vinod and Sigrid, 2012). Nitrogen and P
are fundamental to crop development because they form the basic component of
many organic molecules, nucleic acids, and proteins (Vinod and Sigrid, 2012).
The harvest index values comparison showed that the highest (23.16 t ha-1) was
gained at 0-0 N- P2O5 kg ha-1 which is statistically similar to the indices shown at
0-23, 0-46, 0-69 N- P2O5 kg ha-1 applications (Table 5). The lower harvest indices
were associated with the combinations of the higher nitrogen and any of the
phosphorous rates, the lowest (23.04 %) being at 276-69 N- P2O5 kg ha-1 (Table
[324]
5). The analysis of the agronomic efficiency analysis for nitrogen indicates that
the maximum AE of 15.43 was exhibited at 92 Kg ha-1 N (Table 6). As the N rate
increased the AE was decreasing and finally, a negative value of AE (-0.87) was
recorded to the maximum (368 Kg ha-1) n application. Agronomic efficiency is
usually higher at a low N rate than at a high N rate (Gewaily et al., 2018, Yasuhiro
et al., 2019). In tropical Asia, with proper crop and water management, AEN
should be typically in the range of 20–25 kg ha1 (Yasuhiro et al., 2019). Yoshida
(1981) estimated better agronomic N use efficiency to be 15–25 kg rough rice per
kg applied N in the tropics. Peng et al., (2010) reported that agronomic N use
efficiency was 15 to 18 kg ha-1 N in the dry season in the farmers’ fields in the
Philippines. In China, agronomic N use efficiency was 15–20 kg ha-1 N from
1958 to 1963 and declined to only 9.1kg ha-1 between 1981 and 1983 (Peng et al.,
2010). Since then, agronomic N use efficiency has further decreased in China
because of the increase in N rate (Peng et al., 2010). Generally, fertilizer N use
efficiency of lowland rice is relatively low due to loss of applied N through
leaching, volatilization, and denitrification in the soil-flood water system which
necessitates the need for improved N fertilizer practices to reduce environmental
impacts and increase economic benefits of N fertilization (Fageria and Baligar,
2001). The lower agronomic efficiency at the highest N rates in the current
experiment indicate that emphasis should be given to efficient nitrogen application
methods like the split applications, use of slow N releasing fertilizer sources and
real-time N management to reduce the wastage of N in the rice production system
of the Fogera plain.
Economic management of N fertilizer application is essential for improving crop

productivity, N use efficiency, and environmental sustainability (Yousaf et al.,
2016). Following the CIMYYT (1988) partial budget analysis method, grain and
straw yield adjustments, calculations of total variable costs (TVC), gross benefits
(GB), and net benefits (NB) were performed (Table 7). Dominance analysis was
carried after arranging the treatments in their order of TVC. A treatment was
considered as dominated if it has higher TVC but lower NB than a previous
treatment with lower TVC and higher NB. Non dominated treatments were taken
out and the marginal rate of return (MRR) was computed (Table 8). According to
the CIMYYT (1988) partial budget analysis methodology, treatments exhibiting
the minimum or more MRR (>100%) will be considered for the comparison of
their NB. The highest NB (Birr 98,111.85/ha) with an acceptable level of MRR
(3499.15) was observed at 184-46 N-P2O5 kg ha-1 (Table 9). In agreement with the
present finding Irfan et al., (2016) reported that rice genotypes performed
efficiently at 120 kg N + 90 kg P2O5 ha-1 where highest paddy yield, net
production value, and profit were obtained. The combined application of nitrogen
and phosphorous at 184-46 N-P2O5 kg/ha is the most profitable rate to be
recommended for rice production in Fogera plain.
[325]
Table 2. Analysis of variance (ANOVA) for rice yield and yield components
No of Filled
Total no of tillers/m No of Effective Spikelets/
Source DF Plant height Panicle length Mean Square tillers/m Row panicle GY (t/ha) SY (t/ha) 1000 SW (g) HI (%)
N 4 14080.6** 147.99** 8217.03** 8186.4** 8786.1** 193.7** 3470.9** 29.3** 1231.1**
P 3 7.88 NS 8.13 NS 121.95 NS 136.5 NS 411.3 NS 23.6 ** 202.7 ** 4.4 NS 9.2 NS
N*p 12 27.51 NS 10.6 NS 139.2 NS 136.1 NS 312.1 NS 3.5 ** 37.6 ** 5.2 NS 30.2*
Error 96 40.9 6.2 17918.08 189.6 363.2 0.3 7.4 5.7 14.9
CV% 6.76 12.76 18.79 19.3 22.37 11.61 18.44 7.77 14.0
Table 3. Main effect of nitrogen rates on rice yield and yield components
Plant height Panicle length Total tillers Effective tillers No of Fertile Grain Yield Straw Yield
N (kg/ha) (cm) (cm) number/m number/m spiklets per panicle (t/ha) (t/ha) TSW (G) HI (%)
0 72.7D 17.0C 53.1D 51.6D 74.2C 2.14E 4.07E 30.45BC 34.51A
92 85.8C 19.5B 62.3C 60.7C 88.2B 3.56D 9.63D 30.99AB 28.45B
184 98.3B 20.2AB 71.4B 70.2B 98.6A 5.64C 17.28C 31.65A 25.64C
276 107.2A 20.6A 77.4A 75.7A 101.8A 6.31A 20.67B 30.30BC 23.91D
368 109.0A 20.9A 82.1A 80.7A 102.9A 5.99B 21.78A 29.83C 23.49D
[326]
Table 4. Main effect of Phosphorous rates on rice grain and straw yields
P2O5 (kg/ha) Grain Yield Straw Yield (t/ha)

(t/ha)
0 4.00C 12.58C
23 4.59B 14.41B
46 5.14A 16.5A
69 5.20A 15.26B
7
Grain yield (tha-1)
6
5
4 y = -0.3493x2 + 3.1407x - 0.852
3 R² = 0.9722
2
1
0
0 92 184 276 368
Nitrogen rates (Kg N ha-1)
6
5
4
y = 0.415x + 3.695
3
R² = 0.9147
Grain yield (tha-1)
2
1
0
0 23 46 69
Phosphorous rates (Kg P2O5 ha-1)
Figure 2. Response curves of rice grain yield (t ha-1) to nitrogen and phosphorous application levels
[327]
Table 5. Effects of N and P on Lowland Rice grain yield, straw yield and harvest index
N P GY (t/ha) SY (t/ha) HI (%)

0 0 2.09H 4.11I 33.75A
92 0 3.03G 7.10H 30.50B
184 0 4.19E 13.77F 24.86D-G
276 0 5.31D 18.65D 22.60FG
368 0 5.36D 19.29D 22.03G
0 23 2.13H 3.74I 35.99A
92 23 3.76F 10.47G 27.58CD
184 23 5.53D 16.55E 25.46DEF
276 23 6.40BC 19.57CD 24.93D-G
368 23 5.15D 21.73AB 23.07FG
0 46 2.15H 4.16I 34.33A
92 46 3.79F 10.75G 26.49CDE
184 46 6.61ABC 22.33AB 24.81D-G
276 46 6.52BC 21.37BC 25.03D-G
368 46 6.63ABC 23.90A 24.82D-G
0 69 2.20H 4.25I 33.98A
92 69 3.76F 10.22G 29.24BC
184 69 6.22C 16.45E 27.40D
276 69 7.00A 23.16AB 23.04FG
368 69 6.83AB 22.21AB 24.03EFG
CV % 11.61 18.44 14.0
Table 6. Agronomic Efficiency (AE) of rice
N (kg/ha) Grain Yield (kg/ha) AE

0 2140 -
92 3560 15.43
184 5640 11.30
276 6310 2.43
368 5990 -0.87
[328]
Table 7. Grain and straw yield adjustments, total variable cost, gross and net benefit analysis
TVC GY SY AGY ASY GB NB

N P
(Birr/ha) (t/ha) (t/ha) (t/ha) (t/ha) (Birr/ha) (Birr/ha)
0 0 0 2.09 4.11 1.881 3.699 29832.3 29832.3
92 0 2620 3.03 7.1 2.727 6.39 44482.5 41862.5
184 0 5240 4.19 13.77 3.771 12.393 65780.1 60540.1
276 0 7860 5.31 18.65 4.779 16.785 84658.5 76798.5
368 0 10480 5.36 19.29 4.824 17.361 85957.2 75477.2
0 23 865.52 2.13 3.74 1.917 3.366 29918.7 29053.2
92 23 3158.03 3.76 10.47 3.384 9.423 56991.6 53833.6
184 23 5778.03 5.53 16.55 4.977 14.895 85063.5 79285.5
276 23 8398.03 6.4 19.57 5.76 17.613 98895.6 90497.6
368 23 11018.03 5.15 21.73 4.635 19.557 86040.9 75022.9
0 46 1731.05 2.15 4.16 1.935 3.744 30615.3 28884.2
92 46 3696.05 3.79 10.75 3.411 9.675 57658.5 53962.4
184 46 6316.05 6.61 22.33 5.949 20.097 104427.9 98111.8
276 46 8936.05 6.52 21.37 5.868 19.233 102297.6 93361.5
368 46 11556.05 6.63 23.9 5.967 21.51 106366.5 94810.4
0 69 2596.58 2.2 4.25 1.98 3.825 31320 28723.4
92 69 4234.08 3.76 10.22 3.384 9.198 56721.6 52487.5
184 69 6854.08 6.22 16.45 5.598 14.805 93339 86484.9
276 69 9474.08 7 23.16 6.3 20.844 110062.8 100588.7
368 69 12094.08 6.83 22.21 6.147 19.989 106971.3 94877.2
Table 8. Dominance Analysis
N P TVC (Birr/ha) NB (Birr/ha) Dominance

0.0 0.0 0 29,832.30
0.0 23.0 865.5263 29,053.17 D
0.0 46.0 1731.053 28,884.25 D
0.0 69.0 2596.579 28,723.42 D
92.0 0.0 2620 41,862.50
92.0 23.0 3158.026 53,833.57
92.0 46.0 3696.053 53,962.45
92.0 69.0 4234.079 52,487.52 D
184.0 0.0 5240 60,540.10
184.0 23.0 5778.026 79,285.47
184.0 46.0 6316.053 98,111.85
184.0 69.0 6854.079 86,484.92 D
276.0 0.0 7860 76,798.50 D
276.0 23.0 8398.026 90,497.57 D
276.0 46.0 8936.053 93,361.55 D
276.0 69.0 9474.079 100,588.72
368.0 0.0 10480 75,477.20 D
368.0 23.0 11018.03 75,022.87 D
368.0 46.0 11556.05 94,810.45 D
368.0 69.0 12094.08 94,877.22 D
[329]
Table 9. MRR Analysis
N P TVC (Birr/ha) NB (Birr/ha) MRR (%)
0.0 0.0 0 29,832.30
92.0 0.0 2620 41,862.50 459.168
92.0 23.0 3158.026 53,833.57 2224.998
92.0 46.0 3696.053 53,962.45 23.953
184.0 0.0 5240 60,540.10 426.028
184.0 23.0 5778.026 79,285.47 3484.100
184.0 46.0 6316.053 98,111.85 3499.155
276.0 69.0 9474.079 100,588.72 78.431
References
Amanullah, Khan S-u-T, Iqbal A and Fahad S. 2016. Growth and Productivity Response of
Hybrid Rice to Application of Animal Manures, Plant Residues and Phosphorus. Front.
Plant Sci. 7:1440. doi: 10.3389/fpls.2016.01440.
CIMMYT. 1988. From Agronomic Data to Farmer Recommendations. An Economics
Training Manual. Completely Revised Edition. CIMMYT, Mexico, D. F., Mexico. 79
pp.
CSA (Central Statistical Agency). 2017. Report on Area and Production of Major Crops
(Private Peasant Holdings, Meher Season) The Federal Democratic Republic of
Ethiopia Central Statistical Agency Agricultural Sample Survey Volume I, 2017 /18
(2010 E.C.) April 2018. ADDIS ABABA, Ethiopia.
(2011 E.C.). April 2019. ADDIS ABABA, Ethiopia.
CSA (Central Statistical Agency). 2019. Report on area and production of major crops
(private peasant holdings, meher season). The federal democratic republic of Ethiopia
central statistical agency agricultural sample survey 2018/19 (2011 E.C.). Volume i.
Statistical bulletin 589.
DaquiadoNonilona P. 2019. Growth and Yield Performance of Some Lowland Rice
Varieties Applied with Different Rates of Organic and Inorganic Fertilizers. Asian
Journal of Soil Science and Plant Nutrition. 4(2): 1-11, 2019; Article no.
AJSSPN.42493 ISSN: 2456-9682.
Dong JIA, Jingjing Lu, Jajun SUNＥ, Shuang SONG, Han Du, and Lei HAN. 2016. Effect
of Different Nitrogen Fertilizer Application Strategies onRiceGrrowth and Yield.Asian
Agricultural Research 2016, 8(1):33-39.
Fageria, NK and Baligar VC. 2001. 'LOWLAND RICE RESPONSE TO NITROGEN
FERTILIZATION', Communications in Soil Science and Plant Analysis, 32: 9, 1405 —
1429 To link to this Article: DOI: 10.1081/CSS-100104202 URL:
http://dx.doi.org/10.1081/CSS-100104202.
FAOSTAT (Statistics Division Food and Agriculture Organization of the United
Nations). 2018.
Gebey, T, Berhe K, Hoekstra D and Bogale A. 2012. Rice value chain development in
Fogera woreda based on the IPMS experience. Nairobi, Kenya: ILRI. 23pp.
Gewaily EE, Adel M, Ghoneim and Marvet MA Osman. 2018. Effects of nitrogen levels on
growth, yield and nitrogen use efficiency of some newly released Egyptian rice
[330]
genotypes. Open Agriculture. 2018; 3: 310–318. https://doi.org/10.1515/opag-2018-
0034.
Iqbal Mosud, MahmudaAkter, Jatish Chandra Biswas and Aminul Islam. 2017. Performance
of Nitrogen Phosphorus compound fertilizer in Boro rice. Int J Med Invest 2017; vol 6;
num 1;4-10. http://www.intjmi.com.
Irfan M, MY Memon, JA Shah, M Abbas. 2016. Application of nitrogen and phosphorus in
different ratios to affect paddy yield, nutrient uptake and efficiency relations in rice
(Oryza sativa L.). J. Environ. Agric., 1(2): 79–86.
Landon, JR. 1991. Booker Tropical Soil manual: a handbook for soil survey and agricultural
land evaluation in the tropics and subtropics.
Liu K, Deng J, Lu J, Wang X, Lu B, Tian X and Zhang Y. 2019. High Nitrogen Levels
Alleviate Yield Loss of Super Hybrid Rice Caused by High Temperatures During the
Flowering Stage. Front. Plant Sci. 10:357. doi: 10.3389/fpls.2019.00357.
MoARD. 2010. National Rice Research and Development Strategy of Ethiopia. The Federal
Democratic Republic of Ethiopia, Ministry of Agriculture and Rural Development,
Addis Ababa, Ethiopia. 48 pp.
Molla Haddis and Sofonyas Dargie. 2018. Response of Upland Rice to Nitrogen and
Phosphorus Fertilization on Vertisols of Tigray, Ethiopia. Asian Research Journal of
Agriculture 10(4): 1-7, 2018; ISSN: 2456-561X. DOI: 10.9734/ARJA/2019/47013.
Mulugeta Seyoum and Heluf Gebre-Kidan. 2014. Inherent Properties and Fertilizer Effects
of Flooded Rice Soil. Journal Agronomy 13(2): 72–78.
Murphy HF. 1968. A report on the fertility status of some soils of Ethiopia. Experiment
Bulletin No 44. College of Agriculture, Haileselassie I University, Alemaya,
Ethiopia.
Peng Shaobing, Roland J. Buresh, Jianliang Huang, Xuhua Zhong, Yingbin Zou,
JianchangYang, Guanghuo Wang, YuanyingLiu, Ruifa Hu, Qiyuan Tang, Kehui Cui,
Fusuo Zhang, and Achim Dobermann. 2010. Improving nitrogen fertilization in rice by
site-specific N management. A review. Journal of Agronomy for Sustainable
Development. 30 (2010) 649–656, EDP Sciences, 2010 DOI:10.1051/agro/2010002.
Sah, MK, Shah P, Yadav R, Sah JN and Ranjan R. 2019. Interaction of nitrogen doses and
establishment methods in lowland rice at Parwanipur, Bara, Nepal. Archives of
Agriculture and Environmental Science, 4(1): 113-118, https://dx.doi.org/10.26832/
24566632.2019.0401018.
SAS Institute. 2003. SAS Version 9. 1.2 © 2002-2003. SAS Institute, Inc., Cary, NC.
ShaRada P and Sujathamm P. 2018. Effect of Organic and Inorganic Fertilizers on the
Quantitative and Qualitative Parameters of Rice (Oryza sativa L.)Current agriculture
research journal. ISSN: 2347-4688, Vol. 6, No.(2) 2018, pg. 166-174
Subedi Purushottam, Shrawan Kumar Sah, Santosh Marahattha and Dil Raj Yadav. 2019.
Effects of Need-Based Nitrogen Management and Varieties on Growth and Yield of
Dry Direct Seeded Rice. Pertanika J. Trop. Agric. Sc. 42 (2): 453 - 466 (2019).
Vinod, KK and Sigrid Heuer. 2012. Approaches towards nitrogen- and phosphorus-efficient
rice. AoB PLANTS 2012: pls028; doi:10.1093/aobpla/pls028.
Wang Y, Lu J, Ren T, Hussain S, Guo C, Wang S, Cong R, Li X. 2017. Effects of nitrogen
and tiller type on grain yield and physiological responses in rice. AoB PLANTS 9:
plx012; doi:10.1093/aobpla/plx012.
Yasuhiro Tsujimoto, TovoheryRakotoson, Atsuko Tanaka and Kazuki Saito. 2019.
Challenges and opportunities for improving N use efficiency for rice production in sub-
Saharan Africa, Plant Production Science, DOI: 10.1080/1343943X.2019.1617638.
[331]
Yoshida S. 1981. Fundamentals of rice crop science The International Rice Research
Institute, Los Baños, Laguna, Philippines. 269 Pp.
Yousaf M, Li X, Zhang Z, Ren T, Cong R, Ata-Ul-Karim ST, Fahad S, Shah AN and Lu J.
2016. Nitrogen Fertilizer Management for Enhancing Crop Productivity and Nitrogen
Use Efficiency in a Rice-Oilseed Rape Rotation System in China. Frontiers in Plant Sci.
7:1496. doi: 10.3389/fpls.2016.01496.
Zhao L, Wu L, Wu M and Li Y. 2011.Nutrient uptake and water use efficiency as affected
by modified rice cultivation methods with irrigation. Paddy Water Environment, 9: 25-
32.
[332]
Optimum Nitrogen and Phosphorus Fertilizer
Rates for Upland Rice Production in
North Western Ethiopia
Tilahun Tadesse1*, Zelalem Tadesse1, Habtamu Asega1, and Christian Tafere1
1
Fogera National Rice Research and Training Center, Woreta, Ethiopia
*Corresponding author: E-mail-tilahuntadesse2000@gmail.com
Abstract
An experiment on nitrogen and phosphorous fertilizer rate was conducted on
upland rice at Fogera and Libokemkem woredas in three consecutive cropping
seasons of the years 2015, 2016 and 2017. The treatments were comprised of
factorial combinations of four nitrogen (0,46,92,138 N kg ha -1) and three
phosphorous levels (0,46,92 P2O5 kg ha-1). Data were collected on plant height,
panicle length, and number of total tillers m-2, number of fertile panicles m-2,
thousand seed weight, grain yield, straw yield and harvest index. All collected
data were subjected to analysis of variance. Economic analysis was also carried
out by following CIMMYT (1988) procedures. The results of the experiment
indicated that the main effect of nitrogen application was significantly affecting
plant height, panicle length, total tillers, and number of fertile panicles, grain
yield, and straw yield while that of phosphorous was affecting total tillers,
number of fertile panicles and grain yield. The interaction of nitrogen and
phosphorous were affecting total tillers, number of fertile panicles, grain yield
and straw yield. With regard to the interaction effect, the highest grain yield (5.5
t ha-1) was obtained at 138-46 N-P2O5 kg ha-1. The economic analysis has
further revealed that the combined application of 138-46 N-P2O5 kg ha-1 is the
most profitable treatment. It is thus concluded that the application of nitrogen
and phosphorous fertilizers at rates of 138-46 N-P2O5 kg ha-1 is the best to be
recommended for rainfed upland rice production in the study area and other
similar agro-ecologies.
Keywords: Upland rice, Nitrogen, Phosphorous, profitable.
Introduction
Rice (Oryza sativa L.) is an annual cereal grain and it is the most important
food crops for the world’s population, especially in South Asia, Middle East,
Latin America and West India (Zhao et al., 2011). It is the principal food for
one-third of the world’s population. More than 90% of rice is produced and
consumed in Asia (Subedi et al, 2019). It provides some 700 calories per
person, mostly residing in developing countries. In Ethiopia, rice production
was started three decades ago in the early 1970s and the country has
[333]
reasonable potential to grow various rice types mainly in rainfed lowland,
upland and irrigated ecosystems (Mulugeta and Heluf, 2005). Though rice is
a recent introduction to the country, its importance is well recognized as the
production area coverage of about 10,000 ha in 2006 has increased to over
63,000 ha in 2018 (CSA 2019). The area coverage in domestic rice
production has increased considerably linked with the expansion of
production in the wetland and upland areas with the introduction of suitable
rice varieties for agro-ecologies. In line with the area expansion, the
production levels have been increasing consistently over the years. CSA
(Central Statistical Authority) data indicate that rice production increased
from 71,316.07 tons in 2008 to 171,854.09 tons in 2018 (CSA 2019). The
number of farmers engaged in rice production has also grown year after year.
Rice production has brought a significant change in the livelihood of farmers
and created job opportunities for many citizens in different areas of the
country. Currently, Amhara, Southern Nations Nationalities and Peoples
Region (SNNPR), Oromiya, Somali, Gambella, BeniShangul Gumuz, and
Tigray regions are the rice-producing areas in Ethiopia (MoARD, 2010). The
Amhara region takes the lion’s share of producing the crop and accounted for
74-81% of the area coverage and 78-85% of the production in the years
2016-2018 (CSA 2016, CSA 2017 and 2018). According to the report of
MoARD (2010) the potential rice production area in Ethiopia is estimated to
be about 39,354,190 hectares, of which 5,590,895 ha is highly suitable,
24,910,629 ha is suitable and 8,852,666 is moderately suitable. Most of
Ethiopia’s rice production potential area lies in the western part of the
country.
Upland rice could suitably grow in many parts of the country, the
predominant potential areas are west-central highlands of Amhara Region
(Fogera, Gonder Zuria, Dembia, Takusa, and Achefer), North West lowland
areas of Amhara and Benshangul Regions (Jawi, Pawi, Metema and Dangur),
Gambella Regional state (Abobo and Etang Woredas), South and southwest
Low lands of Southern Nations, Nationalities and Peoples Region (SNNPR)
(Beralle, Weyito, Omorate, Gura Ferda, and Menit), Somali Region (Gode),
Afar and Southwestern highlands of Oromia Region (Illubabora, East and
West Wellega, and Jima Zones) (Dawit Aemu,2015). The national average
yield of rice is about 2.8 t ha-1 (CSA, 2018) which is lower compared to the
world average productivity of 4.6 tones ha-1 (FAOSTAT, 2018). Weeds,
pests, soil nutrient deficiencies and terminal moisture stress are the major
causes of low rice productivity in Ethiopia (MoARD, 2010; Gebey et al.,
2012). Poor soil fertility is among the major factors limiting rice production
in Ethiopia. Nitrogen, phosphorus, and potassium are applied as fertilizers in
[334]
large quantities to rice fields, and a deficiency of either of the nutrient leads
to yield losses (Aamer et al., 2000; Masni and Wasli, 2019; ShaRada et al.,
2018; Subedi et al, 2019). Nitrogen and phosphorus are often cited as the
most limiting nutrients in agricultural soils of Ethiopia (Molla and Sofonyas,
2018). Appropriate fertilizer application is an important management
practice to improve soil fertility and the production of rice (Maneesh et al.
2018). Therefore, a fertilizer experiment was conducted on the upland rice
production of Fogera and Libokemkem Woredas to recommend appropriate
levels of nitrogen and phosphorous rates.

An experiment on nitrogen and phosphorous fertilizer rate was conducted on
upland rice at Fogera and Libokemkem Woredas in three consecutive
cropping seasons of the years 2015, 2016 and 2017. The geographical
coordinates of the experimental site lie between Latitude 11°49’55 North and
Longitude 37° 37′ 40 East. The altitude of the experimental site is 1815
meters above sea level. The study site receives averages mean annual
rainfall, minimum and maximum temperature of 1219 mm, 12.75°C and
27.37°C, respectively. The long-term rainfall data (1986-2017) years
indicated that much of the rainfall appear in July and August (Figure 1).
60 120
Mean Rainfall (mm)
Temperature (°C)
50 100
40 80
30 60
20 40
10 20
0 0
SR18
BK18
HS18
HU16
JM16
AT17
AN17
AN18
MK18
31.93 74.17 15.1
Figure 1. The Rainfall and Temperature condition of the study area for the period 1981-2017
The textural class of the experimental soil was found to be heavy clay with a
pH of 5.87-6.08, which is slightly acidic and it is a preferred range for most
[335]
crops (Table 1). Total nitrogen content was 0.09-0.11%, which is within the
range of low levels (0.02-0.5%) for tropical soils. The organic matter content
of the soil was 2.13-2.39%, which is within a range of medium (2-4%) for
Ethiopian soils as per criteria developed by Murphy (1968). The available P
content of the experimental soil 11.4-25.13 ppm lies in a range of deficiency
(< 20-40mg/kg) for most crops (Landon, 1991).
Table 4. Relevant soil physicochemical properties of the experimental rice field before planting
Soil properties Units Minimum Value Maximum value

Textural class Heavy clay Heavy clay
Chemical properties
pH (H2O) 1:2.5 g soil - 5.87 6.08
Total nitrogen (TN) % 0.09 0.11
Organic carbon (OC) % 1.24 1.33
Organic matter (OM) % 2.13 2.29
Available Phosphorus Ppm 11.4 25.13
The treatments were comprised of factorial combinations of four nitrogen

(0,46,92,138 N kg ha-1) and three phosphorous levels (0,46,92 P2O5 kg ha-1).
The gross size of the experimental plots was 3m x 4m consisting of 15 rows
planted at a spacing of 20 cm apart with a seed rate of 100kg ha-1. The net
plot was made by excluding the left and right outer rows and a plot length of
0.5 m from the top and bottom sides of the plot. The final net plot size was
thus 2.6m x 3m. Data was collected from the net plot area on plant height,
panicle length, number of total tillers m-1 row length, number of fertile
panicles m-1 row length, thousand seed weight, grain yield, straw yield and
harvest index. The rice grain yield and thousand seed weights were adjusted
at 14% standard moisture content. All collected data were subjected to
analysis of variance (ANOVA) using SAS software version 9.0 (SAS-
Institute, 2003). Since the test of homogeneity of variances for each
parameter was non-significant, combined analysis of variance was done over
the years to determine the effects of nitrogen and phosphorous fertilizers on
rice production. Wherever treatment differences are found significant, mean
separation of treatments was executed at F-test and probability levels of 0.01
and 0.05 depending on the results of the ANOVA.
Agronomic efficiency (AE) was calculated to assess the efficiencies of the
applied N rates as follows:
AE= (Gf - Gu / Na) kg rice grain/ kg N fertilizer applied

Where Gf is the grain yield of the fertilized plot (kg), Gu is the grain yield of
the unfertilized plot (kg), and Na is the rate of applied N fertilizer (kg) (Liu
et al, 2019).
[336]
Economic analysis was carried out by following CIMMYT (1988)
procedures by taking all variable costs. The prevailing cost of inputs and
outputs in the year 2019 were considered for the analysis. The cost of Urea
and TSP fertilizers for the stated period at the study area was Birr 13.1 and
14.3, respectively while the price of rice grain and straw was Birr 13.5 and
1.2, respectively.

The analysis of variance indicated that plant height and panicle length were
highly significant (P<0.001) affected by the main effects of nitrogen rates but
not by phosphorous and the interaction of the two (Table 2). Among the
nitrogen rates, the highest plant height was recorded for the highest (138 Kg
ha-1) nitrogen rate while the lowest was recorded at nil N application (Table
3). The highest panicle length exhibited at 138 Kg ha-1 N is statistically at par
with 92 kg ha-1 aN (Table 3), however, the lowest panicle length was
detected at 0kg ha-1 N. In line with the current result, Ghorbannia et al.,
(2012), Shiferaw et al. (2012), Riste et al., (2017) and Molla and Sofonyas,
(2018) reported significant effects of N application on plant height and
panicle length. Shiferaw et al. (2012) observed higher plant height at 92 kg
(113.9 cm) which was statistically at par with 138 kg (113.5 cm) N ha-1.
The increase in plant height in response to the application of N fertilizer was
probably due to enhanced availability of N, which enhanced cell division and
more leaf area resulting in higher photoassimilates and thereby resulted in
more dry matter accumulation (Shiferaw et al. 2012). Similarly, Ghorbannia
et al., (2012) observed that minimum plant height (105.4 cm) was noted for
control (0 kg N ha-1) and a maximum of that (109.3 and 111.3 cm) was for 50
and 100 kg N ha-1. Riste et al., (2017) stated that the highest and significant
panicle length (27.06 cm) was recorded with the application of fertilizer dose
at 60 kg N ha-1 compared to control treatment (No fertilizer). On the other
hand, Molla and Sofonyas, (2018) reported the longest panicles of 20.19 cm
recorded from plots treated with 46 kg N ha-1, while the shortest panicles
were found in the control plots for the main effect of nitrogen. The longest
length of panicle length was observed at 138 kg N ha-1 followed by 46 and
92 kg N ha-1 (Shiferaw et al. 2012).
[337]
Table 2. Analysis of variance (ANOVA) for Plant Height, Panicle Length, Number of Total Tillers,
and Number of Fertile Panicles of upland rice
Source Df Mean Square

PH PL TT/m2 NFP/m2
N 3 1159.23** 24.58 ** 177887.31 ** 170840.91 **
P 2 59.16 NS 0.08 NS 19340.63 * 16846.6*
Rep 2 53.60 NS 0.35 NS 4787.98 NS 2514.45*
Location 4 500.53 ** 9.70 ** 168976.39 ** 137642.37**
N*P 11 42.66 NS 1.75 NS 9913.38 * 7792.7 NS
CV 7.05 6.10 13.24 13.05
PL=Panicle length, Total tillers/m2, NFP=number of fertile panicles/m2
Table 3. Main effects of N and P rates
N (kg/ha) PH PL TT/m2 NFP/m2 GY SY HI

0 57.7D 15.7C 395.5D 387.8D 1.92D 3.08D 38.5D
46 65.4C 16.4B 494.2C 481.0C 3.33C 4.73C 41.3C
92 68.5B 17.3A 546.5B 536.2B 4.59B 5.30B 46.4A
138 71.0A 17.5A 590.7A 576.7A 5.18A 6.24A 45.3A
P2O5 (kg/ha)
0 64.8A 16.7A 492.2B 483.2B 3.66B 4.69A 43.8A
46 66.7A 16.7A 503.4AB 490.4AB 3.93A 5.05A 43.8A
92 65.5A 16.8A 524.6A 512.6A 3.67AB 4.77A 43.5A
PH=Plant height (cm), PL=Panicle length (cm), TT/m2=Total tillers/m2, NFP=number of fertile
panicles/m2, Gy =Grain Yield (t ha-1), SY= Straw yield (t ha-1), HI= Harvest Index (%)
The analysis for number of total tillers and number of fertile panicles has
shown that the main effects of nitrogen highly significantly (P<0.001) and
phosphorous significantly (P<0.005) affected both yield components (Table
1). Moreover, the interaction of N and P significantly (P<0.005) affected the
number of tillers but not the number of fertile panicles (Table 2). The
comparison for the number of total tillers and fertile panicles showed that
among the nitrogen rates the highest values for both were recorded at the
highest (138 kg ha-1) rate while the lowest values were at the 0 kg ha-1 N
(Table 3). Similarly, among the phosphorous rates, the highest values of total
tillers and fertile panicles were exhibited at 92 kg ha-1 P2O5 which are
statistically at par with the vales of 46 kg ha-1 P2O5. The number of total
tillers was significantly responding to the interaction of the two nutrients the
highest number being observed at 138-92 N-P2O5 kg ha-1, while the lower
was at the non (0-0 N-P2O5 kg ha-1) application (Table 5). In conformity with
the current experiment, Kumar et al., (2017) had reported the maximum
number of total tillers m-2 (284.17) with a fertility level of 150: 75: N: P2O5
kg ha-1 and a higher number of effective tillers m-2 at 150: 75 N-P2O5 kg ha-1.
On the other hand, Riste et al., (2017) reported the maximum number of
tillers and the number of panicle m-2 at the rate of 120-90 N-P2O5 kg ha-1.
[338]
The results of Ghorbannia et al., (2012) exhibited that the most fertile tiller
number were obtained at the interaction of 50 kg N ha-1 × 75 and 150 kg P
ha-1, and least tiller number was observed at the interaction of 0 kg N ha-1 × 0
kg P ha-1. According to Molla and Sofonyas, (2018), the application of 69 kg
N ha-1 and 20 kg P ha-1 resulted in the highest seed numbers (126.9) per
panicle of rice than did other rates. Application of fertilizer may have
resulted in optimum levels of nutrients for crop uptake and translocation to
sink thereby expressing superior crop growth and development which
positively reflected in the significantly superior expression of the various
yield attributes (Riste et al., 2017; Kumar et al., 2017). Inferior crop growth
recorded under no fertilizer application may be due to insufficient
availability of nutrients to the crop under the treatment (Riste et al., 2017).
The thousand seeds weight was not affected by the main and interactions of
the nitrogen and phosphorous rates (Table 4). On the contrary, the grain yield
was highly significantly (P<0.001) affected by the main and interaction
effects of nitrogen and phosphorous (Table 4). Concerning the nitrogen rates,
the highest grain yield was shown at 138 Kg ha-1 N while the lowest was
noticed at 0kg ha-1 N. In the case of the phosphorous rates, the highest grain
yield was exhibited at 92 kg ha-1 P2O5 which are statistically at par with the
vales of 46 kg ha-1 P2O5 (Table 3). With regard to the interaction effect, the
highest grain yield (5.5 t ha-1) was obtained at 138-46 N-P2O5 kg ha-1 which
is statistically equivalent with the yield (5.2 t ha-1) of the 138-92 N-P2O5 kg
ha-1 application (Table 5).
Table 4. Analysis of variance (ANOVA) for Thousand Seeds Weight, Grain Yield, Straw Yield and
Harvest Index of upland rice
Mean Square
Source Df TSW GY SY HI
N 3 6.5091435NS 71.8119213** 84.7218519** 297.738958**
P 2 0.1463561NS 2.0826543** 1.5501060NS 46.911670NS
Rep 2 10.8134028* 1.4004861** 1.2509028* 22.257569NS
Location 4 138.9298471** 9.7267320** 72.9610317** 2031.124708**
N*P 11 2.3327546NS 0.4837963** 1.5415741* 56.791458NS
CV 7.1 11.97 23.5 19.27
PL=Panicle length, Total tillers/m2,NFP=number of fertile panicles/m2
[339]
Table 5. Interaction Effects of Nitrogen and Phosphorous Applications on Total Tillers Number,
Grain Yield and Straw Yield
N (kg/ha) P2O5 (kg/ha) Total tillers/m2 Grain Yield (t ha-1) Straw yield (t ha-1)
0 0 380.49D 1.9F 3.0C
0 46 393.82D 1.9F 3.1C
0 92 412.18D 2.0F 3.2C
46 0 436.04CD 3.3E 4.BC
46 46 539.91B 3.5E 5.5AB
46 92 506.71BC 3.2E 4.1BC
92 0 570.58AB 4.6CD 5.5AB
92 46 528.40B 4.9ABC 5.0AB
92 92 540.62B 4.3D 5.5AB
138 0 581.78AB 4.9C 5.6AB
138 46 551.64B 5.5A 6.5A
138 92 638.80A 5.2AB 6.6A
The straw yield was highly significantly (P<0.001) affected by the main
effect of nitrogen but not by phosphorous (Table 4). The interaction of
nitrogen and phosphorous significantly (P<0.005) affected the straw yield
(Table 4). Regarding the nitrogen rates, the highest straw yield was seen at
138 Kg ha-1 N while the lowest was noticed at 0kg ha-1 N. With respect to the
interaction effect, the highest straw yield (6.6 t/ha) was obtained at 138-92
N-P2O5 kg ha-1 which is statistically at par with some of the treatment
combinations (Table 3). The rice harvest index was highly significantly
(P<0.001) affected by the main effect of nitrogen but not by phosphorous and
the interaction (Table 4). The highest harvest index, among the nitrogen
rates, was recorded for 138 Kg ha-1 N which is statistically equivalent with
the value of 92 Kg ha-1 N while the lowest was noticed at 0kg ha-1 N (Table
3). In support of the present finding, Kumar et al., (2017) stated that the grain
and straw yield of rice increased up to application of 150:75 N-P2O5 kg ha-1.
Masni and Wasli, (2019) had also reported that the grain and straw yields of
upland rice were significantly affected and best at 60N and 35kg P ha -1.
Molla and Sofonyas (2018) reported significantly higher grain and straw
yields of upland rice at Tselemti District, north Ethiopia and highest values
were obtained in plots receiving 69 kg N ha-1 and 30 kg P fertilizer ha-1. The
reports of Riste et al., (2017) revealed that paddy and straw yields of rice
were influenced significantly (p < 0.05) under various combinations of N and
P levels. The authors further explained that the treatment where N and P
were integrated at the rate of 120 and 90 kg ha-1 exhibited the highest paddy
yield (4.5 t ha-1) which was 56% more over control. Riste et al., (2017) has
further described that the highest straw yield (9.7t ha-1) was recorded at 150
kg N + 75 kg P2O5 ha-1.
[340]
The better grain and straw yields at the higher rates of N and P nutrients may
be attributed to the fact that application of fertilizer may have resulted in
optimum levels of nutrients for crop uptake and translocation to sink thereby
expressing superior crop growth and development which positively reflected
in the significantly superior expression of the various yield attributes (Riste
et al., 2017). Similar with the observation of highest harvest index at the
current experiment, Worou et al., 2017 had reported higher HI (0.31) of
upland NERICA rice with fertilizer (N at 80 kg ha-1 as urea combined with P
at 80 kg P2O5 ha-1) than without fertilizer (0.21). Higher grain yields in the
fertilizer treatments were associated with a higher harvest index.
The analysis of the Agronomic Efficiency (AE) for the nitrogen indicates
that the maximum AE of 29.74 was exhibited at 46 Kg ha-1 N, then the AE
reduce to 13.92 at 92 Kg ha-1 N, and become lowest (4.32) at 138 Kg ha-1 N
(Table 6). Agronomic efficiency for N is usually higher at a low N rate than
at a high N rate (Gewaily et al., 2018, Yasuhiro et al., 2019). In tropical
Asia, with proper crop and water management, AEN should be typically in
the range of 20–25 kg kg-1 (Yasuhiro et al., 2019). Yoshida (1981) estimated
better agronomic N use efficiency to be 15–25 kg rough rice per kg applied
N in the tropics. Peng et al., (2010) reported that agronomic N use efficiency
was 15 to 18 kg kg-1 N in the dry season in the farmers’ fields in the
Philippines. In China, agronomic N use efficiency was 15–20 kg kg-1 N from
1958 to 1963 and declined to only 9.1kg kg-1 between 1981 and 1983 (Peng
et al., 2010). Since then, agronomic N use efficiency has further decreased in
China because of the increase in N rate (Peng et al., 2010). Generally,
fertilizer N use efficiency of ice at the highest rate is relatively low due to
loss of applied N through leaching, volatilization and denitrification which
necessitate the need for improved N fertilizer practices to reduce
environmental impacts and increase economic benefits of N fertilization
(Fageria and Baligar, 2001). The lower agronomic efficiency at the highest N
rates in the current experiment indicate that emphasis should be given to
efficient nitrogen application methods like the split applications, use of slow
N releasing fertilizer sources and real-time N management to reduce the
wastage of N in the upland rice production system of the study area.
Table 6. Agronomic Efficiency (AE) of rice
N (kg/ha) Grain Yield (Kg/ha) AE

0 1920
46 3288.261 29.74
92 4569.13 13.92
138 5166.087 4.32
[341]
Following the CIMYYT (1988) partial budget analysis method, grain and
straw yield adjustments, calculations of total variable costs (TVC), gross
benefits (GB) and net benefits (NB) were performed (Table 7). Dominance
analysis was carried after arranging the treatments in their order of TVC. A
treatment will be considered as dominated if it has higher TVC but lower NB
than a previous treatment with lower TVC and higher NB (Table 8). Non
dominated treatments were taken out and the marginal rate of return (MRR)
was computed (Table 8). According to the CIMYYT (1988) partial budget
analysis methodology, treatments exhibiting the minimum or more MRR
(>100%) will be considered for the comparison of their NB. The highest NB
(Birr 68,307.5/ha) with an acceptable level of MRR (668.5%) was observed
at 138-46 N-P2O5 kg ha-1 (Table 9). In agreement with the present finding
Irfan et al., (2016) reported that rice genotypes performed efficiently at 120
kg N + 90 kg P2O5 ha-1 where highest paddy yield, net production value and
profit were obtained. The combined application of nitrogen and phosphorous
at 138-46 N-P2O5 kg/ha is the most profitable rate to be recommended for
rice production in Fogera plain
Table 7. Grain and straw yield adjustments, total variable cost, gross and net benefit analysis
N P2O5 TVC GY (t/ha) SY (t/ha) AGY ASY GB NB

(kg/ha) (kg/ha) (Birr/ha) (t/ha) (t/ha) (Birr/ha)
0 0 0.0 1.8867 2.9533 1.69803 2.658 26112.969 26113.0
0 46 1731.1 1.9133 3.0933 1.72197 2.784 26587.359 24856.3
0 92 3462.1 1.9733 3.1867 1.77597 2.868 27417.231 23955.1
46 0 1310.0 3.2533 4.6 2.92797 4.14 44495.595 43185.6
46 46 2386.1 3.5133 5.5133 3.16197 4.962 48640.959 46254.9
46 92 3462.1 3.2267 4.0733 2.90403 3.666 43603.569 40141.5
92 0 2620.0 4.62 5.4467 4.158 4.902 62015.436 59395.4
92 46 3696.1 4.8867 4.9733 4.39803 4.476 64744.569 61048.5
92 92 4772.1 4.26 5.4933 3.834 4.944 57691.764 52919.7
138 0 3930.0 4.8533 5.6267 4.36797 5.064 65044.431 61114.4
138 46 5006.1 5.4533 6.5333 4.90797 5.88 73313.559 68307.5
138 92 6082.1 5.2333 6.5667 4.70997 5.91 70676.631 64594.5
TVC= Total Variable Cost, GY=Grain Yield, SY= Straw Yield, AGY= Adjusted Grain Yield, ASY= Adjusted Straw
Yield, GB= Gross Benefit, NB= Nnet benefit
[342]
Table 8. Dominance Analysis
N (kg/ha) P2O5 TVC

(kg/ha) (Birr/ha) NB DOMINANCE
0 0 0 26,112.97
46 0 1310 43,185.60
0 46 1731.053 24,856.31 D
46 46 2386.053 46,254.91
92 0 2620 59,395.44
0 92 3462.105 23,955.13 D
46 92 3462.105 40,141.46 D
92 46 3696.053 61,048.52
138 0 3930 61,114.43
92 92 4772.105 52,919.66 D
138 46 5006.053 68,307.51
138 92 6082.105 64,594.53 D
D= Dominated
Table 9. Marginal Rate of Return (MRR) Analysis
N (kg/ha) P2O5
(kg/ha) TVC (Birr/ha) NB MRR (%)
0 0 0 26,113
46 0 1310 43,185.60 1303.252
46 46 2386.1 46,254.90 285.2244
92 0 2620 59,395.40 5617.999
92 46 3696.1 61,048.50 153.6196
138 0 3930 61,114.40 28.17443
138 46 5006.1 68,307.51 668.4719
References
Aamer, I, Abbas G and Khaliq A. 2000. Effect of different Nitrogen application Techniques
on the Yield and Yield Components of Fine Rice. International Journal of Agriculture
and Biology 2(3):239–241.
pp.
pp.
(2009 E.C.) May, 2016. ADDIS ABABA, Ethiopia.
(2010 E.C.) April, 2018. ADDIS ABABA, Ethiopia.
(2011 E.C.). April, 2019. ADDIS ABABA, Ethiopia.
[343]
CSA (Central Statistical Agency). 2019. Report on area and production of major crops
(private peasant holdings, meher season). The federal democratic republic of Ethiopia
central statistical agency agricultural sample survey 2018/19 (2011 E.C.). Volume i.
Statistical bulletin 589.
Dawit Alemu. 2015. Rice in Ethiopia: Progress in Production Increase and Success Factors.
6th CARD General Meeting; Ethiopian, Journal at Institute of Agricultural Research, 1-
22.
Fageria, NK and VC Baligar. 2001. 'LOWLAND RICE RESPONSE TO NITROGEN
FERTILIZATION', Communications in Soil Science and Plant Analysis, 32: 9, 1405 —
1429 To link to this Article: DOI: 10.1081/CSS-100104202 URL:
http://dx.doi.org/10.1081/CSS-100104202.
FAOSTAT (Statistics Division Food and Agriculture Organization of the United
Nations). 2018.
Gebey, T, Berhe K, Hoekstra D and Bogale A. 2012. Rice value chain development in
Fogera woreda based on the IPMS experience. Nairobi, Kenya: ILRI. 23pp.
Gewaily EE, Adel M Ghoneim and Marvet MA Osman. 2018. Effects of nitrogen levels on
growth, yield and nitrogen use efficiency of some newly released Egyptian rice
genotypes. Open Agriculture. 2018; 3: 310–318. https://doi.org/10.1515/opag-2018-
0034.
Ghorbannia E, Dastan S, Mobasser HR and Ghanbari-Malidarreh A. 2012. Response of
sensitive to lodging rice cultivar to nitrogen levels in heading stage and phosphorus
rates at the North of Iran. International Research Journal of Biochemistry and
Bioinformatics (ISSN-2250-9941) Vol. 2(6) pp. 142-148, June, 2012 Available online
http://www.interesjournals.org/IRJBB Copyright © 2012 International Research
Journals.
Kumar Sandeep, Sarabdeep Kour, Meenakshi Gupta, Dileep Kachroo1 and Hari Singh.
2017. Influence of rice varieties and fertility levels on performance of rice and soil
nutrient status under aerobic conditions. Journal of Applied and Natural Science 9 (2):
1164 – 1169.
Landon JR. 1991. Booker Tropical Soil manual: a handbook for soil survey and agricultural
land evaluation in the tropics and subtropics.
Liu K, Deng J, Lu J, Wang X, Lu B, Tian X and Zhang Y. 2019. High Nitrogen Levels
Alleviate Yield Loss of Super Hybrid Rice Caused by High Temperatures During the
Flowering Stage. Front. Plant Sci. 10:357. doi: 10.3389/fpls.2019.00357.
Maneesh Bhatt, AP Singh, Veer Singh, DC Kala and Vineet Kumar. 2018. Long-term effect
of organic and inorganic Fertilizers on soil physico-chemical properties of a silty clay
loam soil under rice-wheat cropping system in Tarai region of Uttarakhand. Journal of
Pharmacognosy and Phytochemistry 2019; 8(1): 2113-2118.
Masni, Z and ME Wasli. 2019. Yield Performance and Nutrient Uptake of Red Rice Variety
(MRM 16) at Different NPK Fertilizer Rates. Hindawi International Journal of
Agronomy Volume 2019, Article ID 5134358, 6 pages https://doi.org/10.1155/
2019/5134358.
MoARD. 2010. National Rice Research and Development Strategy of Ethiopia. The Federal
Democratic Republic of Ethiopia, Ministry of Agriculture and Rural development,
Addis Ababa, Ethiopia. 48 pp.
Molla Haddis and Sofonyas Dargie. 2018. Response of Upland Rice to Nitrogen and
Phosphorus Fertilization on Vertisols of Tigray, Ethiopia. Asian Research Journal of
Agriculture. 10(4): 1-7, 2018; Article no.ARJA.47013 ISSN: 2456-561X.
[344]
Mulugeta Seyoum and Heluf Gebre-Kidan. 2014. Inherent Properties and Fertilizer Effects
of Flooded Rice Soil. Journal Agronomy 13(2): 72–78.
Murphy HF. 1968. A report on the fertility status of some soils of Ethiopia. Experiment
Bulletin No 44. College of Agriculture, Haileselassie I University, Alemaya,
Ethiopia.
Peng Shaobing, Roland J. Buresh, Jianliang Huang, Xuhua Zhong, Yingbin Zou,
JianchangYang, Guanghuo Wang, YuanyingLiu, Ruifa Hu, Qiyuan Tang, Kehui Cui,
Fusuo Zhang, and Achim Dobermann. 2010. Improving nitrogen fertilization in rice by
site-specific N management. A review. Journal of Agronomy for Sustainable
Development. 30 (2010) 649–656, EDP Sciences, 2010 DOI:10.1051/agro/2010002.
Riste Khroteü, Gohain T and Noyingthung Kikon.2017. Response of local rice (Oryza sativa
L.) cultivars to recommended NPK fertilizer dose under upland rainfed conditions.
Agric. Sci. Digest., 37 (1) 2017 : 10-15. 0253-150X / Online ISSN:0976-0547.
SAS Institute, 2003. SAS Version 9. 1.2 © 2002-2003. SAS Institute, Inc., Cary, NC.
ShaRada P. and P. Sujathamm. 2018. Effect of Organic and Inorganic Fertilizers on the
Quantitative and Qualitative Parameters of Rice (Oryza sativa L.)Current agriculture
research journal. ISSN: 2347-4688, Vol. 6, No.(2) 2018, pg. 166-174
Shiferaw Nesgea, Heluf Gebrekidan, J.J. Sharma and Tareke Berhe. 2012. Effects of
Nitrogen and Phosphorus Fertilizer Application on Yield Attributes, Grain Yield and
Quality of Rain Fed Rice (NERICA-3) in Gambella, Southwestern Ethiopia. East
African Journal of Sciences (2012) Volume 6 (2) 91-104.
Subedi Purushottam, Shrawan Kumar Sah, Santosh Marahattha and Dil Raj Yadav. 2019.
Effects of Need-Based Nitrogen Management and Varieties on Growth and Yield of
Dry Direct Seeded Rice. Pertanika J. Trop. Agric. Sc. 42 (2): 453 - 466 (2019).
Worou Omonlola Nadine, Thomas Gaiser, Thomas Gaiser, Mouinou Attanda Igué,
Charlemagne Gbemavo and Brice Augustin Sinsin. Responses of Upland NERICA Rice
to Fertiliser Application and Fallow Management in Different Agro-Ecological Zones of
Benin Republic. European Scientific Journal September 2017 edition Vol.13, No.27
ISSN: 1857 – 7881 (Print) e - ISSN 1857- 7431. doi: 10.19044/esj.2017.v13n27p152
URL:http://dx.doi.org/10.19044/esj.2017.v13n27.
Yasuhiro Tsujimoto, Tovohery Rakotoson, Atsuko Tanaka and Kazuki Saito. 2019.
Challenges and opportunities for improving N use efficiency for rice production in sub-
Saharan Africa, Plant Production Science, DOI: 10.1080/1343943X.2019.
1617638.Yoshida (1981).
Yoshida S. 1981. Fundamentals of rice crop science The International Rice Research
Institute, Los Baños, Laguna, Philippines. 269 Pp.
Zhao L, Wu L, Wu M, and Li Y. 2011.Nutrient uptake and water use efficiency as affected
by modified rice cultivation methods with irrigation. Paddy Water Environment, 9: 25-
32.
[345]
[346]
Response of Cotton, (Gossypium hirsutum
L.) Varieties to Different Stages of Topping
Under Irrigated Condition, Ethiopia
Yonas Bekele, Arkebe G/egziabher and Bedane Gudeta
Ethiopian Institute of Agricultural Research (EIAR), Werer Agricultural Research Center,
P.O.Box 2003, Addis Ababa, Ethiopia
Abstract
The objective of this study was to evaluate the impact of topping at different growth
stages on the productivity of cotton and to identify optimum stages of growth for
topping cultivated varieties of cotton under irrigated conditions. The study was
conducted on three cotton varieties during the 2017 and 2018 main cropping seasons
at the experiment site of Werer Agricultural Research Center (WARC). Three cotton
varieties namely, Delta Pine - 90, Weyto - 07 and Werer - 50 and five different growth
stages for topping viz. topping at 50 % flowering stage, topping at 15 days after 50 %
flowering stage, topping at initial boll opening stage, topping at 15 days after initial
boll opening stage, topping at 50 % boll opening stage and no topping treatment
(control) were employed for testing in Randomized Complete Block Design with three
replications. The experiment result showed a significant difference among treatments
for average boll number per plant, plant height, first fruiting branch, seed cotton yield,
and lint yield. There was no significant difference among the treatments for average
boll weight, stem diameter, 100 cotton seed weight and ginning outturn (GOT)
parameters tested in the experiment. The interaction effect between the test varieties
and the different stages of topping was found non-significant for most parameters or
the control treatment gave a better performance as that of topping treatments for some
parameters. Therefore, we will advise to conduct this experiment in the future in other
cotton varieties to know the impact of topping at different cotton growth stages.
Keywords: topping, sympodial branches, variety,
Introduction
Cotton (Gossypium hirsutum L.) is one of the most important commercial
and industrial crops playing a key role in economical and social affairs of the
world. It is considered as a king of fibers, and contributed about 80 percent
of the raw materials requirement for textile industries (Virdia, H.M, 2011).
Despite the severe competition with synthetic fibers, cotton continues to
enjoy a place of prime importance in the textile industry (Kairon et al., 2004;
Martin et al., 1976; Joshi, 1997). Cotton is grown chiefly for its fiber, used in
the manufacture of cloth for the mankind (Singh, 1997). In other words,
cotton lint is the most important seed fiber in the world today and is woven
[347]
into fabrics either alone or combined with other fibers (Purseglove, 1979).
Production of cotton was primarily for home use until Eli Whitney invented
the cotton gin in 1793 (Hoveland, 1980). The invention of saw gin and the
development of the factory system, together with the ease of production and
adaptability to machine manufacture, caused a rapid expansion in the use of
cotton.
In Ethiopia cotton is grown both in irrigated and rain-fed agro-ecologies.

Even if there are so many cotton producers in the country, because of low
production, domestic production cannot feed the need of the country. This is
happened because of low cotton productivity and production system.
Climatic conditions like continuous rains, cloudy weather with high
atmospheric humidity cause excessive vegetative growth at the expense of
fruiting bodies, which leads problems like lodging, boll rot, increased insect
activity and low picking efficiency. Shedding of squares, flowers and bolls
are the major once that can mention at the forefront. In a crop of cotton
shedding of squares, flowers and bolls is a normal phenomenon and all the
squares produced by the plant do not grow into bolls and contribute to yield
(Sudararaj and Thulsidas, 1993; Purseglove, 1979; Arnon, 1972). As a result
of this, only 35-45% of the buds produce mature bolls under usual condition
(Martin et al., 1976; Metcalf and Elkins, 1980; Ustimenko-Bakumovsky,
1983; Arnon, 1972). Conditions that can influence shedding are excess or
lack of moisture, cloudiness, high temperatures, nutrient deficiencies or
imbalance and insect and disease damages (Arnon, 1972; Purseglove, 1979;
Metcalf and Elkins, 1980). The shedding of these fruiting bodies increases
from the center of the main stem to its periphery and from base to top
(Ustimenko Bakumovsky, 1983). One of the effective ways of reducing the
shedding of these fruiting bodies (squares, flowers, and young bolls) is
topping or the removal of the tip of the main stem, alone or together with the
tips of the main branches a few weeks before the bolls begin to open (Arnon,
1972, Ustimenko-Bakumovsky, 1983). The purpose of the topping is to
ensure a redistribution of the nutritive substances in the plant which helps to
check the shedding of fruiting bodies and increase seed cotton yield (Arnon,
1972; Bavilova, 1979; Kerephov, 1982). In the experiments of all Union
Scientific Cotton Research Institute, topping of cotton plants gave an
increment in seed cotton yield from 0.8 to 1.1 t/ha (Bavilova, 1979). In
Central Asia topping of the plants is an important procedure in growing
cotton and the stage of topping ranges from 12-13 to 17-18 sympodial
branches depending upon the growth of the cotton plants (Bavilova, 1979).
So, this research work was designed to identify the appropriate stage for the
[348]
topping of cultivated cotton varieties under irrigated condition of Werer
Agricultural Research Center.

The experiment was conducted at Werer Agricultural Research Center for
two consecutive cropping seasons of 2017 and 2018. It was laid out in
Randomized Complete Block Design (RCBD) with three replications. The
treatments include three cotton varieties, namely Deltapine-90, Werer-50,
Weyto-07 and five stages of topping i.e. topping at 50 % flowering stage,
topping at 15 days after 50 % flowering stage, topping at initial boll opening
stage, topping at 15 days after initial boll opening stage, topping at 50 % boll
opening stage and no topping treatment (control). The spacing used for
sowing was 5m length and 5 rows wide with the spacing of 0.90m between
rows and 0.20m between plants with a total plot size of 22.5 m2. The crop
was planted on 17th and 13th May of 2017 and 2018 respectively. Topping
operations were done on the topmost 3-5 cm. growing point of main stems
clipped by hand as per treatment. Yield attributing characters like plant
height, boll weight, boll number, sympodial branches and other necessary
data were collected from randomly selected 10 plants and the overall
collected data were computed from the inner three rows of each plot. All the
collected data were subjected to DMRT mean separation test procedure in
SAS (Version 9.1).

The average number of bolls per plant in the experiment was 31.4 and it
ranged from 27.3 to 34.1 depending upon the cotton varieties and stages of
topping (Table 2). The figure given in Table 3 represents the actual number
of matured bolls per plant. Even if there is a statistically significant
difference among treatments, the highest boll number per plant 34.1 q/ha was
recorded from Weyto-07 cotton variety with no topping treatment and the
lowest boll number per plant 27.3 was recorded from Werer-50 variety with
topping at 50% flowering stage treatment. It showed that even if there were
statistical differences among treatments, the highest number of bolls per
plant was found from the control treatment. This indicated that the stage of
topping did not affect the number of bolls per plant in this experiment. It
means that the number of bolls per plant showed a negative correlation with
the stage of topping.
[349]
The average boll weight in the experiment was 12.9 gm. and it ranged from
11.5 to 13.9 gm. depending upon the variety and stage of topping (Table 2).
The boll weight did not differ significantly with stages of topping in both
varieties. However, Weyto-07 variety with Topping at 15 days after 50 %
flowering stage interaction gave the least average boll weight of 11.5 gm.
and Deltapine-90 variety with topping at 50 % boll opening stage gave a
better average boll weight of 13.9 gm (Table 4). Nevertheless, even if there
were average boll number differences among treatments, the parameter
average boll weight showed statistically non-significant differences among
treatments.
Table 1. Effect of different stages of topping on number of bolls per plant on different cotton varieties
Variety one (Deltapine – 90) Variety two (Werer -50) Variety three (Weyto-07)
1. No Topping 7. No Topping
2. Topping at 50 % Flowering 8. Topping at 50 % Flowering 13. No Topping
Stage Stage 14. Topping at 50 % Flowering
3. Topping at 15 days After 50 % 9. Topping at 15 days After 50 Stage
Flowering Stage % Flowering Stage 15. Topping at 15 days After 50 %
4. Topping at Initial Boll opening 10. Topping at Initial Boll opening Flowering Stage
Stage Stage 16. Topping at Initial Boll opening
5. Topping at 15 days After Initial 11. Topping at 15 days After Stage
Boll Opening Stage Initial Boll Opening Stage 17. Topping at 15 days After Initial
6. Topping at 50 % Boll Opening 12. Topping at 50 % Boll Boll Opening Stage
Stage Opening Stage 18. Topping at 50 % Boll Opening
Stage
[350]
The average plant height in the experiment was 57.8 cm. and it ranged from
41.1 cm. to 81.0 cm. depending upon the variety and stage of topping (Table
2). Since the stage of topping was done in different plant growth stages, the
statistical difference between plant heights was recorded among treatments.
Therefore, the lowest plant height 41.1 cm and the highest plant height was
81.0 cm was recorded from Deltapine-90 variety with topping at 50%
flowering stage and Werer-50 with no topping treatments respectively (Table
5).
Table 2. Effect of different stages of topping on average boll weight (gm.) on different cotton varieties
Variety one Variety two Variety three (Weyto-07)

(Deltapine – 90) (Werer -50) 13. No Topping
1. No Topping 7. No Topping 14. Topping at 50 % Flowering
2. Topping at 50 % Flowering 8. Topping at 50 % Flowering Stage
Stage Stage 15. Topping at 15 days After 50
3. Topping at 15 days After 50 9. Topping at 15 days After 50 % Flowering Stage
% Flowering Stage % Flowering Stage 16. Topping at Initial Boll opening
4. Topping at Initial Boll 10. Topping at Initial Boll Stage
opening Stage opening Stage 17. Topping at 15 days After
5. Topping at 15 days After 11. Topping at 15 days After Initial Boll Opening Stage
Initial Boll Opening Stage Initial Boll Opening Stage 18. Topping at 50 % Boll
6. Topping at 50 % Boll 12. Topping at 50 % Boll Opening Stage
Opening Stage Opening Stage
[351]
Table 3. Main effect of the different stages of topping on cotton varieties
PH SCY
Source of Variation FFB SD NBPP ABW (gm.) GOT LY HCSW (gm.)
(cm.) (q/ha)
Deltapine – 90 (V1) 48.9 c 5.4 c 1.16 b 31.4 a 13.4 a 36.3 a 38.0 a 13.8 a 8.4 b
Werer – 50 (V2) 65.1 a 6.0 a 1.24 a 30.6 a 13.3 a 32.7 b 37.7 a 12.3 b 8.5 ba
Weyto – 07 (V3) 56.3 b 5.6 b 1.21 a 32.3 a 12.1 b 35.5 a 36.8 b 13.1 a 8.6 a
Mean 57.8 5.7 1.2 31.4 12.9 34.9 37.5 13.0 8.5
LSD 5% 2.4 0.2 0.04 ns 0.8 1.9 0.6 0.7 0.1
No Topping (Control) (T1) 71.6 a 6.0 a 1.2 a 33.5 a 12.5 a 37.6 a 37.8 a 14.2 a 8.4 a
Topping at 50 % Flowering Stage (T2) 47.8 d 5.6 cb 1.2 a 27.8 c 12.7 a 31.6 d 37.1 a 11.8 d 8.6 a
Topping at 15 Days After 50 % Flowering (T3) 51.2 d 5.4 c 1.2 a 29.7 bc 12.8 a 33.0 dc 37.4 a 12.3 dc 8.5 a
Topping at Initial Boll Opening Stage (T4) 55.8 c 5.6 cb 1.2 a 31.2 13.0 a 34.3 bc 37.8 a 12.9 bc 8.5 a
Topping at 15 Days After Initial Boll Opening (T5) 57.8 c 5.8 b 1.2 a bacba
33.1 13.2 a 35.7 bac 37.5 a 13.3 bac 8.5 a
Topping at 50 % Boll Opening Stage (T6) 62.6 b 5.6 cb 1.2 a 33.3 ba 13.3 a 36.8 ba 37.6 a 13.8 ba 8.5 a
Mean 57.8 5.7 1.2 31.4 12.9 34.9 37.5 13.0 8.5
LSD 5% 3.4 0.3 NS 3.7 NS 2.7 NS 1.0 NS
CV % 8.8 6.8 7.6 17.5 13.1 11.5 3.5 11.6 3.4
PH = Plant height (cm.) FFB = First fruiting branch SD = Steam diameter

NBPP = Number of bolls per plant ABW = Average boll wt. (gm.) “ SCY = Seed cotton yield (q/ha)
GOT = Gin out turn LY = Lint yield HCSW = 100 cotton seed wt. (gm.)
 Mean separated by DMRT and columns represented with same letter are not significant at 5%
[352]
Table 4. Interaction Effect of Response of cotton (Gossypium hirsutum L.) varieties to different stage of topping under irrigated condition
Tr. PH ABW SCY HCSW

Treatments FFB SD NBPP GOT LY
No (cm.) (gm.) (q/ha) (gm.)
1 DP-90 + No Topping 59.3 5.6 1.1 33.3 12.7 38.9 38.0 14.7 8.5
2 DP-90 + Topping at 50 % Flowering Stage 41.1 5.5 1.2 27.6 13.2 33.3 37.4 12.5 8.5
3 DP-90 + Topping at 15 days After 50 % Flowering Stage 42.9 5.2 1.1 29.5 13.8 34.9 38.3 13.3 8.4
4 DP-90 + Topping at Initial Boll opening Stage 47.5 5.1 1.2 32.3 13.6 35.3 38.7 13.6 8.5
5 DP-90 + Topping at 15 days After Initial Boll Opening Stage 50.1 5.5 1.2 33.2 13.3 37.2 38.1 14.0 8.3
6 DP-90 + Topping at 50 % Boll Opening Stage 52.6 5.2 1.2 33.1 13.9 38.5 37.8 14.5 8.3
7 Werer -50 + No Topping 81.0 6.7 1.3 33.0 13.0 35.6 37.6 13.3 8.5
8 Werer -50 + Topping at 50 % Flowering Stage 52.7 5.7 1.2 27.3 13.1 29.8 37.4 11.2 8.7
9 Werer -50 + Topping at 15 days After 50 % Flowering Stage 56.9 5.4 1.2 28.5 13.1 30.9 37.6 11.7 8.5
10 Werer -50 + Topping at Initial Boll opening Stage 64.2 6.3 1.2 30.2 13.2 32.3 38.1 12.3 8.5
11 Werer -50 + Topping at 15 days After Initial Boll Opening Stage 65.7 6.1 1.3 32.1 13.9 33.3 37.7 12.5 8.5
12 Werer -50 + Topping at 50 % Boll Opening Stage 70.0 6.0 1.2 32.7 13.5 34.3 37.6 12.9 8.6
13 Weyto - 07 + No Topping 74.3 5.8 1.1 34.1 11.7 38.4 37.8 14.5 8.3
14 Weyto - 07 + Topping at 50 % Flowering Stage 49.7 5.6 1.2 28.5 11.9 31.8 36.6 11.7 8.6
15 Weyto - 07 + Topping at 15 days After 50 % Flowering Stage 53.7 5.6 1.3 31.0 11.5 33.6 36.0 12.0 8.7
16 Weyto - 07 + Topping at Initial Boll opening Stage 55.7 5.3 1.3 32.2 12.3 35.4 36.5 12.9 8.7
17 Weyto - 07 + Topping at 15 days After Initial Boll Opening Stage 57.6 5.6 1.2 33.9 12.5 36.5 36.8 13.4 8.7
18 Weyto – 07 + Topping at 50 % Boll Opening Stage 65.1 5.6 1.3 34.0 12.5 37.6 37.3 14.0 8.6
Mean 57.8 5.7 1.2 31.4 12.9 34.9 37.5 13.0 8.5
LSD 5% 3.4 0.3 NS 3.7 NS 2.7 NS 1.0 NS
CV % 8.8 6.8 7.6 17.5 13.1 11.5 3.5 11.6 3.4
[353]
The reason for these plant height differences was happened because of the topping
of the crop was done in different plant growth stages. Since Deltapine-90 cotton
variety has short plant height nature and topping at 50% flowering stage was the
early stage for the plant growth, due to these reasons the lowest plant height 41.1
cm. was recorded from Deltapine-90 variety with Topping at 50% flowering stage.
Contrarily, since Werer-50 cotton variety has tall plant height nature as that of the
two varieties and no topping operation was done for this variety, the highest plant
height 81.0 cm. was recorded from Werer-50 variety with no topping treatment.
Therefore, even if plant height was showed statistically significant difference
among treatments; the highest 81.0 cm. was recorded from the control treatment
(Werer-50 variety with no topping treatment). Generally, the above observation
showed that early topping restricts plant growth.
Table 5. Effect of different stages of topping on plant height (cm.) on different cotton varieties
Variety one Variety two (Werer -50) Variety three (Weyto-07)

(Deltapine – 90)
1. No Topping 7. No Topping 13. No Topping
2. Topping at 50 % Flowering Stage 8. Topping at 50 % Flowering Stage
3. Topping at 15 days After 50 % 14. Topping at 50 % Flowering Stage
9. Topping at 15 days After 50 % 15. Topping at 15 days After 50 %
Flowering Stage
Flowering Stage Flowering Stage
4. Topping at Initial Boll opening Stage
5. Topping at 15 days After Initial Boll 10. Topping at Initial Boll opening 16. Topping at Initial Boll opening Stage
Opening Stage Stage 17. Topping at 15 days After Initial Boll
6. Topping at 50 % Boll Opening Stage 11. Topping at 15 days After Initial Opening Stage
Boll Opening Stage 13. Topping at 50 % Boll Opening Stage
12. Topping at 50 % Boll Opening
[354]
The first fruiting branch is one of the collected data for this experiment. This
parameter shows statistically non-significant differences among treatments.
However, the average first fruiting branch was 5.7 and the lowest 5.1. The first
fruiting branch data was recorded from Deltapine-90 variety with topping at the
initial boll opening stage and the highest first fruiting branch data 6.7 was
recorded from Werer-50 cotton variety with no topping treatment (Table 2).
However, even if there was a statistical difference between treatments, the highest
data was found from the control treatment (Table 6). So, the result indicated that
topping did not affect the first fruiting branch in this experiment.
Table 6. Effect of different stages of topping on the first fruiting branch on different cotton varieties
Variety one (Deltapine – 90) Variety two (Werer -50) Variety three (Weyto - 07)

2. Topping at 50 % Flowering Stage 8. Topping at 50 % Flowering 14. Topping at 50 % Flowering Stage
3. Topping at 15 days After 50 % Stage 15. Topping at 15 days After 50 %
Flowering Stage 9. Topping at 15 days After 50 %
Flowering Stage
4. Topping at Initial Boll opening Flowering Stage
Stage 10. Topping at Initial Boll opening 16. Topping at Initial Boll opening Stage
5. Topping at 15 days After Initial Stage 17. Topping at 15 days After Initial Boll
Boll Opening Stage 11. Topping at 15 days After Initial Opening Stage
6. Topping at 50 % Boll Opening Boll Opening Stage 18. Topping at 50 % Boll Opening Stage
Stage
Stage
[355]
Stem diameter, 100 cotton seed weight and ginning outturn (GOT) parameters
were collected to assess the effect of topping on the above-listed varieties.
However, all the three cotton varieties showed statistically non-significant
differences among treatments (Table 2). Therefore, it indicated that topping did
not affect the above-listed parameters (Steam diameter, 100 cotton seed weight
and Ginning out turn (Table 7, 8 and 9).
Table 7. Effect of different stages of topping on steam diameter (cm.) on different cotton varieties
Stem Diameter
(cm.)

2. Topping at 50 % Flowering 8. Topping at 50 % Flowering 14. Topping at 50 % Flowering
Stage Stage Stage
3. Topping at 15 days After 50 % 9. Topping at 15 days After 50 15. Topping at 15 days After 50 %
Flowering Stage % Flowering Stage Flowering Stage
10. Topping at Initial Boll opening
4. Topping at Initial Boll opening 16. Topping at Initial Boll opening
Stage
Stage 11. Topping at 15 days After Stage
5. Topping at 15 days After Initial Initial Boll Opening Stage 17. Topping at 15 days After Initial
Boll Opening Stage 12. Topping at 50 % Boll Boll Opening Stage
6. Topping at 50 % Boll Opening Opening Stage 18. Topping at 50 % Boll Opening
Stage Stage
[356]
Table 8. Effect of different stages of topping on 100 cotton seed weight (gm.) on different cotton varieties

2. Topping at 50 % Flowering Stage 8. Topping at 50 % Flowering Stage 14. Topping at 50 % Flowering Stage
3. Topping at 15 days After 50 % 9. Topping at 15 days After 50 % 15. Topping at 15 days After 50 %
Flowering Stage Flowering Stage Flowering Stage
4. Topping at Initial Boll opening 10. Topping at Initial Boll opening 16. Topping at Initial Boll opening
Stage Stage Stage
5. Topping at 15 days After Initial Boll 11. Topping at 15 days After Initial 17. Topping at 15 days After Initial
Opening Stage Boll Opening Stage Boll Opening Stage
6. Topping at 50 % Boll Opening 12. Topping at 50 % Boll Opening 18. Topping at 50 % Boll Opening
Stage Stage Stage
[357]
Table 9. Effect of different stages of topping on ginning out turn on different cotton varieties

Flowering Stage 9. Topping at 15 days After 50 % Flowering Stage
4. Topping at Initial Boll opening Stage Flowering Stage 16. Topping at Initial Boll opening Stage
5. Topping at 15 days After Initial Boll 10. Topping at Initial Boll opening 17. Topping at 15 days After Initial Boll
Opening Stage Stage Opening Stage
6. Topping at 50 % Boll Opening Stage 11. Topping at 15 days After Initial 18. Topping at 50 % Boll Opening Stage
Boll Opening Stage
Stage
The average seed cotton yield depending upon the variety and stages of topping
was 34.9 q/ha and it ranged from 29.8 to 38.9 q/ha. According to the data
illustrated in Table 2, significantly higher yield (38.9 q/ha) was obtained from
Deltapine-90 cotton variety with no topping treatment and the least seed cotton
yield 29.8 q/ha was found from Werer-50 cotton variety with topping at 50%
flowering stage. The data showed that, in all these three cotton varieties, if the
stage of topping delayed, the yield advantage becomes increased. This indicated
that late topping gave better yield as that of earlier topping. However, even if the
statistical analysis showed statistical differences among treatments, the highest
yield was recorded from the control treatment (Table 10). Thus, based on the
above analysis, since no topping treatment gave the highest yield, topping is not
important in this experiment for the listed varieties.
[358]
Table 10. Effect of different stages of topping on seed cotton yield (q/ha) on different cotton varieties
Variety two
Variety one (Deltapine – 90) (Werer -50) Variety three (Weyto - 07)

2. Topping at 50 % Flowering 8. Topping at 50 % Flowering 14. Topping at 50 % Flowering Stage
Stage Stage 15. Topping at 15 days After 50 %
3. Topping at 15 days After 50 % 9. Topping at 15 days After 50 % Flowering Stage
Flowering Stage Flowering Stage
4. Topping at Initial Boll opening 16. Topping at Initial Boll opening
Stage Stage
Stage
5. Topping at 15 days After Initial 11. Topping at 15 days After Initial 17. Topping at 15 days After Initial
Boll Opening Stage Boll Opening Stage Boll Opening Stage
6. Topping at 50 % Boll Opening 12. Topping at 50 % Boll Opening 18. Topping at 50 % Boll Opening
Stage Stage Stage
Lint yield in this experiment showed a statistically significant difference among

treatments. The average lint yield was 13.0 % and it ranged from 11.2 to 14.7. The
highest 14.7 % lint yield was recorded from Deltapine-90 variety with no topping
treatment and the lowest lint yield 11.2 % was observed from Werer-50 cotton
variety with topping at 50% flowering stage (Table 2). Similar to that of seed
cotton yield, if the stage of topping delayed, the lint yield potential also showed
better advantage as that of the earlier topping treatments. However, even if there is
a statistical difference among treatments, the highest lint yield was observed from
no topping treatment (Table 11). This indicated that topping has a negative
advantage for lint yield in this experiment.
[359]
Table 11. Effect of different stages of topping on lint yield on different cotton varieties
Variety two
Variety one (Deltapine – 90) (Werer -50) Variety three (Weyto-07)

Flowering Stage 9. Topping at 15 days After 50 % Flowering Stage
Flowering Stage
4. Topping at Initial Boll opening Stage 16. Topping at Initial Boll opening
5. Topping at 15 days After Initial Boll Stage Stage
Opening Stage 11. Topping at 15 days After Initial 17. Topping at 15 days After Initial Boll
6. Topping at 50 % Boll Opening Stage Boll Opening Stage Opening Stage
12. Topping at 50 % Boll Opening 18. Topping at 50 % Boll Opening
Stage Stage
Conclusions and Recommendations

This experiment was undertaken to evaluate the effect of topping on different crop
maturity stages (Topping at 50 % flowering Stage, topping at 15 days after 50 %
flowering stage, topping at initial boll opening stage, topping at 15 days after
initial boll opening stage, topping at 50 % boll opening stage and no topping
treatment (Control) on three cotton varieties (Deltapine-90, Werer-50 and Weyto-
07). It is known that the main objective of topping in a cotton crop is that to get
better plant architecture, to increase penetration of sunlight through the whole
plant, to decrease insect and disease attack which will be resulted due to extended
young branching habits, to increase photosynthetic efficiency which will be
resulted due to reduced foliage and lodging. However, the interaction effect
between varieties and stages of topping showed non-significant differences for
[360]
most parameters. However, the control treatment showed better performance as
that of topping treatments for some parameters. Therefore, we will advise to
conduct this experiment in the future in other cotton varieties to know the impact
of topping at different cotton growth stages.
Acknowledgments
The authors would like to acknowledge the Ethiopian Institute of Agricultural
Research (EIAR), Werer Agricultural Research Center (WARC) for financial
support and facilitation. The authors are also grateful for other researchers and
technical assistance of the national cotton research project for managing this trial.
References
Arnon, I. 1972. Crop production in dry regions. Vol. 2. Systematic treatments of the principal
crops. Leonard Hill, London. 683 p.
Basnet, KB. 1984. Response of cotton varieties to different levels of nitrogen and salinity in the old
irrigated zone of Hunger Steppe. Ph.D. Dissertation, Patrice Lumumba Peoples Friendship
University, Moscow. 227 p.
Bavilova, PP. 1979. Crop production. Fourth edition. Kolos, Moscow. 516 p.
Bhatt, VS. 1992. Handbook of agriculture. ICAR, New Delhi, India. 532 p.
FAO. 1998. FAO yearbook – 1997. FAO statistics series No. 142, FAO, Rome.
Gomez, KA and Gomez AA. 1983. Statistical procedures for agricultural research. An
international rice research institute book. A Wiley Interscience Publication, John Wiley and
Sons, New York.
Hoveland, CS. 1980. Crop quality, storage and utilization. The American Soc. Agron. & Crop Sci.
Soc. America, USA. 276 p.
Joshi, M. 1997. Hybrid cotton in India. Kalyani Publisher, India. 190 p.
Kairon, MS, Laise DB and Venugopalam MV. 2004. Cotton. In: R. Prasad (ed.) Field Crops
Production, ICAR, New Delhi, India. pp. 646-674.
Kerephov, KN. 1982. Biological bases of crop production. Higher School, Moscow. 406 p.
Martin, J. H., W. L. Leonard and D. L. Stamp. 1976. Principles of field crop production (3rd
ed.). Macmillan Publishing Co., Inc. New York. 1118 p.
Metcalf, DS and Elkins DM. 1980. Crop production: principles and practices (4th ed.). Macmillan
Publishing Co., Inc. New York. 774 p.
Poonia, BL, Singh RP and Jain NK. 2002. Response of upland cotton cultivars to different dates
of planting. Indian J. Agric. Sci. 72(3):171-73.
Purseglove, JW. 1979. Tropical crops Dicotyledons. Longman Group Ltd., London. 719 p.
Sharma, D., N. N. Sharma and S. R. Paul. 1997. Performance of upland cotton varieties for
early sowing conditions in hill zone of Assam. Indian J. Agron. 42(2):377-80.
Singh, C. 1997. Modern techniques of raising field crops. Oxford and IBH Publishing Co., Pvt.
Ltd., New Delhi, India. 523 p.
Sundararaj, DD and Thulasidas G. 1993. Botany of field crops. Macmillan Ltd., India. 583 p.
Ustimenko-Bekumovsky, GV. 1983. Plant growing in tropics and sub-tropics. Mir Publisher,
Moscow. 391 p.
[361]
Coffee, Horticulture, Spices,
Coffee
Assessment of the Effectiveness of Coffee

De-mucilager and Driers for Physical and
Sensorial Coffee Quality
Mikru Tesfa*1 and Nigussie Mekonen1
1
Jimma Agricultural Research Center (JARC). P.O. Box 192, Jimma, Ethiopia
* Corresponding author: mikrutesfa27@gmail.com
Abstract
Coffee Arabica is an essential commodity to the livelihood of millions of Ethiopians
and its quality has critical importance to the coffee industry. The objective of the
present study was to evaluate the effect of newly introduced coffee post-harvest
machinery and driers on coffee raw, cup and total quality as compared to the
conventional processing methods. Different methods, such as normal fermentation
(conventional and drying in sun, normal fermentation and drying in artificial drier,
normal fermentation dried in poly tunnel drier, demucilager pulped and dried in sun,
demucilager pulped and dried in artificial drier and demucilager pulped dried in poly
tunnel were used for coffee sample preparation. The experiment was conducted
during the 2017/18 and 2018/19 cropping season at Limmu Kossa coffee plantation.
Both raw and cup quality parameters were evaluated by a team of certified panelists
at Jimma Agricultural Research Center (JARC) coffee processing and quality analysis
laboratory and analyzed using completely Randomized Design (CRD). Physical coffee
quality parameters such as Bean size, Bean shape and make, color and odor showed
no statistically significant (P>0.05) difference for all preparation methods. The
highest values of cup quality parameters like Acidity (AC), astringency (AS),
bitterness (BI), body (BO), flavor (FL) and overall quality (OAQ) were recorded for
coffee samples prepared using demucilager pulper and dried in tunnel drier. Samples
pulped by demucilager and dried in poly tunnel drier exhibited the highest values,
35.67, 47.50 and 83.17, for raw, cup and total quality, respectively. It seems
important to further study the biochemical composion of samples prepared using in
such new introduced demucilager and driers in the future.
Key words: Artificial drier, Demucilager, Poly tunnel drier
Introduction
Coffee is the major source of foreign currency for Ethiopia and contributes more
than 35% of the total export earnings (FAO/WFP, 2008). Thus, it is a cornerstone
in the export economy of the country and it supports directly or indirectly the
[363]
livelihood of some 15 million people (EEA, 2001). Coffee processing is a very
important activity in the coffee production system and plays a crucial role in
quality determination (Mburu, J.K, 1999). The quality of Ethiopian coffee is
determined by two main factors, namely geographic origin and post-harvest
processing techniques (Musebe R. et al., 2008).
Physical and organoleptic qualities are the most important parameters in the world
coffee trade. It is estimated that 40% of the quality of coffee is determined in the
field (Agricultural practices), 40% at post-harvest primary processing, and 20% at
export processing and handling, including storage (Musebe, R.et al, 2008). This
underscores the importance of primary processing in enhancing the quality and
value of Coffee. For Ethiopian coffee, natural fermentation is recommended as it
improves both raw and roast qualities (Behailu, W.S., et al, 2008). However,
natural fermentation is time-consuming and costly, causes weight loss and is
laborious compared to demucilager. Moreover, natural fermentation requires strict
supervision on adjustment of temperature and time of fermentation, use of good
quality equipment and a substantial amount of water (with acceptable quality) in
order to get a quality coffee product. On the other hand, demucilager is found to
be less costly, less time consuming and less laborious and does not require close
supervision. Moreover, demucilager was reported to avoid coffee weight loss
compared to natural fermentation. Very recently, different coffee processing
stations are employing demucilager in the major coffee growing areas of Ethiopia
taking into account its advantage in handling and cost-effectiveness. According to
some studies in Kenya, Costa Rica and Colombia, both natural fermentation and
demucilager showed no significant influence on the final coffee quality
(Wintegens, J.N., 2004).
However, information on the effects of high-tech processing machineries

(demucilager) on coffee quality of improved Ethiopian coffee varieties is scanty.
Coffee quality problems are mainly associated with improper post-harvest
processing and handling practices such as drying on bare ground, improper wet
processing, storage and transportation, poor agronomic practices like uncontrolled
shade level, lack of stumping, pruning and weeding and poor harvesting practices,
such as stripping and collecting dropped fruits from the ground (D., Nure, 2008).
The risk of contamination is greater during fermentation due to the miss-use of

optimum fermentation time, lack of clean water, bean-soil contact during sun
drying and bean re-humidification under high relative humidity conditions. These
problems could be eliminated with demucilager pulper, optimum fermentation
time and the use of unpolluted water and also artificial drying. However, artificial
drying is an operation of high energy demand and it has been shown that if the
bean’s temperature exceeds 45°C during drying, the coffee’s quality would be
harmed (Sfredo et al., 2005). Among the problems in post-harvest processing and
[364]
handling in the study area, wet coffee washing after fermentation to facilitate the
removal of mucilage before soaking pulped coffee for 24-48 hours to be fermented
naturally in a concrete storage tank is the major one that affects coffee quality.
The fermentation is followed by washing through scrubbing the soaked beans
against the concrete floor, which resembles 3 to 4 rinses of clean water to remove
all traces and decomposed products of the mucilage. This process is time-
consuming and laborious. (Abasanbi A., 2010). Currently, different new machines
remove the skin and pulp from coffee cherries leaving viscous mucilage adhering
to parchment. However, there is limited information on the effect of the new
technologies on the physical and sensory characteristics of coffee brew. Therefore,
the objective was to investigate the effect of newly introduced coffee post-harvest
machineries and driers on a coffee raw, cup and total quality as compared to the
conventional processing methods.

