University of Ghana http://ugspace.ug.edu.gh
GENETIC IMPROVEMENT OF COWPEA (Vigna unguiculata L. WALP) FOR 
EARLINESS AND DROUGHT TOLERANCE  
 
 
 
 
BY 
 
TONY MACINNES ALFRED NGALAMU 
STUDENT NUMBER: 10509367 
 
 
THIS THESIS IS SUBMITTED TO THE UNIVERSITY OF GHANA, LEGON IN 
PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE AWARD OF 
DOCTOR OF PHILOSOPHY DEGREE IN PLANT BREEDING   
 
 
 
 
 
 
 
 
 
 
     
WEST AFRICA CENTER FOR CROP IMPROVEMENT 
COLLEGE OF BASIC AND APPLIED SCIENCES 
UNIVERSITY OF GHANA  
LEGON    
 
DECEMBER, 2018 
i 
 
University of Ghana http://ugspace.ug.edu.gh
DECLARATION  
 
I do hereby declare that except for citations to findings of other researchers, duly cited, this 
work is my original research and that neither part nor whole has been presented elsewhere for 
the award of a degree.   
…………………………………. 
                                                                                      TONY NGALAMU                                                                                                  
STUDENT  
 
…………………………………      
                                                                                                    PROF. KWADWO OFORI  
SUPERVISOR   
 
………………………………….                                                                                                                 
                                                                                       DR. BEATRICE ELOHOR IFIE 
SUPERVISOR 
 
…………………………………. 
DR. JOHN SAVIOUR YAW ELEBLU 
SUPERVISOR 
 
…………………………………..   
PROF. SILVESTRO KAKA MESEKA                                                                                                               
SUPERVISOR 
 
 
 
 
 
 
 
 
 
 
ii 
 
University of Ghana http://ugspace.ug.edu.gh
ABSTRACT 
The number of people suffering from chronic undernourishment in sub-Saharan Africa has 
increased. The potential of the cowpea to be the crop of the future to address food and 
nutrition insecurity is evident in sub-Saharan Africa. This study was carried out to evaluate 
cowpea accessions for early maturity and drought tolerant genotypes with potential for higher 
yields for smallholder farmers in South Sudan. One hundred and six cowpea accessions were 
sourced from different backgrounds. Their genetic diversity was assessed using both agro-
morphological traits and molecular approaches. Nine discriminatory clusters yielded 
cophenetic correlation coefficient values of 0.76. Results of molecular study revealed genetic 
divergence among the assembled cowpea panel creating an opportunity for population 
development through introgression of new alleles. The SNP markers used in this study could 
be utilized to analyse and group collections in the future. Forty-nine accessions selected from 
106 genotypes were screened under well-watered and drought stressed conditions for 
selection of parental lines for population improvement using 7×7 lattice square design. Five 
cowpea accessions from South Sudan and three from Burkina Faso, Niger and Nigeria were 
identified and selected for crosses. Hybridisation was carried out using hand emasculation 
and pollination. The North Carolina II mating design generated 15 F1 populations. The F1s 
were then backcrossed to the local parent to develop BC1F1. The 15 BC1F1 populations with 
their parents as checks were evaluated using an alpha lattice design to advance promising 
genotypes to the next round of backcrossing cycles. The resultant F2 offspring arranged in a 
split-plot design were evaluated to study the combining ability for early maturity and drought 
tolerance under well-watered and drought-stressed conditions using rainout shelters. At 
backcross 3, (BC3F1), the population was selfed twice before running multi-location trials in 
Legon (5°38'N, 0°10'E), Fumesua (6°41′N, 1°28′W), and Nyankpala (9°24′N, 0°59′W), 
Ghana. The study included a total population of 9000 plants, with 3000 plants per 
iii 
 
University of Ghana http://ugspace.ug.edu.gh
experimental site. The trials were laid out in a 5×5 lattice square design with three 
replications and two watering regimes. The plot size was 1.2 m2, the distance between plants 
was 20 cm and 60 cm between rows. Data were collected on agronomic traits and subjected 
to analysis of variance and path analysis. Stability analysis identified best performers for 
earliness, drought tolerance, and seed yield stability. Additive main effects, multiplicative 
interaction (AMMI) stability analysis, and GGE biplot analysis ranked genotypes based on 
their performance in relation to environment, whereas Eberhart and Russell (1966) and Finlay 
and Wricke’s (1963) stability approaches gave holistic information about those genotypes 
with outstanding seed yield stability. Genotype by environment analysis confirmed that 
environment accounted for 63.8% of the variability in the experiments, genotype accounted 
only for 10.8%, and the interaction between genotype and environment accounted for 24.4% 
of the variation. Ten genotypes, A1B×D, A1B×I, A1B×M, BA×D, BA×M, BA×I, L1B×D, 
TA×M, TA×D, and TA×M that mature in 60 days or less after planting were identified. Four 
genotypes, BC×M, L1B×I, TA×M, and A1B×M, were found to combine drought tolerance 
with stable yield potential. Two genotypes, A1B×M and TA×M, demonstrated both early 
maturity and drought tolerance. Five recommendations were drowned from this study for 
future research. 
 
 
 
 
 
 
 
 
iv 
 
University of Ghana http://ugspace.ug.edu.gh
ACKNOWLEDGEMENTS 
My unfathomable appreciation goes to my supervisors Prof. Kwadwo Ofori, Prof. Silvestro 
Meseka, John Eleblu, PhD and Beatrice Ifie, PhD for methodically reading through every 
chapter and making valuable propositions. Additionally, being very caring and understanding 
during epochs I really felt distressed.  I also thank Prof. Emeritus Vernon Gracen and Prof. 
Emeritus Glen Shinn for their undisputed support and valuable leadership. I am grateful to 
Mr. Christian Amenakpor, Mr. Kwabena Bediako and Mr. Michael Teye for their valuable 
assistance, moral support and encouragement. I am also indebted to the management of 
INTRA-ACP, DAAD and ACE for the scholarships, University of Ghana for admission, 
West Africa Centre for Crop Improvement in particular for training. I thank Prof. Eric 
Danquah for guidance and timely support from the inception of the study. Many thanks to the 
management of the University of Juba for release and support. I also wished to recognise 
managements of the Plant Genetic Resource Research Institute and Savanna Agricultural 
Research Institute (SARI), in Ghana, International Institute of Tropical Agriculture, Cowpea 
Improvement Unit (Kano), Nigeria, INERA (Burkina Faso) and ITRA (Togo), for providing 
part of the germplasms used in the study.  The entire Cohort 8 member, I salute you guys. I 
would like to convey my sincere gratitude and thanks to all those who have been instrumental 
during the course of my study at the University of Ghana, and during my stay in Ghana. I 
thank all my friends, Dr. Itai Makanda, Eng. Julius Ngalamu, Four Cousins and Ms. Asia 
MacInnes for their encouragement. My sincere gratitude also goes to the two Farm Managers 
(Crop Science and WACCI) and their subordinates for all the assistance during the field 
work, the laugh and encouragement. To all members of the South Sudan community in 
Ghana particularly Ambassador Amoi Juma Dino, his wife Clara and children Lorika and 
Illam I say “Aro boya” for the comradeship and moral support. To the Love of my life 
Bazilika Tuna and our children Toa Ni, Yau Ni and Daniel, may the almighty God reward 
v 
 
University of Ghana http://ugspace.ug.edu.gh
you for the understanding, care, and support during the coursework, research phase of the 
journey and thesis preparation. Indeed, this accomplishment is a result of support and 
supervision from many actors. It is not possible to mention their names individually, but I 
dully recognised and appreciate your valuable contributions. I will conclude this 
acknowledgement by thanking you and asking God to richly bless you all. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
vi 
 
University of Ghana http://ugspace.ug.edu.gh
DEDICATION 
Unreservedly I dedicate this thesis, which shows all the research work I carried out here at 
the University of Ghana, West Africa Centre for Crop Improvement, to God Almighty, 
specially my uncle and friend late Taban Juma Durufu, my lovely, compassionate and 
empathetic wife Mrs Bazilika Tuna, and to our delightful sons Toa Ni, Daniel and daughters 
Mariam, Yau Ni, my siblings Katra; Joseph, Lemi, Emmanuel, Elinana and beloved parents 
Mr. MacInnes Ngalamu and Ms. Drusilla Galio (My Prayer Warrior).  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
vii 
 
University of Ghana http://ugspace.ug.edu.gh
TABLE OF CONTENTS 
DECLARATION ................................................................................................................... ii 
ABSTRACT ......................................................................................................................... iii 
ACKNOWLEDGEMENTS ................................................................................................... v 
DEDICATION ..................................................................................................................... vii 
TABLE OF CONTENTS ................................................................................................... viii 
LIST OF TABLES ................................................................................................................ xi 
LIST OF FIGURES ............................................................................................................. xiv 
LIST OF PLATES ............................................................................................................... xvi 
LIST OF ABBREVIATIONS ............................................................................................ xvii 
CHAPTER ONE .................................................................................................................... 1 
1.0 GENERAL INTRODUCTION ........................................................................................ 1 
CHAPTER TWO.................................................................................................................... 4 
2.0 LITERATURE REVIEW ................................................................................................. 4 
2.1 Origin, Domestication Distribution and Types of Cowpea .............................................. 4 
2.2 Climatic and Edaphic Factors Influencing Cowpea Production ...................................... 5 
2. 3 Constraints to Cowpea Production .................................................................................. 7 
2.4 Drought Stress and Cowpea Production........................................................................... 8 
2.5 Mechanisms of Tolerance to Drought Stress ................................................................... 9 
2.5.1 Physiological Mechanisms.................................................................................... 9 
2.5.2 Genetic Mechanism ............................................................................................ 10 
2.6 Mechanisms of Drought Stress Avoidance .................................................................... 12 
2.6.1 Morphological Mechanism ................................................................................. 12 
2.6.1.1 Escape (drought avoidance) ............................................................................. 12 
2.6.1.2 Avoidance (dehydration post-ponders) ............................................................ 13 
2.6.1.3 Tolerance (dehydration tolerators) ................................................................... 13 
2.6.2 Physiological Mechanism ................................................................................... 14 
2.6.3 Molecular and Biochemical Mechanism ............................................................ 15 
2.7 Breeding for Tolerance to Drought Stress...................................................................... 16 
2.7.1 Conventional Breeding Efforts ........................................................................... 17 
2.7.2 Molecular Approaches ........................................................................................ 17 
2.8 Significance of General and Specific Combing Ability in Crop Improvement ............. 18 
2.8.1 Combing Ability for Early Maturity in Cowpea ................................................. 19 
2.8.2 Combining Ability for Drought Tolerance in Cowpea ....................................... 20 
2.9 Assessment of Genetic Potential of Cowpea for Breeding Program ............................. 22 
2.9.1 Genetic Diversity Studies in Cowpea ................................................................. 22 
2.9.2 Importance of Genetic Distance in Hybridisation .............................................. 24 
2.9.3 Genetic Information Derived from Allelic Pattern in Germplasm ..................... 24 
2.9.4 Importance of Selection Index in Crop Improvement ........................................ 25 
CHAPTER THREE .............................................................................................................. 27 
3.0 ASSESSMENT OF GENETIC DIVERSITY AMONG COWPEA ACCESSIONS FOR 
EARLINESS AND DROUGHT TOLERANCE ................................................................. 27 
3.1 Introduction .................................................................................................................... 27 
viii 
 
University of Ghana http://ugspace.ug.edu.gh
3.2 Materials and Methods ................................................................................................... 29 
3.2.1 Experimental Materials ....................................................................................... 29 
3.2.2 Experimental Site ................................................................................................ 29 
3.2.3 Morphological Characterisation .......................................................................... 30 
3.2.4 Molecular Characterisation ................................................................................. 31 
3.2.5 Data Collection and Statistical Analysis ............................................................. 32 
3.2.5.1 Morphological data .......................................................................................... 32 
3.2.5.2 Molecular data ................................................................................................. 33 
3.2.5.3 Statistical Analysis ........................................................................................... 34 
3.3 Results ............................................................................................................................ 35 
3.3.1 Morphological ..................................................................................................... 35 
3.3.2 Molecular Genetic Structure ............................................................................... 47 
3.4 Discussion ...................................................................................................................... 53 
3.5 Conclusions .................................................................................................................... 61 
CHAPTER FOUR ................................................................................................................ 63 
4.0 IDENTIFICATION OF DROUGHT TOLERANCE AMONG DIVERSE COWPEA 
GERMPLASM ..................................................................................................................... 63 
4.1 Introduction .................................................................................................................... 63 
4.2 Materials and Methods ................................................................................................... 64 
4.2.1 Genetic Materials and Experimental Site ........................................................... 64 
4.2.2 Data Collection ................................................................................................... 65 
4.2.3 Data Analysis ...................................................................................................... 66 
4.3 Results ............................................................................................................................ 67 
4.3.1 Leaf Wilting Index (LWI), Levels and Types of Drought Tolerance ................. 67 
4.3.2 Wilting Scales (Delayed Leaf Senescence) ........................................................ 69 
4.3.3 Score of Drought Stress Index (DSI) .................................................................. 69 
4.3.4 Soil Moisture Content (SMC) ............................................................................. 69 
4.3.5 Re-growth and Stem Greenness (STG)............................................................... 69 
4.3.6 Correlation between Leaf Wilting Index and Drought Tolerance Traits ............ 74 
4.3.7 Phenotypic Correlation ....................................................................................... 76 
4.4 Discussion ...................................................................................................................... 78 
4.5 Conclusions .................................................................................................................... 81 
CHAPTER FIVE .................................................................................................................. 82 
5.0 COMBINING ABILITY FOR DROUGHT TOLERANCE AND EARLINESS IN 
COWPEA ............................................................................................................................. 82 
5.1 Introduction .................................................................................................................... 82 
5.2 Materials and Methods ................................................................................................... 83 
5.2.1 Experimental Site ................................................................................................ 83 
5.2.2 Combining Ability Study for Drought-stress and Earliness ............................... 83 
5.2.3 Data Collection ................................................................................................... 84 
5.2.4 Data Analysis ...................................................................................................... 85 
5.3 Results ............................................................................................................................ 88 
5.3.1 Response of F2 Populations and Parental Lines to Drought Stress ..................... 88 
ix 
 
University of Ghana http://ugspace.ug.edu.gh
5.3.2 Estimates of General and Specific Combining Ability Effects .......................... 94 
5.3.3 Computation of Earliness in Cowpea ................................................................. 97 
5.4 Discussion .................................................................................................................... 103 
5.5 Conclusions .................................................................................................................. 106 
CHAPTER SIX .................................................................................................................. 108 
6.0 YIELD AND YIELD STABILITY OF COWPEA GENOTYPES ............................. 108 
6.1 Introduction .................................................................................................................. 108 
6.2 Materials and Methods ................................................................................................. 111 
6.2.1 Population Development ................................................................................... 111 
6.2.2 Genetic Materials and Environment of the Test Locations .............................. 112 
6.2.3 Experimental Layout ......................................................................................... 113 
6.2.4 Data Collection ................................................................................................. 115 
6.2.5 Data Analysis .................................................................................................... 117 
6.3 Results .......................................................................................................................... 118 
6.3.1 Crop Performance at Nyankpala ....................................................................... 119 
6.3.2 Crop Performance at Legon .............................................................................. 123 
6.3.3 Crop Performance at Fumesua .......................................................................... 127 
6.3.4 Combined Analysis of Variance ....................................................................... 131 
6.3.5 AMMI Stability Analysis .................................................................................. 136 
6.3.6 GGE Biplot Analysis ........................................................................................ 139 
6.3.7 Adaptability and Yield Stability Analyses ........................................................ 143 
6.4 Discussion .................................................................................................................... 147 
6.5 Conclusions .................................................................................................................. 151 
CHAPTER SEVEN ............................................................................................................ 152 
7.0 GENERAL CONCLUSIONS AND RECOMMENDATIONS ................................... 152 
7.1 General Conclusions .................................................................................................... 152 
7.2 Recommendations ........................................................................................................ 154 
BIBLIOGRAPHY AND APPENDICES ........................................................................... 155 
BIBLIOGRAPHY .............................................................................................................. 155 
APPENDICES .................................................................................................................... 172 
 