The experiment was conducted in Limmu Kossa sites of the Limmu Coffee
Plantation Development Enterprise. Limmu Kossa woreda is located at 7°50 N
latitude and 36° 53 E longitudes in the Southwest part of Ethiopia and 421 Km far
from Addis Ababa. It has an area of 2770.5 Km2. The area has annual rainfall
varies between 1200 and 1850 mm with mean minimum and maximum
temperatures of 7°C-12°C and 25°C-30°C, respectively.
Six coffee bulk prepared samples by different preparation methods. For uniform
processing, only well ripe red coffee cherries were selectively harvested at peak
harvesting time. Six different samples were prepared such as normal fermentation
(conventional) dried by the sun, normal fermentation dried by an artificial drier,
normal fermentation dried by poly tunnel drier, demucilager pulped dried by sun,
demucilager pulped dried by Artificial drier and demucilager pulped dried by poly
tunnel were performed and analyzed by CRD ( Completely Randomized Design).
At Limmu Kossa coffee plantation farm six coffee samples of 8 kg per sample
were prepared based on our treatment.
Data Collection:
Raw quality Evaluation data: During raw quality analysis, 100g, of the green
bean was used for each sample and their screen bean size, shape and make, color
and odor were evaluated by a team of JARC cuppers according to JARC coffee
quality liquoring laboratory manual ( Abrar S. and Negussie M. 2015).
Roasting and Grinding: 100g of green coffee bean per sample was used for
roasting. Each sample was separately roasted until the bean attain medium brown
roast color for (8 minutes on average) by using a roaster machine (Probat BRZ6,
Germany) with a cylinder heat of 180-200°C. When roasting was considered
[365]
medium, the coffee beans were tipped out into the cooling tray and allowed to
cool for about 4 minutes. After cooling and before grinding, the samples were
blown to remove the loose silver skins (chaff). Half of the roasted bean from total
roasted sample per each sample was grounded immediately before brewing using
an electrical coffee grinder (MahlKonig, Germany) with middle adjustment to get
medium size coffee powder.
Brew Preparation: Three cups of the same volume (180ml) and size were used
for each sample to assess cup quality medium level ground coffee powders
weighing 8g were placed in each cup for liquoring. Boiled water was poured into
half volume of the cup containing 8g of coffee powder, and thereafter the volatile
aromatic quality and intensity parameters were recorded by sniffing by well
trained and licensed coffee quality testers. Then, the cups were filled with boiled
water to the brim and left to cool for 3 minutes. Then, the foam of the beverage of
each cup is skimmed off with a spoon. Finally, the brew was made ready for
panelists within 8 minutes.
Cup (Organoleptic) quality evaluation data: This was also carried out at JARC
coffee liquoring laboratory using the standard procedures. Once the beverage was
cooled to around 60°C (palatable temperature), then cup tasting was carried out by
the three trained panelists and licensed Q graders. Three cups per sample were
prepared for a tasting session arranged randomly. Samples were chosen randomly
for testing.
Testers noted sensorial quality attributes following the format and procedure
developed at Jimma Agricultural Research Center (Abrar S. and Negussie M.
2015). Aroma (aromatic quality and intensity), Astringency and Bitterness were
scored using scales ranging from 0 to5. Body, acidity, flavor and overall quality
were scored using scales ranging from 0 to10. Typical flavor (Typicity) was
assessed as an after taste aromatic quality that could vary from winy to flowery
(winy, fragrant, floral, citrus, mocha, fruity and spicy). The mean of each variable
by the panel was used for statistical analysis.
Statistical Analysis: Analysis of variance (ANOVA) was performed to determine

the effects of recently introduced coffee post-harvest machineries on coffee
qualities. Least Significant Difference (LSD) was used for the mean separation.
[366]
Physical (raw) Coffee Quality Attributes
Physical coffee quality attributes (bean size, bean shape and make, color and odor)
are presented in Table 1. There was no significant difference (P>0.05) among the
preparation methods for bean size, bean shape and make, color and odor. The raw
total quality also showed no significant (P>0.05) among samples prepared in
different preparation methods. The lowest value of odor (8.67) was recorded for
samples prepared by normal (conventional) fermentation and dried using
artificially drier.
Table 1. Effect of recently introduced post-harvest machineries and drier on physical (raw) quality attributes of coffee.
Mean of two years observation
Green bean coffee quality parameters

Preparation method Bean size screen Shape and make Color Odor RT (40%)
NFDS 95.50a 12.17a 12.00a 9.67a 33.83a
NFDA 96.50a 12.67a 12.17a 8.67b 33.50a
NFTd 97.20a 12.00a 12.33a 10.00a 34.33a
DPDS 95.50a 12.83a 13.00a 9.67a 35.50a
DPDA 98.33a 12.83a 12.83a 10.00a 35.67a
DPTd 95.00a 12.67a 12.33a 10.00a 35.00a
CV (%) 2.06 5.32 5.99 4.22 3.73
LSD (0.05) 1.60 0.54 0.60 0.33 1.05
Mean values followed by the same letter within a column are not significantly different (P>0.05).
NFDS, NFDA and NFTd= Normal (conventional) fermentation dried in sun, dried
artificially and dried using tunnel drier, respectively, DPDS, DPDA and DPTd =
Demucilager pulped and dried in sun, dried artificially and dried using tunnel
drier, respectively. RT= Raw total
Organoleptic (Cup) Quality Attributes

Results on the effect of recently introduced coffee post-harvest machineries and
drier on Organoleptic (cup) quality attributes are presented in Table 2. The highest
values of cup quality parameters, like acidity (AC), astringency (AS), bitterness
(BI), body (BO), flavor (FL) and overall quality (OAQ) were recorded for coffee
samples prepared using demucilager pulper and dried in poly tunnel drier (Table
2). The highest (47.5) and lowest (43.0) values of total cup quality were recorded
for samples prepared using demucilager pulper and dried in tunnel drier and
conventional fermentation and dried in the open sun, respectively. All prepared
samples showed no statistically significant (P>0.05) effect in the body.
[367]
Table 2. Effect of recently introduced post-harvest machineries and driers on Organoleptic (cup) quality attributes of
coffee. Mean of two years observation
Cup quality parameters

Preparation
method AI AQ AC AS BI BO FL OL CT (60%)
NFDS 3.67b 3.67bc 7.17b 3.67b 3.33c 7.33a 7.00b 7.17b 43.00c
NFDA 3.67b 3.50c 7.33ab 3.67b 3.33c 7.50a 7.33ab 7.33ab 43.67c
NFTd 4.17a 4.17a 7.17b 3.83ab 3.50bc 7.17a 7.00b 7.17b 44.17bc
DPDS 4.00ab 4.00ab 7.83a 4.00ab 4.00ab 7.83a 7.67a 7.67ab 47.00ab
DPTD 3.83ab 3.83abc 7.83a 4.17a 4.17a 8.00a 7.83a 7.83a 47.50a
DPDA 3.83ab 3.67bc 7.33ab 3.83ab 3.67abc 7.33a 7.50ab 7.33ab 44.50bc
CV (%) 6.83 6.19 3.88 6.83 10.16 6.26 3.91 4.77 3.59
LSD (0.05) 0.22 0.19 0.24 0.22 0.3 0.38 0.24 0.28 1.32
Mean values followed by the same letter within column are not significant differences (P >0.05).
AI =aromatic intensity, AQ = aromatic quality, AC = acidity, AS= astringency, BI = bitterness FL= flavour, OL= overall
quality and CT= cup total
Total Quality Values

Though no statistically significant difference (P>0.05) was observed among
samples processed and dried in different preparation methods on raw quality
values shown as (Table 3). Samples pulped in demucilager and dried in poly
tunnel drier was recorded the highest values 35.67, 47.50 and 83.17 for raw, cup
and total quality, respectively.
Table 3. Effect of introduced post-harvest machineries and driers on total quality of coffee.
Mean of two years observation.
Values
Preparation method Raw value (40%) Cup value (60%) Total value (100%)
NFDS 33.83a 43.00c 76.83c
NFDA 33.50a 43.67c 77.17c
NFTd 34.33a 44.17bc 78.50bc
DPDS 35.50a 47.00ab 82.50ab
DPTd 35.67a 47.50a 83.17a
DPDA 35.00a 44.50bc 79.50abc
CV (%) 3.73 2.83 2.83
LSD (0.05) 1.05 1.32 1.84
Mean values followed by the same letter with in a column are not significant difference (P > 0.05).

There was no statistically significant difference (P>0.05) among coffee samples
processed and dried in different preparation methods for raw quality. Samples
pulped by demucilager and dried in poly tunnel drier exhibited the highest values,
35.67, 47.50 and 83.17, for raw, cup and total quality, respectively. Generally, the
result indicated that out of the six coffee samples assessed for raw, cup and overall
quality, those demucilager pulped and dried in the open sun (DPDS) and
demucilager pulped and dried in poly tunnel drier (DPTd) showed very good
overall quality standard and were found to be highly acceptable. So using newly
introduced demucilager and drying technologies had no negative impact on both
[368]
raw and sensory quality attributes and fit to export standards of samples and were
found to be less time consuming and less laborious. However, further study like
evaluating of biochemical composition, cost-benefit analysis should be carried out
to generate complete information on the processing methods and similar
experiments should be carried out in some major coffee-growing areas of the
country to come up with a comprehensive conclusion.
Acknowledgments
We acknowledge the Ethiopian Institute of Agricultural Research (EIAR) and
Horizon coffee plantation PLC for financial supports and sample preparation,
respectively during the study period. We are also thankful to the team of JARC
coffee cuppers.
References
Anwar Abasanbi. 2010. Assessment of coffee quality and its related problems in Jimma zone of
Oromia regional state. M.Sc. Thesis, Jimma University, Ethiopia.
Abrar Sualeh and Negussie Mekonnen. 2015. Manual for Coffee Quality Laboratory ISBN:
978999446605410-8.
Behailu Wolde Senbet, Abrar Sualeh, Nigusie Mekonen and Solomon Endris. 2008. A review of
coffee Processing and Quality Research in Ethiopia .In: Coffee Diversity and Knowledge.
Ethiopian Institute of Agricultural Research, pp: 307-316.
Dessie Nure. 2008. Mapping quality profile of Ethiopian coffee by origin. In: Proceedings of a
National Work Shop four decades of coffee research and development in Ethiopia. 14-17
August 2007, EIAR, Addis Ababa, Ethiopia. pp. 328-333.
EEA(European Environment Agency). 2001. Annual report on Ethiopian economy. 1999/2000,
EEA, Addis Ababa, Ethiopia.
FAO/WFP. 2008. Special report FAO/WFP crop and food supply assessment mission to Ethiopia.
24 January 2008.
Mburu, JK. 1999. Notes on coffee processing procedures and their influence on quality. Kenya
coffee, 64(750): 2861-2867.
Musebe, R, Agwenanda C and Mekonnen M. 2007. Primary Coffee Processing in Ethiopia: In
Africa Crop Science Society; Africa Crop Science Conference Proceedings, 8: 1417-1421.
Sfredo MA, Finzer JRD and Limaverde JR. 2005. Heat and mass transfer in coffee fruits drying.
Journal of Food Engineering 70, 15-25.
Wintegens JN. 2004. Coffee: Growing, processing, sustainable production, a guide book for
growers processors, traders, and researchers, WILEY - VCHVerlag GmbH & Co. KGaA,
Weinheium.
[369]
[370]
Vegetable
Performance of Garlic Varieties under Rain-fed

Cultivation Practices in Libokemkem and
Fogera Areas, South Gondar Zone, Ethiopia
Dessie Getahun1* and Mulat Getaneh1
1
Ethiopian Institute of Agricultural Research (EIAR), Fogera National Rice Research and Training Center
(FNRRTC), Ethiopia, * Corresponding author: desdesgetahun@gmail.com
Abstract
Garlic (Allium sativum L.) is an important edible bulbous crop with unique culinary
and medicinal purposes. It is a major cash crop widely cultivated in Libokemkem and
other districts of the South Gondar zone of Amhara region in Ethiopia. However, the
productivity of garlic in Ethiopia in general and in South Gondar Zone, in particular, is
very low largely due to the use of unimproved local varieties and traditional cultural
practices. Five different improved varieties of garlic were then evaluated for yield and
yield components under rain-fed production practice. This acclimatization and
performance evaluation at three different locations (Angot, Ginaza and Woreta) was
laid in a randomized complete block design with three replications. Variety Adiszemene
local produced consistently high dry bulb yield (55.44, 39.82 and 22.28 quintal per
hectare (q/ha)) at the three locations. Variety Chefe was found to be the lowest yielding
at Angot (16.07 q/ha) and Ginaza (22.56 q/ha), whereas varieties Kuriftu (11.92 q/ha)
and Tsedey (10.06 q/ha) were the lowest yielding at Woreta indicating a profound effect
of environment on yield. The overall result revealed that statistically significant
(P<0.05) high dry bulb yield (39.18 q/ha) was recorded by variety Adizemene local,
followed by variety Holleta (31.32 q/ha). These varieties with consistent high
performance and wide adaptability would then be demonstrated and popularized for
widespread cultivation.
Keywords: Acclimatization, Allicin, bulb, environment, medicine, spice
Introduction
Garlic (Allium sativum L.) is an important edible bulbous crop belonging to the
family Alliaceae along with onion, shallot, leek and chives. It is the second most
widely cultivated Allium after onion and has been used throughout history for
culinary and medicinal purposes (Pandey, 2012). Garlic has a higher nutritive
[371]
value than other bulbs crops (Abou El-Magd et al., 2012). Keeping because of its
medicinal value, especially Allicin of garlic which has antibacterial properties (Al-
Otayk et al., 2009; Sterling and Eagling, 1997), garlic is widely used in all
households throughout the year. According to Amagase et al., (2001) and Iciek et
al., (2009), the unique flavor and health-promoting functions of garlic are
generally attributed to its rich content of sulfur-containing compounds, i.e., alliin,
g-glutamylcysteine, and their derivatives.
Garlic is grown worldwide in all temperate to subtropical and tropical hilly areas
as an important spice and medicinal plant (Pandey, 2012). Clinically, garlic has
been evaluated for lowering blood pressure, cholesterol, and glucose
concentration, as well as for the prevention of arteriosclerosis and cancer (Tsai et
al., 2012). It was further indicated that, epidemiologically, garlic consumption
inversely correlates with the risk of oral, stomach, esophageal, colon, and prostate
cancers. In addition, the biological activities of garlic, including antibacterial,
antithrombotic, antioxidant, immunomodulatory, and anti-diabetic actions and
modulation of drug metabolism, have been extensively investigated. Despite its
various uses for flavoring food and treatment of health problems, garlic breeding
has been hampered for the plants rarely produce seed and are propagated asexually
by cloves or sometimes by top sets. Variety development for better yield and
quality was therefore solely based on clone selection from variations due to
mutation. According to Etoh and Simon (2002), only limited genetic variation can
be introduced via mutations and it is very hard to make significant progress by
mutation breeding alone. However, Khar et al., (2011) and Senula and Keller
(2000) reported that garlic exhibits a wide range of diversity in morphological,
reproductive and bulb traits because of its apomictic nature which leads to the
existence of extensive somatic mutations (Ata, 2005). It was further indicated by
Jabbes et al., (2012) that garlic presents, despite its vegetative multiplication, a
great diversity from agro-morphological and biochemical point of view (Abdoli et
al., 2009; Burba and Gomez Riera, 1997; Hong, 1999; Lallemand et al., 1997;
Messiaen et al., 1993; Ovesna et al., 2007; Stavelìkova, 2008). This variability is
required researchers to benefit from the wide diversity of garlic for the selection of
best genotypes. The genetic diversity serves as a source of genotypes adapted to
local conditions (Baghalian et al., 2005).
Ethiopia with diversified agro-ecological conditions is suitable for garlic

production. South Gondar zone in Amhara region of Ethiopia is potentially
endowed with favorable climatic and soil conditions for the cultivation of garlic
both under rainfed and using irrigation. Garlic in this zone particularly in
Libokemkem district is a major cash crop produced both under rainfed and
irrigation in larger plots similar to field crops such as maize, rice and ‘tef’
(Eragrostis tef). Rainfed garlic in Libokemkem is usually planted from June to
July and harvested in October and the plot is sown immediately with chickpea,
[372]
whereas irrigated garlic is planted in November after harvesting, cereals such as
‘tef’ and maize produced in the rainy season from May to October.
The contribution of this sector towards ensuring the engagement of a large number
of labor with the participation of females and youth could not be undermined since
garlic production is a labor-intensive job. Despite multifaceted uses, garlic suffers
from several problems that caused low productivity and poor quality largely
attributable to the use of unimproved local varieties with poor productivity.
Yeshiwas et al., (2017) also indicated that lack of improved varieties and garlic
rust are the major ones among many contributing factors for low production and
productivity of garlic in Ethiopia. Moreover, due to repeated cultivation of similar
varieties for several years in specific plots, fungal contamination of planting
materials and soil is observed in the vicinity of Libokemkem areas. Besides
applying improved husbandry practices, it is, therefore, crucial to identify
appropriate variety with high productivity and quality suitable to target
environments. With the ultimate objective of supplying healthy planting materials
of alternative varieties, this trial was therefore undertaken to evaluate varieties for
acclimatization and performance studies at Fogera and Libokmekm districts in
northwestern Ethiopia.

The experiment was conducted at three locations: two in Libokemkem district
(Angot and Ginaza kebeles) and one in Fogera district (Woreta kebele) of south
Gondar zone of Amhara region in Ethiopia. Planting materials of released garlic
varieties, Viz., Chefe, Holleta, Kuriftu and Tsedey were obtained from Debre Zeit
Agricultural Research Centre while bulbs of Adiszemene local were bought from
Adiszemene market. Varieties were then laid in randomized complete block
design with three replications in three locations namely, Woreta, Angot and
Ginaza in the South Gondar Region of northwestern Ethiopia.
Land was prepared to a fine tilth through repeated plowing the experimental plot.
Ridges of 20 centimeters (cm) width and 25 cm height along with 40 cm furrow
width were prepared. Cloves (bulb splits) were then planted at both sides of the
ridge with a 10 cm distance between plants. The spacing used was therefore 40 cm
furrow width, 20 cm distance between rows on the ridge and 10 cm between
plants. The effective plot size was 4.8m2 (2.4m x 2m) and contained 160 plants on
two meter long four ridges (8 rows). Planting dates were June 19, 26 and 30, 2018
for garlic trials at Woreta, Angot and Ginaza, respectively. Depending on the
maturity time of varieties, harvesting at these locations was accomplished from
September 20 to October 15, 2018.
[373]
A mix of Nitrate Phosphate sulfur (NPS) (38:19:5) at the rate of 242 kg/ha and
Urea (46%N) fertilizers at the rate of 100 kg/ha were applied. NPS fertilizer was
applied once at planting while Urea was applied in two splits, the first at full
emergence (10-15 days after planting) and the second at one and 1/2 months after
planting. Hand weeding and cultivation (hoeing) were performed throughout the
field evaluation period. Stand count, plant height and record for disease incidence
were made during the vegetative phase. Data was recorded on the maturity date,
bulb size, dry yield per plant and plot. Data were subjected to analysis of variance
and the least significant difference (LSD) was used to compare treatment means
when there was a statistically significant difference (P<0.05).

Significant differences (P<0.05) among varieties were recorded for dry bulb yield
in quintals per hectare, yield per plant and days to maturity. Varieties Adizemene
local (55.44 q/ha) followed by Holleta (42.74 q/ha) and Tsedey (40.88 q/ha) gave
the highest dry bulb yield at Angot kebele of Libokmkem district while Chefe
(16.07 q/ha) was the lowest yielding variety.
Variety Chefe was early in maturity taking only 95 days at Angot. Besides,
Holleta (57.43 cm) was found to be the tallest variety at Angot, though differences
in height among varieties were not statistically significant (Table1). As compared
with other varieties, Adizemene local and Holleta gave the highest yield at Ginaza
kebele (Tables 2). Ayalew et. al ( 2015) also obtained significantly the highest
yield from the local variety at Dabat district of Northwestern Ethiopia further
reflecting that varieties released for certain areas may not suit to other areas.
However, the lowest yielding varieties were Tsedey (10.06 q/ha) and Kuriftu
(11.92 q/ha) at Woreta, and Chefe (22.56 q/ha) at Ginaza kebeles. As compared to
other varieties, Adiszemene local and Chefe (94 days) were early maturing at
Woreta. Furthermore, Adiszemene local (43.63 cm) was shorter in height than
other varieties at Woreta, whereas Chefe (42.65 cm) and Adiszemene local (41.60
cm) were the shortest varieties at Ginza (Tables 2 and 3). On the other hand, trials
conducted during dry periods under irrigation at Dabat District, Northwestern
Ethiopia indicated that Bishoftu Nech was early maturing variety (135 days)
followed by Kuriftu (143 days), whereas MM-98 matured late at 176 days after
planting (Ayalew et al., 2015). This could be attributable to differences in growing
conditions including agro-ecology (temperature, altitude etc.), soil and cultural
practices. Furthermore, Khatun et al., (2014) showed a significant difference
among varieties in plant height, the weight of bulb per plant and the number of
bulbs per square meter which might have increased yield significantly. Knowing
more about variety performance will allow one to make more viable decisions in
the future when it comes to trying different varieties for large scale commercial
production.
[374]
Table 1. Mean dry yield, days to maturity and plant height of garlic varieties at Angot Kebele in 2018 rainy season (June
to September)
Variety Yield (qt/ha) Yield per plant (g) Days to maturity Height (cm)
Adiszemene local 55.44a 19.27a 100.00b 56.17a
Holleta 42.74b 18.52a 103.67bc 57.43a
Tsedey 40.88bc 16.63a 110.00d 56.83a
Kuriftu 30.62 c 11.50 b 104.67c 51.93a
Chefe 16.07d 6.32c 95.00a 50.80a
LSD (0.05) 11.39 3.39 3.72 8.04
CV% 16.29 12.45 1.92 7.82
Means followed with different letters within a column are significantly different at P < 0.05.
Table 2. Mean dry yield, days to maturity and plant height of garlic varieties at Ginaza Kebele in 2018 rainy season
(June to September)
Variety Yield (qt/ha) Yield per plant (g) Days to maturity Height (cm)
Adiszemene local 39.82a 17.18a 98.00ab 41.60b
Holleta 30.51b 13.25b 99.67b 48.55a
Tsedey 25.24bc 14.51b 106.00c 50.27a
Kuriftu 24.71bc 12.81b 106.00c 51.38a
Chefe 22.56c 9.43c 96.00a 42.65b
LSD (0.05) 7.17 2.07 2.57 3.35
CV% 13.37 8.18 1.35 3.79
Means followed with the same letters within a column are not significantly different at p≤0.05.
Table 3. Mean dry yield, days to maturity and plant height of garlic varieties at Woreta Kebele in 2018 rainy season (June
to September)
Variety Yield (q/ha) Yield per plant (g) Days to maturity Height (cm)
Adiszemene local 22.28a 8.72ab 94.00a 43.63b
Holleta 20.72a 9.20a 110.00b 51.26a
Tsedey 10.06b 6.31b 112.00bc 46.18b
Kuriftu 11.92b 7.96ab 114.33c 47.53ab
Chefe 21.11a 8.36ab 94.00a 47.05b
LSD (0.05) 8.3 2.8 3.87 3.95
CV% 25.59 18.32 1.96 4.45
Means followed with the same letters within a column are not significantly different at p≤0.05.
Table 4. Overall mean dry bulb yield, days to maturity and plant height of garlic varieties combined over three locations in
2018 rainy season (June to September)
Variety Yield (q/ha) Yield per plant (g) Days to maturity Height (cm)
Adiszemene local 39.18a 15.06a 97.33b 47.13b
Holleta 31.32b 13.65ab 104.44c 52.41a
Tsedey 25.39c 12.48b 109.33d 51.09a
Kuriftu 22.42cd 10.76c 108.33d 50.28a
Chefe 19.81d 8.04d 95.00a 46.83b
LSD (0.05) 4.56 1.57 1.72 2.88
CV% 17.11 13.56 1.73 6.03
[375]
Table 5. Combined mean dry bulb yield, days to maturity and plant height of garlic varieties in the three locations in 2018
rainy season (June to September)
Location Yield (q/ha) Yield per plant (g) Days to maturity Height (cm)
Angot 37.15a 14.45a 102.67b 54.63a
Ginaza 28.51b 13.43a 101.13a 46.89b
Woreta 17.22c 8.11b 104.87c 47.13b
LSD (0.05) 3.54 1.22 1.33 2.23
Overall highest yield was obtained from Angot followed by Ginaza (Table 5) that
could mainly be attributable to the suitability of soil and environmental condition
for the manifestation of the genetic potential of different varieties. The effect of
different environmental factors on the performance of garlic varieties was also
emphasized by Khan et al., (2018). Raslan et al., (2015) also indicated that garlic
yield and quality vary greatly with variety, location, soil type, agricultural
methods and harvest date. According to Ennes (1990) varieties do not perform
equally in all environments, but some tend to be close to the ideal than others.
Likewise, Foreaker (2015) reported that different types of garlic were chosen for
two separate growing seasons in Alaska. Indeed, varieties response is therefore
measured through interaction of genotypes and the environments.
Overall mean dry bulb yield from the three locations was significantly high in
variety Adiszemene local followed by Holleta. Varieties Chefe and Adiszemene
local were early maturing and the shortest in height (Table.4). Plant height was
strongly correlated with yield at Angot (r=0.84) (Figure 3). This is in conformity
with the reports of Zakari et al., (2017) that the increase in plant heights, number
of leaves and other positively correlated characters increased the amount of
assimilates being produced and translocate to the sink which finally affects the
yield. Insignificant correlation (-0.48 and 0.02) was also observed between height
and yield at Ginaza and Woreta. According to Albuquerque et al., (2017),
correlations among several quantitative parameters evaluated were verified using
phenotypical correlation coefficients because most of them were significant
(r>0.7). Bulb production is correlated with the vegetative growth (Adekpe et al.,
2007). Garlic plants with well-developed vegetative structures have effective
source-drain systems for nutrient and photoassimilate translocation to the bulbs
from the leaves and the pseudo-stems. This property results in a greater potential
for the production of large bulbs (Mathew et al., 2011).
[376]
Figure 1. Yield in quintal per hectare of garlic varieties at Libokemkem and Fogera areas, South Gondar Zone, Ethiopia
Figure 2. Correlation between height in cm and garlic yield in qt/ha (r=0.52)
Figure 3. Correlation between height in cm and yield of garlic in qt/ha at Angot, Ethiopia (r=0.84)
[377]
Varieties Chefe and Adiszemene local were early maturing, while Holleta was
intermediate and Tsedey and Kuriftu were late maturing. On the other hand, Chefe
was the lowest yielding and Adiszemene local was the highest yielding variety.
The yield performance of varieties at Kebele Angot was significantly higher as
compared to the other two kebeles (locations) revealing variations in the
suitability of soil and environmental conditions for the cultivation of garlic. Plant
height was strongly correlated with dry bulb yield at Angot which is the most
appropriate location for garlic cultivation suggesting that plant height could be
used as a selection criterion in a variety development of garlic.
Overall performance revealed that varieties Adiszemene local and Holleta were
found to be the best yielding varieties suitable for cultivation across locations.
These varieties can, therefore, be recommended for on-farm demonstration to
further verify the performance of these varieties with the active participation of
farmers, extension experts and development agents. Popularization of the varieties
and multiplication of healthy planting material for widespread dissemination will,
therefore, rely on feedback from demonstration and divulgation efforts.
Acknowledgment
Authors would like to acknowledge the Cool Season Vegetable Research Program
of Debre Zeit agricultural research center of the Ethiopian Institute of Agricultural
Research for supplying us with planting materials of garlic varieties. We are also
grateful to Mr. Maru Adugna and Desalegn Sisay for their support in managing
field experiments.
References
Abdoli M, Habibi-Khaniani B, Baghalian K, Shahnazi S, Rassouli H, Naghdi BH. 2009.
Classification of Iranian garlic (Allium sativum L.) ecotypes using RAPD marker. Jour. Med.
Plants 8: 45–51.
Abou EM M, El-Shourbagy MT and Shehata, SM. 2012. A Comparative study on the
Productivity of Four Egyptian Garlic Varieties Grown Under Various Organic Material in
Comparison to Conventional Chemical Fertilizer. Australian Journal Basic and Applied
Science 6 (3) 415-421.
Adekpe DI, Shebayan JAY, Chiezey UF and Miko S. 2007. Yield responses of garlic (Allium
sativum L.) to oxidation, date of planting and intra-row spacing under irrigation at Kadawa,
Nigeria. Crop Protection, Guildford, 26 (12):1785-1789
Albuquerque JRAr, Monteiro HNB, Bezerra AAC, Filho CHAM, Lopes ACA and Gomes RLF.
2017. Agro-morphological performance of garlic landraces in Piauí, Brazil. Ciência Rural
47(6).
[378]
Al-Otayk SMI, Motawei and El-Shinawy MZ. 2009.Variation in productive characteristics and
diversity assessment of garlic varieties and lines using DNA markers. J. Meteorology,
Environ and Arid land Agric. Sci 20 (1).
Amagase H, Petesch BL, Matsuura H, Kasuga S, Itakura Y. 2001. Intake of garlic and its bioactive
components. J Nutr. 131:955-962.
Aslam T, Dudi BS, Pandy, AK and Rana MK. 2016. Evaluation of garlic (Allium sativum L.)
genotypes for yield and yield attributing traits under semi arid zone of Haryana (Hisar), The
Asian Journal of Horticulture 11(1).
Ata, AM. 2005. Constitutive heterochromatin diversification of two Allium species cultivated in
Egypt. In Proceedings of the 7th African crop science society conference, Kampala, Uganda,
5–9 December, pp. 225–231.
Ayalew, A, Tadesse D, Medhin ZG and Fantaw S. 2015. Evaluation of Garlic (Allium sativum L.)
Varieties for Bulb Yield and Growth at Dabat, Northwestern Ethiopia. Open Access Library
Journal, 2, e1216. doi: http://dx.doi.org/10.4236/oalib.1101216.
Baghalian K, Sanei MR, Naghavi MR, Khalighi A, Naghgi Badi HA. 2005. Post culture evaluation
of morphological divergence in Iranian garlic ecotypes. Acta Hort. 688:123-125.
Burba JL and Gomez Riera P. 1997. Characterization, adaptation and selection of garlic
germplasm (Allium sativum L.) through the management of dormancy in Mendoza, Argentina.
Acta Hortic. 433: 151-164.
Ennes RA. 1990. Detection and measurement of selection: genetic and ecological approaches.
Plant population genetics, breeding and genetic resources. Sinavar Association Inc.,
Lunderland, Massachusetts, pp. 252-257.
Etoh T and Simon, PW. 2002. Diversity, fertility and seed production of Garlic In: Rabinowitch,
HD and L. Currah (eds), Allium Crop science: Recent advances, CAB International.
Foreaker R. 2015. Garlic Variety Demonstration Trial 2 Year Report 2012-2014, State of Alaska,
Department of Natural Resources, Division of Agriculture, Alaska Plant Materials Center.
Hong CJ. 1999. Fundamental Studies on Crossbreeding in Garlic, Allium sativum L. Kagoschima
University, Kagoshima, Japan. IPGRI, 2001.
Iciek M, Kwiecien I, Wlodek L. 2009. Biological properties of garlic and garlic-derived
organosulfur compounds. Environ Mol Mutagen 50:247-265.
Jabbes N, Ingrid A, Jacques A, Bouthaina Al MD and Cherif H. 2012. Agro-morphological
markers and organo-sulphur compounds to assess diversity in Tunisian garlic landraces.
Scientia Horticulturae.148: 47–54.
Khan H, Hussain MM, Jellani G, Hidayatullah ST, Naseeb T and Mhmood S. 2018. Evaluation of
Garlic Genotypes for Yield and Yield Components in Islamabad, Pakistan Environment.
The Nucleus 55 (1): 22-26.
Khar,A, Kaushik B , Manjusha R J and Lawande KE. 2011. Evaluation of garlic ecotypes for
allicin and other allyl thiosulphinates. Food Chemistry128: 988–996.
Khatun MUS, Ferdous MZ, Islam MK and Anowar MM. 2014. Performance of some high
yielding garlic varieties at two locations of Bangladesh. J. Bangladesh Agril. Univ. 12(2):
235–239.
Lallemand J, Messian, CM, Briad F and Etoh T. 1997. Delimitation of varietal groups in garlic
(Allium sativum L.) by morphological, physiological and biochemical characters. Acta Hortic.
433:123–132.
Messiaen CM, Cohat J, Leroux JP, Pichon M and Beyries A. 1993. Les Allium alimentaires
reproduits par voie végétative. INRA, Paris.
Omnarayan Verma and Bhupendra Thakre. 2018. Evaluation of Garlic Variety for Better Growth
and Higher Yield under Allahabad Agro – Climatic Condition. Int. J. Curr. Microbiol. App.
Sci. 7(02): 2275-2280. doi: https://doi.org/10.20546/ijcmas.2018.702.274
Onvegetables. 2018. Garlic Variety Trial 2017-2018 Results , Ministry of Agriculture, Food and
Rural affairs, Ontario.
[379]
Ovesna J, Kucera L, Kralova J, Leisova L, Stavelikova H and Velisek J. 2007. Genetic diversity
among garlic clones as revealed by AFLP, phenotypic descriptors and S-amino acids level.
Veg. Crops Res. Bull. 66: 105–116, http://dx.doi.org/10.2478/v10032-007-9.
Pandey, UB. 2012. Garlic, Handbook of herbs and spices, Woodhead Publishing.
Raslan M, S AbouZid, M Abdallah, and M Hifnawy. 2015. Studies on garlic production in Egypt
using conventional and organic agricultural conditions. African Journal of Agricultural
Research,Academic journals 10(13): 1631-1635
Senula A and Keller RJ. 2000. Morphological characterization of a garlic core collection and
establishment of a virus-free in vitro genebank. Allium Improvement Newsletter, 10:3–5.
Stavelìkova, H (2008). Morphological characteristics of garlic (Allium sativum L.) genetic
resources collection—Information. Hort. Sci. (Prague) 35: 130–135.
Sterling SJ and Eagling DR. 1997. Agronomics and allicin yield of Australian grown garlic (Allium
sativum), 2nd Int. Symp. on Edible Alliaceae, 555: 63-73.
Tsai, Chia-Wen, Haw-Wen Chen, Le-Yen Sheen and Chong-Kuei Lii. 2012. Garlic: Health
benefits and actions, BioMedicine 2:17-29 .
Yeshiwas Y, Belete Negash, Tegibew Walle, Yohaness Gelaye, Abayneh Melke and Kassahun
Yissa. 2017. Collection and characterization of garlic (Allium sativmL.) germplasm for growth
and bulb yield at DebreMarkos. Ethiopia Journal of Horticulture and Forestry,
academicjournals;10(3):17-26.
Zakari SM, Haruna H and Aliko AA. 2017. Correlation analysis of bulb yield with growth and
yield components of garlic (Allium sativum L.), Nigerian Journal of Basic and Applied
Science 25(1): 58-62.
[380]
Influence of Plant Population on Bulb Yield and
Yield Components of Some True-Seed
Propagated Shallots
Getachew Tabor1*, Fasil Tadesse1, and Yenenesh Asfaw1
1
Ethiopian Institute of Agricultural Research, Debre Zeit Research Center, P. O. Box 32, Bishoftu, Ethiopia;
*
Corresponding author: getachewtf@gmail.com
Abstract
Shallot (Allium cepa var. aggregatum) is one of the vegetable crops used for
condiment. It has been propagated vegetatively. Vegetative propagation has several
disadvantages such as the need for large amount of planting material, short shelf life,
high cost, and transmission of diseases. In order to mitigate the problems, the Debre
Zeit Agricultural Research Center developed two seed propagated shallot cultivars.
Following the development of the cultivars, there was a need to optimize agronomic
practices such as plant population since neither the practices used for bulb
propagated shallot nor that used in onion could be adopted. Therefore, a study was
initiated to determine the appropriate intra-row spacing that could establish an
optimum amount and quality of bulb yield. The experiment was conducted on Vertisol
at Debre Zeit Agricultural Research Center between 2013 and 2017. It comprised
three seed propagated shallot cultivars (DZSHT-193-1A, DZSHT-157-1B and
Vathalam) and four levels of intra-row spacing (5 cm, 10 cm, 15 cm and 20 cm). The
experiment was laid out in randomized complete block design with three replications.
The results of the study showed that the cultivars performed differently in different
years. The main effects of spacing and cultivar were also significant. Plants spaced 20
cm apart had the highest bulb yield per plant and mean bulb weight than those grown
5 cm apart but no difference was observed in bulb yield per ha. However, plants
grown at 5 cm spacing gave more number of small size bulbs than the wider spacing.
Cultivars DZSHT-157-2B and DZSHT-193-1A had better yield than Vathalan.
Therefore, spacing of plants 20 cm apart could enable growers to save seeds and
labor to manage additional seedlings in the nursery and plants in the field.
Keywords: plant population, shallots, Allium cepa var. aggregatum
Introduction
Shallot (Allium cepa var. aggregatum) is one of the important vegetable crops
used for seasoning of different national cuisines in Ethiopia. It had been the
traditional and the major Allium crop produced in the country until the
introduction of onions (Allium cepa var. cepa) in 1970. Its production, however,
has been gradually decreasing due to problems associated with vegetative
propagation. The use of bulbs as propagating material was expensive, bulky to
transport and had high postharvest loss during storage and carries diseases from
one generation to the next. Research efforts made by Debre Zeit Agricultural
[381]
Research Center (DZARC) to mitigate these problems resulted in the release of
two shallot cultivars (DZSHT-193-1A and DZSHT-157-1B) from local
germplasm and introduction and adaptation of a cultivar (Vathalam) which are
seed propagated. Cultural practices such as plant population used for vegetatively
propagated shallot were found to be not appropriate for seed propagated shallots
owing to the apparent difference in growth and development of plants grown from
the two types of planting material. Neither the plant population used for onions
could be adopted because of the developments of lateral shoots in shallots as
opposed to the single bulbs in onions. Therefore, the present study was initiated to
develop an optimum plant population for high yield and quality of bulbs.

The Experiment was undertaken at Debre Zeit Agricultural Research Center.
DZARC is located at 08° 44 N and 38° 58’ E and it has an altitude of 1900 meters
above sea level, masl. The experiment was undertaken for five consecutive dry
irrigated seasons from 2013 to 2017. The trials were planted on Dec. 11, Dec., 29,
Oct. 30, Nov. 1, respectively, in the successive years. Three seed propagated
shallot cultivars (cvs) (DZSHT-193-2B and DZSHT-157-1) (Getachew, 2018) and
cv Vathalam (MoA, 2005) were used for the experiment. Seeds of these cvs were
sown and raised on seedbeds for about six weeks until they attain three to four true
leaves stage and were then transplanted in the research field at DZARC. The field
was prepared with ridge and furrows. The ridges were 20 cm wide and 40 cm
apart from each other. The seedlings were transplanted on both sides of ridges
spaced 20 cm between the two rows for all treatments. The seedlings were spaced
5 cm, 10 cm, 15 cm and 20 cm apart within the row. Di-ammonium phosphate
(DAP) at 200 kg/ha was applied just before planting along with 75 kg urea per ha.
Additional 75 kg/ha of urea was applied four weeks after days to 50% sprouting.
Fungicides Mancozeb (2.5 kg/ha) and Redomil Gold 63.5 WP (2.5 kg/ha) were
applied against downy mildew while karate (1l/ha) was used to manage onion trips
as the pests appear in the field. The plants were irrigated every week using the
ridge and furrow irrigation.
The factorial combinations of the three cvs and four intra-row spacing were
arranged in Randomized Complete Block design with three replications. Data on
yield and yield components were collected and analyzed using PROC GLM
procedure of SAS software (SAS, 2002, SAS Institute Inc., Cary, NC, USA).
Count data on numbers of bulbs were log-transformed before analysis. Whenever
the analysis showed significant, means were separated using the Least Significant
Difference (LSD) at p≤0.05 level of significance.
[382]
Analysis of variance of the data showed that the interaction among year, variety
and spacing, as well as between spacing and variety were not significant in all the
yield and yield component parameters considered. However, the different levels of
plant population cultivars as well as performed differently in different years in
several parameters.
The three cultivars performed differently in the different years (Table 1). Cultivar
Vathalam had the lowest number and weight of small bulbs in most (2013, 2016
and 2017) years. On the other hand, it had a high number and weight of medium
and large size bulbs in all the years except in 2016. DZSHT-157-1B had a higher
number and weight of bulbs of medium and large size bulbs in 2013. Whereas in
2016DZSHT-193-1A and DZSHT-157-1B produced a higher number of small
size bulbs. The total number and weight of bulbs of DZSHT-193-1A and DZSHT-
157-1B were higher than that of Vathalam except in 2015. On the other hand, the
mean bulb weight and bulb yield per plant of the cultivars were not significantly
different across the years. The difference in the performance of the cultivars
across the years was owing to the weather conditions on growth, development, and
disease incidence such as downy mildew.
Significant differences were observed among the plant populations in bulb size,
bulb yield per plant, and number of bulbs produced per m2 (Table 2). Plants grown
20 cm apart produced 60% higher bulb yield per plant as compared to those grown
5 cm apart and 37% higher than those grown 10 cm apart. However, there were no
significant differences between those grown 20 cm and 15 cm apart and between
those grown 10 cm and 5 cm apart. The plants grown 20 cm and 15 cm apart also
produced 31% and 37% bigger bulbs than those grown 5 cm apart, respectively.
However, plants spaced 20 cm, 15 cm and 10 cm had similar mean bulb sizes.
Likewise, the mean bulb size of plants grown 10 cm and 5 cm apart were also
similar. Plants grown 5 cm and 10 cm apart had 46 (72%) and 41 (60%) more
bulblets per m2 than those grown 20 cm apart. Kanton et al., (2003) reported a
decrease in mean bulb weight with an increase in plant density in onions in Ghana.
Similarly, McGeary (1985) reported that mean bulb, and fresh and dry weights
decreased as plant population of White Spanish onion increased from 178 to 1600
per m2.
There was no significant increase in yield per ha with increasing plant population.
The results are in contrast to the reports of Kanton et al., (2003) who found an
increasing total and marketable bulb yields with an increasing population of onion
plants from 37 to 156 plants per m2. They found that above 77 plants per m2 is
optimum to obtain higher marketable bulb yields. Studies conducted on intra-row
spacing of onions also showed a decrease in yield from 36 t to 28 t per ha as intra-
[383]
row spacing increased from 5 to 10 cm (Kahsayet. al., 2013); however, mean bulb
yield increased from 50g to 81g. Fekadu (2015) also reported that bulb propagated
shallot grown at intra-row spacing of 10 cm gave 21% better bulb yield but 6%
smaller bulbs as compared to those spaced 20 cm while maintaining 50 cm
spacing between ridges. Sarah et al., (2015) reported that onion plants (cv Adama
Red) grown on ridges 40 cm apart and at intra-row spacing of 10 cm resulted in
the highest mean fresh and dry bulb weights, fresh and dry biomass yields as
compared to those grown at 7.5 cm and 5 cm intra-row spacing.
The lack of difference in yield per ha with increasing plant population was
attributed to the decrease in mean bulb weight and bulb yield per plant. Shallot
plants that had a wide perimeter of space around them could able to access and
utilize more growth resources such as plant nutrients, water as well as light. The
higher yield in wider spacing could be due to the specific characteristics of
shallots to produce multiple shoots and bulbs, as opposed to the single shoots of
onion. Similarly, Brewster and Salter (1980) did not find significant bulb yield
difference between 30 cm and 15 cm inter-row spacing because of the reversing
effects mean bulb weight with increased spacing.
Further investigation into the different size categories of the bulbs produced
showed significant differences among the spacing on the number and weight of
small size bulbs (Table 3). Plants grown 5 cm and 10 cm apart had 64% and 30%
more number as well as 34% and 18% more weight of small sized bulbs as
compared to those grown 20 cm apart.
Plants grown 10 cm and 15 cm apart had 39% and 23% more number and 23%
and 8% more weight of bulbs as compared to those grown 20 cm apart. Similarly,
those grown 10 cm apart had 15 % and 18% more number and weight of medium
sized bulbs as compared to those grown 5 cm apart. Nonetheless, the difference in
the number and weight of large size bulbs between the spacing treatments was not
significant. Frappell (1973) reported that a trend of progressive shift of size grade
to smaller grades as the density of onion plants increased.
The main effect of cultivar was also significant in the number and weight of small
size bulbs, number of medium size bulbs, and bulb yield (Table4). Cultivars
DZSHT-193-1A and DZSHT-157-1Bhad 59% and 49% higher number and 52%
and 41% higher weight of small size bulbs as compared to Vathalam. Vathalam
had 38% and 29% more number of medium size bulbs as compared to DZSHT-
193-1A and DZSHT-157-1B; however, the difference in weight of bulbs in this
category was not significant among the cultivars. The total yield of these cultivars
showed that DZSHT-193-1A and DZSHT-157-1B had 29% and 26% higher yields
as compared to Vathalam. However, the cultivars did not differ in weight of
medium size bulbs and number and weight of large size bulbs. Moreover, the
[384]
difference in mean bulb size and bulb yield per plant was not significant among
the cultivars. The difference or similarity in the different bulb size categories
among the cultivars could be attributed to the genetic difference of the cultivars.

The results of the current study showed that the different intra-row spacing could
influence the size category of the bulbs produced more than it could influence the
total bulb yield. Plants grown 5 cm apart produced a higher number of small size
bulbs as compared to those grown at wider spacing due to fierce competition
among the plants. Peeling of such bulbs is tedious and time consuming. Plants
which were grown 10 cm and 15 cm apart produced more number of medium and
large size bulbs with no apparent difference in bulb weight and yield. Those
grown 20 cm apart had less number of small but more number of large size bulbs
but with no significant difference from 10 and 15 cm spacing. Therefore, shallot
plants grown using botanical seeds can be produced by spacing them 20 cm apart
resulting in better size bulbs with no apparent loss of yield. These benefits growers
through saving 2.55 kg seeds and labor to manage about 50 seedlings per m2 in the
nursery, transplanting 500 thousand plants per ha in the field2. Moreover, DZSHT-
157-2B and DZSHT-193-1A could benefit growers.
Acknowledgment
The authors acknowledge the Ethiopian Institute of Agriculture Research for
financing the study. Technical staffs of the Cool Season Vegetable Research
Program of the Debre Zeit Agricultural Research Center are acknowledged for
managing the experimental fields and collecting data.
2
Calculated based on 1000 seed weight of 4.08 g and 80% germination
[385]
Table 1. The interaction effects of year of planting and cultivar on different bulb size categories, mean bulb weight and yield of three seed propagated shallot, Debre Zeit.
Total bulb
Year Variety Small (< 25 mm) Medium (25-35 mm) Large (> 35 mm) yield
Weight Mean Bulb Yield/ plant
No. Weight (q/ha) No. Weight (q/ha) No. Weight (q/ha) No (q/ha) weight (g) (g)
2017 DZSHT-157-1B 617.6(2.76)de 111.24(2.02)abc 64.13(1.76)ab 24.29(1.33)ab 40.51(1.56)b 23.27(1.32)abc 722.2(2.84)bcd 158.80(2.18) 22.65(1.34) 39.77(1.57)
DZSHT-193-1A 649.8((2.78)cde 126.81(2.08)ab 67.12(1.78)ab 24.54(1.34)bc 44.45(1.62)b 23.64(1.35)ab 761.3(2.85)bcd 174.98(2.23) 23.98(1.37) 41.91(1.61)
Vathalam 342.1(2.50)g 66.68(1.77)d 93.97(1.91)a 34.14(1.47)a 75.23(1.71)a 44.25(1.51)a 511.3(2.66)de 145.07(2.10) 28.18(1.44) 43.99(1.61)
2016 DZSHT-157-1B 967.8(2.96)ab 145.14(2.15)a 52.79(1.70)abcd 18.68(1.25)abc 23.38(1.18)bc 13.10(0.98)c-f 1044.0(2.99)ab 176.91(2.29) 17.99(1.24) 86.53(1.92)
DZSHT-193-1A 1214.4(3.08)a 144.83(2.15)a 47.23(1.60)bcd 19.75(1.18)bc 3.7(0.28)c 2.34(0.24)g 1265.3(3.09)a 166.92(2.21) 13.24(1.12) 81.79(1.88)
Vathalam 205.2(2.29)h 30.36(1.48)e 50.00(1.36)de 6.47(0.69)e 20.83(1.06)bc 3.66(0.50)ef 276.0(2.41)e 125.89(2.10) 15.38(1.18) 50.43(1.67)
2015 DZSHT-157-1B 728.(2.84)bcd 95.06(1.97)bc 61.81(1.75)ab 18.61(1.25)abc 33.57(1.31)bc 16.47(1.06)b-e 823.4(2.90)abc 130.13(2.11) 16.46(1.21) 63.71(1.80)
DZSHT-193-1A 698.9(2.83)cd 88.86(1.94)bc 36.58(1.23)e 12.68(0.89(de 26.16(1.45)bc 11.67(0.88)def 761.6(2.87)bcd 113.21(2.04) 14.98(1.17) 53.49(1.69)
Vathalam 811.6 (2.90)bc 124.87(2.09)ab 69.68(1.81)ab 24.39(1.35)ab 40.52(1.38)b 21.29(1.15)a-d 921.8(2.95)ab 170.55(2.22) 18.70(1.27) 81.47(189)
2014 DZSHT-157-1B 874.0(2.87)bcd 122.16(2.01)abc 49.2(1.64)bcd 15.86(1.15)bc 32.18(1.39)bc 13.95(1.03)b-e 955.3(2.92)ab 151.97(2.12) 16.72(1.21) 60.31(1.74)
DZSHT-193-1A 667.9(2.78)cde 142.22(2.00)bc 34.73(1.50)cd 13.75(1.06)cd 20.54(1.18)bc 8.75(0.78)ef 723.2(2.82)bcd 164.72(2.09) 23.10(1.26) 70.30(1.71)
Vathalam 731.1(2.84)bcd 101.04(1.97)bc 58.33(1.73)abc 20.97(1.27)abc 21.39(1.12)bc 9.86(0.86)def 1268.3(2.97)ab 131.46(1.59) 14.84(1.12) 44.61(1.62)
2013 DZSHT-157-1B 830.6(2.88)bcd 104.31(1.97)bc 81.24(1.67)abcd 25.83(1.21)bc 44.18(1.45)b 23.47(1.18)abc 956.0(2.95)ab 153.61(2.15) 16.84(1.20) 78.49(1.86)
DZSHT-193-1A 540.1(2.70)ef 118.58(2.05)ab 53.74(1.72)abc 20.76(1.29)ab 45.83(1.63)ab 30.45(1.44)a 639.7(2.78)bcd 169.79(2.22) 28.50(1.43) 99.95(1.97)
Vathalam 442.5(2.59)eg 86.37(1.89)cd 59.17(1.72)abc 20.54(1.27)abc 30.55(1.43)bc 19.93(1.26)abc 532.2(2.68)cde 126.85(2.08) 25.45(1.39) 71.72(1.82)
Mean 688.09(1.93) 107.23(1.86) 58.65(0.84) 20.08(1.08) 33.33(1.30) 17.74(1.04) 810.78(2.85) 145.03(2.12) 19.80(1.26) 64.57(1.75)
CV
(%) (3.37) (3.22) (26.87) (22.04) (39.25) (42.66) (5.43) (8.23) (10.28) (9.94)
[386]
Table 3. The influence of intra-row spacing of true seed propagated shallot plants on number and yield of different bulbsize categories, bulb weight, and yield/plant combined over years
and cultivars at Debre Zeit during 2013-2017
No. of Small (<25 mm) Medium (25-35 mm) Large (>35 mm) Total Bulb yield/ Mean bulb
Spacing plants/ No. Weight (q/ha) No. Weight No. Weight No. of Bulb yield plant (g) weight (g)
(cm) m2 (q/ha) (q/ha) bulblets/m2 (q/ha)
20 16.7 549.1(2.67)c 91.43(1.90)c 48.63(1.58)b 18.42(1.16)bc 31.04(1.32)ab 17.07(1.05) 63.9(2.75)b 126.81(2.0 79.00(1.86)a 21.42 (1.31)a
15 22.2 591.2(2.71)c 107.75(1.94)bc 59.57(1.72)a 19.95(1.24)ab 32.01(1.29)ab 17.38(1.05) 68.3(2.79)b 145.08(2.0
6) 72.02(1.79)a 22.40 (1.30)a
10 33.3 716.59(2.82)b 107.36(1.99)a 67.60(1.77)a 22.74(1.28)a 40.33(1.44)a 21.43(1.16) 82.5(2.89)a 151.52(2.1
9) 57.86(1.73)b
b 19.04 (1.26)ab
5 66.7 895.6(2.89)a 122.40(2.04)a
b 58.78(1.56)b 19.23(1.12)c 30.75(1.14)b 15.08(0.89) 109.7(2.96)a 156.71(2.1
5) 49.39(1.64)c
c 16.34 (1.19)b
LSD0.05 (0.085)** (0.159)* ns ns 20.35(0.066) ns
4) 13.16(0.075) 3.52(0.056)
Mean 688.09(2.77) 107.23(1.97)
(0.106)** 58.65(1.66)
(0.174)** 20.08(1.20) 17.74(1.23) 33.53(1.04) 81.1(2.85) 145.03(2.1 64.57( 1.81) 19.8 (1.26)
CV (%) (5.47) (9.52) (18.63) (23.57) (39.25) (42.66) (5.43) (8.23)
1) ) (9.94) (10.28)
* Means within the same column followed by the same letter are not significantly different at p≤0.05. Means in parenthesis are log transformations;
ns = non-significant difference at p≤ 0.05.
Table 4. The main effects of cultivar on number and yield of different bulb size categories combined over spacing during 2013-2017.
Small (< 25 mm) Medium (25 - 35 mm) Large (> 35 mm) Total bulb yield Mean bulb Yield/plant
Variety
weight (g) (g)
Weight No. Weight (q/ha)
No Weight (q/ha) No Weight (q/ha) No bulb
(q/ha)
DZSHT-157- 803.59(2.84)a 124.26(2.04)a 47.88(1.57)b 18.3(1.15) 28.14(1.17) 15.37(0.94) 830.22(2.89) 157.92(2.15)a 20.76(1.27) 69.89(1.77)
DZSHT-193-
1B 754.2(2.86)a 115.58(2.03)a 61.83(1.70)a 20.65(1.24) 34.76(1.38) 18.05(1.11) 900.18(2.92) 154.28(2.16)a 18.13(1.24) 65.76(1.78)
Vathalam
1A 506.49(2.63)b 81.86(1.84)b 66.23(1.71)a 21.3(1.21) 37.7(1.34) 19.8(1.06) 701.93(2.74) 122.88(2.02)b 20.51(1.28) 58.45(1.72)
Mean 688.09(2.77) 107.23(1.97) 58.65(1.66) 20.08(1.19) 33.53(1.29) 17.74(1.04) 810.78(2.85) 145.03(2.11) 19.8(1.26) 64.57(1.76)
LSD0.05 * * * ns ns ns ns * ns ns
CV (%) (5.47) (9.52) (18.63) (23.57) (39.25) (42.66) (5.43) (8.23) (10.28) (9.94)
* Means within the same column followed by the same letter are not significantly different at p≤0.05. Means in parenthesis are log transformed; ns = non-significant difference at p ≤ 0.05
[387]
References
Brewster JL and Salter PJ. 1980. A comparison of the effects of regular versus random within row
spacing on yield and uniformity of size of spring sown bulb onions. Journal of Horticultural
Sciences (55)3: 235-238.
Fikadu Negese Biru. 2015. Effect of Spacing and Nitrogen Fertilizer on the Yield and Yield
Component of Shallot (Allium ascalonium L.). Journal of Agronomy 14 (4): 220-226, 2015
Frappell, BD. 1973. Plant spacing of onion. Journal of Horticultural Science, Volume 48 (1), 19-
28.
Tabor Getachew. 2018. Development of seed propagated shallot (Allium cepa L var. aggregatum)
varieties in Ethiopia. Scientia Horticulturae. 240:89-93
Kanton RAL, Abbey L, RG Hilla, MA Tabil and ND Jan. 2003. Density Affects Plant
Development and Yield of Bulb Onion (Allium cepa L.) in Northern Ghana. Journal of
Vegetable Crop Production, Volume 8 (2), 15-25.
McGeary, DJ. 1985. The effects of plant density on the shape, size, uniformity, soluble solids
content and yield of onions suitable for pickling. Journal of Horticultural Science.Volume 60,
83-87.
Ministry of Agriculture. 2005. Crop Variety Registry.
SAS. 2002. Statistical Analysis Software. Institute Inc., Cary, NC, USA.
[388]
Evaluation of Lettuce (Lactuca sativa)
Genotypes for Yield and Quality in Ethiopia
Zebenay Dagne1*, Samuel Tilahun1, Ousman Yimer1, Gete Fekadu1, Fekadu
Gebretensay2, Fasil Tadesse Tewolde3, Getachew Tabor3 and Gizachew Atinafu3
1
Holetta Agricultural Research Centre, P.O. Box, 2003 (31 Holetta) Addis Ababa, Ethiopia
2
Kulumsa Agricultural Research Centre, P.O. Box 489, Kulumsa, Ethiopia
3
Deberziet Agricultural Research Centre, P.O. Box 32, Debre Zeit, Ethiopia
*
Corresponding author: dzebenay@gmail.com
Abstract
Development of high yielding, good quality and adapted lettuce varieties is one the
objective of cool season vegetable program. Thus, this study was conducted to evaluate
and select wide adaptable and best quality candidate lettuce genotypes for release.
Multi-location variety trial was conducted during 2016/17 using three genotypes (Rsk-3,
Gsk-3 and GRsk-3) and one lettuce variety as a standard check. This experiment was laid
out in RCBD with three replications and a plot size 3.6m 2was used. The overall
combined analysis across locations and years showed a significant difference in leaf
sizes, plant weight, total yield and mineral composition (p<0.01)among genotypes. In
general, Rsk-3 and Gsk-3 genotypes were found superior in mineral composition, and
also they gave relatively good yield and thus identified for release to be used as
additional lettuce varieties and nutritional source.
Keywords: Lettuce genotypes, leaf size, yield, quality
Introduction
Lettuce (Lactuca sativa L.) is a cool season annual crop that belongs to the
family Astaraceae. It is one of the most popular salads of vegetables in the
world (Mou, 2008). Lettuce is among the major fresh leafy vegetables and its
leaves are commonly found in salad and sandwiches. Lettuce comprises
health-promoting nutrients and biosynthesis of such phytochemicals varying
on the cultivar, leaf color, growing conditions and other environmental factors
(Kim et al., 2016). Lettuce is a very nutritional food with high calcium, iron
and vitamin A content (Antwerpen, 2009; Mou, 2008).
In Ethiopia lettuce production and consumption is playing a major role in

nutrition and income generation among the urban people and small scale
farmers respectively. Lettuce is a nutritious leafy vegetable, rich in mineral
and a source of vitamin. The vegetable contains many important dietary
minerals crucial for human health like iron, zinc which are chronic
malnutrition problems for many children in developing countries and
phosphorous, calcium, magnesium etc. including other health enhancing
compounds (Kim et al., 2016). The nutritional composition of lettuce also
[389]
includes edible portion66%, moisture 93.4%, protein 21g, fat 0.3g, mineral
1.2g, fiber 0.5g, other carbohydrate 2.5g, etc. (Anonymous, 2019). Since it’s
nutritional and other economic importance, a vast number of lettuce varieties
suitable for different agro-ecologies have been released worldwide (Mou,
2008). In our country so far only one single lettuce variety (Tesfa/Maya) has
been released by the national research system and most of the different lettuce
varieties that are produced here are imported by different private seed
companies. Due to the unavailability of a well-organized seed system coupled
with improved lettuce varieties, the cost of lettuce seed is relatively expensive
and unaffordable for many small scale farmers of Ethiopia. Therefore, in
order to avail the technology to producers and farmers, a lettuce line was
introduced from Korea and continuous selection and advancement of the
materials were done at Holetta agricultural research center. The objective of
this research study was to evaluate adaptation, yield and quality of lettuce
genotypes and recommend for production.

Description of the study areas: Multi-location Variety Trial was conducted
in 2016 and 2017 off-season at three locations: Holetta, Debre Ziet, and
Kulumssa Agricultural Research Centers, under irrigation condition.
Table 1: Altitude, rainfall, soil type and maximum and minimum temperature values of experiment sites
Location Altitude Annual rainfall Soil type Temperature

(m) (mm) min max
Holetta 2400 1144 Nitosols 6ºC 22ºC
Deberziet 1900 851 Light soils 8.9°C 28.3°C
Kulumsa 2200 840 Clay soil 10ºC 22ºC
Experiment design and treatments: Three candidate lettuce varieties (Gsk-

3, Rsk-3and RGsk-3) and one standard check (Tesfa/Maya) were used for the
evaluation. The experiment was laid out in RCBD with three replications and
a plot size of 3.6m2 was used. Each plot comprised of five rows with spacing
arrangement of 40cm between rows and 30cm between plants, respectively.
Field management and cultural practices: The seedling was raised in

seedbeds and seed trays and transplanted after three weeks when they
produced 3-4 true leaves. Fertilizer, DAP 200kg ha-1 was applied during
transplanting and UREA 100kg ha-1 was applied in a split application, where
the first 50% of UREA was applied during transplanting and the second 50%
after 30 days of transplanting. All the necessary cultural practices were
applied to all plots uniformly as per the recommendations established for the
respective locations.
[390]
Data collection and analysis: Parameters included in the data collection
were leaf length(cm), head/canopy diameter(cm), leaf weight(g), plant
weight(g), total yield (q/ha) and leaf mineral composition. Data were
collected from the middle rows of the plot and/five plants per plot. Data were
subjected to analysis of variance (ANOVA) using the General Linear Model
of the SAS statistical package (SAS, 2007) version 9.1. All significant pairs of
treatment means were compared using the Least Significant Difference Test
(LSD) at 5% level of significance. For mineral composition analysis, the
LSD test was used for means comparison by SPSS Version 22.

The combined analysis across locations and years showed a highly significant
difference (p<0.01) among varieties for traits like canopy/head diameter, leaf
diameter and leaf length (Table 2). The highest plant diameter (31.44cm and
30.4cm) were obtained from Tesfa and Rsk-3 varieties whereas the least plant
diameter (23.87cm) was recorded from RGsk-3.The highest leaf diameter
(15.06cm) was obtained from RGsk-3 whereas the least plant diameter
(12.78cm) was recorded from Rsk-3. The highest leaf length (20.8cm and
20.5cm) was obtained from Tesfa and Rsk-3 respectively whereas the least
leaf length (14.39cm) was recorded from RGsk-3. The present result agreed
with the finding of Hassan et al. (2012) that lettuce genotypes revealed high
genetic diversity for most traits like leaf length and width.
In 2017 at Holetta Agricultural Research Center, there was a highly

significant difference among varieties in leaf yield (P<0.01). Significantly
maximum plant diameter, leaf diameter and leaf length 48.25, 18.4 and
25.95cm respectively were recorded for Rsk-3 genotypes whereas the least
plant diameter, leaf diameter and leaf length 38.65, 15.6 and 19.7cm were
obtained from Tesfa, Gsk-3 and Tesfa respectively. In 2016 at Kulumsa
agricultural research center, there was a highly significant difference among
varieties in leaf diameter and leaf length (P<0.01). Significantly maximum
leaf diameter and leaf length 18.9 and 22.46cm were recorded from RGsk-3
and-3, respectively whereas the minimum leaf diameter and leaf length 13.3
and 18cm were obtained from-3 and RGsk-3 varieties respectively.
[391]
Table 2. The Performance of the lettuce varieties with respect to their plant sizes at Holetta, Debre Zeit and Kulumssa in 2016 and
2017
Variety 2016
Holetta Debreziet Kulumssa
Plant Leaf Leaf Plant Leaf Leaf Plant Leaf Leaf
diameter diameter length diameter diameter length diameter diameter length
(cm) (cm) (cm) (cm) (cm) (cm) (cm) (cm) (cm)
Tesfa 27.9a 13.10b 21.26 a 33.73a 11.32 b 18.03a 31.733 15.20 b 24.40a
Gsk-3 28.250a 12.54 b 18.86 b 25.00b 10.87 b 15.31b 31.500 13.30c 22.45a
Rsk-3 30.390a 12.54b 19.47b 23.26b 11.82ab 18.00a 29.800 14.00c 22.46a
RGsk-3 24.500b 17.60 a 14.36 c 24.34b 13.03 a 13.73c 29.267 18.90a 18.0b
Mean 27.768 13.94 18.49 26.41 11.86 16.075 30.58 15.35 22.06
LSD (5%) 2.986 1.257 1.223 3.503 1.632 1.470 ns 1.06 2.82
CV 6.289 5.27 3.86 6.50 6.74 4.48 4.50 4.07 6.68
Variety 2017
Holetta Debreziet Kulumssa
Plant Leaf Leaf Plant Leaf Leaf Plant Leaf Leaf
diameter diameter length diameter diameter length diameter diameter length
(cm) (cm) (cm) (cm) (cm) (cm) (cm) (cm) (cm)
Tesfa 38.05b 15.65 b 19.7b 23.20a 10.05a 14.3 a 35.73a 16.43a 3.28a
Gsk-3 38.65b 15.6b 20.4b 20.30 ab 7.13b 11.06b 21.90c 10.10 c 19.35a
Rsk-3 48.25 a 18.4a 25.95a 22.17ab 8.65ab 15.05a 26.950b 11.06bc 19.5a
RGsk-3 28.45c 18.35a 16.3 c 19.46b 10.15a 11.25b 19.75c 11.46b 13.4b
Mean 38.35 17.0 20.58 21.40 9.12 12.85 25.44 12.02 18.85
LSD (5%) 4.441 2.6 3.4 ns 1.886 1.505 2.790 1.23 5.208
CV 6.77 9.10 9.64 9.63 11.43 5.56 6.06 5.35 15.27
Plant weight and total yield performance of lettuce varieties: The

overall combined analysis for total plant weight and total yield of lettuce
varieties across locations and years were highly significant (p<0.01) (Table
4).Based on the combined analysis the highest leaf weight and total yield
(403.46g/plant and 297.40qt/ha) were obtained from Tesfa whereas the lowest
leaf weight and total yield (257.92 g/plant and 201.40qt/ha) was obtained
from RGsk-3. The results of this study were consistent with the finding of
Thakur et al. (2014) who indicated that highly significant differences were
obtained among lettuce genotypes for traits like plant weight and yield.
Similarly, Hassan et al. (2012) indicated that there is diversity in leaf weight
among lettuce genotypes
In 2017 at Holetta and Kulumsa Agricultural Rresearch Center, there was a

highly significant difference among genotypes in leaf weight and total yield
(P<0.01) (Table 3), but in 2016 and 2017 at Debreziet there was no significant
difference in total yield among genotypes. Significantly maximum leaf
weight of 359.67g/plant and 317.46 qt/ha total yield were recorded from Rsk-
3 genotypes whereas the lowest leaf weight 227.15g/plant and 220.7qt/ha total
yield were obtained from Tesfa variety. Similarly, at Kulumsa significantly
maximum leaf weight of 507.9 g/plant and 423.25qt/ha total yield were
recorded from Gsk-3, whereas the lowest leaf weight of 171.64g/plant and
163.27qt/ha total yields were recorded from Gsk-3.
[392]
Table 3. Plant weight and total yield performance of lettuces varieties at Holetta, Debre Zeit and Kulumssa in 2016 and2017
Variety 2016 2017

Holetta Debreziet Kulumssa Holetta Debreziet Kulumssa
Plant Total Plant Total Plant Total Plant Total Plant Total Plant Total yield
weight yield weight yield weight yield weight yield weight yield weight (Q/ha)
(gram) (Q/ha) (gram) (Q/ha) (gram) (Q/ha) (gram) (Q/ha) (gram) (Q/ha) (gram)
Tesfa 337.00a 103.4b 374.30a 331.82 820.00a 365.97a 272.27b 220.77b 270.27a 326.62 507.90a 423.25a
Gsk-3 286.50a 152.8a 200.54b 257.48 690.00b 322.22ab 296.00ab 304.76a 197.10b 179.80 171.64b 163.27b
Rsk-3 174.25b 114.3a 209.66b 284.29 475.00c 262.04b 359.67a 317.46 a 283.05a 180.48 269.32b 224.43b
RGsk-3 184.00b 94.1b 205.91b 259.26 606.67b 250.00b 227.15b 243.06ab 183.00b 201.55 175.71b 180.15b
Mean 245.43 116.1 246.8 284.80 624.61 305.13 284.03 271.5 234.3 222.1 281.1 255.4
LSD (5%) 80.9 36.8 105.2 ns 127.4 75.20 74.66 76.29 42.12 ns 135.1 74.11
CV 19.28 15.02 18.85 17.00 9.68 12.08 12.8 12.8 9.3 18.9 21.9 11.7
[393]
Table 4. The overall combined Performance of the lettuce varieties at Holetta, Debre Zeit and Kulumssa in 2016 and
2017.
Leaf Leaf length Plant weight Total
Plant diameter diameter (cm) (gram) yield
Varieties (cm) (cm) (Q/ha)
Tesfa 31.44a 13.6b 20.8a 403.46a 297.40a
Gsk-3 27.7b 11.8d 18.28b 313.61b 256.14b
Rsk-3 30.4a 12.78c 20.5a 301.77b 242.37b
RGsk-3 23.87c 15.06a 14.39c 257.92c 201.40c
Mean 28.29 13.3 18.4 316.9 251.74
LSD (5%) 1.63 0.8406 1.333 43.39 39.93
CV 8.89 9.73 11.09 19.82 21.44
Proximate Composition: The proximate compositions of the lettuce varieties

were analyzed at Holleta Agricultural Research Center (Figure 1). The candidate
lettuce varieties (Gsk-3 and Rsk-3) surpassed the standard check for the traits like
fresh moisture content, crude protein and Ash total mineral. But, Rsk-3 was
recorded less crud fat than Gsk-3 and Tesfa
100
90
80
70
60
50
40
30
20 Tesfa Lettuce (std check.)
10
0 Gsk-3
Rsk-3
Figure1. Proximate composition of lettuce genotypes at Holetta
Mineral composition of lettuce: From the analysis of the variance, lettuce

varieties showed a highly significant difference (P< 0.01) on mineral composition
of lettuce at Holleta growing conditions (Table 5). According to the results,
Potassium, magnesium and iron had the highest concentrations which were
obtained from Rsk-3 and Gsk-3 varieties. But Rsk-3 and Gsk-3 recorded less Zn
(0.96 + 0.01and1.29 + 0.01) concentration, respectively than Tesfa (2.39 + 0.67)
(3.077). According to Bagdatlioglu et al (2014) reported that Fe and Zn in lettuce
concentrations were found in the range of 4.04–66.13 and 1.86–3.63 mg/kg,
respectively.
[394]
Table 5. Mineral composition of lettuce at Holetta
Sample type K (mg/kg) Ca(mg/kg) Mg (mg/kg) Zn (mg/kg) Fe (mg/kg)

Tesfa Lettuce 696.49+3.53b 58.16+1.18b 30.66+ 0.78c 2.39+ 0.67a 3.03+ 0.01b
Gsk-3 752.10+8.28a 53.53+1.40c 39.48+ 0.63b 0.96+ 0.01b 5.42 + 0.01b
Rsk-3 763.5+1.41a 107.90+ 0.43a 47.41+ 1.64 a 1.29+ 0.01b 64.61+1.47a

This experiment was designed to evaluate lettuce genotypes for their adaptation,
leaf character, yield and quality and they were found different leaf characteristics
and nutritional quality. The overall combined analysis across location and years
showed a highly significant difference among genotypes in plant height, leaf
length, leaf weight, yield and quality (p<0.01). Rsk-3 and Gsk-3 genotypes were
found superior in mineral composition, proximate compositions and also they
gave relatively good yield). Gsk-3 candidate lines have very deep green color and
smooth leaf texture which is important additional lettuce varieties while the other
candidate variety (Rsk-3) has a deep red purplish color which indicates that it has
additional nutritional value, red color lettuce varieties are a rich source of vitamins
specially vitamins A than green color lettuce varieties. Therefore, the genotypes
Gsk-3 and Rsk-3 which were found superior in quality characteristics will be
promoted to official variety release so that they would be produced and used
additional nutritional value.
Acknowledgments
We acknowledge the Ethiopian Institute of Agricultural Research /EIAR), Holetta
agricultural research center for financial support and facilitation. We are also great
full technicians of vegetable research programs at Holetta, Debereziet and
Kulumsa research center. It’s our pleasure to thank KOPIA /KOPIA Ethiopia
center/ for the provision of lettuce genotypes.
Reference
Anonymous. 2019. Fresh Ethiopia. https://www.freshfromethiopia.com/ethiopia-lettuce/
Antwerpen, EG. 2009. Vegetable cultivation a practical hand book
Hassan, SM, Reza MH, Choukan R, Sepahvand N, Khosrowchali M. and Kaviani K. 2013.
Assessment of qualitative and quantitative traits in commercial Iranian lettuce
(Lactucasativa L.) genotypes. Annals of Biological Research, 2012, 3 (9):4352 4361. (http://sc
holarsresearchlibrary.com/archive.html.
Mou. B. 2008. VEGETABLES. Hand book for plant breeding.
Kim, Moo Jung Kim, Youyoun Moon, Dean A. Kopsell , Suejin Park , Janet C. Tou and Nicole L.
Waterland. 2016. Nutritional Value of Crisphead ‘Iceberg’ and Romaine Lettuces (Lactuca
sativa L.). Journal of Agricultural Science, 8(11):1-9.
Thakur, M, Kumar R, Kumar S, Bhardwaj M and Sharma M. 2014. Genetic Divergence Studies in
Lettuce (Lactuca sativa L) Under Protected Conditions in Mid Hills of Himachal Pradesh.
[395]
International Journal of Agriculture Innovations and Research Volume 3, Issue 2, ISSN
(Online) 2319-1473.
Bagdatlioglu N, Nergiz C and Ergonul PG. 2014. Heavy metal levels in leafy vegetables and some
selected fruits. Journal of Consumer Protection and Food Safety. Verbr. Lebensm. (2010)
5:421–42
[396]
Determination of Optimum Plant Population for
Onion (Allium cepa var. cepa L.) Seed Yield and
Quality in Central Rift Valley of Ethiopia
Jibicho Geleto1*, Tesfa Binalfew1, Fekadu Gebretensay Mengistu2, Nigissie Chewaka3,
Shimeles Aklilu1, Yosef Alemu1, Selamawit Ketema1, Melkamu Hinsermu1,
Gebeyehu Wondimu1, and Birhan Abera1
1
Melkassa Agricultural Research Center, P.O.Box 436 Adama, Ethiopia 1
2
Kulumsa Agricultural Research Center, P.O.Box 489, Kulumsa, Ethiopia 2
3
Werer Agricultural Research Center, P.O.Box 2003 A.A, Ethiopia3
*
Corresponding author: budie1977@gamil.com
Abstract
A field experiment was carried out at Melkassa, Kulumsa and Werer Agricultural Research
Centers during 2017, 2018 and 2019 off-season under irrigation to determine optimum plant
population density for high onion seed yield and quality under agro-climatic conditions of
central rift valley areas using Bombay Red onion variety. A factorial combinations of three
intra-row (10, 20 & 30cm), five inter-row (40, 50, 60, 70 & 80cm) spacing and a control
(50x30x20cm, between water furrow, between plant rows on the bed and between plants
respectively) were laid out in RCBD with three replications. Number of flower stalks per
plant, days to 50% flowering, umbel diameter, days to maturity, seed yield per umbel, seed
yield per plant, thousand seed weight, germination percentage and seed yield per hectare
were measured to assess onion seed yield and quality. Net return was also computed to
conclude the economical plant population density for onion seed yield. Analysis of variance
indicated that the interaction effect of inter-row and intra- row spacing brought about a
significant variation on number of flower stalks per plant, umbel diameter, seed yield per
plant, seed yield per umbel and seed yield per hectare; whereas, days to 50% flowering,
days to maturity, thousand seeds weight and seeds germination percentage were not
significantly (p > 0.05) influenced both due to interaction and main effects of the treatment
combinations. The maximum number of flower stalks per plant was recorded from 80x30
(11.6), 70x10 (12.9) and 80x30 (10.3) treatment combinations in 2018 at Melkassa, Kulumsa
and Werer locations respectively. Higher seed yield per plant was obtained from treatment
combinations 80x30 (34.6 g/plant) and 60x30 (14.6g/plant) in 2018 at Melkassa and
Kulumsa respectively; similarly in 2019, treatment combination 80x30 (18.3 g/plant) gave
higher seed yield per plant than the control. The analysis results revealed that the highest
seed yield was obtained from treatment combinations 80x10 (13.5qha -1) and followed by
60x10 (11.6qha-1) in 2017 and from 60x10 (16.6qha -1) in 2019 at Melkassa. Treatment
combination 40x10 gave the highest seed yield both in 2018 and 2019 at Kulumsa; likewise,
the same treatment combination gave the highest seed yield (12qha -1) at Werer in 2019 than
the control. Based on the partial budget analysis, treatment combinations 60x10, 40x30 and
40x10 were superior with respective net income of 1,126,675, 557,324 and 650,000
Ethiopian Birr at Melkassa, Kulumsa and Werer respectively. Therefore, it could be
concluded that Bombay Red onion variety could be planted at optimum spacing of 60 x 10,
40x30 and 40x10cm spacing at these respective locations to attain economical onion seed
yield. However, since spacing is dependent on a number of factors, optimum density should
be assessed and determined periodically by embracing fertilizer and irrigation studies.
Keywords: optimum, yield, quality, spacing, intra-row, inter-row
[397]
Introduction
Onion (Allium cepa L.) production is one of the important agricultural activities
for growers in different agro ecological conditions of Ethiopia with the central rift
valley the belt production region (Lemma and Shimels, 2003, Lemma et al.,
2008). Since its introduction to the country, onion production is known mainly
under irrigated conditions in the production belt areas, though currently the
production of the crop is also expanded to different new areas of the country.
Recently, its production is also started under rainfed conditions (CSA, 2016) and
an improved variety for rainfed production was also released by Melkassa
Agricultural Research Center for growers (MoALR, 2017). The area coverage and
production of the crop are also increasing for the last six years (CSA, 2013-2018).
The area under onion was only 21,865 ha in 2013; whereas, currently it is
increased to 31,673.21ha in 2018, which is about 45% increment. Similarly, the
production was 2,191,886 and 2,938,875.85 quintals in 2013 and 2018
respectively which accounted for 34% increment. This could be due to relative
ease of cultivation techniques and economic, social and nutritional importance of
the crop as the result of increased focus by agricultural development policy of the
country. Hence, the production of the crop showed an increasing trend due to
year-round cultivation. However, Ethiopia is importing a significant amount of
fresh onion bulbs from Sudan since the demand for the crop and its production are
not balanced. The considerable local consumption of the crop especially in homes,
hotels, restaurants, universities and other sectors such as Ethiopian Airlines are
witnessing the high demand for the crop in the country (personal observation).
One of the major onion production constraints is the lack of high-quality seeds and
improper agronomic practices such as plant population density (spacing). Spacing
is a vital operation for ease of field management and high seed yield and quality.
In the early 1980s, repeated studies have been conducted by Melkassa and Werer
Agricultural Research Centers on different spacing trials in response to the
problems identified earlier. Double and single row planting were compared with
20 and 30cm between plants and 80, 100 and 120cm between rows spacing
including furrows for three years (1985-1987). High seed yield (10qha-1) was
obtained in a double row at 80cm spacing between rows compared to 8qha-1 of
single row planting (IAR, 1986, 1987). However, nothing was reported on the
quality of seed produced under these different plant population densities.
Moreover, a spacing of 20, 30, and 40cm between rows on a bed and 30, 40, and
50cm between rows including furrows with 30cm and 20cm intra-row spacing
were studied. A higher seed yield of 9-12qha-1 was obtained at 20 and 30cm
between plants and 30 and 50cm combination. As a result, double row planting of
50X30X20cm (125,000 plants/ha) was recommended for Melkassa and similar
climatic regions (Lemma et al., 1994). However, currently, farmers are using their
[398]
plant population density for onion seed production; instead of double, single row
planting is used by almost all seed growers in the major growing areas of central
rift valley. Different spacing that range from ~ 40 to 80cm between rows including
water furrows is used with about 20cm intra row spacing (Chala a, Personal
Communication and own observation at farmers’ field). The arguments forwarded
by seed growers against double row spacing is its difficulty of field operations
such as weeding and spraying, breakage of seed stalks and loss of seed heads,
lower seed yield and quality and disease problems that incur a higher cost of
production as the result of repeated spraying. Consequently, the onion seed yield
currently obtained by model seed growers is about 6 q ha-1 and the quality is also
poor in most cases (Chala a and Girmab, Personal Communication). Hence, the
objective of this activity was to determine optimum plant population density for
high onion seed yield and quality under current agro- climatic conditions of
central rift valley areas.
a= Chala Gure is a model farmer at Adami Tullu Jido Kombolcha district Edo-
Gojola kebele, former chairman of Maki Batu Union Board. Mobile phone
number: 0916841757/0929242001.
b= Girma Tefera is also a model farmer at Adama district, Melka Hidda kebel
who has been working with the research center for a long period and
currently engaged in onion seed multiplication in the area.

Description of these study areas: The activity was conducted at Melkassa,
Werer which are both representing the dry, low land and Kulumsa which is
representing moist highland areas from 2018 to 2019 during the off-season
(October to June) using furrow irrigation. Melkassa is found in the Ethiopian Rift
Valley, 117 kilometers away from Addis Ababa in the southeast direction, located
at 8o 24’ N and 39o 12’ E. Werer is also found in the Ethiopian Rift Valley, 280
kilometers away from the capital in the Eastern direction and geographically
located at 9o16’N and 40o9’E; whereas, Kulumsa is found in the Central
Highlands of the country, 175 kilometers far from the capital in the South-Eastern
direction.
Table 1: Agro-climatic conditions of the experimental sites
Altitude Rainfall Temperature (oc)

Location Soil Type
(m) (mm) Min Max
Melkassa 1550 818 14 29 Ando-sol
Kulumsa 2200 820 10.5 22.8 Luvisol
Werer 750 590 28 41 Verti-sol & Flui-sol
[399]
Experimental Design and treatments: Multi-location trials were conducted in
Melkassa, Kulumsa and Werer Agricultural Research Centers using Bombay Red
onion variety to study onion seed plant population density for two years (2018 to
2019). A total of 16 treatment combinations of five inter-row spacing including
furrow (40, 50, 60, 70, 80 cm), three intra-row spacing (10, 20, 30 cm) and control
(50x30x20cm, between water furrow, between plant rows on the bed and between
plants respectively) were included in the study (Table 1). Factorial randomized
complete block design (RCBD) with 3 replications were employed, with a plot
size of 9.6 m2 at different inter and intra-row spacing as per the treatment.
Field experimental methods and plant management: Seeds were sown on

well-prepared seedbed of 1-meter width and 5 meters length having a 40cm path
between beds, after 200 g per bed DAP fertilizer was applied. Grass mulch with a
thickness of about 3-5cm was used to cover the seedbeds until seedling
emergence. All routine management practices and seedling protection measures
were applied as necessary. Watering was done using watering cane starting from
day of sowing until seedlings were ready for transplanting. Hand weeding, hoe
cultivation and urea fertilizer application (100g per bed) were also applied.
Seedlings were sprayed with Ridomil Gold M.Z 68 % W.P at the rate of 10-30 g
and profit 72 % EC, commonly named as profenofos at the rate of 30ml per 15
liters capacity knapsack sprayer for controlling leaf diseases and insect pests
respectively. Seedlings were transplanted to the field at 3-4 true leaves stages at
the spacing of 40cm between water furrows, 20cm between plant rows on the bed
and 5cm between plants on well-prepared field for seed bulb production. Furrow
irrigation, weeding, cultivation, fertilization and crop protection measures were
applied in the field as per recommendations for onion bulb production.
At maturity, the bulbs were harvested using forks and left to cure and dry and then
topped using sickles. Bulbs were sorted, graded and stored until planting for seed.
One to two months later, uniform sized bulbs having 4-5cm diameter were
selected and planted on a well-prepared field as per the treatment. Furrow
irrigation was applied every 5 to 7 days as required. Two times hoe cultivation and
three times hand weeding were also made to assure healthy and normal root
development and early plant growth. Fungicides such as Ridomil Gold M.Z 68 %
W.P (3kg/ha), agrolaxyl (3kg/ha); insecticides such as profit (0.75 liters/ha),
agrolambacine (0.4 liters/ha) were sprayed to control purple blotch and thrips
respectively. In order to control cutworm endosulphan was also used on the onion
seed trial field at the early plant growth stage.
Data collected: The data collection was done on plots in the middle rows by
leaving aside plants in the border rows as well as those at both ends of each row.
Data were taken from a sample of 5 (five) plants per plot for most of the
parameters recorded or measured. Data on days to emergence, stand count after
1and 1/2 MAT (months after transplanting), days to 50% bolting, days to 50%
[400]
flowering, number of flower stalks per plant, length of main stalk (cm), number of
flowers per umbel, umbel diameter (mm), no. of seeded fruits per umbel, percent
fruit set per umbel, days to seed maturity, disease score for major onion seed
diseases and insects, stand count at harvest, seed yield per umbel (g), seed yield
per plant (g), seed yield per ha (kg), thousand seeds weight (g) and seeds
germination percentage were taken accordingly.
Statistical analysis: The mean values of the data of the above parameters were
subjected to analysis of variance (ANOVA) using SAS and/or statistix8 software.
For parameters showed significant differences, mean separation was performed
using LSD test at 0.05 of the significance level.
Economic evaluation was computed using the method described by

CIMMYT(1988); where Gross return from each treatment was calculated in
Ethiopian Birr as total return = current seed price per kg. times onion seed yield
obtained from the average of 2018 and 2019 at the respective location; Net return
would also be calculated as a total return - bulb cost of onion seed produced from
respective treatment. Currently, one kg of onion seed bulb is 25 Ethiopian Birr on
average at Melkassa and Werer, 28 Ethiopian Birr at Kulumsa; whereas one kg of
onion seed is 1000, 1200 and 1250 Ethiopian Birr at Melkassa, Kulumsa and
Werer respectively. The average prices were taken from the current local market
price of the area.

Analysis of variance showed that there was highly significant interaction (p <
0.01) between treatment combinations across locations and years for inter and
intra- row spacing on the number of flower stalks per plant, umbel diameter, seed
yield per plant, and seed yield per ha. Significant interaction (p < 0.05) on seed
yield per umbel; and non-significant interaction (p > 0.05) between treatments
across locations and years on, days to 50% flowering, days to maturity, thousand
seeds weight and seeds germination percentage were obtained. Likewise, the main
effect of treatment combinations was not significant on these parameters (Tables
1-10).
[401]
Table 2: Effect of Inter and intra-row spacing on number of flower stalk per plant across locations and years
Melkassa Kulumsa Werer
No Treatments Comb.
2018 2019 2018 2019 2018 2019
1 Single row 40 X 10 cm spacing 5.9 5.3 9.1 7.0 7.3 6.7 6.8
16 Double row 50 X 30 X20 cm 6.5 6.3 9.5 11.6 8.1 4 7.7
spacing (Control)
Mean 8.19 5.40 10.22 9.66 8.01 5.5 7.83
F-test (5%) ** NS NS * NS NS **
CV (%) 20.49 29.37 18.46 19.99 21.22 24.15 22.88
LSD 5.102 NS NS 3.221 1.178
Number of stalks per plant: Analysis of variance showed that there was
significant interaction (p < 0.01) between treatment combinations across locations
and years for inter and intra- row spacing on the number of flower stalks per plant
(Table 2). In 2018 at Melkassa treatment combination 80x30 gave a significantly
higher number of flower stalks (11.6) per plant than the control (6.5) but in 2019
at the same location treatment combination, 40x30 gave a higher number of flower
stalks (9.7) per plant though it was not statistically different from the control (6.3).
On the other hand in 2018 at Kulumsa treatment combination 70x10 gave a better
number of flower stalks (12.9) than the control, but the difference was not
significant. However, in 2019 at the same location, treatment combination 80x10
gave better results though the difference was not significant from the control. At
Werer treatment combination 80x30 gave a higher number of flower stalks per
plant, however, the result was not significantly different from the control. The
result showed that at different locations and years different treatments showed
different responses and this lead to the interaction of treatments. The reasons for
interaction effect could be due to inherent agro-climatic differences exhibited
among the study locations to which the treatment combinations responded
differently. Generally, the result indicated that treatments with wider intra-row
spacing gave a better number of flower stalks per plant than treatment
combinations with narrower intra- row spacing regardless of inter-row spacing
(Table 2). This could be due to less competition for nutrients, air, water and
sunlight among plants in the case of wider intra-row spacing than among the
plants in the narrower intra-row spacing treatments. The higher plant population
led to competition for finite factors like oxygen, light, water and nutrients
(Bleasdale, 1966).
[402]
Days to 50% flowering: There was no significant interaction (p > 0.05) between
treatment combinations across locations and years for inter and intra- row spacing
on days to 50% flowering. The main effects of treatment combinations were also
not significantly affected days to 50% flowering (Table 3). In general, the result
indicated that the treatment combinations with narrow intra- row spacing took a
longer time to reach days to 50% flowering than wider intra-row spacing
treatments regardless of inter-row spacing.
Table 3: Effect of inter and intra- row spacing on days to 50% flowering across locations and years
Days to 50% flowering Days to maturity
No. Treatment
Melkassa Kulumsa Werer Melkassa Kulumsa Werer
Single row 40 X 10 cm 72.2 106.3 70.3 140.0 159 110.8
Single row 40 X 20 cm 71.7 107.0 68.5 140.7 161 109.5
Single row 40 X 30 cm 73.6 105.7 69.5 139.3 160 115.3
Single row 50 X 10 cm 71.0 108.2 69.0 139.3 157 112.3
Single row 50 X 20 cm 69.7 105.5 68.7 139.7 158 112.0
Single row 50 X 30 cm 73.1 106.2 68.7 140.0 159 112.8
Single row 60 X 10 cm 69.2 106.3 70.0 139.7 157 112.2
Single row 60 X 20 cm 72.3 105.5 70.5 140.3 160 115.0
Single row 60 X 30 cm 72.4 105.2 68.8 140.3 160 111.0
Single row 70 X 10 cm 73.1 106.5 68.2 140.0 157 111.2
Single row 70 X 20 cm 71.2 106.7 69.3 140.0 161 109.8
Single row 70 X 30 cm 71.3 106.3 69.5 140.0 162 114.5
Single row 80 X 10 cm 72.1 107.0 70.0 140.7 157 113.5
Single row 80 X 20 cm 72.9 106.7 69.2 139.7 159 111.3
Single row 80 X 30 cm 72.0 106.8 71.5 140.0 161 111.8
Double row 50 X 30 X20 cm 71.9 107.0 69.7 140.0 160 114.3
Mean 71.86 106.43 69.46 140.0 159.3 112.34
F-test NS NS NS NS NS NS
CV (%) 31.05 2.37 3.28 0.59 1.65 5.54
LSD (5%) NS 1.4587 NS NS NS NS
Days to maturity: There was no significant interaction (p > 0.05) between

treatment combinations across locations and years on days to maturity; besides the
main effects of treatment combinations were also not significantly affected days to
maturity at all locations. However, the days to maturity was ranging from 139.3 to
140.7 at Melkassa, 157 to 162 at Kulumsa and 109.5 to 115.3 at Werer; with
average days to seed maturity of 140, 159 and 112 at Melkassa, Kulumsa and
Werer respectively (Table 3). The result showed that onion seed relatively took
shorter days to mature at warmer areas like Werer; whereas, at Kulumsa (cooler
area) it took about one and half months more time to mature regardless of
treatment combinations. This could be due to the fact that at warmer climate the
rate of metabolic activities such as photosynthesis, flowering, seed formation and
maturation are faster than at cooler climates. Generally, from the result it was
observed that intra and inter-row spacing treatment combination didn’t show the
effect on the seed maturity; this is in agreement with the report by Koller and
[403]
Scholl (1968) who indicated that neither row spacing nor seeding rate had any
visible effect on plant maturity.
Yield and yield-related parameters

Umbel Diameter: Analysis of variance showed that there was highly significant
interaction (p < 0.01) between treatment combinations across locations and years
for inter and intra- row spacing on umbel diameter. In 2018 at Melkassa treatment
combination 80x30 highly significantly (p < 0.01) gave bigger umbel diameter
(7.6cm) than 40x10, 50x10, 60x10, 70x10 and 80x10 treatment combinations.
However, it was not significantly different (p > 0.05) from the rest treatment
combinations including control (6.8cm). Besides, in 2019 at the same location,
treatment combination 80x20 significantly (p < 0.05) gave bigger umbel diameter
(8cm) than the control (6.3cm). However, in 2018 at Kulumsa, treatment
combination 6x30 and 70x20 gave wider umbel diameter (6.4cm each) than the
control; whereas in 2019 at the same location, treatment combination 60x20 and
80x20 gave wider umbel diameter (5.3cm each) though there was no significant
difference (p > 0.05) from the control in both years. In addition, in 2018 at Werer,
treatment combinations 40x10 (5.3cm) and 50x30 (5.1cm) gave wider umbel
diameter than the control; however, in 2019 at the same location, treatment
combinations 50x10, 50x20 and 60x20 gave bigger umbel than the control though
there was no significant difference from the control in both years (Table 4). The
result indicated that at each year and location, different treatments give bigger
umbel diameter and hence resulted in the interaction of treatments across locations
and years. The reasons for the interaction effect between treatments could be due
to differences in agro-climatic conditions across locations.
Table 4: Effect of inter and intra -row spacing on umbel diameter (cm) across locations and years
No. Treatment 2017 2018 2019 Com 2018 2019 com 2018 2019 Com Overall
1 Single row 40 X 10 cm 4.9 5.6 5.0 5.2 5.1 4.7 4.9 4.8 4.2 4.5 4.9
2 Single row 40 X 20 cm 5.3 6.6 6.0 6.0 5.4 4.9 5.2 5.1 3.9 4.5 5.3
3 Single row 40 X 30 cm 5.4 6.8 5.7 6.0 5.5 4.8 5.2 5.3 4.2 4.8 5.4
4 Single row 50 X 10 cm 5.5 6.4 6.0 6.0 5.3 4.1 4.7 4.8 4.3 4.5 5.2
5 Single row 50 X 20 cm 5.3 7.1 6.7 6.3 6.0 4.7 5.4 4.9 4.3 4.6 5.6
6 Single row 50 X 30 cm 5.6 6.9 6.3 6.2 6.0 4.8 5.4 5.1 3.9 4.5 5.5
7 Single row 60 X 10 cm 5.4 6.3 6.0 5.9 5.4 4.4 4.9 4.5 3.7 4.1 5.1
8 Single row 60 X 20 cm 5.7 6.8 6.7 6.3 6.3 5.3 5.8 4.8 4.3 4.5 5.6
9 Single row 60 X 30 cm 5.8 7.0 6.0 6.3 6.4 4.8 5.6 5.0 4.2 4.6 5.6
10 Single row 70 X 10 cm 5.0 6.4 6.0 5.8 5.3 4.9 5.1 4.7 4.2 4.4 5.3
11 Single row 70 X 20 cm 5.2 7.2 6.0 6.2 6.4 5.1 5.7 5.1 4.1 4.6 5.7
12 Single row 70 X 30 cm 5.2 7.5 6.0 6.3 6.3 5.1 5.7 4.6 3.8 4.2 5.6
13 Single row 80 X 10 cm 5.0 6.4 6.0 5.8 5.8 4.6 5.2 4.8 4.0 4.4 5.3
14 Single row 80 X 20 cm 5.3 7.1 8.0 6.8 6.1 5.3 5.7 5.1 3.8 4.4 5.9
15 Single row 80 X 30 cm 5.1 7.6 6.0 6.3 6.2 4.7 5.5 4.7 4.1 4.4 5.6
16 Double row 50 X 30 X20 cm (Control) 4.9 6.8 6.3 6.0 5.5 4.8 5.2 4.8 4.0 4.4 5.4
Mean 5.28 6.78 6.17 6.09 5.82 4.8 5.31 4.89 4.06 4.47 5.42
F-test NS ** * ** NS NS ** NS NS NS **
CV(%) 8.58 5.22 11.5 8.43 9.64 9.08 9.52 6.5 12.27 9.35 9.47
LSD(5%) NS 10.77 1.19 0.48 NS NS 0.58 NS NS NS 0.3376
[404]
Besides, the difference between seasons within the same locations could also
affect the treatments differently. Rainfall, day and night temperature, relative
humidity, sunshine, etc… are among the important climatic factors that could
affect treatments differently. Though the treatments highly interacted across
locations and years, generally, the result revealed that treatments with narrow intra
and inter- row spacing showed relatively smaller umbel size than wider intra and
inter- row spacing treatments. This could be due to plant competition for nutrients,
moisture, radiation, air and effect of shading as the result of very narrow inter and
intra -row spacing (Norman, 1963).
Seed yield per umbel: Analysis of variance showed that there was significant
for inter and intra- row spacing on seed yield per umbel; in 2018 at Kulumsa
treatment combination 60x20 highly significantly gave higher seed yield per
umbel (2.8g/umbel) than 60x10, 70x20 and 70x30 treatment combinations.
However, it was not significantly different (p > 0.05) from the rest treatment
combinations including control (1.8 g/umbel).
Table 5: Effect of inter and intra -row spacing on onion seed yield per umbel (g) across locations and years

No. Treatment 2017 2018 2019 Comb 2018 2019 comb 2018 2019 Comb Overall
1 Single row 40 X 10 cm 2.1 2.9 1.7 2.2 1.8 1.7 1.8 1.5 1.5 1.5 1.9
2 Single row 40 X 20 cm 1.2 3.7 2.6 2.5 1.7 2.1 1.9 1.3 1.8 1.6 2.2
3 Single row 40 X 30 cm 2.0 3.6 2.8 2.8 1.9 2.1 2.0 1.5 1.8 1.6 2.3
4 Single row 50 X 10 cm 1.9 4.2 3.3 3.1 2.2 1.6 1.9 1.7 1.5 1.6 2.4
5 Single row 50 X 20 cm 1.8 4.6 3.6 3.3 1.5 1.5 1.5 1.8 1.3 1.9 2.4
6 Single row 50 X 30 cm 1.7 3.8 3.2 2.9 2.2 1.4 1.8 1.8 1.6 1.7 2.3
7 Single row 60 X 10 cm 1.7 3.2 2.4 2.4 1.2 2.1 1.7 2.0 1.7 1.8 2.1
8 Single row 60 X 20 cm 1.9 3.3 2.8 2.7 2.8 2.2 2.5 1.5 1.5 1.5 2.3
9 Single row 60 X 30 cm 1.7 4.8 2.9 3.1 2.0 3.1 2.5 1.9 1.7 1.8 2.7
10 Single row 70 X 10 cm 1.7 3.9 3.0 2.9 1.9 2.2 2.1 1.3 1.5 1.4 2.3
11 Single row 70 X 20 cm 1.7 3.8 2.5 2.7 1.4 1.5 1.5 1.3 1.7 1.5 2.0
12 Single row 70 X 30 cm 1.1 3.7 3.1 2.6 1.1 2.4 1.7 1.3 1.4 1.4 2.2
13 Single row 80 X 10 cm 1.8 3.4 2.5 2.6 2.0 1.7 1.9 1.5 1.7 1.6 2.2
14 Single row 80 X 20 cm 1.8 3.2 3.4 2.8 2.3 2.8 2.5 1.3 1.6 1.5 2.4
15 Single row 80 X 30 cm 1.7 4.0 2.8 2.8 1.8 3.1 2.5 1.4 1.8 1.6 2.5
Double row 50 X 30 X20 cm
1.5 3.8 2.9 2.8 1.8 2.1 2.0 1.7 1.6 1.6 2.3
16 (Control)
Mean 1.70 3.76 2.84 2.76 1.84 2.10 1.97 1.54 1.6 1.57 2.28
F-test NS NS NS NS ** NS * NS NS NS *
CV(%) 35.5 18.73 26.55 25.02 24.01 36.61 32.03 21.12 20.52 20.85 25.95
LSD(5%) NS NS NS NS 1.35 NS 0.7274 NS NS NS 0.3888
On the other hand, though there was no significant difference between treatment
combinations on the parameter at the rest locations, treatment combination 50x20
gave the highest seed yield both at Melkassa across years (3.3g/umbel)and Werer
across years (1.9 g/umbel) (Table 5). Hence, such varying responses of treatments
may have resulted in their interaction effects on the seed yield per umbel across
locations and years. The reason for the interaction of the treatments could be due
[405]
to the variable performance of pollinators, the effect of pesticide on pollinators,
unfavorable climatic conditions for pollination, seed setting and maturation
processes (Lemma, 1998).
Seed yield per plant: Analysis of variance showed that there was highly
significant interaction (p < 0.01) between treatment combinations across locations
and years for inter and intra- row spacing on seed yield per plant. This showed that
the treatment combinations responded differently across locations and years on the
parameter. Accordingly, in 2018 at Melkassa treatment combination 80x30 highly
significantly (p < 0.01) gave higher seed yield per plant (34.6g) than the control
treatment (18.2g). However, in 2018 at Kulumsa, treatment combination 60x30
significantly (p < 0.01) gave higher seed yield per plant (14.6g) than the control
(9.2g). on the other hand in 2019 at the same location, treatment combination
80x10 significantly gave higher seed yield per plant (14.6 g/plant) than most of
the treatment combinations; but it was not significantly different from the control
(10.6g/plant). In addition, in 2019 at Werer, treatment combination 80x30
significantly (p < 0.01) gave the highest seed yield per plant (18.3g) of all
treatment combinations including control (11.9g) (Table 6). The reason for
treatment interaction could be due to differences in soil and climatic conditions,
crop management and pollinating insects of specific locality and/or season.
Regardless of the treatment interaction, generally, the overall result showed that
treatment combinations with wider intra and inter-row spacing gave better seed
yield per plant as compared to treatment combinations with narrow intra and inter-
row spacing.
Table 6: Effect of inter and intra -row spacing on onion seed yield per plant (g) across locations and years
1 Single row 40 X 10 cm 2.5 16.1 6.4 8.3 7.6 6.7 7.2 12.9 8.8 10.8 9.7
2 Single row 40 X 20 cm 2.4 18.5 8.6 9.8 9.2 5.1 7.2 8.1 4.7 6.4 9.0
3 Single row 40 X 30 cm 4.3 22.6 10.8 12.6 7.5 6.3 6.9 9.4 9.1 9.3 11.0
4 Single row 50 X 10 cm 3.2 16.6 6.3 8.7 7.7 5.3 6.5 9.1 10.3 9.7 9.2
5 Single row 50 X 20 cm 3.5 21.8 7.2 10.8 4.0 6.2 5.1 7.7 11.6 9.7 9.8
6 Single row 50 X 30 cm 3.8 23.3 11.5 12.9 9.7 5.1 7.4 11.1 9.9 10.5 11.8
7 Single row 60 X 10 cm 2.6 17.8 8.1 9.5 5.7 10.0 7.9 8.3 6.0 7.2 9.3
8 Single row 60 X 20 cm 3.3 19.5 4.8 9.2 14.1 9.4 11.8 11.8 8.2 10.0 11.3
9 Single row 60 X 30 cm 5.5 28.0 11.4 15.0 14.6 7.1 10.9 10.4 8.8 9.6 13.4
10 Single row 70 X 10 cm 2.1 21.0 7.8 10.3 4.2 9.0 6.6 6.9 12.3 9.6 10.2
11 Single row 70 X 20 cm 3.4 25.1 7.6 12.0 8.3 8.1 8.2 9.3 8.9 9.1 11.2
12 Single row 70 X 30 cm 3.9 31.2 5.7 13.6 9.4 13.6 11.5 4.5 10.1 7.3 12.4
13 Single row 80 X 10 cm 3.2 23.2 9.6 12.0 10.2 14.6 12.4 8.5 9.5 9.0 12.6
14 Single row 80 X 20 cm 3.1 26.1 9.2 12.8 8.7 10.4 9.6 8.2 7.9 8.1 11.8
15 Single row 80 X 30 cm 2.7 34.6 9.4 15.6 13.4 9.6 11.5 10.0 18.3 14.2 15.9
Double row 50 X 30 X20
3.3 18.2 9.4 10.3 9.2 10.6 9.9 9.2 11.9 10.6 11.4
16 cm (Control)
Mean 3.2979 22.73 8.37 11.47 8.98 8.56 8.77 9.1 9.77 9.44 11.25
F-test NS ** NS ** ** ** ** NS ** ** **
CV(%) 41.7 21 38.08 29.95 19.13 30.21 25.02 33.01 12.22 24.45 26.87
LSD(5%) 14.52 3.2143 5.23 4.3106 2.5311 1.9911 2.663 1.988
[406]
Seed yield per ha: Analysis of variance showed that there was significant
for inter and intra- row spacing on seed yield per ha (Table 7). Like most of the
yield related parameters, treatments responded differently during each year at each
location. Accordingly, in 2017 at Melkassa treatment combination 80x10
significantly (p < 0.01) gave higher seed yield per ha (13.5qha-1) than the control
(4.17qha-1); but in 2018 on the same location the above treatment exceeded by
treatment combination 50x10 (16.6qha-1). This treatment gave comparable seed
yield with the control (20 q ha-1); however, there was no significant difference
between them. Moreover, in 2019 on the same site, instead of the above
treatments, treatment combination 60x10 gave higher seed yield (16.6 q ha-1).
However, there was no significant difference from the control (13.6 q ha-1). On the
other hand, at Kulumsa, treatment combination 40x10 gave significantly (p <
0.01) higher seed yield both in 2018 (5.8 q ha-1) and in 2019 (9.4 q ha-1) than
several treatments; though statistically the result was not different from that of
control in both years.
Similarly, at Werer, treatment combination 40x10 gave comparable seed yield per
ha (3.1 q ha-1) (P > 0.05) with the control (3.8 q ha-1) in 2018; whereas,
significantly (p < 0.01) gave higher seed yield (12 q ha-1) than the control (4 q ha-
1
) in 2019 at the same location (Table 7). Most of the plant growth factors such as
soil fertility, PH, Soil moisture and temperature, crop management, pollinating
insects, climatic conditions of the specific locality and cropping season could
affect the treatments differently and resulted in the interaction of treatments for
inter and intra -row spacing. It is known that onion is cross-pollinated crop with
crossing percentage of more than 94% where its pollination is mostly facilitated
by pollinating insects like bees. Therefore the different performances of the
treatments could be due to a low number of pollinators in the area and/or
unsuitable climatic conditions during the time of pollination, and fertilization.
Warmer and drier conditions with ample sunshine are desirable for seed set, maturity and
harvesting; whereas, unfavorable weather conditions such as excessive rainfall and cooler
conditions during flowering lead to diseases and poor seed set (Lemma, 1998) and
(Lemma and Shimeles, 2003). Martin et al., 1976 was also indicated that the fertility
status of the soil, growth pattern of the crop and cultural practices are among the factors
influencing spacing. Regardless of treatment interaction, in general, at all locations and
years, treatment combinations with narrower intra- row spacing gave superior seed yield
per ha regardless of inter-row spacing. Singh and Singh (2002) indicated that closely
spaced crops can intercept more light within a short period gives higher yield as compared
to wider spaced crops. These authors reported that as plant density increases, the amount
of dry matter in vegetative parts also increases; besides, both the biological and economic
yields increase with increasing plant population up to a certain point and subsequently no
addition in biological yield can be obtained and economic yield decreases.
Therefore, in order to come up with conclusive results and recommendation, partial

economic analysis would be described as below.
[407]
Table 7 : Effect of inter and intra -row spacing on onion seed yield (qha-1)across locations and years

1 Single row 40 X 10 cm 8.4 16.1 15.9 13.5 5.8 9.4 7.6 3.1 12 7.6 10.4
2 Single row 40 X 20 cm 5.3 15.7 10.7 10.6 5.1 8.3 6.7 2.1 3.2 2.7 7.5
3 Single row 40 X 30 cm 5.5 13.4 12.1 10.3 4.8 7.5 6.2 2.1 4.3 3.2 7.4
4 Single row 50 X 10 cm 10.3 16.6 12.6 13.2 5.5 6 5.7 2.2 7.9 5 8.5
5 Single row 50 X 20 cm 6.8 16.4 7.2 10.1 3.2 8.4 5.8 2 4.8 3.4 7
6 Single row 50 X 30 cm 4.2 12.6 5.1 7.3 2.6 4.7 3.7 1.6 3.1 2.4 4.9
7 Single row 60 X 10 cm 11.6 15.7 16.6 14.6 3.8 7.6 5.7 2 5.5 3.8 8.5
8 Single row 60 X 20 cm 5.8 12 5.2 7.7 4.3 7.7 6 1.5 3.6 2.6 5.7
9 Single row 60 X 30 cm 5.1 12.5 6.4 8 4.2 5.4 4.8 1.4 2.6 2 5.4
10 Single row 70 X 10 cm 10.3 14.4 11.3 12 3.2 6.7 4.9 2.2 3 2.6 6.8
11 Single row 70 X 20 cm 3.5 12.5 6.7 7.6 2.8 5.9 4.3 1.6 4.8 3.2 5.7
12 Single row 70 X 30 cm 3.1 10.2 3.3 5.5 2.7 6.2 4.4 0.65 2.8 1.7 4.3
13 Single row 80 X 10 cm 13.5 12.5 14.6 13.5 3.6 7.6 5.6 1.5 3.3 2.4 7.2
14 Single row 80 X 20 cm 5.4 11.8 5.8 7.7 3.4 4 3.7 1.5 3.1 2.3 4.9
15 Single row 80 X 30 cm 6.7 9.2 3.9 6.6 3.8 6.2 5 0.98 3.9 2.5 4.7
Double row 50 X 30 4.2 20.4 13.6 12.7 4.9 8.8 6.9 3.8 4 3.9 9.2
16 X20 cm (Control)
Mean 6.856 13.87 9.43 10.06 3.97 6.9 5.44 1.9 4.49 3.19 6.76
F-test ** ** ** ** ** ** ** ** ** ** **
CV(%) 23.7 10.79 21.13 17.11 20.55 22.39 23.36 29.73 13.87 20.79 19.8
LSD(5%) 2.71 4.733 3.323 1.611 2.48 2.576 1.466 1.719 1.038 0.766 0.88
[408]
Economic evaluation: The results of the partial budget analysis showed that
treatment combination 60x10 was superior (1,126,675 birr) and followed by
80x10 (1,100,000 birr) and control (1,020,000 birr) in terms of net income at
Melkassa. However, at Kulumsa, treatment combination 40x30 (557,324 birr)
gave higher net income and followed by the control (548,000 birr) and 60x20
(533,324 birr). At Werer, treatment 40x10 (650,000 birr) was higher in net income
and followed by 50x10 (385,000 birr) and the control (337,500 birr) (Table 8).
Table 8: Partial economic analysis for onion seed yield at Melkassa, Kulumsa and Werer
Net Income (Birr)

No. Treatments Melkassa Kulumsa Werer
1 Single row 40 X 10 cm 850,000 352,000 650,000
2 Single row 40 X 20 cm 810,000 524,000 187,500
3 Single row 40 X 30 cm 863,325 557,324 300,000
4 Single row 50 X 10 cm 920,000 236,000 385,000
5 Single row 50 X 20 cm 810,000 472,000 305,000
6 Single row 50 X 30 cm 596,675 294,676 220,000
7 Single row 60 X 10 cm 1,126,675 310,676 275,000
8 Single row 60 X 20 cm 603,325 533,324 225,000
9 Single row 60 X 30 cm 688,900 451,568 183,333
10 Single row 70 X 10 cm 914,275 267,988 153,572
11 Single row 70 X 20 cm 617,150 356,008 314,285
12 Single row 70 X 30 cm 454,750 421,320 155,357
13 Single row 80 X 10 cm 1,100,000 392,000 150,000
14 Single row 80 X 20 cm 645,000 304,000 212,500
15 Single row 80 X 30 cm 576,675 506,676 262,500
Double row 50 X 30 X20 cm
1,020,000 548,000 337,500
16 spacing (Control)
Seed quality parameters

Thousand seeds weight (g): There was no significant interaction (p > 0.05)
between treatment combinations across locations and years on thousand seed
weight; besides the main effects of treatment combinations were also not
significantly affected the parameter at all locations. However, thousand seed
weight was ranging from 3.6 to 3.9 at Melkassa, 3-3.9 at Kulumsa and 3.8 to 4.3 g
at Werer; with average thousand weights of 3.76, 3.24 and 4.07 g at Melkassa,
Kulumsa and Werer respectively (Table 9).
[409]
Table 9: Effect of inter and intra -row spacing on thousand seeds weight (g) across locations and years
No. Treatment Melkassa Kulumsa Werer Overall

1 Single row 40 X 10 cm 3.8 3.3 4.2 3.8
2 Single row 40 X 20 cm 3.7 3.1 4.2 3.7
3 Single row 40 X 30 cm 3.8 3.1 4.2 3.7
4 Single row 50 X 10 cm 3.9 3.5 3.9 3.8
5 Single row 50 X 20 cm 3.7 3.1 3.8 3.5
6 Single row 50 X 30 cm 3.8 3.4 4.2 3.8
7 Single row 60 X 10 cm 3.8 3.2 4.1 3.7
8 Single row 60 X 20 cm 3.8 3.6 4.0 3.8
9 Single row 60 X 30 cm 3.8 3.2 4.2 3.7
10 Single row 70 X 10 cm 3.5 3.4 4.3 3.7
11 Single row 70 X 20 cm 3.9 2.9 4.0 3.6
12 Single row 70 X 30 cm 3.6 3.6 4.0 3.7
13 Single row 80 X 10 cm 3.8 3.3 4.1 3.7
14 Single row 80 X 20 cm 3.8 2.9 3.9 3.5
15 Single row 80 X 30 cm 3.7 3.1 4.1 3.6
16 Double row 50 X 30 X20 cm (Control) 3.8 3.3 4.2 3.8
Mean 3.76 3.24 4.07 3.69
F-test NS NS NS NS
CV(%) 7.02 14.77 10.71 11.33
Germination percentage: There was no significant interaction (p > 0.05)

between treatment combinations across locations and years on seeds germination
percentage; besides the main effects of treatment combinations were also not
significantly affected the parameter at all locations (Data not shown).
Percent disease infection: Treatment combinations with narrow intra and inter-
row spacing relatively showed higher infection of purple blotch disease than wider
intra and inter-row spacing treatments (Table 10). At Melkassa the disease
infestation was higher for the control treatment. This could be due to the fact that
while plants are closely spaced, the microclimate would be favorable for the
development and occurrence of leaf diseases as the result of shading from adjacent
plants.
[410]
Table 10: Overall mean assessment of disease severity of onion varieties to purple blotch across locations and years
Disease Score (0-5)

No. Treatments Melkassa Kulumsa
1 Single row 40 X 10 cm spacing 0.9 1.6
4 Single row 50 X 10 cm spacing 1 1.3
15 Single row 80 X 30 cm spacing 1.1 1
16 Double row 50 X 30 X20 cm spacing (Control) 1.8 1.4
Mean 0.97 1.06
Note: Scoring scale system
0= No symptom, healthy whole plant parts and 5= ≥76% of leaves area affected
Conclusions
The experiment was conducted at Melkassa, Kulumsa and Werer agricultural
research centers from 2017 to 2019 with 15 treatment combinations and a control
in factorial arrangement using RCBD in three replications to determine optimum
plant population density for maximum onion seed yield and quality under current
agro-climatic conditions of targeted areas. Accordingly, inter and intra -row
spacing combinations affected onion growth, seed yield and quality differently.
Treatment interaction significantly affected the number of flower stalk per plant,
umbel diameter, seed yield per umbel, seed yield per plant and seed yield per ha;
whereas, days to 50% bolting, 50% flowering, percent seed setting, days to
maturity, thousand seeds weight and seeds germination percentage were not
affected due to effect of inter and intra row spacing.
Generally, the result indicated that treatments with wider intra-row spacing gave a
better number of flower stalks per plant; whereas, treatment combinations with
narrower intra- row spacing gave a higher number of flower stalks per plant
regardless of inter-row spacing. Though the treatments highly interacted across
locations and years, generally, the result revealed that treatments with wider intra
and inter-row spacing gave bigger umbel size and better seed yield per plant as
compared to treatment combinations with narrow intra and inter-row spacing.
From the result, it was observed that intra and inter-row spacing treatment
combinations didn’t show an effect on seed maturity at all test locations and years.
According to the result of the current study the results of the partial budget
analysis, treatment combinations 60x10, 40x30, and 40x10 were superior at
[411]
Melkassa, Kulumsa and Werer respectively with respect to net income of
1,126,675, 557,324 and 650,000 birr in that order.
Treatment combinations with narrow intra and inter-row spacing showed higher
infestation of leaf disease infection than wider intra and inter-row spacing
treatments.
Recommendations
Considering the economic value, seed yield, quality, and less disease severity
performances of the treatments at each location the following recommendations
could be given to produce onion seed. Treatment combination with 60cm inter-
row and 10cm intra- row spacing could be recommended for Melkassa and similar
agro-ecologies. However, 40cm inter-row and 30cm intra- row spacing is
recommended for Kulumsa and similar agro-ecologies. Treatment combination
with 40cm inter-row and 10cm intra- row is recommended for Werer and similar
areas sharing the same agro-ecologies.
Since population density is dependent on the changing climatic conditions of the

area, moisture availability, nutrient status of the soil and diseases and insect
prevalence in the locality, optimum planting density should be assessed and
determined periodically through conducting experiments that embrace fertilizer
and irrigation studies.
Therefore in order to improve onion seed production in the country, the current
recommended inter and intra-row spacing should be promoted for onion seed
production in the central rift valley and similar areas of the country.
Acknowledgment
Research for financing this research work. Furthermore, the authors are thankful
to vegetable research technical staffs at Melkassa, Kulumsa and Werer
Agricultural Research Centers for their unreserved effort during the execution of
the trials. We are also thankful to Melkassa, Kulumsa and Werer Agricultural
Research Centers management for supporting and facilitating this research work.
References
Bleasdale JKA. 1966. Plant growth and crop yield. Annals of Applied Biology issue No. 58,
pp.160-162.
CIMMYT. 1988. “From Agronomic Data to Farmer Recommendation: An Economic Work Book”
D.F: pp. 31- 33.
[412]
CSA. 2013. "Report on area and production of major crops, peasant holdings ‘Meher’ season. ."
Central Statistical Authority, Addis Ababa, Ethiopia.
CSA. 2014. "Report on area and production of major crops, peasant holdings ‘Meher’ season. ."
Central Statistical Authority, Addis Ababa, Ethiopia.
CSA, (Central Statistics Agency). 2015. The Federal Democratic Republic of Ethiopia.
Agricultural Sample Survey. Report on Area and Production of Major Crops for Meher
Season. 2014/15 (2007 E.C). Statistical Bulletin No 578.
Season. 2015/16 (2008 E.C).
Season. 2016/17 (2009 E.C).
Season. 2017/18 (2010 E.C). Statistical Bulletin No 586.
Horticultural Department Annual progress Reports. 1978/79, 1979/80, 1980/81 and 1981/82.
Reports on onion seed production spacing trials. Pp. 174-178. In: W. Godfrey-Sam-Aggrey
and Bereket-Tsehai Tuku eds., proceedings of the first Ethiopian Horticultural Workshop, 20-
22 February, 1987. IAR, Addis Ababa, Ethiopia
Koller HR and JM Scholl. 1968. effect of row spacing and seeding rate on forage production and
chemical composition of two sorghum cultivars harvested at two cutting frequencies. Agron. J.
60: 456-459.
Lemma Desalegne, Seifu Gebre Mariam and E. Herath. 1994. Seed production studies on
vegetables with particular reference to onion: p. 276-287. In Proc. of the 2nd National
Horticultural Workshop (E. Herath and Lemma D. eds.) 1-3 Dec. 1992, Addis Ababa, Ethiopia
Lemma Desalegne. 1998. Seed Production Guideline for Tomatoes, Onion and Hot Pepper, IAR,
Addis Abeba, Ethiopia.
Lemma D and Shimels A. 2003. "Research Experiences in Onion Production " Rep. No. 55
Ethiopian Agricultural Research Organization, A.A., Ethiopia
Lemma D, Shimels A, Selamawit K, and Abyote A. 2008. Varital development of major
vegetables in the rift valley region. In "Proceedings of the Inaugural & First Ethiopian
Horticultural Science Society " (D. Lemma, Endale G., Hail Michael K/Mariam, Zenebe W.,
Tefere B., Asfaw Z. and Lakew B., ed.), Vol. I, p. 86. EHSS, Addis Ababa, Ethiopia.
MoALR, (Ministry of Agriculture and Livestock Resource). 2017. Plant Variety Release,
Protection and Seed Quality Control Directorate .Crop Variety Register. Issue No. 20. Addis
Ababa, Ethiopia. Pp.235-244.
Norman AG. 1963. Competition among crop and pasture plants, Adv in Agron. 15: 1114,
Singh N.P., and R.A. Singh, 2002. Scientific crop production, X press Graphics, Delhi-28, 1st ed.,
India
[413]
Effect of Plant Spacing and Harvesting Age on

Growth, Yield and Yield Components of Lavender
(Lavandula angustifolia L.) under Supplemental
Irrigation at Wondo Genet, Southern Ethiopia
Basazinew Degu1* and Sulti Amano1
1
Ethiopian Institute of Agricultural Research, WondoGenet Agricultural Research Center, P.O. Box 198,
Shashemene, Ethiopia; *Corresponding author`s email: basdegu@gmail.com
Abstract
The experiment was carried out at Wondo Genet Agricultural Research Center
experimental field of Southern Ethiopia during 2017 and 2018 cropping seasons with
supplement of irrigation. The objectives of the study were to investigate the optimum
plant spacing and harvesting age for growth, yield and yield components of lavender
under supplemental irrigation conditions. The experiment consisted of four levels of
intra-row spacing (45, 60, 75 and 90 cm) with inter-row spacing of 60 cm and three
levels of harvesting ages after transplant (9, 10 and 11 MAT). The experiment laid
out in Randomized Complete Block Design (RCBD) with three replications as a
factorial arrangement. The result of combined analysis revealed that plant spacing
had a significant effect on the number of branches per plant, fresh leaves and flower
weight/plant, fresh leaves and flower yield/ha and essential oil yield/ha. On the
contrary, it had no significant effect on plant height and essential oil content.
Unlikely, harvesting age had a significant effect on plant height in addition to fresh
leaves and flower weight/plant, flower yield/ha and essential oil yield/ha; but not on
number of branches/plants. The interaction effect of plant spacing and harvesting
age had a significant effect on fresh leaves and flower yield/ha and essential oil
yield/ha. The highest fresh leaves and flower yield/ha (24.56 t) and essential oil
yield/ha (98.72 kg) were obtained from a spacing of 60*45 cm and at 10 MAT.
Whereas, the least essential oil yield/ha (34.42 kg) was obtained from the spacing of
60*90 cm and at 11 MAT. Based on the overall results use of plant spacing 60*45cm
and 10 MAT with irrigation is recommended for lavender for Wondo Genet and its
vicinity.
Keywords: Essential oil content, Essential oil yield, Harvesting age, Lavandula
angustifolia L., Plant spacing
Introduction
Lavender (Lavandula angustifolia L.) is a perennial shrub that belongs to the mint
family Lamiaceae. It is native to southern Europe and the Mediterranean area and
is commercially cultivated in France, Spain, Portugal, Hungary, the UK, Bulgaria,
Australia, China and the USA (Shawl and Kumar, 2000). The plant is cultivated
primarily for its aromatic inflorescence from which the essential oil is isolated,
[415]
although its fresh and dried flowers are also marketed (Renaud et al., 2001). The
essential oil of lavender is commonly used in aromatherapy and massage. Its
major clinical benefits are on the central nervous system (Chu and Kemper, 2001).
Besides, the essential oil lavender is also known for its excellent aroma and is
extensively used in the perfumery, flavor and cosmetic industries. The oil is
known to possess sedative, carminative, anti-depressive and anti-inflammatory
properties (Cavanagh and Wilkinson, 2005). Chu and Kemper (2001) also stated
that Lavender extracts have traditionally been prescribed to treat infertility,
infection, anxiety and fever, and have been used as antidepressants,
antispasmodics, anti-flatulent agents, antiemetic remedies and diuretics. Its
essential oil has gained a strong reputation in aromatherapy and as a holistic
relaxant to treat stress, anxiety, depression, fatigue, or insomnia (Chu and Kemper,
2001). Studies suggest that lavender aroma during recesses prevents deterioration
of work performance (Sakamoto et al., 2005) and might improve memory and
cognition in Alzheimer’s patients (Adsersen et al., 2005).
The growth, yield and yield components of plants are determined by a series of
factors including plant genetic (Shafie et al., 2009), climate, edaphic, elevation,
topography and also an interaction of various factors (Rahimmalek et al., 2009).
Among these factors crop management practices such as harvesting age and plant
spacing are included. Plant spacing and harvesting age affected growth, biomass
and oil yield of Artemisia (Zewdinesh et al., 2011) and Rose-scented geranium
(Haileslassie and Kebede, 2015). Harvesting age affected yield and yield-related
traits of lemongrass (Jimayu et al., 2016; Lulie and Chala, 2016; Jimayu and
Gebre, 2017).
Even though it has many uses, there is a limited finding regarding the agronomic
management practices of lavender cultivating under supplemental irrigation
conditions worldwide including Ethiopia. This would result in the cultivation of
the crop without knowing its appropriate management practices. This leads to a
reduction in the production and productivity of lavender. Though, research is
useful to minimize the reduction of growth, yield and yield components of
lavender. Therefore, the objective of this paper was to investigate optimum plant
spacing and harvesting age for growth, yield and yield component of L.
angustifolia L. under supplemental irrigation condition.

Description of the study area: The study was conducted at Wondo Genet
Agricultural Research Center experimental field, Southern Ethiopia during 2017
and 2018 cropping seasons with supplement of irrigation. The center is located at
latitude 7°19'N and longitude 38°38' E with an altitude of 1780 meters above sea
level (m.a.s.l.). The area receives a mean annual rainfall of 1128 mm with mean
[416]
maximum and minimum temperatures of 26.2 0C and 11 0C, respectively. The soil
textural class of the study area is sandy loam with a pH of 6.4 (Abayneh et al.,
2006). These environmental conditions are conducive for lavender cultivation.
Plant materials, experimental design and field management: Lavender

(Variety WG-Lavender-II) was planted at different intra spaced (45, 60, 75 and 90
cm) with inter-raw spacing of 60 cm at Hawassa in South Ethiopia were harvested
at three different months after transplant (9, 10 and 11 MAT) used for the study.
The experiment was laid by using a randomized complete block design with three
replications as a factorial arrangement. The gross plot size of each treatment was
3.6m x 3.6m. The distance between plots and replications was 1m and 2m,
respectively. Healthy and uniform seedlings were transplanted from the nursery to
the open field condition after two months of establishment. Transplanting of
seedlings was taken place at the commencement of the main rainy season after the
land prepared well. Proper hoeing, watering and weeding were carried out as
required.
Data collection: During the study, the following data were collected. The
characters that demonstrated for data collection are:
Plant height (cm): It was measured in centimeter from the base of the randomly
selected plants to the tip of the main stem by using tape meter; then the average
height was determined.
Number of branches per plant: The total numbers of branches arising from the
main stem were counted manually and the average value was determined.
Fresh leaf and flower weight/plant (g): The average fresh leaf and flower weight
of the randomly selected plants was immediately recorded after the leaves and
flowers were separated from the stem.
Fresh leaves and flower yield (t/ha): All plants from the central rows of each plot
were harvested and fresh leaf and flower yield per net plot was estimated and then
converted into tones per hectare.
Essential oil content (%): Essential oil content was obtained by hydro-distillation,
according to the procedure described by Daniel et al. (2009).The fresh leaves and
flower of lavender were placed in a round bottom flask and subjected to hydro-
distillation in a Clevenger apparatus. Then, harvested plants were separated into
leaf, flower and stem, fresh leaves and flowers having biomass of 300 g composite
sample was charged in the Clevenger apparatus along with 700 ml of water and
trapped for 3 hours. Water was poured into the flask until the plant part submerged
completely. The flask was placed on the heating mantle and the water and plant
[417]
sample could boil for 3 hours and the essential oil was collected and measured by
using pipette reading. The percentage of essential oil content was determined by
the following formula.
( )
Essential oil yield (kg/ha): The volume of essential oil collected in the collecting
tube of the apparatus dehydrated, measured and expressed on weight by weight
(%w/w) fresh basis. Then the essential oil yield/ha was determined by the
following formula.
( )
Data analysis
The collected data were subjected to analysis of variance (ANOVA) using SAS
version 9.0 (SAS, 2000) statistical software packages. Means were separated using
the Least Significant Difference (LSD) procedure at the 5% and 1% level of
significance.
The following meteorological data were collected from Hawassa from 2017 to
2018 cropping seasons (Figure 1).
Mean monthly rainfall (mm)

9.00 8.11
7.49
8.00 7.34 7.31
6.58
7.00
5.55
6.00 5.18 5.35
Rainfall (mm)
4.63
5.00 4.09
201
4.00 2.76 3.46 7
2.61 2.77 2.52
3.00 201
2.00 2.03 8
2.00 1.17 1.16
1.00 0.46
0.00 0.00 0.00
0.00
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Month
Figure 1. Mean monthly rainfall (mm) at Wondo Genet during 2017 and 2018 cropping seasons
(Source: NMA Hawassa Branch)
[418]
Plant height had not significantly (p>0.05) affected by the main effect of plant
spacing. Supporting results obtained by El Naim et al. (2012) who reported that
crop density had no significant effect on plant height of roselle plant. Contrary
result reported by Zewdinesh et al. (2011) on Artemisia and Degu and Tesfaye
(2015) on roselle plant who reported an increase in plant population markedly
would increase plant height. In 2018 cropping season and the combined analysis
result, harvesting age had a significant (p<0.001) effect on plant height (Table 1).
Similar results were reported by Zewdinesh et al. (2011) on Artemisia and
Haileslassie and Kebede (2015) on rose-scented geranium. From the combined
analysis of the result, the highest (84.03cm) and the least (64.63 cm) plant height
was obtained at 10 and 9 MAT, respectively (Table 1). As harvesting age
increased up to 10 MAT, plant heights also increased, and then declined when
harvested at a later age. This might be due to the drying of flower stalk at 11
MAT. The interaction effect of plant spacing and harvesting age did not
significantly (p>0.05) affect plant height (Table 1). A similar result was reported
by Haileslassie and Kebede (2015) on rose-scented geranium.
Table 1: Table 1. Effect of plant spacing and harvesting age on plant height (cm) and number of branches per plant of L.
angustifolia L. at Wondo Genet during 2017 to 2018
Plant height (cm) Number of branches per plant

Combined Combined
Treatment & Statistics 2017 2018 mean 2017 2018 mean
Plant spacing (cm)
60*45 91.31 57.09 74.2 90.24b 168.62 129.43b
60*60 92.09 57.29 74.69 106.09a 177.29 141.69ab
60*75 97.98 54.18 82.19 116.87a 195.04 155.96a
60*90 90.91 55.22 73.07 114.91a 202.09 158.5a
LSD (0.05) 5.65 5.38 10.4 12.84 33.11 19.8
Significance NS NS NS *** NS *
Harvesting age (MAT)
9 91.88 37.37b 64.63b 86.60c 171.17b 128.88b
10 93.02 65.88a 84.03a 111.23b 163.05b 137.14b
11 94.32 64.58 a 79.45 a 123.25 a 223.07 a 173.16a
LSD (0.05) 4.89 4.66 9.01 11.12 28.68 17.15
Significance NS *** *** *** *** ***
Spa * HA NS NS NS NS NS NS
CV (%) 6.21 9.84 14 12.28 18.23 13.84
Means with the same letter at the same column are not significantly different.
Where, MAT = Months after Transplanting, LSD = Least Significance Difference, CV = Coefficient of Variation, Spa =
Plant spacing, HA= Harvesting age, NS= Nonsignificant at 0.05 probability level, * = Significant at 0.05 probability level
and *** =Significant at 0.001 probability level.
The result of number of branches per plant varied in 2017 and 2018 cropping
seasons. In 2017 season, number of branches per plant significantly (p<0.001)
affected by the main effect plant spacing, but not affected in 2018. However, the
[419]
combined analysis result showed, plant spacing had a significant (p<0.05) effect
on the number of branches per plant (Table 1). In line with this similar result was
reported by Zewdinesh et al. (2011) on Artemisia and Haileslassie and Kebede
(2015) on rose-scented geranium. The highest (116.87 and 158.5) number of
branches per plant was obtained when the tested crop was planted at a spacing of
60*75 cm (in 2017) and 60*90 cm (on the combined result), respectively. Based
on the combined result, the number of branches per plant increased in wider
spacing (Table 1). This could be due to; at wider spacing, there will be a reduced
competition of plants for available resources such as light, water and nutrients.
This create favorable condition to increase develop the number of branches in a
given plant. Likewise, harvesting age had a significant (p<0.001) effect on number
of branches per plant (Table 1). This result is in line with the finding of
Haileslassie and Kebede (2015) on rose-scented geranium. Contrasting result was
reported by Zewdinesh et al. (2011) on Artemisia. The combined analysis result
revealed that the highest (173.16) and the least (128.88) number of branches per
plant was obtained at 11 and 9 MAT, respectively (Table 1). As later harvesting
age, the number of branches per plant increased. This might be due to the
development of branches on the main stem from the existing lateral buds. The
interaction effect of plant spacing and harvesting age did not significantly
(p>0.05) affect number of branches per plant (Table 1).A similar result was
reported by Zewdinesh et al. (2011) on Artemisia and Haileslassie and Kebede
(2015) on rose-scented geranium.
In each of the cropping season and combined analysis result, fresh leaves and
flower weight per plant was significantly (p≤0.01) affected by plant spacing
(Table 2). Based on the combined analysis result, the highest (644.67 g) and least
(454.89 g) fresh leaves and flower weight per plant was obtained at 60*90 cm and
60*45 cm plant spacing, respectively (Table 2). At the widest plant spacing
competition of plants for light, water and minerals reduced. This would attribute to
the increase in fresh leaves and flower weight per plant. Harvesting age had a
significant (p<0.001) effect on fresh leaves and flower weight per plant (Table 2).
Similar results were reported by Beemnet et al., (2011) on peppermint and
Tadesse et al., (2016) on stevia. Based on the combined analysis result, the highest
(749.79 g) and least (450.29 g) fresh leaves and flower weight per plant was
obtained at 10 and 9 MAT, respectively (Table 2). At later harvest (11 MAT), the
fresh leaves and flower weight per plant was started to decline. This could be due
to the drying of most of the already developed flower stalks and older leaves on
each plant. The interaction effect of plant spacing and harvesting age did not
significantly (p>0.05) influence fresh leaves and flower weight per plant (Table
2).
Plant spacing had a significant (p<0.001) effect on fresh leaves and flower
yield/ha (Table 2).A similar result was reported by Haileslassie and Kebede
[420]
(2015) on rose-scented geranium and Jimayu et al. (2016) on Lemongrass. Based
on the combined analysis of the result, the highest (16.85 t) and least (11.94 t)
fresh leaves and flower yield/ha was obtained at 60*45 cm and 60*90 cm plant
spacing, respectively. At the closest spacing, there are many plants per unit area.
This would attribute to the increase in fresh leaves and flower yield/ha. Harvesting
age had also a significant (p<0.001) effect on fresh leaves and flower yield/ha
(Table 2). Similar results were reported by Beemnet et al. (2011) on peppermint
and Haileslassie and Kebede (2015) on rose-scented geranium. The combined
analysis result revealed that the highest (19.59 t) and least (11.44 t) fresh leaves
and flower yield/ha were obtained at 10 and 9 MAT, respectively (Table 2). At
later harvest (11 MAT), the fresh leaves and flower yield/ha was started declining.
As to fresh leaves and flower weight per plant, this could be due to the drying of
most of the already developed flower stalks and older leaves on each plant.
The interaction effect of plant spacing and harvesting age did not significantly
(p>0.05) affect fresh leaves and flower yield/ha in 2018. Whereas, it had a
significant (p≤0.05) effect on fresh leaves and flower yield/ha in 2017 and
combined result, respectively (Table 2). The result obtained in this study is
consistent with the report of Haileslassie and Kebede (2015) on rose-scented
geranium and Jimayu and Gebre (2017) on Lemongrass. Based on the combined
analysis result, the highest (24.56 t) fresh leaves and flower yield/ha was obtained
when lavender was planted at 60*45 cm plant spacing and harvested at 10 MAT;
whereas, the least (9.79 t) fresh leaves and flower yield/ha was obtained when it
was planted at 60*90 cm plant spacing and harvested at 9 MAT followed by 11
MAT (Table 4).
Table 2. Effect of plant spacing and harvesting age on fresh leaves and flower weight per plant (g) and fresh leaves and flower yield
(t/ha) of L. angustifolia L. at Wondo Genet during 2017 and 2018 cropping seasons
Fresh leaves and flower weight per plant (g) Fresh leaves and flower yield (t/ha)
Treatment and Statistics
2017 2018 Combined mean 2017 2018 Combined mean
Plant spacing (cm)
60*45 451.04b 458.73c 454.89c 16.70a 16.99a 16.85a
60*60 617.92a 539.48b 578.7b 17.17a 14.99a 16.07a
60*75 651.48a 546.87ab 599.18ab 15.08a 12.66b 13.87b
60*90 672.17a 617.17a 644.67a 12.45b 11.43b 11.94c
LSD (0.05) 98.23 76.99 57.03 2.45 2.09 1.53
Significance *** ** *** *** *** ***
9 577.45b 323.12c 450.29c 14.62b 8.27c 11.44c
10 712.71a 786.88a 749.79a 18.71a 20.47a 19.59a
11 504.30b 511.68b 507.99b 12.72b 13.31b 13.02b
LSD (0.05) 85.24 66.67 49.39 2.12 1.81 1.32
Significance *** *** *** *** *** ***
Spa * HA NS NS NS * NS **
CV (%) 16.83 14.57 10.24 16.33 15.27 10.65
Where, MAT = Months after transplanting, LSD = Least significance difference, CV = Coefficient of variation, Spa = Plant spacing, HA=
Harvesting age, NS = Nonsignificant at 0.05 probability level, * = Significant at 0.05 probability level, ** = Significant at 0.01 probability
level and *** =Significant at 0.001probability level
[421]
Essential oil content was not significantly (p>0.05) affected by plant spacing in
both cropping seasons and combined results (Table 3). Supporting result was
reported by Lule and Chala (2016) who showed that spacing would not affect the
essential oil content of lemongrass. Whereas, harvesting age exerted significant
(p<0.05) effect on essential oil content in the 2018 cropping season. A similar
result was reported by Haileslassie and Kebede (2015) on rose-scented geranium.
The highest (0.51%) and least (0.33%) essential oil content was obtained at 9 and
11 MAT, respectively (Table 3). This could be due to the senescence/drying of
most of the flowers and older leaves of the plant when it was harvested at 11
MAT. But, harvesting age did not significantly (p>0.05) affect essential oil
content in 2017 and the combined result (Table 3). Essential oil content was not
significantly (p>0.05) affected by the interaction effect of both factors (Table 3).
This result in line with, the findings of Beemnet et al. (2011) on peppermint and
Zewdinesh et al. (2011) on Artemisia. Contrasting result was reported by
Haileslassie and Kebede (2015) on rose-scented geranium.
Essential oil yield/ha was not significantly (p>0.05) affected by plant spacing in
the 2017 cropping season. But it did a significant (p<0.01) effect on essential oil
yield/ha in 2018 cropping season and combined result, respectively (Table 3). A
similar result was reported by Zewdinesh et al. (2011) on Artemisia; Haileslassie
and Kebede (2015) on rose-scented geranium and Jimayu et al. (2016) on
Lemongrass. Based on the combined result, the highest (65.04 kg) and least (41.47
kg) essential oil yield/ha was obtained at 60*45cmand 60*90 cm plant spacing,
respectively (Table 3). This might be due to an increase in fresh leaves and flower
yield/ha in closest spacing would attribute to an increase in essential oil yield/ha.
Harvesting age did not significantly (p>0.05) affect essential oil yield/ha in the
2017 cropping season. But it did a significant (p<0.001) effect on essential oil
yield/ha in 2018 cropping season and the combined result (Table 3). A similar
result was reported by Haileslassie and Kebede (2015) on rose-scented geranium.
From the combined result, the highest (78.08 kg) and least (43.92 kg) essential oil
yield/ha was obtained when lavender was harvested at 10 and 11MAT,
respectively (Table 3). This could be due to the senescence/drying of most of the
flowers and older leaves of the plant when it was harvested at 11 MAT. This
would result in the decline of essential oil yield/ha.
The interaction effect of plant spacing and harvesting age did not significantly
(p>0.05) affect essential oil yield/ha in 2017 cropping season, whereas, it had
significant (p<0.05) effect on essential oil yield/ha in 2018 cropping season and
the combined result (Table 3). A similar result was reported by Haileslassie and
Kebede (2015) on rose-scented geranium. The highest (98.72 kg) essential oil
yield/ha was obtained when lavender was planted at 60*45 cm plant spacing and
harvested at 10 MAT; whereas, the least (34.42 kg) essential oil content was
[422]
obtained when it was planted at 60*90 cm plant spacing and harvested at 11 MAT
(Table 4).
Table 3. Effect of plant spacing and harvesting age on essential oil content (%) and essential oil yield/ha (kg) of
L. angustifolia L. at Wondo Genet during 2017 and 2018
Essential oil content (%) Essential oil yield/ha (kg)
Treatment & Statistics Combined
2017 2018 2017 2018 Combined mean
mean
Plant spacing (cm)
60*45 0.36 0.42 0.39 59.41 70.67a 65.04a
60*60 0.32 0.46 0.39 54.58 71.30a 62.94a
60*75 0.33 0.48 0.41 49.4 58.17ab 53.79ab
60*90 0.32 0.39 0.36 39.88 43.06b 41.67b
LSD (0.05) 0.12 0.11 0.08 20.88 19.72 13.06
Significance NS NS NS NS * **
9 0.33 0.51a 0.42 48.59 42.48b 45.43b
10 0.31 0.47a 0.39 60.08 96.07a 78.08a
11 0.35 0.33b 0.35 43.79 44.06b 43.92b
LSD(0.05) 0.1 0.09 0.07 18.08 17.08 11.31
Significance NS ** NS NS *** ***
Spa * HA NS NS NS NS NS *
CV (%) 36.42 24.56 20.86 42.02 33.18 23.94
Where, MAT = Months after transplanting, LSD = Least significance difference, CV = Coefficient of variation, Spa = Plant
spacing, HA= Harvesting age, NS = Nonsignificant at 0.05 probability level, * = Significant at 0.05probability level, ** =
Significant at 0.01 probability level and *** =Significant at 0.001probability level
Table 4. Interaction effect of plant spacing and harvesting age on fresh leaves and flower yield (t/ha) and essential oil
yield (kg/ha) of L. angustifolia at Wondo Genet during 2017 and 2018 cropping seasons
Fresh leaves and flower yield (t/ha) Essential oil yield (kg/ha)
Treatment and
Harvesting age (MAT) Harvesting age (MAT)
Statistics
9 10 11 9 10 11
Plant spacing (cm)
60*45 11.67efg 24.56a 14.32d 44.55bc 98.72a 53.73bc
60*60 12.88def 21.14b 14.20de 49.41bc 98.27a 41.14bc
60*75 11.43fg 17.46c 12.71def 51.50bc 61.58b 48.28bc
60*90 9.79g 15.20cd 10.83fg 36.26c 53.73bc 34.42c
LSD (0.05) 2.65 22.62

The combined result revealed that plant spacing had a significant effect on the
number of branches per plant, fresh leaves and flower weight/plant, fresh leaves
and flower yield/ha and essential oil yield/ha. The highest number of branches per
plant (158.5) and fresh leaves and flower weight/plant (644.67 g) was obtained at
the spacing of 60*90 cm; fresh leaves and flower yield/ha (16.85 t) and essential
oil yield/ha (65.04 kg) was obtained at a spacing of 60*45cm. Whereas, harvesting
age had a significant effect on all tested parameters except essential oil content.
[423]
The highest plant height (84.03 cm), fresh leaves and flower weight/plant (749.79
g), fresh leaves and flower yield/ha (19.59 kg), and essential oil yield/ha (78.08
kg) were obtained at 10 MAT; however, the least plant height (64.63cm), fresh
leaves and flower weight/plant (450.29 g), fresh leaves and flower yield/ha (11.44
kg) were obtained at 9 MAT; and essential oil yield/ha (43.92 kg) was obtained at
11 MAT followed by 9 MAT. Moreover, the interaction effect did not
significantly affect all the tested parameters except fresh leaves and flower
yield/ha and essential oil yield/ha. The highest fresh leaves and flower yield/ha
(24.56 t) and essential oil yield/ha (98.72 kg) were obtained when lavender was
planted at 60*45 cm and harvested at 10 MAT. Whereas, the least essential oil
yield/ha (34.42 kg) was obtained when it was planted at 60*90 cm and harvested
at 11 MAT. Therefore, to get the highest essential oil yield/ha of lavender at
Wondo Genet and a place where having the same agro-ecologies under
supplemental irrigation conditions, it is better to plant it at a spacing of 60*45 cm
and harvested at 10 MAT is highly recommended.
Acknowledgment
Ethiopian Institute of Agricultural Research (EIAR) and Wondo Genet
Agricultural Research Center are acknowledged for the financial and technical
support of this study. We are also a thanks and appreciation to Mr. Yigermal
Molla and Mr. Admasu Usur for their technical support.
References
Abayneh Esayas, Demeke Tafesse and Ashenafi Ali. 2006. Soils of Wondo Genet Agricultural
Research Center. National Soil Research Center, 67 p.
Adsersen A, Gauguin B, Gudiksen L and Jager AK. 2005. Screening of plants used in Danish folk
medicine to treat memory dysfunction for acetylcholinesterase inhibitory activity. Journal of
Ethnopharmacology 104:418-422.
Beemnet MK, Solomon AM and Jaime ATda Silva. 2011. Agronomic Characters, Leaf and
Essential Oil Yield of Peppermint (Mentha piperita L.) As Influenced by Harvesting Age and
Row Spacing. Medicinal and Aromatic Plant Science and Biotechnology 5(1) 49 – 53.
Cavanagh HMA and Wilkinson JM. 2005. Lavender essential oil: a review healthcare infection.
Aust. Infect. Control 10, 35–37.
Chu CJ and KJ Kemper. 2001. Lavender (lavandula ssp.). Retrieved July 24, 2008 from
http://www.longwoodherbal.org/lavender/lavender.pdf.
Degu Basazinew and Tesfaye Bizuayehu. 2016. Effects of inter and intra row spacing on growth,
yield and yield components of Roselle (Hibiscus sabdariffa L.). International Journal of
Advanced Biological and Biomedical Research, 5(1), 260-274.
El Naim AM, Khaliefa EH, Ibrahim KA, Ismaeil FM, Zaied MB. 2012. Growth and yield of
Roselle (Hibiscus sabdariffa L.) as influenced by plant population in Arid tropic of Sudan
under rain-fed. Int. J. Agr. Forest., 2(3), 88-91.
Getachew Jimayu and Aynalem Gebre. 2017. Influence of harvesting age on yield and yield
related traits of lemongrass (Cymbopogon citratus L.) varieties at Wondo genet, southern
Ethiopia. Acad. Res. J. Agri. Sci. Res. 5(3): 210-215.
[424]
Haileslassie Gebremeskel and Kebede Woldetsadik. 2015. Effect of plant spacing and harvesting
age ongrowth, biomass and oil yield of rose-scented geranium (Pelargonium graveolens L.
Herit). Basic Research Journal of Agricultural Science and Review ISSN 2315-6880 Vol. 4(3)
pp. 094-101
Jimayu Getchew, Gebre Aynalem and Mengesha Beemnet. 2016. Response of Lemongrass
(Cymbopogon ctratus L) Varieties for Different Intra and Inter row spacing at Wondo genet,
Southern Ethiopia. Acad. Res. J. Agri. Sci. Res. 4(6): 279-284.
Lulie Belstie and Chala Mideksa. 2016. Influence of Plant Population Density on Growth and
Yield of Lemon Grass (Cymbopogon citratus L.) at Wondo Genet, South Ethiopia. Acad. Res.
J. Agri. Sci. Res. Vol. 4(3), pp. 76-84.
Rahimmalek M, Bahreininejad B, Khorrami M, and Sayed TBE. 2009. Genetic variability and
geographical differentiation in Thymus daenensissubsp. daenensis Cleak, an endangered
aromatic and medicinal plant as revealed by Inter Simple Sequence Repeat (ISSR) markers.
Biochem. Genet, 47, 831–842.
Renaud ENC, Charles DJ and Simon JE. 2001. Essential oil quantity and composition from 10
cultivars of organically grown lavender and lavandin. J. Essent. Oil Res., 13, 269–273.
Sakamoto, R., K. Minoura, A. Usui, Y. Ishizuka and S. Kanba, 2005. Effectiveness of aroma on
work efficiency: Lavender aroma during recesses preventsdeterioration of work performance.
Chemical Senses30:683-691.
SAS inst. 2000. SAS (Statistical Analysis System), 1996. SAS/STAT. Guide Version 9. SAS,
Institute Inc. Raleigh, North Carolina: USA.
Shafie MSB, Zain HSM and Shah MS. 2009. Study of genetic variability of Wormwood capillary
(Artemisia capillaris) using inter simple sequence repeat (ISSR) in Pahang region, Malaysia.
Plant Omics J, 2(3), 127–134.
Shawl AS, Kumar S. 2000. Potential of Lavender oil industry in Kashmir. J. Med. Arom. Plants,
22:319-321. J. Med. Arom. Plant Sci. 22 (2000) 319.
Tadesse Nibret, Gebere Aynalem, Lulie Belstie and Hordofa Melkamu. 2016. Influence of plant
population density on growth and yield of Stevia (Stevia rebaudiana Bertoni L.) at Wondo
Genet South Ethiopia. Acad. Res. J. Agri. Sci. Res. 4(6): 321-329
Zewdinesh Damtew, Bizuayehu Tesfaye and Daniel Bisrat. 2011. Leaf, Essential Oil and
Artemisinin Yield of Artemisia (Artemisia annua L.) As Influenced by Harvesting Age and
Plant Population Density. World Journal of Agricultural Sciences 7 (4): 404-412, 2011.
[425]
[426]
Effect of Harvesting Age and Plant Spacing on
Growth, Yield and Yield Component of Lavender
(Lavandula angustifolia L.) under Rainfed
Condition at Hawassa, Southern Ethiopia
Basazinew Degu1* and Sulti Amano1
1
Ethiopian Institute of Agricultural Research, Wondo Genet Agricultural Research Center,
P.O. Box 198, Shashemene, Ethiopia; *Corresponding author`s E-mail: basdegu@gmail.com
Abstract
The experiment was carried out at Hawassa, Southern Ethiopia during 2017 and
2018 cropping season under rainfed condition to investigate the optimum harvesting
age and plant spacing for growth, yield and yield component of lavender under
rainfed condition. The experiment consisted of four levels of intra-row spacing (45,
60, 75 and 90 cm) with inter-row spacing of 60 cm and three levels of harvesting ages
after transplanting (9, 10 and 11 MAT). The experiment laid out in Randomized
Complete Block Design (RCBD) with three replications as a factorial arrangement.
The results of combined analysis of variance revealed that spacing had a significant
influence on the number of branches per plant, fresh leaves and flower weight/plant,
fresh leaves and flower yield/ha and essential oil yield/ha. The highest number of
branches/plant (180.7) and fresh leaves and flower weight/plant (755.39 g) was
obtained at a spacing of 60*90 cm; fresh leaves and flower yield/ha (21.12 t) and
essential oil yield/ha (119.2 kg) was obtained at a spacing of 60*45cm. Whereas,
harvesting age exerted a significant influence on essential oil content and essential oil
yield in kg/ha, but not on the other traits. The highest essential oil content (0.67 %)
and essential oil yield/ha (115.5 kg) were obtained at 11 MAT. Moreover, the
interaction effect did not significantly influence all the tested parameters. Therefore,
it was concluded that further investigation is needy by adding additional levels of
harvesting age above 11MAT to investigate at which age the essential oil yield/ha will
be declined. But, for the time being until another investigation obtained, at Hawassa
and a place where having similar agroecology, it is highly recommended to plant
Lavender at a spacing of 60*45 cm and harvest it at 11 MAT to get the highest
essential oil yield/ha.
Keywords: Essential oil content, Essential oil yield, Harvesting age, Plant
spacing
Introduction
Lavender (Lavandula angustifolia L.) is a perennial shrub that belongs to the
mint family Lamiaceae. It is native to southern Europe and the
Mediterranean area and is commercially cultivated in France, Spain,
Portugal, Hungary, the UK, Bulgaria, Australia, China and the USA (Shawl
and Kumar, 2000). It is cultivated for its aromatic inflorescence from which
[427]
the essential oil is isolated, although its fresh and dried flowers are also
marketed (Renaud et al., 2001).
The essential oil of this species is commonly used in aromatherapy and

massage. Its major clinical benefits are on the central nervous system (Chu
and Kemper, 2001). The essential oil is also known for its excellent aroma
and is extensively used in the perfumery, flavor and cosmetic industries. The
oil is known to possess sedative, carminative, anti-depressive and anti-
inflammatory properties (Cavanagh and Wilkinson, 2005). Chu and Kemper
(2001) also stated that Lavender extracts have traditionally been prescribed
to treat infertility, infection, anxiety and fever, and have been used as
antidepressants, antispasmodics, anti-flatulent agents, antiemetic remedies
and diuretics. Its essential oil has gained a strong reputation in aromatherapy
and as a holistic relaxant to treat stress, anxiety, depression, fatigue, or
insomnia (Chu and Kemper, 2001). Studies suggest that lavender aroma
during recesses prevents deterioration of work performance (Sakamoto et al.,
2005) and might improve memory and cognition in Alzheimer’s patients
(Adsersen et al., 2005).
The growth, yield and yield components of plants are determined by a series
of factors including plant genetic (Shafie et al., 2009), climate, edaphic,
elevation, topography and also an interaction of various factors (Rahimmalek
et al., 2009). Among these factors crop management practices such as
harvesting age and plant spacing are included. Plant spacing and harvesting
age affected growth, biomass and oil yield of Artemisia (Zewdinesh et al.,
2011) and Rose-scented geranium (Haileslassie and Kebede, 2015).
Harvesting age affected yield and yield-related traits of lemongrass (Jimayu
et al., 2016; Lulie and Chala, 2016; Jimayu and Gebre, 2017).
Even if it has many uses, there is a limited finding regarding the agronomic
management practices of lavender cultivating under rainfed conditions in
worldwide including Ethiopia. This would result in the cultivation of the crop
without knowing its appropriate management practices. This leads to a
reduction in the production and productivity of lavender. Though, research is
useful to minimize the reduction of growth, yield and yield components of
lavender. Therefore, the objective of this paper was to investigate optimum
harvesting age and plant spacing for growth, yield and yield component of
Lavandula angustifolia L. under rainfed condition.