 
 
 
 
 
 
x 
 
University of Ghana http://ugspace.ug.edu.gh
LIST OF TABLES 
 
Table 3.1: Showing region, countries and number of cowpea genotypes assembled .............. 30 
 
Table 3. 2:  Score of qualitative characteristics of cowpea accessions .................................... 33 
 
Table 3.3:  Frequency distribution of qualitative variables for cowpea accessions studied .... 37 
 
Table 3.4: Mean values of agro-morphological traits of 106 cowpea accessions ................... 43 
 
Table 3.5: Principal components loading for 106 cowpea accessions variables ..................... 45 
 
Table 3.6:  Mean Fst value and expected heterozygosity within cluster ................................. 48 
 
Table 3.7: Allele-frequency divergence among clusters (Net nucleotide distance) ................ 49 
 
Table 4.1: Eleven variables used to categorise the drought tolerance of the 49 cowpea 
accessions .................................................................................................................. 65 
 
Table 4.2:  Mean squares of 10 measured traits of 49 cowpea genotypes ............................... 67 
 
Table 4.3:  Drought tolerance scores* and moisture levels in 49 cowpea genotypes .............. 68 
 
Table 4.4:Descriptive statistic for quantitative parameters measured under drought stress     
condition ................................................................................................................... 70 
 
Table 4.5:  Mean of traits of 49 cowpea genotypes evaluated across two environments ........ 72 
 
Table 4.6:  Distribution of 49 cowpea accessions in five clusters according to quantitative 
traits........................................................................................................................... 73 
 
Table 4.7:  Mean values of cowpea agronomic traits measured under stressed and well-
watered conditions .................................................................................................... 74 
 
Table 4.8: Associations among 10 drought related traits ......................................................... 74 
 
Table 4.9: Phenotypic correlation coefficients between morphological, grain yield and sub- 
yield components of cowpea genotypes measured under drought condition ........... 77 
 
xi 
 
University of Ghana http://ugspace.ug.edu.gh
Table 5.1: Farmers preferred drought tolerant cowpea genotypes used as parental lines in 
North Carolina II mating design ............................................................................... 84 
 
Table 5.2: Format of analysis of variance for North Carolina mating design II ...................... 86 
 
Table 5.3: Mean squares for responses of 15 F2 populations, eight parental lines and two 
checks subjected to drought stress and well-watered conditions at reproductive stage
................................................................................................................................... 88 
 
Table 5.4: Estimated mean squares for GCA and SCA for delayed leaf senescence .............. 88 
 
Table 5.5: Mean squares for 15F2 cowpea population, eight parents and two checks subjected 
to drought stress at reproductive stage ...................................................................... 90 
 
Table 5.6: Means of yield and yield components in F2 cowpea population and their eight 
parents evaluated under drought-stress and well-watered conditions....................... 91 
 
Table 5.7: Mean squares, variance components for the responses of 15 F2 population 
evaluated under well-watered and drought stress condition ..................................... 93 
 
Table 5.8: Estimates of general combining ability effects of yield and yield components for 
both male and female ................................................................................................ 95 
 
Table 5.9: Estimates of specific combining ability effects for crosses under well-watered and 
drought stress condition ............................................................................................ 96 
 
Table 5.10: Mean squares for 15 F2 population, eight parents and two checks measured under 
well-watered condition.............................................................................................. 99 
 
Table 5.11: Estimates of general and specific combining ability among eight parental lines 
and their F2 for the number of days to the first flower, days to 50 % flowering, days 
to the first mature pod and 95% maturity in cowpea .............................................. 100 
 
Table 5.12: Mean square for performance and genetic components for days to the first flower, 
days to 50 % flowering, maturity in 15 crosses of cowpea .................................... 102 
 
Table 6.1: List of 15 BC3F3 cowpea genotypes, eight parents and two checks evaluated in 
three locations in 2018 ............................................................................................ 114 
 
Table 6.2: Description of test sites for 25 cowpea genotypes evaluated across three diverse 
agro-ecologies in Ghana ......................................................................................... 116 
xii 
 
University of Ghana http://ugspace.ug.edu.gh
 
Table 6.3: Seed yield performance of 25 genotypes and other agronomic traits measured 
under drought stressed and well-watered conditions at Nyankpala ........................ 121 
 
Table 6.4: Means from analysis of variance measured under drought stress and well-watered 
conditions at Nyankpala 2018 ................................................................................. 122 
 
Table 6.5: Seed yield performance of 25 genotypes and other agronomic traits measured 
under drought stressed and well-watered conditions at Legon ............................... 125 
 
Table 6.6: Means from analysis of variance measured under drought stress and well-watered 
conditions at Legon 2018 ........................................................................................ 126 
 
Table 6.7: Seed yield performance of 25 genotypes and other agronomic traits measured 
under drought stressed and well-watered conditions at Fumesua ........................... 129 
 
Table 6.8: Means from analysis of variance measured under drought stress and well-watered 
conditions at Fumesua 2018.................................................................................... 130 
 
Table 6.9: Mean squares from the combined analysis of variance of seed yield and yield 
components measured under drought stress and well-watered conditions ............. 133 
 
Table 6.10: Mean squares from the combined analysis of variance for seed yield and yield 
components of 25 cowpea genotypes measured across three test sites under drought 
stress and well-watered conditions ......................................................................... 135 
 
Table 6.11: AMMI analysis of variance for seed yield across locations ............................... 137 
 
Table 6.12: Means estimates of adaptability and phenotypic stability for 25 cowpea 
genotypes evaluated across three locations under drought stressed and well-watered 
conditions ................................................................................................................ 144 
 
Table 6.13: Overall mean performances of 25 cowpea genotypes evaluated across sites under 
drought stressed and well-watered conditions ........................................................ 146 
 
 
 
 
 
 
xiii 
 
University of Ghana http://ugspace.ug.edu.gh
LIST OF FIGURES  
Fig. 3. 1 Dendrogram resulting from the analysis of 106 cowpea accessions (based on 13 
phenotypic characters) obtained by UPGMA using the overall distance of 
Euclidean. The cophenetic coefficient (r) was 0.76 .................................................. 42 
 
Fig. 3. 2 Eigenvalue and principal components responsible for variability among 106 cowpea 
accessions .................................................................................................................. 46 
 
Fig. 3. 3 Display of population structure of the 106 cowpea genotypes .................................. 47 
 
Fig. 3. 4 Expected heterozygosity and mean fixation index (Fst) for each cluster .................. 48 
 
Fig. 3. 5 Phylogenetic tree depicting genetic similarities and dissimilarity among 106 cowpea 
accessions .................................................................................................................. 50 
 
Fig. 3. 6 Genetic variation of 106 cowpea accessions analysed by SNP marker at K=6 ........ 52 
 
Fig.4.1 Dendrogram resulting from the analysis of 49 cowpea accessions (agro-
morphological traits) obtained by Ward method using the overall distance of 
Euclidean................................................................................................................... 75 
 
Fig 5. 1 Chlorophyll content of the genotypes under drought stress condition and re-watering 
at 30, 45 and 60 days after planting .......................................................................... 89 
 
Fig. 6. 1 Estimated means for the twenty-five genotypes selected under drought stressed and 
well-watered    conditions at (P<0.05) .................................................................... 118 
 
Fig. 6. 2 Boxplot estimated means for the significant effects under drought stress (DS) 
conditions at (P<0.05) ............................................................................................. 119 
 
Fig. 6. 3 Estimated mean squares of seed yield across sites under drought stressed (DS) and 
well-watered (WW) conditions ............................................................................... 131 
 
Fig. 6. 4 Path diagram cause (traits studied) and effect (seed yield) relationship of 25 cowpea 
genotypes ................................................................................................................ 132 
 
Fig. 6. 5 Estimated means of overall seed yield of 25 cowpea genotypes at (P<0.05) .......... 134 
 
Fig. 6. 6 AMMI 1biplot for seed yield of 25 cowpea genotypes denoted by numbers and six 
environments using genotypic and environmental scores ...................................... 138 
 
Fig. 6. 7 Polygon view of genotype by environment interaction of 25 cowpea genotypes ... 139 
xiv 
 
University of Ghana http://ugspace.ug.edu.gh
 
Fig. 6. 8 Discrimination power and representativeness of test sites ...................................... 141 
 
Fig. 6. 9 Mean vs Stability view of 25 cowpea genotype main effect and GEI effect. ......... 142 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
xv 
 
University of Ghana http://ugspace.ug.edu.gh
LIST OF PLATES 
 
Plate 3. 1: Plant growth habits observed: Apagu 1B, Laduni 1A and Titinwa A representing 
erect (A), climbing (B) and prostrate (C) types of cowpea accessions respectively 36 
 
Plate 3. 2: Cowpea accessions Laduni 1B (left) and Apagu 1B (right) pigmented and non-
pigmented pods respectively ..................................................................................... 38 
 
Plate 3. 3: Variability in cowpea terminal leaflet shapes observed ........................................ 39 
 
Plate 3. 4: Cowpea accessions Apagu 1B (A), AGRAC−316 (B) and Beledi C (C), showing 
erect, pendant and 30°−90° types of pod attachment respectively ........................... 40 
 
Plate 3. 5: A mixture of cowpea seeds from different accessions showing diversity of seed 
characteristics ............................................................................................................ 40 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
xvi 
 
University of Ghana http://ugspace.ug.edu.gh
LIST OF ABBREVIATIONS 
 
AEC Average Environment Axis 
AFLP Amplified Fragment Length Polymorphisms 
AMMI Additive Mean Effect and Multiplicative Interaction 
AMP Arithmetic Means Productivity Index 
ANOVA Analysis of Variance 
AQPs Aquaporins 
BC Backcross 
BSCGD Broad Sense Coefficient of Genetic Determination 
CAPS Cleaved Amplified Polymorphic Sequence  
CCC Cophenetic Correlation Coefficient 
CGIAR Consultative Group on International Agricultural Research  
CPMV Cophenetic Correlation Coefficient 
CRI Crops Research Institute 
CV Coefficient of Variation  
DAP Days After Planting  
DF Degree of Freedom 
DII Drought Intensity Index 
DLS Delayed Leaf Senescence  
DNA Deoxyribonucleic Acid 
DS Drought Stressed 
DSI Drought Stress Index 
DSI Drought Susceptibility Index 
ENV Environments  
FAO Food and Agriculture Organization  
Fst Fixation index 
GBS Genotype by Sequencing  
GCA General Combining Ability 
GEI Genotype by Environment Interaction  
GGE Genotype and Genotype by Environment 
GLYD Grain Yield 
GMP Geometric Mean Productivity 
GR Genotype Response  
GCV Genetic Coefficient of Variation 
GWAS Genome Wide Association Selection 
HI Harvest Index 
IBOGR International Board for Plant Genetic Resource 
IDC Iron Deficiency Chlorosis 
IITA International Institute of Tropical Agriculture 
INERA Institut de l’Environnement et de Recherches Agricoles 
INRAN Institut national de la recherche agronomique 
IPCA Interaction Principal Component 
ITRA Institut Togolaise de Recherche Agronomique 
LAI Leaf Area Index 
LEA Late Embryogenesis Abundant Proteins 
LSD Least Significant Difference 
LWI Leaf Wilting Index 
MAB Marker Assisted Breeding 
xvii 
 
University of Ghana http://ugspace.ug.edu.gh
MAF Minor Allele Frequency  
MAS Marker Assisted Selection  
NARO National Agricultural Research Organisation 
NCII North Carolina II Mating Design 
NDVI Normalised Difference Vegetation Index 
NJ Neighbour-Jointing  
NPP Number of Pods Per Plant 
NSCGD Narrow Sense Coefficient of Genetic Determination 
NSP Number of Seeds Per Pod 
OA Osmotic Adjustment  
PCA Principal Component Analysis 
PGRRI Plant Genetic Resources Research Institute 
QTLs Quantitative Trait Loci 
RAPD Random Amplified Polymorphic DNA 
RELP Restriction Fragment Length Polymorphism   
SARI Savanna Agricultural Research Institute 
SCA Specific Combining Ability  
SEM Structural Equation Modelling 
SI Stress Intensity 
SMC Soil Moisture Content  
SNPs Single Nucleotide Polymorphisms    
SPAD Chlorophyll Meter 
SSA sub-Sahara Africa 
SSRs Simple Sequence Repeats 
STG Stay Green 
STI Stress Tolerance Index 
TOL Stress Tolerance 
UCR University of California, Riverside  
UG University of Ghana 
UJ University of Juba 
UPGMA Unweighted Pair−Group Average 
WACCI West Africa Centre for Crop Improvement 
WW Well-Watered  
WSC Water Soluble Carbohydrate 
xviii 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
CHAPTER ONE 
1.0 GENERAL INTRODUCTION 
 
Cowpea (Vigna unguiculata (L) Walp) is a diploid (2n = 2x =22) and an important grain 
legume capable of producing appreciable yields under 500 mm of rainfall in sub-Saharan 
Africa (SSA).  It is estimated that Africa produces more than 96.1% of the cowpea in the 
world with a grain production of 6.99 million tonnes under 12.32 million hectares 
(FAOSTAT, 2016). Compared to other annual crops grown in drought-prone areas in SSA, 
cowpea could be one of the best options to grow, although drought stress is still a major 
production constraint for the crop (Fatokun et al., 2012). The crop provides food for both 
humans and livestock and also serves as a dependable revenue generating commodity 
(Ajeigbe et al., 2008). Cowpea is widely grown by nearly all the smallholder farmers in all 
agro-ecological zones across SSA for its grain. In South Sudan, cowpea ranks third and first 
among the legumes and is produced mostly by smallholder farmers. 
Poor cowpea yield remains one of the major production problems in SSA where yields are as 
low as 0.4 ton/ha (Haruna and Usman., 2013). However, higher yields of 2.0 t/ha and above 
have been recorded when improved cultivars are grown as pure stands under high rainfall and 
adequate management practices (Agyeman et al., 2014). Collection of adapted farmers’ 
preferred varieties and introduction of improved varieties from international and national 
research institutions could provide a huge genetic pool for a functional cowpea breeding 
program in South Sudan.  
Yield reductions in cowpea have been attributed to a number of biotic and abiotic stresses. 
Drought stress is the most important abiotic stress disrupting cowpea production in the SSA 
countries. This prevalence could be due to high variability in amount and distribution of 
rainfall during the cropping season. Both intermittent and terminal drought stresses occur in 
1 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
South Sudan but the terminal drought is the most important because it impacts directly on 
pod formation and filling. Terminal drought generally leads to significant reduction in yield 
(Singh, 2007) and poor seed quality (Beebe et al., 2013). 
Extensive research efforts have been made to improve the efficiency of selection for drought 
tolerance based on specific physiological traits and yield. However, a good number of the 
approaches failed due to genotype by environment interaction effect (GEI) and the lack of 
precise and efficient drought screening methods. Several researchers (Mai-Kodomi et al., 
1999a; Singh et al., 1999, Fatokun, et al., 2012) have indicated that screening cowpea 
genotypes for tolerance to drought at seedling stage has a good potential of differentiating 
among contrasting entries.  
The recurrent frequent occurrence of drought stress in the Iron-stone plateau, semi-arid and 
floodplains agro-ecological zones of South Sudan causes cowpea varieties to flower and 
mature faster, but with poor seed yield and low harvest index (Ngalamu et al., 2015). Early 
maturing varieties need to be developed for both leaf and high grain yield in order to address 
the threat posed by drought stress. Moreover, breeding for earliness and drought tolerance 
should be the research focus of cowpea breeding in South Sudan where its agricultural 
research program and activities are still in early developmental stages. There is little 
information however on the genetic divergence for earliness and drought tolerance in the 
existing adapted varieties preferred by farmers in South Sudan (Ngalamu et al., 2017).  
The new developments in plant molecular genetics have provided powerful tools for diversity 
studies and marker-assisted selection of polygenic traits such as drought tolerance and grain 
yield, therefore, complimenting conventional breeding tools. There is the need to develop 
new varieties with farmer and consumer preferred traits as well as climate change resilient. 
Thus, there is a need to screen and identify drought tolerant genotypes that could be use as 
2 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
the source of a gene conferring tolerance to drought stress for introgression into adapted 
farmers preferred early maturing varieties. Furthermore, gene action controlling earliness and 
tolerance to drought stress must be determined or confirmed to provide direction for 
improvement.   
The overall research goal of the study was to identify promising cowpea genotypes that are 
early maturing and combining tolerance to drought with high yield to ensure food and 
nutrition security in South Sudan. 
The specific objectives were to:  
1. assess the genetic diversity among adapted farmer's preferred varieties and improved 
varieties from international agricultural centres for earliness and drought tolerance, 
2. identify sources of drought tolerance from selected accessions for introgression into 
adapted cowpea genotypes, 
3. introgress gene of drought tolerance into locally adapted early maturing varieties, 
4. determine the gene action controlling earliness and drought tolerance, and  
5. assess the performance of the backcross populations developed and their parents 
under drought-stressed and well-watered conditions.  
 
 
 
 
 
 
  
3 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
CHAPTER TWO 
2.0 LITERATURE REVIEW  
2.1 Origin, Domestication Distribution and Types of Cowpea 
Cowpea is a Dicotyledonea, belonging to the family Leguminosae, tribe Phaseoleae, genus 
Vigna, and section Catiang (Padulosi and Ng, 1997). The genus Vigna which comprises 
approximately of 85 species is divided into seven subgenera; Haydonia, Ceratotropis, 
Macrorhycha, Lasiocarpa, Sigmoidotropis, Vigna and Plectotropis (Marechal et al., 1978). 
Cowpea and its cross-compatible wild relatives belong to the subgenus of Vigna. 
The exact location or centre of cowpea origin is still difficult to determine; several 
assumptions are linked to it. Some studies tend to associate the centre of origin and 
domestication of cowpea to cytological evidence and botanical, cultural practices, 
geographical information (distribution) and archaeological records (Padulosi and Ng, 1997). 
Ng (1995) reported that the evolution process in V. unguiculata, change it from a perennial 
plant to an annual crop and from purely outcrossing to inbreed. Furthermore, Padulosi and 
Ng, (1997) stated that the process of evolution and domestication of cultivated cowpea (sub-
specie unguiculata) came about as a result of successive selection of undomesticated species. 
Through these processes, the species lost some of its traits such as seed dormancy and pod 
dehiscence and in return gained improvement in the number of seeds setting per pod. To 
establish cowpea centre of domestication, through analysis of chloroplast deoxyribonucleic 
acid (DNA) polymorphism, Nigeria is suggested as centre of cowpea domestication 
(Vaillancourt and Weeden, 1992). However, using amplified fragment length polymorphism 
(AFLP), the centre of domestication was proposed to be north-east Africa (Coulibaly et al., 
2002). Nevertheless, the global cowpea producing countries are in Africa, with Nigeria as a 
leading producer and consumer, producing over 3.03 million metric tons of grain on 3.61 
million hectares (ha), followed by Niger 1.99 million metric tons on 5.19 million ha 
4 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
(FAOSTAT, 2016). These two countries are followed by Burkina Faso, Mali, Togo, Senegal, 
Ghana, Benin, Chad and Cameroon in West Africa, and Central Africa, in that order. In the 
East and Southern Africa, the order is Somalia, Sudan, Kenya, Uganda, Tanzania, Zimbabwe, 
Botswana, Zambia, and Mozambique. However, countries in Far East including India, and 
China to mention few do produce cowpea. West Indies, Haiti and Cuba in Central America, 
Brazil in South America and the USA in North America also do produce cowpea.  Whereas 
in South Sudan, data on cowpea production level is very scared and if available it is 
inconsistent and not reliable (Ngalamu et al., 2015). 
Cowpea exhibits a number of growth habits, erect, semi-prostrate, prostrate and climbing 
type. However, preference for the growth habits varies from one country, region to the other. 
For instance, in Niger, the prostrate type is preferred by majority of farmers because it 
provides more leaves as feed for livestock. Whereas, in South Sudan, growth habit is not an 
issue since both the leaves and grains are consumed. Wang et al. (2006) reported that the 
growth habit in cowpea is associated with competitive ability. Thus, the exhibited growth 
habit by cowpea made the crop competitive to weeds. In addition, cowpea can be used in soil 
conservation and improvement of soil fertility. Roberts et al. (2005) observed that cowpea 
suppresses nematode and other pathogens. 
2.2 Climatic and Edaphic Factors Influencing Cowpea Production 
Cowpea thrives in broad types of soils and nutrient status and soil pH ranging from 5.5 to 8.3. 
Well-drained sandy loams and sandy soils were reported to be the best. It can tolerate salinity 
to some extent, but it tolerates soils high in aluminium (Ngalamu et al., 2014). It is observed 
to flourishes largely under humid conditions. Some genotypes were reported to be tolerant to 
heat and water deficit, however it is sensitive to frost (Davies et al., 1991). When soil 
temperature is above 1.9°C, cowpea germinates rapidly, but lower temperatures were 
observed to slow germination. This is evident when cowpea is grown under a well-watered 
5 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
and stressed condition or two irrigation regimes (Davies et al., 1991). Severe drought stress is 
of common occurrence in the Sahel zone of SSA. Comparing the performance of cowpea, 
pearl millet (Pennisetum glaucum) and peanut (Arachis hypogaea) grown in the zone, 
cowpea was observed to continue to stay longer and produce appreciable more yield than the 
other two crops when they are grown in the same fields under drought condition (Singh and 
Matsui, 2002). 
Delayed leaf senescence (DLS), an agronomic feature in cowpea, indicates a certain level of 
drought tolerance at the reproductive stage in erect cowpea genotypes. Flowering in cowpea 
starts at approximately 35 days after planting and producing a yield of 2.0 t/ha in two months. 
An additional 1.0 t/ha of cowpea grain could be obtained by genotypes that completes its 
lifecycle in 100 days from planting. These dynamics is ascribed to the ability of a cowpea 
plant to produce a second flush of pods (Hall et al., 2003). 
Photoperiod has no direct or little effect on cowpea leaf appearance (Ishiyaku et al., 2005). It 
has no effect either on branching and internode elongation. In contrast, it was found to have 
great effect on the reproductive development of genotypes, although some genotypes are 
unresponsive to photoperiodism (Ellis et al., 1994). However, cowpea responds to 
photoperiod in a quantitative way of short-day plants (SDP). Singh and Ntare (1985) reported 
that the longer the photoperiodic duration than the critical value, flowering is not affected. 
Thus, day neutral cowpea genotypes are preferred in breeding programs. In addition, several 
scientists (Craufurd et al., 1997, Ishiyaku et al., 2005) observed that long photoperiod during 
reproductive stage reduces flowering and pod production and eventually grain yield. Craufurd 
et al. (1997) reported that the neutrality of cowpea to photoperiod with respect to the onset of 
flowering could be markedly sensitive with pod production, although it is logical to assume 
that long days affect cowpea post flowering growth and development.    
6 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
2. 3 Constraints to Cowpea Production 
Despite being extensively grown in SSA, the crop is constrained by a number of insect pests 
such as pod suckers, aphids, thrips, and many others. Viral, bacterial and fungal diseases also 
devastate cowpea production. Other production constraints are drought, flood, and parasitic 
weeds including Striga gesnerioides (Wild) (Reiss and Bailey, 1998). In South Sudan, 
drought, lateness, low yield, lack of improved seeds and extension service were ranked as the 
major cowpea production constraints (Ngalamu et al., 2015). 
Drought stress significantly reduces number of pods per plant resulting in a reduction in the 
cowpea seed yield. Belko et al. (2013) found that an individual seed weight in cowpea was 
significantly higher under drought stress, but this does not compensate for the loss in the 
number of pods. Inherently, leaf area measured at the end of the drought stress cycle was 
strongly associated with seed yield (Belko et al., 2013).  
The detrimental effects of water deficit at any of these reproductive stages of flowering or 
pod filling in cowpea cannot be overcome by re-watering. Although reproductive activity in 
cowpea plants re-watered at the flowering stage resumes, most of the newly formed pods fail 
to reach maturity due to inadequate moisture supply resulting from excessive transpiration 
(Ahmed and Suliman, 2010). Timely planting is so important that the reproductive stages 
(flowering and pod filling) coincide with the period when water is most available (Ahmed 
and Suliman, 2010). 
According to Chiulele (2010), yield reduction in cowpea due to terminal drought stress is 
estimated to range from 21 to 30%. However, the yield reduction in cowpea production 
depends on both the location and duration of the cropping season (Sabaghnia et al., 2006). 
Most farmers’ varieties are susceptible to drought, hence dry spells in farmer’s field result in 
yield reduction. Drought can occur at anytime, anywhere and at any crop growth stage. 
Cowpea plants are most prone to damage due to drought stress during flowering and pod 
7 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
setting stages (Bahar and Yildirim, 2010). As such, it is necessary to enhance the degree of 
tolerance to drought stress in the adapted cowpea genotypes in order to obtain high and stable 
yields.  
2.4 Drought Stress and Cowpea Production 
Extensive agricultural losses in the world are attributed to agricultural drought stress (Bruce 
et al., 2002;  Tuberosa and Salvi, 2006). Agricultural drought stress normally results in a 
reduction in grain yield. It was observed to be the correlation between the influenced stage of 
crop development, severity, duration of drought stress and the level of drought susceptibility 
of genotype (Lorens et al., 1987). Effect of drought stress can be severe in field crops 
particularly at flowering phase where it causes a whole range of abnormality from abortion 
and lost pollen viability in cowpea, sterility in rice (Lafitte et al., 2004), whereas in maize, 
growth and embryo fertilization are reduced (Bruce et al., 2002; Earl and Davis, 2003). The 
occurrence of drought throughout the vegetative phase of the plant has a detrimental effect on 
yield as well. This reduction in yield is attributed to reduced leaf area which reduces the 
interception of photosynthetically active radiation (PAR), that in turn, truncates radiation use 
efficiency of the photosynthetic system and subsequent reduction in harvest index (Earl and 
Davis, 2003).  
Yield losses in cowpea are mainly due to the adverse effects of drought stress on the 
photosynthetic pathways and activity. Leaf senescence tends to disrupt photosynthetic 
apparatus as a result photosynthesis declines during grain filling. The incidence of drought, 
high temperatures, and excessive irradiance was observed to worsen the degeneration of 
photosynthetic apparatus (Martinez et al., 2003). Drought tolerant and relatively high 
yielding cowpea genotypes were observed to have enhanced adaptations mechanisms 
(morphological and physiological) that improve their photosynthesis stimulation, uptake of 
water and reduction of water loss under drought stress condition (de Souza et al., 1997). 
8 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
2.5 Mechanisms of Tolerance to Drought Stress 
2.5.1 Physiological Mechanisms  
 
According to Passioura (2002), the efforts to develop drought-tolerant genotypes in the past 
were obstructed by the polygenic nature of the trait as well as lack of in-depth knowledge of 
the physiological mechanisms influencing seed yield under water deficit condition. Breeders 
can focus on high yield under optimum and drought conditions to improve drought tolerance 
in high yielding genotypes (Danquah and Blay, 1999). As water deficit levels rise, the 
interaction between genotype and water deficit adversely affects plant traits that influences 
yield reduction. As a result, the pathway involved in the interaction could be explored in the 
development of a drought-tolerant population. This approach is variedly not being optimal, 
because of the polygenic nature of yield, high influence of genotype × environment 
interaction and low heritability (Babu et al., 2003). The understanding of both physiological 
and molecular bases may help the breeder identify and target core traits that limit yield under 
drought conditions. 
Cattivelli et al. (2008) concluded that having in-depth knowledge of the molecular and 
physiological basis of tolerance to drought stress in field crop of interest may aid the breeder 
to identify the core traits that influence yield. Molecular technology is a huge asset to the 
conventional breeding programs and will accelerate the pace of crop improvement. The 
advancement in molecular biology platforms enhances the ability to locate and sequence 
genes of interest. Efficient utilization of the molecular tool in the introgression of quantitative 
trait loci (QTL), selection or genetic transformation of hard to introgress desired QTL 
strongly depends on plant breeders' sound thoughtful insight on the yield determining 
physiological procedures (Araus, 2002). Numerous traits are involved in drought tolerance 
and can be measured at different physiological, biochemical, and morphological levels 
(Hamidou et al., 2007). A number of screening methods are used to assess water-use 
9 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
efficiency and drought tolerance. The most common are wooden boxes, pots, hydroponic and 
field screenings (Singh et al., 1999; Ogbonnaya et al., 2003). Regularly measured traits under 
drought-stressed conditions are the leaf area index, chlorophyll stability index, relative water 
content, diffusion pressure deficit, carbon isotope discrimination, and root/shoot ratio (Singh 
et al., 1999). 
Developing drought-tolerant genotypes could be through classical breeding approaches.  
Relying on water−use efficiency of the genotypes is the most common method used to 
determine early plant responsiveness to drought stress conditions. Genotypes exhibit 
considerable variations under drought stress which is a gene linked trait. Taiz and Zeiger 
(2002) observed that drought tolerant genotypes had reduced transpiration through restriction 
of stomatal opening or reduction of the leaf area or both. Costa et al. (1997) reported that 
genotypes that have reduced water loss and adjust their organ size or withstand the biomass 
production under water-limited conditions are susceptible to drought stress. Condon et al. 
(2004) proposed three key procedures to be considered in water−use efficiency hybridization 
in crops. These three processes are (i) enhancing level of available water uptake, (ii) boosting 
production of biomass per unit of water transpired, and (iii) dividing of biomass produced in 
the harvested product.  
2.5.2 Genetic Mechanism  
 
The whole plant crop genomes are involved in the population development when using 
conventional breeding and then identification and selection of outstanding recombinants from 
among many segregating populations. This traditional approach is cumbersome and time-
consuming, demanding a number of crosses, generations, and cautious field evaluation, and 
occurrence of tight linkage of unwanted traits with the wanted loci and makes the process 
hard for a breeder to accomplish a wished breeding outcome (Xu and Crouch, 2008). The 
advent of molecular biology resulted into development of advanced technologies, such as 
10 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
numerous molecular breeding strategies and types of molecular markers which offer 
opportunities for geneticists and breeders to overcome many of the difficulties encountered 
when utilizing conventional breeding approaches.  Khan et al. (2010) reported that field crop 
adaptation to drought stress is multiplex trait control by a number of gene or QTLs of small 
effects. Several researchers embrace the fact that breeding for drought tolerance necessitates 
amalgamation of both conventional and molecular approaches (Chaves et al., 2003; Blum, 
2011).  
The molecular procedure or genomics offer an exceptional opportunity for examining 
quantitative traits in their single genetic determinants (QTLs), consequently laid the 
foundation for genetic engineering (Salvi and Tuberosa, 2005) and marker-assisted selection 
(MAS) (Morgante and Salamini, 2003). The application of molecular markers in plant 
breeding for drought tolerance can be divided into three key categories: 
1) Fingerprinting (characterisation of germplasm); 
2) Recognition and characterization of genome regions (QTLs) associated with the 
expression of the desired trait; and 
3) Accelerated crop improvement via MAS approach.  
Advancement in molecular biology has cropped new scientific disciplines such as plant 
functional genomics. This discipline focuses on the study of genes functions. As a result, 
substantial progress has been made in area of sequencing plant genome including sequencing 
of an approximately 620 Mbp of cowpea genome. The other advent in plant breeding 
molecular approaches are the detection of genome-wide single nucleotide polymorphism 
(SNP) by means of Genotype by Sequencing (GBS), genetic diversity analysis of a 
worldwide germplasm using genome-wide association (GWAS), and SNP markers linked to 
morpho-agronomic traits identification in crops such as cowpea, maize, rice and barley ability 
to withstand biotic incidences, as well as tolerance to abiotic stress. When breeding for seed 
11 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
quality in cowpea, MAS and GWAS are tools a breeder can rely on for selection. Selection of 
some agronomic parameters such as plant growth habit, coloration of the dry pod, placement 
of pods, the pattern of mature seed, seed coat colour, the pattern of eye colour, flower colour, 
seed protein content, and sugar content can possible be obtained through GWAS. Muchero et 
al. (2013) reported that SNP markers are also used to identify some agronomic traits: grain 
yield, 100-seed weight and seed size; beside the biotic and abiotic stresses such 
resistance/tolerance to cowpea aphid, iron deficiency chlorosis (IDC), cowpea wilt (Fusarium 
oxysporum f. sp tracheiphilum), cowpea mosaic virus (CPMV), bacterial blight 
(Xanthomonas axonopodis pv. vignicola), and low phosphorus uptake efficiency are being 
studies for relations. 
 