Description of the study area: The study was conducted at Hawassa
Agricultural Research Center experimental field, Southern Ethiopia during
[428]
2017 and 2018 cropping seasons under rainfed condition. The center is
located at latitude 7°05'N and longitude 39°29' E with an altitude of 1700
meters above sea level (masl). The area receives a mean annual rainfall of
964 mm with a mean maximum and minimum temperatures of 27.340C and
12.940C, respectively. The soil textural class of the study area is sandy loam
with a pH of 7.2. These environmental conditions are conducive for lavender
cultivation.
Plant materials, experimental design and field management: Lavender

(Variety WG-Lavender-II) was planted at different intra spaced (45, 60, 75
and 90 cm) with inter-raw spacing of 60 cm at Hawassa in South Ethiopia
were harvested at three different months after transplant (9, 10 and 11 MAT)
used for the study. The experiment was laid by using a randomized complete
block design with three replications as a factorial arrangement. The gross
plot size of each treatment was 3.6m x 3.6m. The distance between plots and
replications was 1 m and 2 m, respectively. Healthy and uniform seedlings
were transplanted from the nursery to the open field condition after two
months of establishment. Transplanting of seedlings was taken place at the
commencement of the main rainy season after the land prepared well. Proper
hoeing, watering and weeding were carried out as required.
Data collection: During the study, the following data were collected. The
characters that manifested for data collection are:
Plant height: It was measured in centimeter from the base of the randomly
selected plants to the tip of the main stem by using tape meter; then the
average height was determined.
Number of branches per plant: The total numbers of branches arising from
the main stem were counted manually and the average value was determined.
Fresh leaf and flower weight/plant (g): The average fresh leaf and flower
weight of the randomly selected plants was immediately recorded after the
leaves and flowers were separated from the stem.
Fresh leaves and flower yield (t/ha): All plants from the central rows of each
plot were harvested and fresh leaf and flower yield per net plot was estimated
and then converted into tones per hectare.
Essential oil content (%): Essential oil content was obtained by hydro-
distillation, according to the procedure described by Daniel et al. (2009).The
fresh leaves and flower of lavender were placed in a round bottom flask and
subjected to hydro-distillation in a Clevenger apparatus. Then, harvested
[429]
plants were separated into leaf, flower and stem, fresh leaves and flowers
having biomass of 300 g composite sample was charged in the Clevenger
apparatus along with 700 ml of water and trapped for 3 hours. Water was
poured into the flask until the plant part submerged completely. The flask
was placed on the heating mantle and the water and plant sample could boil
for 3 hours and the essential oil were collected and measured by using pipette
reading. The percentage of essential oil content was determined by the
following formula.
( )
Essential oil yield (kg/ha): The volume of essential oil collected in the
collecting tube of the apparatus dehydrated, measured and expressed on
weight by weight (%w/w) fresh basis. Then the essential oil yield/ha was
determined by the following formula.
( )
Data analysis: The collected data were subjected to analysis of variance

(ANOVA) using SAS version 9.0 (SAS, 2000) statistical software packages.
Means were separated using the Least Significant Difference (LSD)
procedure at the 5% and 1% level of significance.
The following meteorological data were collected from Hawassa from 2017
to 2018 (Table 1).
[430]
Table 1. Mean monthly maximum and minimum temperature (oC); and Mean monthly rainfall (mm) at Hawassa
trial site during2017 and 2018 cropping seasons
Mean monthly temperature (oC) Mean monthly rainfall

Month 2017 2018 (mm)
Max. Min. Max. Min. 2017 2018
January 29.00 9.75 29.18 11.43 0.00 0.14
February 29.90 14.49 30.23 13.28 3.16 2.19
March 31.56 14.49 28.34 13.93 2.94 4.22
April 31.60 14.78 26.65 15.63 2.95 9.20
May 28.40 15.82 27.76 15.65 7.78 3.55
June 27.73 15.49 25.12 15.50 1.95 6.30
July 26.10 16.03 34.16 15.16 5.49 7.59
August 26.61 15.53 34.08 15.20 3.84 5.41
September 26.34 15.56 25.68 14.34 6.01 2.20
October 27.94 15.12 27.75 13.86 4.29 1.68
November 28.49 11.69 27.50 13.04 0.21 4.99
December 28.32 9.43 28.29 11.65 0.00 0.22
(Source: NMA Hawassa Branch)

Plant height (cm): Plant height was significantly (p<0.05) affected by plant
spacing in the 2017 cropping season. In the 2017 cropping season, the
highest (85.11 cm) and least (77.33 cm) plant height was obtained at 60*75
cm and 60*45 cm, respectively (Table 2). Whereas, it didn`t significantly
(p>0.05) affected by plant spacing in the 2018 cropping season and the
combined result (Table 2).Supporting result obtained by ElNaim et al. (2012)
who showed that crop density had no significant effect on plant height of
roselle. Contrasting result was reported by Zewdinesh et al. (2011) on
Artemisia and Degu and Tesfaye (2015) on Roselle who showed that an
increase in planting population markedly would increase plant height.
Harvesting age did not significantly (p>0.05) affect plant height (Table 2).
Contrasting result was reported by Zewdinesh et al. (2011) on Artemisia and
Haileslassie and Kebede (2015) on rose-scented geranium. This could be due
to the consistency of plant height at each level of harvesting age. The
interaction of plant spacing and harvesting age had a significant (p<0.05)
effect on plant height in the 2018 cropping season (Table 2). The highest
(79.8 cm) plant height was obtained when lavender was planted at plant
spacing of 60*90 cm and harvested at 10 MAT; while the least (65.73 cm)
was obtained when it was planted at plant spacing of 60*90 cm and
harvested at 11 MAT (Table 5). Based on the combined result, the interaction
effect did not significantly (p>0.05) affect plant height (Table 2). A similar
result was reported by Haileslassie and Kebede (2015) on rose-scented
geranium.
[431]
Table 2. Effect of plant spacing and harvest age on plant height (cm) and number of branches per plant of
L. angustifolia L. at Hawassa during 2017 and 2018
Plant height (cm) Number of branches per plant

Treatment and Combined Combined
Statistics 2017 2018 mean 2017 2018 mean
Plant spacing (cm)
60*45 77.33b 72.04 74.69 97.64 174.5b 136.09c
60*60 78.40b 73.22 75.81 107 210.76a 158.88b
60*75 85.11a 71.42 78.27 113 221.51a 167.26ab
60*90 82.31ab 71.44 76.88 120.2 241.2a 180.7a
LSD (0.05) 5.65 5.6 4.74 16.29 32.39 20.41
Significance * NS NS NS ** **
Harvest age (MAT)
9 78.22 71.02 74.62 104.60 220.28 162.44
10 83.75 72.72 78.23 113.02 222.80 167.91
11 80.4 72.37 76.38 110.77 192.92 151.84
LSD (0.05) 4.89 4.85 4.1 14.11 28.05 17.67
Significance NS NS NS NS NS NS
Spa*HA NS * NS NS NS NS
CV (%) 7.15 7.95 6.34 15.22 15.63 12.99
Means with the same letter at the same column are not significantly different. Where, MAT = Months after
transplanting, LSD = Least significance difference, CV = Coefficient of variation, Spa = Spacing, HA= Harvesting
age, NS = Nonsignificant at p<0.05, * = Significant at p<0.05 and ** =Significant at p<0.01 probability level
Number of Branches per plant: Number of branches per plant didn`t

significantly (p>0.05) affected by plant spacing in 2017 cropping season;
whereas, it was significantly (p<0.01) affected by plant spacing in 2018
cropping season and the combined result (Table 2). Based on the combined
result, the highest (180.7) and least (136.09) number of branches per plant
was obtained at 60*90 cm and 60*45 cm plant spacing, respectively (Table
2). At the widest spacing, there is reduced competition of plants for light,
water and nutrients as compared to the closest spacing. Though, this would
increase the number of branches per plant. The main effect of harvesting age
did not significantly (p>0.05) affect the number of branches per plant (Table
2). Supporting result was reported by Zewdinesh et al. (2011) on Artemisia.
Likewise, the interaction of both factors did not significantly (p>0.05) affect
the number of branches per plant (Table 2). A similar result was reported by
Zewdinesh et al. (2011) on Artemisia and Haileslassie and Kebede (2015) on
rose-scented geranium.
Fresh leaves and flower weight/plant (g): Fresh leaves and flower
weight/plant was significantly (p<0.01) affected by plant spacing during the
cropping seasons (Table 3). The same result was reported by Tadesse et al.
(2016) on Stevia. Based on the combined result, the highest (755.39 g) and
least (570.08 g) fresh leaves and flower weight/plant was obtained at 60*90
cm and 60*45 cm plant spacing, respectively (Table 3). As plant spacing
increases, fresh leaves and flower weight/plant was increased and vice versa.
This might be due to reduced competition of plants for light, water and
[432]
nutrients in widest spacing would result in an increase in fresh leaves and
flower weight/plant.
Fresh leaves and flower weight/plant were significantly (p<0.001) affected

by harvesting age in 2017 cropping season. The highest (598 g) and least
(456.02 g) fresh leaves and flower weight/plant was obtained at 60*90 cm
and 60*45 cm plant spacing, respectively (Table 3). Whereas, it didn`t
significantly (p>0.05) affected by harvesting age in 2018 cropping season
and the combined result (Table 3). Likewise, fresh leaves and flower
weight/plant were significantly (p<0.05) affected by the interaction of plant
spacing and harvesting age in the 2017 cropping season (Table 3). The
highest (782.05 g) fresh leaves and flower weight/plant was obtained when
lavender was planted at a spacing of 60*90 cm and harvested at 10 MAT;
whereas, the least (441.1 g) fresh leaves and flower weight/plant was
obtained when it was planted at a spacing of 60*45 cm and harvested at 10
MAT (Table 6). However, in the 2018 cropping season and the combined
result showed that the interaction of both factors did not significantly
(p>0.05) affect fresh leaves and flower weight/plant (Table 3).
Fresh leaves and flower yield (t/ha): Plant spacing had a significant
(p<0.001) effect on fresh leaves and flower yield/ha during the cropping
seasons and combined result (Table 3). Supporting result was reported by
Haileslassie and Kebede (2015) on rose-scented geranium and Jimayu et al.
(2016) on Lemongrass. Based on the combined result, the highest (21.12 t)
and least (13.99 t) fresh leaves and flower yield/ha was obtained at 60*45 cm
and 60*90 cm plant spacing, respectively (Table 3). At the closest spacing,
fresh leaves and flower yield was increased and vice versa. This might be
due to the presence of many plants per unit area, this would result in an
increase in fresh leaves and flower yield/ha.
As to fresh leaves and flower weight/plant, fresh leaves and flower yield/ha
was significantly (p<0.01) affected by harvesting age in 2017 cropping
season. The highest (15.09 t) and least (12.11 t) fresh leaves and flower yield
/ha were obtained at 10 and 9 MAT, respectively (Table 3). Whereas,
harvesting age had not a significant (p>0.05) effect on fresh leaves and
flower yield/ha in 2018 cropping season and the combined result (Table 3).
This result is contradicted with the findings of Beemnet et al. (2011) on
peppermint and Haileslassie and Kebede (2015) on rose-scented geranium.
The interaction of plant spacing and harvesting age not significantly (p>0.05)
affect fresh leaves and flower yield/ha (Table 3). A similar result was
reported by Beemnet et al. (2011) on peppermint. On the contrary, the
contradicted finding was reported by Haileslassie and Kebede (2015) on
rose-scented geranium.
[433]
Table 3. Effect of plant spacing and harvest age on fresh leaves and flower weight per plant (g) and fresh leaves
and flower yield (t/ha) of L. angustifolia L. at Hawassa during 2017 and 2018
Fresh leaves and flower weight/plant (g) Fresh leaves and flower yield (t/ha)
Treatment & Statistics
2017 2018 Combined mean 2017 2018 Combined mean
Plant spacing (cm)
60*45 451.89b 688.28b 570.08b 16.74a 25.49a 21.12a
60*60 498.06b 837.22ab 667.64a 13.84b 23.26ab 18.55b
60*75 565.93a 856.09a 711.01a 13.1bc 19.82bc 16.56b
60*90 615.42 a 895.36a 755.39 a 11.4c 16.58c 13.99c
LSD (0.05) 67 149.5 87.89 1.96 3.63 2.21
Significance *** * ** *** *** ***
Harvest age (MAT)
9 456.62b 807.17 631.89 12.11b 21.04 16.58
10 598a 815.17 706.58 15.09a 21.17 18.13
11 543.86a 835.38 689.62 14.1a 21.65 17.88
LSD (0.05) 58.02 129.47 76.11 1.7 3.14 1.92
Significance *** NS NS ** NS NS
Spa*HA * NS NS NS NS NS
CV (%) 12.86 18.67 13.3 14.6 17.44 12.9
Where, MAT = Months after transplanting, LSD = Least significance

difference, CV = Coefficient of variation, Spa = Spacing, HA= Harvesting
age, NS = Non significant at p<0.05, * = Significant at p<0.05, ** =
Significant at p<0.01 and *** =Significant at p<0.001probability level
Essential Oil Content (%): Essential oil content was significantly (p<0.05)
affected by plant spacing in the 2018 cropping season. The highest (0.54%)
and least (0.42%) essential oil content were obtained at 60*75 cm and 60*60
cm plant spacing, respectively (Table 4). Whereas, plant spacing had not a
significant (p>0.05) effect on essential oil content in the 2017 cropping
season and combined mean (Table 4).
Based on the combined result, harvesting age had a significant (p<0.001)

effect on essential oil content (Table 4). A similar result was reported by
Haileslassie and Kebede (2015) on rose-scented geranium. The highest
(0.67%) and least (0.54%) essential oil content was obtained when lavender
was harvested at 11 and 9 MAT, respectively. When harvesting age
increases, essential oil content also increased under rainfed conditions. This
could be due to the presence of more stressed leaves at 11 MAT. However,
the interaction of the plant spacing and harvesting age did not significantly
(p>0.05) affect essential oil content (Table 4). A similar result was obtained
by Beemnet et al., (2011) on peppermint and Zewdinesh et al. (2011) on
Artemisia. Contrasting result was reported by Haileslassie and Kebede
(2015) on rose-scented geranium.
[434]
Essential Oil Yield/ha (kg): The combined result revealed that the main
effect plant spacing and harvest age had a very highly significant (p<0.001)
effect on essential oil yield/ha (Table 4).Supporting result was reported by
Zewdinesh et al. (2011) on Artemisia; Haileslassie and Kebede (2015) on
rose-scented geranium and Jimayu et al. (2016) on Lemongrass. The highest
(119.2 t) and least (80.06 t) essential oil yield/ha was obtained when lavender
was planted at a spacing of 60*45 cm and 60*90 cm, respectively (Table 4).
This could be due to the presence of many plants per unit area contribute to
an increase in fresh leaves and flower yield/ha thereby increased in essential
oil yield/ha. The highest (115.5 t) and least (81.63 t) essential oil yield/ha
was obtained when Lavender was harvested at 11 and 9 MAT, respectively.
As to essential oil content, essential oil yield/ha increased when harvesting
age increases. This could be due to an increase in essential oil content that
would contribute to an increase in essential oil yield/ha. However, the
interaction effect of plant spacing and harvesting age did not significantly
(p>0.05) affect essential oil yield/ha (Table 4). Supporting result was
reported by Zewdinesh et al. (2011) on Artemisia. Contradicted results were
reported by Beemnet et al., (2011) on peppermint and Haileslassie and
Kebede (2015) on rose-scented geranium.
Table 4. Effect of plant spacing and harvest age on essential oil content (%) and essential oil yield/ha (kg) of
L. angustifolia L. at Hawassa during 2017 and 2018
Essential oil content (%) Essential oil yield/ha (kg)
Treatment & Statistics Combined
2017 2018 2017 2018 Combined mean
mean
Plant spacing (cm)
60*45 0.67 0.50ab 0.59 112.28a 126.11a 119.20a
60*60 0.69 0.42b 0.56 95.49ab 96.34bc 95.91b
60*75 0.69 0.54a 0.62 90.89 bc 106.19 b 98.54b
60*90 0.67 0.51a 0.59 76.54c 83.59c 80.06c
LSD (0.05) 0.08 0.09 0.07 17.67 14.83 9.5
Significance NS * NS ** *** ***
Harvest age (MAT)
9 0.64b 0.43b 0.54b 76.46b 86.80b 81.63c
10 0.67ab 0.45b 0.56b 101.56 a 94.74 b 98.15b
11 0.74a 0.60a 0.67a 103.37a 127.64a 115.50a
LSD (0.05) 0.07 0.07 0.06 15.3 12.84 8.23
Significance * *** *** ** *** ***
Spa*HA NS NS NS NS NS NS
CV (%) 12.66 17.61 11.68 19.27 14.72 9.88
Where, MAT = Months after transplanting, LSD = Least significance difference, CV = Coefficient of variation, Spa
= Spacing, HA= Harvest age, NS = Non significant at p<0.05, * = Significant at p<0.05, ** = Significant at p<0.01
and *** =Significant at p<0.001
[435]
Table 5. Effect of plant spacing and harvesting age on plant height (cm) of L. angustifolia L. at
Hawassa during 2018
Plant height (cm)

Treatment and Statistics Harvesting age(MAT)
9 10 11
Plant spacing (cm)
60*45 71.93abc 68.87bc 75.33abc
60*60 75.8ab 71.53abc 72.33abc
60*75 67.53bc 70.67abc 76.07ab
60*90 68.8 bc 79.8a 65.73c
LSD (0.05) 9.62
Where, MAT = Months after transplanting and LSD = Least significance difference
Where, MAT = Months after transplanting and LSD = Least significance difference
Table 6. Effect of plant spacing and harvesting age on fresh leaves and flower weight/plant
(g) of L. angustifolia L. at Hawassa during 2017
Fresh leaves and flower weight/plant (g)

Treatment & Statistics Harvesting age (MAT)
9 10 11
Plant spacing (cm)
60*45 446.88e 441.10e 467.69de
60*60 444.57e 532.79bcde 516.82cde
60*75 483.10de 636.05b 578.63bcd
60*90 451.91e 782.05a 612.31bc
LSD(0.05) 116.05

The study showed that the highest number of branches/plant (180.7) and
fresh leaves and flower weight/plant (755.39 g) was obtained at the spacing
of 60*90 cm; fresh leaves and flower yield/ha (21.12 t) and essential oil
yield/ha (119.2 kg) was obtained at a spacing of 60*45cm. Harvesting age
also had a significant influence on essential oil content and essential oil
yield/ha; but it did not significantly influence plant height, number of
branches per plant, fresh leaves and flower weight/plant and fresh leaves and
flower yield/ha. The highest essential oil content (0.67 %) and essential oil
yield/ha (115.5 kg) were obtained at 11 MAT; however, the least values
(0.54 % and 81.63 kg, respectively) were obtained at 9 MAT. The interaction
of plant spacing and harvesting age had not a significant influence on plant
height, number of branches per plant, fresh leaves and flower weight/plant,
fresh leaves and flower yield/ha, essential oil content and essential oil
yield/ha. Therefore, it was concluded that further investigation is needy by
adding additional levels of harvesting age above 11MAT to investigate at
which age the essential oil yield/ha will be declined. But, for the time being
until another investigation obtained, at Hawassa and a place where having
similar agroecology, it is highly recommended to plant Lavender at a spacing
of 60*45cm and harvest it at 11MAT to get the highest essential oil yield/ha.
[436]
Acknowledgment
Agricultural Research Center are acknowledged for the financial and
technical support of this study. We are also a thanks and appreciation to Mr.
Yigermal Molla and Mr. Admasu Usur for their technical support.
References
Adsersen A, B Gauguin, Gudiksen L and Jager AK. 2005. Screening of plants used in
Danish folk medicine to treat memory dysfunction for acetylcholinesterase inhibitory
activity. Journal of Ethnopharmacology 104:418-422.
Beemnet MK, Solomon AM and Jaime AT da Silva. 2011. Agronomic Characters, Leaf and
Essential Oil Yield of Peppermint (Mentha piperita L.) As Influenced by Harvesting
Age and Row Spacing. Medicinal and Aromatic Plant Science and Biotechnology 5(1)
49 – 53.
Cavanagh HMA and Wilkinson JM. 2005. Lavender essential oil: a review healthcare
infection. Aust. Infect. Control 10, 35–37.
Chu CJ and KJ Kemper. 2001. Lavender (lavandula ssp.). Retrieved July 24, 2008 from
http://www.longwoodherbal.org/lavender/lavender.pdf.
Daniel Bisrat, Solomon Abate and Wossen Kebede. 2009. Laboratory manual for plant
products analysis. Technical manual. Ethiopian Institute of Agricultural Research,1
(23): 4 – 6.
Degu Basazinew and Tesfaye Bizuayehu. 2016. Effects of inter and intra row spacing on
growth, yield and yield components of Roselle (Hibiscus sabdariffa L.). International
Journal of Advanced Biological and Biomedical Research, 5(1), 260-274.
El Naim, AM, Khaliefa, EH Ibrahim, KA Ismaeil, FM Zaied, MB. 2012. Growth and yield
of Roselle (Hibiscus sabdariffa L.) as influenced by plant population in Arid tropic of
Sudan under rain-fed. Int. J. Agr. Forest., 2(3), 88-91.
Haileslassie Gebremeskel and Kebede Woldetsadik. 2015. Effect of plant spacing and
harvesting age on growth, biomass and oil yield of rose-scented geranium (Pelargonium
graveolens L. Herit). Basic Research Journal of Agricultural Science and Review ISSN
2315-6880 Vol. 4(3) pp. 094-101
Jimayu Getachew, Gebre Aynalem and Mengesha Beemnet. 2016. Response of Lemongrass
(Cymbopogon citratus L) Varieties for Different Intra and Inter row spacing at Wondo
genet, Southern Ethiopia. Acad. Res. J. Agri. Sci. Res. 4(6): 279-284
Rahimmalek M, Bahreininejad B, Khorrami M, Sayed TBE. 2009. Genetic variability and
geographical differentiation in Thymus daenensissubsp. daenensis Cleak, an endangered
aromatic and medicinal plant as revealed by Inter Simple Sequence Repeat (ISSR)
markers. Biochem. Genet, 47, 831–842.
Renaud ENC, Charles DJ, Simon JE. 2001. Essential oil quantity and composition from 10
cultivars of organically grown lavender and lavandin. J. Essent. Oil Res., 13, 269–273.
Sakamoto R, K Minoura, A Usui, Y Ishizuka and S Kanba. 2005. Effectiveness of aroma on
work efficiency: Lavender aroma during recesses prevents deterioration of work
performance. Chemical Senses30:683-691.
SAS inst. 2000. SAS (Statistical Analysis System), 1996. SAS/STAT. Guide Version 9.
SAS, Institute Inc. Raleigh, North Carolina: USA.
Shafie MSB, Zain HSM and Shah MS. 2009. Study of genetic variability of Wormwood
capillary (Artemisia capillaris) using inter simple sequence repeat (ISSR) in Pahang
region, Malaysia. Plant Omics J, 2(3), 127–134.
[437]
Shawl AS, Kumar S. 2000. Potential of Lavender oil industry in Kashmir. J. Med. Arom.
Plants, 22:319-321. J. Med. Arom. Plant Sci. 22 (2000) 319.
Tadesse Nibret, Gebere Aynalem, Lulie Belstie and Hordofa Melkamu, 2016. Influence of
plant population density on growth and yield of Stevia (Stevia rebaudiana Bertoni L.) at
Wondo Genet South Ethiopia. Acad. Res. J. Agri. Sci. Res. 4(6): 321-329
Zewdinesh Damtew, Bizuayehu Tesfaye and Daniel Bisrat. 2011. Leaf, Essential Oil and
ArtemisininYield of Artemisia (Artemisia annua L.) As Influenced by Harvesting Age
and Plant Population Density. World Journal of Agricultural Sciences 7 (4): 404-412,
2011.
[438]
Determination of Optimum Spacing and Harvesting
Age of Citron Scented Gum (Eucalyptus citriodora
H.) for Maximum Biomass and Essential Oil Yield at
Wondo Genet, South Ethiopia
Desta Fikadu Egata1* and Wondimkun Dikir Desta2
1
Ethiopian Institute of Agricultural Research Department of Horticulture Wondo Genet Agricultural Research
Center, P.O. Box: 198, Shashemene, Ethiopia; 2Ethiopian Institute of Agricultural research Department of
Agronomy Wondo Genet Agricultural Research Center; *Corresponding author email:
destafikadu55@yahoo.com
Abstracts
The experiment was conducted at Wondo Genet south Ethiopia for three years. The
study was aimed for the determination of optimum plant spacing and harvesting age
on E.citriodora. The treatment consists of three levels of harvesting ages and four
levels of plant spacing. The character that manifested for data collection is: plant
height, leaf weight/plant stem weight/plant leaf to stem ratio, leaf and stem yield/ha,
essential oil content and essential oil yield/ha were collected. The whole plant had
been cut using sharp saw at 20 cm height above the ground and essential oil was
extracted from the leaf part. The results of the study revealed that harvesting age had
a very highly significant effect on plant height, fresh leaf to stem ratio, essential oil
content and essential oil yield/hectare. Spacing had a very high significant effect on
fresh leaf yield/hectare and fresh stem/hectare. The interaction effect of spacing and
harvesting age had a significant effect on essential oil content. The highest leaf
weight/pant was obtained with a spacing of 1.5 m and 2 m between plants and raw
respectively. Maximum fresh leaf yield/hectare was obtained with narrower spacing
(1.5 m between plant and row) and the highest essential oil yield/hectare was
obtained when the plant harvested at 10 months after transplanting but it was not
significantly different with harvesting age at 9 months after transplanting. Based on
the overall result use of plant spacing 1.5*2m and 9 to 10 MAT were recommended
for citron scented for Wondo Genet and its vicinity.
Keywords: Citron scented gum, harvesting age, planting space, Leaf and
essential oil
Introduction
Citron scented gum (Eucalyptus citriodora Hook) is belongs to the genus
Eucalyptus and family Myrtaceae (Orwa et al., 2009). It is commonly known as
‘Lemon-Scented Eucalyptus’ or ‘Lemon-Scented Gum and ‘Shitobahirzaf’ in
Amharic (Demel, 2000). The lemon-scented leaves originated from Australia
(Chen et al., 2007). It is extensively planted and coppiced for the extraction of
essential oil that is rich in citronellal and used in perfumery and as a flavouring
agent (Chen et al., 2007).
[439]
It is a multi-purpose crop used for; timber, charcoal, firewood, grown as
ornamental but commercially mainly cultivated for its essential oil distilled from
the leaves (Orwa et al., 2009). Its oil is widely used in a number of perfumery
formulations, toiletries and as disinfectants. The leaves reported to possess
antiseptic properties and are used in the treatment of various skin diseases. Oil of
E. citriodora reported to possess antibacterial, antifungal, ascaricidal and insect
repellent activities (Singh et al., 2002; Verbal et al., 2009; Luqman et al., 2008).
The oil is also observed to be phytotoxic and has the potential to be used as
herbicide (Batish et al., 2004, Singh et al., 2006). In traditional medicine, the
essential oil is used for the treatment of joint pains (Buchman et al., 1979). Due to
its aforementioned numerous advantages, E. citriodora is one of the most
commonly grown species across the world (Batish et al, 2006).
In Ethiopia, many different Eucalyptus species were introduced to during the reign
of Emperor Menilek II in 1895 to supply fuelwood and construction wood (Mesfin
and Wubalem, 2014). Its cultivation gradually spread throughout the country with
the effort of different organizations such as academic, research and development
institutions including Alemaya College of Agriculture, Institute of Agricultural
Research (IAR) and Chilalo Agricultural Development Unit (CADU) (Amare,
20010). Currently, different eucalyptus species were being grown for its multi-
purpose advantage to Ethiopian public and Wondo Genet Agricultural Research
Center was adapted and registered one variety of Citron scented gum for its aroma
of essential oil at a national level. Leaf and essential oil yield is highly influenced
by; location, season, nature of soil, age of plant and planting population density
(Sefidkon et al., 2009). Even though different Eucalyptus tree species had been
grown in Ethiopia for over a century and people developed traditional knowledge
for eucalyptus management, still there is knowledge and information gap to obtain
optimum essential oil yield of Citron scented gum (FAO, 2009). Despite different
Eucalyptus species reach for harvest after 5-8 years after planting for its wood
product, the determination of proper spacing and harvesting age of the leaves is
highly required for higher essential oil yield of E. citriodora. Because of seldom
research activities undertaken on this plant in Ethiopian condition for essential oil,
this experiment was aimed to determine optimum plant spacing and harvesting age
of citron scented gum.

The experiment was conducted at Wondo Genet Agricultural Research Center for
three years starting from 2017 up to 2019. This trial was consists of three levels of
harvesting ages (8, 9 and 10 months after transplanting) and four levels of plant
spacing (1.5 m*1.5 m, 1.5 m*2 m, and 2 m*2.5 m between plant and row,
respectively. Treatments arranged in factorial Randomized Complete Block
Design (RCBD) with three replications. Twelve treatments combined and each
[440]
treatment combination was assigned randomly within the block. A single plot area
was 80 m2 having 8m length and 10m width. A spacing of 1 m and 2 m was
maintained between plants and rows respectively.
Seedlings were raised in the nursery bed for three months using plastic pots and
transplanted on the main field at the onset of the rainy season. Hoeing and
weeding was done as required. Supplemental irrigation was applied during the
dry seasons when rain was not enough as well as where there was no rain at all.
The whole plant had been cut using sharp saw at 20 cm height above the ground
during harvesting. Harvesting was done for three consecutive years. The harvested
herb was distilled for the production of essential oil using hydro distillation in a
clavenger type apparatus according to Gunter (1972).
Both agronomic and chemical traits data were collected for this study. The
character that manifested for data collection are: plant height, leaf weight/plant
(kg), stem weight/plant (kg), leaf to stem ratio, leaf yield/ha (kg), stem yield/ha
(kg), essential oil content (%) and essential oil yield/ha were collected. The
collected data were subjected to analysis of variance (ANOVA) using SAS
version 9.3 (SAS, 2002) statistical software packages. Means were separated using
the Least Significant Difference (LSD) procedure at the 5% and 1% level of
significance.

Plant height: Plant height was very highly significantly (p<0.01) influenced by
age, plant population density and harvesting Cycle (Table 1). As plant population
density, harvesting age and harvesting cycle increased plant height was increased
(Table 2). Maximum plant height was obtained when plants planted with spacing
narrower (1.5 m x 1.5 m) and lower plant height were obtained from wider
spacing (2 m x 2.5 m) between plant and row respectively. This is due to plant
grown on narrower spacing was highly competed with each other for the sunlight.
This result was in line with the result reported by (Gebremedhin, 2015) on haricot
bean that Plants with a dense population were grown rapidly higher than small
population per a given cultivation land area.
Fresh leaf weight kg/plant: The result of this study indicated, fresh leaf
weight/plant was not significantly affected by plant spacing and harvesting age (p
> 0.05). Even though Fresh leaf weight/plant was not significantly affected by
plant population density and harvesting age, it was significantly influenced by the
harvesting cycle (Table 1). Fresh leaf weight per plant was increased from
harvesting cycle one (1.76 kg) to harvesting cycle three (4.12 kg) (Table 2). This
is maybe due to plants perennial character that plant coppiced and produce a
number of branches that bear tremendous leaves than the plant preexisting harvest.
[441]
Maximum fresh leaf weight/plant (2.79 kg) was recorded with the spacing of 1.5
m and 2 m between plant and row respectively harvested at the age of 10 months
after transplanting and minimum fresh leaf weight/plant (2.42 kg) was obtained
from the wider spacing (2 m and2.5m) between plant and row respectively
harvested at age of 8 months after transplanting (Table 2). This result revealed that
E.citriodora requires specific plant and row spacing for optimum leaf weight/plant
that not too narrow and wider. As the harvesting cycle increased fresh leaf
weight/plant also highly increased (from 1.23 kg to 4.72 kg). From this
experimental observation, optimum spacing for Citron scented gum for essential
oil production is 1.5 m * 2 m between plant and row respectively. Maximum leaf
weight/plant was obtained when harvested at the age of 10 months after
transplanting. This is due to the root of a plant well organized absolved enough
water and nutrient from deep soil. Ramamurthy et al., 2016 also reported a similar
result that older E.citriodora was a high yielder than the younger ones.
Fresh leaf yield/hectare/harvest: The result indicated that harvesting age had
no significant influence but plant population density highly significant effect (P<
0.01) on fresh leaf yield ha-1(Table 1). Even though harvesting age had no
significant effect on fresh leaf yield, it was increased as the harvesting age
increased (Table 2). Maximum leaf yield ha-1 of the single harvest was obtained at
age of 10 months after transplanting (7.37 tones) and the minimum was obtained
when harvested at age of 8 months after transplanting (6.55 tones) (Table 2). This
experiment was revealed that E. citriodora fresh leaf yield ha-1was increased as
plant population density becoming increased. Maximum leaf yield ha-1(7.41tons)
was obtained at narrower (1.5m*1.5m) spacing but not significantly different with
the spacing of 1.5m*2m between plant and row respectively. This is maybe due to
the maximum number of plants/hectare. Minimum Fresh leaf yield ha-1(5.04 tons)
was obtained from wider spacing (2m*2.5m) or lower population density. Akhtar
et al., (2008) was reported oppose this result that narrower spacing was lower
herbage biomass this is maybe due to further late harvesting age decreases leaf
yield since E. citriodora when plant grows further it highly compete for water and
nutrient. Harvesting cycle also highly influenced leaf yield at (P<0.01) and as
harvesting cycle increased leaf yield increased from 4.86 tons to 10.60 tons. This
experiment was confirmed that harvesting age and plant spacing highly affect leaf
of citron scented gum yield per hectare as reported by Ramamurthy et al., (2016).
Fresh stem weigh/plant: Fresh stem weight significantly influenced by plant

population density (planting space), harvesting age and cycle (Table 1). Stem
weight/plant was increased as harvesting age increased from 8 to 10 months after
transplanting. This is due to the eucalyptus perennial tree is a plant that rapidly
grow or increase its height and size after transplanting. Low fresh stem
weight/plant (2.23 kg) was obtained when the plant was harvested at 8 months
after transplanting and maximum weight (2.63 kg) was obtained at 10 months age
after transplanting. Even though wider spacing had no significant difference,
[442]
maximum weight was recorded with the spacing of 1.5 m and 2 m between plant
and row respectively. Akhtar et al., (2008) reported similar findings on plant
spacing that wider spacing the lower number of trees decrease stem or wood yield
per hectare.
Fresh leaf to stem ratio: Fresh leaf to stem ratio significantly affected (P<0.01)
by harvesting age and cycle (Table 1). Fresh leaf to stem ratio was decreased as
harvesting age increased and maximum Fresh leaf stem was recorded at the
spacing of 1.5 m and 2 m between plant and row respectively. Lower fresh leaf to
stem ratio was obtained at the early stage and late harvesting age. The maximum
fresh leaf to a ratio (1.33) was obtained at 9 months after transplanting. This result
revealed that too narrower space and too wider is not proper for E.citriodora leaf
production for essential oil. Since E. citriodora essential is extracted from leaves,
leaf to stem ratio is very important to get optimum essential oil yield. Similar
result reported by Zewdinesh et al., (2012) on rosemary that as harvesting age
increased leaf to stem ratio was increased. This result indicated that stem growth
change was proportionally higher than leaf growth. Based on this finding, leaf to
stem ratio can be used as a harvesting age indicator for citron scented gum
produced for essential oil extraction.
Essential oil content: The result indicated essential oil content of Citron scented gum
was highly influenced by harvesting age and harvesting cycle as well as by the interaction
of harvesting age and plant population density (Table 1). Maximum essential oil cont
(1.41 %) was obtained when harvested at 9 months age after transplanting having the
spacing 1.5 meters and 2 meters between plant and row respectively (Table 3). As plant
spacing becoming wider essential oil content becoming higher this is due to wider space
receive high sunlight than narrower spaces which highly influence essential oil content as
reported by Etemadi et al., (2018) that essential oil content decreased with increasing
shade levels the vice versa was reported by Saleh M. (1973) that essential oil content was
increased when plants are grown under full sunlight.
Essential oil yield/hectare/harvest: The essential oil yield/hectare/harvest had

significantly (p<0.01) influenced by plant population density, harvesting age and
harvesting cycle (Table 1). The essential oil yield was increased as harvesting age
increased from 71.31 kg to 94.43 kg with a mean (84.99 kg). This is due to plant leaf
yield increased as the age of a plant increased and also essential oil content is low at an
early stage. Maximum essential oil yield ha-1(94.43kg) was obtained at late harvest (10
MAT) and minimum (71.31kg) at early harvesting age (8 MAT). Plant spacing also the
other source of variation for E.citriodora essential oil yield that maximum yield was
obtained with the spacing 1.5*2m this due to maximum essential oil content was obtained
using this spacing. The harvesting cycle had also a very significant difference on essential
oil yield (Table 1). As the harvesting cycle increased from one to three essential oil yield/
hectare/harvest increased from 47.53 kg to 152.55 kg, this is due to fresh leaf yield and
essential oil content was increased as harvesting cycle increased that directly associated
with essential oil yield of the plant. This is in line with the result reported by Mengesha et
al., (2011) on Japanese mint and peppermint.
[443]
Table1. Analysis of variance for tested traits of Citron scented gum to determine optimum plant population density and harvesting age
SV DF PH PBN SBN FLWP FSWP FLSR FLYHH FSYHH EOC EOYHH

REP 2 0.034 32.7 1612.009 0.47 0.44 0.03 2774906.9 968146 0.02 161.89
HA 2 1.84*** 64 3811.74 1.87 1.46 0.33** 6157092.6 6941882 0.32** 5295.6***
SP 3 0.26** 140.63 3883.25 0.81 0.87 0.15 36148298.4*** 27208204*** 0.19 1808.27*
HC 2 8.31*** 483.76*** 55620*** 63.68*** 142.82*** 4.41*** 360950094.9*** 923917083*** 2.57*** 123724.56***
HA*SP 6 0.82 102.38 589.14 0.52 0.62 0.03 4565248.8 6864679 0.29*** 1414.05
Error 92 21.63 53.54 1270.04 0.415 0.41 0.049 3382937 3140780 0.062 704.63
CV (%) 9.5 19.46 24.55 24.91 26.52 16.95 26.44 27.66 21.41 31.23
Mean 2.31 37.6 145.15 2.59 2.42 1.31 6955.297 6406.768 1.16 84.99
Where: SV (Source of variance), DF (Degree of freedom), PH (Plant height), PBN (Primary branch number/plant), SBN(secondary branch number/plant). FLWP (Fresh leaf weight/plant),
Fresh stem weight/plant, FLSR (Fresh leaf to stem ratio) FLYHH (Fresh leaf yield/hectare per harvest), FSYHH (Fresh leaf yield/hectare per harvest), EOC (essential oil content), and
EOYHH (essential oil yield/ harvest)
[444]
Table 2.Mean value of traits tested to determine effect of plant population density and harvesting age of Citron scented gum for optimum essential oil yield
Harvesting age PH(m) PBN SBN FLWP(kg) FSWP(kg) FLSR FLYH(kg) FSYH(kg) EOC (%) EOYHH(kg)
Mean Mean Mean Mean Mean Mean Mean Mean Mean Mean
8 MAT 2.07c 36.27a 134.79ab 2.40b 2.23b 1.37a 6545.4a 5957.7a 1.06b 71.31b
9MAT 2.34b 37.58a 145.30b 2.52b 2.38ab 1.36a 6948a 6427.4a 1.24a 89.23a
10 MAT 2.51a 38.94a 155.37a 2.84a 2.63a 1.20b 7372.5a 6835.2a 1.19a 94.43a
LSD(0.05) 0.10 3.43 16.68 0.30 0.30 0.10 861.01 829.62 0.12 12.43
Spacing
1.5m*1.5m 2.38a 39.19a 153.81a 2.47ab 2.44ab 1.24b 8143.7a 7413.6a 1.04b 86.51ab
1.5m*2m 2.35a 38.91a 150.44a 2.79ab 2.56a 1.33ab 7471.0ab 6794.4ab 1.17ab 93.69a
2m*2m 2.34a 38.04ab 148.93a 2.68ab 2.51a 1.25b 6768.9b 6375.5b 1.21a 85.84ab
2m*2.5m 2.16b 34.25b 127.43b 2.42b 2.16b 1.24a 5437.6c 5043.6c 1.23a 73.93b
LSD(0.05) 0.26 3.96 19.26 0.35 0.35 0.12 994.21 957.97 0.13 14.35
Harvesting cycle
HC1 2.06a 33.88c 111.15c 1.76b 1.23b 1.50a 4863.9b 3390b 0.97b 47.53b
HC2 2.00a 41.21a 136.12b 1.88b 1.30b 1.51a 5403.7b 3574.8b 1.05b 54.89b
HC3 2.86b 37.70b 188.19a 4.12a 4.72a 0.90b 10598.3a 12255.5a 1.47a 152.55a
LSD(0.05) 0.10 3.43 16.68 0.30 0.30 0.10 861.01 829.62 0.12 12.43
Where: PH (Plant height), PBN (Primary branch number/plant), SBN (secondary branch number/plant). FLWP (Fresh leaf weight/plant), Fresh stem weight/plant, FLSR (Fresh leaf to
stem ratio) FLYHH (Fresh leaf yield/hectare per harvest), FSYHH (Fresh leaf yield/hectare per harvest), EOC (essential oil content), and EOYHH (essential oil yield/hectare/harvest)
[445]
Table 3. Mean value of the interaction effect of plant spacing and harvesting age on tested traits to determine optimum population density and harvesting age of citron scented gum
Level of HA Level of SP PH PBN SBN FLWP FSWP FLSR FLYH FSYH EOC EOYHA
8MAT 1.5*1.5m 2.11 34.3 138.39 2.24 2.14 1.36 7309.2 6063.86 0.87 65.72
8MAT 1.5*2m 2.22 42.02 148.05 2.84 2.7 1.41 7888.39 7462.47 0.97 82.69
8MAT 2*2m 2.05 37.91 134.39 2.31 2.18 1.3 5794.88 5434.24 1.2 70.22
8MAT 2*2.5m 1.88 30.86 118.34 2.22 1.91 1.4 5189.27 4870.17 1.18 66.62
9MAT 1.5*1.5m 2.42 42.89 160.24 2.56 2.6 1.29 8472.35 7999.8 0.93 80.06
9MAT 1.5*2m 2.35 36.41 143.32 2.55 2.29 1.33 7090.54 6128.44 1.41 107.07
9MAT 2*2m 2.47 36.98 157.78 2.85 2.72 1.29 7502.25 7136.16 1.23 99.08
9MAT 2*2.5m 2.13 34.04 119.85 2.12 1.93 1.53 4726.87 4445.27 1.39 70.72
10MAT 1.5*1.5m 2.62 40.38 162.81 2.61 2.59 1.08 8649.66 8177.12 1.33 113.73
10MAT 1.5*2m 2.45 38.29 155.42 2.97 2.69 1.26 7434.02 6792.33 1.13 91.31
10MAT 2*2m 2.52 39.22 159.16 2.88 2.63 1.17 7009.47 6555.96 1.19 88.23
10MAT 2*2.5m 2.48 37.86 144.09 2.92 2.62 1.27 6396.66 5815.4 1.12 84.46
Where: HA (Harvesting age), SP (spacing), PH (Plant height), PBN (Primary branch number/plant), SBN (Secondary branch number/plant). FLWP (Fresh leaf weight/plant), Fresh stem
weight/plant, FLSR (Fresh leaf to stem ratio) FLYHH (Fresh leaf yield/hectare per harvest), FSYHH (Fresh leaf yield/hectare per harvest), EOC (Essential oil content), and EOYHH
(Essential oil yield/ harvest)
[446]
E.citriodora is grown for its multi-purpose advantage and its growth depends on
our purpose and it requires 5 – 8 years to reach the first harvest for its wood
advantages but it can be harvested less than one year for its essential oil. Since the
leaf is the main economical part of E. citriodora for essential oil extraction, can be
harvested at any time of age but optimum essential oil yield obtained at 9 to 10
month after transplanting. The highest essential oil yield/hectare/harvest was
obtained when the plant harvested at 10 months after transplanting (94.43 kg
ha/harvest). Maximum essential oil content was obtained at the age of 9 months
after transplanting with a spacing of 1.5*2 m between plant and row respectively.
In this experiment, 1.5*2m spacing can be recommended as optimal spacing for
essential oil content and essential oil yield/hectare. Our experiment had the
limitation of quality parameters in which the essential oil of Citron scented gum
can be affected by population density and harvesting age. Therefore we
recommend that this experiment should be repeated with testing the effect of
spacing and harvesting age on the essential quality by determination of essential
oil composition using Gas Chromatography Spectrometer (GC-MS).
Acknowledgement
Agricultural Research Center are acknowledged for the financial and technical
support of this study. We are also a thanks and appreciation to Mr. AberaWaritu,
Mr. BiraraTilahun, Mr. DamtewTsegaye and Mr. Bekri Melka for their technical
support.
References
Amare Gatahun. 2010. Eucalyptus farming in Ethiopia: The case of Eucalyptus farm and village
woodlots in Amhara Region. In L. Gil, W. Tadesse, E. Tolosana, & R. López (Eds.)
Eucalyptus species management, history, status and trends in Ethiopia (pp. 206-221).
Proceedings from the congress held in Addis Ababa. September 15-17, 2010. Addis Ababa,
Ethiopia.
Batish DR, Singh HP, Setia N, Kohli RK, Kaur S, Yadav SS. 2007. Alternative control of little
seed canary grass using Eucalyptus oil. Agron. Sustain. Dev., 27: 171-177
Beemnet Mengesha Kassahun, Jaint. A. Teixeira da Silva, Solomon Abate Mekonen . Mekonen.
2011. Agronomic characters, leaf and essential oil yield of peppermint (Mentha Piperiata L.)
as influenced by harvesting age and row spacing. Medicinal and aromatic plant science and
biotechnology, 5(1) 49-53.
Chen J, Craven LA. 2007. Myrtaceae: Flora of China, Science Press (Beijing) and Missouri
Botanical Garden Press, 13: 321.
Javaid Akhtar, Saqib ZA, Qureshi RH, Haq MA, Iqbal MS, and Marcar NE. 2008. The effect of
spacing on the growth of Eucalyptus camaldulensis on salt-affected soils of the Punjab,
Pakistan
[447]
Luqman S, Dwivedi GR, Darokar MP, Kalra A, Khanuja SPS. 2008. Antimicrobial activity of
Eucalyptus citriodoraessential oil. Int. J. Essential Oil Therap., 2(2): 69-75.
Nematollah Etemadi, Sedigheh Rezai, Ali Nikbakht, Mostafa Yousefi, and Mohamad Mahdi
Majidi. 2018. Effect of Light Intensity on Leaf Morphology, Photosynthetic Capacity, and
Chlorophyll Content in Sage (Salvia officinalis L.) Horticultural Science and Technology
36(1):46-57.
Orwa C, A Mutua, R Kindt, R Jamnadass and A Simons. 2009. "Agroforestree Database: a tree
reference and selection guide version 4.0." World Agro forestry Centre, Kenya.
Saleh M. 1973. Effects of light upon quantity and quality of Matricaria chamomilla L. oil. Planta
Med 24:337-340.
SAS inst. 2002. SAS (Statistical Analysis System). 1996. SAS/STAT. Guide Version 6.12. SAS,
Institute Inc. Raleigh, North Carolina: USA
Sefidkon F, Bahmanzadegan A, Assareh MH, Abravesh Z. 2009. Seasonal variation in volatile oil
Of Eucalyptus species in iran. J. Herbs, Spices Med. Plants, 15: 106-120.
Singh H. 2002. Antifungal activity of the volatile oil of Eucalyptus citriodora. phytother., 73(3):
261-262.
Singh, Nidhi Setiaa, Daizy R. Batisha, Harminder P.Singhb, Nidhi Setiaa, Shalinder Kaura, and
Ravinder K. Kohli. 2006. Chemical Composition and Phytotoxicity of Volatile Essential Oil
from Intact and Fallen Leaves of Eucalyptus citriodora.
Solomon A, Beemnet M. 2011. Row spacing and harvesting age affect agronomic characteristics
and essential oil yield of Japanese mint (Mentha arvensis L.). Medicinal and Aromatic Plant
Science and Biotechnology 5 (1), 74-76
Teketay, Demel. 2000. Facts and experience on eucalyptus in Ethiopia and elsewhere: ground for
making wise and informed decision. Workshop on Eucalyptus Dilemma held on 15
November 2000 at Ghion Hotel, Addis Ababa.
Ramamurthy V, Munnu Singh, Akashta Srinivas, Silpasree AC, Ramesh Babu, Naidu LG, Kand
Prakasa Rao EV. 2016. Effect of age of plantation and season on leaf yield, content and
composition of Eucalyptus citriodora Hook oil and soil properties in semi-arid conditions of
Karnataka. Res. on Crops, 17(1): 112-117. Verbel JO, Nerioa LS, Stashenkob EE. 2009.
Bioactivity against Triboliumcastaneum Herbst (Coleoptera tenebrionidae) of Cymbopogon
citratusand Eucalyptus citriodoraessential oils grown in Colombia. Pest Manage. Sci., 66:
664–668.
Zewdinesh Damtew Zigene, Beemnet Mengessha Kassahun, Tsion Tesema Ketaw. 2012. Effects
of Harvesting Age and Spacing on Leaf Yield and Essential Oil Yield of Rosemary
(Rosmarinus officinalis L.). The African Journal of Plant Science and Biotechnology (6):9-
12.
[448]
Enhanced Income Generation through Sweet Basil
(Ocimum basilicum L.) and Maize (Zea mays L.)
Intercropping at Wondo Genet, South Ethiopia
Nebret Tadesse1*, Midhekisa Chala2 and Basazinew Degu1
1
Wondo Genet Agricultural Research Center, EIAR, P.O. Box: 198, Shashemene, Ethiopia.
2
Ambo Agricultural Research Center, EIAR, P.O. Box: 37, Ambo, Ethiopia.
*
Corresponding author email: nibrettadesse@gmail.com,
Abstract
Field experiments were conducted at Wondo Genet research center in 2017 and 2018
cropping seasons. The aim of the study was to determine the optimum population
density of sweet basil intercrop with maize for better income generation. The
experiment was laid out in randomized complete design using six intercropping ratios
= T1= sole basil, T2= maize sole, T3=100% maize + 100% Sweet Basil, T4=100%
Maize + 75% Sweet Basil, T5=100% Maize+50% Sweet Basil, T6= 100% Maize
+25% Sweet Basil. The results of the study revealed, combined mean grain yield of
maize was not significantly affected by the treatment of intercropping ratio of basil,
whereas the mean herbage yield of basil was significantly affected by the
intercropping ratio of basil proportion in maize. The combined mean yield advantage
in terms of total LER indices was greatest in the cases of 100% sweet basil into 100%
maize intercropping (1.66), this might be attributed to more efficient total resource
exploitation and greater overall production. Intercropping basil population in 100%
and 75% into 100% maize gave higher mean grain yield of maize, basil herbage and
economically profitable as compared to sole maize and sole basil cropping. We
recommend intercropping 75%, up to 100% sweet basil population in maize is would
be profitable for small scale farmers who practice mono-cropping around wondo
genet and similar agro-ecology.
Keywords: Economic Advantage, Intercropping ratio, Land Equivalent Ratio,

Maize, and Sweet Basil.
Introduction
Intercropping as a method of sustainable agriculture is the growing of two or more
crops during the same season in the same area to utilize common limiting
resources better than the species grown separately, and hence it is as an efficient
resource use method (Ghosh et al., 2006). Now a day this method has become one
of the popular methods in the agricultural system due to the more efficient use of
resources and its role in the reduction in weeds interference and other pests (Chen
et al., 2012; Lithourgidis et al., 2011). It should be kept in mind that intercropping
has a long history of food production in the world.
[449]
According to Seran and Brintha (2010), the intercropping system gave a higher
cash return to smallholder farmers than growing as the mono-crop. Intercropping
maize with cowpea has been reported to increase light interception in the
intercrops, reduce water evaporation, and improve conservation of the soil
moisture compared with maize grown alone (Ghanbari et al, 2010).
Medicinal plants are considered as a source of health products, essential oils and
other natural aroma chemicals in the national and international markets (Sujatha et
al., 2011). In addition, many studies have emphasized the use of medicinal plants
in intercropping systems, such as cassava/okra, Okra with chili pepper, saffron,
and chamomile, were also reported by Muoneke and Mbah, (2007), Uddin and
Odebiyi, (2011). Besides, intercropping of medicinal plants with other cereal/pulse
crops is vital to enhance yield and quality of crops. For example, intercropping
peppermint (Menthapiperita) plants with soybean produced a significantly higher
amount of essential oil when compared to monoculture plants and the oil yield
increased by 50% (Maffei and Mucciarelli, 2003).
In many studies, one of the most common reasons for the adoption of
intercropping is yield beneficial which is justified by the greater resource
exploitation by intercrops than monoculture (Poggio, 2005). The more efficient
exploitation of resources in intercropping happens because the component crops
use the resources either at different times or obtain resources from different parts
of soil or aerial environment (Echarte et al., 2011). Regarding yield advantage, in
a field study, Esmaeili et al. (2011) reported that the total seed yield of barley
(Hordeum vulgare) and annual medic (Medicago cutellate) improved in some of
the intercropping ratios when compared to the monoculture of either crop. Another
advantage of intercropping systems is weed suppression (Baniket al., 2006). The
reduction of weed growth by crop interference is a viable alternative to reduce the
reliance on herbicide application in weed management (Poggio, 2005). Even
though maize is a major food crop produced in south Ethiopia, there is information
on intercropping with aromatic plants such as sweet basil. In Ethiopia, sweet basil
is locally known as “Bessobila” in Amharic and different Ethiopian ethnic group
has a different name for sweet basil, (Helen, 1980; Gebrehiwot, et al., 2015).
Common basil or sweet basil (O. basilicum L.) is most widely used due to its high
economic value, popularity, and demands among these economically important
species of basil (Simon et al., 1999). Ethiopian farmers conventionally cultivate
and use this crop for house consumption, provide for the local market and some
investors started the exportation of sweet basil to different countries (Yimer,
2010). As explained by Lulie (2014), intercropping spearmint with maize at a
42cm inter-row spacing of spearmint followed by 36cm inter-row spacing is
biologically efficient, economically feasible, more profitable and provide
opportunities for the producers than planting alone. Therefore, the main
objectives of this study were to assess the influence of intercropping sweet basil
[450]
with maize on yield and yield component of maize and basil yield and to diversify
maize-sweet basil based farming practices by the inclusion of sweet basil for
additional cash generation.

Description of the study area: The study was conducted at Wondo Genet
Agricultural Research Center’s fields, in Southern Ethiopia during the 2017 and
2018 growing seasons. Wondo Genet is located 270 km south of Addis Ababa and
14km southeast of Shashemene. The geographical coordination of the area is 7019’
N latitude and 380 38’ E longitudes with an altitude of 1780 meters above sea
level (m.a.s.l). The site receives a mean annual rainfall of 1128 mm with
minimum and maximum temperature of 110C and 260C, respectively. The soil is a
sandy clay loam with an average pH of 7.2.
Plant materials, experimental design and field management: The

experiment was laid out in a randomized complete block design with three repeats.
The treatments consisted with six additive series intercropping combinations viz:
T1= sole Basil (60cm inter-row with 40cm intra-row spacing ), T2=Sole maize
(80cm x 30cm), T3= 100% maize + 100% Sweet Basil (60cm inter-row x 40cm
intra-row), T4=100% Maize+75% Sweet Basil (Jumping two rows of maize + one
row of sweet Basil with 50cm intra-row), T5=100% Maize+50% Sweet Basil
(jumping two row of maize + one row of Sweet Basil with 40cm intra-row), T6=
100% Maize +25% Sweet Basil ( jumping two row of maize + one row of sweet
Basil with 60cm intra row spaced. Maize variety BH546 was used and spaced at a
distance 30cm x 80cm with 41663ha-1plant population and the companion crop
was Sweet Basil. Maize seeds were sown at the rate of two seeds per hill and
simultaneously uniform sweet basil seedlings transplanted at the height of around
15 cm or two months after sowing according to the treatments. Maize was thinned
to one plant per hill after three weeks. Gross plot size of the experiment was 16m 2
(4m x 4m) and the distance between the plots and blocks were 1 and 1.5 m apart,
respectively. Each plot consisted of five rows of maize. The net central unit areas
(2.4m x 3.4m=8.16 m2) of each plot excluding the borders were used for data
collection. Recommended agronomic practices were followed, and fertilizers were
applied at rates of 50 kg urea and 46 kg P ha-1.
Data collection and measurements: Data for maize on plant height (cm), cobs
per plant, cob length (cm), and data for sweet basil on plant height (cm) were
determined from 10 randomly sampled plants per plot during physiological
maturity. The maize grain yield, leaf and inflorescence weight per plant (gm), dry
herbal yield (kg ha-1), and Oil yield (kg ha-1) of each crop of the component crop
in the intercrop was determined from the harvestable plots and converted to kg ha-
[451]
1
. Sweet basil leaf and inflorescence oil extraction were done using the hydro-
distillation method in a Clevenger type apparatus.
Data analysis: The collected data were subjected to analysis of variance

(ANOVA) using SAS version 9.3 (SAS, 2012) statistical software packages.
Treatment means were separated using the Least Significant Difference (LSD)
procedure at the 5% and 1% level of significance.
Land equivalent ratio (LER) was determined using the formula defined by Willey
(1991)
LER = La + Lb = Yab/ Yaa + Y ba/Ybb
Where Lm and Lb are the LERs for the individual crops of the system
Ymb = Intercrop yield of crop ‘Maize’
Ybm = Intercrop yield of crop ‘Sweet Basil’
Ymm = Pure stand crop yield of ‘Maize’
Ybb = Pure stand crop yield of ‘sweet basil’
The economic benefits of maize-basil intercropping were determined by

calculating the monetary advantage index (MAI) using the formula established by
Willey (1979). Prevailing local market prices of maize and basil were taken for
economic analysis and the prices of maize and basil were 5.5 and 30
ETB/Ethiopian Birr per kg) respectively, during the experiment.
MAI = (Value of combined intercrops) x (LER-1)/LER

Where MAI = Monetary advantage index
LER = Land equivalent ratio

Maize components: Among all measured parameters of maize for two years only
in the first year; the above-ground biomass had a significant difference. The
highest above-ground biomass (18706 kg ha-1) was recorded from sole maize
followed by 25% of sweet basil intercropping in maize (13483 kg ha-1) and the
lowest (10502 kg ha-1) was obtained from 100% of sweet bail with 100% maize, it
indicated that intercropping highest present of sweet basil in maize affect the
biomass of maize. However, the second year and the combined mean were not
significantly affected above-ground biomass of maize (Table 1). On the contrary,
according to Ghanizadeh et al. (2014) reported, the total yield of sole maize was
more than their intercrops in maize: sweet basil: borage intercropping. Further,
Girma (2015) reported, the highest biological yield obtained from monoculture
than intercropped. The higher yield of monoculture compared to intercrop may be
due to minimal disruption of the plant’s habitat (Banik et al., 2006).
[452]
Table 1: Influence of Maize- Sweet Basil intercropping on yield and yield components of maize for two consecutive years
Treatments Plant height(m) Number of ear per plant

Years 2017 2018 Combined mean 2017 2018 Combined mean
Sole Maize 2.56 2.37 2.57 1.3 1.3 1.4
100% Maize+100% Sweet Basil 2.71 2.33 2.36 1.7 1.5 1.3
100% Maize+ 75% Sweet Basil 2.76 2.23 2.58 1.3 1.2 1.4
LSD (0.05) ns ns Ns ns ns ns
CV% 7.6 4.7 5.1 18.3 24.2 23
Table 2: Influence of Maize- Sweet Basil intercropping on yield and yield components of maize for two consecutive years
Treatments Aboveground biomass (kg ha-1) Harvest index

2017 2018 combined 2017 2018 Combined mean
Sole Maize 18706a 10045 12969.9 0.34 0.54 0.45
100%Maize+100%Basil 10502b 9868 11750.0 0.63 0.54 0.47
100%Maize+ 75% Basil 10577b 9742 12311.4 0.65 0.52 0.46
100%Maize+50% Basil 12967b 9693 11940.4 0.45 0.47 0.46
100%Maize+25% Basil 13483ab 8044 11520.2 0.45 0.53 0.50
LSD 0.05 5245.8* ns ns ns ns ns
CV% 21.0 13.1 7.9 27.1 14.9 8.9
The other parameters; - plant height (cm), number of ear per plant, ear length
(cm), grain yield, and harvest index were not showed significant differences
among the intercropped ratio of basil in 100% of maize or in another word those
parameters don’t influence by the intercropping population density of sweet basil;
this could be due to the reason that intercropping sweet basil in maize at an early
stage or immediately after maize emerged doesn’t affect maize growth parameters
and grain yield; on the other hand, before the canopy of maize formed and
competition occurred sweet basil harvesting time reached within two and half
month (Table 1, 2, and 3).
Table 3: Influence of Maize- Basil intercropping on yield and yield components of maize for two consecutive years
Treatments Ear length (cm) Grain yield (kg ha-1)

Years 2017 2018 Combined 2017 2018 Combined mean
mean
Sole Maize 17.7 20.1 19.52 6262 5276.9 5769.4
100%Maize+100%Basil 17.5 19.9 18.4 6079.6 4262.2 5751.8
100%Maize+ 75% Basil 18.7 19.3 18.9 6441.2 5163.4 5609.3
100%Maize+50% Basil 18.6 19.6 19.2 5811.2 4777.4 5487.3
100%Maize+25% Basil 18.6 19.6 19.2 6424.3 5424.1 5343.3
LSD (0.05) ns ns Ns ns ns ns
CV% 2.9 3.18 2.25 12.6 19.24 14.0
Basil component
In each year and the combined mean, plant height was significantly (P<0.05)
affected by the intercropping ratio of sweet basil. Sweet bail planted at the ratio of
100% basil with 100% maize gave higher mean plant height (50.5cm), followed
by decreasing consecutively with a decreased ratio of basil (Table 4). Such an
[453]
increase in plant height with increased plant density may be explained by
increased activity of stem growth hormone for plant sunlight competition. This
finding was consistent with the data reported by Girma (2015) who observed that
taller basil plants were achieved from the high intercropping population density of
basil-maize intercropping. Significantly higher combined mean fresh leaf yield
per plant (40.1 kg p-1) was obtained from sole basil. Even though statistically were
not significant differences observed between the intercropping ratio of basil in
maize from 25% up to 100%; here are increments of value when the percent of
basil decreased in maize (Table 4). This might be due to the shade effect of maize
and competition. Dry herbage yield and oil yield were significantly influenced by
the intercropping ratio of basil in maize. The combined mean higher dry herbage
yield (1560 kg ha-1) and oil yield (7.95 kg ha-1) were obtained from the sole basil
followed by the intercropping ratio of 100% basil in 100% maize for both and the
lowest combined mean dry herbage (485 kg ha-1) and oil yield (3.3 kg ha-1) were
obtained from the intercropping ratio of 25% basil in maize treatment.
Table 4: Influence of Maize- Basil intercropping on yield and yield components of Basil for two consecutive years
Plant height (cm) Leaf weight per plant (gm)

Treatments 2017 2018 Combined Mean 2017 2018 Combined mean
Sole basil 43.2c 41.3ab 42.2cd 44.2a 36.1 40.1a
100%Maize+100%Basil 53.9a 47.06a 50.5a 14.1b 27.3 20.7b
100%Maize+75%Basil 51.9a 40.5b 46.2b 14.8b 25.4 20.7b
100%Maize+50%Basil 55a 35.6b 45.3bc 18.7b 27.8 23.24b
100%Maize+25%Basil 47.5b 36.2b 41.8d 17.2b 40.1 28.6b
LSD 0.05 4.2* 6.1* 3.4* 9.2* ns 11.4*
CV% 3.5 5.9 3.03 22.3 29 22.6
This higher dry herbage yield and oil yield insole and in 100% basil in 100%
maize treatment attributed to competition free and the accommodation of the high
number of basil per unit area contribute to producing high biomass of herbage
yield (Table 5). This finding was consistent with the result of Girma (2015) who
reported higher dry herbage yield at maize: sweet basil (1:1) intercropping.
Table 5: Influence of Maize- Basil intercropping on yield and yield components of Basil for two consecutive years
Dry herbage weight (kg ha-1) Oil yield(kg ha-1)

Treatments 2017 2018 Combined Mean 2017 2018 Combined mean
Sole basil 1926.6a 1501a 1560a 7.8aa 8.05 7.95a
100%Maize+100%Basil 544.9b 1331ab 938.1b 6.1a 4.8 4.62ab
100%Maize+75%Basil 532.3b 793.5cd 662.9bc 3.5b 6..0 4.5ab
100%Maize+50%Basil 423.4b 956.8bc 644.6bc 2.5bb 66.6 4.2ab
100%Maize+25%Basil 332.4b 386d 485c 6.1a 2.3 3.3b
LSD 0.05 407.04* 494.1* 348.2*** 2.6* Ns 4.2*
CV% 28.7 26.4 21.5 26.3 30.1 28.2
[454]
Total Land Equivalent Ratio (LER)
Intercropping ratio of sweet bail to maize crop had a significant (P<0.05) effect on
the total land equivalent ratio in both years and the combined mean (Table 6). In
2017 cropping year maximum LERt (1.64) was obtained from 75% of basil
intercropping in maize, which was statistically similar with total land equivalent
ratio (1.45) obtained from 100% basil in 100% maize and 50% basil intercropping
in 100% maize treatments. The lowest LERt (1.07) was recorded from 25% basil
intercropping in 100% maize treatment. In 2018 cropping year higher LERt (1.86)
was obtained from 100% basil intercropping in 100% maize treatments followed
by LERt(1.54) which was obtained from 75% of basil intercropping in 100%
maize which were statistically similar with the LERt obtained from 50% basil
intercropping in 100% maize treatment. This total land equivalent ratio difference
between the years could be due to the favorable environment condition found in
cropping year.
All combined mean LERt intercropping ratio of proportion was above the unity
(1.00) indicating the advantages of intercropping over sole cropping. The
combined mean yield advantage in terms of total LER indices was greatest in the
cases of 100% basil into 100% maize intercropping (1.66) which might be
attributed to more efficient total resource exploitation and greater overall
production. This indicated that an additional 0.66 ha (66%) more area would have
been needed to get equal yield to planting maize and basil in pure stands. The
intercropping ratio of basil in maize in additive series from 25%, 50%, 75% and
100% in maize cropping were showed yield increment advantage in 6%, 44%,
59% and 66% consecutively over the sole cropping. A similar finding was
reported by Girma (2015) on Maize- basil intercropping. Also, Higher LERt were
reported on different crops by Abate and Alemayehu, 2018; Takele and
Mohammed, 2014; Hirpa, 2014 on intercropping fenugreek, field pea, haricot bean
and sorghum in maize compared to monocropping of maize in Ethiopia.
Monetary advantage Index (MAI)

MAI also used to evaluate the economic advantages of the intercropping system.
The monetary advantage index was significantly (p<0.05) influenced by the
intercropping ratio of basil in maize in both years and the combined mean. In the
2017 cropping year intercropping ratio of 100% basil into 100%, maize gave
significantly higher economic return (22659 birr ha-1); which were statistically
similar with MAI obtained from 75% basil intercropping in 100% maize and 50%
basil intercropping. The lowest economic return (2332 birr ha-1) was obtained
from the intercropping ratio of 25% basil in 100% maize this could be due to the
low number of basil population that might result in less biomass production in
terms of basil yield that affect the economic return.
[455]
Table 6: Land equivalent ratio and monetary advantage of Intercropping Maize and Basil Crops as affected by additive %
proportion for two consecutive years
Treatments The total land equivalent ratio Monetary advantage index (birr ha-1)
2016/17 2017/18 Combined 2016/17 2017/18 Combined mean
mean
100%Maize+100%Basil 1.45a 1.86a 1.66a 22659a 31631a 27145a
100%Maize+75%Basil 1.64a 1.54ab 1.59a 16194a 22961ab 19577.5b
100%Maize+50%Basil 1.44a 1.43b 1.44b 15623a 15191b 15407b
100%Maize+25%Basil 1.07b 1.06c 1.06c 2332b 1484c 1908c
LSD( 0.05) 0.28* 0.22* 0.17* 10535* 9174.6* 6439.3*
CV% 10.24 7.4 5.7 28.9 32.3 31.2
In the second year maximum economic return (31631 birr ha-1) was obtained from
the 100% basil intercropping ratio in 100% maize and the lowest economic return
(1484 birr ha-1) from 25% basil intercropping in maize. Similarly, the combined
mean economic return was recorded from 100% basil intercropping in 100%
maize and the lowest economic return obtained from 25% basil intercropping in
100% maize. These results suggest that the intercropping ratio could improve the
smallholder farmers’ incomes and total efficiency of the system with low input
investment by varied plant population and time. A similar finding was reported by
Girma (2015) on Maize- basil intercropping.
Conclusion
Based on the results of this study; the mean grain yield of maize was not
significantly affected by the treatment of the intercropping ratio of basil, whereas
the mean herbage yield of basil was significantly affected by the intercropping
ratio of basil proportion in maize. Intercropping basil population in 100% and
75% into 100% maize gave higher mean grain yield of maize, basil herbage and
economically profitable as compared to sole maize and sole basil cropping.
Therefore we recommend intercropping 75%, up to 100% sweet basil population
in maize is would be profitable for small scale farmers who practice mono-
cropping around Wondo genet and similar agro-ecology.
Acknowledgment
We would like to acknowledge the Crop Research Directorate and Wondo Genet
Agricultural Research Center for providing all the necessary facilities and support
during the execution of this study. Our sincere appreciation goes to Gizachewu
Atinafu, Cherenet Tefera, BiraraTilahun, Zerihun Jonba, Bersome’eso, and
Alemitu Teka for providing all the necessary facilities and support during the field
and laboratory works.
[456]
References
Abate M and Alemayehu G. 2018. Biological Benefits of Intercropping Maize (Zea mays L) with
Fenugreek, Field Pea and Haricot Bean Under Irrigation in Fogera Plain, South Gonder Zone,
Ethiopia. Agriculture, Forestry, and Fisheries. Vol. 7, No. 1,pp. 19-35,(2018).doi:
10.11648/j.aff.20180701.14
Banik P, Midya A, Sarkar BK, Ghose SS. 2006. Wheat and chickpea intercropping systems in an
additive series experiment: Advantages and weed smothering. European Journal of
Agronomy 24: 325-332,(2006).
Chen Y, Luan C, Shi X. 2012. Xanthium suppression under the maize-sunflower intercropping
system. Journal of Integrative Agriculture 11(6): 1026-1037,(2012).
Echarte L, Della Maggiora A, Cerrudo D, Gonzalez VH, Abbate P. 2011. Yield response to plant
density of maize and sunflower intercropped with soybean. Field Crop Research 121: 423-
429,( 2011).
Esmaeili A, Sadeghpour A, Hosseini SMB, Jahanzad E, Chaichi MR, Hashemi M. 2011.
Evaluation of seed yield and competition indices for intercropped barley (Hordeum vulgare)
and annual medic (Medicago scutellata). International Journal of Plant Production 5(4): 395-
404, (2011).
Gebrehiwot H, Bachetti R and Dekebo A. 2015. Chemical composition and antimicrobial activities
of leaves of sweet basil (Ocimumbasilicum L.) herb. Int J Basic Clin Pharmacol. 4(5):869-
875, (2015).
Girma A. 2015. Yield Advantage and Economic Benefit of Maize Basil Intercropping under
Different Spatial Arrangements and Nitrogen Rates. Scholarly Journal of Agricultural
Science Vol. 5(8), pp. 296-302(2015).
Ghosh PK, Mohanty M, Bandyopadhyay KK, Painuli DK, Misra AK. 2006. Growth, competition,
yield advantage, and economics in soybean/pigeon pea intercropping system in semi-arid
tropics of India, I. Effect of subsoiling. Field Crop Research 96: 80-89, (2006).
Ghanbari A, Dahmardeh M, Siahsar BA and Ramroudi M. 2010. Effect of maize (Zea mays L.)
cowpea (Vigna unguiculata L.) intercropping on light distribution, soil temperature and soil
moisture in an environment," J. Food Agr Environ, vol. 8, pp.102-108(2010).
Ghanizadeha H, Bicharanlou B, Zaefarian F, Bagheri M. 2014. Study of intercropping of maize
with sweet basil and borage. Cercetări Agronomiceîn Moldova Vol. XLVII , No. 2 (158)
(2014). DOI: 10.2478/cerce-2014-0013.
Helen D. 1980. The cultivated basils. Buckeye Printing, Independence, Mo(1980).
Hirpa T. 2014. Effect of intercrop row arrangement on maize and haricot bean productivity and the
residual soil. World Journal of Agricultural Sciences, 2 (4): 069- 077,( 2014).
Lithourgidis AS, Vlachostergios DN, Dordas CA, Damalas CA. 2011. Dry matter yield, nitrogen
content, and competition in pea-cereal intercropping systems. European Journal of
Agronomy, 34: 287-294,(2011).
Lulie B, Getu H and Bogale T. 2014. Intercropping of Maize (Zea mays L.) with Spearmint
(Mentha spicata L.) as Supplementary Income Generation at Wondo Genet Agricultural
Research Center, South Ethiopia. International Journal of Recent Research in Life Sciences.
Vol. 1 pp: 31-43.(2014).
Maffei M, Mucciarelli M. 2003. Essential oil yield in peppermint/soybean strip intercropping.
Field Crop Research 84: 229-240, (2003).
Muoneke CO and Mbah EU. 2007. Productivity of cassava/okra intercropping systems as
influenced by okra planting density. African Journal of Agricultural Research 2(5): 223-
231(2007).
Poggio SL. 2005. Structure of weed communities occurring in monoculture and intercropping of
field pea and barley. Agriculture, Ecosystems and Environment 109: 48-58, (2005).
SAS Rev. 2012. Version 9.3 Institute inc., Cary, NC, USA (2012).
[457]
Seran TH and Brintha I. 2010. Review on maize-based intercropping. Journal of Agronomy, vol. 9,
no. 3, pp. 135-145, (2010).
Simon, JE, Morales MR, Phippen WB, Vieira RF, and Hao Z. Basil. 1999 A source of aroma
compounds and a popular culinary and ornamental herb. In: J. Janick (ed.), Perspectives on
new crops and new uses. ASHS Press, Alexandria, VA p. 499–505(1999).
Sujatha S, Bhat R, Kannan C, Balasimha D. 2011. Impact of intercropping of medicinal and
aromatic plants with organic farming approach on resource use efficiency in areca nut (Areca
catechu L.) plantation in India. Industrial Crops and Products 33: 78-83(2011).
Takele A and Mohammed J. 2016. Time of Haricot Bean Intercropping into the Maize-Based
Cropping Systems under Conservation Tillage in the Rift Valley of Ethiopia. Ethiop. J.
Agric. Sci. 26(2) 11-26 (2016).
Uddin RO, Odebiyi JA. 2011. Influence of intercropping on incidence, abundance and severity of
pest damage on okra (Abelmoschus esculentus (Linn.) Moench) (Malvaceae) and chili pepper
(Capsicum frutescence Linn.) Journal of Agricultural Science 3(3): 63-66, (2011).
Yimer M. 2010. Home Gardens of Ethiopia (2010).
Willey RW. 2991. Evaluation and presentation of intercropping advantages. Exp. gric.21:119-
(1991).
Willey RW. 1979 Intercropping-its importance yield advantages. Agron. J. 71(2): 115-119, (1979).
[458]
Potato
Bridging Attainable Yield Potentials of Potato

through Nutrient Management in
Northwestern Ethiopia
Tadele Amare1*, Zerfu Bazie1, Erkihun Alemu1, Amlaku Alemayehu1, AbereTenagne1,
Bitewlgn Kerebih1, Yasin Taye1, Assefa Derebe1 and Tesfaye Feyissa2
1
Adet Agricultural Research Center, 2Amhara Regional Agricultural Research Institute (ARARI)
*
Corresponding author: tadele17b@yahoo.com
Abstract
Ethiopian highlands have a vast potential to increase the production and productivity
of potato. The northwestern Ethiopia (Gojam and Gondar) is the major potato
producers. The yield of potato is far below its attainable potential (less than 10 tone
ha-1). Soil fertility is one of the major yield-limiting factors of potato in Ethiopia. An
intensive on-farm research was conducted for three consecutive years to find
biological and economical optimum amounts of nutrients for the major potato
producing areas of northwestern Ethiopia (Yilmna Densa, Farta and Laygaint
districts). A factorial experiment with four levels of nitrogen (46, 92, 138, 194 kg ha-
1
), three levels of P2O5 (46, 69, 92 kg ha-1) and a pilot treatment with no NP nutrient
inputs were used. A randomized complete block design (RCBD) with three
replications was employed. The findings of the research showed that more than 40
tonnes ha-1of attainable yield potential of potato for the study areas through nutrient
management. The yield was increased significantly with an increased rate of nitrogen
at all sites while the response to phosphorus was not strong. The yield difference was
insignificant between the rates of phosphorus for most of the sites and the years. The
financial analysis of the finding indicated that applying 138/46 (N/P2O5kg ha-1)
brought a marginal rate of return (Birr/Birr) 70.9 while 138 N /69P2O5 resulted in
10.7 for Yilmana Deans. For Laygaint and Farta districts, 138 N/46 P 2O5 gave a
marginal rate of return 24.3. Therefore, based on the farm gate price of potato and
the cost of fertilizer 138 N / 69 P2O5 kg ha-1 is recommended for Yilmana Densa and
similar districts while Farta and Laygaint 138 N / 46 P 2O5 kg ha-1 is recommended.
Keywords: Potato, Soil fertility, Farta, Laigaint, Yilmana Densa
Introduction
Ethiopia has a vast potential to increase the production and productivity of potato
(Solanum tuberosum L.) especially in the highlands (Gebremehdin et al, 2012;
Haverkot et al., 2012). About 70% of the cultivated land of Ethiopia is suitable to
potato production (FAO, 2008) but only 2% of the potential has been used (Adane
et al., 2015). About 40% of potato producers of the country are in South Gonder,
North Gonder, East Gojam, West Gojam and Agew Awi zones of the Amhara
region (Adane et al., 2015) where Adet Agricultural Research center is mandated.
Potato is the fourth crop globally in terms of production and area coverage
[459]
(FAOSTAT, 2007). It also ranks first among root and tuber crops in Ethiopia
(CSA, 2016).
Potato is a cheap and nutritive food security crop, because of its high production
of yield with good nutritive values per unit area and time than other major cereal
crops. Potato was one of the major strategic crops to the United Nation’s
Millennium Development Goals of achieving food security and eradicating
poverty. Moreover, 2008 was recognized as the year of potato by the United
Nations. Its contribution to food security with stable prices might be continued as
the price of potato mainly depends on local demand and supply than global trade.
It is short cycle and early maturing have additional advantages of double cropping
and crop intensification than other crops that take longer days for maturity. The
productivity of potato in Ethiopia is below 10 tonnes per hectare (Adane et al.,
2015; Asresie et al., 2015; Haverkort et al, 2012). Baye and Gebremehidin (2012)
indicated that released potato varieties have achievable yield potentials up to 54
tonnes/ha in Ethiopia under farm conditions. Haverkot et al., (2012) reported the
highest yield report of potato (64 tonne/ha around Shashemene). Soil fertility is
one of the factors that limit agricultural productivity in Ethiopia including potato
(Adane et al., 2012; Degefu and Mengistu, 2017). Therefore, the research was
conducted to increase the productivity of potato through nutrient management.
Materials and Method

Study sites: Yilmana Densa is located at about 42 km from Bahir Dar on the way
to Addis Abeba through Mota. The district is characterized by highland and mid-
altitude. It has a unimodal type of rainfall with an average annual of 1240 mm,
while the minimum and maximum monthly temperatures are 9.3OC and 25.7OC,
respectively. June, July and August receive the largest shares of the annual
rainfall. The local geology is characterized by volcanic basalt flows and Cenozoic
pyroclastic fall deposits (Anteneh, 2009). The Vertisols and Nitisols are
dominated the mid-altitude of Yilmana Densa. The farming system is a mixed
crop-livestock type. Tef is the most dominant cereal crop in the Vertisols while
potato, tef, barley and wheat are important crops for the highlands. The district is
one of the agriculturally potential places of the region. Animal fattening is
becoming a feasible business model of the district that is transforming the farming
community to a better life. The district is one of the highly populated districts of
the region that led to very small farmland per capita. While the districts of Farta
and Lai Gaint are parts of the south Gondar Admiration zone of the Amhara
National Regional State. The farming system of the districts is also a mixed crop-
livestock type. The dominant crops in these districts are wheat, potato, barley, faba
bean and triticale.
[460]
Treatment setup: A factorial experiment with four levels of nitrogen (46, 92,
138, 194 kg/ha) and three levels of phosphorus (46, 69, 92 P2O5 kg/ha) were used
for the study. One pilot treatment with no nutrient input (control) was used to
evaluate the yield of potato without fertilizer application as there are
circumstances farmers use no fertilizer; making the total number of treatments 13.
Treatments were replicated three times and arranged in a randomized complete
block design (RCBD). Nitrogen was applied by splitting it three times (one third at
planting, one third at earthen up and final one third at flowering) while the whole
dose of phosphorus applied at planting. Gudene was the variety used for the study.
Disease-free tubers were raised from tissue culture of ARARI and grown at Adet
Agricultural Research Center in the screen house and used for the study. The
distance between plants and rows was 0.3m and 0.75m respectively. The size .of
each experimental plot was 13.5m2 (3m*4.5m) and data were collected from the
central 4 rows. The ridges were parallel to the slopes of the farms. Earthening up
was uniformly done at the ages of 2-3 weeks after germination. Redomil at the
rate of 3 kg ha-1 was applied uniformly against late blight. Data on the biological
yield of potato as well as the price of potato and fertilizer were collected.
Soil sampling and analysis: Composite soil samples were collected at the depth
of 0-20 cm before planting and from each treatment after harvesting (after two
months of harvesting). Collected soil samples were air-dried and ground by mortar
and pestle. Soil pH was determined in a 1:2.5 soil to water suspension following
the procedure outlined by Sahilemedihn and Taye (2000). Soil organic carbon
content was determined by the wet digestion method using the Walkley and Black
procedure (Nelson and Sommers, 1982). Total nitrogen was determined using the
Kjeldahl method (Bremner and Mulvaney, 1982) while the available phosphorus
was determined following the Olsen procedure (Olsen and Sommer, 1982). The
exchangeable potassium was measured by flame photometer after extraction of the
samples with 1N ammonium acetate at pH-7 following the procedures described
by Sahilemedihn and Taye (2000).
Data analysis: The effect of independent variables (Nitrogen, phosphorus and

their interaction) on the dependent variable (yield of potato) was statistically
tested. The analysis was made for each site and year and combined as well.
Analysis of variance (ANOVA) was carried out to assess the difference between
treatments. Upon the existence of a significant difference for ANOVA, (p < 0.05),
further analysis of mean separation was carried out using Duncan’s Multiple
Range Test (DMRT). Graphical analyses were also employed to evaluate response
curves over different doses of nitrogen and phosphorus nutrients. The partial
budget analyses were done based on (CIMMYT, 1988). The cost of NPS and urea
was 1284.05 and 1158.58 Birr per 100 kg; respectively. The farm gate price of
potato (Birr 100 kg ha-1) was 487 at Yilmana Densa and 450 Birr at Laigaint and
Farta.
[461]
Soils of the study sites: The soil parameters indicative to soil fertility were
analyzed for the composite soil samples collected before planting as summarized
below. The pH of the soil (pH of water) was above five that ranged from 5 to 5.3
for the case of Yilmana Densa and south Gonder it was ranged from 5.1 to 5.4
both sites are moderately acidic (FAO, 1984). Soil acidity is not a critical yield-
limiting factor for the study sites especially in the upper parts of the landscapes as
andosol is dominating in that parts of the districts with pH higher than our report
here. Moreover, potato is tolerant to acid soils compared to other sensitive crops
and hence pH was not a limiting factor for our research. The soil organic carbon
(SOC) contents of the study sites ranged from 0.96 to 1.75% for Yimana Densa
and 1.14 to 2.01% for south Gondar. Accordingly, the soils of Yilmana Densa
were below the critical limits and south Gondar ranges from below to critical
levels(Loveland and Webb, 2003; Murphy, 2014). Generally, soils of the study
area were low in SOC that limits nutrients supplies including N and P (Murphy,
2014). Therefore, for optimum production of potato nutrients must be supplied in
the forms of synthetic fertilizer, organic fertilizer, or in the form of synthetic
integrated with organic fertilizers. The exchangeable potassium of the soil was
ranged from 0.69 to 0.88 cent mole/kg of the soil for Yilmana Densa while for
South Gondar it ranged from 0.46 to 0.62 cent mol/kg of the soil. The result of soil
analysis on soil potassium was above the critical limits for all the study sites (IPI,
2016), indicating that potassium is not a priority fertilizer for the production of
annual crops including potato with its current state as supported with the finding
of other annual crops (Tadele et al., 2018). The available phosphorus of Yilmana
Densa was in a range of 9 to 10 ppm while for south Gondar it ranged between 10
and 20 ppm. Accordingly, better response to phosphorus application was observed
from Yilmana Densa than south Gondar districts.
The response of potato to the application of nutrients (NP) was higher as shown in
Figure 3. The response to nitrogen was higher than phosphorus (Figure 1, Table 1
and 2). Figure 1 showed that there was an increase in yield of potato for the
application of nitrogen even at the rates of 194 kg Nha-1for all sites and for all
seasons (Figure 3: A and B). The yield gap between the rates of nitrogen was
higher and uniformly increased for nitrogen than phosphorus. However, the yield
response to the applied phosphorus was not as strong as the response to nitrogen
(Figure 3: C and D). The higher yield gap for the application of phosphorus was
observed between the control (no fertilizer) and the treatments with phosphorus
(Figure 3: C and D) compared to the response to nitrogen (Figure 3: A and B). For
the applied phosphorus the yield was increased with a low slope coefficient. The
Analysis of variance (ANOVA) results of the research showed a significant yield
differences (p <5%) for all sites and years to nitrogen (Table 1 and 2). The
maximum tuber yield (40 t ha-1) of potato was found in Yilmana Densa at
[462]
Chinkulit Kebele in the 2018/2019 cropping season using 194 kg of Nha-
1
compared to 11 t ha-1 of tuber yield from control from the same location in the
same year (data is not shown here). The average maximum yield of potato for the
three years in Yilmana Densa for nitrogen was 29.2 t ha-1by applying194 kg of
Nha-1 compared to 8.8 t ha-1 tuber yields from treatments without nutrient (Table
1). While the maximum potato tuber yield (297.5 t ha-1) in south Gondar was
recorded at Tsegur Kidanimiret, in Farta district by applying194 kg Nha-
1
compared to 11.6 t ha-1of potato tuber yield without nutrient input from the same
site in the same season (2018/19 cropping season (data not shown). The three-year
average tuber yield of potato in south Gondar was ranged from 7.54 t ha-1 to 22.08
t ha-1without nutrient input and applying 194 kg N.ha-1, respectively (Table 2).
40 30
A C
30
20
Yield (t/ha)
Yield (tl/ha)
20
10
10
0 0
0 46 92 138 184 230 0 23 46 69 92 115
25
30
20
25 D
Yield (t/ha)
20 15
B
Yield (t/ha)
15 10
10 5
5 0
0 23 46 69 92 115
0 P2 O 5 (KG/ha)
0 46 92 138 184 230
Nitrogen (kg/ha)
Figure 1. Response of potato to N (A Yilmana Densa and B South Gondar) & P (C Yilmana Densa and D south
Gondar).Where: is year one, is year two is year three and is the mean
[463]
Table 1. Potato response to applied NP at Yilmana Densa (tuber yield tha-1) over the years
Nitrogen (kg/ha) Year 1 Year 2 Year 3 Mean

46 15.48 15.52 17.18 16.06
92 23.16 21.64 22.46 22.21
138 24.81 26.37 29.61 26.93
194 26.61 27.01 34.11 29.16
LSD 2.97 2.3 3.18 1.7
Phosphorus (P2O5 Kgha-1)
46 22.17 21.63 25.61 22.98
69 21.94 22.97 25.51 23.41
92 23.43 23.31 26.41 24.38
LSD NS NS NS NS
CV 24.4 18.8 22.8 23
N*P NS NS NS NS
Control(0N and 0P) - 7.53 10.13 8.83
Table 2. Potato response to applied NP at South Gondar (tuber yield t ha-1) over the years
NitrogenKg ha-1 Year 1 Year 2 Year 3 Mean

46 13.23 12.33 12.74 12.76
92 16.91 14.99 17.50 16.61
138 18.81 22.82 21.00 20.89
194 19.28 24.23 22.51 22.08
LSD 4.30 2.55 3.11 1.97
P2O5 Kgha-1
46 17.50 17.59 18.17 17.82
69 16.50 18.49 18.71 18.01
92 17.17 19.70 18.43 18.43
LSD NS NS NS NS
CV 36.3 20.9 33 31
N*P NS NS NS NS
Control (0N and 0P) - 8.06 7.01 7.54
The response of tuber yield to phosphorus was much less than to the responses for
nitrogen for all the sites; resulting in a non-significant yield difference between
phosphorus rates (Table 1 and 2). The smallest rates of phosphorus (46 kg P2O5ha-
1
) resulted in comparable tuber yield of potato as that of the maximum rates (69
and 92 kg P2O5ha-1) for all the sites and more specifically to south Gondar. Our
finding on nitrogen was in line with the findings of Yassin (2017) who
recommended 110 kg Nha-1 for the major potato producing areas of Ethiopia
although his recommendation was much lower than our finding. However, his
recommendation of phosphorus for the production of potato for the major
producing areas of the country (90 kg P2O5ha-1) disagreed with our findings. He
recommenced a higher rate that was two times our findings. Melkamu and Muluye
(2018) under Koga irrigation scheme reached conclusions of different rates of
phosphorus recommendations for different sites and varieties. They recommended
102 kg P2O5ha-1for variety Belete under low phosphorus conditions and for
Gudene 69 kg P2O5ha-1. For areas with sufficient phosphorus, they recommended
69 kg P2O5 ha-1 for Belete and 52P2O5ha-1 for Gudene. However, our main
[464]
difference with their finding was the growing season that is our experiment was in
the main season while their experiment was entirely dependent on irrigation.
Besides, the authors (Melkamu and Muluye, 2018) reached into conclusions with
only two sites experiment for a single irrigation season while our data was
sufficient to capture the three rainy seasons. They claimed more than 30 ppm of
available phosphorus at Koga that might be a technical error from our experiences
at Koga. On the other hand, the recommendations by Asresie et al., (2015)was
somewhat in line with our finding with phosphorus they recommended 69 kg
P2O5ha-1 while their recommendation to nitrogen was much lower than our finding
as they recommended 80.80 kg Nha-1for the potato growing areas of northwestern
Amhara Region ( South Gondar and Gojam areas). The recommendation of
nitrogen made by Zelalem et al., (2009) to the Vertisols of Debrebirhan area (138
kg Nha-1) was in line with our findings to all sites of our study. They
recommended 20 kg P2O5ha-1; lower than by half to our findings.
The finding showed that the yield of potato at Yilmana Densa could be improved
by more than threefold (greater than 300%) and at south Gondar by more than two
and a half folds (greater than 250%) through nutrient applications (N) and its
implication on food and nutritional security is tremendous as the study represents
40% of the potato producers of the country (Adane et al., 2015). There is a huge
growing demand of potato for the local consumption and one of the best strategies
to boost production and productivity of potato in the farming systems of the
country especially for areas that represent this research is through nutrient
management as reported by Adane et al., (2012) as well as Degefu and Mengistu,
2017.
Partial budget analysis: Partial budget analysis was employed to identify the economical
optimum rates of NP fertilizer for the production of potato. Farmgate prices for potato
were 4.87 and 4.50 Birr kg-1 of potato for Yilmana Dens and south Gondar; respectively.
The cost of fertilizer was 1284.05 and 1158.58 for NPS and Urea; respectively. The
significant difference in the biological yield was reflected in the economical responses of
the partial budget analysis. Some of the treatments with N/P2O5 kg ha-1 (46/69, 46/92,
92/69, 194/46 at Yilmana Densa and 46/69, 46/92, 92/92, 138/69, 138/92 and 194/46 at
south Gondar) were dominated and discarded from the analysis.
Accordingly, the highest marginal rate of return (70.9 birr/ton) for Yilmana Densa was
found at the rates of 138 kg Nha-1 and 46 Kg P2O5ha-1 (Table 3). For south Gondar the
maximum marginal rate of return (24.3 Birr/Birr) was found with treatments 138 kg Nha -1
and 46 Kg P2O5ha-1(Table 4). The optimum rate of NP fertilizers for potato production
with this research disagrees with the one recommended by Yasin (2017). With the value
of the yield collected with research, the economical optimum rate could be continuously
updated with the situation of farm gate price of potato and the cost of fertilizer. Based on
the findings of this research, for Yilmana Densa and similar areas 138 kg Nha-1 and 46 Kg
P2O5ha-1as first option and 138 kg Nha-1 and 69 Kg P2O5ha-1are economically feasible
rates while for south Gondar 138 kg Nha-1 and 46 Kg P2O5ha-1a profitable rate.
[465]
Table 3. Partial budget analysis for Yilmana Densa
NPS Urea Cost of Cost of Urea TC Yield TR MC MR MRR

Treatment (Kgha-1) (kgha-1) NPS (Birr) (Birr) (Birr ) (ton) (Birr ) (Birr (Birr ) (Birr/ton)
0/0 0 0 0 0 0 8.83 43002.1 - - -
46/46 121.0 50 1553.7 579.3 2133.0 16.84 82010.8 2133.0 36875.7 17.3
92/46 121.0 150 1553.7 1737.9 3291.6 22.34 108795.8 1158.6 25626.4 22.1
92/92 242.0 100 3107.4 1158.6 4266.0 23.15 112740.5 974.4 2970.3 3.0
138/46 121.0 250 1553.7 2896.5 4450.2 25.87 125986.9 184.2 13062.2 70.9
138/69 181.6 225 2331.8 2606.8 4938.6 27.04 131684.8 488.5 5209.4 10.7
138/92 242.0 200 3107.4 2317.2 5424.6 27.87 135726.9 485.9 3556.2 7.3
194/69 181.6 346.7 2331.8 4016.8 6348.6 29.66 144444.2 924.1 7793.2 8.4
194/92 242.0 321.7 3107.4 3727.2 6834.6 30.59 148973.3 485.9 4043.2 8.3
Table 4. Partial budget analysis for South Gondar
Treatment NPS Urea Cost of Cost of Urea TC Yield TR MC MR MRR

(Kg) (Kg) NPS (Birr) (Birr) (Birr ) (ton) (Birr ) (Birr (Birr ) (Birr/Ton)
0/0 0 0 0 0 0 7.54 33930.0 - - -
46/46 121 50 1553.7 579.3 2133.0 13.289 59800.5 2133.0 25870.5 12.1
92/46 121 150 1553.7 1737.9 3291.6 16.294 73323.0 1158.6 13522.5 11.7
92/69 181.6 125 2331.8 1448.2 3780.1 17.388 78246.0 488.5 4923.0 10.1
138/46 121 250 1553.7 2896.5 4450.2 210.0 94500.0 670.1 16254.0 24.3
194/69 181.6 346.7 2331.8 4016.8 6348.6 221 99450.0 1898.4 4950.0 2.6
194/92 242.0 321.7 3107.4 3727.2 6834.6 234 105300.0 486.0 5850.0 12.0

To improve the productivity of potato through NP nutrient management-intensive
on-farm research was conducted for three consecutive rainy seasons in the
northwestern Ethiopia. The finding of the research proved that the yield was
improved by two and half fold for the case of south Gondar and by threefold for
the case of Yimna Densa. The findings of the research have tremendous
implications on bridging the nutritional and food securities of the nation in general
and the northwestern parts of the country in particular as 40% of potato growers
are concentrated in this part of the country and potato is better off in nutritional
contents. The result revealed that soil fertility management plays major role in the
productivity of potato. The research indicated that the response to nitrogen was
stronger than the response to phosphorus. The tuber yield of potato was not
reduced to the extent of 194 kg Nha-1, it rather increased with increasing N rates,
indicating we should focus on nitrogen fertilizers to maximally increase the
productivity and profitability of potato production. With current increasing trends
of potato demands throughout the county, the application of maximum rates of
nutrients could lead to a more productive, profitable and sustainable potato
business. For Yilmana Densa and similar niches, 138 kg Nha-1 and 46 kg P2O5ha-1
[466]
is recommended as the first option followed by 138 kg Nha-1 and 69 kg P2O5ha-1
as the second option. For south Gondar, 138 kg Nha-1 and 46 kg P2O5ha-1 is
recommended. For south Gondar, the productivity was lower than Yilmana Densa
that needs further work on other management aspects. Furthermore, the
recommendations could be subjected to change upon fluctuations on the farm
price of potato and the cost of fertilizers.
Acknowledgment
We appreciate farmers for allowing our experiments to be carried out on their
farms. The research financially supported by the Agricultural Growth Program
(AGPII) and potato commodity. We extend our deepest gratitude to the potato
team of Adet Agricultural Research Center for supplying disease-free potato
tubers. We also thank Abrham Aweke and Zimie Abaw for their support in
analyzing the soil samples at Adet Agricultural Research Center.
References
Adane Hirpa, Meuwissen MPM, Van der Lans IA, Lommen WJM, Oude Lansink AGJM, Admasu
Tsegaye and Struik PC. 2012. Farmers’ opinion on seed potato management attributes in Ethiopia:
A conjoint analysis. Agronomy Journal, 104:1413–1424.
Yassin Esmael. 2017. A review status of potato production and marketing in Ethiopia. Advanced
Journal of Agricultural Economics, Extension and Rural Development ISSN: 6723-2418 Vol. 2
(1), pp.9-14
Anteneh Zewde. 2009. Geological Controls in the Formations and Expansion of Gullies over Hillslope
Hydrological Processes in the Highlands of Ethiopia, Northern Blue Nile Region. (Master’s thesis)
Cornell University, NY, USA.
Asresie Hassen, Alemu Worku, Molla Tafere, Mekonen Tolla, Abel Ahmed, Seferew Dagnew,
Yihenew G.Selassie, Desallegn Molla and Tesfaye Abebe. 2015. Best Fit Practice Manual for
Potato Production and Utilization, BDU-CASCAPE working paper 8.
Bremner, JM, and Mulvaney CS. 1982. Nitrogen total. In A. L. Page, R. H. Miller, D. R. Keeney
(Eds.), Methods of soil Analysis, vol. 2, American Society of Agronomy, Madison, Wis. (1982),
95-624.
Degefu Kebede and Mengistu Ketema. 2017. Why do Smallholder Farmers Apply Inorganic
Fertilizers below the Recommended Rates? Empirical Evidence from Potato Production in Eastern
Ethiopia. Adv Crop Sci Tech 5 (2): 265.
FAO.1984. FAO Fertilizer and plant nutrition bulletin 9. Food and Agricultural Organization of the
United Nations, Rome, Italy.
Gebremedhin Woldegiorgis, Schulz S, Baye Berihun. 2013. Seed potato production and
dissemination, experiences, challenges and prospects. Proceedings of the National Workshop on
Seed Potato Tuber Production and Dissemination, 12-14 March 2012, Bahir Dar, Ethiopia.
Haverkot A, van Koesveld F, Schepers H, Wijnands J, Wustman R and Zhang X. 2012. Potato
prospects for Ethiopia: On the road to value addition. Applied plant research Publication number
528. Wageningen University and Research Center, The Netherlands.
IPI (International Potash Institute). 2016. Technical manual on potash fertilizer: Use for soil fertility
experts and development agents. DOI 10.3235/978-3-905887-22-8
Loveland P and Webb J. 2003. Is there a critical level of organic matter in the agricultural soils of
temperate regions: A review. Soil and Tillage Research 70, 1 -18.
[467]
Melkamu Alemayehu and Muluye Jember. 2018. Optimum rates of NPS fertilizer application for
economically profitable production of potato varieties at Koga Irrigation Scheme, Northwestern
Ethiopia. Cogent food and Agriculture
Murphy BW. 2014. Soil organic matter and soil function- review of the literature and underlying data.
Department of the Environment, Canberra, Australia.
Nelson DW and Sommers LE. 1982. Total carbon, organic carbon and organic matter. In A. L. Page
(Ed.), Methods of Soil Analysis. Part 2: Chemical and Microbiological Properties (pp. 539-579),
Agron. 9. Madison, Wisconsin.
Olsen SR and Sommers LE. 1982. Phosphorus. In A. L. Page (Ed.), Method of soil analysis, pp.403-
430, Part 2: Chemical and microbiological properties. Agronomy 9. Madison, Wisconsin.
Tadele Amare, Zerfu Bazie, Erkihun Alemu, Asmare Wubet, Birhanu Agumas, Mengistu Muche,
Tesfaye Feyisa and Desalew Fentie. 2018. Crops Response to Balanced Nutrient Application in
Northwestern Ethiopia. Blue Nile Journal of Agricultural Research: 1 (1), pp 1-14.
Zelalem Ayichew, Tekalign Tsegaw, and Nigussie Dessacha. 2009. Response of potato (Solanum
tuberosum L.) to different rates of nitrogen and phosphorus fertilization on Vertisols at Debre
Berhan, in the central highlands of Ethiopia. African Journal of Plant Science, 3(2), 016-024.
[468]
Evaluation Performance of Potato Clones in the
Mid-highland Areas of Ethiopia
Shamil Alo1, Engida Tsegaye2, Abraham Negara3 and Alemu Worku4*
1
Tepi Agricultural Research Center, 2Hawasa Agricultural Research Center, 3Backo Agricultural Research
Center and 4Adet Agricultural Research Center; *Corresponding Author: alemuworku32@yahoo.com
Abstract
The aim of this work was to evaluate the performance of advanced potato clones for
tuber yield and yield components. Seven advanced clones and one standard control
cultivar broadly grown in Ethiopia were compared in Bako, Hawasa and Tepi research
stations. The design of the experiment was a randomized complete block with three
replications of four rows of 10 plants each. Combined analysis over locations and
seasons revealed that the clone CIP-398220.510 gave the highest total tuber yield of
298 .68 qt/ha followed by CIP-398180.292 (272.41 qt/ha). Whereas the lowest tuber
yield (133.41 qt/ha) was from CIP-398193.650. Moreover, the lowest percentage of
late blight infestation was from CIP-398220.510 and CIP-398180 .292. As a result
clones, CIP-398220.510 and CIP-398180.292 were selected as a candidate for variety
verification trial after seed multiplication for release to maximize potato yield in the
mid highland areas of the country. Despite the satisfactory tuber yields obtained by
these genotypes under humid lowlands in comparison with other genotypes, we believe
that their potential was not achieved and higher yields could be obtained with better
crop management. Therefore, there is a need to carry out complementary studies on
agronomic techniques aiming at increasing the yields obtained so far.
Keywords: Potato Genotypes, Mid Highland areas, Tuber yield
Introduction
As in other developing countries, potato crop (Solanum tuberosum L.) is
increasingly becoming a cash crop in Ethiopia. However, the national meher
season production estimated at 48483.12 t year-1 (CSA, 2010) is still low
compared to the quantity required for feeding more than a hundred millions of
people. Unfortunately, this gap is expected to grow bigger and bigger due to the
increase in population and urbanization. Therefore, introducing potato crops in
humid lowlands agrarian systems by determining the performance of different
potato genotypes on main agro-physiological characteristics under field conditions
is important. That is why we evaluated a couple of CIP (International Potato
Center) genotypes with the expected potential for humid lowlands adaptability.
Our objective, in the long run, is to be able to recommend to farmers and other
stakeholders varieties to grow in the humid lowlands of the south and southwest
regions of the country.
[469]
In Ethiopia, one of the major factors contributing to this low productivity is lack
of improved high yielding, late blight resistance potato cultivars with wide
adaptability and stability in tuber yield. The potato research program in Ethiopia
is evaluating potato genotypes advanced by the International Potato Center (CIP)
for late blight resistance, tuber yield and other agronomic traits. The breeding
program at CIP advanced the potato clones with field resistance for late blight and
other improved agronomic traits (Population B) (Landeo et al., 1995). However,
the yield performance and adaptation of the clones to tropical Africa, particularly
to the Ethiopian humid mid highlands are not known. Thus, the objective of this
study was to evaluate the performance of potato genotypes for tuber yield and
yield-related traits in the mid-high lands of Ethiopia and to develop high yielding
varieties with good tuber qualities.