2.6 Mechanisms of Drought Stress Avoidance  
The ability to withstand drought stress should reflect a balance amid the three mechanisms of 
escape; avoidance and tolerance at the same time supporting appreciable productivity 
(Agbicodo et al., 2009). 
2.6.1 Morphological Mechanism 
Cowpea drought tolerance morphological mechanism encompasses adjustment at tissue, 
molecular, physiological and whole plant system. Intrinsic single or combination of changes 
in cowpea determines its ability to withstand moisture stress conditions. The morphological 
adaptation mechanisms under drought conditions are as follows: 
2.6.1.1 Escape (drought avoidance) 
 
Escape may be defined as the preparedness of field crops to accomplish their lifespan ahead 
the occurrence of severe water deficit. In other words, drought avoiders, mature rapidly 
before the onset of drought (Agbicodo, 2009). Persistent drought incidences in cowpea 
producing areas have compelled farmers to plant early maturing varieties. Cowpea varieties 
12 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
that are early maturing tend to best escape terminal drought, albeit they are exposed to 
intermittent drought stress at vegetative stage and the yield is severely affected. According to 
Fatokun et al. (2012), intermittent drought stress reduces yield by up to 50 to 67%. This has 
serious implication on farmers’ ability to produce sufficient quality of cowpea grain for food 
security and household incomes. 
2.6.1.2 Avoidance (dehydration post-ponders) 
 
Dehydration avoidance is a measure of an extent to which plant water status is kept during 
drought stress. Drought tolerant field crops that have the ability to maintain leaf turgor 
through adjustment of osmotic pressure or increasing rate of water uptake and lessening the 
rate of water loss. The incidence of drought stress during the vegetative phase of crop 
development was observed to have little effect on grain production when successive 
conditions at the sites are favourable to encourage recovery (Fang and Xiong, 2015). 
Nevertheless, acute water deficit during floral initiation and flowering phase can result in 
almost comprehensive detachment of formed flowers and formation of immature pods, 
leading to 100% yield loss (Agbicodo et al., 2009). 
2.6.1.3 Tolerance (dehydration tolerators) 
Tuberosa (2012) defined drought tolerance as the preparedness of a crop to withstand drought 
stress and grow and yield satisfactorily. One of the important preconditions for gainful 
phenotyping of drought tolerance is the identification of key functional features that 
contribute to tolerance. Thus, understanding the metabolic and regulatory function of drought 
tolerance in crop plants is likely to provide valuable information that would help in 
suggesting approaches for modification through genetic manipulation.  
Earlier studies (Agbicodo et al., 2009) have shown that drought-tolerant cowpea lines are of 
two types, Type 1 and Type 2. Under drought stress, in 'Type l', plants stop growth but 
preserve moisture and keep all the leaves and growing tips alive for long period of time, 
13 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
whereas in 'Type 2', plants assemble moisture from the lower leaves to the growing tips. This 
process in Type 2 results into senescence and death of lower leaves, whereas, the tips of the 
leaves remain alive for an even longer period of time compared to ‘Type 1' plants. 
Knowledge about the inheritance of these two types of drought tolerance traits would 
facilitate their use in cowpea improvement programs.  
2.6.2 Physiological Mechanism  
Several physiological processes are involved in plant adaptation to dry environments (Fahad 
et al., 2017). However, cowpea exhibited little osmotic adjustment. Drought tolerance is a 
multiplex trait and its manifestation leans on exertion and reciprocation of different 
morphological characters such as early maturity, reduction of leaf area, rolling of leaves, 
production and deposition of wax in the epicuticular, effectual rooting system, stay−green, 
and yield stability. Physiological mechanisms such as reduced rate of transpiration, osmotic 
adjustment, high water−use efficiency, and stomatal closure were founded to be beneficial to 
field crops under drought stressed condition (Sinclair and Ludlow, 1986; Ludlow and 
Muchow, 1990). Not much was studied about the genetic mechanisms that condition these 
biochemical processes including proline accumulation, an increment in nitrate reductase 
activity and amplified storage of carbohydrate. Recognition of physiological traits that are 
positively and strongly correlated drought stress may be used in an index for selecting 
drought-tolerant genotypes. Plant physiologists have measured various plant characteristics, 
for example, stomatal orientation, osmotic pressure adjustment, water use efficiency, 
characteristics of the roots, broken off leaf water loss, and leaf water potential that correlate 
well with drought tolerance. Levitt (1972) grouped drought resistance into two categories 
based on physiological traits: dehydration avoidance and tolerance. According to Farooq et 
al. (2017), morpho-physiological traits such as osmotic adjustment (OA), deep roots 
architectural systems, early maturity, and wax deposition on epicuticular layer allow plants to 
14 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
sustain hydration status (avoidance). On the contrary, qualities such as accumulation of 
molecular protectants and remobilization of stem water soluble carbohydrates (WSC) permit 
plants to maintain proper functionality in severely dehydrated states (tolerance) (Afshari et 
al., 2013). 
 The most central establishment responsible for tolerance to drought is antioxidation, 
scavenging defense system, osmotic adjustment and osmoprotection. The mechanisms behind 
the genetic variation are not clear since it is only displayed as a response to the occurrence of 
drought. Nevertheless, it was assumed to be a complex physiological mechanism. High 
osmotic adjustment levels coupled with a low critical relative of water content of leaves are 
observed when drought stress mounts (Sinclair and Ludlow, 1986; Ludlow and Muchow, 
1990). However, –18 bar leaf water potential was observed in cowpea under extreme 
dehydration avoidance (Turk and Hall, 1980). The extreme dehydration avoidance of cowpea 
may be explained partially by those mechanisms. Cowpea was observed to be sensitive to soil 
water deficit, partially closes their stomata even before the differences in the leaf water 
potential are identified (Bates and Hall, 1981). Additionally, OA allows plant leaf to maintain 
turgor at a lower water potential by decreasing the rate of decline in leaf water content 
(Ludlow and Muchow, 1990).  
2.6.3 Molecular and Biochemical Mechanism 
Water stress due to water deficit is common in natural environments and occurs due to lack of 
rainfall. According to Nguyen and Blum (2004), drought stress results in damage of plant cell 
membranes (detrimental effect on the protoplasm). Removal of water from the protoplasm 
results in shrinkage and high concentration of solutes that may have serious structural and 
metabolic consequences in the plant. The integrity of the cell membrane and proteins are also 
affected (growth and development adversely affected). Late embryogenesis abundant (LEA) 
proteins, aquaporins (AQPs), and molecular chaperones play crucial roles in an osmotic 
15 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
adjustment in plant cells. During seed development, abundant proteins are formed. LEA 
proteins are low molecular weight proteins weighing between 10 and 30 kDa. They possess 
essential amino acids rich in lysine, glycine, serine and lacks cysteine and tyrosine (Nguyen 
and Blum, 2004). They are extremely high thermal stable proteins (hyper−hydrophilic) 
proteins and able to survive in an aqueous state even though under boiling conditions. LEA 
proteins are known to be playing protection role, when plants are grown under stress 
condition, through the biological macromolecules produced; redirecting distribution of 
intracellular water and it binds to inorganic ions to escape the tissue injury as result of 
accumulation and high concentrations of ions under drought stress conditions (Gosal et al., 
2009). LEA proteins do avoid extreme desiccation of plant tissues and regulate the expression 
of other genes by binding to nucleic acids (Gosal et al., 2009).  
2.7 Breeding for Tolerance to Drought Stress  
Management of drought stress effects can be achieved through the development of trait 
specific genotypes coupled with good agronomic practices (gap), such as planting date; plant 
stand and soil health (Rouf et al., 2012; Mariani and Ferrante, 2017). Breeding drought 
tolerant genotypes will help cowpea withstand drought at different stages (early, intermittent 
or terminal) of drought stress. Numerous breeding strategies have been developed to ensure 
sustainable cowpea production through genetic improvement of adapted farmers preferred 
high yielding genotypes. Institutions such as International Institute of Tropical Agriculture 
(IITA) and University of California Riverside (UCR) have extensively worked to develop 
genotypes with high level of water-use efficient cowpea that respond to water deficit. The 
best two core approaches cowpea breeders use to employ are: (1) selection of desired 
materials, as is the case of conventional breeding using molecular approaches, and (2) 
priming and hormonal applications inducing otherwise susceptible plants (Farooq et al., 
2009; Fathi and Tari, 2016). 
16 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
2.7.1 Conventional Breeding Efforts 
Through conventional breeding, genetic variability for drought tolerance among genotypes 
can be identified and those introduced can either be adopted of introgressed into the adapted 
local varieties to develop what is preferred by farmers. The conventional breeding approach 
has a number of limitations. This approach is time-consuming because a number of selections 
have to be carried out and the capital cost is huge (sequential evaluation). In addition, there is 
a high likelihood of introgressing a number of undesired genes into the agronomically 
desirable genotype.   
Cowpea breeders equally use physiological traits to aid in selection, designing methods for 
their assessments. However, a good number of the approaches failed because of the genotype 
by environment interaction effect (GEI) and also due to lack of precise and efficient drought 
screening methods. Sheshshayee et al. (2003) reported that selection for drought tolerance 
based on a physiological trait needs a comprehensive knowledge about the nature of the trait, 
its responsiveness to environment and contribution to seed yield. Earliness in cowpea 
production areas that are prone to drought stress is undoubtedly a core trait. Genotypes that 
mature in 50 to 60 days after planting are the most desired. In addition, traits such as the 
number of days to first flower, 50% flowering, first mature pod, and 95% maturity could be 
used in selection for earliness since it provides the required genetic information (Owusu et 
al., 2017).   
2.7.2 Molecular Approaches 
Marker-assisted breeding (MAB) is considered as one of the efficient approaches due to its 
ability to examine the entire genomic regions of a crop under water deficit condition. Thi 
Lang and Buu (2008) reported that tolerance to drought stress is governed by many minor 
genes (polygenes) that have an additive effect. The loci on a chromosome containing genes 
with a known QTL could be exploited either through direct selection under drought stressed 
17 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
condition (this can occur naturally or through simulation) or via QTL mapping and the results 
can be used for MAS (Ashraf et al., 2008). QTL mapping is feasible and allows fast-tracking 
of the trait, number of genes involved. Although conventional breeding is still being used but 
is challenging due to the complexity of tolerance to drought stress and difficulties associated 
with selection based phenotypic data. QTL mapping has proved to be essential for drought 
stress tolerance improvement. Some of the DNA used approaches are RFLPs, RAPDs, CAPS, 
SSRs, AFLPs, and SNPs. Its undoubted application was observed in improving crops such as 
maize, barley, cotton, rice and sorghum, with cowpea being intensively studied (Bernier et 
al., 2008). 
2.8 Significance of General and Specific Combing Ability in Crop Improvement  
 
Combining ability analysis is a useful biometrical tool for identifying outstanding combiners 
from the crosses, and aid logical selection of parental lines for crop improvement programs. 
However, the genetic background and performance of particular parental lines would 
certainly illustrate the line as a good or poor combiner. Hence, having thoughtful and logical 
statistical information about the nature of gene effects and how they are expressed in terms of 
combining ability is essential for crop improvement. Higher SCA values for a trait designates 
dominant gene effects, and higher values of GCA effects determines a greater role played by 
additive gene effects in governing the trait under study. General combining ability estimates 
the average contribution of an inbred to hybrid performance in a series of hybrid 
combinations and SCA is the contribution of an inbred to hybrid performance in a cross with 
a specific other inbred concerning its contribution in crosses with an array of other inbreds. If 
both the GCA and SCA values are not significant, then epistatic gene effects may play a vital 
role in determining these traits (Fehr, 1993). A significant GCA and SCA mean squares for 
traits studied at each site and across sites specifies the importance of both additive and non-
additive gene effects in the inheritance of the traits. 
18 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
2.8.1 Combing Ability for Early Maturity in Cowpea 
Xu et al. (2009) reported that early maturing cowpea varieties produce pods and mature in 55 
days and this pattern of maturity is essential in addressing the hunger gap period in SSA. 
Ojomo (1971) crossed an exotic early flowering genotype to adapted local late flowering 
genotypes and reported that flowering date is governed by two major gene pairs, early 
flowering being dominant over late flowering. Ishiyaku et al. (2005) reported that earliness is 
under polygenic control with additives × additive (I), additive × dominance (j), dominance × 
dominance interaction effects (epistatic effects). 
The genetics of earliness heritability is estimated using a simple statistical method that 
measures variance (phenotypic) amongst F2 population developed from the hybridization 
process between two sets of districted parental lines. Thus, the overall phenotypic variance 
among the F2 population is made up of both the genotypic and environmental variances (Xu 
et al., 2009). Furthermore, they reported that an estimated mean of the phenotypic variance 
among the parental lines could be used to compute environmental variance. The difference 
between phenotypic and environmental variances of the F2 individuals gives the genetic 
variance. However, the estimated heritability may be defined as the ratio of the genetic 
variance to the total phenotypic variance. This type of analysis has a narrow inference space 
because it depends entirely on the genetic differences between two particular parental inbred 
lines selected. Hence, this type of heritability cannot be generalized to other populations 
(Lynch and Walsh, 1998). The most commonly used computation method is the method that 
estimates broad−sense heritability with large inference space and does not require any 
hybridization procedure to estimate heritability. It can be carried out by analysing multiple 
genotypes for traits of preference using a simple analysis of variance (Singh and Chaudhary, 
1985).  
19 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
2.8.2 Combining Ability for Drought Tolerance in Cowpea 
Knowledge of genes controlling the trait of interest is essential for crop improvement 
program. Hinkossa et al. (2013) reported that having information about the effect and extent 
of gene action controlling the trait of interest must be well understood and determined. Traits 
such as drought tolerance and yield are polygenic in nature. Numerous genes control drought 
tolerance and affected by environmental factors which are not transmissible from parents to 
offspring. It’s therefore, absolutely necessary to determine the genetic factors conditioning 
these traits in order to establish an efficient breeding program.   
Singh and Chaudhary (1985) reported that there are three types of gene effects:  epistatic, 
additive and dominance. The dominance and epistatic effects constitute the non-additive part. 
The dominance gene effect can either be ambidirectional, a condition where numerous genes 
impact phenotype and dominance are in dissimilar directions reliant on the gene and it could 
either be unidirectional, dominance in one direction or positive and negative dominance at 
different gene loci (Kearsey and Pooni, 1996). Interaction of alleles at different loci is known 
as Epistasis. There are two concepts about epistatic gene action. Thus, it’s important for both 
plant breeders and bio-statisticians to distinguish between the concepts of statistical epistasis 
and physiological epistasis. Physiological epistasis is referred to as the effect of interactions 
between loci on the phenotype of an individual. This type of epistasis is a property of 
particular genotypes, and its values are independent of gene frequency. Moreover, statistical 
epistasis is the genetic variance within a population that can be attributed to interactions 
among loci. This type of epistasis is a population level phenomenon, and unlike previous 
epistasis, the latter changes as gene frequencies change. At some gene frequencies there may 
be physiological epistasis, but no statistical epistasis. The other scenario is, if there is 
statistical epistasis there must also be physiological epistasis (Goodnight, 1995). 
20 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
 The effects of additive genes are reflected by the extent to which offsprings resemble their 
parents, as reflected in narrow−sense heritability (Derera, 2005).  Estimation of relative 
proportion of additive genetic effects (general combining ability of a line) and non−additive 
genetic effects (specific combining ability of a population SCA) controlling the drought 
adaptive traits and their interactions with the environment is useful for designing breeding 
programs and assembling germplasm for population advancement (Shahi and Singh, 1985). 
Information about GCA and SCA of parental lines are useful and aids in the interpretation of 
the genetic basis underlying inheritance pattern of the trait of interest. Several researchers 
(Acquaah, 2012; Alidu et al., 2013) reported significant GCA effect for cowpea grain yield 
under drought-stress condition, suggesting that yield under drought-stress condition can be 
improved by exploiting additive gene effects. Ayo-Vaughan et al. (2013), in a separate study 
on combining ability for seed and pod in cowpea, reported that additive gene effects 
significantly controlled the number of pods per plant, pod length, number of seeds per pod 
and 100-seed weight. Thus, upon successful introgression of drought tolerance genes into the 
farmers preferred varieties, South Sudan will have its first early maturing, drought tolerant 
lines for a future release.  Being it a notable feature in self-pollinating crops such as cowpea, 
the skill for accurate identification of parental combinations required for generation of 
superior pure lines for farmers to adopt is very crucial to the success of the breeding 
programme. Efforts have been directed towards developing and evaluating methods of 
envisaging cross potential in early generations (Thurling and Ratinam, 1987).  Several 
scientists (Khan et al., 2010; Acquaah, 2012; Nduwumuremyi et al., 2013) reported that 
proper choice of good mating designs and selection of suitable parental lines are imperative 
to rewarding crop improvement schemes dependent on number of dynamics, for instance, 
objectives of the study, cost, time, space and other biological limitations.   
21 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
According to Singh and Chaudhary (1985), the mating design is a procedure for producing 
progenies for further selection. Theoretically and practically, plant breeders and geneticists 
use diverse types of mating designs and activities for targeted purposes. The postulate of 
mating designs facilitates: (1) provision of data on the gene controlling trait being studied; (2) 
populations development to be used as basis for selection and advancement progenies into 
prospective varieties; (3)   estimation of genetic gain; and (4) provision of data required for 
evaluation of the parental lines used in the improvement programme (Acquaah, 2012; 
Nduwumuremyi et al., 2013). Thus, combining ability is the capacity of an individual to 
transmit superior traits to its progenies. Hence it is a gene governing the trait of interest to be 
improved.  The estimation of general combining ability for a particular genotype be governed 
by the mating design, but essentially it is the deviance of the progeny average from the mean 
of the parental lines evaluated (Acquaah, 2012). There is limited information on the 
combining ability of lines that have been previously used in the cowpea breeding programs in 
IITA for the development of drought-tolerant lines. This has tremendously resulted in the 
slow phase of developing improved drought tolerant cowpea genotypes (Agbicodo et al., 
2009).  
 
2.9 Assessment of Genetic Potential of Cowpea for Breeding Program 
2.9.1 Genetic Diversity Studies in Cowpea 
Genetic diversity is the variation of inheritable characters existing among alleles of genes 
within individuals of species' populations and plays an important role in evolution by 
allowing adaptation in a new environment (Charles et al., 1997). It is a basic requirement for 
any successful crop improvement program. Assemblage, introduction, and evaluation of 
cowpea accessions are mandatory since it avails a grander scope for using the genetic 
multiplicity. A quantitative assessment of the genetic divergence among the germplasm and 
the comparative contribution of different traits towards the genetic divergence is essential and 
22 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
creates an effective genetic pool for a breeder in the hybridization program.  This genetic 
diversity could be assessed by three approaches: physiological, morphological, or 
phenological approach. These criteria could equally be used to select accessions with 
improved adaptation to dry environments (Blum, 1988). Conventionally, genetic diversity is 
estimated by measuring variation in phenotypic or qualitative traits such as the number of 
days to first flowering, time to maturity, plant type, flower colour, seed type, seed colour, 
seed size, hilum colour, and quantitative agronomic traits. However, this approach is often 
limited and expression of quantitative traits is subject to strong environmental responsiveness 
(Baye et al., 2011). Thus, having in-depth knowledge of phenotypic variation and 
relationships among genotypes will assist breeders to develop appropriate breeding strategies 
and to create the most adaptive and productive cultivars. The study on germplasm genetic 
variation exploiting, morphological, phonological and agronomic traits would be useful in the 
development of new varieties with better adaptation to both biotic and abiotic stress factors, 
as well as for high yield potential. The bigger the number of accessions collection, the higher 
the opportunity for selection of the most suitable parental lines for breeding objectives after a 
complex evaluation of samples, for adaptation capacity to broad or specific agro-ecological 
conditions (Hamidou et al., 2007; Agbicodo et al., 2009). It also provides the opportunity to 
identify and select accessions with good levels of tolerance to abiotic stress factors 
(Agbicodo et al., 2009). 
The inevitability for finding out genetic divergence among the genotypes are more evident 
because of two main reasons: i) genetically diverse parents if included in the hybridization 
program are likely to produce high heterotic effect; and ii) a broad spectrum of variability 
could be expected in the segregating generation of crosses involving distantly related parents.  
23 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
2.9.2 Importance of Genetic Distance in Hybridisation  
The genetic distance may originate from differences in the genetic complement or genomic 
difference of individuals. This difference is known as a genotypic or genetic variation. It 
could also be induced by exposure of the individuals to different values of treatment for 
particular environmental parameters during development (Aremu, 2011). This type of 
variation is termed environmental variation and it is not a heritable component of variation. 
However, genotypic variation is heritable and when acted upon by selection pressure 
evolutions do occur (Aremu, 2011). 
One of the major statistical tools used in estimating genetic distances is multivariate analysis. 
Multivariate analysis provides the possibility of gathering many variables into one analysis. 
Genetic distance based on phenotypic characters is one of the main multivariate techniques 
used to provide criteria for choosing parents. Genetic distance between genotypes is a way to 
predict the genetic variability among hybrid combinations (Cruz and Regazzi, 2001). In 
addition to genetic distance studies, it is worth noting that the genotypes selected for crosses 
should possess high individual performance, adaptability and stability features for yield. 
When these requirements are fulfilled, there is a high probability of selecting transgressive 
genotypes due to the occurrence of heterosis and the action of complementary dominant 
genes (Bertan et al., 2007). 
2.9.3 Genetic Information Derived from Allelic Pattern in Germplasm 
Allelic diversity is an estimate of the average number of alleles per locus. This is denoted by 
Na= Total number of alleles/Number of a studied locus. Allelic diversity depends on the 
population size. n is the total number of an allele in a population, whereas Na is the average 
number of alleles per locus. Thus, the genetic information provided encompasses inheritance, 
nature of inheritance of the trait of interest and frequency of alleles or gene of interest in the 
population. This implies that when the homozygosity is high, heterozygosity will be low.  In 
24 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
addition, the genetic relatedness of a population is normally measured by genetic distance 
(Jakobsson et al., 2008).  
2.9.4 Importance of Selection Index in Crop Improvement 
The estimation of genetic progress and selection of the best genotypes can be performed 
using both direct and indirect selection or classical index which is based on the sum of ranks. 
Direct and indirect selection is applicable if gains are from a single trait targeted in the 
selection, and other traits of secondary significance. These secondary traits may have 
favourable or unfavourable effects. The classical selection index was proposed by Smith 
(1936) and Hazel (1943) involves a linear combination of numerous parsimoniously essential 
traits whose weighting coefficients are estimated in order to exploit the correlation between 
the genotypic aggregate and the index. The genotypic aggregate is computed by another 
linear combination, involving the genetic values that are weighted by their respective 
economic weights (Cruz et al., 2012). 
Estimate of heritability along with the genetic coefficient of variation (GCV) is more useful 
in predicting the effect of selection than heritability values alone. The GCV determines the 
degree of genetic variability expressed by a plant for a trait in a population (Ahsan et al., 
2015). Selection is the retention of desired genotypes and the elimination of undesirable ones. 
It is an important process in breeding for improvement of one or more plant attributes. 
Drought susceptibility index (DSI), is another criterion for selection used as a degree of 
drought tolerance of the selected genotypes and was computed according to the procedure of 
Fischer and Maurer (1978) as:  
DSI = GR/DI,   
Where: GR = genotype response, calculated as GR = (1 – Ym1/Ym2),   
Where:  Ym1 and Ym2 are mean yields of all genotypes in severe drought stress and well-
watered conditions respectively.       
25 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
DI = drought intensities, computed as DI = (1 – Y1/Y2),   
Where:  Y1 and Y2 are respective yields of each selected hybrid under severe drought stress 
and well-watered environments.  
DI value ranges from 0 to 1, where lower values are considered as drought tolerant and high 
values as susceptible. 
The Johnson et al. (1955) method determined the genetic correlation and analysis of variance 
computed each of the indices. They classified stress intensity (SI) into mild, moderate and 
severe. Stress intensity is mild when yield reduction is between 0 and 25%, moderate when 
yield reduction falls between 25 and 50% and severe when yield reduction is between 50 and 
100%. 
 
 
 
 
 
 
 
 
 
 
 
 
  
26 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
CHAPTER THREE 
 
3.0 ASSESSMENT OF GENETIC DIVERSITY AMONG COWPEA ACCESSIONS 
FOR EARLINESS AND DROUGHT TOLERANCE 
 
3.1 Introduction 
Crop improvement requires genetic divergence, a degree of heritability and intense selection. 
However, the responsiveness of genotypes to selection warrants loss of variability as 
adaptability improves (Simmonds and Arthur, 2003). High genetic variability in a collection 
creates a high chance for the selection of superior cultivars as compared to low diversity 
populations. Charles et al. (1997), reported that the evolution and genetic improvement of 
species depend on the genetic diversity present in the germplasm assembled.   
Conventionally, agro-morphological traits such as growth habit, flower colour, yield potential 
and tolerance to stress were exploited in cowpea diversity study. However, morphological 
markers do not show the inherent genetic relationships in assembled genetic stock. This is 
attributed to the influence of environment in the expression of some of the genes. Thus, the 
knowledge about the structure of the assembled germplasm is limited (Wamalwa et al., 
2016). Additionally, good accession collection could be maximally utilised in population 
development using molecular markers. The markers may be used to fingerprint the collection 
and also act as indicators of uniformity (Simmonds and Arthur, 2003). Therefore, efficient 
utilisation of phylogenetic resources by plant breeder will help agricultural producers meet 
the global food requirement. These resources are abundant in many undesirable genes and are 
primarily useful as sources of a few desirable genes they possess (Vernooy et al., 2013). 
In Africa and particularly in sub-Saharan countries, the majority of the population live in 
rural areas and rely solely on crop production for subsistence. The socio-economic 
development in the African rural environment tends to be stagnant due to many agricultural 
production constraints (Brooks, 2013). Most rural populations in SSA are constantly faced 
27 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
with a reduction and erratic distribution of rainfall during cropping seasons and, as a result, 
their food and nutrition security status are weakened. According to Ouedraogo et al. (2017), 
the problem faced by rural subsistent farmers could be tackled through diversification of 
agriculture through proper utilization of all genetic resource and development of new 
efficient and durable production systems.  
In South Sudan the food shortage created by the “Man-made famine” of 2016 to July 2017 
prompted the breeding programs in the country, such as the cowpea breeding program in the 
University of Juba, to expedite the process of providing seed companies and farmers 
(community-based organisations) with improved and high yielding varieties in order to 
address the food crisis. Thus, genetic diversity research is a decisive base for further cowpea 
improvement in this defining era of climate change. Hence, it will enhance selection of 
parental lines having the targeted genes of interest; hence the breeding objectives can be 
attained in a short period of time (Collard and Mackill, 2008). 
Therefore, the broad objective of this study was to characterise the assembled cowpea 
germplasm for useful agro-morphological traits towards the development of high yielding, 
early and drought tolerant cowpea varieties.  
The specific objectives were to: 
a) assess genetic diversity among germplasm collection using agro-morphological traits 
and SNP markers; and 
b)  identify early and drought tolerant accessions to be used as parent for breeding work. 
 
 
 
  
28 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
3.2 Materials and Methods 
 
3.2.1 Experimental Materials 
 
One hundred and six cowpea accessions (pure lines) were collected in 2016 from six 
countries. Five of these countries are in West Africa and only one in East Africa (Fig. 3.1).  
Twenty-one of the cowpea accessions (21) were collected from the Equatoria region of South 
Sudan, (37) from Savanna Agricultural Research Institute (SARI), Plant Genetic Resources 
Research Institute (PGRRI) and West Africa Centre for Crop Improvement (WACCI) in 
Ghana, (10) from Institut de l’Environnement et de Recherches Agricoles (INERA) in 
Burkina Faso, (1) from Institut national de la recherche agronomique (INRAN) in Niger, (20) 
from Institut Togolaise de Recherche Agronomique (ITRA) in Togo, and (17) from IITA in 
Nigeria. The accessions from South Sudan were purified by planting row of each accession 
and rouging out off-types. Most of the accessions assembled for this study were not photo-
period sensitive except for Apagu 2A from South Sudan.  
3.2.2 Experimental Site   
 
The phenotypic evaluation was carried out at Crop Science Farm of the University of Ghana. 
One young trifoliate leaf from each plant of 106 cowpea accessions were harvested.  DNA 
extraction was carried out in the laboratory of the Biotechnology Centre in the College of 
Basic and Applied Sciences, University of Ghana. The extracted DNA were then shipped to 
TEXAS A &M AgriLife Genomics and Bioinformatics Services, College Station, TX, USA 
for genotyping where Novaseq 6000 sequencing system- Illumina was used. 
 
 
 
29 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
Table 3.1: Showing region, countries and number of cowpea genotypes assembled 
Region  Country  Number of Accession  
East Africa South Sudan 21 
West Africa Ghana 37 
 Nigeria 17 
 Niger 1 
 Togo 20 
 Burkina Faso 10 
 
 
3.2.3 Morphological Characterisation 
 
The 106-cowpea diversity panel from Burkina Faso, Ghana, IITA, Niger, South Sudan and 
Togo were evaluated in August 2016, using an augmented design with 103 of the accessions 
as unreplicated entries and three (Songotra, Padi-tuya and Dan lla) as replicated entries or 
checks. Each of the accessions was planted on a four-row plot with an area of 6.0 m2. The 
trial was planted in a field comprising of ten blocks and 13 plots per block. The crop was 
grown in the minor season under rain-fed conditions with supplemental irrigation, and no 
chemical fertilizer was applied. A botanical insecticide “Attack,” a non-systemic and highly 
active insecticide that controls a wide range of insect pests such as thrips, beetles, aphids was 
used. Its active ingredient is ‘emamectin benzoate’ which accounts only for 5 % of the 
ingredients was applied at 1.5 litres per hectare at all growth and development stages to 
reduce fall armyworm and other insect pest damage. In addition, Hercules 50 SC, a contact 
pesticide was used to control insect pest.   
Harvesting was done as soon as pods turned yellow colour completely and were sun-dried. 
The experiment was repeated at the same site (University of Ghana) experimental farm in 
September 2017, and the average of the two evaluations was used for analysis.  
 
 
30 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
3.2.4 Molecular Characterisation 
 
DNA was extracted from the leaf samples using the Hexadecyltrimethylammonium bromide 
(CTAB) protocol (Porebski et al., 1997). Leaf samples in the Eppendorf tubes were 
macerated into a fine paste in 500µL of CTAB buffer. The macerated samples were incubated 
in a water bath for 15 minutes at 55℃. After which the incubated samples were centrifuged at 
12000 rpm for 5 min. 200 µL of the supernatant was transferred from each tube to new 
Eppendorf tubes. 250 µL of Chloroform: Iso Amyl Alcohol (24:1) was added to each tube 
and inverted to obtain a milky solution. The solutions were centrifuged at 13000 rpm for 1 
min. After which, 200 µL of the supernatants were transferred into a new set of Eppendorf 
tubes. 50µL of 7.5 M ammonium acetate followed by 500 µL of ice-cold absolute ethanol 
were added to each tube. The tubes were inverted slowly to precipitate the DNA in the 
solution. The tubes were centrifuged again for 13000 rpm for 1min, and the supernatant in 
each of the tubes was poured entirely off so that the DNA pellets would remain at the bottom. 
The DNA was subsequently washed twice with 500 µL of ice-cold 70% ethanol by 
centrifuging at 13000 rpm for 1 min. The DNA in the tubes were allowed to dry for about 15 
minutes, after which they were dissolved in 100 µL of 1x Tris-Acetate 
ethylenediaminetetraacetic acid (TAE) buffer and stored in −80℃ freezer. Sample libraries 
was sequenced using the Illumina HiSeq 2500. Sequence cluster identification, quality 
prefiltering, base calling and uncertainty assessment were done in real time using Illumina's 
HCS 2.2.68 and RTA 1.18.66.3 software with default parameter settings. Sequencer. bcl 
basecall files were demultiplexed and formatted into fastq files using bcl2fastq 2.20.0 script 
conFig.ureBclToFastq.pl. 
 