Description of the study area: The field experiment was conducted under
rainfed conditions during the year 2017 and 2018 main cropping season at Backo,
Hawasa and Teppi Research Centres, in South and Sowth western Ethiopia. These
areas are low lands and known for their high rainfall annually. Teppi is located at
an elevation of 1200 meters above sea level. The research station receives an
annual rainfall of 1559 mm with maximum and minimum temperatures of 29.70C
and 15.50C respectively. The soil of the experimental site is radish brown sandy
clay loam classified as Nitosol with a pH range of 5.6 to 6.
Experimental treatment and design: A total of 7 potato genotypes were

introduced from the International Potato Center (CIP) and one standard cultivar
was used for this experiment. The experiment was laid out in a Randomized
Complete Block Design (RCBD) where each genotype was replicated three times
at each location. Each plot was 3m x 3m = 9 m 2 wide consisting of four rows,
which accommodated 10 plants per row and thus 40 plants per plot. The spacing
between plots and adjacent replication was 1 m and 1.5 m, respectively.
Experimental data collection and analysis: Several parameters were recorded

from the planting date to harvest time. The major parameters collected are listed as
follows: plant height, number of stems per hill, number of tuber per plant, average
tuber weight and tuber yield and analyzed using SAS.
Results and Discussion:

The result of the analysis of variance for tuber yield and yield components of
seven advance potato genotypes studied in the above locations is presented below.
The analysis of variance for each location showed significant differences among
the genotypes for tuber yield and yield components (P<_0.05). This shows the
[470]
existence of large variability among the genotypes for the parameters considered.
The individual and combined analysis of variance for tuber yield and yield-related
parameters revealed highly significant differences (P<_0.05) among genotypes in
all test locations. At Bako except stem number per plant significant variation has
been recorded at all the tested parameters. The largest tuber was recorded from the
standard check ‘belete’ with an average tuber weight of 172.93gm tuber while the
lowest was from CIP-398098.119 (97.27gm tuber). The highest tuber yield
(386.5qt/ha) was recorded from CIP-398180.292 followed by CIP-398190.605
which is 385.71 qt/ha (Table 1).
Table 1. Mean tuber yield and yield components of genotypes combined over two years at Bako
Plant Stem Tuber Average Marketable Total Tuber

height Number/pl Number/ Tuber Tuber Yield(Qt/ha)
No Genotypes (cm) ant Plant weight(gm) Yield(Qt/ha)
1 CIP-398190.605 65.00bc 5 7.4a 116.80b 380.20a 385.71a
2 CIP-398098.119 84.00ab 5 8.8a 97.27b 138.14c 142.86c
3 CIP-398190.200 83.00ab 6 7.0a 112.07b 328.87ab 334.52ab
4 CIP-398193.650 67.00bc 4 5.80ab 89.67bc 98.28d 100.00d
5 CIP-398180.292 60.00c 7 8.40a 107.47bc 384.50a 386.50a
6 CIP-398220.510 67.4bc 5 5.60ab 121.70b 258.95b 260.32b
7 CIP-399208.620 54.00c 4 2.80d 103.03bc 132.45c 134.11c
8 BELETE 94.00a 4 6.20ab 172.93a 320.52ab 322.22ab
Mean 68.73 4.38 6.5 115.11 255.24 258.28
CV 9.7 20 24.8 28.1 25.01 24.4
The analysis of variance showed a highly significant difference due to genotypes

and environments for all characters at Hawasa in both seasons. Genotypes also
performed significantly with tuber yield-related characters. At Hawsa the highest
tuber yield was recorded on CIP-398220.510(399.3 qt /ha) followed by CIP-
399208.620(397.5 qt/ha) whereas on average more number of tubers (14.25) was
recorded on CIP-398220.510(Table 2).
Table 2. Mean tuber yield and yield components of potato genotypes combined over two years at Hawasa
Plant Stem Tuber Average Marketable Total Tuber

height Number/ Number/ Tuber Tuber Yield(Qt/ha)
No Genotypes (cm) plant Plant weight(gm) Yield(Qt/ha)
1 CIP398190.605 64.67c 4.33b 12.75 76.7ab 192.6abc 235.8ab
2 CIP398098.119 76.00c 4.67ab 8.13 141.5c 202.2abcd 218.6ab
3 CIP398190.200 74.63e 5.00b 10.92 142.3c 236.9c 271.1abc
4 CIP398193.650 52.00a 4.00a 10.58 79.3ab 174.1a 209.3a
5 CIP398180.292 65.33c 4.00a 11.12 108.3abc 231.5c 281.1bc
6 CIP398220.510 57.67b 6.33c 14.25 119.0bc 374.2e 399.3d
7 CIP399208.620 60.33b 4.67ab 10.92 133.9c 369.1e 397.5d
8 BELETE 53.00a 4.00a 9.96 104.6 162.0a 204.7a
Mean 63.89 4.59 11.48 107.7 244.7 28.1
CV 2.6 11.9 22.9 24.7 13.8 13.6
At Teppi the highest tuber yield was recorded on CIP-398220.510(236.42 qt /ha)

followed by CIP-398098.119 (169.13 qt/ha) whereas on average more number of
[471]
tubers (12.2) was recorded on CIP-398098.119 (Table 3). As compared to other
study sites, the performance of genotypes at Teppi is lower due to high late blight
pressure influence during both seasons.
Table 3. Mean tuber yield and yield components of potato genotypes combined over two years at Teppi
No Genotypes Plant Stem Tuber Marketable Total Tuber

height (cm) Number/plant Number/Plant Tuber Yield(Qt/ha)
Yield(Qt/ha)
1 CIP-398190.605 41.4ab 4.6a 8.4bc 143.15c 145.31c
2 CIP-398098.119 53.2a 3.0b 12.2a 168.56bc 169.13bc
3 CIP-398190.200 35.6bc 2.4bc 11a 160.24bc 161.73bc
4 CIP-398193.650 35.2bc 2.6bc 6.6c 89.12d 91.53cd
5 CIP-398180.292 35.8bc 2.6bc 6.8c 147.89bc 149.63bc
6 CIP-398220.510 55.4a 2.8bc 11.8a 235.68a 236.42a
7 CIP-399208.620 55.6a 2.6bc 7.4bc 136.18c 137.03c
8 BELETE 54.8a 2.4bc 9.4bc 101.42d 102.47cd
Mean 45.87 2.87 9.2 148.78 149.16
CV 16.34 14.67 18.5 15.02 15.68
The over location and years combined analyses of variance (Table 4) revealed that
there was significant (P<0.05) among genotypes for all parameters considered in
the study. Among the genotypes, CIP398220.510 and CIP398180.292 showed
better performance in total tuber yield which gave 298.68 qt/ha and 272.41 qt/ha
respectively. The more number of tubers per plant (10.55) was recorded from
genotype CIP-398220.510 and the largest average tuber weight was recorded on
the standard variety Belete which is138.77gm per tuber basis (Table 4).
Table 4. Mean tuber yield and yield components of potato genotypes combined over location and years at Bako, Hawasa
and Teppi
No Genotypes Plant Stem Tuber Average Marketable Total Tuber
height Number/ Number/ Tuber Tuber Yield(Qt/ha)
(cm) plant Plant weight(gm) Yield(Qt/ha)
1 CIP398190.605 57.02bc 4.64 9.52ab 96.75bc 238.65b 255.61b
2 CIP398098.119 71.07a 4.22 9.71ab 119.39ab 169.63bc 176.86bc
3 CIP398190.200 64.41ab 4.47 9.64ab 127.19ab 242.00b 255.78b
4 CIP398193.650 51.40bc 3.53 7.66b 84.49c 120.50c 133.61c
5 CIP398180.292 53.71bc 4.53 8.77b 107.89bc 254.63ab 272.41ab
6 CIP398220.510 60.16ab 4.71 10.55a 120.35ab 289.61a 298.68a
7 CIP399208.620 56.64bc 3.76 7.04b 118.47ab 212.58bc 222.88bc
8 BELETE 67.27a 3.47 8.52b 138.77a 194.65bc 209.80bc
Mean 59.50 3.95 9.06 111.41 216.24 145.18
CV 9.55 15.52 21.73 26.40 17.94 17.89
Conclusion
Tuber yield performance of genotypes across locations and seasons identified CIP-
398220.510 and CIP-398180.292 relatively stable genotypes in Bako, Hawasa and
Teppi areas. The two genotypes CIP-398220.510 and CIP-398180.292 were the
two superior genotypes that produced the highest tuber yield as well as the lowest
[472]
late blight percentage. Therefore, these clones are selected as a candidate for a
variety verification trial having a clean and enough amount of seed for release to
maximize potato yield in the mid highland areas of the country.
Acknowledgment
We would like to acknowledge the financial support provided by the Ethiopian
Institute of Agricultural Research. We highly appreciate and acknowledge the
effort made by respective research centers’ field assistants for their unreserved
support and substantial contribution to accomplish this study.
References
CSA. 2019. The Federal Democratic Republic of Ethiopia Agricultural Sample Survey on farm
Management Practices (Private Peasant Holdings, Meher Season) Volume III Addis Ababa.
FAO. 2010. Quality declared planting material: Protocols and standards for vegetatively propagated
crops. FAO plant protection and production paper number 195. Food and Agricultural
Organization of the United Nations, Rome, Italy. ISSN 0259-2517
Kooman PL. 1995. Yielding ability of potato crops as influenced by temperature and daylength. Ph.D.,
Wageningen Agricultural University.
Lung’aho C, Nderitu SKN, Kabira JN, El-Bedewy Ra, Olanya OM. and A Walingo. 2006. Yield
performance and release of four late blight tolerant potato varieties in Kenya. Journal of
Agronomy 5, 57 - 61.
Yan W and Kang MS. 2003. GGE biplot analysis: a graphical tool for breeders, geneticists, and
agronomists CRC Press, New York.
Yan W and Tinker NA. 2006. Biplot analysis of multi-environment trial data: principles and
applications. Can J of Plant Sci 86: 623-645.
[473]
[474]
Tuber Yield Performance and Stability of Potato
Genotypes in the Highlands of Ethiopia
Baye Berihun1, Animut Tarik2*, Alemu Worku2, Manamino Workayehu2, Melkamu
Eneyew2, Abebe Chindi3, Tesfaye Abebe3, Zerihun Kebede4,
Fekadu G/Tensay5, Wubet Awoke1 and Gebre Alamir1
Woldiya University1, Adet Agricultural Research Center2, Holeta Agricultural Research Center3,
Debre Birhan Research Center4, Kulumsa Research Center5
*
Corresponding e-mail: animuttarik2001@gmail.com
Abstract
Potato (Solanum tuberosum L.) is one of the most widely grown root and tuber crops
in the mid and high altitude areas of Ethiopia. It has both dietary and income-
generating roles to producers. However, the yield obtained at farm levels is low
compared with research sites. One of the main reasons for such low yield of potato in
the country is shortage of high yielding and disease resistant variety. To this effect,
potato national variety trial was conducted at Adet, Holetta , Kulumsa, Sinana and
Debre Birhan with the main objective of identifying high yielding and relatively stable
potato genotypes during 2016 and 2017 main cropping season. In this experiment, ten
genotypes were tested along with the standard and local checks in randomized
complete block design with three replications on a gross plot size of 9 m2 planted at a
spacing of 75cm × 30cm between rows and plants, respectively. Combined analysis of
variance over locations and years revealed that CIP-396034.268 gave the highest
tuber yield of 37.56 t/ha followed by CIP-392657.171(35.64 t/ha) and standard check
Belete (37.52 t/ha). Whereas the lowest tuber yield (15.64 t/ha) was from CIP-
396029.250. AMMI analysis identified that two of the interaction principal components
axis were significant at P<0.01% and explained 77.01% of the genotype by
environment interaction sum square. Moreover, the lowest percentage of late blight
infestation was recorded from CIP-392657.171. As a result clones CIP-396034.268
and CIP-392657.171 were selected as a candidate for a variety verification trial in the
year 2019 for release to maximize potato yield in the country.
Keywords: Potato, highland, tuber yield, GGE
Introduction
Crop genotypes grown in different environments encounter considerable
environmental fluctuations in yield performance as a result of genotype-by-
environment interaction (GE). GE is defined as a phenomenon that, phenotypes
respond to genotypes differently for different environmental factors. If there is no
GE, a single trial would be sufficient for genotype evaluation and a single cultivar
would prevail all over the world (Gauch and Zobel, 1996). However, GE is a
common phenomenon in field trials conducted under natural conditions. It affects
breeding progress; because it obscures the demonstration and selection of superior
genotypes across environments (Ebdon and Gauch, 2002).
[475]
Major interaction can be expected when there is a wide variation among genotypes
for morph-physiological traits and wide variation between environments due to
climatic, soil, biotic and management factors. Bocianowski et al. (2011) reported
that the main effects of cultivar and location were significant for yield and traits
directly influencing yield. Often 80% of yield variation is caused by the
environment, while genotype and GEI cause 10% of the variation each (Yan
2001). In potato, a large contribution of the environment that influenced the tuber
yield of potato has been reported by many authors (Baye, 2017; Gedif and
Yigzaw, 2014; Yigzaw et al., 2008). Of the environment components, location and
genotype by location interaction were the most important source of tuber yield
variation (Gedif and Yigzaw, 2014). On the other hand, Baye (2017) reported that
the, environment; particularly the production season was the major contributor for
the total tuber yield variation in potato, indicating potato tuber yield can be
affected by various spatial and temporal environmental factors.
As GE is expected in genotype evaluation, the selection of genotypes with

superior yield and stability requires a multi-environment trial (MET) (Roozeboom
et al., 2008). The stability of a genotype is its ability to realize a maximum degree
of genetic potential in a wide range of environments (Annicchiarico, 2002). The
most desired type of genotype is the one that combines high yield with stable
performance in various environments (Piepho, 1996).
In Ethiopia, potato is considered one of the most important food crops that
contributes not only to the mitigation of food security but also serving as an
important cash-generating crop. Despite its importance, the productivity is low
(12.3 t ha-1) as compared to the world average 19.4 t ha-1. One of the major factors
contributing to this low productivity is a shortage of improved high yielding, late
blight resistant potato cultivars with wide adaptability and stability of performance
in tuber yield. The potato research program in Ethiopia is evaluating potato
genotypes advanced by the International Potato Center (CIP) for late blight
resistance, tuber yield and other agronomic traits. The breeding program at CIP
advances potato clones that have field resistance for late blight and other improved
agronomic traits (Population B) (Landeo et al., 1995). However, the yield
performance and adaptation of the clones to tropical Africa, particularly to the
Ethiopian high lands are not known as the environment under which they have
been advanced are different to other countries as usually manifested by
inconsistent performance and release of different varieties in different countries
from among the same set of dispatched clones from CIP. Thus, the objective of
this study was to evaluate the performance and stability of potato genotypes for
tuber yield and yield-related traits in the high lands of Ethiopia; and to develop
high yielding varieties with good tuber qualities.
[476]
Location description: Potato field experiments were conducted in different
corners of Ethiopia, at Adet (North West Ethiopia), Holetta (Central Ethiopia),
Bekoji and Sinana (South East Ethiopia) and D/Birhan (North-East Ethiopia)
during 2016 and 2017 main rainy season. Based on the Ethiopian agro-ecological
zone classification, Adet is located at the mid high land agroecological belt, while
Holetta, Sinana, Bekoji and D/Birhan clustered under the high land agro-
ecological belt. The testing locations are considered as representative testing
locations by the national potato research program to develop high yielding and
stable potato varieties for the Ethiopian high land potato growing areas (Table 1).
Table 1. Environmental descriptions of the experimental locations
Altitude Ave. rain Temperature (oC) Soil type Soil pH

Locations (masl) fall (mm) Min Max
Adet 2240 1271 8.8 25.2 Nitosol 5.20
Holetta 2400 1041.4 6.70 21.70 Nitosol 5.60
Bekoji 2700 1350 10 22.5 Nitosol 6.12
Sinana 2400 1380 9 21 Pelvic vertisol 6.05
D/Birhan 2800 1260 3 20 Clay 5.83
Genetic materials: A total of 10 potato genotypes were used in this study. The
genotypes were CIP-392797.22, CIP-395017.242, CIP-392657.171, CIP-
396034.268, CIP-391058.175, CIP-395017.229, CIP-393220.54, CIP-392633.64,
Belete and local check. The first 8 genotypes were advanced from the
International Potato Center (CIP) for late blight resistance, tuber yield and other
agronomic traits. Belete is a CIP sourced potato variety. The variety was
developed and as, late blight resistant, widely adaptable, and high yielding variety
used as a standard check.
Experimental design: The experiment was lied out in Randomized Complete

Block Design (RCBD) with three replications. Each experimental plot consisted of
40 tubers planted in a 3m x 3m plot with an inter-row and intra-row spacing of
0.75m and 0.30m, respectively. Application of pest and disease management,
weeding and ridging and other cultivations was done following the
recommendation of each location and when required. Nitrogen and Phosphorus
fertilizers were applied based on the recommendations of each test location, and
applied in the form of urea and di-ammonium phosphate (DAP). The whole P
source was applied at planting while N application was split in two: a week after
emergence and at the start of flowering. Planting and harvesting were carried out
from the end of May to mid-June, and from September to October, respectively,
depending on the planting and harvesting time at each test location.
[477]
Trait measurements: The traits were measured similarly at each experimental
location. Days to emergence (DTE) recorded as the number of days from planting
to when 50% of the plants emerged; it was assessed daily until all the plots had
more than 50% emergence. Plant height (PH) was measured as the distance in cm
from the soil surface to the tip of the growing main stem when 50% of the
genotypes flowered. Stem number plant-1 (SNPP) was the number of stems of a
genotype counted before the plant canopy declined as those only emerged below
the soil surface. Harvesting was carried out once, when the last cultivars reached
physiological maturity, and 16 plants per plot were harvested and used to evaluate
tuber traits. Tuber number plant-1 (TNPP) was recorded as the number of tubers
collected from a mature plant at harvest. Average tuber weight (ATW) was
recorded as the ratio of the weight of tubers per plant to the total number of tubers
per plant at harvest. Tuber yield plant-1 (TYPP) (the average tuber yield plant-1)
was calculated taking the tuber yield (fresh weight) of all harvested plants from
the net plot of a cultivar divided by the number of plants harvested. The yield
estimated per plant base was adapted to yield ha-1.
Data analysis: Combined analyses of genotypes across locations were also

performed after carrying out the homogeneity of error variance following the
procedure of Gomez and Gomez (1984). Statistical analysis was carried out using
GenStat software version 18.1 (VSN international limited, 2015) for ANOVA and
GGE bi-plot analysis. The adjusted total tuber yield means of each trial was used
in a combined analysis of variance to evaluate the main effect of environment (E),
genotype (G), and genotype-by-environment interaction (GE) variances. Further
partitioning and analysis of GE interaction was carried out with GGE bi-plot
model using GGE software (Yan, 2001). GGE bi-plot analysis was used to
visualize the relationship between test locations and genotypes, and to determine
“which-won-where” lot. It also reveals the stability of genotypes; genotypes
located near the bi-plot origin are considered as widely adapted genotypes while
genotypes located far are considered as being specifically adapted. The GGE bi-
plot was constructed using the first two principal components (PC1 and PC2)
derived from environment-centered tuber yield data (Yan et al., 2000).
Environment-centered data is the data with the grand mean and the environmental
effects removed from the data as defined by the following GGE model:
Yij     j   1 i 1 i 1   2  i 2   ij
Where Yij is the measured mean of genotype
j2
i(=1,2,….,n) in environment j(=1,2…,m), μ is the grand mean, βj is the main effect

of environment j, λ1 and λ2 are the singular values (SV) for the first and second
principal component (PC1 and PC2), respectively, ξi1 and ξi2 are eigenvectors of
genotype i for PC1 and PC2, respectively, ŋ1j and ŋ2j are eigenvectors of
environment j for PC1 and PC2, respectively, εij is the residual associated with
genotype i in environment j.
[478]
Analysis of variance: Combined analysis of variance result of each location
over two years indicated significant (P<0.01) genotypic differences for plant
height, stem number per plant, number of tubers per square meters, average tuber
weight tuber yield (Tables 2). Combined analysis over the location in a year as
well as combined over locations and years discovered the presence of significant
(P<0.01) difference among tested genotypes in tuber yield (Table 3). Across
locations and season, the highest total tuber yield (37.56t/ha) was obtained from
genotype CIP-396034.268. However, the two top-performing genotypes (CIP-
396034.268 and CIP-392657.171) gave higher tuber yield across six environments
i.e., at Adet, Debre Birhan and Holetta where the verification trial will be done
(data not shown). In contrast, the lowest tuber yield (24.39 t/ha) was harvested
from CIP-395017.229 (Table 3). This is mainly attributed to a genetic difference
in yielding ability and tolerance to late blight of potato.
Table 2. Mean Performance of potato genotypes for tuber yield and yield-related traits across ten environments
Genotypes PH SNPP TNPP ATW MTY/ha TTY/ha

CIP-392797.22 62.3ef 3.86b 7.47f 86.22ab 24.8d 28.96c
CIP-395017.242 67.42cd 3.66bc 9.13cd 92.4a 26.52cd 32.33bc
CIP-392657.171 65.13cdef 3.90b 8.57def 89.6ab 30.67b 35.64ab
CIP-396034.268 76.02a 3.84b 9.60bcd 90.77ab 34.72a 37.56a
CIP-391058.175 65.8cde 3.99ab 9.68bcd 71.48c 24.55d 27.91c
CIP-395017.229 61.07fg 3.80b 8.91cde 69.9c 24.39d 27.95c
CIP-393220.54 63.61def 3.73b 10.04bc 72.08c 30.06bc 33.82b
CIP-392633.64 68.37bc 3.53bc 7.80ef 69.04c 25.95d 29c
Belete(st.check) 71.68b 4.3a 11.40a 77.95bc 32.08ab 37.52ab
Local check 57.56g 3.22c 10.48ab 46.86d 17.34e 19.83d
Mean 65.9 4 9 77 27.11 30.83
CV (%) 12.1 24.7 24.9 35.3 28.1 28.2
Sig.level ** ** ** ** ** **
**: significant level at p≤.01
AMMI analysis: AMMI analysis identified that two of the interaction principal
components axis were significant at P<0.01% and explained 77.01% of the
genotype by environment interaction sum square (Table 4). AMMI adjusted and
re-ranked tuber yield of each genotype by their respective IPCA axis score and
environmental IPCA axis scores, and thereby brought about a significant change
in the ranks of genotypes across environments. The relative contribution of each
environment and genotypes to the G X E interaction were measured based on the
magnitude of respective IPCA 1 score, which is measured as the their
perpendicular distance from the benchmark, IPCA 1 = 0 (Figure 1).
The genotype main effect (Genotype) plus genotype by environment

interaction(GGE biplot) analysis revealed that first principal component(Axis 1)
[479]
and second principal component (Axis2) accounted for 51.28% and 25.73% of
GGE sum of squares of tuber yield, respectively explained a total of 77.01%
variation (Figure 1). This result showed that there was a differential yield
performance among potato genotypes across testing environments due to the
presence of genotype by environment interaction (GEI). Moreover, the GGE bi-
plot classified the testing locations into three mega-environments. Hence, Adet,
Debre Birhan and Holetta grouped in the same mega-environment where the two
best-performed potato genotypes G3 (CIP-392657.171) and G4 (CIP-396034.268)
lied in these locations. Whereas Bekoji and Sinana areas grouped in a separate
sub-region; accordingly genotype G9 (Belete) and G7 (CIP-393220.54) were the
winning genotypes at Bekoji and Sinana, respectively. This implies that Bekoji
and Sinana were different mega-environments which may require potato varieties
for maximum tuber yield.
Suitability of genotypes for a particular environment using the “Which-Won-

Where” function of a GGE biplot: Figure 1 shows the which-won-Where view of
this study. These test environments fall in three of them. The vertex genotype for
the sector which encompasses Adet, Holetta and Debre Birhan imply that CIP-
396034.268 genotype was the winning genotype for those environments and the
variety CIP-392657.171was the second top yielder genotype across the three
locations. The standard check Belete was the best performing genotype at Bekoji
area; whereas genotype 7(CIP-393220.54) was better at Sinana. Hence, for these
environments (Sinana and Bekoji areas) might need specific recommendations for
indicated genotypes in the vertex (Figure 1).
[480]
Table 3. Mean performance of potato genotypes for tuber yield at each environment and across environments
Environments
Genotypes AD2016 AD2017 BE2016 BE2017 DB2016 DB2017 HO2016 HO2017 SI2016 SI2017 Combined
CIP-392797.22 17.92bc 48.16b 12.59c 43.54ab 34.19bcd 15.12c 37.68bc 46.87a 24.35abc 24.94bc 30.54cd
CIP-395017.242 19.78b 30.06c 31.39b 55.40a 31.61cd 26.43bc 44.99ab 36.65bcde 16.65c 20.57cd 31.35cd
CIP-392657.171 36.20a 63.85a 26.40b 36.84ab 29.72cd 23.79bc 47.65ab 46.07a 20.98bc 24.90bc 35.64ab
CIP-396034.268 38.48a 47.13b 25.00b 30.95bc 48.71ab 47.33a 56.89a 42.65ab 17.67c 20.77cd 37.56a
CIP-391058.175 9.16c 45.97b 26.85b 30.12bc 29.24cd 27.36bc 22.40d 40.25abcd 19.93bc 23.85bc 27.51de
CIP-395017.229 7.67c 22.23c 28.89b 27.97bc 27.65d 19.15bc 39.45bc 33.58cbd 24.52abc 28.44abc 25.95e
CIP-393220.54 20.12b 43.54b 31.03b 34.53abc 36.70bcd 28.77b 31.16cd 42.15abc 31.07a 34.99a 33.41bc
CIP-392633.64 20.49b 45.77b 8.89c 14.83c 42.53abc 17.55bc 46.33ab 29.64e 27.39ab 31.31ab 29.09de
Belete 25.12b 46.47b 43.25a 38.55ab 52.4a 26.99bc 40.52bc 32.26de 27.56ab 31.49ab 37.52a
Mean 21.66 43.69 26.03 34.75 36.97 25.83 40.79 38.90 23.35 26.81 31.88
C.V(%) 27.4 17.8 18.3 39 23.1 29.9 18.9 13.1 22 22 24
P *** *** *** * * ** ** ** * ** ***
AD2016, Adet in 2016; AD2017, Adet in 2017; BE2016, Bekoji in 2016; BE2017, Bekoji in 2017; DB2016, Debre-Birhan in 2016; DB2017, Debre-Birhan in 2017; HO2016, Holetta in 2016;
HO2017, Holetta in 2017; SI2016, Sinana in 2016; SI2017, Sinana in 2017. C.V (%), coefficient of variation, P (probability of significance level) where, * significant at 0.05; ** significant
at 0.01 and *** significant at 0.001.
[481]
Table 4. AMMI Anova for tuber yield (t/ha) of potato genotypes across environments
Source of Sum of Degree of Mean of

F F-PROB Explained variation (%)
variation squares freedom Squares
Environment(ENV) 9627.398 4 2406.849 31.73651 0
Genotype(GEN) 8095.417 9 899.4907 11.86061 0
ENV*GEN 11070.83 36 307.523 4.05497 0
PC1 2466.546 12 205.5455 2.70344 <0.00191 51.28
PC2 1673.761 10 167.3761 2.20142 <0.0183 25.73
Residuals 18959.62 250 75.8385
AMMI: additive main effect and multiplicative interaction; PC1: principal component axis1; PC2: principal component axis 2
1=CIP-392797.22, 2=CIP-395017.242, 3=CIP-392657.171,4=CIP-396034.268,5=CIP-391058.175,6=CIP-

395017.229,7=CIP-393220.54,8=CIP-392633.64,9=Belete(st.ck) and10=Local check
Figure 1. Which-won-where view of GGE bi-plot of potato genotypes across locations and over years

Analysis of variance revealed that the parameters considered had a significant
difference (P≤0.05) among the tested potato genotypes. Two genotypes, CIP-
392657.171 and CIP-396034.268, gave higher tuber yield in most testing
environments across environments. Accordingly, genotype CIP-396034.268 gave
significantly higher tuber yield (45.2t/ha) over the standard check (Belete) at Adet,
Holeta and D/Birhan followed by CIP-392657.171(39.7t/ha) since the two potato
genotypes were relatively stable across six environments. GGE bi-plot classified
the tested locations in two mega-environments along with its winner genotypes.
Accordingly, Adet, Holetta and D/Birhan considered as high yielding
environments where clustered in the same sub-region (mega-env’t); whereas
Bekoji and Sinana were low yielding environments for this particular experiment.
The two selected genotypes (CIP-396034.268 and CIP392657.171 ) had also
[482]
shallow eye depth and good tuber shape. Therefore, the two best-performed
genotypes CIP-392657.171 and CIP-396034.268 verified along with checks for
release in 2019 across locations(Adet, Holetta and Debre Birhan); and evaluated
by variety releasing technical committee. At least one potato variety will be
released and meant for large scale production to farmers by the year 2020.
Acknowledgment
The authors greatly acknowledge the Ethiopian Institute of Agricultural Research
(EIAR) and Amhara Agricultural Research Institute (ARARI) and the financial
support provided by the Agricultural Growth Program project (AGP-II). We also
acknowledge researchers who involved during the execution and data collection of
the national variety trial.
References
Annicchiarico P. 2002. Genotype × Environment Interactions Challenges and Opportunities for Plant
Breeding and Cultivar Recommendations.
Baye Berihun. 2017. Potato Breeding for Nitrogen-Use Efficiency: Constraints,
Achievements J. Crop Sci. Biotech. 2018. (September) 21 (3) : 269-281.
Bocianowski.J, Kozak M, Liersch A, Tartanus M, Piotto FA and Azevedo RA. 2011. Genetic
divergence is not the same as phenotypic divergence.
Ebdon JS and Gauch HG. 2002. Additive Main Effect and Multiplicative Interaction Analysis of
National Turf grass Performance Trials: II.Cultivar Recommendations, Crop Sci. 42:497–506.
FAO. 2018. Food and Agriculture Organization of the United Nations, Rome, Italy. www.fao.org ›
faostat.
Genstat software version 18.1 edition. 2015.VSN international limited, 2015.
Gauch HG and Zobel RW. 1996. AMMI analysis of yield trials.Pages 1–40 in M. S. Kang and H. G.
Gauch, eds. Genotype-byenvironment interaction. CRC Press, Boca Raton, FL.
Gedif M, Yigzaw D and Genet T. 2014. Genotype-environment interaction and correlation of some
stability parameters of total starch yield in potato in Amhara region, Ethiopia. Journal of plant
breeding & crop science. 6(2):31-40.
Roozeboom KL, Schapaugh WT, Vanderlip RL and Tvinstra MR. 2008. Testing Wheat in Variable
Environments: Genotype, Environment, Interaction Effects, and Grouping Test Locations. Crop
Science 48(1):317
Landeo JA, Gastelo M, Pinedo H and Flores F. 1995. Breeding for horizontal resistance to late blight in
potato free of Rgenes. In Phyto phthora infestans 150, 268–274 (Eds
L.J.Dowley,E.Bannon,L.R.Cooke,T.Keane and E. O’Sullivan). Dublin: Boole Press.
Piepho HT. 1996. A Monte Carlo Test for Variance Homogeneity in Linear Models. Biomedical
journals.
Yigzaw D, Fentahun M and Tesfaye A. 2008. Performance stability analysis of potato varieties under
rainfed and irrigated potato production systems in Northwestern Ethiopia. Eth.J.Sci. and Technol.
5(2):90-98.
Yan W. 2001. GGEbiplot – a Windows application for graphical analysis of multi-environment trial
data and other types of two way data. Agron. J. 93: 1111–1118.
Yan W. 2002. Singular value partitioning for biplot analysis of multi-environment trial data. Agron J.
94: 990–996.
[483]
[484]
Adaptation of Korean Potato
(Solanum tuberosum L.) Varieties
under Irrigation
Abebe Chindi1*, Kasaye Negash1, Egata Shunka1, Gebremedhin W/Giorgis1, Tesfaye
Abebe1, Fikadu Gebretinsay2, Niguse Abebe3,
Wasu Mohammed4 and Zerihun Kebede5
1
Ethiopian Institute of Agricultural Research/EIAR/, Holetta Agricultural Research Centre,
Department of Horticulture, P.O.Box 2003, Addis Ababa, Ethiopia: 2Ethiopian Institute of Agricultural
Research Centre/EIAR/, Kulumsa Agricultural Research Centre: 3Tigary Agricultural Research Institute
(TARI), Mekelle Agricultural Research Centre: 4 Haramaya Universities
5
Amhara Region Agricultural Research Institute (ARARI), Debre Berihan Agricultural Research Centre;
*Corresponding author: abechindii@yahoo.com, abechindi@gmail.com
Abstract
The adaptability of crops can vary from location to location depending on the agro-
ecology of a particular area. Therefore, it is essential to conduct a location-specific
adaptation trial to identify suitable variety/varieties. Accordingly, an adaptation trial of
potato varieties was conducted in the central highlands of Ethiopia to evaluate the
performance and adaptability of introduced and improved varieties in different potato
growing agro-ecologies of the country. Treatments include five potato varieties
introduced from the Republic of South Korea (Chubak, Seohong, Goun, Jowon, and
Haryung) and three nationally released varieties (Awash, Gudanie and Belete) were
evaluated for their vegetative growth performance and tuber yield under irrigation
conditions. The experimental field was laid out in Randomized Complete Block Design
(RCBD) with three replications. The results from Holetta and Kulumsa revealed that
plant height, main stem number, average tuber number and average tuber weight per
plot were highly significant. Belete variety had the highest plant height (65.43cm) and
stem numbers of 6.58 per hill whereas; Chubak variety had the lowest plant height
(26.22 cm) and stem number 3.57 per hill among the other varieties. Potato variety
Haryung gave the highest average tuber number (52.73/ m2) followed by Gown (42.39).
The highest total yield (30.08 t/ha) and marketable tuber yield (17.72t/ha) recorded for
Gudanie variety followed by Haryung (27.09 t/ha) & (18.48 t/ha) and Seohong
(25.54t/ha) & (19.44t/ha), respectively. Finally, the result of the study revealed that the
genotype and growing environment has a great influence on yield and yield components
of potato tubers. It is concluded that the study demonstrated the effect of varietal
differences on the growth potential of potato varieties under irrigation or for short rain
fed areas. Thus, in both seasons among the introduced potato varieties Haryung and
Seohong were well adapted and gave a comparative tuber yield with nationally released
potato varieties and can be used as alternative varieties in potato producing regions of
the country especially under irrigation and short rainfed conditions.
Keywords: Potato, Adaptation, Vegetative growth, Total tuber yield, marketable

tuber Yield, irrigation
[485]
Introduction
In Ethiopia, potato (Solanum tuberosum L.) has a promising prospect in improving

the quality of the basic diet in both rural and urban areas (Abebe et al., 2017).
Potato is an important crop for smallholder farmers in Ethiopia, serving both as a
cash crop and food security crop. It is one of the root crops widely grown in the
country because of increasing demand and emerging markets that have provided a
great opportunity for resource-poor farmers to generate additional income (Bekele
et al., 2011). It is a short duration crop that can mature within a short period. It
contains practically all essential dietary constituents like carbohydrates, essential
nutrients, protein, vitamins, and minerals (Sriom et al., 2017). Potato production
has been considered as the priority compared to other food crops because of its
contribution to food security, income generation and double cropping advantages
and its utilization in different forms (Lung’aho et al., 2007; Muthoni. and
Nyamongo, 2009). It is one of the strategic crops, enhancing food security and
economic benefits to the country. The adaptability of crops can vary from location
to location depending on the agro-ecology of a particular area. Therefore, it is
essential to conduct location-specific adaptation trial to identify suitable potato
variety/varieties (Addis et al., 2017). But lack of well-adapted cultivars to various
abiotic stresses is one of the production problems that account for low yield and
small area cropped to the nation (Seifu and Betewulign, 2017). Accordingly, an
adaptation trial of potato varieties was conducted in central highlands of Ethiopia
to identify potato varieties that is better in adaptation, yield and other agronomic
characteristics, and pest and disease tolerant. Despite high potential production
environments and marked growth, the national average potato yield in Ethiopia is
13.9 t ha-1 (CSA, 2018) which is lower than the experimental yields of over 35 t
ha-1 (Baye and Gebremedhin, 2013) and world average yield of 20 t ha-1
(FAOSTAT, 2019) as well as other top potato producing countries in Africa. The
low yields are the result of a number of production constraints mainly involving
abiotic and biotic stress factors (Hirut, 2015). Among the biotic constraints late
blight, bacterial wilt, virus diseases and potato tuber moth constitute the major
threats to potato production, while the abiotic stresses include soil nutrient
deficiency, frost, drought, erratic rainfall, and air and soil high temperature
especially in marginal areas (Gildemacher et al., 2009; Semegn et al., 2015).
There are many complicated reasons for this low yield of potato in the country.
Lack of good quality seed, soil fertility, unbalanced mineral nutrition, inadequate
application of fertilizers, pests and disease, irregularity of water supply and
traditional irrigation schemes and schedules are the main reason which accounts
for the low productivity of potato (Bezabih and Mengistu, 2011). Although the
irrigated potato production system contributed the lion’s share both in the country
and the region, its productivity (3.7 t ha-1) was lower than the rainfed (10.5 t ha-1)
[486]
system (CSA, 2003). This could be due to differences in climatic conditions and
production constraints of the two production systems. Furthermore, in Ethiopia
researchers have never released improved varieties for the irrigated potato
production system. The prevailing average monthly maximum temperature is
higher in the irrigated potato production system than in the rainfed system. The
average monthly minimum temperature is low and causes frost injury to the plant
during the irrigated potato production system. Therefore, irrigated potato
production is affected both by the prevailing higher maximum and lower
minimum temperature compared to the rainfed potato production system (Yigzaw
et al, 2008). Yield is a complex trait in potato and is generally considered to have
low heritability (Lynch and Tai, 1979). Hence, indirect selection could be a useful
strategy to bring considerable genetic improvement on potato tuber yield.
Therefore, the objective of this study was to evaluate the performance and
adaptability of released potato varieties and commercial potato introduced from
Korea for fresh/table purposes under irrigation production systems.

Description of the study area: The field experiment was conducted under
irrigation conditions during the year 2017/18 and 2018/19 cropping season at
Holetta, Kulumsa, Debre Birhan, Mekelle and Haramaya, in various agro-
ecologies of potato growing regions of Ethiopia. Holetta agricultural research
center is located at 090 00’N, 380 30’E at an altitude of 2400 m.a.s.l. It is 29 km
away from Addis Ababa on the way to Ambo and characterized by a mean annual
rainfall of 1041.4 mm, a mean relative humidity of 58.70%, and mean maximum
and minimum temperature of 21.70 °C and 6.70 °C, respectively. The main rainy
season is from June to September, which accounts for 70% of the rainfall while
the remaining thirty percent is from February to April (EIAR, 2017). The soil of
the center is red Nitosol, which is characterized by an average organic matter
content of 1.8%, Nitrogen 0.17%, pH 5.24, and phosphorus 4.55 ppm (Gemechu,
2007). Haramaya University research farm is located at 2020 meters above sea
level, 9°41''N latitude and 42°03''E longitude. The area has a bimodal rainfall
distribution with mean annual rainfall of 760 mm (Belay et al. 1998). The long
rainy season extends from June to October and accounts for about 45% of the total
rainfall. The mean maximum temperature is 23.4°C while the mean minimum
annual temperature is 8.25°C (Tekalign, 2011). The soil of the experimental site is
a well-drained deep alluvial with a sub-soil stratified with loam and sandy loam.
Kulumsa agricultural research center is located at 8°01’N, 39°09’E at an altitude
of 2200 m.a.s.l with 10.5 ° C minimum and 22.8 ° C temperature, respectively.
The annual rainfall is 850 mm. Similarly, Adet agricultural research center is
located 11°16’N, 37°29’E at an altitude of 2240 m.a.s.l with 12°C minimum and
24°C maximum temperature, respectively. The annual rainfall is 1250 mm. And
Mekele agricultural research center is located 13.3N, 39.4E, with an altitude of
[487]
1970 m.a.s.l with 9.7°C minimum and 26.6 °C maximum temperature,
respectively. The annual rainfall is 550 mm. Debre Berhan is located 9°45′N
latitude 39°31′E longitude, 130km far from Addis Ababa in the North direction. It
is situated on plateaus in the central Ethiopia highland at an average elevation of
between 2800 and 2845 above sea level. The temperature of Debre Berhan is in
average between 6.6°c-24°c; and the average rainfall is 964mm and the climate is
highland (Seyom, 2007).
Experimental treatment and design: The performance of eight potato varieties

was assessed in this study. Five of them are varieties of commercial potato
varieties introduced from the Republic of South Korea and three nationally
released potato varieties were evaluated for adaptability and tuber yield and yield
components using irrigation. A total number of eight (8) potato genotypes were
used for the experiment. The experiment was laid out as a Randomized Complete
Block Design (RCBD) with three replications. Each plot was 3.0m x 3.0m = 9m2
wide consisting of four rows, which accommodated 10 plants per row and thus 40
plants per plot. The spacing between plots and adjacent replication was 1m. At
each site, medium-sized (39-75g) (Lung’aho et al., 2007) and well-sprouted tubers
were planted using irrigation during Feb. 2017 and 2018 for two consecutive years
at the spacing of 75 cm between ridges and 30 cm between tubers. Fertilizer was
applied as the recommendation made by Holetta Agricultural Research Centre,
which Phosphorus and Nitrogen fertilizer was applied at the rate of 92 kg P2O5 ha-
1
and 110 kg ha-1, respectively (Berga et al., 1994). All other cultural practices
were applied according to Holetta Agricultural Research Centre recommendation.
For data estimation, tubers were harvested from middle rows, leaving the plants
growing in the two border rows as well as those growing at both ends of each row
to avoid edge effect (EARO, 2004).
Data were collected on plant emergence, plant height, as well as a number of stem
per plant during the vegetative growth stage, number of tubers per plant, tuber
yield(t/ha), average tuber weight (ATW) in gram, average tuber number
(ANT)/plant, were recorded. Quality parameters such as dry matter content and
specific gravity were taken during harvesting. The data were subjected to analysis
of variance (ANOVA) following the standard procedure given by (Gomez and
Gomez, 1984). After fitting the ANOVA model for those significant response
variables, the mean separation was carried out using the LSD method at a 5%
level of significance. All the statistical analyses were carried out using SAS-9.2
statistical software package (SAS, 2009).

Results of analysis of variance (ANOVA) indicated that five growth characters for
eight (8) improved potato varieties were shown in Table 1. Accordingly, all the
[488]
growth parameters considered revealed a highly significant difference (P<0.01)
among the tested varieties. During 2017/18 off-season using irrigation, the highest
plant height was recorded for Gudanie variety followed by Belete. The highest
main stem number was observed for Chubak followed by Seohong and Haryung.
As indicated in (Table 1) the highest average tuber number was recorded for
Awash variety followed by Seohong and Gudanie. The highest average tuber
weight was recorded for Awash variety followed by Seohong and Belete. In line
with the study, Kena et al. (2018) reported that the interaction effect of variety,
location and year showed significantly different (p<0.05) on the number of main
stem per plant. The difference in plant height among the varieties might be
associated with genetic differences, which may lead to the variable performances
in growth and development (Habtam, 2012). The differences might be due to
varietal effect and plant canopy which determine the main stem to different
locations (Azad et al., 1997). In the irrigated potato production system, marketable
tuber yield showed positive and statistically significant association with number of
main stems per plant, plant height and number of tubers harvested per plot
(Yigzaw et al., 2008).
In addition, De la Morena et al. (1994) described that the difference in number of

number main stem among the varieties might be due to the inherent genotypic
variation in the number of buds per tuber which is in turn influenced by the size of
the tubers, physiological age of the seed, storage condition, and number of viable
sprouts at planting, sprout damage at the time of planting and growing conditions.
Similar to other parameters, the highest marketable tuber number was obtained
from variety Awash followed by Gudanie and Chubak. Even though there were
inconsistent results with regard to the growth parameters, the introduced potato
varieties showed comparable average tuber number, tuber weight as well as
marketable tuber number. The result at Haramaya indicated that during 2017/18
the highest total and marketable tuber yield was recorded from variety Seohong
followed by variety Haryung which gave a total and marketable tuber yield.
However, the lowest total and marketable tuber yield was obtained from the
variety Goun. Thus, the yield differences between these varieties may be related to
their genetic makeup in the efficient utilization of inputs like nutrient as reported
by (Tisdale et al., 1995).
[489]
Table 1: Performance of eight potato varieties for Plant height (PH), Main Stem Number (MSN), Average Tuber Number
(ATN), Average Tuber Weight (ATW) and Marketable Tuber Number at Holetta, D. Berihan & Kulumsa during
2017/18.
No Varieties PH (cm) MSN/plant ATN/m 2 ATW (g) MTN/plot

1 Chubak 56.71b 6.27a 11.61bc 75bc 141.33ab
2 Seohong 47.56c 5.93ab 12.91ab 89a 140.88ab
3 Goun 40.47 d 3.64e 7.32 f 57 d 97.33c
4 Jowon 33.27e 3.62e 9.60de 65cd 85.66c
5 Haryung 35.98de 5.49abc 8.32ef 71bc 103.22c
6 Awash 55.47b 4.51de 14.03a 91a 152.55a
7 Gudanie 64.60a 5.20bcd 12.39ab 65cd 141.44ab
8 Belete 63.22a 4.82cd 10.57cd 82ab 124.00b
CV (%) 11.30 19.66 17.50 19.37 16.41
LSD
5.32 0.92 1.80 0.14 19.20
(0.05%)
P-Value <.0001 <.0001 <.0001 <.0001 <.0001
Means followed by the same letters with in the same column are statistically non-significant at p<0.05 according to the
least significant difference (LSD) test.Var= Variety, Loc= location, Var*Loc= Interaction of variety with location, CV=
Coefficient of Variation
There was highly significant (P< 0.01) variation among the tested varieties with
respect to total tuber yield, marketable tuber yield, dry matter content and specific
gravity among the evaluated varieties as indicated in (Table 2). During 2017/18
the over locations results showed highly significant (P< 0.01) variation for the
tested clones however the interaction effects of varieties and locations were non-
significant. The highest total and marketable tuber yield was recorded from variety
Seohong followed by Chubak. The lowest total and marketable tuber yield was
obtained from the variety Goun. With regard to dry matter content and specific
gravity, the highest result was obtained from Belete followed by Haryung, Goun
and Seohong in descending order (Table 2). However, except Chubak variety, the
other tested genotypes and Belete gave similar specific gravity values. In line with
the result, Getachew (2016) found that varieties Belete and Gudanie produced
tubers with higher dry matter content. Inconsistent with this result, Tai and
Coleman (1999) reported that dry matter content is subjected to the influence of
both the environment and genotypes. In the same manner, Rytel et al., (2013)
reported that the quality of potato tubers and their chemical composition is
influenced by genetics, soil fertility, weather conditions and chemical treatments
that are applied. Similarly, Habtamu et al., (2016) also reported the highest
specific gravity (1.0967) for variety Belete evaluated at three locations of eastern
Ethiopia. In general, the current investigation also agreed with these different
scholars findings. Kena et al. (2018) described that the variation in total yield of
potato genotypes at different locations may be due to a response of the genotypes
to growing environmental factors. This suggestion is in agreement with other
authors who reported that yield differences among genotypes were attributed both
by the inherent yield potential of genotypes and growing environment as well as
the interaction of genotype x environment(Kena et al., 2018).
[490]
Table 2: Performance of eight potato varieties for Marketable Tuber Yield (MTY), Total Tuber Yield (TTY), and Dry matter
content (DM) and Specific gravity at Holetta, D. Berhan, Kulumsa, Mekele & Haramaya during 2017/18.
No Varieties MTY (t/ha) TTY (t/ha) DM (%) 3
SG(g/cm )
1 Chubak 30.95b 33.52bc 20.83c 1.07c
2 Seohong 37.95a 41.78a 23.45ab 1.08ab
3 Goun 20.07 d 22.49d 24.31 a 1.08ab
4 Jowon 23.84cd 27.20cd 21.53bc 1.08b
5 Haryung 28.61bc 31.44bc 24.47a 1.09a
6 Belete 31.07b 34.38b 25.19a 1.09ab
CV(%) 29.07 28.13 12.61 1.22
P-value Clo <.0001 <.0001 0.0005 0.0010
Loc <.0001 <.0001 <.0001 <.0001
Clo*Loc NS NS NS NS
Means followed by the same letters with in the same column are statistically non- significant at p<0.05
according to the least significant difference (LSD) test.Var= Variety, Loc location, Var*Loc= Interaction of
variety with location, CV= Coefficient of Variation
This experiment was repeated during 2018/19 at the same locations except
Haramaya University. The data from the remaining testing locations of Holetta,
Kulumsa, Debre Birhan and Mekelle Agricultural research centers summarized in
Table 3. The ANOVA table indicates that the growth parameters were highly
significant (P<0.01) for the evaluated varieties as well as for locations except
marketable tuber yield. There was a highly significant (P<0.01) difference for the
interaction effect of varieties and locations also. The highest main stem number
was recorded for variety Haryung and Gudanie followed by Belete and Seohong.
The lowest main stem number was recorded by Chubak variety. The highest
average tuber number was produced by variety Haryung followed by Gown and
Gudanie, respectively. The lowest average tuber number was recorded by Belete
variety. The highest total and marketable tuber yield recorded for Gudanie variety
followed by Haryung and Seohong, respectively (Table 3). Similarly, other
researchers also investigated that marketable yield was significantly varied by
variety, location and genotypes x environment interaction (Kumar et al., 2007).
Table 3: Performance of eight potato varieties forMain stem number (MSN), Average tuber number (ATN), Average tuber
weight (ATW), Marketable tuber Yield (MTY) and Total tuber yield (TTY) for at Holetta & Kulumsa-2018/19.
No Varieties MSN ATN/m2 ATW/plant MTY (t/ha) TTY (t/ha)
1 Chubak 3.57 b 37.37b 50.19c 14.73bc 19.28c
2 Seohong 6.27a 30.09c 51.27c 19.44a 25.54b
3 Goun 6.19a 42.39b 78.21a 16.12ab 23.60bc
4 Jowon 4.18b 41.23b 81.86a 19.17a 22.77bcd
5 Haryung 6.68a 52.73a 49.14c 18.48ab 27.09ab
6 Awash 5.90 a 36.79b 60.55 b 12.19 c 18.53d
7 Gudanie 6.68a 40.43b 59.76b 17.72ab 30.08a
8 Belete 6.58a 29.78c 62.68b 14.98bc 24.65b
CV (%) 18.81 12.31 10.49 18.29 14.72
P-value LC <.0001 0001 <.0001 NS <.0001
VAR <.0001 <.0001 <.0001 0.0025 <.0001
LC*VAR 0.0076 0.0016 <.0001 <.0001 <.0001
Means followed by the same letters with in the same column are statistically non-significant at p<0.05 according to the
least significant difference (LSD) test.Var= Variety, Loc=location, Var*Loc= Interaction of variety with location, CV=
Coefficient of Variation
[491]
The implication of the study showed that the varieties studied had good and
promising agronomic and tuber yield traits useful for breeding and utilization
purposes, particularly towards irrigation production system. The studied varieties
indicated that there was genetic variability and the varieties can be utilized for
various agro-ecologies based on their performance. In addition to the nationally
released potato varieties, those introduced Haryung and Seohong which adapted
very well and produced comparable tuber yield was recommended as alternative
varieties under irrigation production. The limitation of the study was that, those
nationally released varieties were not properly included across locations to exploit
their potential yield under irrigation and need further evaluation.

The study showed the presence of genetic variability among potato genotypes for
tuber yield and adaptability for various agro-ecologies of potato production. The
result of the study justifies that different varieties had different genetic potential
across locations. However, the findings indicate that there was no superior
genotype for all tested characteristics as compared to the nationally released
varieties. The mean analysis showed that Haryung and Seohong varieties
produced equivalent total and marketable tuber yield across a location that was
comparable to nationally released varieties, Gudanie and Belete under Holetta and
Kulumsa conditions for irrigation production system. Thus, the introduced potato
varieties showed comparable tuber yield with the nationally released varieties for
tuber yield, dry matter and specific gravity and other related agronomic traits at
Holetta, Kulumsa, Debre Birhan and Mekelle conditions. Among the tested potato
varieties, Haryung and Seohong were recommended for adaptability and better
tuber yield and quality across testing locations under the irrigation production
system.
Acknowledgment
We would like to thank the Ethiopian Institute of Agricultural Research (EIAR)
through Agricultural Growth Program (AGP_II) for financing this research and
Holetta Agricultural Research Centre (HARC) to facilities and operate the
research. It’s our pleasure to thank RDA of South Korea and International Potato
Centre (CIP) for the provision of potato germplasm. Administration and research
technicians involved in the activity are acknowledged.
References
Abebe Chindi, Egata Shunka, Atsede Solomon, Gebremedhin W/Giorgis, Ebrahim Seid, Lemma
Tessema. 2017. Participatory potato seed production: a breakthrough for
[492]
food security and income generation in the central highlands of Ethiopia. Open
Agriculture. 2: 205–212.
Addis Shiferaw, Dessalegn Regassa and Wakene Tigre. 2017. Irish Potato (Solanum Tuberosum L.)
Variety Evaluation at Bule Hora District of Borena Zone. Global Journal of Science Frontier
Research: D Agriculture and Veterinary. Volume 17 Issue 2 Version 1.0. Online ISSN: 2249-
4626 & Print ISSN: 0975-5896.
Azad S, Nehra B, Khurana S and Singh N., 1997. Influence of plant density and geometry on
growth and yield in seed crop of potato. Indian Potato Assoc 24: 17-23.
Belay, S. C., Wortman, W. and Hoogen, G. 1998. Haricot bean agro-ecology in Ethiopia: definition
using agro-climatic and crop growth stimulation models. Africa Crop Science Journal,6:9-18.
Berga, L., Gebremedhin, W.G., Teriessa, J., Bereke-Tsehai, T., and Yaynu, H. 1994. Potato
improvement research. In: Proceedings of the 2nd National Horticultural Workshop of Ethiopia.
1-3 Dec. 1992, Addis Ababa. EIAR, Addis Ababa.
Baye Berihun and Gebremedhin Woldegiorgi, 2013. Potato research and development in
Ethiopia: achievements and trends. In: G. Woldegiorgis, S. Schulz and B. Berihun, editors, Seed
Potato Tuber Production and Dissemination Experiences, Challenges and Prospects. EIAR and
ARARI, Bahir Dar, Ethiopia. Pp. 35-44.
Bezabih, E. and Mengistu,. 2011. Potato value chain analysis and development in the case of Tigray
and SNNPR Regions, Ethiopia. International Potato Centre (CIP-Ethiopia), Addis Ababa.
CSA (Central Statistical Agency), 2018. Agricultural sample survey of area and production of crops of
2018/2019 (2010 E.C.) in Ethiopia. Volume I. Report on area and production of major crops.
Private peasant holdings-Meher season. Statistical bulletin 586. April 2018. Addis Ababa,
Ethiopia. 53pp.
Central Statistical Agency (CSA), Ethiopian Agricultural Sample Survey Report of 2015/2016.
Statistical Report on Farm Management Practices Private Peasant holdings, Meher season, Addis
Abeba, Ethiopia, 2015, Volume I, 17-19.
Central Statistical Agency (CSA), Ethiopian Agricultural Sample Survey Report of 2003.
Statistical Report on Farm Management Practices Private Peasant holdings, Meher season, Addis
Abeba, Ethiopia, 2003, Volume III, 17-19.
De la Morena, Guillen, I.A. and Garcia del Morel, L.F. 1994. Yield development in potatoes as
influenced by cultivar and the timing and level of nitrogen fertilizer. American Potato Journal,
71:165-173.
EIAR. 2017. Holetta agricultural research center progress report 2017. EIAR/ HARC.
Ethiopian Agricultural Research Organization (EARO), 2004. Directory of released crop varieties and
their recommended cultural practices. Ethiopian Agricultural Research Organization, Addis
Ababa, 36 p.
FAOSTAT (Food and Agricultural Organization Statistic). 2019. World food and agricultural
organization data of statistics. Rome, Italy. FAO, Bulletin, Rome, 10:275.
Gemechu Keneni. 2007. Phenotypic diversity for biological nitrogen fixation in Abyssinian field pea
(Pisumsativum var. abyssinicum) germplasm accession. Doctoral Dissertation Addis Ababa
University, Addis Ababa, Ethiopia.
Getachew Kahsay. 2016. Response of potato (Solanum tuberosum L.) varieties to nitrogen and blended
fertilizers under irrigation at Maichew, Southern Tigray, Ethiopia. MSc Thesis, Haramaya
University, Haramaya.
Gildemacher PR, Demo P, Barker I, Kaguongo W, Woldegiorgis G, Wagoire WW and Struik PC.
2009. A description of seed potato systems in Kenya, Uganda and Ethiopia. American journal of
potato research, 86(5), 373-382.
Gomez KA and Gomez AA. 1984. Statistical Procedures for Agricultural Research, 2nd edition, John
Wiley and Sons, New York, p.680.
Habtamu Gebreselassie, Wassu Mohammed and Beneberu Shimelis. 2016. Evaluation of potato
(Solanum tuberosum L.) varieties for yield and yield components in Eastern Ethiopia. Journal of
Biology, Agriculture and Healthcare, 6(5): 146-154.
[493]
Habtam Setu. 2012. Response of Potato (Solanum tuberosum L.) to the application of Phosphorus and
Potassium fertilizers at Assosa, Benishangul Gumuz Regional State, Western Ethiopia. MSc.
Thesis, Haramaya University, Haramaya.
Haverkort A, van Koesveld F, Schepers H, Wijnads J, Wustman R and Zhang X. 2012. Potato
Prospects for Ethiopia: on the road to value addition. Praktijkonderzoek Plant and Omgeving
(Applied Research). The Netherlands, PPO Publication no. 528, p.66.
Hirut Betaw. 2015. Genetic analyses of drought tolerance and resistance to late blight among potato
genotypes. 120 Pp. A thesis submitted in partial fulfillment of the requirements for the degree of
Doctor of Philosophy (PhD) in Plant Breeding. African Centre for Crop Improvement (ACCI),
School of Agriculture, Earth and Environmental Sciences, College of Agriculture, Engineering
and Science, University of KwaZulu-Natal, Republic of South Africa.
Kena Kena. 2018. Adaptability and Performance Evaluation of Potato (Solanum tuberosum L Varieties
under Irrigation in West and Kellem Wollega Zones. J Advan Plant Sci 1: 213.
Kidest Firde, Habtam Setu, Tenagne Eshete, Tajebe Mosie, Getaneh Sileshi and Edeo Negash. 2019.
Growth, Yield, and Fruit Quality Performance of Peach Varieties. Ethiop. J. Agric. Sci. 29(2) 45-
58.
Kumar S, Khade HD, Dhokane VS, Bethere AG and Sharma A. 2007. Irradiation in Combination With
Higher Storage Temperatures Maintains Chip-Making Quality of Potato. J Food Sci. 72:6.
Lung’aho C, Lemaga B, Nyongesa M, Gildermacher P, Kinyale P. Demo P and Kabira J. 2007.
Commercial seed potato production in eastern and central Africa. Kenya Agricultural Institute.
140p.
Lynch DR and Tai GCC. 1979. Yield and yield components response of eight potato genotypes to
water stress. Crop Science, 29 (5): 1207-1211.
Muthoni J and Nyamongo DO. 2009. A review of constraints to ware Irish potatoes production in
Kenya. Journal of horticulture and forestry vol. 1(7). Kenya.
Rykaczewska K. 2013.The impact of high temperature during growing season on potato cultivars with
different response to environmental stresses. Amer. J. Plant Sci., 4; 2386-2393.
Rytel E, Lisinska G and Tajner-Czopek A. 2013. Toxic compound levels in potatoes are dependent on
cultivation methods. ACTA Alimentaria, 42(3): 308-317.
SAS Institute. 2009. The SAS System for Windows. SAS Inst., Cary, NC.
Seifu Fetena and Betewulign Eshetu. 2017. Evaluation of released and local potato (Solanum
tuberosum L.) varieties for growth performance. J. Agron., 16: 40-44.
Seyom E. 2007. The assessment of physico-chemical parameters of river Beressa in Debre Berhan
(North Shoa) for sustainability drinking water, Addis Abeba, Ethiopia
Sriom DPM, Priyanka R, Devraj S, Rajat KS and Sudhir KM. 2017. Effect of Different Levels of
Nitrogen on Growth and Yield in Potato (Solanum tuberosum L.) CV. Kufri Khyati.
Int.J.Curr.Microbiol.App.Sci; 6(6): 1456-1460.
Tai GCC and Coleman WK. 1999. Genotype environment interaction of potato chip color.
Canadian Journal of Plant Science, 79: 433–438.
Tekalign T. 2011.Genotype x environmental interaction for tuber yield, dry matter content and specific
gravity in elite tetraploid potato (Solanum tuberosum L.) genotypes. East Africa Journal of
Sciences, 5 (1): 1-5.
Tisdale SL, Nelson W, Beaton JD and Havlin JL. 1995. Soil fertility and fertilizers. 5th edition.
Macmillan, New York.
Yigzaw Dessalegn, Fentahun Mengistu and Tesfaye Abebe. 2008. Performance Stability Analysis of
Potato Varieties under Rainfed and Irrigated Potato Production Systems in Northwestern
Ethiopia. Eth.J.Sci & Technol. 5(2):90-98, 2008.
[494]
Collection and Characterization of Local Potato
Cultivars from Eastern and Western Ethiopia
Semagn Asredie1*, Alemu Worku2,
Manamino Workayehu2, and Egeta Shunka3
1
Amhara Agricultural Research Institute, Bahir Dar, Ethiopia; 2Adet Agricultural Research Center, Adet,
Ethiopia; 3Holetta Agricultural Research Center, Holetta, Ethiopia;
*
Corresponding author: sekol2009@gmail.com
Abstract
Characterizing and assessing the importance and spatial distribution of local potato
genetic resources grown in the country is an important first step for conservation,
improvement and other uses of the crop. This study was conducted to collect,
characterize and assess the potential benefits, limitations and areas of adaptation of
local potato cultivars from eastern and western Ethiopia using morphological markers.
Local potato cultivars were collected from three districts in East Harerghe (Kombolcha,
Haramaya and Kufa Chelle), two districts in West Shewa (Jeldu and Chelia) and one
district in East Wollega (Jimma Arjo) using morphological descriptors and farmers
perception. Fourteen cultivars were collected and planted in isolation under screen
house conditions at Adet Agricultural Research Center. They were characterized using
key morphological markers such as leaf color and morphology, flower color and tuber
characteristics. At the same time, they were tested for major diseases using the ELISA
method and those showed negative results were conserved in TC- lab at Amhara
Regional Agricultural Research Institute (ARARI). In East Harerghe, from seven
cultivars collected, three of them are found to be morphologically distinct and the rest
are duplicates known by different names. According to the perception of the farmers and
our observation, the local cultivars dominate potato production. The widely grown local
cultivars in this area are also widely grown in southern Ethiopia including Shasemene.
In west Ethiopia, seven cultivars of the collections in are found to be morphologically
distinct. However, the shares of the production of these local cultivars are very low.
Instead of improved varieties are widely grown. From the local cultivars, ‘Chala’ is
morphologically identical to ‘Menagesha, an improved variety. Five of these cultivars
are rarely grown while ‘Dalicho’ is produced by some farmers in this part of the
country. This study shows that widely demanded local varieties are produced in a wider
range in Ethiopia. These cultivars are important assets for potato improvement and
other uses. The characterization in this study is merely based on morphological markers,
it may not be reliable; it should be supported by molecular markers.
Keywords: Local cultivars, East Harerghe, West Shewa, East Wolega,

Morphological markers.
Introduction
Potato production in Ethiopia is dominated by ‘local cultivars’ of unknown origin.
Local varieties are farmer varieties that have not been formally tested through
variety selection procedures by breeding programs. These cultivars were
presumably developed through a farmer-driven process of evaluation and selection
of varieties/clones introduced from outside sources. The percentage of total potato
[495]
acreage planted to local cultivars is about 73-77% in Ethiopia (Labarta et al. 2012;
Kolech et al. 2015).
Local potato cultivars usually yield less than improved varieties under optimal
conditions but because of their relative stability and unique suitability for specific
uses, are nevertheless important in providing food security at the household level
(Barnaud et al. 2007). Local varieties of buffer risks associated with unfavorable
environments (Padgham 2009). According to Padgham (2009), the natural
selection processes that farmers used to identify (and ultimately adopt) local
varieties enhances the value of these genetic resources for future crop
improvement. In Ethiopia, farmers prefer to use local varieties for dry season
production. Since local varieties have long been thought to have low yield
potential (Kidane Mariam 1979; Lemaga 1983), they have received little attention
by the Ethiopian research system.
A full assessment of local cultivars is an important first step for crop improvement
in developing countries (Williams et al., 1991). The success of any plant breeding
effort to provide varieties that are adapted to local abiotic and biotic stresses
depends on a continual use of locally adapted varieties and wild relatives of
closely related species for genetic enhancement (Padgham 2009). As Bradshaw et
al. (2006) noted, better-adapted cultivars typically result from crosses between
diverse sets of parents with complementary features, where at least one parent is
adapted to local growing conditions. Varieties developed with the use of existing
local varieties as parents are more likely to be adapted to local growing conditions.
Therefore, characterizing and assessing the importance and spatial distribution of
these genetic resources is very important to conserve and use them for crop
improvement and other uses. As Williams et al. (1991) noted the first step in crop
improvement in developing countries should be a full assessment of the local
material. That is, whether local cultivars have considerable variation or not in
relation to key traits. Based on this, our potato genetic resources need to be well
assessed and conservation measures need to be in place. So far, local varieties
from the northwest and southern Ethiopia have been collected and assessed using
morphological and molecular markers. The potential benefits, limitations and
areas of adaptation of these cultivars have been studied. However, to make this
study complete, the varieties from east and west Ethiopia have given high concern
due to the probability of getting drought tolerant varieties especially in this part of
the region. According to CSA (2014) potato is widely grown in these areas
especially in East Harerghe and West Shewa. Thus, this study has been conducted
to characterize varieties from eastern and western Ethiopia using morphological
markers and assess the potential benefits, limitations and areas of adaptation of
varieties grown in these parts of Ethiopia.
[496]
Methodologies
The collection work was conducted in major potato districts in east and west
Ethiopia during 2016-2017. In east Ethiopia, districts from East Harerghe zone
such as Haramaya, Kombolcha, Gurawa and Kufa Chelle were included in the
collection work. In West Ethiopia, districts in West Shewa (Dendi, Jeldu, and
Chelia), Horogudru Wollega (Horo) and East Wollega (Jima Arjo) were included.
These collection sites were selected after consulting CSA (2014) and different
research centers (Haramaya University, Holetta and Bako Agricultural Research
Centers).
Questionnaire checklists were organized and used during the local variety
collection. The names of each variety were provided by local farmers. Forty tubers
each of six local varieties were collected from East Harerghe (Haramaya,
Kombolcha, Kufa Chelle) and forty tubers of the other seven varieties were
collected from West Shewa (Jeldu, Chelia) and East Wollega (Jima Arjo) (Figure
1).To establish whether a variety with a given name grown in one district is
similar or different across districts, or within a district, a sample of 25 tubers of
each variety grown in each district were collected from farmers’ fields. These
tubers were multiplied in a screen house at the Adet Agricultural Research Center
for morphological testing and genetic fingerprinting. The varieties were tested for
6 viruses and bacterial wilt at ARARI TC lab using ELISA method.
Consequently, from the tested cultivars, ‘Abaminimine’ did not pass the test.
Morphological assessment: The vines, flowers and tubers of local cultivars were
characterized using descriptors for cultivated potato (Huaman et al. 1977). For leaf
and stem characteristics, stem color, shape of stem cross-section, abaxial and
adaxial leaf pubescence, leaf dissection, leaf size, leaf color, growth habit and
branching habit were recorded. Flower characteristics measured included flower
color, degree of flowering, and tendency to premature flower abscission. Tuber
characteristics recorded after harvest includes tuber skin color, tuber flesh color,
tuber skin texture, secondary tuber color, tuber shape, and storability.
Local varieties in East Ethiopia: Improved varieties were not common in this
region; local cultivars dominate the potato production. Seven local cultivars such
as Ashe, Shentemtella, China, Jarso, Betie, Dadafa and Jiru Belina were collected
in East Harerghe (Haramaya, Kombolcha, and Kufa Chelle). The cultivars were
morphologically characterized and their adaptation areas were located. The
characteristics of these varieties are indicated in Table 1. From the collected local
cultivars, only three of them are found to be morphologically distinct, the rest are
duplicates known by different names. ‘Ashe’, is a local variety widely grown in
East Harerghe and a highly demanded one. It is known by different names such as
[497]
Ashe, Shentemtella, Dedefa, and Tuluma and morphologically similar to ‘Nech
Abeba’, a widely grown variety in Shashamene area. Similar findings were also
reported in collections in Southern Ethiopia including Shashemene (Kolech et al.
2016). ‘Ashe’ is preferred by farmers for its early maturity, quick sprouting (for
multiple cropping), high tuber yield and good taste. Another widely grown local
variety in East Harerghe is ‘Jarso’. This cultivar is morphologically similar to
‘China’ or ‘Achire China’ in southern Ethiopia. It is characterized by early
maturity and relatively later sprouting than ‘Ashe’. This cultivar is the second
widely grown variety after ‘Ashe’. ‘Belina’, is another important cultivar in Kufa
Chelle district of East Harerghe which is known by its pink stem color.
Potato productivity in this part of the country is above the world potato average
(19.3 ton/ha) (CSA 2014) probably due to the productivity of the local varieties
and/or the production practices the farmers adopt to grow the crop. The planting
method is different from the practice known in other parts of the country. Farmers
in the area use a narrow spacing (50 cm between rows and 12-15 cm between
plants) as a contrary to 75cm *30cm plant spacing adopted in other parts of the
country for potato production. Since the area receives inadequate rainfall for the
crop, the farmers usually use water conservation measures to save the available
water.
Local varieties in West Ethiopia: Unlike East Ethiopia, Potato production in

west Ethiopia is dominated by improved varieties such as Menagesha and Gudene.
Especially ‘Gudene’ is widely grown in most potato producing areas in Ethiopia
due to their better taste, wide adaptation and long storability (Kolech et al. 2015).
Both varieties were released for production by Holetta Agricultural Research
Center. Although we collected several local cultivars, the share of them for
production in this part of Ethiopia is very low. Seven cultivars such as Chala, Aba
minminie, Yeferenji, England, Genet, Misre and Dalicho were collected from
West Shewa (Jeldu, Chelia) and East Wollega (Jima Arjo). Based on
morphological markers, most of these cultivars were found to be morphologically
distinct (Table 1). ‘Chala’, a cultivar collected from Jeldu, was morphologically
similar to Menagesha variety. ‘Genet’, a germplasm tested during the 1990s by
research centers for release but rejected by Variety Release Committee, was one
of the cultivars collected at Jima Arjo district, in East Wollega. The unique
characteristics of this variety are large rose pigmentation on the tubers, long
dormancy and early maturity.
According to the farmers’ perception in the area, all of these local cultivars except
Dalicho were grown rarely and believed to be older. ‘Dalicho’ is still grown by
some farmers at Chelia district. Some of these cultivar' names such as ‘Yeferenji’
were also recorded in northwest Ethiopia (Kolech et al. 2019). The planting time
in west Ethiopia is similar to that of northwest Ethiopia. The farmers in this part of
[498]
Ethiopia plant potatoes from March to April. Planting method at Oroguduru
Wollega, however, is different from other areas in Ethiopia; farmers plant potatoes
in raised heaps (separate hills). This needs further study so that such a planting
method is economically feasible to the farmers in the area.