 
 
31 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
3.2.5 Data Collection and Statistical Analysis 
 
3.2.5.1 Morphological data 
 
The following quantitative data were recorded. 
1. Plant height at 20 days after planting was measured using a meter ruler from the soil 
level to the base of meristem of the mother plant measured in centimetre (cm). 
2. Number of days to first flower was recorded as number of days from sowing to the 
day when first flower bud opened. 
3. Leaf area index (LAI) was computed as the ratio of foliage area to ground area (is a 
one-sided green leaf area per unit ground area in broadleaf canopies). 
4. Days to 50% flowering recorded as number of days from sowing to day when 50% of 
plants flowered. 
5. Days to first mature pods were measured as the number of days from sowing to day 
when 50% of plants had mature pods. 
6. Pod length was recorded in cm as mean of the 10 longest mature pods from 10 
randomly selected plants. 
7. Number of locules per pod was counted from the 10 pods measured for pod length. 
8. Days to 95% maturity was recorded from seedling emergence to harvest of each 
genotype per plot (duration). 
9. Number of main branches computed as the average of number of branches whose 
origin is in the leaf axils on the main stem counted from 10 randomly selected plants 
recorded in the 8th week after sowing.  
10. Number of pods per plant recorded as mean number of mature pods from 10 randomly 
selected plants.   
11. Stem diameter (mm) was measured starting from early vegetative stage to the first 
harvest using a digital calliper from 10 randomly selected plants.  
32 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
12. 100−seed weight (g) was determined by randomly counting 100 seeds from a bulked 
seed and weighed using a digital weighing−scale.   
13. Total seed weight (GYLD) was measured on plot bases after harvest and converted 
into kg per hectare (kg ha-1). 
14. Delay leaf senescence (DLS) as visual scoring on a scale of 1−5, where 1 = totally 
green and turgid and 5 = completely yellow to brown almost dead.  
15. Drought stress index as a visual score on a scale of 1−9, where 1 low susceptibility, 5 
medium susceptibility, and 9 highly susceptible. 
 
Table 3. 2:  Score of qualitative characteristics of cowpea accessions  
S/ No Major characteristics  When scored 
01. Growth habit Observed in the 6th week after planting 
02. Growth pattern  Observed in the 6th week after planting 
03. Plant pigmentation  Observed in the 6th week after planting 
04. Immature pod pigmentation Observed in the 6th week after planting 
05. Terminal leaflet shape Observed in the 6th week after planting 
06. Flower colour When 50% flowering is attained 
07. Raceme position  When peduncles have reached full length  
08. Pod attachment to peduncle When pods are full grown 
09. Pod curvature When pods are full mature 
10. Pod colour Mature pod 
11. Seed shape After harvest 
12. Eye pattern  After harvest 
13. Eye colour After harvest 
14. Testa texture After harvest 
 
 
3.2.5.2 Molecular data  
 
Raw SNP reads for quality was checked using FastQC (Andrews, 2010). Then the raw data 
were processed by the pipeline dDocent v2.2.6 (Puritz et al., 2014). The dDocent combines 
several developed tools into one single pipeline specifically tailored for GBS/RAD 
sequencing data for variation discovery. Generally, the raw sequencing data was first 
processed with a quality filter using the TrimGalore tool (Krueger, 2015) to remove Illumina 
sequencing adapter and trim the low-quality bases (Phred score < 10) on the end of reads. 
The quality filtered reads were mapped back to a de novo assembly reference constructed 
33 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
with dDocent. Only reads with >3X coverage and that are present in more than 10% of total 
samples were selected for de novo assembly. For the assemblies, BLAST (version 2.6.0) with 
e-value = 10^-5 was used to identify and remove assemblies that matched with chloroplast or 
ribosomal DNA in other plant species.  
Reads for each sample were mapped to the de novo assembly sequences using the 
BWA−MEM algorithm with the fairly conservative default parameters (Mikheenko et al., 
2018). Alignment files generated for each sample were then processed by the program 
FreeBayes (Garrison and Marth, 2012), a Bayesian-based variant detection approach, to 
detect SNPs from the aligned reads of all samples. 
Additional quality filtering was performed to minimize the calling of false SNPs due to 
sequencing error, paralogs, or artifacts of library preparation. The raw variant call file (VCF) 
was filtered using vcftools v0.1.15 (Danecek et al., 2011) to a minimum quality score of 30, 
with a minimum genotype depth set to 3 reads, no more than 20% missing data per SNP, and 
minimum mean depth of coverage (DP) of 20. Only bi-allelic SNP with a minimum minor 
allele frequency (MAF) of 0.05 or more were retained for downstream analysis. 
3.2.5.3 Statistical Analysis 
 
Statistical software GenStat 18th Edition was used to compute means of the traits which were 
then used for principal component analysis (PCA).  XLSTAT (2017) version was used to 
generate dendrogram and clusters using the phenotypic data. The SNP data was analysed 
using the Darwin Software Version 6 and STRUCTURE Version 2.3.3 to assess the genetic 
structure of the cowpea accessions, PCA and Neighbour-Joining (NJ) tree. SNP marker data 
were subjected to structural analysis using data from 3148 markers to genetically 
characterized 106 cowpea accessions. Bayesian clustering analysis was carried out in the 
STRUCTURE programme (Pritchard et al., 2000) to evaluate the population structure. Four 
34 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
independent runs were performed for each number of assumed subpopulations (K) ranging 
from 1 to 12, using the admixture model correlated for all frequencies. A burn-in period of 
5,000 and Monte Carlo Markov Chain (MCMC) run length of 10,000 for each run for all 
ranges (1 to 12) of K were carried out.  Evanno’s method (Earl and vonHoldt, 2012;  Evanno 
et al., 2005) was used to estimate the best number of subpopulations (K) that explains the 
structure of the genotypes. The run with the maximum log likelihood was used to assign 
genotypes into subpopulations based on their membership probability. The run was repeated 
for the best K at a burn-in period of 50, 000 and an MCMC run length of 100, 000 with one 
repeat. Genotypes with membership probability greater than or equal to 0.35 were put into 
same subpopulation whilst genotypes with membership probability less than 0.35 were put 
into mixed groups.  
 
3.3 Results 
 
3.3.1 Morphological  
  
The frequency distribution (percent) of variability in qualitative traits is presented in Table 
3.3 as observed on the accessions studied. About ninety-six percent (96.2%) of the accessions 
were determinate, and only 3.8% were indeterminate. Diverse growth habits were exhibited 
by the various cowpea lines in the collection. Erect types were the most observed (35.8 %); 
semi-erect (19.8%), intermediate (18.9%), semi-prostrate (9.4%), prostrate (4.7%) and 
climbing (11.3%) were all observed (Plate 3.1)  
 
 
 
 
35 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
.        
  A                                                     B                                                  C  
Plate 3. 1: Plant growth habits observed:   Apagu 1B, Laduni 1A and Titinwa A representing erect (A), climbing 
(B) and prostrate (C) types of cowpea accessions, respectively   
 
 
Variable degree of plant stem pigmentation (Plate 3.2) was observed in the panel with those 
plants moderately pigmented at the base and tips of petioles accounting for 50%, and 10.4% 
were not pigmented.  
Immature pods without anthocyanin pigmentation in the panel accounted for 49.1%, whereas 
those with solid pigmentation accounted for 12.3%.  Mature and dried pods pigmentation also 
varied in the collection. Dark tan accounted for 60.4% of the total (Table 3.3). Accessions 
with white, violet and mauve-pink petals accounted for 43.4%, 55.7%, and 23.6%, 
respectively. 
 
 
 
 
 
 
 
 
 
  
36 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
Table 3. 3: Frequency distribution of qualitative variables for cowpea accessions studied  
Descriptor and classes                                                                                       Frequency of class (%) 
 0 1 2 3 4 5 6 7 8 9 10 
Growth habit: 
1= Acute erect; 2= Erect; 3= Semi-erect; 4= 
Intermediate; 5= Semi–prostrate; 6=Prostrate; 
7=Climbing  - -  35.8 19.8 18.9 9.4 4.7 11.3  - -  -  
Growth pattern:  
1= Determinate and 2= Indeterminate  - 96.2 3.8  - -  -  -  -   -  -  - 
Plant stem pigmentation: 
 0=None; 1= Very slight; 3= Moderate at the base 
and tips of petioles; 5= Intermediate; 7= Extensive; 
9= Solid   10.4 21.7 -  50 -  8.5  - 6.6  - 2.8  - 
Immature pod pigmentation:  
0= None; 1= Pigmented tip; 2= Pigmented sutures; 
3= Pigmented valves, green sutures; 4= Splashes of 
pigment; 5= Uniformly pigmented; 6= Other 49.1 26.4 12.3  - -  12.3  -  -  -  -  - 
Leaf shape: 
1= Globose; 2= Sub–globose; 3= Sub–hastate; 4= 
Hastate   - 0.9 84.9 5.7 8.5  - -  -  -  -  -  
Flower colour: 
1= White; 2= Violet; 3= Mauve–pink; 4= Other 
(specify in the descriptor) -  43.4 55.7 23.6  -  -  -  -  -  -  - 
Raceme position:  
1= Mostly above canopy; 2= In upper canopy;  
3= Throughout canopy   - 46.2 44.3 9.4 -   -  -  - -  -  -  
Pod attachment to peduncle:  
3= Pendant; 5= 30 – 90° down from erect; 7= Erect  -  -  - 19.8  - 66  - 14.2  -  -  - 
Pod curvature:  
0= Straight; 3= Slightly curved; 5= Curved; 7= 
Coiled 6.6  - -  67.9  - 19.8 -  5.7  -  -  - 
Pod colour:  
1= Pale tan or straw; 2= Dark tan; 3= Dark brown; 
4= Black or dark purple; 5= Other  - 60.4 21.7 5.7 12.3  -  -  -  -  -  - 
Seed shape: 
1= Kidney; 2= Ovoid; 3= Crowder; 4=Globose; 5= 
Rhomboid  - 17 13.2  - 7.6 53.8  -  -  -  -  - 
Eye pattern:   
0= Absent; 1= Very Small; 2= Kabba group; 3= 
Narrow eye; 4= Small eye; 5= Holstein group; 6= 
Watson group; 7= Self-coloured; 8= Other   - 40.6  -  - 39.6 13.2  - 6.6  -  -  - 
Eye colour: 
0= Eye absent (white, cream); 1= Brown splash or 
grey; 2= Tan Brown; 3= Red; 4= Green; 5= Blue to 
black; 6= Blue to black spots or mottle; 7= 
Speckled; 8= Mottled; 9= Mottled and speckled; 
10= Other 89.6 6.6 2.8  - 0.9  -  -  -  -    - 
Testa texture: 
 1= Smooth; 3= Smooth to rough; 5= Rough; 7= 
Rough to wrinkled; 9= Wrinkled   - 53.8  - 26.4 -   -  - 19.8  -  -  - 
 
 
 
 
37 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
Pigmented                                                             Non−pigmented 
Plate 3. 2: Cowpea accessions Laduni 1B (left) and Apagu 1B (right) pigmented and non-pigmented pods 
respectively   
 
Accessions exhibited the three main raceme positions included mostly above the canopy, in 
the upper canopy and throughout canopy in the collection. These three observed raceme 
positions accounted for 46.2%, 44.3% and 9.4%, respectively (Table 3.3).  
Four distinct terminal leaflet shapes were observed among panel members with sub−globose 
accounting for 84.9% of the total observation and 8.5% hastate. Similarly, hastate and 
globose terminal leaflets were noticeable.  Still, they were less frequent than sub-globose and 
sub-hastate terminal leaflet orientation (Plate 3.3).    
 
 
38 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
                                                                                                                                      
A=Hastate                                                                                     B=Globose   
       
                                          
C=Sub−globose                                                                                      D=Sub−hastate 
Plate 3. 3: Variability in cowpea terminal leaflet shapes observed  
 
Genotypes exhibited three main types of pod attachments; pendant 30-90° and erect in the 
collection (Plate 3.4). The most observed pod attachment type was 30−90° for 66% of the 
population. 
The 106 cowpea accessions showed various eye colours with 89.6% of the accessions having 
white or cream eye colour, 6.6% with brown splash or grey, 2.8%, and 0.9% have tan brown 
and green eye colour, respectively (Table 3.3). Four seed shapes were observed with 
rhomboid accounting for 58.8% of the collection, 17% of the accessions had kidney shape, 
13.2% were ovoid, and 7.6% of the panel members had globose seed shape (Plate 3.5).  
 
 
39 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
       
A                                                     B                                                     C 
Plate 3. 4: Cowpea accessions Apagu 1B (A), AGRAC−316 (B) and Beledi C (C), showing erect, pendant and 
30°−90° types of pod attachment respectively    
 
Four noticeable pod curvatures were observed in the collection; 67.9% were slightly curved, 
19.8% curved, 6.6% with straight orientation and 5.7% were coiled (Plate 3.4). 
 
Plate 3. 5: A mixture of cowpea seeds from different accessions showing diversity of seed characteristics 
 
40 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
The assembled accessions displayed different eye patterns with 40.6% of the accession 
having very small eye pattern, 39.6% showed small eye pattern. Holstein group accounted for 
13.2% of the variability, and 6.6% were observed to be self-coloured. About fifty-four 
percent (53.8%) of the accessions had smooth testa texture, 26.4% were smooth to rough, and 
19.8% had rough to wrinkled testa texture (Table 3.3).  
Genetic diversity analysis based on agro-morphological data of the 106 accessions using 
cluster analysis revealed that the accessions are genetically diverse. This genetic divergence 
is shown by a dendrogram based on 13 phenotypic characters by unweighted pair-group 
average (UPGMA) method using the overall Euclidean distance. The Euclidean dissimilarity 
coefficient ranged from 0.0 to 1.4 with the cophenetic correlation coefficient (CCC) of 0.76. 
The resultant dendrogram grouped the 106 cowpea accessions into two main clusters, with 
nine sub-clusters at agro-morphological level (Fig. 3.1). The smallest sub-clusters 3, 6, 8 and 
9 had one accession each, cluster 5 had 2 accessions, and cluster 7 with 5 members (Table 
3.4). Sub-cluster 1, 2 and 3 were the largest with 35 members in cluster 1, 39 accessions in 
cluster 2 and 21 members in cluster 3. Majority of cowpea accessions from South Sudan were 
members of sub-cluster 1 with only two accessions in sub-cluster 2. The mean performance 
values of measured agro-morphological traits in 106 cowpea accessions in the clusters 
revealed that members in cluster 1 registered the highest grain yield of 503.34 kg/ha and 
cluster 8 accessions had the lowest grain yield (101.00 kg/ha). 
 
 
 
 
41 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
Dendrogram
1.4
1.2
1
0.8
0.6
0.4
0.2
0
 
Fig. 3. 1  Dendrogram resulting from the analysis of 106 cowpea accessions (based on 13 phenotypic characters) obtained by UPGMA using the overall distance of 
Euclidean. The cophenetic coefficient (r) was 0.76 
 
42 
 
Dissimilarity
BLK452
GH2338
Rec064
GH2200
IT96D604
Padi_tuya
WAC91
Gorom
Kvx40481
WAC21
IT10K8177
Kvx74511
Rec083
IT98K10921
Rec003
GH3701B
Rec108
WAC19
Rec005
Danila
Mouride
IT08K125100
IT98K2058
Rec046
WAC Q6
GH4524
IT97K56818
WAC115
Rec062
Apagu2A
Apagu1B
Laduni1A
Laduni1B
TitinwaA
BelediA
TitinwaB
AGRAC116
GH5038
Rec014
GH7875
WAC39
TitinwaC
BelediC
AGRAC316
Rec017
AGRAC216
Laduni2B
Rec007
IT98D1399
GH7220
GH7245
GH6045
Rec074
Apagu1C
GH5346
GH5346B
Rec016
BelediB
IT08K1493
Laduni2A
WAC101
GH2347
Rec059
Apagu1A
IT96D610
IT08K15024
AmosV
IT93K4521
IT93K5031
Rec039
Rec021
WAC32
IT08K15012
Rec105
WAC 81
IT08K18011
Rec049
Rec041
GH3689
GH3701A
Rec009
MangalaA
IT99K57321
KNI
IT98K5061
IT99K57211
GH7228
IT97K49935
GH2306
GH2307
GH3668
IT08K18724
IT98K5031
Songatra
Blackeye
Kvx303096G
BLK19
IT99K1122
Local 3
IT97K56818
B
GH5043
KVX3964525
Pobe
GH4527
MangalaB
University of Ghana http://ugspace.ug.edu.gh
 
 
 
 
Table 3. 4: Mean values of agro-morphological traits of 106 cowpea accessions 
 Yield 100−seed Plant Stem Leaf Number Days to Days to Number of Number Number Number Pod 
Number of clusters (kg/ha)  weight (g) height diameter area of first 50% days first of pods of of days to length 
(cm) (mm) index branches flower flowering pod per locules 95% (cm) 
per plant plant per pod maturity 
Cluster 1 (35) 503.3 13.8 12.5 12.3 0.2 5.6 36.0 38.0 42.5 42.6 15.4 61.0 15.7 
Cluster 2 (39) 271.3 14.4 10.7 11.0 0.1 4.9 37.0 40.0 44.7 33.0 13.2 64.0 15.0 
Cluster 3 (1) 277.0 14.4 49.0 10.3 0.2 4.8 43.0 49.0 54.0 22.0 15.5 66.0 17.7 
Cluster 4 (21) 190.9 13.4 9.8 10.4 0.1 4.5 38.0 41.0 45.5 30.7 13.7 64.0 14.9 
Cluster 5 (2) 158.0 17.4 9.9 11.1 0.2 4.3 38.0 43.0 47.0 16.9 11.7 63.0 12.7 
Cluster 6 (1) 119.0 11.2 10.1 9.6 0.1 4.4 33.0 39.0 40.0 28.9 11.7 65.0 10.9 
Cluster 7 (5) 150.0 13.7 11.2 11.2 0.1 5.2 40.0 43.0 47.3 35.7 13.2 65.0 14.7 
Cluster 8 (1) 101.0 10.8 13.7 7.8 0.1 4.0 33.0 37.0 40.0 36.9 14.1 63.0 13.8 
Cluster 9 (1) 146.0 13.9 23.8 9.2 0.3 11.4 42.0 45.0 49.0 14.7 14.8 76.0 15.7 
Footnote: Number of entries per cluster are shown in brackets  
 
43 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
 
The Eigenvalues and percentage of explained variability by principal component showed that 
the first three principal components accounted for 56% of the total variation (Fig. 3.2 {a}). 
Vectors for variability were scattered on the biplot. The most discriminative traits included 
stem diameter, number of days to first flower, number of days to 50% flowering, number of 
days to first pod, pod length, leaf area index, number of locules per pod, number of pods per 
plant, number of branches per plant, number of days to 95% maturity and yield (Table 3.5).  
Informative vectors for evaluation of cowpea accessions included vector length which was the 
longest relative to other vectors. Plant height at 20 days after planting and 100−seed weights 
were the least discriminating vectors as their vector length was the shortest relative to other 
vectors (Fig. 3.2 {b}). Vectors for variability were scattered on the biplot. The most 
discriminative traits in terms of contribution to variation among 106 cowpea accessions 
included stem diameter, number of days to first flower, number of days to 50% flowering, 
number of days to first pod, pod length, leaf area index, number of locules per pod, number of 
pods per plant, number of branches per plant, number of days to 95% maturity and yield (Table 
3.5 and Fig. 3.2 {a}). 
 
 
 
 
 
  
 
44 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
 
Table 3. 5: Principal components loading for 106 cowpea accessions variables 
Variables  Mean StDev PC1 PC2 PC3 PC4 PC5 
Yield  297.51 158.14 -0.35 -0.18 0.11 -0.13 0.05 
100-seed  14.00 3.30 0.03 -0.11 0.44 -0.45 -0.09 
Plant height  11.23 4.18 -0.08 -0.27 0.13 -0.05 -0.61 
Stem diameter  11.10 1.79 -0.17 -0.40 -0.06 -0.14 0.45 
LAI  0.145 0.05 -0.14 -0.41 -0.07 -0.00 -0.12 
Branch  5.08 1.07 -0.15 -0.27 0.17 -0.24 -0.32 
Flower1  37.28 2.67 0.40 -0.29 -0.11 -0.08 0.10 
Flower50  40.17 2.80 0.40 -0.29 -0.08 -0.07 0.02 
Firstpods  44.94 3.35 0.41 -0.25 -0.04 -0.11 0.12 
NPP  35.12 11.44 -0.28 -0.14 0.11 -0.22 0.46 
Locules  13.55 2.70 -0.26 -0.24 -0.25 0.45 -0.06 
Mat95  63.45 4.39 0.30 -0.21 -0.22 0.09 -0.11 
Podlenth  14.74 2.26 -0.21 -0.33 -0.12 0.31 0.02 
DSI  3.62 2.76 -0.11 0.12 -0.52 -0.46 -0.02 
DLS  2.32 0.49 -0.15 0.07 -0.56 -0.34 -0.19 
StDev    2.01 1.62 1.36 1.14 1.05 
Variance    4.02 2.61 1.84 1.29 1.10 
Prop. Var Exp   0.27 0.17 0.12 0.09 0.07 
Cumulative   0.27 0.44 0.56 0.65 0.72 
Yield (grain yield), 100-seed (100-seed weight), Plant height (plant height at 21 days after planting), Stem 
diameter, Leaf area index, Branch (number of branches per plant), Flower1 (number of days to first flower), 
Flower50 (number of days to 50% flowering), Firstpods (number of days to first mature pod), NPP (number of 
pods per plant), Locules (number of locules per pod), Mat95 (number of days to 95% maturity), Podlenth (Pod 
length), DSI (Drought stress index) and DLS (delayed leaf senescence) 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
45 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
 
Scree plot Factor loadings (axes PC1 and PC2: 45.33 %)
4 100 0.8
STEM_DIAMETER
PODLENTH LAI
3.5
0.6
LOCULES
80
YIELD BRANCH
3 PODN
0.4 PLANT_HEIGHT
FLOWER1
FRFSLTOPWOD50S
2.5
60
0.2 SDW100
MAT95
2
0
40
1.5
-0.2
1
20
-0.4
0.5
0 0 -0.6
F1 F2 F3 F4 F5 F6 F7 F8 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
axis PC1 (27.08 %)
  
 
 
 
(a) Scree plot graph of factors responsible for variability among study accessions  (b) Projection of the 13 characteristics in axe 1 and 2 
 
Fig. 3. 2 Eigenvalue and principal components responsible for variability among 106 cowpea accessions  
 
 
46 
 
Eigenvalue
Cumulative variability (%)
PC2 (18.25 %)
University of Ghana http://ugspace.ug.edu.gh
 
 
3.3.2 Molecular Genetic Structure 
 
The generated genetic structure from the cluster analysis of the SNPs markers gave 
information on the germplasm studied (Fig. 3.3). The cluster revealed the existence of a 
considerable degree of genetic diversity and reflecting a reasonable amount of genetic 
dissimilarity among the genotypes. However, there were six distinct clusters with the 20 
cowpea accessions from South Sudan were genetically associated with all the members of the 
six clusters with only AGRAC−216 being admixtureThe output of the analysis designated by 
each of the six colours represented a population/cluster; red is cluster 1 predominantly 
accessions from Burkina Faso, green is cluster 2 (Nigeria), blue is cluster 3 (Ghana), yellow 
is cluster 4 (South Sudan), pink is cluster 5 (Niger), and blue as cluster 6 (Togo). Cowpea 
accessions from South Sudan revealed relatedness to all members of the six clusters (Fig. 
3.3). 
Fig. 3. 3 Display of population structure of the 106 cowpea genotypes 
 
The highest values of expected heterozygosity approximately 0.3 were observed in cluster 5 
and the lowest was in cluster 4 0.1 (Fig. 3.4).  An estimated Fst value of 0.75 was obtained in 
cluster 1 and 0.28 in cluster 1 (Fig. 3.4). 
 
47 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
 
Fig. 3. 4 Expected heterozygosity and mean fixation index (Fst) for each cluster 
 
The expected heterozygosity among members of a cluster revealed that cluster 5 had the 
highest value of 0.31 with cluster 4 having the lowest (0.12). The inferred inferences used in 
the population structure of the 106 cowpea accessions revealed a peak delta K=6.  Whereas, 
the mean value of the fixation index of the 106 cowpea accessions members of six clusters 
ranged from 0.29 to 0.73 (Table 3.6). 
Table 3. 6: Mean Fst value and expected heterozygosity within cluster 
Cluster Number of Members Expected heterozygosity Mean value of Fst 
Cluster 1 30 0.25 0.29 
Cluster 2 21 0.17 0.54 
Cluster 3 18 0.2 0.47 
Cluster 4 7 0.12 0.73 
Cluster 5 13 0.31 0.33 
Cluster 6 17 0.2 0.47 
  
The analysis of allelic divergence among the clusters revealed that the longest net nucleotide 
distance existed between members of cluster 4 and cluster 5 (0.15). However, the shortest 
divergence (0.08) was observed between cluster 1 and cluster 6 (Table 3.7). 
 
48 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
Table 3.7: Allele-frequency divergence among clusters (Net nucleotide distance) 
Clusters 1 2 3 4 5 6 
1 -          
2 0.09 -    
3 0.09 0.13 -   
4 0.12 0.14 0.12 -  
5 0.09 0.10 0.13 0.15 - 
6 0.08 0.10 0.10 0.12 0.10 - 
 
The hierarchical cluster analysis of the 106 cowpea accessions revealed node numbering 
begun at 107. The weighted Neighbour-Joining was obtained from tree construction with a 
bootstrap analysis of 1000. The average edge distance between bootrapped trees was 0.40 
with 0.29 as 5-percentile and 0.50 as 95-percentile. The dissimilarity index used was the 
Euclidean procedure and the values on the node of the phylogenetic tree are probabilities 
indicating relatedness of the accessions (Fig..3.5).  However, there are six distinct clusters 
with the majority (38) of the cowpea accessions in cluster one, 20 cowpea accessions in 
cluster two, 16 in cluster three. Similarly, cluster four had six members, cluster five had 10 
accessions, and six had six members. 
49 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
99 Rec041
56 Rec105
80 Gorom
96 GH7220
74 GH2307
IT08K18724
W AC101
37 46
36 IT99K1122
Rec021
19 48
33 W AC3226 Songotra
Mouride
AGRAC216
100 IT10K8177
76 Laduni 1A
63 Titinwa C22
100 IT93K503192 IT96D610
BLK19
18 100 Apagu1C
55 GH5346 B
6 Apagu 2B
42 38 100 Rec039
72 GH3668
7 GH2306
Rec064
W AC Q6
1
25 GH5038
AGRAC116
2
23 KVX3964525
GH2338
W AC19
74 Kvx74511
100 GH5043
60 IT08K15012
51 100
Rec016
Apagu 2A
26 Pobe
GH7245
44 95 Rec074
93 Rec007
100 Laduni 2A
99 Rec005
30 KNI
100 Laduni 2B
90 Mangala Mangala B
92 Rec017
52
98 IT96D604
94 IT08K125100
IT97K56818
Beledi B
97 Rec059
100 Beledi C
96 GH5346
64 Mangala Mangala A
1 Rec108
68 100 AmosV
100 W AC115
40 89 Titinwa A
GH6045
Laduni 1B
74 GH4527
29 Local 3
3 14 Apagu1B
55 IT98K2058
IT98K5031
17 100 Kvx40481
55 IT98K10921
100 IT98D139919 100 IT08K15024
0 W AC39
GH2200
35 Rec062
58 IT99K57211
19 W AC 81
Rec083
10 100 IT97K56818
5 50
BLK452
IT08K1493
53 15 GH7875
IT98K5061
Rec046
8
54 GH3689
0 28
Blackeye
30 Dan lla
Apagu1A
2
100 Beledi A
99 Padi_tuya
83 GH3701 A
Titinwa B
60
100 W AC21
71 IT97K49935
GH4524
100 Rec014
99 IT08K18011
100 W AC91
GH2347
27 100 AGRAC316
IT99K57321
100 Rec009
94 Rec049
73 Kvx303096G
46 65
GH3701 B
GH7228
100 Rec003
IT93K4521
  
Fig. 3. 5 Phylogenetic tree depicting genetic similarities and dissimilarity among 106 cowpea accessions 
 
 
50 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
The principal coordinates are calculated for the first five axes (with positive eigenvalues). 
Dissimilarity in the collection was calculated from the data file using Cluster Darwin V6 
(Perrier et al., 20003). Drought tolerant accessions grouped in one cluster; susceptible 
genotypes were assigned in another. Equally, high yielding accessions were in different 
clusters so as the late maturing group. Colour code: Togo (Black), Ghana (Blue), Niger 
(Brown), Nigeria (Green), Burkina Faso (Red) and South Sudan (Yellow) (Fig.3.6). 
 
 
 
 
 
 
 
 
51 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
Factorial analysis: (Axes 1 / 2)
30
Kvx74511
25
GH5043
IT08K15012
Rec016
20
ApRaegcu00 25A WAC101
Mangala Mangala B Rec017
Laduni 2BKNI Beledi B15
Laduni 2ARec007Mangala Mangala A IT08K18724
Rec074 Rec041
GH7245 IT96D604Titinwa B
IT98K10921 Pobe 10 GH7220
ITIT9088DK113590924 Songotra Rec105
Kvx4A0p4a8g1u1B Local 3 IT99K1122
GH3701 A
ReGcH1503846
IT97K5IT68918G8KH54053214
Rec059 5 GH2338 Mouride
Beledi C WAC19
GH36A89GITR9A7KC43919635
Laduni 1B Gorom
H4527 GIHT20200GH2347 8K125100
GH2307
IT98KKv2Wx03A5W08C3RA03eC99c620G10 WAC115KVX3964525
-15Dan lla-1G0H3701 B-5 Amo5sV 10 15 20 25 30 35 40 45
Apagu1
IT9R3eKRc4e0
A
5c4209109 G IT98K5061 AGRAC216
Beledi A TitiGnHw6a
H0 74A2528 IT99K57321
Padi_Btulaycakeye WAC Q6
-5
WAC91
Rec046 GH7875
IT08KR1e80c10114 GH5038
BLK452
IT99K57211 -10
Rec062 WAC 81 AGRAC116
Rec083
IT97K56818
-15 WAC32 Rec021
IT08K1493
GH2306 Rec064
-20
IT1L0aKd8u1n7i7 1A Apagu 2BIT93K5031
IT96BDL61K019
ApGaHg5u314C6 B-25
TitinwRae cC039
GH3668
-30
 
Fig. 3. 6 Genetic variation of 106 cowpea accessions analysed by SNP marker at K=6 
 
 
 
 
 
 
 
 
  