Characterizing the available genetic resources in the country is an important first
step for crop improvement, conservation and other uses of the crop. Realizing this,
the potato cultivars in major potato growing areas of east and west Ethiopia were
collected and characterized using morphological markers. At the same time, data
on the production status and adaptation areas of these cultivars were recorded. In
east Harerghe, the local cultivars (mainly ‘Ashe’ and ‘Jarso’) are widely grown.
These cultivars are also known by different local names in the same administrative
zone and even in Shashemene and some areas in the country. These two cultivars
have specific characteristics such as early maturity and short dormancy which suit
well to the potato production system in the area. ‘Ashe’ is also known by ‘Nech
Abebe’, ShentemTella, Tuluma and Dedefa. In this part of the country, the
improved varieties are grown only by a few farmers.
In the contrary, improved varieties such as Gudene and Menagesha dominate

potato production in west Ethiopia (West Shewa, East Wollega and Horoguduru
Wollega zones). These varieties are characterized by their high yielding, late blight
resistant and long dormancy traits which fit well to the production system of this
part of Ethiopia. Hence, the impact of the effort of Holetta Agricultural Research
Center is highly prominent. Although six older and unique local cultivars are
recorded in west Ethiopia, they are rarely grown or grown by few farmers. But,
they have important characteristics such as good taste, long dormancy, a high
number of tubers and good for irrigation agriculture.
In general, at least nine unique cultivars are recorded in east and west Ethiopia.
These cultivars are now conserved at Amhara Agricultural Research Institute tissue
culture laboratory for future use. This characterization is, however, merely based
on morphological markers; it should be supported by molecular markers in order to
get reliable data. Together with the previous 15 unique collections (identified by
both morphological and molecular markers in previous studies in the northwest and
South Ethiopia), these are important assets for potato improvement as varieties
developed with the use of the existing local cultivars as parents are more likely to
be adapted to local growing conditions.
[499]
Table 1: Local names of Ethiopian local cultivars, along with known duplicate names, variety characteristics and areas of adaptation
Known duplicate names (where Collection

Cultivar name Adaptation Variety characteristics
grown) districts
Ashe ShentemTela, Dedefa, Tuluma Lay Kombolcha Widely grown in East Harerghe Flower color: white; Short flower stalk; Predominant tuber skin color: white; Tu
(East Harerghe) and Nech Abeba highlands, East Arsi shape: Round; medium deep eyes; Dormancy: Quickly sprouting; Maturity: ea
(Shashemene vareity) maturing (3 months); Other characteristics: many stems; quick cooked, the skin ea
peeled after cooked.
Jarso Key Abeba/Key shull/ Achire China Haramaya, East Harerghe highlands, South Flower color: pink; Predominant tuber skin color: white; Tuber shape: oblong, shal
(South Ethiopia) Kombolcha Ethiopia eyes; Maturity: early, maturing; tastier than Ashe cultivar; Dormancy: quick sprouting
Belina Kufa Chelle East Harerghe highlands Large leaves, blue and winged stems, round tubers,
Chala Menagesha (released variety) Jeldu and Dendi Widely grown in west Shewa and Red flower, purple pigment on the eyes, widely grown, tasty, loose apart when cooke
East Wollega
Aba minimine Chelia Rarely grown at Chelia, West White flower, late maturing, water-efficient, give good yield under irrigation, tasty, ma
Shewa but small tubers. Old variety
Yefereji Chelia Rarely grown at Chelia, West White flower, white tubers, Oval shaped tubers, medium taste
Shewa
Dalicho Jima Arjo Grown by some farmers in East White flower, late maturing, tasty, long stolon,
Wollega (2460 m.asl)
England Jima Arjo Rarely grown in Jima Arjo, East Red flowers, late maturing, oval tubers, good for irrigation.
Wollega
Misire Jima Arjo Rarely grown in Jima Arjo, East Red, round and small tubers; pink sprouts.
Wollega
Genet tested at research centers during Jima Arjo Rarely grown in Jima Arjo, East without flower, rose large pigmentation on the tubers, predominantly white tubers, o
1990s but not released Wollega tubers, long dormancy, early maturing
[500]
Acknowledgment
We are highly indebted to Adet Agricultural Research Center and Ethiopian
Institute of Agricultural Research/EIAR/ for financing this work. Special thanks go
to Mamaru Yeshambel and Bahir Dar tissue culture team members for variety
disease testing and screen house seed multiplication.
References
Barnaud A, Deu M, Garine E, Mckey D, Jolly HI. 2007. Local genetic diversity of sorghum in a village
in northern Cameroon: structure and dynamics of landraces. Theor appl genet. 114: 237-248.
Bradshaw JE, Bryan GJ and Ramsay G. 2006. Genetic resources (including wild and cultivated
Solanum species) and progress in their utilization in potato breeding. Potato Research, 49(1), 49–
65.
CSA (Central Statistics Agency). 2014. Agriculture sample survey 2013/14 (2006 EC): Report on area,
production and farm management practice of Belg season crops for private peasant holdings.
Volume V: Statistical Bulletin 532, Addis Ababa, Ethiopia.
Huaman Z, Williams T, Salhuana W and Vincent L. 1977. Descriptors for the cultivated potato.
International Board for Plant Genetic Resources. FAO.
Kidane Mariam H. 1979. Preliminary assessment of the responses of potato genotypes in the eastern,
southern and central regions of Ethiopia. Eth. J. Agr. Sci., 1(1): 41-47.
Kolech SA, Halseth D, De Jong W. Perry K, Wolfe D, Tiruneh FM and Schulz S. 2015. Potato Variety
Diversity, Determinants and Implications for Potato Breeding Strategy in Ethiopia.
Kolech SA, Halseth D, Perry K, Wolfe D, Douches DS, Coombs J. De and Jong W. 2016. Genetic
diversity and relationship of Ethiopian potato cultivars to germplasm from the USA, Europe and
International Potato Center. American Journal of Potato Research. 93(6): 609–619
Kolech SA, De Jong W, Halseth D and Schulz S. 2019. Understanding farmer needs and unlocking
local genetic resources for potato improvement: a case study in Ethiopia. Afri. J. Food Agric. Nutr.
Dev. 19 (1): 13883-13905. DOI: 10.18697/ajfand.84.BLFB1012.
Labarta R, Wambungu S, Yirga C, Mugabo J, Nsabimana J, Mulwa C, Schulz S, Lorochelle C, Alwang
J, Andrad R, and Yigezu Y. 2012. Report on Objective 2 of the Diffusion and Impacts of Improved
Varieties in Africa (DIIVA) Project. International Potato Center (CIP), Nairobi, Kenya.
Lemaga B. 1983. Country program report on Ethiopian potato research activities. Presented at the
Regional Workshop on Potato Development and Transfer of Technologies in Tropical Africa.
Addis Ababa, Ethiopia, 22 August to 1 September 1982..
Padgham J. 2009. Agricultural development under a changing climate: opportunities and challenges for
adaptation. Joint department discussion paper –Issue 1. World Bank.
Williams CN, Uzo JO and Peregrine WTH. 1991. Vegetable production in the tropics. Intermediate
Tropical Agriculture Series. Longman Scientific and Technical, Essex.
[501]
[502]
Root and Tuber Crops
Performance and Stability Study of Newly Bred

Orange Fleshed Sweet potato Genotypes
in Ethiopia
Fekadu Gurmu1*, Bililign Mekonen1, Yitages Kuma2,
Shiferaw Mekonen1 and Zeritu Shashego1
1
South Agricultural Research Institute, Hawassa Research Centre, P.O.Box 6, Hawassa, Ethiopia
2
Ethiopian Institute of Agricultural Research, Werer Research Center, P.O.Box 2003, Addis Ababa, Ethiopia
*
Corresponding author: fekadugb@gmail.com
Abstract
Sweetpotato (Ipomoea batatas L. Lam) is among very valuable root crops in Ethiopia.
Especially the orange-fleshed sweet potatoes (OFSP) are rich in β-carotene and they
are good sources of vitamin A. However, in Ethiopia, the acceptance of the OFSP
varieties by farmers was low due to their low root dry matter content (RDMC).
Therefore, a crossing block was established in 2013 to cross selected OFSP varieties
with white-fleshed varieties in order to improve the RDMC of the OFSP varieties and
thereby increase the adoption of the varieties. Nursery and preliminary yield trials
were conducted in 2014 and 2015, and finally, national variety trials were conducted
at four locations for two years (2016 and 2017) using 12 selected genotypes. The
verification trial was conducted at four locations for one season. The analysis of
variance revealed that there was a significant difference (p < 0.01) among the varieties
for total root yield and RDMC. Two genotypes, Ukr/Eju-10 and Ukr/Eju-13 have been
identified as best candidates for release based on their fresh root yield of 28.7 and 26.8
t ha-1, RDMC of 31.9 and 31.4%, resistance to sweet potato virus disease with low
score of 1.4 and 1.6, respectively. The flesh color of the two varieties is deep orange
with high β-carotene contents of 12.48 and 9.51 mg 100 g -1, respectively as compared
to the standard check (Kulfo), which had 6.91 mg 100 g-1. The two newly bred
candidate varieties were ranked 1st and 2nd based on farmers’ preference taste-test.
Therefore, the two candidate varieties can be the potential OFSP varieties that can
replace Kulfo, which is currently under production in Ethiopia.
Keywords: β-carotene, candidate varieties, dry matter, root yield, farmers’

preference test.
Introduction
Sweetpotato (Ipomoea batatas L. Lam) is among very valuable root crops in
Ethiopia and it is widely grown in the eastern, southern and southwestern parts of
the country (Belehu, 2003, Tofu et al., 2007, Gurmu et al., 2015a). The roots are
used as a staple food while the vines and leaves are used as animal feed. The
orange-fleshed sweet potato (OFSP) varieties are rich in β-carotene (pro-vitamin
A). These varieties are playing a crucial role in preventing vitamin A deficiency in
Africa which otherwise leads to blindness and maternal and pre-school children
[503]
mortality in many of the developing countries (WHO, 2009; Wang et al., 2011;
Gurmu et al., 2015b).
The orange-fleshed sweet potato (OFSP) varieties are the cheapest staple food,
easily accessible and year-round sources of vitamin A (Hagenimana et al. 2001;
Mwanga et al. 2003; Kapinga et al., 2005; Burri 2011; Gurmu et al. 2015a;
Gurmu et al., 2017). Most of the orange-fleshed sweet potato varieties contain
3000-16000 μg 100 g-1 of β-carotene and this contributes to 250-1300 μg 100 g-1
retinol activity equivalents (RAE) (Kapinga et al., 2010). Therefore, according to
Kapinga et al. (2005), the addition of 100 g OFSP in a daily diet can prevent VAD
in children and significantly reduce the death of mothers. Christina (2007) also
indicated that a 100 g OFSP can provide more than 6500 µg β-carotene, which is
equivalent to more than 1000 µg RAE. Therefore, 100 g of OFSP day-1 exceeds
the RDA to prevent VAD. As a result, depending upon the color intensity of the
OFSP variety used, and taking into account losses of β-carotene during cooking,
which accounts for an approximately 20% loss through boiling (van Jaarsveld et
al., 2006), a quarter of a cup to one cup of boiled and mashed sweet potato meets
the RDA of vitamin A of a young child (Hagenimana et al., 2001; Fleshman et al.,
2011; Gurmu et al., 2015a). OFSP is also a good source of energy, which is
equivalent to 293-460 kJ 100 g-1 (Woolfe, 1992; Hagenimana et al., 2001).
However, in Ethiopia, the white-fleshed sweet potato varieties are commonly
grown by farmers and farmers prefer them due to their high root dry matter
content as compared to the OFSP varieties. Therefore, improving the dry matter
contents of the OFSP and evaluating their performance with farmers’ participation
is very vital to increase their acceptance by farmers. With this rationale, the
current study was conducted to evaluate the newly bred OFSP varieties for root
yield, root dry matter and β-carotene contents and to identify the best and farmer
preferred candidate varieties for release.

Study Areas: A crossing block was established at the South Agricultural Research
Institute in Hawassa city, Ethiopia. Hawassa is located at 7o 03’ 54” N and 38o 28’
59” E with an altitude of 1750 meters above sea level. It is found at 275 km south
of the Ethiopian capital, Addis Ababa. The field evaluation was conducted for two
years (2016-2017) at four locations (Dilla, Halaba and Hawassa) in the Southern
Nation, Nationalities and Peoples’ Regional State and Werer in Afar Regional
State in Ethiopia. The details of the test sites are summarized in Table 1.
Plant Materials: Twelve orange-fleshed sweet potato genotypes selected from
diallel cross families were evaluated along with one released OFSP check variety
(Kulfo). The list of the genotypes is presented in Table 2. The genotypes were
selected based on their high RDMC, orange flesh color (β-carotene content) and
fresh root yield.
[504]
Experimental Design and Field Layout: The national variety trial was
conducted using a randomized complete block design with three replications. The
experimental plots consisted of a four-row plot of three meters long for each
genotype. The recommended inter-row spacing of 0.60 and intra-row spacing of
0.3 meter that resulted in a total of 10 plants per row and 40 plants per plot was
used for the field experiments. The experiments were planted in July to August
and harvested from November to December each year. Only plants in the two
central rows were considered for data recording and harvesting. The experiment
was conducted under rainfed conditions and did not receive irrigation. No
fertilizers were applied at any of the test sites and weeding and cultivation were
done as recommended for the crop in Ethiopia (ARC 2015).
Table 1. Description of the experimental sites
Annual Mean annual

Altitude rainfall temperature (oC) RH
Location Code (masl)* Coordinates (mm) Min Max (%)
Dilla DIL 1519 06o22’49’’N, 38o18’25’’E 1354.6 12.9 28.1 65.0
Halaba HAL 1772 07o18’38’’N, 38o05’38’E 928.8 14.6 28.6 58.3
Hawassa HAW 1700 07o03’54’’N, 38o28’59’’E 1046.3 13.3 27.6 62.1
Werer WER 750 09o16’55’’N, 40o09’23”E 590.0 26.7 40.8 -
*masl = meter above sea level, RH = relative humidity.
Source: National Meteorological Agency, Hawassa Main Branch.
Table 2. Description of sweet potato genotypes used for the study
No Genotypes Short names Origin

1 Ukrewe x Ejumula-10 Ukr/Eju-10 Ethiopia
2 Ukrewe x Ejumula-13 Ukr/Eju-13 Ethiopia
3 Resisto x Temesgen-14 Res/Tem-14 Ethiopia
4 Resisto x Temesgen-23 Res/Tem-23 Ethiopia
5 Jewel Jewel CIP-Kenya
6 Carrot Dar Carrot Dar CIP-Kenya
7 Resisto x PIPI-21 Res/Pip-21 Ethiopia
8 SPK004/6/6 SPK004/6/6 Uganda
9 Maputha-1 Maputha-1 CIP-Kenya
10 SPK004/6 SPK004/6 Uganda
11 Tainung-15 Tainung-15 CIP-Kenya
12 Kulfo Kulfo Ethiopia
Data Collection: Fresh root yield was the major traits followed by root dry matter
and β-carotene contents. All the data samples were taken from each replicate at
each location. Fresh root yield (t ha-1) was measured from two central rows and
expressed as harvested fresh root weight in kg plot-1 and later converted to tonnes
per hectare. RDMC estimated as a percentage of root dry weight (g) to fresh root
weight (g) where 100- 200 g samples were taken from roots of sampled plants in
the plot and the samples were dried in an oven at 80oC for 48 hours. β-carotene
content was analysed at the International Potato Center (CIP) in Uganda using
Near Infrared Reflectance Spectrometry (NIRS) and expressed as mg 100 g-1 on
dry weight basis.
[505]
Data Analysis
Analysis of variance: A combined analysis of fresh root yield, RDMC and β-
carotene content data across the environments were analysed using GenStat 14th
edition (Payne et al. 2011) and SAS version 9.3 (SAS Institute Inc. 2003)
statistical packages. The following statistical model was used for the combined
analysis of variance over environments:
Y ijkl    G i  L j
 Y k  GL ij
 GY ik
 LY jk
 GLY ijk
 B l ( j )   ijkl
Where: Yijkl is observed value of genotype i in block l and of location j and year k,
μ is grand mean, Gi is effect of genotype i, Lj is location effect, Yk is year effect,
GLYijk is the interaction effect of genotype i with location j and year k, Bl(j) is the
effect of block l in location j, єijkl is error (residual) effect of genotype i in block l
of location j and year k.
G x E and stability analysis: G x E and stability analyses were conducted using a

GGE bi-plot procedure (Yan 2001; Yan 2002; Yan et al., 2007) in GenStat 14th
edition.
The model for a GGE bi-plot (Yan, 2002; Yan et al., 2007) based on singular
value decomposition (SVD) of t principal components is:
t
Y ij   i   j
  k 1
 k  ik  jk
  ij
Where: Y ij is the performance of genotypei in environment j,  is the grand mean,

 j is the main effect of environment j, k is the number of principal components
(PC);  k is singular value of the kth PC; and  ik and  jk are the scores of ith
genotype and jth environment, respectively for PCk;  ij is the residual associated
with genotype i in environment j.

Mean Performance of the Genotypes for Fresh Root Yield: The combined
analysis of variance showed that there was a highly significant (p < 0.001)
difference among the four test environments and the 12 sweet potato genotypes for
fresh root yield (Table 3). The result also showed that genotype x environment
interaction (GEI) was highly significant (p < 0.001) for this trait. The treatment
effects had accounted for the large portion of the total sum of squares, which was
29.3%, followed by location (22.3%) and Treatment x location interaction
(13.4%). Year effect accounted only for a very small portion of the total sum of
[506]
squares (0.4%). The three-way interaction effects accounted for 11.2%. Although
the higher portion of the total sum of squares was accounted for by the genotypic
effects, however, the analysis showed the presence of significant G x E interaction
effects which necessitated stability analysis.
The mean fresh root yield of the genotypes over two years at the four locations is
presented in Table 4. Dilla was the best environment for fresh root yield during
2016 with a mean yield of 24.8 t ha-1 while in 2017 Werer was the best with a
mean of 28.3 t ha-1. Halaba was the least yielding environment during both years.
Genotypes Eju/Ukr-10 and Eju/Ukr-13 were the highest yielding genotypes across
the locations and years with a mean of 28.7 and 26.8 t.ha-1, respectively. The two
genotypes showed a yield advantage of 168.9 and 150.9%, respectively over the
standard check (Kulfo). Res/Tem-23 and Kulfo both gave the lowest yield with a
mean of 10.7 t ha-1.
The presence of a highly significant difference between the four test locations and
the two years indicates that the genotypes performed differently across locations
and over years. The significant genotype by environment interaction also indicates
the significance of environmental effects (locations and years) in the performance
of the sweet potato genotypes for fresh root yield. Similar findings were reported
by various authors, who indicated that sweet potato genotypes are sensitive to
environmental variations (Wolfgang et al., 2005; Mbwaga et al., 2008; Chiona,
2009; Osiru et al., 2009; Moussa et al., 2011).
Table 3. Analysis of variance of three traits of sweet potato genotypes evaluated across six environments in 2016 and
2017
Sum of squares Mean squares %SS

Source of variation df (SS) (MS)
Treatment (Trt) 11 8928.2 811.7** 29.3
Block (B) 2 23.6 11.8ns 0.1
Location (Loc) 3 6813.2 2271.1*** 22.3
Year 1 132.0 132.0* 0.4
Trt x Loc 33 4087.8 123.9* 13.4
Trt x Year 11 1447.5 131.6* 4.7
Loc x Year 3 1890.2 630.1** 6.2
Trt x Loc x Year 33 3427.9 103.9* 11.2
Error 190 3769.7 19.8 12.4
Total 287
*,** & *** significant at p < 0.05, p < 0.01 & p < 0.001 probability levels, respectively; ns = not significant, df = degrees of
freedom, trt = treatment, loc = location
[507]
Table 4. Total storage root performance of sweetpotato clones tested across locations and over years (ton/ha)
2016 2017
Overall Yield advantage over
Genotype Dilla Halaba Hawassa Werer Mean Dilla Halaba Hawassa Werer Mean mean standard check (%)
Ukr/Eju-10 31.8 17.7 25.1 37.1 27.9 27.6 17.1 27.8 45.5 29.5 28.7 168.9
Ukr/Eju-13 30.3 22.7 25.8 27.4 26.6 24.8 16.5 27.9 38.8 27.0 26.8 150.9
Res/Tem-14 31.0 13.5 22.5 29.5 24.1 20.3 10.5 13.9 23.6 17.1 20.6 92.8
Res/Tem-23 9.3 13.4 18.2 7.1 12.0 13.2 6.5 7.6 10.3 9.4 10.7 0.2
Jewel 20.1 17.9 13.9 27.1 19.8 25.2 6.8 21.5 27.4 20.2 20.0 87.1
Carrot Dar 27.1 17.7 14.8 28.3 22.0 17.5 6.3 22.4 22.7 17.2 19.6 83.5
Res/Pip-21 32.2 24.2 19.1 16.3 23.0 21.2 12.7 17.3 38.9 22.5 22.7 113.1
SPK004/6/6 31.4 15.1 20.5 10.2 19.3 21.6 12.3 20.7 31.6 21.6 20.4 91.4
Maputha-1 21.5 9.0 13.4 6.6 12.6 23.4 11.4 23.3 21.8 20.0 16.3 52.7
SPK 004/6 26.4 7.4 5.4 11.5 12.7 19.1 14.2 15.3 31.9 20.1 16.4 53.4
Tainung-15 20.0 5.6 5.4 26.6 14.4 19.5 11.7 15.7 32.2 19.8 17.1 60.0
Kulfo 16.8 2.3 9.2 2.1 7.6 18.3 10.4 11.2 15.1 13.7 10.7 -
Mean 24.8 13.9 16.1 19.1 18.5 21.0 11.4 18.7 28.3 19.8 19.2
LSD (5%) 8.7 8.6 10.7 12.2 5.3 7.1 6.6 8.7 11.3 5.8 10.3
CV 20.6 33.5 34.3 35.8 33.5 25.2 29.4 31.5 33.2 30.1 33.4
[508]
Mean Performance of the Genotypes for Root Dry Matter and Β-Carotene
Contents: The mean root dry matter and β-carotene content of the genotypes is
presented in Table 5. The mean root dry matter content of the genotypes ranged
from 22.9% to 31.9%. The highest mean root dry matter content was recorded
from three genotypes: Eju/Ukr-10 (31.9%), Eju/Ukr-13 (31.4%) and Res/Tem-14
(31.5%). The lowest mean root dry matter content (22.9%) was recorded from the
check variety, Kulfo. β-carotene content of the genotypes ranged from 2.4 to 12.4
mg 100 g-1. Four genotypes showed a mean β-carotene content that is greater than
10 mg 100 g-1. The highest β-carotene content (12.4 mg 100 g-1) was recorded
from a genotype Ukr/Eju-10 (Table 5).
Table 5. Mean dry matter and beta-carotene contents of sweet potato clones evaluated over the years
Dry matter content (%)

Beta-carotene content
S. No Genotypes 2016 2017 Mean (mg 100 g-1) (2016)
1 Ukr/Eju-10 32.1 31.8 31.9 12.4
2 Ukr/Eju-13 31.3 31.6 31.4 9.5
3 Res/Tem-14 31.8 31.3 31.5 8.5
4 Res/Tem-23 25.3 24.2 24.8 7.3
5 Jewel 28.2 30.4 29.3 11.3
6 Carrot Dar 28.6 31.8 30.2 10.5
7 Res/Pip-21 29.0 31.6 30.3 2.4
8 SPK004/6/6 30.2 30.3 30.2 8.5
9 Maputha-1 28.5 27.8 28.2 8.8
10 SPK 004/6 28.6 29.6 29.1 9.6
11 Tainung-15 28.6 27.6 28.1 10.3
12 Kulfo 23.3 22.5 22.9 8.3
Mean 28.7 29.2 29.0 9.0
LSD (5%) 2.1 2.4 2.2 2.5
CV 10.2 8.3 8.5 6.3
Stability Analysis: Figure 1 displays the GGE-biplot analysis of fresh root yield.
The genotypes 1 (Ukr/Eju-10) and 2 (Ukr/Eju-13) had relatively higher PC1
values indicating that these genotypes had high average fresh root yield. On the
other hand, the genotypes 4 (Res/Tem-23) and 12 (Kulfo) had the lowest PC1
values which indicates that they had low fresh root yields. Among the test
genotypes, only four of them showed a fresh root yield that was below average
(Fig. 1). Among the test environments, Werer had a relatively higher PC1 score
showing that this environment was high yielding environment during both years as
compared to the other environments (Fig. 1).
The stability of the genotypes across the test environments is displayed in Figure
2. Genotypes 1 (Ukr/Eju-10) and 2 (Ukr/Eju-13) had a shorter absolute length of
projection on the Y-axis showing that they stable genotypes. On the other hand,
genotypes with a longer absolute length of projection on the Y-axis such as 8
(spk004/6/6), 7(Res/Pip-21), 10 (spk004/6), 6 (Carrot Dar), 3 (Res/Tem-14) and 5
(Jewel) were less stable genotypes.
[509]
According to Yan (2001), large positive PC1 scores for genotypes indicate that
those genotypes had relatively higher mean values. The average environment
coordinate (AEC) is the line that passes through the bi-plot origin and separates
genotypes with mean values below average and above average (Yan et al., 2000,
Yan, 2001). Those genotypes to the right of this line are high yielders, while those
to the left are low yielders (Yan, 2001; Yan and Kang, 2003; Gurmu et al., 2012).
Therefore, eight genotypes had high fresh root yields that were above average and
two of them (Ukr/Eju-10 and Ukr/Eju-13) were best yielders.
According to Yan (2001), genotypes with PC2 scores near zero are the most stable
genotypes. Accordingly, two genotypes, Ukr/Eju-10 and Ukr/Eju-13, which had
PC2 scores near zero were found the most stable genotypes among the tested
genotypes. These genotypes also had the highest mean fresh root yields.
4.00
2.00 8
10 7
WER17
9
HAL17 DIL16
12
PC2 - 13.84%
DIL17HAW17 2
0.00 1
HAL16
HAW16
11
4 5
3
-2.00 6
WER16
-4.00
-2.00 0.00 2.00 4.00
PC1 - 63.21%
Figure 1. GGE-biplot showing eight environments (four locations x two years) and 12 orange-fleshed sweetpotato
genotypes evaluated for fresh root yield.
DIL 16= Dilla 2016; DIL 17 = Dilla 2017; HAL 16 = Halaba 2016; HAL 17 = Halaba 2017; HAW 16 = Hawassa 2016; HAW
17 = Hawassa 2017; WER 16 = Werer 2016; WER 17 = Werer 2017
[510]
8
10 7
WER17
9
HAL17 DIL16
12
PC2 - 13.84%
DIL17HAW17 2
HAL16
HAW16 1
11
4 5
3
6
WER16
PC1 - 63.21%
Figure 2. Mean performance of 12 orange-fleshed sweet potato genotypes and their stability across eight environments
for fresh root yield.
DIL 16= Dilla 2016; DIL 17 = Dilla 2017; HAL 16 = Halaba 2016; HAL 17 = Halaba 2017; HAW 16 = Hawassa 2016; HAW
17 = Hawassa 2017; WER 16 = Werer 2016; WER 17 = Werer 2017;
Farmers’ Preference Test: The mean farmers’ preference for orange-fleshed

sweet potato genotypes evaluated across locations is presented in Table 6. The
criteria for selecting the genotypes were the root flesh color, sugariness of the
boiled root, test, powderiness, texture and overall acceptance of the boiled roots.
More than 40 farmers had participated in the evaluation of the genotypes. The
panelist farmers were given cards of different colors: green for best, yellow for
medium and red for bad and they ranked the genotypes based on the above
criteria. Accordingly, the genotypes (Ukr/Eju-10) and (Ukr/Eju-13) were ranked
1st and 2nd by the farmers. The OFSP variety (Kulfo) that is currently under
production in Ethiopia and used as a check variety in this study was ranked 8 th.
The result demonstrated that the two candidate varieties met the farmers’
preference in terms of the above criteria which is very important criteria for the
acceptability of OFSP varieties.
[511]
Table 6. Mean farmers’ preference for sweetpotato clones evaluated across locations
Parameters
Powederin Overall Rank
S. No Genotypes Color Sugariness Test ess Texture acceptance Average
1 Ukr/Eju-10 4.7 4.0 4.7 4.3 4.6 4.7 4.5 1
2 Ukr/Eju-13 3.1 3.4 3.1 3.4 3.3 3.4 3.3 2
3 Res/Tem-14 2.7 2.5 2.7 2.6 3.3 2.5 2.7 4
4 Res/Tem-23 1.9 1.9 1.7 1.9 2.3 1.8 1.9 10
5 Jewel 1.6 1.9 1.8 1.6 2.2 1.9 1.8 11
6 Carrot Dar 1.7 1.7 1.7 1.6 2.3 1.9 1.8 12
7 Res/Pip-21 2.5 2.7 2.6 2.8 2.7 2.7 2.7 5
8 SPK004/6/6 2.6 2.7 2.7 3.0 2.7 2.7 2.7 6
9 Maputha-1 2.5 2.3 2.8 2.6 2.2 2.7 2.5 7
10 SPK 004/6 2.3 2.2 2.6 2.5 2.2 2.5 2.4 9
11 Tainung-15 3.0 2.7 2.9 2.9 2.7 2.7 2.8 3
12 Kulfo 2.2 2.6 2.4 2.5 2.6 2.7 2.5 8
Mean 2.6 2.5 2.6 2.6 2.8 2.7 2.6
[512]
Conclusion
The root dry matter content of OFSP varieties greatly influences the acceptance
and adoption of varieties by farmers. Among the 12 OFSP genotypes evaluated
across four locations and over two years, two candidate genotypes were selected
based on their better root yield, dry matter content and β-carotene content. These
genotypes had wider adaptability and stability across environments. More
importantly, the genotypes were also selected by farmers as best ones and ranked
1st and 2nd among the tested genotypes. Therefore, the genotypes were presented
for the national variety verification for release. The genotypes are potential
candidates that could replace Kulfo variety, an OFSP variety that is currently
grown in Ethiopia.
Acknowledgment
The South Agricultural Research Institute (SARI), Ethiopian Institute of
Agricultural Research (EIAR) and International Potato Center (CIP) are duly
acknowledged for their financial support.
Reference
ARC. 2015. Sweetpotato production and field management in Ethiopia, Production Manual.
Awassa Agricultural Research Center, Hawassa, Ethiopia, 2015.
Belehu T. 2003. Agronomical and Physiological Factors Affecting Growth, Development and
Yield of Sweetpotato in Ethiopia. Ph.D. Thesis, University of Pretoria, Pretoria.
Burri BJ. 2011. Evaluating sweet potato as an intervention food to prevent vitamin A deficiency.
Comprehensive Reviews in Food Science and Food Safety. 10, 118-130.
Chiona M. 2009. Towards Enhancement of β-carotene Content of High Dry Mass Sweetpotato
Genotypes in Zambia. Ph.D. Thesis, University of KwaZulu-Natal, Pietermaritzburg.
Christina SL. 2007. Nutrient and Sensory Quality of Orange Fleshed Sweetpotato. MSc Thesis,
University of Pretoria, Pretoria.
Gurmu F, Lire E, Asfaw A, Alemayehu F, Rezene Y, Ambachew D. 2012. GGE-biplot analysis of
grain yield of faba bean genotypes in southern Ethiopia. Electronic Journal of Plant Breeding.
3(3), 898-907.
Gurmu F, Hussein S, Laing M. 2015a. The potential of orange-fleshed sweet potato to prevent
vitamin A deficiency in Africa. International Journal for Vitamin and Nutrition Research.
84(1-2), 65-78.
Gurmu F, Hussein S, Laing M. 2015b. Diagnostic assessment of sweet potato production in
Ethiopia : constraints, post-harvest handling and farmers’ preferences. Research on Crops.
16(1), 104-115.
Hagenimana V, Low J, Anyango M, Kurz K, Gichuki ST, Kabira J. 2001. Enhancing vitamin A
intake in young children in Western Kenya: orange-fleshed sweet potatoes and women farmers
can serve as key entry points. Food and Nutrition Bulletin. 22, 370-387.
Kapinga R, Anderson P, Crissman C, Zhang D, Lemaga B, Opio F. 2005. Vitamin A partnership
for Africa: a food-based approach to combat vitamin A deficiency in sub-Saharan Africa
through increased utilization of orange-fleshed sweet potato. Chronica Horticulturae. 45(3),
12-14.
[513]
Kapinga R, Tumwegamire S, Ndunguru J, Andrade MI, Agili S, Mwanga ROM, Laurie S, Dapaah
H. 2010. Catalogue of Orange Fleshed Sweetpotato Varieties for Sub-Saharan Africa.
International Potato Center (CIP), Lima, Peru.
Mbwaga, Z, Mataa M, Msabaha M. 2008. Quality and yield of orange-fleshed sweet potato
(Ipomoea batatas) varieties grown in different agro-ecologies. In: Kasem, Z.A., editors. The
8th African Crop Science Society (ACSS) Conference Proceeding, El-Minia, Egypt. 27-31
October, 2007. African Crop Science Society, pp. 339-345.
Moussa SAM, Hala AAE, Nashwa IAE. 2011. Stability study of sweet potato yield and its
component characters under different environments by joint regression analysis. Journal of
Horticultural Science and Ornamental Plants. 3, 43-54.
Mwanga ROM, Odongo B, Niringiye C, Zhang D, Yencho GC, Kapinga R. 2003. Orange fleshed
sweet potato breeding activities in Uganda. In: The 6th Conference of the African Crop
Science Society (ACSS) Conference Proceeding, 12-17 October, 2003, Nairobi. African Crop
Science Society, Kampala, Uganda.
Osiru MO, Olanya OM, Adipala E, Kapinga R, Lemaga B. 2009. Yield stability analysis of
Ipomoea batatas L. cultivars in diverse environments. Australian Journal of Crop Science. 3,
213-220.
Payne RW, Murray DA, Harding SA, Baird DB, Soutar DM. 2011. GenStat for Windows (14th
Edition) Introduction. VSN International, Hemel Hempstead, UK.
SAS Institute Inc. 2003. Version 9.1. SAS Institute Inc., Cary, NC.
van Jaarsveld PJ, Wet De, Harmse M, Nestel E, P Rodriguez-Amaya DB. 2006. Retention of β-
carotene in boiled, mashed orange-fleshed sweet potato. Journal of Food Composition and
Analysis. 19(4), 321-329.
Wolfgang JG, Manrique K, Zhang D, Hermann M. 2005. Genotype x environment interactions for
a diverse set of sweet potato clones evaluated across varying ecogeographic conditions in
Peru. Crop Science. 45, 2160-2171.
Woolfe JA. 1992. Sweetpotato: An Untapped Food Resource. Cambridge University Press,
Cambridge, United Kingdom.
Yan W. 2001. GGE Biplot: A Windows application for graphical analysis of multi-environment
trial data and other types of two-way data. Agronomy Journal. 93, 1111-1118.
Yan W. 2002. Singular-value partitioning in biplot analysis of multi-environment trial data.
Agronomy Journal. 94, 990-996.
Yan W, Cornelius PL, Crossa J, Huntt LA. 2001. Two types of GGE biplots for analyzing multi-
environment trial data. Crop Science. 41, 656-663.
Yan W and Kang MS. 2003. GGE Biplot Analysis: A Graphical Tool for Breeders, Geneticists,
and Agronomists. CRC Press, Boca Raton, Florida.
Yan W, Kang MS, Ma B, Woods S, Cornelius PL. 2007. GGE Biplot vs. AMMI analysis of
genotype-by-environment data. Crop Science. 47, 643-655.
[514]
Verification of Introduced Early Maturing
Orange Fleshed Sweetpotato Varieties for
Registration in Ethiopia
Fekadu Gurmu1*, Bililign Mekonen1, Yitages Kuma2,
Shiferaw Mekonen1 and Zeritu Shashego1
1
South Agricultural Research Institute, Hawassa Research Centre, P.O.Box 6, Hawassa, Ethiopia; 2Ethiopian
Institute of Agricultural Research, Werer Research Center, P.O.Box 2003, Addis Ababa, Ethiopia;
*Corresponding author: fekadugb@gmail.com
Abstract
Sweetpotato is one of the drought tolerant crops that are widely grown in moisture
stressed areas. In addition, some sweet potato varieties can mature early and can
escape the effect of prolonged drought. In Ethiopia, sweet potato varieties are
grouped into three groups. These are early maturing (< 120 days), medium maturing
(121-150 days) and late-maturing (>150 days). The objective of the study was to
identify high yielding and early maturing orange-fleshed sweet potato varieties
among introduced varieties. Four early maturing varieties have been introduced
from Uganda and evaluated for one season. Then the verification trial was
conducted across four locations for one year. Based on the result of the analysis of
variance, there was a highly significant difference among the tested varieties for root
yield (p < 0.01). However, there was no significant difference among the tested
varieties in terms of days to maturity. Nevertheless, the candidate varieties, Vita and
Kabode, showed better performance than the check variety Kulfo in terms of root
yield, root dry matter content (RDMC), resistance to sweet potato virus and flesh
colour. The two varieties had fresh root yield of 18.3 and 21.1 t ha-1, RDMC of 28.7
and 30.6%, resistance to sweet potato virus disease with low score of 1.2 and 1.1,
respectively. The two varieties had also intermediate orange flesh colour (better
amount of vitamin A) as compared to the light orange flesh colour of Kulfo variety.
Farmers also ranked the two varieties, Vita and Kabode as 2 nd and 1st, respectively.
Therefore, the two candidate varieties can replace Kulfo, which is currently under
production in Ethiopia.
Keywords: β-carotene, candidate varieties, dry matter, earliness, root yield,

farmers’ preference test.
Introduction
Food insecurity is a major problem in developing countries including Ethiopia.
Cereals are the main food sources in most places in Ethiopia which leads to a
nutrition imbalance on top of food insecurity. Micro-nutrient deficiency is also the
other global health problem, especially in low-income countries of the world
(Welch, 2002; Knez and Graham, 2013). Deficiency of micro-nutrients such as
vitamin A and minerals, especially iron (Fe) and zinc (Zn), affects nearly two
[515]
billion people worldwide (Allen et al., 2006; Tulchinsky, 2010). The deficiencies
increase susceptibility to various diseases. Pregnant and lactating mothers, and
young children, are greatly affected by nutrient deficiencies since they need
relatively high levels of vitamins and minerals (Nabakwe and Ngare, 2004; WHO,
2009b).
Most communities in Ethiopia are deficient in micro-nutrients such as vitamin A,

iron and zinc. Vitamin A (VAD) leads to blindness, retarded growth and death. It
is estimated that some 3 million children in SSA under the age of 5 years suffer
from partial or total blindness as a result of VAD. The deficiency is reaching
alarmingly high proportions in some regions of the country (Demissie et al., 2009;
WHO, 2009). A very high prevalence rate of clinical VAD was reported in some
parts of the country (Demissie et al., 2009).
Different strategies have been used to control VAD. The strategies include
Vitamin A supplementation of large doses in the form of capsules, fortification of
commonly consumed food items such as oil, sugar, breakfast cereals and grain
flour, dietary diversification, which includes eating food items naturally rich in
pro-vitamin A such as yellow/orange root crops, leafy vegetables and
yellow/orange fruits. The last strategy is bio-fortification in which crop varieties
that have high levels of pro-vitamin A so that everybody growing them could have
access for this vitamin through consuming them. Orange fleshed sweet potato
varieties are among these crops that can help curb deficiency related to vitamin A.
Orange fleshed sweet potatoes (OFSPs) are relatively cheap and easily accessible
sources of vitamin A. They can be used to combat the problem of VAD affecting
the poor community of most sub-Saharan African countries, including Ethiopia.
During the past research years; sweet potato variety development program of the
country has released some varieties that have varying level of pro-vitamin A
contents. However, such kind of efforts are continuous as better levels are always
searched through breeding works.
In general, sweet potato is considered as one of the drought tolerant crops that are
widely grown in moisture stressed areas. One of the mechanisms for its tolerance
is early maturing and escaping from the effect of terminal drought. Accordingly,
in Ethiopia, sweet potato varieties are grouped into three categories. These are
early maturing (< 120 days), medium maturing (121-150 days) and late-maturing
(>150 days). The objective of the study was to identify high yielding and early
maturing orange-fleshed sweet potato varieties. Therefore, there is an urgent need
to replace the existing OFSP varieties with new high yielding and early maturing
varieties that have good dry matter and β-carotene contents.
[516]
Study Areas: The verification trial was conducted for one year (2018) at four
locations (Dilla, Halaba and Hawassa in the Southern Nation, Nationalities and
Peoples’ Regional State and Werer in Afar Regional State in Ethiopia). The details
of the test sites are summarized in Table 1.
Plant Materials: Two orange-fleshed sweet potato genotypes that were

introduced from Uganda were evaluated in the verification trial along with the
check variety (Kulfo). The list of the varieties is presented in Table 2. The
varieties were selected based on their high RDMC, orange flesh color (β-carotene
content) and fresh root yield.
Table 1. Description of the experimental sites
Altitude Annual Mean T (oC) RH

Location Code (masl)* Coordinates RF(mm) Min Max (%)
Dilla DIL 1519 06o22’49’’N, 38o18’25’’E 1354.6 12.9 28.1 65.0
Halaba HAL 1772 07o18’38’’N, 38o05’38’E 928.8 14.6 28.6 58.3
Hawassa HAW 1700 07o03’54’’N, 38o28’59’’E 1046.3 13.3 27.6 62.1
Werer WER 590.0 26.7 40.8 -
750 09o16’55’’N, 40o09’23”E
*masl = meter above sea level, RH = relative humidity.
Source: National Meteorological Agency, Hawassa Main Branch.
Table 2. Description of sweet potato genotypes used for the study
No Varieties Other names Origin

1 Vita NASPOT 9 O Uganda
2 Kabode NASPOT 10 O Uganda
3 Kulfo Kulfo Ethiopia
Experimental Design and Field Layout: The national variety verification trial
was conducted using a 10m x 10m plot for each variety and replicated over
locations. At each location, there was one on-station trial and two on-farm trials.
The spacing between each row and plants was 0.60 and 0.3 meter, respectively
which resulted in a total of 555 plants per plot. The experiments were planted in
July to August and harvested in November to December. The experiment was
conducted under rainfed conditions and did not receive irrigation, except at Werer.
No fertilizers were applied at any of the test sites and weeding and cultivation
were done as recommended for the crop in Ethiopia (ARC 2015).
Data Collection: Fresh root yield was the major traits followed by root dry matter
and β-carotene contents. All the data samples were taken from each plot at each
location. Fresh root yield (t ha-1) was measured from the whole plot except the two
border rows and expressed as harvested fresh root weight in kg plot-1 and later
converted to tonnes per hectare. RDMC estimated as a percentage of root dry
weight (g) to fresh root weight (g) where 100- 200 g samples were taken from
[517]
roots of sampled plants in the plot and the samples were dried in an oven at 80oC
for 48 hours and RDMC was measured as percent after constant weight reached
by dividing weight of dried flesh to fresh weight. Root flesh colour was recorded
using a sweet potato colour chart.
Data Analysis: A combined analysis of fresh root yield, RDMC and β-carotene
content data across the environments were analysed using GenStat 14th edition
(Payne et al. 2011) and SAS version 9.3 (SAS Institute Inc. 2003) statistical
packages.

Mean Performance of the Genotypes for Fresh Root Yield: The combined
analysis of variance showed that there was a highly significant (p < 0.001)
difference among the four test environments and the three sweet potato varieties
evaluated for fresh root yield. The mean fresh root yield of the varieties at the four
locations is presented in Table 3. Varieties Vita and Kabode had a mean root yield
of 18.3 and 21.1 ha-1 as compared to the check Kulfo which yielded only 12.6 t ha-
1
. The two varieties showed a yield advantage of 45.2 and 67.5%, respectively
over the standard check (Kulfo).
Mean performance of the varieties for resistance to SPVD, days to

maturity, root dry matter content and flesh colour: The mean SPVD score,
days to maturity, farmers’ preference rank, dry matter content and flesh colour of
the varieties is presented in Table 4. The mean SPVD score of the varieties ranged
from 1.0 (no visible symptom) for varieties Vita and Kabode to 3.9 (highly
affected leaves) for Kulfo variety. The root dry matter content of the varieties
ranged from 21.6% for variety Kulfo to 30.8% for variety Kabode. The two
candidate varieties, Vita and Kabode had higher RDMC than the standard/positive
check variety, Kulfo. Kulfo had the shortest days to maturity (116 days) at Werer
while Vita had the longest days to maturity (125 days) at Hawassa. But all the
three varieties are grouped under the early maturing group when their means are
considered. The two varieties had also intermediate orange flesh colour (better
amount of vitamin A) as compared to the light orange flesh colour of Kulfo
variety. Farmers also ranked the two varieties, Vita and Kabode as 2nd and 1st,
respectively (Table 4).
[518]
Table 3. Root yield (t/ha) performance of early maturing OFSP varieties evaluated at four locations in VVT (2018)
Locations
Dilla Halaba Hawassa Werer
Genotypes Mean on- Mean
On- On- Farmer On- On-
Farmer 1 Farmer 2 Farmer 2 Farmer 1 Farmer 2 Farmer 1 Farmer 2 station on-farm
station station 1 station station
VITA 17.3 22.9 21.6 15.5 16.3 17.5 18.6 17.8 16.5 21.6 20.7 18.0 18.3 18.9
Kabode 20.2 26.7 23.6 18.5 18.9 19.3 22.2 19.5 17.2 23.4 22.4 21.8 21.1 21.2
Kulfo 11.5 12.5 11.3 13.3 15.4 14.7 12.5 13.2 13.5 13.2 13.6 12.1 12.6 13.3
Table 4. SPVD score, days to maturity, farmers preference rank, dry matter content and flesh colour data of early maturing OFSP varieties evaluated at four locations in VVT (2018)
Locations
Genotypes Dilla Halaba Hawassa Werer
SPVD DM FPR DMC FC SPVD DM FPR DMC FC SPVD DM FPR RDMC FC SPVD DM FPR DMC FC
VITA 1.5 121 2 28.3 IO 1.2 122 2 29.1 IO 1.6 125 2 28.8 IO 1.0 118 1 - IO
Kabode 1.3 123 1 30.2 IO 1.0 122 1 30.8 IO 1.2 122 1 30.7 IO 1.0 118 2 - IO
Kulfo 3.5 123 3 22.9 LO 3.3 119 3 23.2 LO 3.9 120 3 21.6 LO 3.1 116 3 - LO
SPVD = Sweetpotato virus disease (1-5, 1 = resistant and 5 = susceptible); DM = Days to maturity; FPR = Farmers preference rank; DMC = Dry matter content; FC = Flesh colour, IO =
Intermediate orange; LO = Light orange.
NB: As the flesh colour intensity increases from light orange to deep orange, the beta-carotene and hence the vitamin A content increases tremendously.
[519]
Conclusion
The OFSP varieties under production in Ethiopia had low root dry matter that
greatly influenced the acceptance and adoption of varieties by farmers. Three
early-maturing sweet potato varieties have been introduced from Uganda and
evaluated for one season where two best varieties, Vita and Kabode have been
selected for verification. These two varieties had been evaluated in the verification
trial along with the check variety, Kulfo. The two varieties had high root yield,
high dry matter content, low virus disease score, and good preference by farmers.
They had also intermediate orange flesh colour (better amount of vitamin A) as
compared to the light orange flesh colour of Kulfo variety. Therefore, these
varieties are potential candidates that could replace Kulfo variety, an OFSP variety
that is currently grown in Ethiopia.
Acknowledgment
The financial support of the South Agricultural Research Institute (SARI),
Ethiopian Institute of Agricultural Research (EIAR) and the International Potato
Center (CIP) is highly acknowledged.
References
ARC. 2015. Sweetpotato production and field management in Ethiopia, Production Manual.
Awassa Agricultural Research Center, Hawassa, Ethiopia, 2015.
Belehu T. 2003. Agronomical and Physiological Factors Affecting Growth, Development and
Yield of Sweetpotato in Ethiopia. PhD Thesis, University of Pretoria, Pretoria.
Burri BJ. 2011. Evaluating sweet potato as an intervention food to prevent vitamin A deficiency.
Comprehensive Reviews in Food Science and Food Safety. 10, 118-130.
Chiona M. 2009. Towards Enhancement of β-carotene Content of High Dry Mass Sweetpotato
Genotypes in Zambia. PhD Thesis, University of KwaZulu-Natal, Pietermaritzburg.
Christina SL. 2007. Nutrient and Sensory Quality of Orange Fleshed Sweetpotato. MSc Thesis,
University of Pretoria, Pretoria.
Gurmu F, Hussein S, and Laing M. 2015a. The potential of orange-fleshed sweet potato to prevent
vitamin A deficiency in Africa. International Journal for Vitamin and Nutrition Research.
84(1-2), 65-78.
Gurmu F, Hussein S, Laing M. 2015b. Diagnostic assessment of sweet potato production in
Ethiopia : constraints, post-harvest handling and farmers’ preferences. Research on Crops.
16(1), 104-115.
Hagenimana V, Low J, Anyango M, Kurz K, Gichuki ST and Kabira J. 2001. Enhancing vitamin A
intake in young children in Western Kenya: orange-fleshed sweet potatoes and women farmers
can serve as key entry points. Food and Nutrition Bulletin. 22, 370-387.
Kapinga R. Anderson P. Crissman C. Zhang D. Lemaga B. Opio F. 2005. Vitamin A partnership
for Africa: a food based approach to combat vitamin A deficiency in sub-Saharan Africa
through increased utilization of orange fleshed sweet potato. Chronica Horticulturae. 45(3),
12-14.
[520]
Kapinga R, Tumwegamire S, Ndunguru J, Andrade MI, Agili S, Mwanga ROM, Laurie S, Dapaah
H. 2010. Catalogue of Orange Fleshed Sweetpotato Varieties for Sub-Saharan Africa.
International Potato Center (CIP), Lima, Peru.
Mbwaga Z, Mataa M, Msabaha M. 2008. Quality and yield of orange fleshed sweet potato
(Ipomoea batatas) varieties grown in different agro-ecologies. In: Kasem, Z.A., editors. The
8th African Crop Science Society (ACSS) Conference Proceeding, El-Minia, Egypt. 27-31
October, 2007. African Crop Science Society, pp. 339-345.
Mwanga ROM. Odongo B, Niringiye C, Zhang D, Yencho GC, Kapinga R. 2003. Orange fleshed
sweetpotato breeding activities in Uganda. In: The 6 th Conference of the African Crop Science
Society (ACSS) Conference Proceeding, 12-17 October, 2003, Nairobi. African Crop Science
Society, Kampala, Uganda.
Payne RW, Murray DA, Harding SA, Baird DB, Soutar DM. 2011. GenStat for Windows (14th
Edition) Introduction. VSN International, Hemel Hempstead, UK.
SAS Institute Inc. 2003. Version 9.1. SAS Institute Inc., Cary, NC.
Woolfe JA. 1992. Sweetpotato: An Untapped Food Resource. Cambridge University Press,
Cambridge, United Kingdom.
[521]
[522]
Participatory Orange-fleshed Sweet potato Variety
Selection in Ethiopia: Integrating Farmers’ and
Researchers’ Criteria for Variety Selection
Fekadu Gurmu1, Abiyot Aragaw2, Roland Brouwer2,
Bililign Mekonen1 and Zeritu Shashego1
1
South Agricultural Research Institute, Hawassa Research Centre, P.O.Box 6, Hawassa, Ethiopia
2
International Potato Center, P.O.Box 5, Hawassa, Ethiopia; *Corresponding author: fekadugb@gmail.com
Abstract
Sweetpotato is one of the food security crops mainly produced by smallholder farmers
in Ethiopia. Orange-fleshed sweet potato (OFSP) can be one of the cheapest sources of
vitamin A for resource-poor farmers in Ethiopia. Although there have been some
research efforts to release OFSP varieties in Ethiopia using only researchers’ criteria,
the adoption of these varieties by smallholder farmers was generally low, especially in
southern Ethiopia. Therefore, the current study was conducted to assess farmers’
criteria for the selection of OFSP varieties in Ethiopia in order to complement
farmers’ and researchers’ criteria during variety development. The study was
conducted at 13 sites of the Sidama and Gedeo zones of the Southern Region using a
mother and baby trial design. At each study site, Farmers Training Centers (FTCs) and
farmer fields were used as replications. A total of seven varieties were used for the
study and more than 1009 farmers (561 male and 448 female) participated in the
varietal assessment. Farmers’ preferences for agronomic as well as postharvest and
taste-test data were recorded using focused group discussion and preference rankings.
The farmers were allowed to rank the seven OFSP varieties from 1 (best) to 7 (least)
according to the different perceived agronomic and taste criteria. Accordingly, the
mean ranks per study site and for each variety were estimated. Similarly, the standard
scientific parameters (yield and yield components) were used to evaluate the seven
varieties. The data were analysed using GenStat and SPSS statistical packages. The
results indicated that there were significant differences (p<0.01) among the tested
varieties and the locations (villages). Farmers selected the varieties based on
agronomic assessment as root yield, vine thickness, resistance to drought, early
maturity, resistance to disease and insect pests, vine length, root shape and size (round
and medium) and cooking time (short), in that order. For the taste-test of boiled roots
and leaves, they ranked the criteria as powdery (high dry matter), good texture, low
sugariness, color of boiled roots (deep orange) and generally good test. Based on the
above farmers’ selection criteria and other relevant researchers’ data such as high
root yield (> 26 ton ha-1), root dry matter content (> 30%), beta-carotene content (> 9
mg 100 g-1), resistance to sweet potato virus disease expressed as a score (< 1.5)
where1= immune and 5 = susceptible, four varieties, namely Ukr/Eju-10, Ukr/Eju-13,
Vita and Kabode have been identified for release.
Keywords: Criteria, Farmers’ preferences, OFSP, Participatory variety selection,

Vitamin A.
[523]
Introduction
Sweetpotato is one of the food security crops in Ethiopia grown mainly by
smallholder farmers in the south, east and western parts of the country (Tadesse,
2006; Tofu et al., 2007; Gurmu et al., 2015a). Production of sweet potato is
mainly confined to two regions in Ethiopia: the Southern Nation, Nationalities,
and Peoples Region (SNNPR) and Oromia Region. The crop covered over 53,400
ha with a production of over 1.85 million tons in the 2017/18 cropping season
(CSA, 2015). Sweetpotato is rich in macro and micro-nutrients, among which
vitamin A is the major one in the orange-fleshed types. Vitamin A is a very
essential element for human health. It protects eye health, supports immunity,
relieves inflammation, keeps skin glowing, contains cancer-fighting properties,
boosts bone health, reduces cholesterol, aids in reproduction and development,
promotes tissue repair and prevents urinary stones. On the contrary, vitamin A
deficiency (VAD) leads to blindness, retarded growth and death. It is estimated
that some three million children in sub-Saran Africa (SSA) under the age of 5
years suffer from partial or total blindness as a result of VAD (Kapinga et al.,
2005, 2010; WHO, 2009; Gurmu et al., 2015b). It affects especially pre-school
children and pregnant and lactating mothers.
VAD is a serious health problem in most communities in Ethiopia, particularly

affecting, young children and mothers. The deficiency is reaching alarmingly high
proportions in some regions of the country (Tofu et al., 2007; Demissie et al.,
2009; WHO, 2009). Demissie et al. (2009) reported 23-48% VAD prevalence
among children, pregnant and lactating mothers in Ethiopia. A very high
prevalence rate of clinical VAD, close to 8%, was reported in some parts of the
country (Demissie et al., 2009). Among the different strategies used to combat the
problem of VAD, consuming plants that are naturally rich in vitamin A is the
major one. Foods from plant sources are easily available and provide vitamin A in
the form of β-carotene. One of the natural vitamin A-rich plants is orange-fleshed
sweet potato (OFSP).
OFSPs are relatively cheap and easily accessible sources of vitamin A that are
used to combat the problem of VAD affecting the poor community of most sub-
Saharan African countries, including Ethiopia. Some OFSP varieties have been
released in Ethiopia and have been distributed to farmers (Tofu et al., 2007).
However, the acceptance of these varieties by farmers is very low, mainly due to
the low root dry matter content (RDMC) of the OFSP varieties (Tadesse, 2006;
Gurmu et al., 2015a). Farmers prefer to grow the white-fleshed sweet potato
(WFSP) varieties that have high RDMC. However, the WFSP varieties have no β-
carotene and this limits the efforts that are being made to combat VAD. Currently,
OFSP varieties with high dry matter contents have been developed by the
Hawassa Agricultural Research Center of the South Agricultural Research
[524]
Institute for evaluation and release for wider production. Therefore, the current
study was conducted to test the performance of the newly bred varieties and some
introduced varieties along with standard checks for adaptation and to test the
farmers’ preference for the varieties in selected districts of Sidama and Gedeo
zones of the Southern Nationa, Nationalities and Peoples’ Region (SNNPR).
To this effect, CIP has collaborated with SARI through the EU-financed project
“Quality Diets for Better Health” to identify new climate-smart and vitamin A rich
sweet potato varieties to be disseminated in SNNPR.
Study sites: The study was conducted at 24 kebeles on households’ farm plots of
the Sidama and Gedeo zones of the SNNPR. At each kebele, farmers training
centers (FTCs) and farmer fields were used as replications. Therefore, there were a
total of 48 sites that were used for the study. The description of the sites (FTCs) is
given in Table 1.
Table 1. Description of the study areas (FTCs)
No. Name of Kebele Woreda Altitude

(Household) (District) Zone (masl)* Coordinates
1 D/Morocho Aleta Chuko Sidama 1711 06°39.227’ 038°18.073’
2 D/Elelicho ‘’ ‘’ 1787 06°39.193’ 038°20.210’
3 L/Haytala ‘’ ‘’ 1693 06°37.067’ 038°17.184’
4 Miridicha ‘’ ‘’ 1660 06°34.395’ 038°16.365’
5 Rufo Waeno ‘’ ‘’ 1817 06°33.921’ 038°19.638’
6 Dibicha ‘’ ‘’ 1623 06°30.030’ 038°18.639’
7 Mekela ‘’ ‘’ 1783 06°31.098’ 038°20.302’
8 L/Honcho ‘’ ‘’ 1830 06°30.529’ 038°22.445’
9 Halo ‘’ ‘’ 1928 06°33.122’ 038°23.086’
10 Gure ‘’ ‘’ 1876 06°35.839’ 038°21.804’
11 Chichu Woyama ‘’ ‘’ 1682 06°35.331’ 038°16.570’
12 Chuko Woyama ‘’ ‘’ 1782 06°35.227’ 038°18.276’
13 Chichu Dilla zuria Gedeo 1532 06021’48.7’’ 038018’15.9’’
14 Sisota ‘’ ‘’ 1674 06023’05.0’’ 038019’22.3’’
15 Andida ‘’ ‘’ 1867 06022’22.3’’ 038021’15.0’’
16 Amba ‘’ ‘’ 1975 --- ---
17 Holena ‘’ ‘’ 2134 06021’27.1’’ 038021’00.4’’
18 Otilicho ‘’ ‘’ 2078 06024’10.1’’ 038023’17.3’’
19 Tumata Wonago ‘’ 1570 06 20’12.7’’
0 038016’34.6’’
20 Kara Soditti ‘’ ‘’ 1852 06018’03.8’’ 038013’32.5’’
21 Bele Bukisa ‘’ ‘’ 1778 06017’27.1’’ 038015’05.6’’
22 Tokicha ‘’ ‘’ 1736 06017’55.5’’ 038017’03.1’’
23 Hase Haro ‘’ ‘’ 1641 06018’15.1’’ 038017’14.9’’
24 Mekonisa ‘’ ‘’ 1893 06 16’42.2’’
0 038013’34.6’’
masl = meter above sea level; min = minimum; max = maximu
Plant materials: A total of seven varieties have been used for the study. Two of
the varieties were elite OFSP genotypes bred in Ethiopia and three were released
in Uganda and introduced from Kenya. Two varieties widely grown in Ethiopia
[525]
were used as checks in the evaluation process. The list of the varieties is indicated
in Table 2.
Table 2. Description of sweet potato varieties used for the study
Predominant
No Breeding names Farmer names flesh-colour RDMC
1 Ukrewe x Ejumula-10 Alamura DO 31.8
2 Ukrewe x Ejumula-13 Dilla DO 32.4
3 SPK 004/6 Vita IM 29.6
4 SPK 004/6/6 Kabode IM 30.3
5 NASPOT 13 NASPOT 13 DO 26.8
6 Kulfo Kulfo PO 22.5
7 Awassa-83 Awassa-83 W 33.3
IO = intermediate orange, DO = dark orange, PO = pale orange, W = white,
RDMC = root dry matter content.
Experiential design and agronomic practices: The design used for the study
was mother and bay design. In the mother trials, all the seven varieties were
planted at the same place and replicated across FTCs and farmers. Therefore, the
mother trials were planted on 12 FTCs and 12 farmers’ fields, comprising 24 total
numbers of trials in Aleta Chuko woreda. In Dilla zuria and Wonago weredas, the
trials were planted on 6 FTCs and 6 farmers’ fields in both weredas, which
resulted in 24 total numbers of trials. In the baby trials, two of the improved
varieties were given to each farmer to plant along with his local variety.
Accordingly, all the seven varieties were given in pairs to over 500 farmers for the
baby trials. The total plot size for each variety in the mother trials was 25m 2.
Regarding the agronomic practices, 0.60 meter and 0.30-meter spacing were used
between each row and each plant, respectively. Therefore, the total number of
plants in each plot was 139 plants. However, for data collection, the border rows
and the first and last plants from each side of the rows were excluded. No fertilizer
was applied and the trials were hand weeded as deemed necessary.
Data collected
Agronomy data: Data were recorded on the following qualitative and quantitative
traits: Sweetpotato virus disease (SPVD), recorded on singe plant basis, using a
scale of 1 to 9, where 1 = No visible virus symptoms; 3 = Faint mosaic; 5 =
Mosaic clear; 7 = Heavy mosaic and stunted; 9 = Necrosis (dead patches) (Hahn,
1979; Mwanga et al., 2002; Ndunguru et al., 2009); Days to 50% maturity,
recorded when leaves of 50% of the plants were turned yellow; Vine length (cm),
measured by taking the vertical length of the vine from the ground level to the tip
of the main shoot of the plants; Vine internode length (cm), measured on three
internodes located in the middle section of the vine of five randomly selected
[526]
plants in a plot; Vine thickness: measured on three vines in the middle section of
the vine of five randomly selected plants in a plot and measured using a caliper;
Aboveground fresh weight (ton ha-1), measured as a weight of the above ground
parts of the two central rows converted to ton ha-1; Storage root length (cm),
recorded as a mean vertical length from the bottom to the tip of roots of ten
randomly selected plants in a plot; Storage root girth (cm), expressed as the mean
diameter of the middle portion of roots of five randomly taken plants in a plot and
measured using a caliper; Number of storage roots per plot, measured as a number
of storage roots produced by the whole plant in a plot; Individual root weight (g),
estimated by dividing storage root weight per plot by the number of storage roots
per plot; Fresh root yield, expressed as harvested fresh roots weight of two central
rows converted to tonnes ha-1; Predominant skin colour, recorded on a scale of 1-9
as described by Huaman (1991) and (1999), where 1 = White, 2 = Cream, 3 =
Yellow, 4 = Orange, 5 = Brownish orange, 6 = Pink, 7 = Red, 8 = Purple red, 9 =
Dark Purple; Predominant flesh colour, recorded in scale of 1-9 as described by
Huaman (1991) and (1999), where 1 = White, 2 = Cream, 3 = Dark cream, 4 =
Pale yellow, 5 = Dark yellow, 6 = Pale orange, 7 = Intermediate orange, 8 = Dark
orange, 9 = Strongly pigmented with anthocyanin;
Farmers’ preference data: Farmers’ preferences for agronomic as well as

postharvest and taste-test data were recorded using focused group discussion and
preference rankings. The farmers were allowed to rank the seven OFSP varieties
from 1 (best) to 7 (least) according to the different perceived agronomic and taste
criteria. Accordingly, the mean ranks per study site and for each variety were
estimated.
Data analysis: For this report, only root yield data were considered for analysis
among the whole phonological and agronomic data that were collected in the
course of the study. Mean of the FTCs and farmers’ fields were used as kebele
data for analysis, making the total number of locations (kebeles) 24 i.e (12
Kebeles in Aleta chuko and 12 in Dilla zuria and Wonago wereds). A combined
analysis of data across the 12 locations was conducted using GenStat 14 th edition
(Payne et al., 2011) statistical package. Then the data was re-analyzed at zonal
level and the results were presented accordingly.
The following statistical model was used for the combined analysis of variance
over environments:
Y ijk    G i  E j
 GE ij
 R k   ijk
Where: Yijk is observed value of genotype i in block k of environment j, μ is grand

mean, Gi is effect of genotype i, Ej is environment or location effect, GEij is the
interaction effect of genotype i with environment j, Rk is the effect of replication k
[527]
in environment j, єijk is error (residual) effect of genotype i in block k of
environment j.

Analysis of variance (ANOVA): The result of the ANOVA indicated that there
was a significant difference (p<0.01) among the varieties tested across the 24
locations (villages) as indicated in Table 3. The replication effect was also
significant (p<0.01), indicating the presence of significant differences among the
24 test locations (villages). The analysis was also conducted at zonal level using
zones as locations and kebeles as replications within a zone. The ANOVA at the
zonal level also indicated that there was a significant difference (p<0.01) among
the varieties and the locations i.e the two zones (Table 4). The replication effect
was still significant (p<0.01), indicating the presence of significant differences
among the 12 villages in each zone. The treatment by location interaction was not
significant, indicating that each variety performed consistently across the two
locations (zones).
Table 3. ANOVA of root yield (ton ha-1) of sweet potato varieties tested during belg season of
2018 across 24 kebeles of three wereds in Sidama and Gedeo zones of SNNPR
Sources of variation d.f. SS MS

Replication 23 9359.68 406.94**
Treatment 6 2253.98 375.66**
Residual 138 5097.48 36.94
Total 167 16711.14
Table 4. ANOVA of root yield (ton ha-1) of sweet potato varieties tested during
belg season of 2018 across 12 kebeles, each in Sidama and Gedeo zones of SNNPR
Sources of variation d.f. SS MS
Replication 11 4545.31 413.21**
Location 1 801.38 801.38***
Treatment 6 2253.98 375.66**
Treatment x Location 6 201.97 33.66ns
Residual 143 8908.50 62.30
Total 167 16711.14
Mean performance of the varieties for root yield: The mean root yield
performance of the varieties across the 24 kebeles are presented in Table 5. The
highest mean root yield of 25.6 t ha-1 was recorded from the variety Kabode across
the 24 kebeles, followed by Kulfo, which gave 24.5 t ha-1. However, there was no
significant difference between the six varieties, except with one of the check
varieties (Awassa-83), where all the varieties significantly performed better than
this variety. Awassa-83 is a dominantly grown sweet potato variety in Ethiopia.
Among the test locations (kebeles), Chichu gave the highest root yield of 35.5 t ha-
1
, followed by Mekonisa (33.8 t ha-1) and Chichu Woyama (31.7 t ha-1). The
lowest yielding kebels were Otilicho and Holena with mean root yields of 5.8 and
[528]
6.8 t ha-1, respectively. This low mean root yield may be attributed to the altitude
of the kebels, which is relatively high. Sweetpotato is a low land crop and it needs
a relatively higher temperature for better root yield performance. Negeve et al.
(1992) stated that root yield, number of roots and proportion of marketable roots
decline with increasing altitude. The crop performs better between minimum and
maximum temperature ranges of 15 and 330C, respectively (Negeve et al., 1992;
Ramirez, 1992). When the temperature drops below 100C, the growth of the crop
is severely retarded. The crop is sensitive to frost (Negeve et al., 1992) and
therefore its cultivation is confined to the tropics and warm temperate regions.
Therefore, it is very important to cluster the kebeles according to their altitude
before analysing the over locations data.
Table 5. Mean root yield of sweet potato varieties tested during belg season of 2018 across 24 kebeles of t
hree wereds in Sidama and Gedeo zones of SNNPR
No. Name of Kebele Woreda Genotypes
(Household) (District) Zone 1 2 3 4 5 6 7 Mean
1 D/Morocho Aleta Chuko Sidama 33.2 22.7 30.3 30.2 26.2 21.1 27.6 27.3
2 D/Elelicho ‘’ ‘’ 15.6 13.1 21.2 14.5 24.5 14.7 17.3 17.3
3 L/Haytala ‘’ ‘’ 23.9 25.5 30.2 33.4 34.0 18.3 30.4 27.9
4 Miridicha ‘’ ‘’ 20.6 27.7 34.0 35.5 32.4 20.3 25.5 28.0
5 Rufo Waeno ‘’ ‘’ 22.4 16.3 19.9 26.6 20.5 9.8 15.3 18.7
6 Dibicha ‘’ ‘’ 24.1 23.1 28.1 31.3 21.1 32.0 10.2 24.3
7 Mekela ‘’ ‘’ 19.2 18.5 21.2 25.3 16.1 20.1 12.2 19.0
8 L/Honcho ‘’ ‘’ 13.8 14.6 28.7 23.9 24.1 22.0 17.6 20.7
9 Halo ‘’ ‘’ 14.7 22.1 18.0 27.0 31.9 13.1 3.9 18.6
10 Gure ‘’ ‘’ 21.5 21.3 20.8 20.2 19.9 33.8 8.2 20.8
11 Chichu Woyama ‘’ ‘’ 21.9 31.7 29.3 41.7 29.1 50.6 17.3 31.7
12 Chuko Woyama ‘’ ‘’ 22.2 31.3 26.3 30.2 19.6 35.2 5.7 24.4
13 Chichu Dilla zuria Gedeo 38.5 34.1 42.9 39.1 41.7 27.2 25.3 35.5
14 Sisota ‘’ ‘’ 20.7 17.3 17.7 31.0 18.8 33.9 19.0 22.6
15 Andida ‘’ ‘’ 9.8 7.6 9.0 14.5 19.9 15.3 10.0 12.3
16 Amba ‘’ ‘’ 5.0 27.0 11.3 20.2 10.9 23.3 17.6 16.5
17 Holena ‘’ ‘’ 4.2 8.4 10.0 5.8 5.1 11.2 2.6 6.8
18 Otilicho ‘’ ‘’ 3.8 5.3 9.4 8.2 3.6 8.5 1.6 5.8
19 Tumata Wonago ‘’ 22.7 19.6 21.5 28.4 32.8 14.5 16.4 22.3
20 Kara Soditti ‘’ ‘’ 27.8 18.1 22.2 28.7 19.4 40.4 16.1 24.7
21 Bele Bukisa ‘’ ‘’ 7.2 9.2 15.7 18.3 8.0 20.4 8.5 12.5
22 Tokicha ‘’ ‘’ 8.1 7.4 19.1 23.8 21.6 35.2 8.9 17.7
23 Hase Haro ‘’ ‘’ 12.6 22.4 17.2 16.7 16.5 18.9 6.1 15.8
24 Mekonisa ‘’ ‘’ 27.1 41.0 29.3 38.9 40.2 48.6 11.6 33.8
Mean of genotypes 18.4 20.2 22.2 25.6 22.4 24.5 14.0 21.0
1 = Alamura; 2 = Dilla; 3 = Vita; 4 = Kabode; 5 = NASPOT-13; 6 = Kulfo; 7 = Awassa-83
Farmers’ preference test: The farmers have selected the varieties based on
agronomic assessment as root yield, vine thickness, resistance to drought, early
maturity, resistance to disease and insect pests, vine length, root shape and size
(round and medium) and cooking time (short), in that order (Table 6). For the
taste-test of boiled roots and leaves, they ranked the criteria as powdery (high dry
matter), good texture, low sugariness, color of boiled roots (deep orange) and
overall test of the leaves (Table 7).
[529]
Table 6. Pair-wise ranking of the agronomic and postharvest criteria by farmers
Criteria B C D E F G H Points Rank

A A C D E A G A 3 5
B C D E B G B 2 6
C C C C G C 6 2
D D D G D 5 3
E E G E 4 4
F G F 1 7
G G 7 1
H 0 8
A = Resistance to disease and insect pests; B = Vine length; C = Vine thickness; D = Resistance to drought; E = Early
maturity; F = Root shape and size; G = Root yield; H = Cooking time
Table 7. Pair-wise ranking of the criteria for taste-test of boiled roots and leaves
Criteria N O P Q Points Rank

M N O P M 1 4
N O N N 3 2
O O O 4 1
P P 2 3
Q 0 5
M = Color of boiled roots (deep orange); N = Good texture; O = Powdery (high dry matter);
P = Low sugariness; Q = Overall test of the leaves
Conclusion
Vitamin A deficiency (VAD) is a serious health problem in most communities in
Ethiopia, particularly affecting, young children and mothers. Consuming plants
that are naturally rich in vitamin A is a sustainable approach to combat the
problem of VAD. One of the natural vitamin A-rich plants is orange-fleshed sweet
potato (OFSP). OFSPs are relatively cheap and easily accessible sources of
vitamin A that are used to combat the problem of VAD affecting the poor
community of most sub-Saharan African countries, including Ethiopia. Some
OFSP varieties have been released in Ethiopia and have been distributed to
farmers. However, the acceptance of these varieties by farmers is very low, mainly
due to the low root dry matter content of the OFSP varieties. Currently, OFSP
varieties with high root dry matter contents have been developed by the Hawassa
Agricultural Research Center and evaluated in collaboration with CIP in 2018 belg
(short rain) season in 24 kebeles of three districts in Sidama and Gedeo zones of
SNNPR using Mother and Baby Design.
Based on the above farmers’ selection criteria and other relevant researchers’ data
such as high root yield (> 26 ton ha-1), root dry matter content (> 30%), beta-
carotene content (> 9 mg 100 g-1), resistance to sweet potato virus disease
expressed as a score (< 1.5) where1= immune and 5 = susceptible, four varieties,
namely Ukr/Eju-10,Ukr/Eju-13, Vita and Kabode have been identified as
candidate varieties for release.
[530]
Acknowledgment
The authors acknowledge the South Agricultural Research Institute for providing
the test materials and for technical support. The International Potato Center is also
highly acknowledged for the provision of financial and technical support.
References
CSA. 2015. Ethiopia Agricultural Sample Survey 2014/2015: Report on Land Utilization (Private
Peasant Holdings, Meher Season). Central Statistical Agency (CSA), Federal Democratic
Republic of Ethiopia, Addis Ababa, Ethiopia.
Demissie T. Ali A, Mekonnen Y, Haider J and Umeta M. 2009. Demographic and health-related
risk factors of subclinical vitamin A deficiency in Ethiopia. Journal of Health and Population
Nutrition 27: 666-673.
Gurmu F, Hussein S and Laing M. 2015a. Diagnostic assessment of sweet potato production in
Ethiopia : constraints, post-harvest handling and farmers’ preferences. Research on Crops
16(1): 104-115.
Gurmu F, Hussein S and Laing M. 2015b. The potential of orange-fleshed sweet potato to prevent
vitamin A deficiency in Africa. International Journal for Vitamin and Nutrition Research
84(1-2): 65-78.
Kapinga R, Anderson P, Crissman C, Zhang D, Lemaga B and Opio F. 2005. Vitamin A
partnership for Africa: a food based approach to combat vitamin A deficiency in sub-Saharan
Africa through increased utilization of orange-fleshed sweet potato. Chronica Horticulturae
45(3): 12-14.
Kapinga RS, Tumwegamire S, Ndunguru J, Andrade MI, Agili S, Mwanga ROM, Laurie S and
Dapaah H. 2010. Catalogue of Orange Fleshed Sweetpotato Varieties for Sub-Saharan
Africa. International Potato Center (CIP), Lima, Peru.
Negeve JM, Hahn SK and Bouwkamp J.C. 1992. Effect of altitude and environments on sweet
potato yield in Cameroon. Tropical Agriculture-Trinidad 69: 43-48.
Ramirez GP. 1992. Cultivation, harvesting and storage of sweet potato products. In: David, M. and
N. Solveig, editors. Proceedings of the FAO Expert Consultation on Roots, Tubers, Plantains
and Bananas in Animal Feeding, Cali, Colombia. 21–25 January 1991. FAO, Rome. p 203-
215.
Tadesse T. 2006. Evaluation of Root Yield and Carotene Content of Orange-fleshed Sweetpotato
Clones Across Locations in the Southern Region of Ethiopia. MSc Thesis, Hawassa
University, Hawassa.
Tofu A, Anshebo T, Tsegaye E and Tadesse T. 2007. Summary of progress on orange-fleshed
sweet potato research and development in Ethiopia. In: Proceedings of the 13th International
Society for Tropical Root Crops (ISTRC) Symposium, 9-15 November, 2007, Arusha.
ISTRC, Arusha, Tanzania.
WHO. 2009. Global Prevalence of Vitamin A Deficiency in Population at Risk from 1995-2005.
WHO Global Database on Vitamin A Deficiency. World Health Organization, Geneva,
Switzerland.
[531]
[532]
Effects of Soil Amendments on Growth and
Biomass Yield of Early Generation Seeds of Sweet
potato (Ipomoea batatas (L.) Lam)
Grown in Net Tunnels
Daniel Markos1 and Fekadu Gurmu1
1
Hawasa Agricultural Research Center, P.O. Box 06, Hawasa, Ethiopia
Abstract
Early generations vines of sweet potatoes were multiplied in insect-proof net tunnels
that have rapid multiplication beds. The net tunnels are meant to inhibit the entrance
of virus transmitting insects such as aphids and whiteflies whereas rapid
multiplication beds require suitable soil amendments to support growth of vigorous
vines. To this end, farmyard manure, wood ash, sawdust, compost, coffee husk and
control (soil only) were compared during belg and meher seasons of 2016 to 2017 in
Chefe testing site of Hawasa Agriculture Research Center. Significantly higher
numbers of vines were obtained due to farmyard manure (4686 vines/plot) and coffee
husk (4602 vines/plot) compared to the control (2683 vines/plot). Significantly
longer internodes were recorded due to farmyard (100.6% greater) compared to the
control. Pencil root length, pencil root width and below-ground biomass were not
affected significantly (P<0.05) by the difference in soil amendments (P<0.05) in this
experiment. The percentage increment in above-ground biomass due to farmyard
manure, coffee husk and compost were 62.01%, 53.56% and 49.87%, respectively.
The increment in above-ground biomass was negative due to sawdust. Results
showed that soil pH showed 8.9% and 1.4% increment due to the application of
wood ash and sawdust whereas there was 9.3, 5.9 and 0.9 decrease in pH due to use
of compost, coffee husk and farmyard manure. All soil amendments contributed for
moisture conservation compared to control but a coffee husk. 47.0%, 31.2%, 30.3%
and 26.5% increase in cation exchange capacity was observed by the end of the
second year due to the use of coffee husk, wood ash, sawdust and farmyard manure
medium, respectively. The potassium content was increased by 47.0%, 45.5% and
35.9% due to the use of wood ash, farmyard manure and coffee husk medium,
respectively. Thus, 50% farmyard manure, 50% coffee husk or 50% compost are
recommended for better growth of pre-basic sweet potato vines in net tunnels.
Keywords: Coffee husk, compost, farmyard manure, internodes, pencil roots, vine
yield
Introduction
Sweet potato (Ipomia batatas L.) Lam) is grown in over one hundred countries in
the tropical, sub-tropical and temperate climates. The total annual production of
sweet potato is 106–110 million tons with the plantation area of 8.1 million
hectares (FAO 2008; 2011). It is a staple food for many people in South-East Asia,
[533]
Latin America and Africa (including Ethiopia). However, owing to changing
climate and food habits, there was growing demand for sweet potato planting
materials in Ethiopia. Sweetpotatoes produced in tissue culture laboratories are
generally small miniature plantlets with very weak root and shoot systems. These
were hardened off in screen houses and net tunnels to adapt the harsh
environments prevailing in the field. These materials are called early generation
(pre-basic) seed. Thereafter, sweet potatoes are propagated using vine cuttings
obtained either from freshly harvested plants or from the nursery. However,
recurrent use of vines can cause increased weevil infestation and subsequent yield
reduction. Vines obtained from a nursery should be healthy and vigorous for
maximum root production. Healthy sweet potato cuttings respond very well to soil
amendments and applied fertilizers under growing conditions (Onunka et al.,
2012; Udoh et al., 2014). Whiteflies and aphids are the major vectors for sweet
potatoes feathery and mottled viruses, the two chronic diseases that resulted in
total sweet potatoes failure in southern Ethiopia from 2010 to 2014 growing
seasons. This necessitated the multiplication of early generation seeds in net
tunnels. However, sweet potato vine business in Ethiopia is dominated by the
informal sector; demand for quality planting materials of improved varieties is low
and inconsistent. This was mainly due to a lack of demand for quality declared
seed (QDS from root producers, which was in turn due to lack of awareness, low
willingness to pay, and lack of registered QDS producers (SPHI, 2014). Therefore,
there was a need to multiply large amounts of plantlets in Eastern and Southern
Africa including Ethiopia in order to respond to the ever-increasing demand of
virus-free sweet potato planting materials required in a very short period. This was
mainly because sweet potato growers have limited access to timely availability of
quality sweet potato seed which intern contributes to sub-optimal root yields
(Srinivasulu et al., 2017).
Increasing vine production per unit area and per unit time could be obtained
through increasing plant population, growing in high rainfall and temperature
conditions, and using nitrogenous fertilizers (Meyreles et al., 1977; Maniam et al,
2012). Irrigating twice a day (early morning and late afternoon) has been practiced
for watering in nonrainy times. Depending on the growing conditions, harvesting
can be done about 80 to 100 days from the time of planting or last cutting (Onunka
et al., 2012 and Udoh et al., 2014). Soil amendments with organic manure
improve physical properties, microbial activity and nutrient availability (Snyman
et al., 1998). Sas-Paszt et al. (2014) compared different mulching materials and
elaborated that only sawdust mulch significantly increased the number of apple
fruit in size diameter class of 7.0 - 7.5 cm compared with the control. They also
noted that the best pH and organic matter results were observed with the use of the
compost, cow manure and the mycorrhizal substrate, whereas the concentrations
of P, K and Mg, most of microelements and soil organic matter were also
elevated. However, crops like sweet potato rarely benefit from that favor mainly
[534]
because they were branded wrongly poor man’s crop requiring little or no external
inputs. The bulkiness of the organic amendment can be reduced through proper
management during preparation, application, storage, handling and enrichment,
and this helps to reduce drudgery. Soil macro- and micronutrient mining and
organic matter depletion could also result from an extensive farming system with
less nutrient replenishment, crop residue removal, the inefficiency of soil nutrient
input, excessive tillage, overgrazing, deforestation, and soil erosion. Despite
frequent planting and harvesting the vines for planting purposes, there were not
any soil amendments implemented in those growing systems. As a result, there
were complex signs and symptoms of nutritional disorders observed among
growing vines. Therefore, there is a need to identify soil amendments to be used
for sweet potato production in net tunnels. Thus, this experiment was designed to
determine suitable soil amendments that maximize the vine production of sweet
potatoes in net tunnels.

Description of the Study Area: The study is carried out at Hawasa Agricultural
Research Center Hawasa Chefe Kote Jebesa research Station (Altitude 1700 m
a.s.l, 07o03’51.8” N latitude and 038o28’52.2” E longitude) inside net tunnels used
for early generation seed multiplication from 2016 to 2017. The area receives a
total of 1000 to 1200 mm rainfall in bimodal raining pattern with short rains (also
called belg rains) coming from April to May whereas long rains (also called meher
rains) coming from July to October. The mean minimum and maximum
temperatures of the area were 16.50C and 29.20C, respectively.
Experimental design and field management: The experiment consisted of six

treatments arranged in a Randomized Complete Block Design with each treatment
replicated three times in belg and meher season. The treatment included 50%
farmyard manure (10 kg/m2), 50% saw dust (10 kg/m2), 50% wood ash (10
kg/m2), 50% compost (10 kg/m2), 50% coffee husk (10 kg/m2) and 100% soil
medium. Wood ash was brought from villages and farmyard manure was obtained
at 50% moisture from dairy farms in Hawasa. The compost was prepared
following heap composting procedures. Coffee husk was obtained from coffee
pulper in Yirgalem area. The sawdust was obtained from Hawasa chip wood
factory. As all of these materials are by-products, payments were done only for
labor during collection. The composition of soil medium would behalf and that of
the added medium would behalf in all treatments except the control and mixed
well with the soil through tillage operation. Uniform depth of each of the
treatment medium was established in each of the beds by mixing the medium
uniformly with the topsoil. Each treatment was laid in a plot size of 2 m x 1 m in a
raised bed, and plants were spaced by 20 cm x 10 cm row by plant spacing. The
test sweet potato variety called Kulfo (LO3232) was grown in all of the beds on
[535]
the same dates. 50g of urea was top-dressed uniformly in each rapid multiplication
beds one month after planting. For belg season cropping, the soil amendments was
incorporated to the soil in the first week of February and the vines were planted in
the first week of March every year. Soil amendments were incorporated in June
and planting of vines was carried out in July as far as meher planting was
considered. The crop was hand-weeded weekly from the second week until the
sixth week when weed invasion was less problematic due to leaf canopy closure.
Harvesting was done in the fourth month.
Sampling and measurements

Plant sampling and measurements: Plants were selected at random from the
centre row (datum row) and two plants at each end of the centre row were
considered as buffers. Vine length/branch, branch number/plant and leaf
number/plant were measured in second, third and fourth of the month of planting.
Branch number/plant was counted from randomly selected five plants and divided
for number of sampled plants to get branch number/plant. Leaf number per plant
was counted from randomly from five plants grown datum rows of a plot. Vine
length was measured from twenty branches that were sampled randomly.
Subsequently, number of vines per plant and number of vines/ha was calculated as
follows;
Number of vines/plant = (mean number of branches/plant)*(mean length of vines /branches)

30cm
Number of vines/ha = (Number of vines/plant) (number of plants/ha)
The number of internodes/plant was counted from randomly selected five plants
and divided for number of sampled plants to get number of internodes/plant. The
length of the internodes was measured from twenty vines that were sampled
randomly and divided for number of sampled vines to get mean length of
internodes/plant. Yield of aboveground biomass (YAGB) was measured by
weighing all above-ground vegetative parts of plants found in datum rows by the
end of the fourth month during harvesting. The yield of below-ground biomass
(YBGB) was measured by weighing pencil roots and adventitious roots after
harvesting plants found in datum rows by the end of the fourth month during
harvesting. Pencil root length (PRL) and pencil root width (PRW) were measured
from randomly selected ten pencil roots harvested from the net plot area.
Soil sampling and measurements: Regarding soil physic-chemical properties,

EC was measured using ES ISO 11265: 2014 procedures and pH-H2O was
measured through ES ISO 10390: 2014 procedures. The percentage of sand, clay
and silt were analyzed using Boycous Hydrometer and texture determined using
the textural triangle. Total nitrogen was determined using ES ISO 11261:2015
procedures. CEC was analyzed using the ammonium acetate method. B, Ca, Co,
Fe, K, Mg, Mn, Na, P, S, Zn, Al, Mo and Si were determined using Mehlich-3
[536]
procedures in Horticoop Ethiopia soil and water analysis laboratory in Debrezeit,
Ethiopia.
Data analysis: The collected data were statistically analyzed using the PROC
ANOVA function of SAS 9 computer software and treatment means were
separated using LSD at a probability level of 5% and 1%, whenever treatments
were significant.

Mean squares
Season: The mean square values showed that the effects of the season were
significant on VLTM (P<0.05), VLThm (P<0.05), VLFM (P<0.05), BNFM
(P<0.01), LNTM (P<0.05), LNThm (P<0.001), LNFM (P<0.001), soil moisture
(P<0.01), PRL (P<0.05), PRW (P<0.001), YAGB (P<0.001) and YBGB (P<0.05).
The effect of season was strong and significant on most growth and biomass
components (Annex table 1 and 2). This could be due to variation in rainfall in the
growing environment (Fig 1). However, the effects of the season was not
significant (P<0.05) on BNTM, BNThM, NoI and LoI.
Medium of Growth: The mean square values depicted that the medium of growth
did not affect significantly (P<0.05) most growth and biomass components of
sweet potatoes grown in net tunnel. However, effects of soil amendments were
significant on soil moisture (P<0.01), YAGB (P<0.01), VLThM (P<0.05),
BNThM (P<0.05), LNThM (P<0.05) and LoI (P<0.05) (Annex table 1 and 2).
[537]
Season X Medium of Growth: The two-way interaction effect of season x
medium of growth were significant on BNFM (P<0.05), LNTM (P<0.05), LNFM
(P<0.05) and LoI (P<0.01). However, the effects of season x medium of growth
were non- significant on growth and biomass yield of early generation sweet
potatoes grown in net tunnel (Annex table 1 and 2).
Effects of season on growth and biomass of sweet potatoes: The LNThM,

LNFM, VLTM, VLThM and LNTM were significantly higher in belg season than
meher season. However, VLFM, NBTM, BNThM, BNFM, PRL, PRW, YAGB
and YBGB were significantly higher in meher season than belg season (Table 2).
Thus plants grown in meher season are more branchy (14.74 to 24.2%), has lesser
leaf number, longer and wider pencil roots besides higher above ground and
below-ground biomass compared to plants grown in belg season. This could be
because meher seasons mark long rains compared to belg season that marks short
rains in the area. Thus, meher season vine production would result in 50.87%
more yields compared to belg season vine production. However, the positive
increment of vine length/plant (15.46%) and leaf number/plant (58.9%) in belg
season with 26, 555kg/ha above ground biomass production showed that sufficient
amount of vines could be produced using belg rains for the supply of farmers and
vine multipliers that require vines in meher season.
Table 1. Means of sweet potato growth and biomass as affected by season
Belg season Meher season % change LSD

VLTM (cm) 73.47 62.11 15.46 6.75*
VLThM (cm) 100.9 87.6 13.18 5.75**
VLFM (cm) 112.3 126.6 -11.30 9.73*
BNTM 8.0 9.9 -19.19 NS
BNThM 13.3 15.6 -14.74 NS
BNFM 15.0 19.8 -24.24 1.71**
LNTM 67.3 27.6 58.99 17.7*
LNThM 125.6 36.0 71.34 16.7**
LNFM 122.8 42.4 65.47 9.8**
PRL (cm) 7.5 12.1 -38.02 2.75*
PRW (cm) 3.4 21.1 -83.89 2.54**
YAGB (kg/ha) 26555 54055 -50.87 10605**
YBGB (kg/ha) 3961.1 7044.4 -43.77 1345*
VLTM- vine length after two months, VLThM-vine length after three months,
VLFM-vine length after four months, BNTM – branch number after two months,
BNThM - branch number after three months, BNFM - branch number after four
months, LNTM – leaf number after two months, LNThM – leaf number after three
months, LNFM-leaf number after four months, YAGB – yield of above-ground
biomass, YBGB-yield of below-ground biomass, PRL-pencil root length, PRW-
pencil root width, ns- denotes presence and absence of significant difference at 5%
level of probability, * and ** denotes significance at 5 and 1% level of probability.
[538]
Soil physical and chemical properties: The soils of the study area are loam in
texture. The macronutrient (N, P, K, Ca, Mg) and micronutrients (Fe, Zn, Na, B,
Cu, Fe, Si, Al, Mn, Mo and Co) contents were also analyzed and reported. Results
of soil analysis after vine harvesting showed that Fe, Zn, Na, B, Cu, Fe, Si and pH
increased significantly due to 50% wood ash treatment compared to other soil
medium. Sawdust produced significantly higher amounts of Co, N, and soil
moisture compared to other medium used to grow sweet potatoes. Coffee husk
produced superior EC and CEC compared to other soil medium amendments.
FYM resulted in high EC, CEC, K, S but intermediate Zn content of the media.
Compost resulted in low pH, CEC, K but intermediate N, P and Ca (Annex table 3
and 4). When pH is considered, there was 8.9% and 1.4% increment in pH due to
the application of wood ash and sawdust whereas there was 9.3, 5.9 and 0.9
decrease in pH due to the use of compost, coffee husk and farmyard manure.
Thus, organic residues used in the study prevented low pH and subsequently poor
plant performance, which would otherwise have caused poor root growth, which
intern is associated with H+ ion injury, toxicities of aluminum (AI) or manganese
(Mn) and deficiencies of calcium (Ca), phosphorus (P) or magnesium (Mg). This
agrees with findings of Sas-Paszt et al. (2014) who noted the best pH and organic
matter results from the use of the compost, cow manure and the mycorrhizal
substrate, whereas the concentrations of P, K, Mg, most of microelements and soil
organic matter were also elevated due to use of organic resources. Results of these
study were in line with Bedada et al (2016) who reported 0.83 mg/kg B, 251
mg/kg Mn, 2.41 mg/kg Cu, 143 mg/kg Fe and 18.1 mg/kg Zn for compost used in
Beseka area of Ethiopia using Mehlich-3 procedure for micro-nutrients. The
micronutrient concentrations were greater compared to untreated plots.
All soil amendments contributed to moisture conservation compared to soil

medium but coffee husk. The use of each of these soil amendments improved the
N content compared to soil medium. Sawdust produced superior N, coffee husk,
compost and farmyard manure were intermediate and wood ash was the least as
far increment in N content was considered. The use of wood ash and farmyard
manure medium increased the P content by 43% and 33.4% (Annex Table 4).
Application of coffee husk, farmyard manure and wood ash increased the EC
content by 169.2%, 146.2% and 76.9% respectively compared to the soil medium;
However, there was 23.1% decrease in EC due to use of sawdust in the soil
amendments. 47.0%, 31.2%, 30.3% and 26.5% increase in CEC was observed by
the end of the second year due to the use of coffee husk, wood ash, sawdust and
farmyard manure medium, respectively. The K content was increased by 47.0%,
45.5% and 35.9% due to the use of wood ash, farmyard manure and coffee husk
medium, respectively (Annex Table 5). The current result is in par with the
finding of Kundu et al. (2006) who reported more moisture, enhanced mineral N
(29-87%), P (1.4-12.6%) and K (16-36%) availability when applied for dry season
sweet potato from soil amendments.
[539]
Growth components
Vine length, branch number and leaf number: The length of vines/branch
showed variation is due to treatment effect since the fourth month. Significantly
longer (P<0.05) vines were measured from plots treated with coffee husk and
farmyard manure. The length of vines was shortest due to the treatment of medium
with wood ash. The number of branches/plant increased as the crop stayed in the
field. Significantly more (P<0.05) number of branches/plant were measured in the
fourth month of the growth of vines from plots treated with compost and farmyard
manure. The numbers of branches/plant recorded were the least from the control
plot. The number of leaves/plant did not show significant variation (P<0.05) in the
second and third months. However, the number of leaves was significantly higher
(P<0.05) due to use of compost and coffee husk medium compared to the control.
The numbers of leaves obtained due to the application of farmyard manure were
intermediate (Table 3). The absence of measurable variation in the second and
third months unlike that of the fourth month could be due to uniformity of soil
amendments in the provision of weed suppression and moisture conservation in
the aforementioned months. Conversely, the statistical difference measured on
vine, branch and leaf traits in the fourth month was due to early mineralization of
nutrients from farmyard manure and coffee husk compared to sawdust and wood
ash.
[540]
Table 3. Effects of soil amendments on vine length/branch, branch number/plant and leaf number/plant as affected by soil amendments over three month growth period
Treatments Vine length/branch (cm) Branch number/plant (cm) Leaf number/plant

Two month Three month Four month Two month three month four month two month three month four month
50% farm yard manure 79.9 96.9 120.0 9.7 15.4 23.43 46.1 81.0 126.3
50% saw dust 66.1 89.7 103.3 8.3 14.4 22.7 41.2 83.0 98.83
50% wood ash 59.5 92.3 95.5 7.7 13.8 19.5 56.3 81.6 122.8
50% compost 69.8 94.7 114.1 9.0 13.8 24.2 46.1 70.3 131.6
50% coffee husk 65.6 100.8 129.1 9.4 14.7 14.7 43.1 78.8 138.6
100% soil medium 65.7 91.3 111.8 9.4 14.8 14.4 51.9 90.0 118.8
LSD ns ns 22.4* ns ns 5.48* ns ns 14.5*
CV (%) 8.3 13.9 16.5 17.5 13.4 17.6 11.9 18.6 15.3
CV – coefficient of variation, LSD – least significant difference number of vines, * and ns- denotes presence and absence of significance difference at 5% level of probability, respectively.
[541]
Number of vines: The number of vines harvested per plot is a function of the
length of vines/branch, the number of branches/plant and the number of
plants/plot. Considering vine length of 30 cm for early generation seeds and
assuming plot size of 10 m2, the growth of vines was slow in the first two months,
increases by two-fold in the third month and reaches its climax by doubling the
figures in the fourth month in kulfo variety. Results showed that a significantly
higher number of vines was obtained from farmyard manure (93,720vines/plot)
and coffee husk (92,040.7vines/plot) medium compared to the control
(53,664vines/plot) (Fig 2). The current finding of coffee husk conformed with
Tenaw (2006) who elaborated integrated use of coffee by-product and nitrogen
fertilizer increased N uptake and water use of maize in Hawasa, Southern
Ethiopia. The superior performance of manure was also due to increased nutrient
uptake efficiency (Teklu and Hailemariam, 2009).
four month three month two month

Vine number (103 /10m2)
Fig 2. Increment in vine yield over four month period for variety Kulfo
Internodes number and length: The number of internodes found in each vine
did not differ statistically due to the change in soil amendments in this experiment.
However, internode length showed significant (P<0.05) variation due to
differences in soil amendments (Table 4). Significantly (P<0.05) longer internodes
were recorded due to the farmyard medium (100.6% greater) compared to the
control. The change in length of internodes due to coffee husk and compost was
32.5% and 27.1%, respectively compared to the control. The increment in
internodes length due to the use of wood ash and sawdust medium was slightly
higher than the control but inferior to that of farmyard manure, compost and
coffee husk.
[542]
Table 4. Effects of soil amendments on growth and biomass yield of sweetpotatoes
Treatments NoI LoI % PRL PRW YAGB % YBGB % change

/plant (cm) change (cm) (cm) (kg/ha) change (kg/ha)
50% Farm yard manure 149.2 3.33 100.6 10.04 2.88 51,166.6 62.01 4625.0 -7.5
50% saw dust 147.1 1.86 12.0 9.58 3.77 27,250.0 -13.72 5083.3 1.7
50% wood ash 150.3 1.83 10.2 10.38 3.39 36,000.0 13.98 5300.0 6.0
50% compost 145.4 2.11 27.1 11.27 2.83 47,333.3 49.87 7383.3 47.7
50% coffee husk 129.3 2.20 32.5 9.71 4.00 48,500.0 53.56 5625.0 12.5
100% soil medium 139.9 1.66 0.0 8.25 3.66 31,583.3 0.00 5000.0 0.0
LSD ns 1.12* ns ns 15712.3* ns
CV (%) 13.8 15.6 19.5 15.6 14.3 7.3
NoI – Number of internodes, LoI – Length of internodes, PRL-pencil root length, PRW- pencil root width, YAGB – yield of
above-ground biomass, YBGB-yield of below-ground biomass, * and ns- denotes presence and absence of significant
difference at 5% level of probability, respectively.
Biomass yield and its components: Pencil root length, pencil root width and
below-ground biomass were not affected significantly (P<0.05) by the difference
in soil amendments considered in this experiment (Table 4and Fig 3). The sweet
potato roots produced in net tunnels were miniature of roots regardless of the soil
amendments, and hence called pencil roots despite growth for three to four
months. The roots did form long but thin storage roots, possess a huge amount of
adventitious lateral roots but did not develop fully because of the deliberate
nutrition, temperatures and spacing factors that favored the growth of above-
ground biomass compared to storage roots (below-ground biomass). This result is
against the findings of Sas-Paszt et al. (2014) who compared different mulching
materials and elaborated that only sawdust mulch significantly increased the
number and size of apple fruit and could be attributed to perennial nature of the
test crop that would benefit from prolonged mineralization of the organic residue.
The above-ground biomass differed significantly (P<0.05) due to the change in

soil amendments. Results showed that significantly higher above-ground biomass
resulted due to farmyard manure medium compared to the control. Results of
above-ground biomass from farmyard manure were comparable to coffee husk and
compost. The percentage increment in above-ground biomass due to farmyard
manure, coffee husk and compost was 62.01%, 53.56% and 49.87%. The
increment in the above-ground biomass was negative due to sawdust medium.
This agrees with numerous authors who explored that the production of tubers is
negatively correlated with that of the leaves and with the total aerial parts
(Pushkaram et a.,l 1976; Enyi 1977; Kamalam et al., 1977). The reduction in root
yield and increment in the production of vegetative parts was due to the
decreasing distance between plants and the use of nitrogenous soil amendments
(Gamboa 1962; Mandal et al., 1971). The current finding agrees with Meniam et
al. (2012) who reported the use of farmyard manure or compost at the time of
preparation of field nursery beds.
[543]
Below ground biomass Above ground biomass
Yield (103 kg/ha)
Fig 3. Bar chart showing proportion of above and below ground biomass

Sweetpotato vine business in Ethiopia is dominated by the informal sector;
demand for quality planting materials of improved varieties is low and
inconsistent. This was mainly due to a lack of demand for quality declared seed
(QDS) from root producers, which was in turn due to lack of awareness, low
willingness to pay, and lack of registered QDS producers. Hence, Hawasa
Agriculture Research Center was multiplying and delivering early generation
seeds of sweet potatoes to commercial vine multipliers. However, soils in Hawasa
area lack many nutrients essential for plant growth. This resulted in complex signs
and symptoms of nutritional disorders observed among growing sweet potato
vines. To combat this constraint, there was a need to identify locally available soil
amendments (farmyard manure, sawdust, wood ash, compost and coffee husk) for
sweet potato vine production in net tunnels. The use of each of these soil
amendments improved the N content compared to soil medium. Results showed
that sawdust produced superior N; coffee husk, compost and farmyard manure
were intermediate and wood ash was the least as far as an increment in N content
was considered. The use of wood ash and farmyard manure increased the P
content by 43% and 33.4%. FYM resulted in high EC, CEC, K and S but
intermediate Zn content. Compost resulted in low pH, CEC and K but
intermediate N, P and Ca. Pencil root length, pencil root width and pencil root
yield were not significantly (P<0.05) different in the experiment. Significantly
higher above-ground biomass (P<0.05) was obtained from farmyard manure,
coffee husk and compost. Significantly taller (P<0.05) vines were measured from
plots that received coffee husk and farmyard manure. Significantly higher leaf
number was recorded from plots that received coffee husk, compost and FYM.
The largest branches were from plots that received compost, FYM and sawdust.
[544]
The best internodes length was obtained from plots treated with FYM. It is
obvious from this study that the practice of using wood ash and sawdust for sweet
potatoes vine multiplication was not conducive as it was associated negatively
with above-ground biomass production and shortened the length of the internodes
of the Kulfo variety. Farmyard manure increased vine yield by 62.01% whereas it
decreased root yield by 7.5%. Coffee husk increased vine production by 53.56%
whereas increased root yield by only 12.5%. This was mainly due to earlier
mineralization, enhanced nutrient uptake efficiencies and moisture use. Hence,
this finding shows that vine multipliers and sweet potato farmers can prepare rapid
multiplication sweet potato nursery to produce sufficient vines to a hectare of land
through the use of either farmyard manure or coffee husk in both seasons in
Hawasa and areas of similar agro-ecologies.
Acknowledgment
Research (EIAR) and Southern Agriculture Research Institute (SARI) for
sponsoring and hosting the experiment, respectively and Sweetpotato Action for
Security and Health in Africa (SASHA) project for funding the soil analysis of
this study.
References
Bedada W, Lemenih M, Karltun E. 2016. Soil nutrient build-up, input interaction effects and plot-
level N and P balances under long term addition of compost and NP fertilizer. Agriculture,
Ecosystems and Environment, 218:220-231
Enyi, B. 1977. Analysis of growth and tuber yield in sweet potato (Ipomoea batatas) cultivars,
Journal of Agricultural Science, 88:42.
FAO. 2008. FAO Statistical Databases. Food and agriculture organization of the United Nations.
http://faostat.fao.org.
FAO. 2011. FAO Statistical Databases. Food and agriculture organization of the United Nations.
http://faostat.fao.org.
Gamboa M. 1962. Estudio sobre fertilizacion y distancias de siembra en el cultivo del camote
(Ipomaea batatas L) en la estacion experimental agricola Fabio Baudrit Moreno Tesis
Ingeniero Agronomo San Jose Universidad de Costa Rica 64p.
Mandal R, Singh K, Maini S and Magoon M. 1971. Response of sweet potato to plant spacing and
nitrogen fertilization, Indian Journal of Agronomy 16:85.
Maniyam N, Gangadharan B and Susantha K. 2012. Sweet Potato Agronomy, Regional Centre of
Central Tuber Crops Research Institute, India.
Meyreles L, Mac L and Preston T. 1977. Cassava forage as a source of protein: effect of
population density and age at cutting, Tropical Animal Production 2:15.
Kamalam P, Biradar R, Hrishi N and Rajendran P. 1977. Path analysis and correlation studies in
sweet potato (Ipomoea batatas Lam), Journal of Root Crops 3:5.
Kundu D, Singh R and Roy C. 2006. Effect of rice straw mulch and irrigation on nutrient
availability in soil and tuber yield of sweet potato (Ipomoea batatas L.) in coastal Orissa. In:
Naskar SK, Nedunchezhiyan M, Rajasekhara Rao K, Sivakumar PS, Ray RC, Misra RS,
Mukherjee A (Eds) Root and Tuber Crops: in Nutrition, Food Security and Sustainable
[545]
Environment, Regional Centre of Central Tuber Crops Research Institute, Bhubaneswar,pp
117-122.
Onunka N, Chukwu E. Mbanasor O and Ebeniro C. 2012. Effect of organic and inorganic manures
and time of application on soil properties and yield of sweet potato in a tropical Ultisol.
Journal of agriculture and social research (JASR) vol. 12, no. 1, 2012.
Pushkaram K, Nair P and Gopakumar K. 1976. Analysis of yield and its components in sweet
potatoes (Ipomoea batatas L) Agricultural, Research Journal of Kerala.
Sas-Paszt L, Pruski KZ, Urawicz E, Sumorok B, Derkowska E and Gluszek S. 2014. The effect of
organic mulches and mycorrhizal substrate on growth, yield and quality of Gold Millennium
apples on M.9 rootstock. Can. J. Plant Sci. 94:281_291.
SPHI. 2014. Sweet Potato Seed Systems Community of Practice Meeting Report, April 28-29,
2014, Kigali, Rwanda.
Srinivasulu R, Lydia N, Kimenye, Margaret M. 2017. Strategies for the development of the sweet
potato early generation seed sector in eastern and Southern Africa, Open Agriculture. 2017 (
2): 236–243.
Snyman H, Jong D and Aveling T. 1998. The stabilization of sewage sludge applied to agricultural
land and the effects on maize seedling, Water science technology, 38 (2): 87-95.
Tenaw Worekayehu. 2006. Effect of coffee residue and cropping system on crop yield and
Physico-chemical properties of the soil in Southern Ethiopia, PhD Thesis, and the University
of Putra Malaysia.
Teklu E and Hailemariam T. 2009. Agronomic and economic efficiency of manure and urea
fertilizers use on vertisols in Ethiopian Highlands. Agricultural sciences in China 8:352-360
Udoh Akinjoba. 2014. Outcome of phosphorus fertilizer on tuber yields vegetative growth and
phosphorus uptake of sweet potato (Ipomoea batatas). International Journal of Manures and
Fertilizers, Vol. 3 (7), pp. 558-560.
[546]
Annex table 1. Mean square of growth and biomass for sweetpotatoes grown in net tunnels under medium of growth experiment
Source of variation Degree of VLTM VLThM VLFM BNTM BNThM BNFM LNTM LNThM
freedom
Season (s) 1 1160* 1590* 1837* 33.2ns 23.6ns 209.7** 14228* 72316***
Error 4 105 76 219 6.1 28.3 6.8 725 651
Medium (m) 5 279ns 99* 641ns 3.5ns 2.2* 12.9ns 191ns 244*
S xM 5 154 ns 173 ns 391 ns 11.0ns 41.4 ns 63* 251* 375ns
Error 20 76 117 57 6.7 47.1 23.2 108 226
CV(%) 8.3 13.9 6.5 17.5 13.4 17.6 11.9 8.6
VLTM- vine length after two months, VLThM-vine length after three months, VLFM-vine length after four months, BNTM – branch number after two months, BNThM - branch number after
three months, BNFM - branch number after four months, LNTM – leaf number after two months, LNThM – leaf number after three months, LNFM-leaf number after four months
Annex able 2. Mean square of growth and biomass for sweetpotatoes grown in net tunnels under medium of growth experiment (Continued)
Degree of Soil
Source of variation LNFM NoI LoI PRL PRW YAGB YBGB EC
freedom moisture
Season (s) 1 58201*** 84235** 812ns 1.6ns 189* 2815*** 6806250000** 85562500** 0.000
Error 4 223 3844 689 0.89ns 17.1 15 261034722 4201180 0.000
Medium (m) 5 222ns 3874** 370ns 0.98ns 5.9ns 27.8ns 599477777* 5750111ns 0.061***
S xM 5 385* 163ns 0.803** 2.2ns 7.8ns 27.7ns 25166666ns 3261000ns 0.000
Error 20 161 883 394 0.17 8.5 38.5 190976388 154115972 0.000
CV (%) 15.4 4.6 13.8 15.6 9.5 25.5 24.5 31.3 1.73
PRL-pencil root length, YAGB – Yield of above ground biomass, YBGB- yield of below ground biomass, NoI – number of internodes, LoI-length of internodes
[547]
Annex table 3. Effects of soil amendments on micronutrients contents of soils
Treatments Fe (mg Mn (mg Zn (mg Mo (mg Co (mg Na (mg B (mg Cu (mg Fe (mg Al (mg Si (mg
/kg) /kg) /kg) /kg) /kg) /kg) /kg) /kg) /kg) /kg) /kg)
50% Farm yard manure 76.1 205.5 18.1 0.123 1.17 123.3 0.05 2.52 76.2 690.1 814.0
50% saw dust 31.8 266.5 14.4 0.130 1.21 107.5 0.32 2.22 76.1 755.6 857.0
50% wood ash 103.0 180.2 23.4 0.133 1.05 256.9 0.65 2.82 104.0 702.1 917.0
50% compost 87.1 198.2 15.1 0.137 1.11 84.7 0.05 2.15 102.7 788.7 845.6
50% coffee husk 78.4 210.2 16.4 0.123 1.15 65.2 0.05 2.46 79.5 725.1 804.3
100% soil medium 75.1 421.4 14.5 0.133 1.13 71.6 0.05 1.84 73.9 795.5 874.0
LSD 34.6** 2.52** 0.38** ns 0.017** 2.49** ns 0.17** 0.87** 7.45** 14.6**
CV (%) 19.2 0.90 1.98 3.72 0.97 1.31 16.3 4.6 0.64 0.62 1.07
Annex table 4. Effects of soil amendments on Macronutrients, EC, CEC, pH and moisture of soils
Treatments pH N P S
H20 % change Soil moisture (%) % change (%) % change (mg/kg) % change (mg/kg) % change
50% Farm yard manure 7.33 -0.9 217.0 17.3 0.19 58.3 535.1 33.4 25.8 59.3
50% saw dust 7.50 1.4 237.5 28.4 0.45 275.0 339.6 -15.3 13.80 -14.8
50% wood ash 8.06 8.9 218.8 18.3 0.13 8.3 573.6 43.0 39.9 146.3
50% compost 6.71 -9.3 190.6 3.0 0.22 83.3 421.3 5.0 22.8 40.7
50% coffee husk 6.96 -5.9 169.6 -8.3 0.24 100.0 389.0 -3.0 20.3 25.3
100% soil medium 7.40 0.0 185.0 0.0 0.12 0.0 401.1 0.0 16.2 0.0
LSD 0.201** 33.8* 0.11** 4.59** 0.67**
CV (%) 2.3 14.6 1.3 0.92 2.53
[548]
Annex table 5. Effects of soil amendments on Macronutrients, EC, CEC, pH and moisture of soils
Treatments EC (mS CEC (Meq/ K (mg Ca (mg Mg (mg

/cm) % change 100g soil) % change /kg) % change /kg) % change /kg) % change
50% Farm yard manure 0.32 146.2 24.08 26.5 2709.6 45.5 3410.3 -3.6 558.1 32.5
50% saw dust 0.10 -23.1 24.81 30.3 1872.6 0.6 3567.6 0.9 431.0 2.3
50% wood ash 0.23 76.9 24.98 31.2 2738.3 47.0 4207.3 18.9 887.3 110.6
50% compost 0.17 30.8 19.07 0.2 1751.6 -5.9 3681.0 4.1 425.3 0.9
50% coffee husk 0.35 169.2 27.99 47.0 2530.0 35.9 3527.3 -0.3 463.3 10.0
100% soil medium 0.13 0.0 19.04 0.0 1862.3 0.0 3537.3 0.0 421.3 0.0
LSD 0.104** 0.095** 33.4** 44.04** 3.37**
CV (%) 1.72 0.36 0.95 0.75 0.56
[549]
[550]
Evaluation of Cassava (Manihot esculenta Crantz)
Genotypes for Total Cyanide Content, Storage
Tuber and Starch Yield in South-Western Ethiopia
Tewodros Mulualem1*, Neim Semman1 and Getachew Etana1
1
Jimma Agricultural Research Center, P.O. Box 192, Jimma, Ethiopia
*
Corresponding author: tewodrosmulualem@gmail.com
Abstract
Cassava (Manihot esculenta) is one of the main food crops that significantly contribute
to food security and poverty alleviation in South and Southwest Ethiopia. Cyanide
present in cassava roots poses a health challenge in the use of cassava for food. It is
therefore important to identify the sweet and bitter types for food and industry. In this
study, 11 farmers preferred genotypes were assessed for storage tuber yield, cyanide
content and starch contents at Jimma, Metu and Tepi during 2015-2017 cropping
seasons. From the genotypes analyzed, AAGT-108 (128.35 mg/kg) and AAGT-189
(118.5 mg/kg) are identified for bitter type (high HCN content). Based on storage tuber
yield, genotype AAGT-108, AAGT-189 and AAGT 192 produced the highest tuber yield
in all tested locations. The starch content varies from 5.11 (Qulle at Metu) to 18.40
(AAGT-108 at Tepi). Similarly, genotypes AAGT-108, AAGT-189 and AAGT-192 are the
best materials at Jimma as compared to others and gave the mean starch content of
15.22, 16.98 and 16.61 %, respectively. Based on the overall result of hydrogen cyanide
and starch contents, genotype AAGT-108 and AAGT-189 are recommended for bitter
type (high HCN content). On the contrary, the rest of the genotypes have sweet type due
to low hydrogen cyanide content below 100 mg/kg and are recommended for human
consumption.
Keywords: Cassava, cyanide, starch, storage tuber, yield
Introduction
Cassava (Manihot esculenta Crantz,) is the sixth most important food crop
globally, in terms of annual production (FAOSTAT, 2010). The crop is mainly
grown in the tropics, including sub-Saharan Africa, Asia, the Pacific Islands, and
Central and South America (Lebot, 2009) for staple food of an estimated 800
million people worldwide (FASTAT, 2013). The crop can survive adverse
conditions such as infertile soil, drought, pests and diseases (El-Sharkawy, 1993;
Bokanga, 1999) and plays several important roles in Africa serving as a rural
staple food, famine-reserve crop, cash crop for households and as a raw material
for feed and industrial manufacturing (Nweke et. al., 2002). It has been identified
as a potentially valuable source of food for addressing food security in developing
countries (Montagnac et al., 2009) and a major source of food calories for about
two of every five Africans (Serge et al., 2013). Although reliable statistical
information on the distribution and production of cassava in Ethiopia is lacking,
[551]
the crop has been cultivated, particularly, in the South, South West, and Western
parts to overcome hunger and make a significant contribution in the diets of the
people (Tewodros and Zelalem., 2015).
The starchy tuberous roots are the main food source, which are a high amount of
carbohydrates, fibers, and low-level fats and protein, a good proportion of
essential amino acids that make them a good dietary source (Montagnac et. al.,
2009). Conversely, the wider utilization of cassava in Ethiopia is inadequate; due
to information on the biochemical composition of cassava is meager. Further,
limited exposure to high doses of hydrogen cyanide is a health risk in major
growing regions where cassava is a staple food source (Getachew et. al., 2012).
Besides, cassava in itself is not a balanced food and malnutrition occurs when
cassava is consumed alone as a staple food. Studies of nutritional composition on
cassava as a food are considerable significance since it may help to identify long-
forgotten food resources (Tewodros and Biruk, 2012). In this regards, few attempt
was made to understand the proximate composition and anti-nutritional factors of
the underutilized tubers of cassava to make edible tubers as the safe food sources
for mass consumption (Tesfaye et. al., 2017). In contrast to cultivated tubers, little
is known about the proximate composition and reasons to expect that some of the
species differ in composition from common varieties. Furthermore, several species
of cassava also have two cyanogenic glycosides, linamarin and a small amount of
lotaustralin, which are catalytically hydrolyzed to release toxic hydrogen cyanide
(HCN) toxic to human beings.
The level of total cyanide contents of different varieties and plant parts of cassava
is 1–1550 mg HCN equivalents/kg fresh material = ppm, in the root parenchyma
and 900–2000 ppm in the root cortex (peel) (Cardoso et al., 2005). Cassava leaves
contain 20–1860 ppm of total cyanide (Bradbury and Denton, 2011). The World
Health Organization (WHO) set a safe limit of 10 ppm total cyanide for cassava
flour (FAO/WHO, 1995). This maximum limit has been adopted in different
countries for example, 10 ppm for cassava chips by Food Standards Australia and
New Zealand (Fsanz, 2009), while 40 ppm is the limit in Indonesia (Djazuli and
Bradbury, 1999). Internationally, the Codex Standard for ‘sweet cassava’ (those
varieties with low levels of cyanogens) is 50 ppm (fresh weight basis, FAO/WHO,
2005), but many countries have yet to formally adopt recommended limits
(Kolind-Hansen and Brimer, 2009).
Despite its food security and industrial importance, there are no efforts so far done
in the nutritional composition; industrial value and information on the biochemical
composition of cassava genotypes are scarce. Furthermore, the glycoside attributes
of the existing genotypes have never been assessed and the level of anti-nutritional
factors on cassava at the country level is still unknown; which hinders the wider
utilization and researchers to access the biochemical composition of cassava in the
[552]
country. Cognizant of these facts, the present study was designed to evaluate the
yield, hydrogen cyanide and starch contents of cassava genotypes collected from
Southwest Ethiopia for use and conservation.

Study areas: A field studies were conducted during the 2015/16 and 2016/17
growing seasons at Jimma, Tepi and Metu agricultural research centers. The detail
descriptions of tested sites are presented (Table 1).
Table 1. The geographical description of the study sites

Location Altitude Latitude Longitude Rainfall Temperature (0C)
(m.a.s.l.) (mm) Maximum Minimum
Jimma 1753 7o 40.00' N 36o 47’.00’ E 1521.1 26.2 12.1
Metu 1550 8°18′ .00' N 35°35′ .00’ E 1520 28.0 12.2
Tepi 1200 7° 3' .00' N 35° 18' .00’E 1685.9 29.9 15.4
Source: JARC, 2010
Plant materials, experimental design and management: A total of 11 cassava

genotypes were collected from major growing areas of Southwest Ethiopia. The
experiment was laid out in a randomized complete block design (RCBD) with
three replications. The gross plot size for each treatment was 4m x 4m, using inter-
row spacing of 1.5m and intra-rows spacing of 1m. Cuttings of the same size and
age were used as planting material. One month after planting, seedlings were
earthed up followed by frequent weeding. All other agronomic practices were
followed according to the recommendations.
Data collection: Data were collected from six plants from each plot and the
average values were used for data analysis. The characters that are used for data
collection were: tuber length (cm), tuber fresh weight (t/ha), hydrogen cyanide
content (mg/kg fresh storage sample) and starch content (% fresh storage sample)
and tuber dry weight (t/ha).
Determination of starch content: The starch content was determined by a

modified method of Asaoka et al. (1992). A cassava storage tuber was collected
from tested materials from each location.500 g of fresh storage tuber was taken
from the distal, middle and apical sections of washed tubers from each of the
middle four plants randomly. After peeling, the tuber cut into small pieces with a
stainless steel knife and then milled using an electric grinder, to obtain fine
powder cassava flour. The starch particle was isolated using excess pure water.
The slurry was filtered through a 1mm sieve mesh. The residue was washed three
times with 500 ml of water each time to remove remnants of starch. The filtered
was allowed to settle for 2 hours for each wash before decanting the liquid. Hence,
[553]
starch was suspended and non-starch materials removed by decanting the
supernatant. Subsequently, the starch was dried in an oven at 30-33ºC until
constant weight attained and dried. The amount of dried starch obtained from one
kg of fresh cassava tuberous roots was weighed and expressed as a percent of the
fresh storage tuberous.
Determination of hydrogen cyanide Content: Determination of hydrogen

cyanide contents (HCN mg/Kg). The duplicate samples were harvested from the
fields and transported immediately to the laboratory (Food and nutritional
laboratory Addis Ababa). The collected samples were cleaned, peeled and washed
with potable water. Samples from the distal, middle and apical sections of peeled
tubers were cut into cube. The acid titration method (AOAC, 2005) for the
determination of hydrocyanic acid in beans was used. One hundred (100) ml of
H2O was added to 25 g of the sample in a 500 ml Kjedahl flask for steam
distillation. The distillate was collected in 20 ml 0.02N. AgNO3 acidified with 1
ml HNO3. The apparatus was adjusted so that the tip of the condenser dipped
below the surface of the liquid in the receiver. After 150 ml had passed over,
excess AgNO3 was titrated with 0.02 KSCN using Fe alum an indicator. The
results were calculated as fresh matter basis as follows:
100/W ×0.27)] × [(VAgNO3 = sample HCN/100g mg

1 ml of 0.01 N AgNO3 = 0.27 mg of HCN
Where: VAgNO3 Volume of silver nitrate = [(20-(2×V KSCN))]
V KSCN Volume of potassium thiocyanate consumed in and W: weight of sample
Data analysis: The collected data were analyzed by using SAS statistical
Software package (version 9.0 of SAS Institute Inc, 2000). Both quantitative and
qualitative data were subjected to analysis of variance (ANOVA) using the RCBD
procedure as suggested by Gomez and Gomez (1984). Means were separated
using the Least Significant Difference (LSD) procedure at the 5% and 1% level of
significance.

Mean performance of Storage tuber and length: The mean storage tuber yield
(t/ha), storage tuber length (cm) hydrogen cyanide (mg/kg fresh tuber weight) and
starch contents (%) of tested cassava genotypes and locations are presented in
Table 1. The result of the study revealed that there were significant (p<0.05)
differences among cassava genotypes collected from southwest Ethiopia. Based on
the combined mean performance of Jimma, genotype AAGT-108 (70.54 t/ha),
AAGT-189 (61.49 t/ha) and AAGT 192 (61.45 t/ha) were produced the highest
storage tuber yield. Similarly, genotype AAGT-108, AAGT-192, AAGT-028 and
AAGT-189 gave the highest performance at Metu and Tepi with storage tuber
[554]
yield of 77.8 t/ha, 76.4 t/ha, 74.15 t/ha and 58.6 t/ha, and 78.55, 66.78, 72.99, and
77.59 t/ha, respectively. Likewise, storage tuber length of tested genotypes and
locations showed similar trends as storage tuber yield. Genotypes AAGT-189,
AAGT-108 and AAGT-191 produced the longest storage tuber with 56.75, 52.40
and 52.90 cm, respectively. However, released varieties Qulle and Kello provided
the longest storage tube at Metu and Tepi. Based on overall performance
genotypes AAGT-028, AAGT-189 and AAGT-192 are the best-performed
genotypes in all tested locations.
Mean performance of hydrogen cyanide and starch content: The mean

hydrogen cyanide and starch contents are important quality parameters in cassava
breeding. In this study, genotypes AAGT-108, AAGT-189 and AAGT-192
produced the highest mean hydrogen cyanide contents with the value of 105.35,
118.5 and 11.7 mg/kg of fresh tuber for Jimma, 71.05, 75.50 and 76.05 mg/kg
fresh sample for Metu and 128.35, 88.35 and 76.25 mg/kg for Tepi, respectively.
The starch contents of established genotypes with tested locations are presented in
Table 2. At Jimma, genotypes AAGT-108, AAGT-189 and AAGT-192 are the
best materials as compared to others and gave the mean starch content of 15.22,
16.98 and 16.61 % per one kg. of fresh storage tuber cassava sample. Similarly,
the genotype performed well at Metu and Tepi with a value of 10.18, 10.83 and
10.91 % at Metu and 18.4, 12.7 and 15.65% per kilogram of fresh storage tuber
cassava sample, respectively.
Discussion
In the present study, the mean storage tuber yield showed highly significant
differences (p<0.05) among cassava genotypes from southwest Ethiopia, this
suggested, the presence of a high degree of genetic variability in the materials
evaluated and the existence of considerable genetic diversity among cassava
genotypes for selection. The result of this study is similar to the report of
Tewodros and Getachew, (2013) who reported cassava genotypes collected from
southwest Ethiopia had significant difference storage tuber yield and related traits.
Similarly, Tesfaye et al., (2017) also reported a high significant (p<0.01)
difference among cassava genotypes tested at Hawassa, Amaro, Jima and Sekota
areas of Ethiopia. The storage tuber length also varies significantly (p<0.05)
among tested cassava genotypes. The longest tuber length is obtained from
genotypes AAGT 189, Qulle, AAGT 108 at Jimma, Metu and Tepi with values of
56.75, 53.10 and 40.40 cm, respectively. The length of the tuber is highly affected
by the soil texture of cassava grown. This result is supported by the report of
Tewodros and Yared, (2015), who reported the storage tuber length of cassava
grown in clay soil and high moisture stress areas of southern Ethiopia are reduced
significantly. Besides, Tesfaye et al, (2017) reported variety Kello produced
longest mean storage tuber of 40.67 cm grown in major growing areas of Ethiopia.
[555]
However, in this study, the storage tuber length obtained from Jimma (48.90) and
Metu (53.10 cm) is higher than the report of Tesfaye et al, (2017).
The hydrogen cyanide and starch content are the important quality parameter in
cassava breeding. According to Montagnac et al., (2009) and André et al., (2016),
cassava genotypes are classified in two groups, according to the difference on
toxicity and palatability of the storage tubers: sweet (slightly sweet; that is, plants
that have less than 100 mg/kg cyanuric acid) or bitter (those have a perceptible
bitter taste due to the high concentrations of HCN, that is, values over 100 mg/kg
fresh sample). There is no morphological characteristic that may tell these two
groups apart (Dufour et. al., 1988). In the present study, cassava genotypes AAGT
189, AAGT 191 and AAGT 108 produced the highest cyanogenic
glycosidecyanide content with a value of 118.5, 88.9 and 128.35 mg/kg fresh
sample, respectively. The result obtained from this study was similar to the report
of André et al., (2016), who reported, the cyanide contents ranged from 15-225
mg/kg of cassava collected from Brazil. On contrarily, the result obtained from
this study was higher than the report of Ezeigbo et al., (2015) reported the cyanide
contents of cassava genotypes ranged from 36.65-62.57 grown in Abia state of
Nigeria. Similarly, the mean starch contents of cassava genotypes ranged from
5.11-18.4%. The lowest starch content obtained from genotype Qulle at Metu and
the highest mean starch content was collected from genotype AAGT 108 at Tepi.
The result obtained from this study was lower than the study of Muleta and
Mohammed, (2017) who reported the starch contents ranged from 62-91%.
Similarly, the starch content obtained from in this study almost similar to the study
of Ezeigbo et al, (2015) reported the starch contents of five cassava genotypes
from Nigeria ranged from 17.48-20.62%.
[556]
Table 1. Mean fresh storage tuber yield (t/ha) and tuber length (cm) of cassava genotypes of in tested locations and years
Genotypes Storage tuber yield (t/ha) Storage tuber length (cm)

Jimma Metu Tepi Jimma Metu Tepi
2015 2017 Mean 2015 2017 Mean 2015 2017 Mean 2015 2017 Mean 2015 2017 Mean 2015 2017 Mean
Qulle 68.47 62.8 65.64 62.9 34.6 48.75 71.88 26.06 48.97 41.7 56.1 48.90 55.0 51.2 53.10 35.1 38.7 36.90
Kello 48.01 44.4 46.21 75.7 34.4 55.05 77.71 31.18 54.45 44.0 53.4 48.70 47.3 43.4 45.35 30.9 38.8 34.85
45/72NR 62.01 54.9 58.46 53.3 41.3 47.30 37.92 77.60 57.76 45.0 45.7 45.35 46.0 34.8 40.40 31.8 34.0 32.90
45/72NW 51.41 54.5 52.96 54.1 58.5 56.30 77.46 75.75 76.61 43.3 53.5 48.40 39.7 35.8 37.75 32.5 40.5 36.50
AAGT 028 38.50 39.9 39.20 80.1 68.2 74.15 68.42 77.56 72.99 47.3 46.3 46.80 39.3 36.5 37.90 27.6 41.3 34.45
AAGT 108 73.38 67.7 70.54 78.4 77.2 77.80 79.59 77.51 78.55 51.0 53.8 52.40 41.7 44.2 42.95 28.9 51.9 40.40
AAGT 189 58.48 64.5 61.49 45.1 72.1 58.60 76.71 78.46 77.59 58.0 55.5 56.75 35.0 52.2 43.60 29.7 37.8 33.75
AAGT 191 47.38 52.2 49.79 65.4 78.9 72.15 64.58 53.99 59.29 56.0 49.8 52.90 40.7 38.5 39.60 27.1 31.9 29.50
AAGT 192 74.40 48.5 61.45 80.6 72.2 76.40 51.88 81.67 66.78 47.0 49.4 48.20 44.7 46.0 45.35 31.1 37.9 34.50
AAGT 200 52.50 67.3 59.90 74.2 62.1 68.15 64.84 50.48 57.66 61.0 45.4 53.20 35.0 41.9 38.45 30.9 34.5 32.70
Local 40.88 64.1 52.49 64.3 52.0 58.15 70.46 74.77 72.62 53.3 45.1 49.20 43.0 43.1 43.05 28.1 33.4 30.75
Mean 55.95 56.44 56.20 66.7 59.2 62.95 67.4 64.1 65.75 49.8 50.3 50.05 42.5 42.5 42.50 30.3 38.2 34.25
LSD (5%) 7.17 19.1 12.71 37.2 21.1 11.87 17.18 25.67 17.93 19.5 13.7 8.71 16.2 11.8 9.14 6.2 8.6 7.41
[557]
Table 2. The mean hydrogen cyanide content (HCN mg/kg) and starch content (%) from one kg. fresh tuber sample of cassava genotypes over locations and years
Genotypes Hydrogen cyanide contents Starch content (%)

Jimma Metu Tepi Jimma Metu Tepi
2015 2017 Mean 2015 2017 Mean 2015 2017 Mean 2015 2017 Mean 2015 2017 Mean 2015 2017 Mean
Qulle 51.10 52.73 65.64 62.9 34.6 48.75 71.88 26.06 48.97 41.7 56.1 48.90 55.0 51.2 53.10 35.1 38.7 36.90
Kello 55.00 42.87 46.21 75.7 34.4 55.05 77.71 31.18 54.45 44.0 53.4 48.70 47.3 43.4 45.35 30.9 38.8 34.85
45/72NR 77.90 56.70 58.46 53.3 41.3 47.30 37.92 77.60 57.76 45.0 45.7 45.35 46.0 34.8 40.40 31.8 34.0 32.90
45/72NW 59.30 60.30 52.96 54.1 58.5 56.30 77.46 75.75 76.61 43.3 53.5 48.40 39.7 35.8 37.75 32.5 40.5 36.50
AAGT 028 87.50 67.50 39.20 80.1 68.2 74.15 68.42 77.56 72.99 47.3 46.3 46.80 39.3 36.5 37.90 27.6 41.3 34.45
AAGT 108 113.1 98.75 70.54 78.4 77.2 77.80 79.59 77.51 78.55 51.0 53.8 52.40 41.7 44.2 42.95 28.9 51.9 40.40
AAGT 189 128.6 108.4 61.49 45.1 72.1 58.60 76.71 78.46 77.59 58.0 55.5 56.75 35.0 52.2 43.60 29.7 37.8 33.75
AAGT 191 66.00 59.20 49.79 65.4 78.9 72.15 64.58 53.99 59.29 56.0 49.8 52.90 40.7 38.5 39.60 27.1 31.9 29.50
AAGT 192 118.6 112.8 61.45 80.6 72.2 76.40 51.88 81.67 66.78 47.0 49.4 48.20 44.7 46.0 45.35 31.1 37.9 34.50
AAGT 200 66.90 44.70 59.90 74.2 62.1 68.15 64.84 50.48 57.66 61.0 45.4 53.20 35.0 41.9 38.45 30.9 34.5 32.70
Local 47.90 51.10 52.49 64.3 52.0 58.15 70.46 74.77 72.62 53.3 45.1 49.20 43.0 43.1 43.05 28.1 33.4 30.75
Mean 79.26 68.65 56.20 66.7 59.2 62.95 67.4 64.1 65.75 49.8 50.3 50.05 42.5 42.5 42.50 30.3 38.2 34.25
LSD (5%) 2.3 1.1 12.71 37.2 21.1 11.87 17.18 25.67 17.93 19.5 13.7 8.71 16.2 11.8 9.14 6.2 8.6 7.41
[558]
The yields of cassava are highly affected by genotype and location (environment)
which indicated the yields and quality performance of genotypes vary from one
location to another. Similarly, the hydrogen cyanide content also varied from one
location to another within and among genotypes. Based on the overall mean result
of hydrogen cyanide and starch contents, genotype AAGT-108 (128.35mg/kg) and
AAGT-189 (118.5 mg/kg) are recommended for bitter type (high HCN content).
As a result, the National Variety Release Committee officially released genotype
AAGT-108 (Melko-108) as a bitter type for production. On the contrary, the rest
of the genotypes have sweet types due to their cyanide content below 100 mg/kg
and are recommended for human consumption after full processing.
Acknowledgments
Research/Jimma Agricultural Research Center (JARC) for the financial support of
this study. The Authors also extend their acknowledgment for the collaborating
agricultural research center (Tepi) for their implementation of national variety and
eventually the verification trial.
References
André RF, Maria AN, Almir JR, Caroline P, Jeandro A1, Tamires B, Rosilene K, Luis A.B.L, and
Helenice DSS. 2016. Cassava Flour: Quantification of Cyanide Content. Food and Nutrition
Sciences,7, 592-599
AOAC 2005. Association of Official Agricultural Chemist, Official Methods of Analysis, 17thedn,
Vol. 1 and 2. Horowitz edition intern, Maryland, USA. Washington DC. USA.
Ashoka MJ and Richard JE.1992. Effects of cultivar and growth stage on the gelatinization
properties of cassava (Manihot esculanta Cratz) starch. J.Sci. Agri. 59: 53-58
Bokanga M. 1999. Cassava: Post-harvest operations. Information Network on Post-Harvest
Operations.In Post-Harvest Compendium, pp.1-26.
Bradbury JH, Denton IC. 2011. Mild methods of processing cassava leaves to remove cyanogens
and conserve key nutrients. Food Chemistry (in press)
Cardoso AP, Mirione E, Ernesto M, Massaza F, Cliff J, Haque MR and Bradbury JH. 2005.
Processing of cassava roots to remove cyanogens. Journal of Food Composition and Analysis,
18, 451-460.
Djazuli M and Bradbury JH. 1999. Cyanogen content of cassava roots and flour in Indonesia.
Food Chemistry, 65, 523-525.
Dufour DL. 1988 Cyanide Content of Cassava (Manihot esculenta, Euphorbiaceae) Cultivars Used
by Tukanoan Indians in Northwest Amazonia. Economic Botany, 42, 255-266.
http://dx.doi.org/10.1007/BF02858929
El-Sharkawy MA. 1993. Drought-tolerant cassava for Africa, Asia, and Latin America, Bioscience,
43(7): 441-451.
Ezeigbo OR, Ukabi CF, Ike-Amadi A and Ekaiko MU. 2015. Determination of Starch and Cyanide
Contents of Different Species of Fresh Cassava Tuber in Abia State, Nigeria. British
Biotechnology Journal, 6(1): 10-15.
[559]
FAO/WHO. 1995. Codex standard for edible cassava flour. Codex Standard 176-1989. Rome,
Italy: Food and Agriculture Organisation and World Health Organisation of the United
Nations.
FAO/WHO. 2005. Codex standard for sweet cassava. Codex Standard 238-2003. Rome: Food and
Agriculture Organisation and World Health Organisation of the United Nations
FAO. 2010. The world cassava economy: facts, trends and outlooks. Rome: Food and Agriculture
Organization of the United Nations and International Fund for Agricultural Development
FAO (Food And Agriculture Organization of The United Nation). 2013. Save and grow: Cassava.
A guide to sustainable production intensification.
Fsanz. 2009. Proposal P1002Hydrocyanic Acid in Ready-to-Eat Cassava Chips, First Review
Report. Food Standards Australia New Zealand.
Getachew W, Tewodros M and Taye K. 2012. Postharvest handling utilization and food
preparation of cassava in Amharic Version. Ethiopian Institute of Agricultural Research,
Gomez KA and Gomez AA. 1984. Statistical Procedures for Agricultural Research. 2 nd ed. John
Wiley and Sons, inc., New York
Kolind-H, L and Brimer L. 2009. The retail market for fresh cassava root tubers in the European
Union (EU): the case of Copenhagen, Denmark - a chemical food safety issue? Journal of the
Science of Food and Agriculture, 90, 252-256.
Lebot V. 2009. Tropical root and tuber crops: cassava, sweet potato, yams and aroids. Wallingford,
UK: CABI.
Montagnac J, Davis C and Tanumihardjo S. 2009.Processing techniques to reduce toxicity and
anti-nutrients of cassava for use as a staple food. Comprehensive review in food science and
food safety. 8:1-8.
Nhassico D, Muquingue H, Cliff J, Cumbana A and Bradbury JH. 2008. Rising African cassava
production, diseases due to high cyanide intake and control measures. Journal of the Science
of Food and Agriculture, 88, 2043-2049.
Nweke F, Spencer D and Lynam J. 2002.The cassava transformation: Africas best kept secret.
Muleta F and Mohammed A. 2017. Determining of cyanide concentration levels in different
cassava varieties in selected iodine deficiency disorded (IDD) areas of wolaita zone, southern
Ethiopia. Journal of natural sciences research.7:3.
Serge S, Traore S. and Brou K. 2013.Assessment of cyanide content in cassava (manihot
esculenta crantz) varieties and derived products from Senegal.International journal of nutrition
and food sciences 2:5, 225-231
SAS Institute. 2000. Statistical Analytical Systems SAS / STAT user’s guide version, 8(2) cary NC
:SAS institute inc.
Tadesse T, Bekele A, Tsegaye E, Getachew W, Tewodros M, Beshir W and Messele G.
2017. Performance of cassava (Manihot esculanta Cratz) clones in potential and low moisture
stressed areas of Ethiopia. African Journal of Agricultural Research, 12(20): 1738-1746.
Tewodros M. and Biruke A. 2012. Cassava (Mannihot esculenta Cranz) Varieties and Harvesting
Stages Influenced Yield and Yield Related Components. IISTE, Journal of Natural Science
Research,. 2.(10):122-128.
Tewodros M. and Getachew W. 2013. Evaluation of the Adaptability and Acceptability of
improved Cassava (Manihot esculenta Crantz) varieties in Southwest Ethiopia. Greener
Journal of Agricultural Sciences, 3 (8): 658-662,
Tewodros M. and Yared D. 2015.Farmers appraisal, manifestation and scaling up of improved
cassava technologies in moisture stressed areas of the Southern Ethiopia. Journal of Genetic
and Environmental Resources Conservation, 2015, 3(1):100-105
Tewodros Mulualem and Zelalem Bekeko. 2015. Assessment of conventional breeding on cassava
and its physiological adaptive mechanisms: Implication for moisture stress. Asian Journal of
Agricultural Research, 9(2): 38-54
[560]
Spices
Chemical Composition of Essential Oils from

Released Coriander Variety
(Coriandrum sativum L.) Grown In Ethiopia
Biruk Hirko1, Sileshi Abera1, Haimanot, Mitiku2 and Belay Gezahagn1
1
Ethiopian Institute of Agricultural Research, Tepi Agricultural Research Centre P. O. Box 34, Tepi, Ethiopia
2
Ethiopian Institute of Agricultural Research, Debrezeit Agricultural Research Centre,
P. O. Box 32, Debrezeit, Ethiopia
Abstract
The objective of this study was to determine the chemical composition of two coriander
variety (Indium and Dinqnesh) produced in Ethiopia. The essential oil content was
obtained by hydro distillation method with Clevenger apparatus. The chemical profiles
of the two varieties were identified using GC-MS. The essential oil content of the dried
seeds was 0.60% for Dinqnesh and 0.50% for Indium varieties, respectively. About
thirty-four and fifty-one different compounds were identified from the essential oil of
Indium and Dinqnesh varieties, respectively. Linalool was found to be the principal
constituent for both varieties. The variety Indium had higher linalool content (76.45%)
than Dinqnesh (58.85%). Other major components identified were Geranyl acetate
(5.60% for Indium and 4.04% for Dinqnesh variety), γ-terpinene (6.76% for Dinqnesh
variety and 2.83% for Indium) and α-pinene (6.15% for Dinqnesh and 4.90% for
Indium) were identified. Bicycle [2,2,2] heptan-2one,1,7,7-trimethyl-, (1S) and
Benzene, 1-methyl-3-(1-methylethyl) were a compounds found only in Dinqnesh
Variety and yielded 6.54% and 1.28%, respectively. By comparing the chemical
composition of both variety essential oil constituents to other origin or standard
requirements, both varieties meet the potential for nutritional and export standard
requirements.
Keywords: Coriander, essential oil, chemical composition, quality
Introduction
The flavour and aroma of produced from coriander seed is mainly due to its
essential oil constituents. Coriander seed oil is found among the 20 major essential
oils (Lawrence, 1992) and its commercial need depends on its physical
characteristics, chemical constituents and aroma (Smallfield et al., 2001). Linalool
is the principal volatile compound found in seeds, containing more than 50% of
the total essential oil (Ramadan and Morsel, 2003). Coriander is an economically
important spice, due to its biological active compounds found in it. The
accumulation and chemical composition of essential oil in plants are governed by
different factors: environmental (Rakic and Johnson, 2002; Sriti et al. 2011),
genetic (Zheljazkov et al., 2008; Ebrahimi et al., 2010) as well as cultivation
practices (Zheljazkov et al., 2008).
[561]
Since the chemical compositions of essential oils are the basis for their
exploitation, more attention has to be paid to the variation in the constituents.
Coriander has been cultivated in different agro-ecology of Ethiopia. The essential
oil compositions of coriander fruit samples from Bale and Gondar regions of
Ethiopia is reported by Nigist and Birhanu (1998). So far two coriander varieties
namely Indium and Dinqnesh which released by Tepi and Debrezeit Agricultural
Research Centers collaboratively; their essential oil compositions is not yet
studied. Therefore, this study was conducted to determine the essential oil yield
and compositions of the two released varieties of coriander.
Materials and Method

Sample collection: The seeds of coriander varieties were collected from Kulumsa
Agricultural Research Centre during January, 2018 and were transported to Tepi
Agricultural Research Centre for the essential oil extraction.
Sample extraction: Essential oil extraction from the coriander seed was
performed by the Hydro-distillation in the Clevenger apparatus. 100 g seeds of
each powder of coriander varieties with three replicates were set separately in
round bottom flask of 2000 ml in which 1000 ml distillate water was added. The
process of hydro-distillation lasted for four consecutive hours and oil obtained was
stored in a dark glass container until the time of chromatographic separation.
Gas chromatography mass spectrometry (GC-MS) analysis: GC-MS analysis

of essential oil composition was identified at Jije analytical testing service
laboratory. Essential oil compound identification was depending on the
comparison of their mass spectra with data of CG-EM (Nist 62 lib.) (Mclafferty
and Stauffer, 1989); and a retention index of kovats (Adams, 1995). The analysis
conditions had a chromatographic system 7890B GC with split/splitless mode
inlet; equipped with DB-5MS capillary column (30 m x 0.25 mm x 0.25 mm) and
detector 5977A MS. The GC experimental conditions functioned with an inlet
temperature of 260oC with an injection volume of 1μl; column flow of carrier gas
Helium (1ml/min) and with a programmed oven temperature of 40°C hold for 3
minutes, then 4°C/minutes to 90°C hold for 3 minutes by 4oC/minutes to 170 °C
hold for 3 minutes, at 6°C/minutes to 230oC hold for 4 minutes, 10°C/minutes to
270°C hold for 1 minute. The total running time is 46.5 minutes.

Essential oil yield (%): Hydro-distillation of coriander seeds resulted in 0.60%
and 0.50% Essential oil for Dinqnesh and Indium varieties, respectively. The
result obtained in this study is comparable with many authors report of oil yield
between 0.2 and 1.3% among 36 different Coriander accessions from Austria
[562]
(Dobos and Novak, 2005); between 0.1 to 0.36 % for Iranian accessions (Nejad et
al., 2011); less than 1% (Telci et al., 2006) and between 0.5 to 2.5% (Mahendra
and Bisht, 2011). Contradictorily, Zheljazkova et al. (2008) reported the essential
oil content 0.8 to 2.2% from different cultivars of coriander native to Atlantic
Canada which is greater than the result obtained in this study.
Chemical composition of oil: A total of 51 and 34 compounds were identified

from Dinqnesh and Indium varieties, respectively and presented in table 1. The
essential oil tested contained high amount of linalool (76.45% for Indium variety
and 58.85% for Dinqnesh variety followed by α-pinene (6.15% for Dinqnesh
variety and 4.90% for Indium), γ-terpinene (6.76% for Dinqnesh variety and
2.83% for Indium), geraniol (4.04% for Dinqnesh and 1.22% for Indium), and
geranyl acetate (5.60% for Indium and 4.04% for Dinqnesh variety).
Table.1. Chemical composition of two varieties essential oils of coriander (coriandrum sativum L.) seeds
S/n Name of compounds Chemical composition (%)

Dinqnesh Indium
1 3-carene 0.029 -
2 Bicyclo[3,1,0]hex-2-ene, 2-methyl-5-(1-methylethyl)- 0.043 -
3 α-pinene 6.146 4.903
4 Camphene 1.073 0.211
5 Bicyclo[3,1,0]hexane, 4-methyl-1-(1-methylethyl)- 0.335 0.132
6 β- pinene 0.479 0.461
7 β-myrcene 0.906 0.314
8 α-phellandrene 0.026 -
9 α-terpinene 0.068 -
10 Benzene, 1-methyl-3-(1-methylethyl)- 1.286 -
11 D-limonene 3.034 0.723
12 Eucalyptol 0.049 -
13 β-cis-Ocimene 0.033 0.026
14 γ-terpinene 6.761 2.827
15 Cis-linalooloxide 0.127 0.168
16 Cyclohexene, 1-methyl-4-(1-methylethylidene) 0.791 -
17 Trans-Linalool oxide (furanoid) 0.029 0.093
18 Linalool 58.855 76.452
19 Bicycle[2,2,2]heptan-2one, 1,7,7-trimethyl-, (1S) 6.542 -
20 Cyclohexanol, 1-methyl-4-(methylethenyl)- 0.034 -
21 Citronellal 0.146 0.057
22 Pinocarvone 0.046 -
23 α-Campholenal 0.028 -
24 Endo-Borneol 1.763 0.429
25 Terpinen-4-ol 0.233 0.082
26 L-alpha-Terpineol 0.656 0.163
27 Bicyclo[3,1,1]hept-3-ene-2-one,4,6,6-trimethyl- 0.063 -
28 2,6-Octadien-1-ol, 3,7-dimethyl-, (Z)- 0.032 -
29 Citronellol 0.339 -
30 2,6-Octadienal, 3,7-dimethyl-,(Z)- 0.038 -
31 Geraniol 4.022 1.197
32 Citral 0.048 0.027
33 Thymol 0.021 0.034
34 1-Naphthalenol, decahydro-4a-methyl- 0.029 0.056
35 Myrtenyl acetate 0.371 0.109
[563]
36 6-Octen-1-ol, 3,7-dimethyl-, acetate 0.028 -
37 Geranyl acetate 4.034 5.609
38 Dodecanal 0.034 0.047
39 Caryophyllene 0.069 0.113
40 2- Dodecenal,(E) 0.14 0.547
41 Benzene, 1-(1,5-dimethyl-4-hexenyl)-4-methyl- 0.042 -
(1S,5S)-2-Methyl-5-((R)-6-methylhept-5-en-2-
42 0.036 0.058
yl)bicycle[3,1,0]hex-2-ene
Cyclohexane, 3-(1,5-dimethyl-4-hexenyl)-6-methylene-
43 0.038 0.113
,[S-(R*,S*)]-
(3S,3Ar,3Br,4S,7R,7aR)-4-isopropyl-3,7
44 dimethyloctahydro-1H- 0.032 -
cyclopenta[1,3]cyclopropa[1,2]benzene-3-ol
45 aR-Turmerone 0.40 1.049
46 Tumerone 0.151 0.209
47 (E)-Tetradec-2-enal 0.027 0.285
48 Curlone 0.165 0.285
49 Tetradecanoic acid 0.136 0.032
50 2-pentadecanone, 6,10,14-trimethyl- 0.023 0.045
51 n-Hexadecanoic acid 0.164 0.119
52 Oleic acid 0.039
Total 100 97.014
In this study Bicycle[2,2,2] heptan-2one,1,7,7-trimethyl-, (1S) and Benzene, 1-

methyl-3-(1-methylethyl) were a compounds found only in Dinqnesh Variety
yielded 6.54% and 1.28%, respectively, (Table 1). From a total of 51 compounds
identified from Dinqnesh variety; 17 compounds were not found in Indium
variety. However, the percentage of linalool in Indium was very high (76.45%)
when compared to the linalool percentage of Dinqesh variety (58.85%).
Differently from identified compounds, Oleic acid is found only in Indium variety.
Raal et al. (2004) analysed the oil of Coriander seeds from different geographical
origins of Europe and the major constituent of the oils were reported as linalool
(58.0−80.3%), γ-terpinene (0.3−11.2%), α-pinene (0.2−10.9%), p-cymene
(0.1−8.1%), camphor (3.0−5.1%) and geranyl acetate (0.2−5.4%). The essential
oil content of coriander consists mainly linalool (50 to 60%) and about 20%
terpenes (pinenes, γ-terpinene, myrcene, camphene, phellandrenes, α -terpinene,
limonene, and cymene) as reported by Telci et al. (2006). In this study, the other
compounds identified in the coriander seed oil are D-limonene (0.73% for Indium
and 3.03% for Dinqnesh variety), aR- turmerone (1.05% for Indium and 0.40%
for Dinqnesh variety) and indo-borneol (0.43% for Indium and 1.75% for
Dinqnesh variety). Bicycle [2, 2, 2] heptan-2one, 1, 7, 7-trimethyl-, (1S) and
Benzene, 1-methyl-3-(1-methylethyl) were the major compounds only found in
Dinqnesh variety (Table 1).
The constituent of oil in Indium variety was higher for linalool (76.45%), Geranyl
acetate (5.61%) and aR-Turmerone (1.04) than for Dinqnesh variety. This implies
Indium variety is more selective because of its higher principal constituents of
linalool found in coriander. In other case, these compounds are responsible for the
characteristics of fragrance and aroma and important for antimicrobial and
[564]
antioxidant activity. However, among the common chemical constituents, the
Dinqnesh variety oil contains the highest γ-terpinene (6.67%), Bicycle [2,2,2]
heptan-2one, 1,7,7-trimethyl-, (1S) (6.54%), α-pinene (6.15%), Geraniol (4.02%),
D-limonene (3.03%), Endo-Borneol (1.76%), Benzene, 1-methyl-3-(1-
methylethyl)-(1.28%) and camphene (1.07%) than the Indium variety. The results
showed the percentage of compounds constituents obtained within the same range
to other country reports as compared with other author reports. This small
variation obtained might be due to climate, soil condition and genotype difference.
Fig.1. GC chromatogram of indium variety essential oil
Fig.2. GC chromatogram of Dinqnesh variety essential oil
Conclusion
Essential oil composition depends on genetic and ecological conditions affecting
the plants. The results obtained indicated variation on coriander seed essential oil
[565]
from two varieties cultivated in the same agro-ecological conditions. In our study
Linalool (58.85−76.41%), Geranyl acetate (4.04−5.60%), γ-terpinene
(2.83−6.76%) and α-pinene (4.90−6.15%) were found as major components are
responsible for the character of fragrance and aroma of the coriander. The results
showed the Ethiopian coriander had met the international standard for essential oil
and is competitive for export to the international market.
Acknowledgements
The author would like to thank the Ethiopian Institute of Agricultural Research for
funding the research and JIJE Analytical Testing Service Laboratory for GC-MS
analysis.
References
Adams RP. 1995. Identification of essential oil components by gas chromatography/mass
spectroscopy. Carol Stream: Allured, 1995. 469p.
Aissaoui A, Zizi S, Israili ZH and Lyoussi B. 2011. Hypoglycemic and hypolipidemic effects of
Coriandrum sativum L. in Merionesshawi rats, Journal of Ethnopharmacology, 137(1), 652–
661.
Bown D. 1995. The Royal Horticultural Society Encyclopedia of Herbs and Their Uses, Dorling
Kindersley Ltd. London, p. 424.
Delaquis PJ, Stanich K, Girard B and Mazza G. 2002. Antimicrobial activity of individual and
mixed fractions of dill, cilantro, coriander and eucalyptus essential oils, International Journal
of Food Microbiology, 74(1–2), 101–109.
Dobos G, Novak J. 2005. Comparison of the composition of the essential oil of some winter-
annually cultivated coriander accessions (Coriandrum sativum L.). Zeitschrift fur Arznei- and
Gewurzpflanzen;10:144–145.
Ebrahimi ES, Hadian J and Ranjbar H. 2010. Essential oil compositions of different accessions of
Coriandrum sativum L. from Iran. Natural product research, 24(14), pp.1287-1294.
Lawrence, BM. 1992. A planning scheme to evaluate new aromatic plants for the flavor and
fragrance industries. In: Janick, J.; Simon, J.E., editors. New Crops: exploration, research, and
commercialization. Proceedings of the Second National Symposium; John Wiley and Sons,
Inc., NewYork. p. 620-627.
Mahendra P, and Bisht S. 2011. Coriandrum sativum: A daily use spice with great medicinal
effect. Pharmacognosy Journal, 3 (21): 84–88.
Mclafferty FW and Stauffer DB. 1989. The wiley/NBS registry of mass spectral data. New York:
Hardback, 7872p.
Mhemdi H, Rodier E, Kechaou N, Fages J. 2011. A supercritical tuneable process for the selective
extraction of fats and essential oil from coriander seeds. J. Food Eng. 105(4):609-616.
Mohammadi S, and Saharkhiz MJ. 2011. Changes in essential oil content and composition of
catnip (NepetacatariaL.) during different developmental stages. J. Essent. Oil Bear. Pl. 14 (4):
396–400.
Msaada K, Hosni K, Taarit BM, Chahed T, Kchouk EM and Marzouk B. 2007. Changes on
essential oil composition of Coriander (Coriandrum sativum L.) fruits during three stage of
maturity”, Food Chemistry, 102, 1131-1134.
Raal A, Arak E and Orav A. 2004. Chemical composition of coriander seed essential oil and their
conformity with EP standards. Agraarteadus 15:234–239.
[566]
Rajeshwari CU, Siri S and Andallu B. 2012. “Antioxidant and anti arthritic potential of coriander
(Coriandrum sativum L.) leaves”, e-SPEN Journal, 7(6), 223–228.
Silva F, Ferreira S, Queiroz JA, Fernanda CD. 2011. Coriander (Coriandrum sativum L.) essential
oil its antibacterial activity and mode of action evaluated by flow cytometry. J. Med.
Microbiol. 60:1479-1486.
Small field BM, Van Klink JW, Perry NB, Dodds G. 2001. Coriander spice oil: effects of fruit
crushing and distillation time on yield and composition. Journal of Agriculture and Food
Chemistry;49:118-123.
Sriti J, Wannes WA, Talou T, Vilarem G amd Marzouk B. 2011. Chemical composition and
antioxidant activities of Tunisian and Canadian coriander (Coriandrum sativumL.) fruit. J.
Essent. Oil Res. 23 (4): 7–15.
Telci I, Gul TO and Sahbaz N. 2006. Yield, essential oil content and composition of Coriandrum
sativumvarieties (var. vulgare Alef and var. microcarpumDC.) grown in two different
locations. Journal of Essentl Oil Research; 18: 189-193.
Zheljazkov VD, Pickett KM, Caldwell CD, Pinocock JA, Roberts JC and Mapple- Beck L. 2008.
Cultivar and sowing date effects on seed yield and oil composition of coriander in Atlantic
Canada. Industrial Crops Prod., 28: 88–94.
[567]
[568]
Evaluation of Improved Turmeric Boiler
at Different Boiling Durations
Biruk Hirko1*, Haimanot Mitiku2, Sileshi Abera1, Abukiya Getu1 and Belay Gezahagn1
1Ethiopian Institute of Agricultural Research, Tepi Agricultural Research Centre P. O. Box 34, Tepi, Ethiopia
2Ethiopian Institute of Agricultural Research, Debrezeit Agricultural Research Centre, P. O. Box 32, Debrezeit, Ethiopia
*Corresponding author: birukhirko@gmail.com
Abstract
Turmeric boiling is the most tedious, laborious and time-consuming practice. A study
was conducted to evaluate the improved and traditional boiler at different boiling
durations. The experiment consisted of four levels of boiling durations (25, 35, 45 and
60 minutes) and two types of turmeric boiler laid out in a 4x2 factorial arrangement
using Completely Randomized Design (CRD) with three replications. The result
revealed that in the modified boiler the material handling system is easy for evacuating
the boiled rhizomes to the drying chamber; no trampling of cooked rhizomes, less
fuelwood and water requirements compared to the traditional method. In other case
average loss of 2.26% oleoresin recorded in the traditional boiling method when
compared to the improved method. As the duration of the boiling increased the heat
utilization also increased due to the long time stay of rhizomes and finally essential oil
and oleoresin leached along with the boiled water which is more prevalent by
traditional methods. The results of the present study showed that improved boiler
processing turmeric for 25 minutes was superior as it minimizes quality loss, save fuels,
water and time. Therefore the result is good news for the turmeric processors which
includes farmers and small scale processers.
Key words: boiling, curing, duration, quality, turmeric
Introduction
Turmeric (Curcuma longa L.) is a rhizomatous herbaceous perennial plant of the
Zingiberaceae family. It is native to tropical South Asia but, now a day it is widely
cultivated in the tropical and subtropical regions of the world. The deep orange-
yellow powder of turmeric is prepared from boiled and dried rhizomes of the plant
(Li et al., 2011).
The volume and value of the export of spices from Ethiopia for the period 2009/10
were 293 tons and 284,000$, respectively (Ethiopian Export Promotion Agency of
the Ministry of Trade and Industry, 2011).Turmeric is an important spice in every
Ethiopian kitchen and commonly used as an ingredient in the preparation of local
sauces 'Wot' (Edossa, 1998; Girma et al., 2008). Turmeric preparation for market
is the most laborious and time-consuming practice. Normally, after harvesting, the
rhizomes are prepared for market in the cured and dried form.
The curing process involves boiling the fresh rhizomes in water, drying in the sun
and finally polishing. In south-western Ethiopia, turmeric curing is done using
[569]
traditional means which is causing many losses and difficulties. During cooking,
turmeric rhizomes are added into a clear barrel filled with clean water leaving 5-7
cm gap and packed or covered with dry turmeric leaves. Boiling continued with
uniform fire for 45-60 minutes (Girma et al, 2008). When the white fume comes
out, easily pierced with a wooden stick and typical turmeric aroma smell comes
out, then it is presumed that turmeric is enough boiled. Boiled rhizomes are pulled
out of the barrel with the help of a wooden comb with a long handle and allowed
to leach the water through the small perforates. The boiled rhizomes are spread on
mesh wire for drying.
As the boiling duration increase, the loss of curcumin and oleoresins contents
increased (Kamble and Soni, 2009). Boiling at the optimal level is important as
overcooking spoils the colour and undercooking renders the product brittle
resulting in the breakage of rhizomes during drying and polishing. Also, the
traditional handling of boiled rhizomes causes trampling, mud mixing, scorching;
leading to quality and quantity loss. Furthermore, the current turmeric boiling
barrel in south-western Ethiopia is not suitable for easily evacuating the boiled
rhizomes to the drying chamber. Therefore, the modified barrel having perforated
pan and easy to take in or out rhizome was made and compared to the traditional
method.

Turmeric boiler design and specification: Turmeric boiling barrel and the
perforated drum was designed and made at Tepi Agricultural Research Centre and
its detail was given in Table 1. It was a simple modification of the existing boiling
barrel used by farmers to reduce the time of boiling, fuel required and the heat
loss. The mild steel barrel was punched on periphery made of 1cm diameter at
10cm spacing up to 30-cm height of the barrel and a 15-cm upper portion was kept
intact. On the periphery of barrel 2.5 cm thick layer of glass wool insulation was
provided. The capacity of the barrel was 36 kg.
Table1. Specification of modified turmeric boiler
Boiling pot (Mild Steel) measurement Perforated barrel Measurement

(cm) (cm) (Mild Steel)
Top diameter of pot 60 cm Top Diameter 45 cm
Bottom diameter of pot 45 cm Height 45 cm
Height 60 cm Thickness 0.45 cm
Thickness 0.45 cm Diameter of perforations 1 cm
Insulation Thickness 2.5 cm Capacity 36 kg
Diameter of lid 45 cm
Thickness of lid 0.15 cm
[570]
Turmeric rhizome of Dame Variety was collected from the Tepi Agricultural
Research Centre seed multiplication plot during the 2017 main cropping season. A
total of 36 Kg of turmeric finger rhizome was used in one running for each
experimental unit. Turmeric rhizomes were harvested when matured, extra soil
attached removed, washed; roots were cut, added to the prepared barrel and filled
in the perforated barrel. Then, water was filled in the boiling barrel up to ½
heights and covered with lid. The water was filled to the barrel and rhizomes
added, then it was set in a way that there was no compressing load condition and
boiled till steam comes out and then time was noted. When the specified time of
boiling was reached the barrel was lifted with the help of a wooden stick passing
through the rings and kept back in the boiling pot on the furnace for the second
round. In this way, each sample was boiled in the traditional open pan and
improved boiler for the durations; 25, 35, 45 and 60 minutes and spread on mesh
wire for drying. The fuelwood required per batch, labour required and ease of
material handling during curing were compared.
Essential oil determination: Turmeric the essential oil content was obtained by
hydro distillation with Clevenger apparatus held to a round bottom flask. The
polished sample was ground to pass No 20 (850µ) sieve size (AOAC, 2000).
About a hundred gram of dried turmeric powder was transferred quantitatively in
2000 mL flask. 1000 mL of water was added to a flask fitted with an electrical
heating mantle and Clevenger apparatus with a condenser. The stand road and
stand base were connected to the heating mantle by clump holder. A continuous
flow of water was maintained through a hose tube and the flask was heated to
boiling and maintained a reflux rate of 1 to 2 drops per second. The distillation
was carried out for 4-5 h for the isolation of volatile oil (Sayyad and Chaudhari,
2010). The flask was rotated occasionally to wash down any material adhering to
the upper part of the walls. Turmeric essential oil is lighter than the water and was
collected from the collector at the top of layer. All oil and vapor mixture was got
condensed and collected by using separator funnel having volume count up to15
ml reading and quantified as volume-weight basis and expressed in the unit of
ml/100g.
( )
( )
( )
Oleoresin content: A hot continuous extraction (Soxhlet) method was used in the
laboratory to extract oleoresin following the method described by ICS-UNIDO,
(2008). As cited by Parthasarathy et al. (2008) oleoresin was determined by the
solvent extraction method using hexane (95%) as an organic solvent for 4 to 5
hours. 10 g of powdered turmeric embedded in filter paper. 250 mL of hexane was
used for each sample extraction. Each extraction process was carried out using
acetone about 250 mL at a temperature of 56oC for 4 hours. Solvent removal from
[571]
the miscella was done by pressure rotary vacuum evaporator at 40oC and 90 RPM.
When the last traces of acetone were evaporated, the flask was placed in a hot air
oven at 110±2ºC until two consecutive weightings taken at 11/2-hours intervals did
not differ by more than 1 mg (ASTA, 1997). Finally, the flask was cooled in
desiccators and then weighed and quantified as a percent weight-weight basis
based on the formula described by ASTA (2002) and Daniel et al. (2008).
( )
( )
( )
Experimental design and analysis: The experiment consisted of four levels of

boiling durations (25, 35, 45 and 60 minutes) and two types of turmeric boiler laid
out in a 4x2 factorial arrangement using Completely Randomized Design (CRD)
with three replications. Data obtained were subjected to analysis of variance
(ANOVA) and Least Significant Difference (LSD) was used to separate means.
Data are analysed using SAS software 9.2 Version.

Effect of boiling duration on physical characteristics of turmeric rhizome:
The effects of boiling duration on rhizome color and breakage of turmeric rhizome
after boiling were presented in table 2. Boiling had an effect on the color of
turmeric products due to the heat effect cause leaching of its yellow pigments in
boiling water and color loss. Minimum color change was observed for shorter
boiling durations due to its shorter period stay of rhizome during boiling. The
rhizomes boiled for 25 and 35 minutes shows uniform yellow color of inner color
but more uniform color for rhizome boiled for 45 and 60 minutes.
Table 2.Effect of boiling duration on physical characteristics of turmeric rhizome
Boiling duration Rhizome color Rhizome breakage (after

(minutes) (inner color) boiling)
25 Uniform yellow easy
35 Uniform yellow easy
45 Uniform yellow more easy
60 More Uniform more easy
yellow
Boiling for prolonged duration resulted in soften rhizome which might lead to
damage and rubbing during handling before drying. Also, the breakages of
rhizomes were easy for 25 and 35 minutes and easier for 45 and 60 minutes boiled
turmeric rhizomes. Similarly, Shinde et al (2011) reported turmeric rhizomes
steam-cooked for 15, 20, 25 and 30 minutes attains non-uniform yellow color with
[572]
separate yellow core, uniform yellow color with visible core, uniform yellow color
with soft rhizomes and uniform yellow color with very soft rhizomes respectively.
Fuel, labour, water and time requirements of traditional and improved

boiler: Handling of improved turmeric boiler with rhizomes was easier than the
traditional one as presented in table 3. Wooden stick passing through the rings
kept in the improved boiling barrel simplified the time and easy put in or out the
rhizome during processing. No trampling takes place in an improved method,
whereas in the traditional method there was a high risk of damage on rhizomes
during removing the product from the barrel. Only a pair of people were required
for handling improved boiler. However, a minimum of two people needed for
traditional boiling process. Also, the fuelwood required for boiling 36kg of
turmeric rhizome is 14kg whereas it is 20kg for traditional one (Table 3). In
addition, more water (80L/36kg of batch mass was needed for traditional
processing while very low water (54 L/36kg of mass batch required for improved
boiler. Improved boiler also had the advantage of very minimal risk of rhizome
damage and the difficulties of fumes during rhizome out by fork for more than six
minutes from the traditional boiling. However, in the improved boiler, the boiled
rhizomes were pulled out by piercing a stick between two rings of the pan.
Table 3. Fuel, labour, water and time requirements for turmeric boiling
Traditional boiler Improved boiler

High fuel requirements 20kg/36kg of mass batch Less wood requirements of 14kg/36 kg of mass batch
High water consumption of 80L/36kg of batch Low water requirement of 54L/36kg
Difficult to material handling Easy to handle during processing
More than six minute consuming while taking out Less time needed while taking out rhizome from
rhizome from vessel vessel
Minimum of two peoples required Two peoples enough during boiling
More risk of rhizome damage (10%) during taking out No risk of rhizome damage during taking out rhizome
rhizome
Oleoresin content (%): There was a significant (P<0.05) difference in oleoresin

content among boiling durations (Table 4). The result indicated that turmeric
rhizomes boiled for 25 minutes in both improved and traditional method retains
the highest oleoresin content (12.03% and 10.82%), respectively. It was shown
that as boiling duration increased, then, there was a decrement of oleoresin content
irrespective of the boiling method. About 10.05% loss was observed due to the
traditional boiling method. Similarly, the oleoresin content in the rhizomes loss
due to prolonged boiling duration ranged from 12.84% to 22.94%.
This study ratified that more reduction in oleoresin was due to extended boiling
duration than the boiling method used. As the duration of the boiling increased the
heat utilization also increased, rhizomes absorb more amount of heat from the
boiler due to long boiling duration and finally oleoresin might have leached along
with the boiled water which more prevalent by traditional method. Similar to this
[573]
Kamble and Soni (2009) reported the more the cooking time the more loss of
curcumin and oleoresins. The same trend was reported by Shibru et al. (2017) the
highest mean oleoresin yield (19.27%) was obtained from mother rhizomes boiled
at 100oC temperature for 30 min, whereas from finger rhizomes mean values of
13.83%, 13.02%, 11.53% and 11.37% were attained from finger rhizomes boiled
at 100oC for 30, 45, 60 and 75 minutes, respectively showing the effect of longer
processing time on the quality.
Essential oil content: The Essential oil content exhibited significant (P <0.05)
difference due to boiling duration between two groups of boiling method (Table
4). The highest essential oil content 4.73 ml/100mg from improved boiler method
boiled rhizome for 25 minutes showed non-significant difference from traditional
method boiled rhizome for 25, 35 and 45 minutes. However, the lowest essential
oil values of 3.97 ml/100g and 3.80ml/100g were recorded from samples boiled
for 60 minutes regardless of boiling method. The results obtained showed that
high essential oil was attained when turmeric rhizome was boiled for the shortest
duration. The study implies that extended boiling resulted in the loss of essential
oil and the processor should take high care during processing. The amounts of
essential oil and oleoresin in the rhizomes might be affected by heat magnitudes of
the traditional boiler due to the direct contact of rhizome with boiler and the loss
might be more as boiling duration increased.
However, improved boiler has no direct contact and the loss due to heat might be
reduced. Tisserand and Balaces (1995) studied that atmospheric oxygen can
change the chemical composition of volatile oil, and this decomposition is speeded
up by both heat and light. The stage at which the boiling is stopped largely
influences the final product (Purseglove et al., 1988).
Table. 4. Effect of boiling method on turmeric oleoresin and essential oil content
Boiling method Boiling durations Oleoresin content (%, w/w) Essential oil content (ml/100g)
25 minutes 10.82ab 4.63ab
35 minutes 10.29bc 4.50ab
Traditional method
60 minutes 9.43c 3.97cd
25 minutes 12.03a 4.73a
Improved method
45 minutes 9.91bc 4.33bc
60 minutes 9.27c 3.80d
LSD (5%) 1.3 0.38
CV (%) 7.3 4.96
Kamble and Son (2009) reported a higher percentage of essential oil and curcumin
is retained in the rhizomes boiled for 25 minutes than those from 35 minutes and
45 minutes. Contradictorily other authors reported different duration of boiling
i.e.45-60 minutes (Girma et al., 2008) and 30 minutes (Shibru et al., 2017).
[574]
Traditional processing of turmeric has been very tedious, requires much labour
and leads to many post-harvest quality losses. Therefore, the study was undertaken
to minimize the output cost of water, fuel, time and to evaluate the effective
boiling method. Improved method of boiling turmeric rhizome for 25 minutes is
yielding about 4.73ml/100g essential oil and 12.03% oleoresin content. Again, the
time, fuel and water-saving by using improved boiler during processing encourage
the farmers to use it. So, the producers and processors can attain utilization of the
improved turmeric boiler would worth a lot to contribute to the industry by
improving the efficiency of operation and quality.
Acknowledgment
The authors are indebted to the Ethiopian Institute of Agricultural Research for
funding budget and laboratory and field assistance of Tepi Agricultural Research
Centre.
Reference
AOAC (Association of Official Analytical Chemists). 2000. Official Methods of Analysis. 17th
edition, Washington D.C.
ASTA. 1997. Official analytical methods of the American Spice Trade Association, 4 th Edition
ASTA (American Spice Trade Association). 2002. A concise guide to spices, herbs, seeds and
extractives, Washington DC, pp. 48–50.
Daniel Bisrat, Solomon Abate and Wossen Kebede. 2008. Laboratory manual for plant products
analysis. Technical manual. 1(23):3-10.
Edossa Ettisa. 1998. Spices research achievements and experiences, Research report No. 33.
Institute of Agricultural Research, Addis Ababa Ethiopia.4-25.
Girma H and Kindie T. 2008. The effects of seed rhizome size on the growth, yield and economic
return of ginger (ZingiberofficinaleRoscae.). Asian Network for Scientific Information.Asian
Journal of Plant Sciences, 7(2):213-217.
ICS-UNIDO (International Center for Sience- United Nation Industrial Development
Organization). 2008. Extraction Technologies for Medicinal and Aromatic Plants.ICS-
UNIDO, AREA Science Park, Padriciano 99, 34012 Trieste, Italy.
Kamble KJ and Soni SB. 2009.A study of improving turmeric processing.Journal of Agricultural
Science, 22(1): 137-139.
Li S, Yuan W, Deng G, Wang P, Yang P, Aggarwal BB. 2011. Chemical composition and product
quality control of turmeric (Curcuma longa L.). Pharmaceutical Crops, 2(1):28-54.
Parthasarathy AV, Chempakam B and Zachariah TJ. 2008. Chemistry of Spices. Kingslynm, UK:
Cabi, 97:8.
Purseglove JW, Brown EG, Green CL and Robbins, SRJ. 1988. Spices, Vol I and II, Longman
Scientific and Technical, London, UK.
Sayyad FS and Chaudhari RS. 2010. Isolation of Volatile Oil from Some Plants of Zingiberaceae
Family and Estimation of their Anti-bacterial Potential. Journal of Current Pharmaceutical
Research, 4(1): 1-3.
[575]
Shibru ZF, Ali M, Girma H and John B. 2017. Influence of boiling temperature levels and
durations on the essential oil content, oleoresin content, and colour value of different rhizome
set types of turmeric (curcuma longa L.). Journal of Science, 7(3): 112-123.
Shinde G, Kamble KG, Harkari M and More G. 2011.Process optimization in turmeric heat
treatment by design and fabrication of blancher. International Conference on Environmental
and Agriculture Engineering IPCBEE. IACSIT Press, Singapore, v.15.pp 31-46.
Tisserand R and Balaces T. 1995. Essential Oils Safety. Bell and Brain Ltd., Glasgow.
[576]
Evaluation of the Effects of Different Curing and
Drying Methods on Quality of Black Pepper
(Piper nigruml.) in Southwestern Ethiopia
Biruk Hirko1*, Haimanot Mitiku2, Sileshi Abera1 and Abukiya Getu1
1
Tepi Agricultural Research Center P.O.Box 34, Tepi, Ethiopia; 2Debrezeit Agricultural Research Center,
P.O.Box 32, Debrezeit, Ethiopia; *Corresponding author: birukhirko@gmail.com
Abstract
Black pepper (Piper nigrum L.), the king of spices is very much dependent on the post-
harvest methods. Pungency and aroma are the most important qualities of pepper,
which is attributed to oleoresin and volatile oil respectively. Conventional open sun
drying and shade drying was tested on different pre and post-treatment curing of black
pepper processing. Quality improvement and reduction in drying time were evident for
the sun-dried samples. Approximately 5-9 days were taken for open sunlight dried
samples whereas in shade drying it took about 9-14 days. The highest oleoresin content
(8.15% w/w) was recorded for direct open sunlight dried samples, while in shade dried
samples it was low as (6.96%) by weight. The highest essential oil content of (2.00%
v/w) was obtained from Blanched berries for 2minutes and sun-dried samples.
Blanching followed by sun-drying and direct sun-drying of berries had the highest
acceptability for overall acceptability. Therefore, Physico-chemical quality analyses
and physical evaluation standards proved that the open sun drying is efficient to
produce black pepper of export quality.
Keywords: black pepper, blanching, drying methods, quality
Introduction
Black pepper (Piper nigrum L.) belongs to the family Piperaceae. It is a perennial
woody climbing liana. It is native to India, Indonesia, Malaysia, South America
and West Indies but is also widely cultivated in the tropical regions. Black pepper
is a universal table condiment used to flavor all types of cuisines worldwide. It is
christened as the ‘King of Spices’ (Mathew et al., 2001, Srinivasan, 2007).
Black pepper products include pepper powder, oil and oleoresin. Pepper oil and
oleoresin are produced in response to the increased quality consciousness,
preference for natural flavors and inconsistency in the quality of raw materials
(Zacharriah, 1989). After flowering, a six to eight months maturation period is
needed for harvesting. Harvesting is done during December to February for low
land crops and January to April for hill grown pepper. Post-harvest technology has
a tremendous role in the quality improvement of spices (Pruthi, 1993; Purseglove
et al., 1981). In the case of pepper, spikes are removed from the vines and they are
kept as such for a day for de-spiking. The most common pre-treatment before
drying adopted in India is blanching (Anonymous, 1997).
[577]
The green colour of pepper berries converts into black colour due to the enzymatic
browning by fermentation and oxidization of phenolic compounds which appear in
the pepper berries (Dhas and Korikanthimath, 2003). The green pepper berries are
generally being blanched in hot water (80oC) for 1 - 2 minutes in producing black
pepper for a uniform black colour as well as better flavor quality and aroma
(Bunchol, 2011).
Postharvest technology covers several unit operations after harvesting of black

pepper, such as curing, drying, cleaning or garbling, quality evaluation, packaging
and storage. For the preparation of good quality black pepper, a simple blanching
process has been developed by (CFTRI, 1980). In the Central Food Technological
Research Institute, the pepper berries/spikes are dipped in boiling water for one
minute and then sun-dried to get a quality product. Drying characteristics of
specific products should be determined to improve the quality. Black pepper is
usually harvested and processed traditionally under variable sanitary conditions
which lead to contamination of products by dirt and dust and also resulting in poor
quality products with less oleoresin, aroma and ultimately resulting in low market
price in our country. Blanching was not practiced in our country. Therefore, the
objective of this study was to determine the effects of blanching and drying
methods on the quality of black pepper.

Matured black pepper berries that were on the process of turning to yellow colour
from fully-mature green fruits harvested for this trial. It took 6-7 months from
flowering to harvesting. Sample sources were from one of the released varieties
from Tepi Center (Gacheb) was used for this experiment. In total, 100Kg of black
pepper samples were collected from Tepi Agricultural Research Centre breeder
seeds multiplication in January 2017 main harvesting season.
Curing: The harvested black pepper products were divided into two portions. One
portion was the berries detached and the other was the intact spikes with berries.
Then samples for hot water treatments were done by immersing the pepper
samples in boiling water at 800C for 2 minutes following the methods (Bunchol,
2011) with the help of a clean gunny bag, then taken out, drained and spread on a
mesh wire for drying. Other samples that did not require blanching or hot water
treatments were directly taken to each drying method (sun drying or shade drying.
Drying: Each differently treated black pepper samples were loaded onto wire-
mesh trays of equal dimension of 50cm*70cm for both shade and sun-dried
samples. Both portions were then divided again into two drying methods of open
sun drying and shade drying. Polyethylene sheets were used to cover samples
during night, mixed and re-spread daily. After the completion of drying berries
[578]
were removed from spikes and sifted. Finally, the dry weight of the product was
recorded and percent recovery was worked out. Extraneous materials were
counted and removed from the dried samples. Samples were done in triplicate.
Experimental design and analysis: The experiment consisted of eleven curing

methods laid out in using Completely Randomized Design (CRD) with three
replications. Data obtained were subjected to analysis of variance (ANOVA) and
Least Significant Difference (LSD) was used to separate means. Data are analysed
using SAS software 9.2 Version.
Data collected
Dry recovery (%): was calculated from fresh weight (W1) taken after harvesting
and the dried weight (W2) taken after drying and expressed in percentage.
( )
Dry recovery ( ) , Where, W1= fresh weight (kg) of
the sample;
W2= dried weight (kg) of sample
Extraneous matter (%): All materials other than black pepper berries,
irrespective of whether they are of like stems and leaves or minerals and sand
were weighted and given as % by weight (Anbalagan et al., 2010).
Color: the colours of the dried black pepper samples were determined visually to
three classification light black, black-brown and black.
Oleoresin (%): A hot continuous extraction (Soxhlet) method was used to extract
oleoresin. Parthasarathy et al. (2008) stated oleoresin was determined by the
solvent extraction method using hexane as an organic solvent for 4 to 5 hours.
About 10 g of black pepper powder embedded in filter paper placed in glass
columns blocked with non-absorbent cotton below which a volumetric flask (500
ml) was kept to collect the extract. A volume of 250 ml of hexane was used for
each sample extraction. Solvent removal at each step was done by a rotary vacuum
evaporator. Then the flask was placed in a hot air oven at 110±2ºC until two
consecutive weightings taken at 11/2hours intervals did not differ by more than 1
mg (ASTA, 1997). Finally, the flask was cooled in a desiccator and then weighed
and quantified as percent weight-weight basis based on the formula described by
ASTA, (2002) and Daniel et al. (2008).
( )
( )
( )
[579]
Essential oil content: Turmeric essential oil content was obtained by hydro
distillation with Clevenger apparatus held to a round bottom flask. About a
hundred gram of dried turmeric powder was transferred quantitatively in 2000 ml
flask. A volume of 1000 ml of water was added to a flask fitted with an electrical
heating mantle and Clevenger apparatus with a condenser. The distillation was
carried out for 4-5hr for the isolation of volatile oil (Sayyad and Chaudhari, 2010).
All oil and vapor mixture was collected by using separator funnel and quantified
as the volume-weight basis, expressed in the unit of ml/100g.
( )
( )
( )
Days to dry: The two modes of drying (open sun and shade drying) differ in
drying duration (Table 1), in which T2 (blanching berries for 2min and sun
drying) and T10 (sun drying of spikes 2 days then detached berries and blanching
for 10 min and sun drying) took the shortest (5 days) drying duration. This implies
blanching and sun-drying reduce the drying duration required. Similarly, the
highest air temperature received in open sun leads to the second shortest (6 days)
to drying for T4 (blanching spikes for 2min and sun drying) and T6 (direct sun-
drying of berries after harvest). However, T9 (direct shade drying of spikes after
harvest) needed the longest (14 days) to dry. In this study blanching and sun-
drying significantly decreased the drying duration needed for black pepper.
Effective drying could be achieved in a few days in sunny days, while it could
take more than a week in shade drying. The report of Kaymak (2002) showed the
partial destruction of the pepper cell walls by blanching facilitating water transfer.
Sun-dried pepper berries take 4-5 days for proper drying, depending upon the
climatic conditions (Krishnamurthy et al., 1999 and Patil, 1989). Blanching
accelerates the drying and browning of the berries, but the cost of fuel for heating
water may be an additional factor for farmers and small scale processors.
Dry recovery (%): Results from the current investigation showed that there was a
significant difference among the various curing and drying methods (p<0.05) on
average dry weight recovered (Table 1). In the present investigation, dry recovery
of sun-dried sample dryer was low when compared to shade drying. This might be
due to low moisture content attained by sun drying. Maximum dry recovery of
(32.63%) was recorded from blanched spikes for 2 minutes and shade dried
sample (T5) but, statistically no significant difference with the values obtained for
berries blanched for 2 minutes and shade dried (T3) and direct shade-dried berries
after harvest (T8). Whereas, the minimum result (29.93%) was recorded from
berries blanched for 2 minutes and sun-dried (Table 1). Rain, low sunlight,
remoistening during night and humid days may affect drying recovery and
property of black pepper. The curing and drying method had a significant role on
[580]
dry recovery of black pepper and predominated by drying method in which most
of the shade dried samples showed the highest recovery percentage.
Table 1.Effect of blanching method and drying method on physical characteristics of black pepper
Days Dry recovery Extraneous

Trt Treatments Colour
to dry (%) matter (% )
T1 1-2 days heaping spikes and then sun drying
7 30.80bc 0.77cd Black brown
berries
T2 Blanching berries for 2min and sun drying 5 29.93c 0.46e Black brown
T3 Blanching berries for 2min and shade drying 9 32.53a 0.81cd Black brown
T4 Blanching spikes for 2min and sun drying 6 31.36abc 0.73cd Black brown
T5 Blanching spikes for 2 min and shade drying 11 32.63a 1.23ab Black brown
T6 Direct sun drying of berries after harvest 6 31.28abc 0.65de Black
T7 Direct sun drying of spikes after harvest 9 32.33ab 0.76cd Black
T8 Direct shade drying of berries after harvest 13 32.61a 1.15b Black brown
T9 Direct shade drying of spikes after harvest 14 31.48abc 1.40a Light black
T10 Sun drying of spikes for 2 days then detached
berries and blanching for 10 min and sun 5 30.13c 0.90c Black brown
drying
T11 Sun drying of spikes 2 days then detached
berries and blanching for 10 min and shade 8 31.12abc 1.27ab Black brown
drying
LSD (5%) 1.57 0.19
CV (%) 2.94 12.48
This may be due to low air temperature received during under shade drying favour
less and slow moisture loss when compared to direct sun drying. This might
finally lead to the formation of mold on the sample as drying duration was
prolonged due to unfavorable weather conditions. An increase in moisture content
might be enhancing the dry recovery of black pepper and fetch more money for
producers. However black pepper with high moisture content increases the
chances for the development of mold and attack by microorganisms. Thomas and
Gopalakrishnan (1992) also reported a size decrease, shrink in size and wrinkles in
the skin after drying.
Extraneous matter (%): This study showed that drying in the sun generally
achieved the required minimum extraneous matter less than 1% (Indian Standard)
than shade-drying. Direct shade-dried spikes after harvest(T9) had the highest
extraneous matter weighted and is significantly different (P<0.05) from other
treatment except for (T11) and (T5), (Table 1).However, blanching berries for
2minutes and sun-drying resulted in the lowest extraneous matter (0.46%). This
may be due to sun drying is free from shade, leaves and unwanted material or
drops from the shade tree which could lead to a re-wetting of the samples. In the
latter case, the samples also stand the risk of further recontamination before they
are dry enough to withstand microbial growth. Generally, sun drying reduces
extraneous matter that might be observed as the drying period is faster. This effect
is better as practiced with blanching. Similarly, IPC (2008) reported hot water
blanching to reduce contaminants and increase drying rate. This finding suggests
[581]
that direct sun-drying of berries alone and blanching of berries for 2 minutes
followed by sun-drying providing black pepper of comparing the value of the
Indian standard (<1%) extraneous matter.
Colour: The colours of dried black were assigned to the nearest colour of light
black, black-brown and black visually.
Essential oil: Data of chemical quality attributes was affected by the curing and
drying method as presented in (Table 2). The effect of curing and drying method
on the essential oil content of the black pepper had shown a significant difference
(P<0.05) among treatments. The highest three values (2.00, 1.87 and 1.87ml/100g)
were blanched samples of (T2), direct sun-dried berries (T6) and T1 (heaped
sample), respectively showed no significant difference. The four lowest values
(1.66, 1.63, 1.53 and 1.67ml/100g of T5, T8, T9, and T11, respectively) did not
show a significant difference among the samples dried in shade drying regardless
of curing method. This result was supported by Jacob et al (1985) blanching
improves the colour but, affords pepper fruits depleted of significant quantum of
volatiles. Before drying of green berries blanching in 80-82°C hot water helps to
avoid loss of oil, beyond that temperature quality deterioration may result.
Table 2. Effect of curing and drying method on essential oil and oleoresin content of black pepper.
Essential oil Oleoresin

Trt Treatments description
(ml/100g) content (%)
T1 1-2 days heaping spikes and then sun drying berries 1.87abc 7.85a
T2 Blanching berries for 2min and sun drying 2.00a 8.08a
T3 Blanching berries for 2min and shade drying 1.73bcd 7.31ab
T4 Blanching spikes for 2min and sun drying 1.76bcd 7.52ab
T5 Blanching spikes for 2 min and shade drying 1.66de 6.96b
T6 Direct sun drying of berries after harvest 1.90ab 8.15a
T7 Direct sun drying of spikes after harvest 1.72cd 7.62ab
T8 Direct shade drying of berries after harvest 1.63de 8.10a
T9 Direct shade drying of spikes after harvest 1.53e 8.05a
T10 Sun drying of spikes 2 days then detached berries
1.75bcd 8.06a
and blanching for 10 min and sun drying
T11 Sun drying of spikes 2 days then detached berries
1.67de 7.99a
and blanching for 10 min and shade drying
LSD (5%) 0.18 0.87
CV (%) 6.00 6.61
Oleoresin content (%): The effect of curing and drying methods on oleoresin
content of the black pepper did not show a significance difference (P>0.05) among
many treatments (Table 2). The effect of processing on black pepper oleoresin
quality ranged from 6.96% to 8.10%. However, they showed a significant
difference (P<0.05) with T5 (blanching spikes for 2 min and shade drying)
attained the lowest 6.96% oleoresin content. The less variation among treatments
implies processing method did not affect black pepper oleoresin content.
[582]
In the international market oleoresin of black pepper is the most desired value-
added extract from black pepper. The minimum oleoresin percentage as per the
USFDA (United States Food and Drugs Administration standard) is 7.5% by
weight and is comparative with the present investigation. In other cases the
quantity and quality of oleoresin depend to a great extent on the geographical
origin of spice (Shankarikutty et al., 1982). The same for oleoresin content in this
study, Govindarajan (1977) stated the delay in drying time and degree of
microbial contamination can easily affect the aroma quality though no great
changes may occur in the physicochemical characteristics. Moreover, current
results are in support of the previous report by Joy et al. (2002) and Purseglove et
al. (1981) that quality improvement of black pepper and more other spices is
strongly positively correlated with shortening of drying duration.

This study was conducted to determine the effect of curing and drying methods on
black pepper product qualities. Blanching berries for 2minutes and sun drying, and
direct sun-drying of berries after harvest resulted in higher essential oil and
oleoresin content, shorten the drying time and eliminates the formation of mold
development. Berries should be separated from spikes before drying black pepper
as it can take time and difficulties during separation and finally, resulted in the
impurity of the product. Different postharvest activities with a role in the
reduction of drying duration are highly advised.
Acknowledgment
The authors are indebted to the Ethiopian Institute of Agricultural Research for
funding budget for this research. Valuable support of Tepi Agricultural Research
Centre staff is gratefully acknowledged.
References
Anbalagan G, Prabakaran AR and Gunasekaran S. 2010. Spectroscopic characterization of Indian
standard sand. Journal of applied spectroscopy, 77(1), pp.86-94.
ASTA (American Spice Trade Association). 1997. Official method of analysis. Englewood cliffs,
NJ, USA.
ASTA (American Spice Trade Association). 2002. A concise guide to spices, herbs, seeds and
extractives, Washington DC, pp. 48–50.
Bunchol AJ. 2011. Interview with Food Technologist Officer Malaysian Pepper Board Sarawak.
CFTRI. 1980. (Central Food Technological Research Institute), Annual Report for 1978-79.
Mysore. p.187.
Dhas P, and Korikanthimath V. 2003. Processing and quality of black pepper – A review. Journal
of Spices and Aromatical Crops, 12(1), 1–13.
ICS-UNIDO (International Center for Science- United Nation Industrial Development
Organization). 2008. Extraction Technologies for Medicinal and Aromatic Plants.ICS-
UNIDO, AREA Science Park, Padriciano 99, 34012 Trieste, Italy.
[583]
IPC. 2008. Finalized IPC Good Agricultural Practices for Pepper (Piper nigrumL.): International
Pepper Community.
Jacob CV, Varghese M, Joyand T and Mathuua T. 1985. Focus on pepper technology: Drying
ofpepper. Indian Spices 22(3): 27-29. Spices Board (Govt. Of India).
Joy CMI, George PP and Jose KP. 2002. Drying of Black Pepper (Piper nigrum L.) Using solar
tunnel dryer, Pertanika J. Trap. Agric. Sci. 25(1): 39-45.
Kaymak-Ertekin F. 2002. Drying and rehydrating kinetics of green and red peppers. Journal of
Food Science, 67(1), 168–175.
Mathew PJ, Mathew PM and Kumar V. 2001. Graph clustering of Piper nigrum L.(black
pepper). Euphytica, 118(3), pp.257-264.
Parthasarathy AV, Chempakam B and Zachariah TJ. 2008. Chemistry of Spices. Kingslynm, UK:
Cabi, 97:8.
Patil RT. 1989. Drying studies on black pepper. Journal of Food Sci. Technol. 26(4): 230-231.
Pruthi JS. 1993. Major Spices of India: Crop Management and Post-Harvest Technology. Indian
Council of Agriculture Research, New Delhi.
Purseglove JW, Brown EG, Green CL, Robins SRJ. 1981. Spices. Longman Group Limited.
London. 2: 644-735.
Shankarikutty B, Narayanan CS, Rajaraman K, Sumathikutty MA, Omankutty M and Mathew AG.
1982. Oils and oleoresins from major spices. Journal of plantation crops 10(1): 1-20.
Srinivasan K. 2007. Black pepper and its pungent principle-piperine: a review of diverse
physiological effects. Critical reviews in food science and nutrition, 47(8), pp.735-748.
Thomas PP and Gopalkrishnan N. 1992. A process for the unwrinkled dry ball shaped green
pepper. Indian Spices, 29(4), pp.9-10.
Zacharia TJ. 1998. Quality analyses of major spices. Pafai Journal 20(2): pp 25-31.
[584]
Fruit Crops
Performance of Commercial Avocado Varieties in

Raya Azebo, Southern Zone of
Tigray Region, Ethiopia
Haile Abebe1*, Wakuma Biratu1, Kidane Tesfay1, Molla Berhe1,
and Haileselassie Gebremeskel1
1
Ethiopian Institute of Agricultural Research, Mehoni Agricultural Research Center. P.O.Box. 47. Mehoni,
Tigray, Ethiopia; *Corresponding author email: zehaylians@gmail.com
Abstract
Avocado is among the subtropical fruit crops produced in Ethiopia; however, the
production and Productivity of avocado is influenced by different factors; of which
lack of improved and well-adapted varieties are the major one. So, an experiment
was conducted at Fachagama from 2013 to 2018 to evaluate and select the best
adaptable avocado variety/ies at Raya Valley. Six avocado varieties namely Ettinger,
Fuerte, Pinkarton, Hass, Naba and Bacon were laid in RCBD and replicated three
times. To achieve the objective, the growth, phenological and yield data were
collected and analyzed using SAS software. Accordingly, above graft union stem
girth diameter was significantly (P≤0.01) affected by variety, while tree height and
canopy diameter were not significantly ascertained by variety. All the tested avocado
fruit yield parameters were significantly (P≤0.01) influenced by variety throughout
the three years except marketable fruit yield tree-1 and the total yield tree-1 which
was not significantly (P≥0.5) determined by variety during 2017 fruit harvesting
year. Variety exerted significant (P≤0.01) variation on fruit length (cm) fruit weight
(g) and seed weight respectively, thought the three years (2016-2018), the two years
(2017 and 2018) and the two years (2016 and2018). The highest girth diameter
(61.44mm) was obtained on Nabal variety. Likewise, the highest (15.51 and18.88kg
tree-1) total yield was recorded in Pikerton and Bacon varieties respectively during
2016 and 2018. From the result, Pinkerton and Bacon are recommended for Raya
Valley and other areas having similar agro-ecologies. However, it is important to
consider the water requirements of these crops.
Key word: Avocado, Yield, Growth, Varieties
Introduction
Avocado (Persea americana Mill.) is a native tree of Central American countries,
the northern coast of South America and the West Indies (Chen et al., 2007;
Bayram et al., 2012; Flores et al., 2019). It has been reported that the cultivation
of avocado for commercial purposes began in California and Florida and then in
Israel, South Africa and Chile. Although a range of avocado cultivars are grown,
Hass is one of the most world's widely grown, imported and exported cultivar
[585]
(Shepherd and Bender, 2002). According to Garner et al. (2008) this fruit is
characterized by excessive flower and fruit abscission.
According to Fulgoni et al. (2013) and Flores et al. (2019) avocado is a fruit with
a high nutritive value and contains monounsaturated fatty acids (MUFA) dietary
fiber, essential nutrients and phytochemicals. However, the outstanding
compositional feature is the high-fat content, which varies significantly between
different cultivars due to climatic and soil differences (Flores et al., 2019). It has a
bland nature with a particularly high nutritional value containing 15-30% oil,
vitamins viz. vitamin A, B6, B12, C, K, E, Folacin, and Niacin), minerals, high
caloric value 123-387 kcal/kg of edible avocado, and has low sugar content
(Wong et al., 2010; Villa-Rodríguez et al., 2011; Mooz et al., 2012; Tan et al.,
2017).
In Ethiopia, the tree was first introduced in 1938 by private orchardists in Hirna
and Wondogenet and gradually adapted to different agro-ecologies of the country
(Edossa, 1997, Zekariyas, 2010). The introduced avocado varieties include
“Ettinger”, “Fuerte”, “pinkarton”, “Hass” “Naba “and “Bacon” which are known
for bearing uniform and good quality fruits (Ketema et al., 2010). According to
CSA report (CSA, 2017), in 2015 and 2016, in Ethiopia, the total production of
avocado was 538245.79 and 649,821.04 quintals, respectively which was
increased by 20.73 %. In the same year, the area covered by avocado was
16665.45 and 17834.58 ha, respectively which was increased by 30.51 percent.
Similarly, in Tigray, the crop is now produced by more than four thousand farmers
who collectively farm more than 10.64 ha (CSA, 2017).
Avocado is a fruit grown mainly in warm temperate and subtropical climates

throughout the world (Tan et al., 2017; Flores et al., 2019). However, limited
cultivation areas, high nutritional value and a distinctive flavor are the causes for
high prices for avocados in world markets (Poudel et al., 2018).
In Ethiopia, particularly in the Raya valley of Tigray region, the absence of

improved and adapted avocado varieties are limiting the production, productivity
and profitability of the crop where the crop is highly demanded and the agro-
ecology also favor the crop production. So far, there were no studies done to select
the adaptable and high yielding varieties in Tigray region in general; rather some
nurseries were established and distributing seedlings of some fruit crops (Mango
and Avocado) varieties which are not studied and recommended to the local areas.
Taking into account the above-mentioned gaps, this study aimed to evaluate the
performance of six avocado varieties for their growth, yield and fruit quality
parameters at the Raya valley.
[586]
Materials and Methodology
Description of the Experimental Site: A study was conducted at the Mehoni
Agricultural Research Center (MhARC) Fachagama testing site in the Raya
Valley, Northern Ethiopia. It is located at 668 Km from the capital Addis Abeba
and about 125 Km south of Mekelle, the capital city of Tigray regional state.
Geographically, the experimental site is located at 12° 41'50'' North latitude and
39° 42'08'' East longitude with an altitude of 1578 m.a.s.l. Data from the
meteorological class of the center shows a mean annual rainfall of 539.32 mm
with an average minimum and maximum temperature of 12.81 and 23.24°C,
respectively. The soil textural class of the experimental area was clay loam with
pH of 7.9.
Experimental Materials: Seedlings of six improved Avocado varieties namely

Ettinger, Fuerte, pinkarton, Hass, Naba and Bacon were collected from Melkassa
Agricultural Research Center, Ethiopia to test the performance of the varieties at
Fachagama during July 2013.
Design and Experimental Procedure: A completely randomized block design

(RCBD) and replicated three times. Seedlings were planted in a plot having
7*7*6m (294m2) size i.e. 7m between plants and rows and 6 plants per plot and
the net area of the experimental area was 5292m2(294*3*6). All agronomic
practices like, cultivation, weeding and harvesting were done at standard time. Pits
or holes (60cmx60cm) for planting were prepared two months before planting and
the subsoil and topsoil was kept separately. Well decomposed manure with the
topsoil was mixed and fills back the prepared pits. To help the planters to exactly
locate the center of the pit during planting a stake were placed at the center of the
hole. And the seedlings were planted on the center of the hole and pressed the soil.
Data Collection and Measurements: Measurement of avocado traits beginning

from 2016 to 2018 data on Tree height, average canopy diameter, average tree
trunk diameter (Above Graft union), fruit diameter (cm): fruit length (cm),
average fruit weight (gr), average juice weight (gr), average seed (stone) weight
(gr), marketable yield, unmarketable yield and total fruit yield (kg) were collected.
Data Analysis: Collected data were subjected analysis of variance (ANOVA)

following a procedure appropriate to a randomized complete block design as
suggested by (Gomez and Gomez, 1984). Dunkens’s multiple range test (DMRT)
was used for comparison of means at P ≤0.05.
[587]
Tree Height at First Harvest and Canopy Diameter (m): The analysis revealed
that tree height at first fruit harvesting year (2016) was not significantly (P≥0.05)
influenced by avocado varieties (Table 1). However, the longest and shortest tree
heights were recorded, on Fuerte and Pinkerton varieties respectively (Table 4).
Variety exerted non-significant effects on avocado canopy spread during the first
fruit harvesting year (2016) as indicated in the Anova table (1). Even though,
numerically, the maximum (3.23m) and minimum (2.53 m) canopy spreads was
recorded on Fuerte and Nabal varieties (Table 4). The results of this finding
disagreed with the result of Chen et al. (2007) who reported the significant effects
of avocado genotypes on plant height and canopy diameter.
Above graft Union Girth Diameter (mm): At first fruit harvesting season
(2016), variety maintained significant (P≤0.01) effects on above graft union girth
diameter (Table 1). Nabal variety showed the maximum, which however was not
statistically different from Fuerte and Bacon; while, Pinkerton showed the
minimum girth diameter yet statistically not different from Ettinger and Hass. This
finding is in agreement with Chen et al. (2007) who reported that there is a
significant difference in the canopy girth diameter among Gwen, Bacon, Fuerte,
and Zutano varieties.
Marketable Fruit Number: The marketable fruit number was significantly (P≤
0.01) affected by avocado varieties in all seasons of fruit harvesting season (2016,
2017 and 2018 years) (Table 2). Significantly the highest number of fruits per tree
was harvested from Pinkerton variety in 2016 and 2017cropping years
respectively. In contrast, the lowest number of fruits per tree was recorded on
Fuerte during 2016 and 2017 cropping years respectively. In 2018, the maximum
number of fruits per tree was harvested from Hass variety; while the lowest
number of fruits per tree was recorded on Nabal variety (Table 5). Similarly,
Bayram et al. (2012) found that, significant difference among avocado cultivars.
This could be attributed to inherent characters and wider adaptability of the
variety.
Marketable Fruit Yield Tree-1 (kg): Avocado variety exerted significant (P≤
0.01) effects on the marketable yield tree-1 during 2016 and 2018; however,
marketable yield tree-1 was non-significantly (P≥0.5) affected by variety in 2017
cropping year (Table 2). Pinkerton variety gave the maximum fruit yield during
the 2016 and 2018 fruit harvesting seasons, which however was not statistically
different from Bacon variety during the 2018 fruit harvesting year (Table 5). The
result is in agreement with the findings of Bayram et al. (2012) who reported a
significant effect of variety on marketable fruit yield tree-1. Fruit quality is one of
the most important factors that determine the desirability of avocados during the
[588]
marketing and which are highly linked with the agro-ecological condition under
which the crop is produced, management practices made and the genetic
contribution.
Unmarketable Fruit Number Tree-1: Anova table (2) indicated the significant
(P≤ 0.01) effects of varieties on unmarketable fruit number tree-1 in all fruit
harvesting year (206, 2017 and 2018). The highest unmarketable fruit number per
tree was obtained from Nabal Variety on the 2016 fruit harvesting year. The
maximum number of unmarketable fruits per tree was observed on Bacon which
was statistically similar to Ettinger and Pinkerton varieties (Table 5).
Unmarketability in this case is attributed to smaller sized fruit which may be the
matter of genetic characters and management practices.
Unmarketable Fruit Yield Tree-1 (kg): Unmarketable fruit yield of avocado was
significantly (P≤ 0.01) altered by variety during 2016, 2017 and 2018 crop
harvesting years (Table 2). Significantly the highest unmarketable fruit yield per
tree was recorded on Pinkerton variety during 2016; though Fuerte and Ettinger
gave no unmarketable yield yet statistically not different from Bacon (Table 5).
During 2017, the highest unmarketable yield was recorded on Fuerte followed by
Hass; however, no unmarketable yield was recorded on Pinkerton, Bacon, Nabal
and Ettinger (Table 5). Similarly, the largest and lowest unmarketable fruit yield
was recorded, respectively from Bacon and Fuerte varieties during the 2018 fruit
harvesting season.
Total Fruit Yield Tree-1 (kg): The total yield of avocado was significantly (P≤
0.01) determined by variety throughout 2016 and 2018, while non-significant
(P≥0.5) effects of variety on yield were recorded during 2017 cropping year
(Table 2). Significantly highest and lowest fruit yield was recorded from Pinkerton
and Nabal varieties during the 2016 fruit harvesting year. Besides, during the 2018
fruit harvesting year, the maximum fruit yield was obtained on Bacon variety;
which is statistically not different from Pinkerton variety. However, the lowest
fruit yield was recorded from Nabal Variety during 2018 (Table 5).
In agreement with the present (2016 and 2018) finding, the significant effects of
variety on the cumulative yield of avocado was previously reported by Bayram et
al. (2012). Lovatt et al. (2015) also reported that a mean yield of Hass avocado is
51 kg/tree at a typical California planting density condition. The annual yields of
all the cultivars lack consistency, and in some years the yield was negligible. This
variability in the production of the three cultivars is probably reflecting a tendency
to biennial bearing on some varieties. The yielding capacity of fruits crops are
attributed to the agroecological condition, genetic potential and management
practices just like other crops.
[589]
Fruit Length (cm): The study shewed fruit length was significantly (P≤0.01)
influenced by variety throughout 2016, 2017 and 2018 crop harvesting seasons
(Table 3). As presented in Table (6), Ettinger variety gave the highest fruit length
during 2016, yet statistically not different from Bacon and Pinkeron. Fuerte was
also devoted the maximum fruit length which is statistically identical with Ettinger
in 2017. Fuerte (13.02 cm), Pinkerton (12.77 cm) and Ettinger (12.41 cm)
varieties gave the highest fruit lengths which are significantly similar to each of
them during 2018. The current results are in agreement with the findings of Poudel
et al. (2018), who reported the significant effects of avocado genotypes on fruit
length in Nepal. Similarly, Bayram et al. (2012) have also reported that there was
variation in fruit length among avocado genotypes.
Fruit Diameter (cm): Anova table (3) showed that the significant (P≤0.01)
effects of variety on avocado fruit diameter during 2017; however, fruit diameter
was not statistically affected (P≥0.05) by avocado varieties in 2016 and 2018 crop
harvesting years. Significantly the largest fruit diameter was found on Hass
variety during the 2017 fruit harvesting season; however, the lowest fruit diameter
was obtained from Fuerte variety yet statistically not different from Bacon,
Ettinger, and Pinkerton and Nabal varieties (Table 6). In line with the results
during 2017, Poudel et al. (2018) also reported that significant variations among
avocado genotypes for fruit diameter. Likewise, Bayram et al. (2012) reported that
fruit width difference between Reed and Hass.
Average Fruit Weight (g): Variety exerted highly significant (P≤0.01)

differences on average fruit weight during 2017 and 2018, while on the 2016 crop
harvesting year, fruit weight was significantly (P≤0.05) determined by variety
(Table 3). Statistically, the highest fruit weight was obtained from Ettinger variety
during the 2016 fruit harvesting year, yet statistically not different from Bacon,
Pinkerton and Fuerte varieties. Similarly, Nabal showed the maximum fruit weight
on the 2017 fruit harvesting year but statistically not different from Pinkerton
variety. In addition to this, Nabal gave the highest fruit weight during 2018, yet
statistically not dissimilar from Bacon variety (Table 6). Significant effects of
variety on individual fruit weight of avocado were also previously reported by
Bayram et al. (2012) and Poudel et al. (2018). Likewise, Gregoriou and
Economides (1991) reported that fruits of 'Ettinger' had the highest weight and
fruits of 'Hass,' the lowest. This is highly dependent on the size of the fruit which
is influenced by the management practices and the inherent characters of the
varieties.
Average Fruit Flesh (Juice) Weight per Fruit (g): Highly significant (P≤0.01)
effects of avocado variety on flesh weight was exerted during 2016 and 2018 crop
harvesting years, although on 2017, flesh weight was significantly (P≤0.05)
affected by variety (Table 1). Statistically, the highest juice (flesh) weight was
[590]
recorded from Pinkerton variety during 2016, yet it was statistically not differed
from Ettinger and Bacon varieties. Likewise, the maximum flesh (juice) weight
was recorded from Nabal variety in the 2017 fruit harvesting year. In addition to
this, during the 2018 crop harvesting year, the most prominent juice weight was
obtained on Fuerte variety (Table 6). But then, Nabal variety gave the lowest fruit
flesh (juice) weight in 2016 and; in 2017 and 2018, Hass variety gave the lowest
juice weight (Table 6).
Average Stone (seed) weight Fruit-1(g): The average stone (seed) weight of
avocado was significantly (P≤0.05) ascertained in 2016, 2018 and 2017 year (3).
The maximum avocado seed weight was found on Ettinger during the 2016 crop
harvesting year, which however was not statistically different from Pinkerton and
Bacon (Table 6). Similarly, during 2017, the highest avocado seed weight was
recorded on Nabal variety but it was statistically similar Hass variety (Table 6).
Bacon showed the highest seed weight, yet statistically similar from Ettinger as
indicated in (Table 6). However, during 2016 and 2018, Nabal variety gave the
minimum seed weight; while during 2017, the minimum seed weight was recorded
on Hass variety (Table 6). The present results accorded with the reports of Poudel
et al. (2018) who reported a highly significant difference among avocado
genotypes for seed weight. Similarly, Gregoriou and Economides (1991) reported
that the highest tone weight in Ettinger fruits and lowest stone weight in Hass
fruits.

According to the present study, among the tested tree growth parameter, only
above graft union stem girth diameter was significantly (P≤0.01) affected by
varieties; while tree height and canopy diameter was not significantly ascertained
by varieties. All the tested avocado fruit yield parameters were significantly
(P≤0.01) influenced by variety throughout the three years except marketable fruit
yield tree-1 and the total yield tree-1 which was not significantly (P≥0.5)
determined by variety during 2017 fruit harvesting year. Variety exerted
significant (P≤0.01) variation on fruit length (cm), fruit weight (g) and seed
weight, thought the three years (2016-2018), the two years (2017 and 2018) and
the two years (2016 and 2018) respectively. The highest girth diameter (61.44mm)
was obtained on Nabal variety. Likewise, the highest total yield was recorded on
Pikerton and Bacon varieties during 2016 and 2018 respectively; yet the statistical
difference was not observed with Pinkerton during 2018.
Fruit length was significantly (P≤0.01) influenced by variety throughout 2016,

2017 and 2018 crop harvesting seasons and Ettinger variety gave the highest fruit
length during 2016, yet statistically not different from Bacon and Pinkeron. Fuerte
was also devoted the maximum fruit length which was statistically at par with
[591]
Ettinger variety during 2017. Statistically, the highest fruit weight was obtained
from Ettinger variety during 2016 fruit harvesting year, yet statistically not
different from Bacon, Pinkerton and Fuerte varieties. Similarly, Nabal showed the
maximum fruit weight in the 2017 fruit harvesting year but statistically not
different from Pinkerton variety.
The maximum avocado seed weight was found on Ettinger during the 2016 crop
harvesting year, which however was not statistically different from Pinkerton and
Bacon. Similarly, during 2017 the highest avocado seed weight was recorded on
Nabal variety. Bacon showed the highest seed weight, yet statistically similar to
Ettinger.
Thus, regarding marketable fruit yield and number, total yield per tree, average
single fruit weight and juice weight the varieties Pinkerton and Bacon are selected
as compared to the others; so both of them are recommended for growers in Raya
valley and other areas having similar agro-ecologies. However, it is important to
study the water requirements in Ethiopia at all.
Table1. Mean squares from the first year (2016) analysis of variance for the tree growth performance of avocado varieties
at Raya valley condition
Tree Growth Parameters

Source of Tree height at First Canopy Diameter (m) Above Graft union Girth
Variance D.F Harvest (m) Diameter(mm)
Treatment 5 0.282ns 0.212ns 180.1896**
Error 10 0.215 0.332 29.777
CV 20.34 21.25 10.11
*, **, and *** significant at P≤0.05, p≤0.01 and p≤0.001 probability levels respectively; ns= not significant, DF= Degree of
freedom; SOV= Source of Variation;
[592]
Table 2. Mean squares from the three years (2016, 2017 and 2018) analysis of variance for avocado yield parameters at Raya valley condition
Mean Squares
Marketable Fruit Number Tree-1 Marketable Yield Tree-1 (Kg) Unmarketable Fruit Number Tree-1 Unmarketable Yield Tree-1 (Kg) Total Yield Tree-1(Kg)
SOV 2016 2017 2018 2016 2017 2018 2016 2017 2018 2016 2017 2018 2016 2017 2018
TRT 485.33** 771.66** 1048.82** 66.95** 28.52ns 58.84** 2.66** 7.26** 30.22** 0.096** 0.57** 2.03** 67.79** 21.977ns 73.24**
Error 11.22 69.63 68.76 1.21 9.55 4.09 0.001 0.02 2.37 0.001 0.001 0.04 1.219 9.352 3.865
CV 10.91 14.11 18.29 12.92 18.99 16.83 11.99 15.30 29.16 20.26 11.25 16.37 12.75 18.48 14.89
*, **, and *** significant at P≤0.05, p≤0.01 and p≤0.001 probability levels respectively; ns= not significant, DF= Degree of freedom; SOV= Source of Variation;
Table 3. Mean squares from the three years (2016,2017 and 2018) analysis of variance for avocado yield parameters at Raya valley condition
SOV Mean Squares
Fruit Length (cm) Fruit Diameter (cm) Average single Fruit Weight (gr) Average single fruit Juice Wt. (gr) Average Seed (stone) Wt. (gr)
2016 2017 2018 2016 2017 2018 2016 2017 2018 2016 2017 2018 2016 2017 2018
TRT 17.796** 8.759** 8.653** 2.845ns 4.984** 0.459ns 9783.42* 10275.8** 6310.256** 4213.12** 5376.15* 3874.48** 978.64** 457.76* 10.44**
Error 2.773 0.690 0.505 0.915 0.639 0.125 2374.89 1174.81 197.423 223.69 1203.52 122.23 53.91 109.84 15.00
CV 15.31 7.45 6.22 13.06 10.76 5.18 19.51 14.28 4.94 10.38 21.88 5.50 13.96 20.10 10.44
*, **, and *** significant at P≤0.05, p≤0.01 and p≤0.001 probability levels respectively; ns= not significant, DF= Degree of freedom; SOV= Source of Variation;
Table 4. Mean tree growth performance of avocado varieties grown at Raya Valley
Varieties Tree Growth Parameters
Tree height at First Harvest (m) Canopy Diameter (m) Above Graft union Girth Diameter (mm)
Pinkerton 1.95 2.62 43.927c
Hass 2.25 2.56 50.46bc
Fuerte 2.71 3.23 61.40a
Nabal 2.06 2.53 61.44a
Bacon 2.59 2.74 59.42ab
Ettinger 2.10 2.58 47.17c
LSD 0.84 1.048 9.93
CV 20.34 21.25 29.98
Error Mean Square 0.21 0.33 29.78
Column means with followed by the same letter are not significantly different at P≤0.05 probability level VAR: Variety, Err.MSq. Error Mean Square
[593]
Table 5. Mean performance of avocado varieties for yield and yield components during 2016, 2017 and2018 fruit harvesting years at Raya Valley
VAR Marketable Fruit Number Tree-1 Marketable Yield Tree-1 (Kg) Unmarketable Fruit Number Tree-1 Unmarketable Yield Tree-1 (Kg) Total Yield Tree-1(Kg)
2016 2017 2018 Mean 2016 2017 2018 Mean 2016 2017 2018 Mean 2016 2017 2018 Mean 2016 2017 2018 Mean
Pinkerton 49.92a 81.00a 49.89bc 60.27 15.10a 18.23 17.04a 16.79 1.83a 0.00c 7.00a 2.94 0.41a 0.00c 1.00c 0.47 15.51a 18.23 18.04a 17.26
Hass 35.11b 73.00ab 73.51a 60.54 6.13d 14.32 12.64b 11.03 1.10b 2.00b 4.14b 2.41 0.16c 0.33b 0.76c 0.42 6.29d 14.65 13.40b 11.44
Fuerte 14.44d 44.25de 32.75d 30.48 4.45de 11.02 10.62b 8.70 0.00c 3.67a 3.05bc 2.24 0.00d 1.08a 0.86c 0.65 4.45de 12.10 11.47b 9.34
Nabal 20.00d 39.33e 22.33d 27.22 3.05e 18.30 5.11c 8.82 2.00a 0.00c 0.67c 0.89 0.33b 0.00c 0.20d 0.18 3.37e 18.52 5.31c 9.07
Bacon 36.58b 59.49bc 57.75b 51.27 12.20b 19.10 16.44a 15.91 0.06c 0.00c 8.72a 2.93 0.02d 0.00c 2.44a 0.82 12.22b 19.097 18.88a 16.73
Ettinger 28.08c 57.72cd 35.75cd 40.51 10.11c 16.71 10.22b 12.34 0.00c 0.00c 8.11a 2.70 0.00d 0.00c 1.91b 0.64 10.11c 16.71 12.12b 12.98
LSD 6.093 15.18 15.09 1.999 5.62 3.68 0.18 0.26 2.80 0.057 0.05 0.36 2.009 5.56 3.58
CV 10.91 14.11 18.29 12.93 18.99 16.83 11.99 15.30 29.16 20.26 11.25 16.37 12.75 18.48 14.89
Err. M.Sq. 11.218 69.63 68.76 1.207 9.55 4.09 0.010 0.021 2.37 0.001 0.0007 0.04 1.219 9.35 3.87
Table 6. Mean performance of avocado varieties for fruit quality components during 2016, 2017 and 2018 fruit harvesting years at Raya valley
Fruit Length (cm) Fruit Diameter (cm) Average single Fruit Weight (gr) Average single fruit Juice Wt. (gr) Average Seed (stone) Wt. (gr)
VAR
2016 2017 2018 Mean 2016 2017 2018 Mean 2016 2017 2018 Mean 2016 2017 2018 Mean 2016 2017 2018 Mean
Pinkerton 12.11abc 11.31bc 12.77a 12.06 8.17 7.21b 6.79 7.39 292.67ab 269.42ab 293.18bc 285.09 189.16a 159.21bc 213.39b 187.25 62.65ab 61.08a 31.44bc 51.72
Hass 9.48bcd 9.85cd 8.38c 9.24 6.37 9.89 a 6.1 7.45 206.0bc 143.23 c 196.85d 182.02 128.99 b 99.59 c 135.18c 121.25 37.97c 28.80 b 30.39c 32.39
Fuerte 9.18cd 13.24a 13.02a 11.81 8.64 6.46 b 6.89 7.33 224.33abc 251.56b 329.92a 268.6 112.32bc 154.75bc 242.37a 169.81 52.48b 48.89a 37.54b 46.3
Nabal 7.79d 11.68b 10.90b 10.12 6.46 7.70 b 7.19 7.12 170.67c 318.99a 283.36c 257.67 94.87c 230.43a 202.36b 175.88 24.81c 61.73a 28.41c 38.32
Bacon 12.45ab 8.53d 11.31b 10.76 7.68 6.63b 7.15 7.15 296.00a 217.12b 308.92ab 274.01 166.56a 142.39bc 216.44b 175.13 65.64ab 57.34a 50.02a 57.66
Ettinger 14.24a 12.33ab 12.14ab 12.9 6.62 6.68 b 6.81 6.7 309.00a 239.72b 294.42bc 281.04 172.82a 164.90b 196.61b 178.11 71.93a 55.08a 44.83a 57.28
LSD 3.03 1.51 1.29 1.74 1.45 0.64 88.66 62.36 25.562 27.21 63.11 20.12 13.36 19.07 7.05
CV 15.31 7.45 6.22 13.06 10.76 5.18 19.51 14.28 4.94 10.38 21.88 5.5 13.96 20.1 10.44
Err.
2.77 0.69 0.51 0.92 0.64 0.12 2374.89 1174.81 197.42 223.69 1203.52 122.25 53.91 109.84 15
M.Sq.
[594]
Reference
Bayram S, Arslan MA, Turgutoğlu E and Erkan M. 2012. The performance of some
avocado cultivars under Mediterranean coastal conditions in Turkey. Journal of Food,
Agriculture & Environment 10 (2): 588-592.
Central Statistical Agency (CSA). 2017. Agricultural sample survey 2016/17 (2009 E.C.)
Report on area and production of major crops. Statistical bulletin 586, volume 1.
Addis Abeba, Ethiopia
Chen H, VETM, Ashworth S Xu and Clegg M. 2007. Quantitative Genetic Analysis of
Growth Rate in Avocado. J. Amer. Soc. Hort. Sci. 132(5), 691–696.
Coit J Eliot. 1935. Marketability of Avocado Varieties. California Avocado Association
1935 Yearbook 20: 59-64
Garedew W and Tsegaye B. 2010. ‘Trends of Avocado (Persea americana) Production
and its Constraints in ManaWoreda, Jimma Zone: A Potential Crop for Coffee
Diversification. Journal of Trends in Horticultural Research, 1: 20-26.
Garner LC, VETM Ashworth, MT Clegg and CJ Lovatt. 2008. The Impact of Outcrossing
on Yields of ‘Hass’ Avocado. J. Amer. Soc. Hort. Sci. 133(5):648–652.
Edinéia Dotti Mooz, ED, NM Gaino, MYH Shimano RD, Amancio MHF, Spoto. 2012.
Physical and chemical characterization of the pulp of different varieties of avocado
targeting oil extraction potential. Ciênc. Tecnol. Aliment., Campinas, 32(2), 274-280.
Flores M, Saravia C, Vergara CE, Avila F, Valdés H and Ortiz-Viedma J. 2019. Avocado
Oil: Characteristics, Properties and Applications. A Review. Molecules 24, 2172
doi:10.3390/molecules24112172
Ketema Yilma, Kahsay Berhe, Berhanu G/Medhin, Dirk Hoekstra, Desalegne Ayele and
Kidanemariam G/Hawriat. 2010. A Farmer-based fruit seedling supply system in Dale
Pilot Learning Woreda (PLW): Experiences from IPMS. Dale Woreda Office of
Agriculture and Rural Development, Yirgalem, Ethiopia.
Lovatt C, Zheng Y, Khuong T, Campisi-Pinto S, Crowley D and Rolshausen P. Yield
characteristics of ‘Hass’ avocado trees under California growing conditions. Actas
Proccedings, cultural management and techniques, California Irrigation Management
Information System (3): 336.
National Agricultural Research Institute. January. 2004. Postharvest Care and Market
Preparation. Technical Bulletin No. 18.
Poudel K, Sah MK, Mandal JL and J Shrestha. 2018. Fruit characterization of different
avocado (Persea americana Mill.) genotypes in eastern mid-hills of Nepal. Journal of
Agriculture and Natural Resources (2018) 1(1): 142-148
Shepherd J and Bender G. 2002. A history of the avocado industry in California.
California. Avocado Society, Yearbook. 85, 29-50.
Tan C and S Tan. 2017. Influence of Geographical Origins on the Physicochemical
Properties of Hass Avocado Oil. J. Am. Oil Chem. Soc. 94, 1431-1437.
Tsegay Hailu. 2017. Determinations of Optimum Maturity Indices and Storage life for
Selected Avocado Cultivars Grown in Wendo-genet (Southern Ethiopia). A Thesis.
Addis Ababa University, Addis Ababa Ethiopia pp.13
Wong MC Requejo-Jackman and A Woolf. 2010. What is unrefined, extra virgin cold-
pressed avocado oil? Inform, 21, 189-260.
[595]
Victor FL, Dreher M and AJ Davenport. 2013. Avocado consumption is associated with
better diet quality and nutrient intake, and lower metabolic syndrome risk in US
adults: results from the National Health and Nutrition Examination Survey.
Villa-Rodríguez JA, FJ Molina-Corral, JF Ayala-Zavala, GI Olivas and GA González
Aguilar. 2011. Effect of maturity stage on the content of fatty acids and antioxidant
activity of ‘Hass’ avocado. Food Res. Int. 44, 1231-1237.
Gregoriou C and CV Economides. 1991. Performance of Ettinger, Fuerte, and Hass
Cultivars of Avocado on Two Rootstocks in Cyprus. California Avocado Society
1991 Yearbook 75:87-92
[596]
Evaluation of Table Grape Varieties
for Yield and Quality Traits
Muluken Demelie1, Fasil Tadesse Tewolde1,
Gadisa Lellisa1 and Etsegent Keflu1
1
Ethiopian Institute of Agricultural Research, Debre Zeit Agricultural Research Center, Debre Zeit, Ethiopia1
Abstract
Ethiopia has a potential for grapevine production to harvest twice a year and has a
tradition of wine making that stretches back to the reign of the Queen of Sheba.
Currently, Ethiopia produces its signature wine from two commercial wineries.
Grapevine research has been started at Debre Zeit Agricultural Research Center
and released eleven varieties for wine (eight) and raisin (three) production.
However, there is no variety released specifically for table grape production while
the national demand for table grape is increasing. This research aimed to evaluate
table grape genotypes on yield and yield related traits. The study was conducted in
Debre Zeit and central rift valley at Awash Winery vineyard with six table grape
genotypes. Genotypes were evaluated using a randomized complete block design
with three replications. Data on yield, yield components and quality-related traits
were collected and subjected to analysis of variance using SAS 9.2. The result of
analysis of variance revealed that there was a highly significant (P<0.01) variation
among genotypes for all traits except total soluble solid (TSS) which showed
significant difference (P<0.05). The fruit yield ranges from 1.97 kg/vine up to 6.23
kg/vine. The genotype Flame Tokay and Awash Nigest gave the highest fruit yield
12.46 t/ha and 6.23 kg/vine, 11.9 t/ha and 5.97 kg/vine, respectively. These varieties
have red and pale yellow colors. The cluster weight of both Flam Tokay and Awash
Nigest was 0.77 and 0.65 kg, respectively, which satisfies table grape quality
standard. Berry characteristics; length and diameter were range from 18.22 mm to
22.60 and 15.09 to 19.33 mm, respectively. Seed/berry is an economically important
and distinctive attribute for table grape to which the highest seed/berry value was
2.23 while the lowest was 171. The result showed that the highest and lowest TSS
value was 19.52 and 16.97 Brix from Flam Tokay and Awash Nigest variety,
respectively. The highest sugar content, 11.9 was recorded in Flam Tokay and the
lowest, 7.8 in Awash Nigest cultivar. The total acidity (g/L of H 2SO4) recorded from
3.78 to 6.54 as well as pH value was recorded from 3.19 to 3.93. The varieties of
‘Flam Tokay and Awash Nigest showed high yield and quality suitable for table
grape production in Ethiopia. Further research and development shall be done to fill
the increasing demand for export and domestic market for table grape.
Key words: Table grape, fruit yield, fruit quality, cluster weight and TSS
Introduction
Grapevine (Vitis vinifera L.) is one of the world’s major fruit crop and has the
earliest recorded history (Uddin et al., 2011). Grape is used for making wine,
raisin and for fresh consumption, intended for table use (Bernadine, 2011, Guo et
al., 2017). It has gained popularity for its high nutritive value, excellent in taste,
[597]
multipurpose use and higher economic return (Kamiloglu, 2011). Also the best
raw materials for manufacturing fruit creams without added sugar (Azzouz et al.,
2016). Hence, viticulture has been an important economic activity and also vital in
generating employment in the production of table grapes (Saeed et al., 2017).
Grapevines are adapted to a wide range of agro-ecologies. However, heavy clay
soil, shallow, sodic or poorly drained soils should be avoided (Uddin et al., 2011).
Ethiopia has been growing grapes and producing its signature wine for almost half
a century. It has ample potential for grapevine production and the possibility of
harvesting twice a year. This is due to the proximity to the equator, which has high
plateaux and verdant valleys as well as six climatic zones which are perfect for
grape production. With a tradition of honey wine making that stretches back to the
reign of the Queen of Sheba, the ancient wine-making tradition is gradually being
taken over by the modern form of winemaking. Currently, only two local wine
producers in the country are Castel winery and Awash winery (Marie, 2016).
Grapevine research has been started at Debre Zeit Agricultural Research Center
since, the 1980s with the introduction and collection of genotypes. National
grapevine research program released eleven grapevine varieties; eight for wine
and three for raisin production (DZARC, 2015). Recently, due to an increasing
demand for table grape, cultivation has been established at Koka, Ziway, Meki
and Raya owned by private businesses. Table grape-producing companies such as
Luna, Almeta and Elfora have started table grape production in Ethiopia (Personal
observation, 2018). National table grape production and productivity are at the
infant stage and insignificant to satisfy the local demand while it has the potential
to export for the nearby markets. This is due to the lack of improved technologies,
limited familiarity with table grape cultivation and less attention for Research and
development in the country. Therefore, this research was conducted to evaluate
table grape genotypes for yield and yield-related traits and select candidate
varieties that satisfy international table grape standard.
Methodology
This study was conducted in the central rift valley at Debre Zeit and Awash
Winery vineyard in Merti Jeju of Ethiopia. The experiment consisted of six table
grape varieties which were Awash Nigest, Flame Tokay, Netch Dube, Netch
Debulbul, Key Dube and Cardinal. Varieties were evaluated using RCBD design
with three replications. The experiments consisted of 10 vines per plot in each
replication and were planted with a spacing of 2 m * 2.5 m between plants and
rows, respectively. Vines were trained on cordon trellis with spur pruning
(pruning twice a year in late July and early February) and harvested twice a year
(December and June). Pruning was performed to leaves, 2 –3 buds per spur. Data
collected from the central vine were on yield, yield components and quality-
related traits. Yield and yield components analysis viz: number of cluster/vine,
[598]
cluster weight, fruit yield (kg/vine). Berry analysis viz: weight, length and
diameter. Fruit quality analysis viz: TSS, pH, sugar content and titrable acidity.
All agronomic practices for grapevine production were applied through the whole
experimental period. Collected data were subjected to analysis of variance using
SAS 9.2 and means separation was done using the Least Significant Differences
(LSD) at p = 0.05.

Analysis of variance result for thirteen parameters of yield, yield related and
quality traits in six table grape genotypes is presented in Tables 1, 2 and 3. The
result revealed a highly significant difference (P<0.01) among genotypes for all
traits except total soluble solid (TSS) which showed a significant difference
(P<0.05). The mean values of table grape genotypes for 5, 4 and 4 fruit yield and
yield components, berry characteristics and quality traits are given in Table 1,
Table 2, and Table 3 respectively.
Yield and cluster characteristics

The fruit yield (t/ha and kg/vine) and bunch characteristics (cluster weight,
number of cluster and number of berries per bunch) are presented in Table 1. The
analysis of variance indicates that fruit yield per vine ranges from 1.97 to 6.23
kg/vine. The genotype Flame Tokay gave the highest fruit yield 12.46 t/ha and
6.23 kg/vine, followed by Awash Nigest 11.93 t/ha and 5.97 kg/vine. Cardinal
performed poorly in terms of fruit yield 3.93 t/ha and 1.97 kg/vine. The fruit yield
and bunch weight differed significantly among genotypes. This research result
partially agreed with a study (Uddin et al., 2011) who reported that average fruit
yield per vine ranged from 7.67 to 21.67 Kg. Anirudh et al., (2014) also reported
that fruit yield ranges from 4.39 to 25.07 kg/vine. This research result indicates
that 50% of the genotypes gave yield greater than the range of those obtained at
the international grape growing areas as reported by Zankov and Todorove (1989),
who got 12.4 t ha-1 (Uddin et al., 2011). Yield of grapevine depends on genetic
factors that encode fruitfulness (including fertility of vines under spur pruning of
1year shoots and alternate bearing), age, healthiness, nutrition of plants and
climate conditions (especially temperature and rainfall). Weather conditions
influenced flower and fruit development in the current year and differentiation of
inflorescence primordial for the next year (Jerzy, 2014).
The result indicated that cluster weight ranges from 0.43 to 0.77 kg among
evaluated genotypes. Amongst all genotypes, Flam Tokay and Netch Dube gave
higher cluster weight 0.77 and 0.73. Uddin et al., (2011) reported that cluster
weight of different varieties ranged from 0.21 to 0.60 kg. Anirudh et al., (2014)
also stated that the cluster weight of all studied cultivars was big (more than 0.50
kg) and the highest being 0.89 kg, while the lowest values were 0.65 kg. The
[599]
number of clusters/vine ranged from 4.91 to 11.92. Awash Nigest and Flam Tokay
were found to have higher numbers of clusters per vine, 11.92 and 9.69 (Table 1).
Differences in the number of clusters/vine depends on soil type, green growth
density, fertility factor and number (Mahmoud, et al., 2015). The result showed
that the number of berries per cluster was 64.56 to 142.22 Netch Dube, Netch
Debulbul and Awash Nigest had the highest number of berries per cluster 142.22,
132 and 131.55, respectively. This result disagrees with Uddin et al., (2011) who
found a maximum number of berries per cluster (300) and minimum (172).
Varieties with more berries per bunch are more attractive to consumers and have
high market value as compared to varieties with fewer berries per bunch (Uddin et
al., 2011).
Table 1: Combined Statistical analysis of yield and yield-related traits of table grape varieties evaluated at Debre Zeit and
Merti, 2008 - 2017
Fruit Yield Fruit Yield Cluster weight No No. of berries/
Variety
(qt/ha) (kg/Vine) (kg) cluster/Vine cluster
Awash Nigest 119.43 5.97 0.65 11.93 131.55
Cardinal 39.31 1.97 0.43 4.91 64.56
Flame Tokay 124.55 6.23 0.77 9.69 106.17
Key Dube 52.62 2.63 0.56 5.40 97.39
Netch Debulbul 106.44 5.32 0.69 8.80 132.00
Netch Dube 114.73 5.74 0.73 8.94 142.22
Mean 92.83 4.64 0.64 8.28 112.31
CV 24 24.9 22.04 18.18 9.19
F-test (0.05) ** ** ** ** **
Berry Characteristics: Berry characteristics were studied in terms of length,

diameter and number of seeds per berry. All these parameters are significantly
different within the evaluated genotypes in this study. The berry weight was range
from 3.26 to 5.52 gm. The genotype of Flame Tokay 5.52 gm. while Netch
Debulbul recorded the smallest berry weight 3.26 gm. These results partially in
agreement with Anirudh et al., (2014) who reported that berry weight was found
1.3 to 4.80 gm. and El-Morsy et al., (2017) who found the berry weight was 4.92
to 8.24 gm. within evaluated cultivars. The berry length was highest recorded in
Flame Tokay 22.60 mm while the smallest berry length was recorded in Awash
Nigest 18.22 mm. Anirudh et al., (2014) who reported that the berry length was
found to be 20 to 21.7 mm while El-Morsy et al., (2017) who reported that the
highest values of berry length was obtained 31.5 mm and the lowest berry length
obtained 22.4 mm. The data analysis showed that Netch Dube has the highest
diameter of 19.33 mm while Netch Debulbul showed the lowest berry diameter
15.09 mm. Anirudh et al., (2014) reported the berry diameter of range from 19.7
to 20.7 mm. With respect to the number of seeds per berry, Cardinal and Awash
Nigest have a low number of seeds/berry 1.71 and 1.78 while Netch Dube has the
highest number of seed per berry 2.23. Large-sized berries are usually seeded. The
small-sized seedless grapes are sprayed with growth regulators (gibberellic acid)
to increase the size (Weaver, 1972). The presence of seeds is associated with the
[600]
size of the berries. In general, table grapes have very distinct characteristics and
the most prominent ones are seedless and attractiveness (Asfaw, 2013). Color
important characteristics of grapes and play an important role in marketing. The
best performing varieties have red and pale yellow colored and they are very
attractive in general appearance and deserve a place in the market. The color is an
important quality attribute in table grapes berries contain anthocyanins only in
their berry skins. Anthocyanins are the most prominent pigments in grape skin
(Kamiloglu, 2011).
Table 2. Combined Statistical analysis of the main morphological characteristic of berry/ fruit table grape varieties
evaluate at Debre Zeit and Merti from 2008 – 2017
Berry
Berry Weight Berry diameter No. of seed/ Berry color after
Varieties length
(gm.) (mm) berry maturity
(mm)
Awash Nigest 3.34 18.22 16.07 1.78 Pale yellow
Cardinal 4.82 21.61 18.93 1.71 Red
Flame Tokay 5.51 22.60 17.57 2.05 Red
Key Dube 5.26 20.15 17.75 2.10 Red
Netch Debulbul 3.26 18.37 15.09 2.20 White
Netch Dube 4.90 21.18 19.33 2.23 White
Mean 4.52 20.35 17.46 2.01
CV 12.62 10.29 9.89 16.03
F-test (0.05) ** ** ** **
Quality Characteristics:Total soluble solids (TSS), PH, Sugar content and total
acidity also varied significantly among the genotypes (Table 3). The result
showed that total soluble solid (TSS) were recorded the highest value from Flam
Tokay variety 19.52 Brix and the lowest value was recorded from Awash Nigest
variety 16.97 Brix. This result in agreement with Shawkat, et al (2016) who
reported that total soluble solid was recorded the highest value 17.43 Brix and the
lowest values were recorded 10.47 Brix and Anirudh et al.,(2014) who reported
that the highest TSS was found 18.23 Brix and the lowest TSS (14.30 Brix. The
indicator of grape maturity is the sugar content, determined as the total soluble
solids content in the berry juice and it is measured on a degree-Brix scale. Table
grape growers need to measure the sugar content because it is connected with
grape technological maturity. Soluble solids content depends on the cultivar and
production area. In the world, there are marketing standards for table grapes
cultivars grown. According to the International Organization Vine and Wine
(OIV, 2008) table grapes with a Brix degree equal to or above 16 are considered
as ripe (Maante et al., 2015). The highest sugar content was found in the cultivars
of Flam Tokay 11.90 while the lowest sugar content was recorded in Awash
Nigest cultivar 7.8. Sweetness differed in different sugars; it was higher in
fructose than glucose and sucrose. Hence the sugar variety and content may
influence the sweetness of grape berries. Sugars and organic acids are easily
influenced by culture technique, climate, soil and other environmental conditions
(Guo et al, 2017).
[601]
This study result showed that total acidity (g/L of H2SO4) highest recorded was
found in Netch Dube 6.54 and the lowest value was recorded in Netch Debulbul
3.78. Despite, the importance of total acidity (TA) to the overall flavor, growers
mainly use TSS (sweetness) as an indicator of ripeness and most of the
commercial varieties are considered mature when TSS ranges from 15 to 18%.
Chemical composition of grapes, similar to other species, is influenced by several
factors such as maturity, genotype and growing conditions (Shawkat, et al., 2016).
In this study, the highest PH value was recorded from Flam Tokay 3.93 and the
lowest value was recorded in Awash Nigest variety 3.19. High juice pH to be
undesirable for the production of quality wines as it results in reduced colour
stability and poor taste. PH indicates the strength of the acids present (Guo et al.,
2017). Different factors affect table grapefruit quality are: distribution of berries
on the bunch, uniformity in shape and size, attractive appearance, free of berries
from wind damages, spray residues, insect infestations, or fungal disorders. Other
factors that affect quality include eating quality, which are judged by ripeness,
flavor and texture of the berry and absence or presence of seeds (Asfaw, 2013).
Table 3. Statistical analysis of table grape quality evaluated at Merti in 2017
Total acid in
Variety Sugar content g/L of H2SO4 PH TSS
Flam Tokay 11.90 5.15 3.93 19.52
Awash Nigest 7.80 5.34 3.19 16.97
Netch dube 11.80 6.54 3.29 17.74
Cardinal 11.70 5.03 3.34 18.62
Netchdub 10.40 3.78 3.58 18.23
Keydube 9.40 6.13 3.29 18.65
Mean 10.50 5.33 3.43 18.29
CV 5.07 0.67 0.99 9.07
F-test (0.05) ** ** ** *

The varieties of Flam Tokay showed higher yield and quality while and Awash
Nigest showed higher yield. Awash Nigest disease resistance than other varieties.
These varieties are suitable for table grape for the domestic and export market.
TSS of these varieties is greater than the international table grape quality standard
which is 14 -16°Brix at the time of harvest. Cluster weight at the time of harvest
was also greater than 0.65 kg, which is one of the economically important traits
for table grape quality. Flam Tokay and Awash Nigest varieties are better varieties
for table grapes production in the study area and other similar agro-ecologies.
Demand for table grape has been increasing and the cultivation of introduced
varieties. The industry could expand further through adaptation of research
findings and due to favorable agro-ecologies for harvesting twice a year. It needs
to adopt new varieties and technologies to fill market niches, maximize yield and
quality of table grape production and productivity in Ethiopia. The Research
[602]
Institute, police makers, Wineries Company and Agricultural office emphasize
grapevine research and development to fill the increasing of export and domestic
demand.
Acknowledgment
The authors wishes to express their deep gratitude to Debre Zeit Agricultural
Research Center facility and support the research work, Ethiopian Institute of
Agricultural Research funding budget, Awash Winery for providing experimental
land and research field management and Castel Winery conducting the quality
analysis. Special thanks to Maseresha Yizegawe farm manager of Awash winery
grape farm and Richard Autor General Manager of Awash winery on the effort
and they gave us to accomplish this research work.
Reference
Anirudh Thakur, Arora NK and Som Pal Singh. 2014. Evaluation of some grape varieties in the
arid irrigated region of North-west India. Acta Hort. 785, ISHS 2008.
Asfaw Zelleke, 2013. The Grapevine. Published by: EIAR Printed by: e t h - c a n a p l c. Addis
Ababa ISBN 978- 99944-53-74-x.
Azzouz S, Hermassi I, Toujani M, Belghith A. 2016. Effect of drying temperature on the
rheological characteristics of dried seedless grapes. Food and Bio products processing, 100,
246–254.
El-Morsy, F. M.; Aisha S. A. Gaser and Magda N. Mohamed, 2017. Morphological Description
and Evaluation of Six Newly- Introduced Grape Cultivars under Egyptian Conditions. J. Plant
Production, Mansoura Univ., Vol. 8(11): 1059 – 1070.
El-Morsy FM, Aisha SA Gaser and Magda N Mohamed. 2017. Morphological Description and
Evaluation of Six Newly- Introduced Grape Cultivars under Egyptian Conditions. J. Plant
Production, Mansoura Univ., Vol. 8(11): 1059 – 1070.
Guo Yinshan, Niu Zaozhu, Su Kai, Zhao Jia, Ren Zhihua, Zhao Yuhui, Guo Quan, Gao Hongyan
and Guo Xiuwu. 2017. Composition And Content Analysis Of Sugars And Organic Acids For
45 Grape Cultivars From Northeast Region Of China. Pak. J. Bot., 49(1): 155-160.
Jerzy Lisek. 2014. Evaluation Of Yield And Healthiness Of Twenty Table Grapevine Cultivars
Grown In Central Poland, Journal of Horticultural Research, vol. 22(1): 101-107.
Kamiloglu O. 2011. Influence of Some Cultural Practices on Yield, Fruit Quality and Individual
Anthocyanins of Table Grape Cv. ‘Horoz Karasi. The Journal of Animal & Plant Sciences,
21(2): 2011, Page: 240-245, ISSN: 1018-7081.
Maante M, Vool E, Ratsep R and Karp K. 2015. The effect of genotype on table grapes soluble
solids content. Agronomy Research 13(1), 141–147.
Mahmoud A Aly, Thanaa M Ezz, Mohammed M Harhash, Salah E El- Shenawe and Ahmed
Shehata. 2015. Performance of some Table Grape Cultivars Grafting on Different Rootstocks
in El-Nubaria Region. Asian Journal of Crop Science, 7: 256-266.
Marie-Luce Kühn. 2016. The Emergence of “Old-New” Wine Producing Regions: Ethiopia,
Moldova & Turkey.
Saeed Ahmed, Sergio R Roberto, Ronan C Colombo, Renata Koyama, Muhammad Shahab, and
Reginaldo T. Souza. 2017. Performance of the new seedless grape ‘BRS Isis’ grown in
subtropical area. BIO Web of Conferences 9, 01030, 40th World Congress of Vine and Wine.
[603]
Shawkat Mustafa Atrushy, Shayan Akram Mustafa and Faraidwn Karim Ahmad. 2016. Evaluation
of some seeded cultivars of grape in different location of Kurdistan region, Iraq Kurdistan
Region -Republic of Iraq. Kufa Journal For Agricultural Sciences (279 -266) :8 (2).
Tilahun Hordofa , Michael Menkir, Sileshi Bekele, Teklu Erkossa, 2014. Irrigation and Rain-fed
Crop Production System in Ethiopia. https://www.researchgate.net/publication/267831688
Uddin MM, Shah KU, Rahman, Alam U and Rauf MA. 2011. Evaluation of local and exotic
grapes germplasm at mingora, Swat. Sarhad J. Agric. 27(4): 553-556.
Weaver RJ. 1972. Plant Growth Substances in Agriculture. W.H. Freeman and Co., San Francisco.
[604]

Article 4 Proceeding2019 Crop

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Article 4 Proceeding2019 Crop

Uploaded by

Copyright:

Available Formats

Results of

የኢትዮጵያ የግብርና ምርምር ኢንስቲትዩት

Proceedings of the National Conference on

Copy Editing and Design:

Evaluation and Stability Analysis of Cotton (Gossypium hirsutum L.)

Genetic Diversity Study of G. barbadense L. and Interspecific

Evaluation of Yield and Fiber Qualities Performance of Long-staple Cotton

Yield Performance of Faba Bean (Vicia faba L) Genotypes for High

Yield Performance of Large-Seeded Faba Bean Genotypes for Potential

Evaluation of Large Red Bean Genotypes for Lowland Areas of Ethiopia

Evaluation of Mung Bean Genotypes for Lowland Areas of Ethiopia

Introduction and Evaluation of Quality Protein Maize Hybrid Varieties for

Participatory Demonstration and Evaluation of Improved Maize Technologies in

Evaluation of Lowland Rice (Oryza sativa L.) Varieties in the Rainfed-Lowland

Statistical Analysis of Multi-Environment Trial Data to Enhance

Yield Potential and Stability of Irrigated Spring Bread Wheat Genotypes in

Agronomy and Physiology

Determination of NP Fertilizer Requirement for Newly Released

Effect of NP Fertilizer Rates and Plant Population Density on

Evaluation of QPM Maize Variety under Integrated Uses of Different

Determination of Nitrogen and Phosphorous Fertilizer Rates on Lowland Rice Production

Response of Cotton, (Gossypium hirsutum L.) Varieties to Different Stages of

Coffee, Horticulture, Spices Medicinal and Aromatic Plants

Influence of Plant Population on Bulb Yield and Yield Components of

Medicinal and Aromatic Plants

Enhanced Income Generation through Sweet Basil (Ocimum basilicum L.)

Evaluation Performance of Potato Clones in the Mid-highland Areas of Ethiopia

Adaptation of Korean Potato (Solanum tuberosum L.) Varieties under Irrigation

Verification of Introduced Early Maturing Orange Fleshed Sweetpotato Varieties for

Participatory Orange-fleshed Sweetpotato Variety Selection in Ethiopia:

Effects of Soil Amendments on Growth and Biomass Yield of Early Generation

Evaluation of Cassava (Manihot esculenta Crantz) Genotypes for Total

Evaluation of Improved Turmeric Boiler at Different Boiling Durations

Evaluation of the Effects of Different Curing and Drying Methods on

Evaluation of Table Grape Varieties for Yield and Quality Traits

Diriba Geleti (PhD)

Evaluation of Food Barley Genotypes for Grain

Keywords: ANOVA, GGE biplot, Stability, Grain yield

In general, barley remains an important crop in the Ethiopian agriculture because

The National Barley Research program at Holetta and Kulumsa in collaboration

Materials and Methods

Table 1. Description of testing sites

No. Testing sites Altitude (m) Rainfall (mm) Longitude Latitude

Table 2. Lists of genotypes and environments used for the experiment

Combined analysis of variance showed that genotypes, environment and genotype

Significant at 5% and 1% probability level, ns=non significant, §these traits were

Stability of the test genotypes across environments

The “which-won-where” pattern of a genotype by environment data set is an

Materials and Methods

Yij = µ + Gi +Bj + eij and Yijk = µ + Gi +Ej + GEij +Bk(j) + eijk.

Trt Genotype Year Trt Genotype Year Loc Year Env

Results and Discussion

SV DF DHE DMA PLH SC(DF)§ NB(DF)§ TKW HLW GYLD

Extract Protein Beta glucan

Keywords: Stability analysis, GGEbiplot, AMMI analysis, moisture stress

Material and Methods

Entry Set I Entry Set II

Yijly = µ + Yy + Ll + (R/YL)jyl + Gi + GLil + GYiy + GLYily + eijrly

Results and Discussion

SV DF DHE DMA TKW PLH DF HLW DF GYLD

Figure 3: The which-win-where view of the GGEBiplot analysis of the Experiment I

GGEBiplot analysis for Environment II

Adane Chofere1, Shimeles Gezahegn1, Workeneh Mekasa3,