52 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
3.4 Discussion 
 
The use of plant diagnostic traits to identify varieties has been a classical taxonomic 
approach for both varietal purity testing and identification. Plant morphological traits such as 
plant stem pigmentation, stem diameter, plant growth pattern, plant growth habit, plant height 
and the number of branches per plant were scored, measured, and analysed which facilitated 
classification of accessions into different groups.  
Growth habits of the 106 accessions varied greatly as genotypes with the most being acute 
erect (35.8 %) and the least prostrate (4.7%). The majority of Ghanaian accessions had semi-
prostrate growth habit, some had intermediate or prostrate growth habit, and only 1% was 
found to be erect. This result agrees with the finding of Doumbia et al. (2014) reported that 
the germplasm from Mali had more than 80% semi-prostrate growth habit. However, 
according to Wang et al. (2006), semi-prostrate growth habit has three benefits, competitive 
ability with weeds, ease of harvesting and high biomass (forage for livestock). Growth habit 
in cowpea is controlled by a single gene (Lachyan and Dalvi, 2013) and is easily selected for.   
The growth patterns of the genotypes were variable thus they were grouped into a 
determinate type which accounted for 96.2% of the total collection and indeterminate type 
constitute 3.8% of the total collection. Nkouannessi (2005) and Sarutayophat et al. (2007) 
obtained groupings of higher percentage of determinate growth type in cowpea. Similar 
finding was reported in green gram by Jain and Khare (2002). However, they found that 
determinate and indeterminate expressions of plant growth pattern were stable, but semi-
determinate expression was highly influenced by environment.  
Plant stem pigmentation is another qualitative trait that breeders explore in order to quantify 
the level of genetic divergence. The 106 cowpea accessions displayed different levels of 
pigmentation which varied with the stocks, the base and tips of petioles, and stem. This 
finding agrees with the results obtained by Cobbinah et al. (2011) who reported a similar 
53 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
trend of variability for the trait in some cowpea genotypes evaluated in Ghana. Karkannavar 
et al. (1991) proposed a digenic inheritance mode of stem pigmentation with complementary 
gene action and a 17.4% crossover rate was recorded between a complementary gene for 
stem pigmentation and one for stipule colour. 
Leaf shape is a prominent qualitative trait that breeders use as a morphological marker in 
diversity studies. About eighty-five percent (84.9%) of the assembled accessions had sub-
globose leaf shape, 0.9% were globose. Sub-hastate leaf shape accounted for 5.7% and 
hastate for 8.5% of the total variability present in the collection. Doumbia et al. (2014) 
reported similar patterns of divergence in cowpea accessions from Mali with more than 50% 
having sub-globose leaf shape, but other accessions from Ghana were different. The 
Ghanaian collection was predominantly hastate and sub-hastate.  Uguru (1996) found that the 
rhomboid and lanceolate leaf shapes were genetically stable. But Pottorff et al. (2012) 
reported that hastate leaf shape is dominant to the more common sub-globose leaf shape. 
They further reported that the hastate leaf shape was incompletely dominant to (ovate) 
globose leaf shape.   
Plant height is an important characteristic which aids breeders in differentiating the 
genotypes. Based on this trait, genotypes were grouped into: dwarf, medium and tall. The 
106 genotypes exhibited plant height ranging from 9.8 to 49.00 cm. A large percentage 
(66.04%) of the genotypes had medium height. Although, plant height was influenced by 
high additive genetic variance (Patil and Navale, 2006), it may also by influenced by 
environment. 
Based on the number of days to 50 percent flowering which ranged from 37 days (GH 4527) 
to 49 days (Apagu 2A), three classes of maturity groups were identified. Among the 106 
genotypes, 76 accessions which were members of cluster 1, 2, 6 and 8 matured early (<41 
DAP), 29 members of cluster 4, 5 7 and 9 were medium (41−49 DAP), and the only member 
54 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
of cluster 3 was late maturing (>49 DAP). This finding corroborates with the result of Aremu 
et al. (2007) who found similar trend for the trait with cowpea accessions collected from the 
West Africa sub-region. The variability observed in the number of days to 50% flowering 
could be attributed to a minor complex gene (Weiss, 1971; Adu-Dapaah et al., 1988) with a 
tendency for dominance of early flowering. Ikram (2004) reported that earliness in cowpea 
was under the influence of additive gene action more than dominance effects. In addition, 
environmental conditions were observed to affect flowering. The number of days to 95% 
maturity varied from 61 to 76 days. Thirty-five accessions (cluster 1 members) were early 
(<60 days), one accession was late (>65 days) and remaining accessions were medium 
(60−65 days). Nevertheless, Patil and Navale (2006) found that maturity in cowpea is 
controlled by dominant gene action. In addition, environmental conditions such as soil 
moisture level, temperature, and relative humidity influence the growth and development of 
crop plants.  
Based on variation observed in the flower colour, the genotypes were grouped into 
accessions with white flowers accounting for 43.5%, whereas 55.7% of accessions had violet 
flowers and 23.6% were mauve-pink. The variations in flower colour can be attributed to the 
genetic factors. Genes determine the colour of the petal by producing or blocking 
anthocyanin pigmentation. Othman et al. (2006) reported that flower colour was simply 
inherited when different types of colour were considered separately, and two genes were 
observed to control flower colour. Patel et al. (2016) found that a single recessive gene-
controlled flower colour in cowpea. 
Raceme position and pod attachment to peduncle in cowpea genotypes differed significantly. 
However, 46.2 % of the accessions had their racemes positioned above the canopy, 44.3% 
had their raceme in the upper canopy, and 9.4% were observed to have their raceme 
positioned throughout the canopy. Higher raceme position enables easy discernibility of pods 
55 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
during harvest. With respect to pod attachment, 19.8% of the accessions had pendent pod 
attachment, 66% showed 30º−90º down from erect and remaining 14.2% had erect pod 
attachment. The expression of these traits is genetically controlled and heritable (Jain and 
Khare, 2002). Pod characteristics influence the yielding ability of cowpea.  
Number of branches per plant determines the pod bearing ability which contributes to yield. 
Hence identification and selection of genotypes with more branching ability is desirable. In 
this study, genotypes exhibited varying degree of number of branching ranging from 4.0 (GH 
4527) to 11.4 (Mangala Mangala B). Based on this trait, the 106 cowpea accessions were 
grouped into fewer (<4), medium (4-6) and high (>6) branches per plant. Accession GH 4527 
had the least number of branches per plant, while one hundred and four accessions had 
medium number of branches per plant, and Mangala Mangala B had the highest number of 
branches per plant. Patil and Navale (2006) reported that number of branches per plant is 
controlled by dominant gene action indicating the presence of intra-allelic interaction in the 
trait. Thus, heterosis could be the best breeding approach for improving the number of 
branches per plant. Other breeding approaches such as isolation of recombinants in advance 
segregating population could also be useful. The distinction in the number of branches per 
plant is influenced by genetic factors and environmental conditions, cropping seasons, seed 
rate and spacing (Weiss, 1971).  
Number of pods per plant were variable among the genotypes, with Laduni 1B having the 
highest number of pods (63). Based on the variability of this character, the genotypes were 
grouped as less (<15), moderate (25.7), or high (>42) pod bearing types. Patil and Navale 
(2006) found that number of pods per plant was governed by additive gene action. In 
addition, the distinction in the number of pods per plant could also be attributed to pod 
bearing ability of the genotype, environmental responsiveness and soil nutritional status.  
56 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
Equally, pod length varied among the accessions panel, and classified the accessions into 
those with short pod length (<11 cm), medium (11−15 cm), and long (>15 cm) pod. Thirty-
seven accessions had long pod length, while 68 were medium and one short. Singh et al. 
(1997) reported that pod length is moderately to highly heritable with an average value of 0.7 
percent signifying pod length in cowpea is under genetic influence. Patil and Navale (2006) 
reported that pod length in cowpea is under the influence of high additive variance which 
means that the trait is controlled by additive gene action.  
Pod colour at maturity is another essential trait.  The results of grouping of the cowpea 
accessions based on pod colour at maturity showed that 60.4% of accessions were straw 
coloured, 21.7% had dark tan pod colour, 5.7% had dark brown, and 12.3% were black or 
dark purple. Similar groupings were earlier reported by Nwofia and Emeka (2014) and 
Sarutayophat et al. (2007) in cowpea, Jain and Khare (2002) in green gram. Asare et al. 
(2011) reported a dominant allelic factor controlling pod colouration.  
Pod curvature ranks among the most used traits that influence producer and consumer 
preference when the pods are tender. Diverse pod curvatures were found in this study. 
Slightly curved pod orientation accounted for 67.9% of the population, 19.8% were curved, 
5.7% had coil pod curvature, and 6.6% were straight. Nwofia and Emeka (2014) reported that 
pod curvature is controlled by dominant gene action. Furthermore, they stated that coiled 
pods were completely dominant over the straight pods. 
Significant variation was observed in seed shape for different cowpea accessions, with 17% 
of the genotypes having kidney seed shape, 13.2% ovoid shaped, 7.6% globose shape and the 
remaining rhomboid shape. Nkouannessi (2005) and El Naim et al. (2012) reported that the 
seeds of the same cultivar might vary in shape. Furthermore, they stated that the shape may 
be influenced by the position of seed in the pod which may be influenced by environmental 
conditions during pod filling stage.  
57 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
The genotypes varied in eye colour and pattern. Eye colour included white-cream eye 
colouration which was the most observed (89.6%), brown splash or grey eye colour, red tan 
brown and blue to black eye colour. The eye pattern commonly found in the 106 accession 
were very small eye pattern (40.6%), small (39.6%), holstein, and self-coloured (eye covers 
the whole seed). The seed testa texture varied significantly with 53.8% having smooth testa 
texture, 26.4% smooth to rough and 19.8% rough to wrinkle. Nkouannessi (2005) found seed 
testa texture ranging from rough to wrinkle. A similar trend of smooth to rough seed coat 
texture was reported by Adewale et al. (2011) in cowpea. 
There was a significant difference among the accessions for hundred seed weight. The value 
ranged from 10.8 g (GH 4527) to 17.4 g (WAC91 and Gorom). Based on this, genotypes 
were grouped as light (<13.0 g), medium (13−15 g) and heavy (>15 g) seed weight. Among 
the genotypes, only one had light 100-seed weight, 103 were medium, and two cultivars had 
heavy 100-seed weight. This variation among the genotypes could be due to inherent 
genotypic differences and environment and could be used for yield improvement. Rashwan 
(2010) reported a cowpea population had highly significant positive additive × dominance 
epistatic effect for this trait. The genotypic variation in test weight may also be due to the 
varied capacity of the genotype to utilize the reserved assimilates.  
Seed yield per plant varied significantly within and among genotypes. Based on this, the 106 
genotypes were grouped into three categories; low, moderate, and high yielders. Seed yield 
per hectare ranged from 503.3 kg (cluster 1 members) to 101.00 kg (cluster 8) with an 
average yield of 212.9 kg. Grain yield differences depend on a number of factors, and the 
most notable ones are number days to maturity with late and early maturity genotypes having 
high and low seed yield, respectively. Yield also depends on the inherent capacity of 
genotype, seed size, the influence of cultural practices and environmental conditions. 
58 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
Siddique and Gupta (1991) reported that additive gene effects were significant in 
conditioning grain yield in cowpea.  
Among the 13 quantitative characters studied, almost all the traits are important characters 
contributing to the diversity among the 106 accessions. These findings corroborate with the 
results obtained by Backiyarani et al. (2000) who found that every trait studied should add 
into genetic divergence displayed by accessions. 
To both consumers and farmers, the most important traits of cowpea grain/seed are seed 
colour, seed size, cooking ability, testa and seed coat. Accessions from cluster 1 and 8 had 
the most desired features high yield, maturity, seed size, and colour. South Sudanese farmers 
and consumers preferred varieties are AGRAC−116, AGRAC−216, AGRAC−316, Apagu 
1B, Titinwa A, and Titinwa B, Beledi A, Beledi C, and Laduni 1B. Accession from West 
Africa, especially members of clusters 2, 4 and 7 (Padi−tuya, Songotra, Black eye, 
IT93K−503−1, Dan lla and Mouride) with white seed colour, distinctive black and brown 
eyes, smooth, rough, smooth to rough seed coat. Individually accessions with bigger seed 
size were AGRAC−116, AGRAC−216, Apagu 2A, Beledi A, Black eye, WAC91, Rec 003, 
IT08K−149−3, Pobe Local, and WAC21. The 100−seed weight of the accessions ranged 
between 19.0–23.0 g.  Omoigui et al. (2006) also reported a range of 20.0−22.0 g for 
100−seed weight in cowpea. 
All crop improvement programs in South Sudan are still at their initial stages. This is the 
pioneer study that for the first time ever assessed the genetic potential of some cowpea 
accessions from South Sudan using molecular marker. To date, there is no genotypic 
information about the genetic variability existing in them. Knowledge of the genetic value of 
each accession will provide good evolutionary information about these accessions and is step 
towards an effective breeding program (Becerra et al., 2017) of cowpea in South Sudan. 
59 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
Huynh et al. (2013) reported that grouping accessions into groups of relatedness will help 
population development and enhance germplasm management. 
The SNP marker analysis grouped the 106 cowpea accessions into six clusters in the 
dendrogram. However, the phenotypic analysis showed nine clusters. The differences in 
results between the two approaches may be attributed to the ability of molecular markers to 
provide insights into genetic relationships within and between sub-populations. However, 
morphological markers are measurable and can be influenced by GEI effect. This finding 
agrees with results of Karampatakis et al. (2017) who reported a significant difference 
between phenotypic and genotypic diversity within same population. The principal 
coordinate analysis gave a clear insight about the diversity in the 106-cowpea accession as 
each and every set had a unique trait (tolerance and susceptibility to drought stress, high yield 
and lateness. This collection may constitute a useful stock for any future cowpea 
improvement in the country.   
The maximum likelihood and Bayesian analysis revealed that genotypes were grouped based 
on a threshold set. Those members with probability level greater than or equal to 0.35 were 
grouped into the same subpopulation (cluster), whilst genotypes with probability level less 
than 0.35 were put into an admixture group. The observed admixtures could be the result of 
inbreeding. However, one of the most important evolutionary forces, migration, was found to 
be responsible for the gene flow among accessions. This result agrees with the finding of Ali 
et al. (2015), who attributed diversity to exchange of materials between breeder, natural and 
artificial selection, genetic drift and environmental variation. This flow gives rise to 
individuals known as admixtures. Admixtures provide essential information about the genetic 
variation needed for selection. McTavish and Hills, (2014) found that, admixtures could 
hinder opportunities for local adaptation. Conversely, Buerkle and Lexer, (2008) reported 
admixtures could be utilised as basis for genetic mapping. The fixation index of the 
60 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
accessions revealed wide range of genetic differentiation. Jakobsson et al. (2008) used Fst as 
an evolutionary parameter with an assumption of SNP markers being polymorphic taking 
into consideration the component of genetic variation within a cluster. The Fst value less than 
0.05 means there was little genetic difference among members in a cluster, 0.05 to 0.15 as 
moderate genetic difference, 0.15 to 0.25 showing the great genetic difference and greater 
than 0.25 as an indication of very great genetic difference. In this study, the mean expected 
heterozygosity in the population was 0.21. Low heterozygosity value implies that the 
collection had a narrow genetic base. This finding corroborated with those of other 
researchers (Asare et al., 2011; Ali et al., 2015).  
3.5 Conclusions 
 
The genetic distance among and within individual accessions resulted in their separation into 
nine distinct clusters with a cophenetic correlation coefficient of 0.76. The first three 
principal components accounted for 65% of the total variability in the panel, indicating that 
all the traits scored varied significantly. These accessions could serve as important breeding 
stock for the cowpea development, particularly in South Sudan and Togo where limited 
breeding activities for cowpea are undertaken. The most critical yield components were the 
number of pods per plant, seeds per pod and 100−seed weight. Individually, high value for 
these traits were found in GH 4527, Laduni 1B, AGRAC−216, AGRAC−316, Titinwa A, 
Titinwa B, Beledi A, Beledi B and Apagu 1B and in six advanced breeding lines Songotra, 
Padi−tuya, Black eye, Dan lla, Mouride, IT93K−503−1 and IT97K−499−35. The accessions 
that are early to medium maturing are GH 4527, BLK 452, Titinwa A, Titinwa B, Laduni 1B, 
Beledi A, Beledi C, AGRAC−316 and those that showed a high level of drought tolerance 
were IT98K−503−1, IT97K−499−35, Dan lla and Mouride. Furthermore, the molecular study 
confirmed the existence of genetic divergence among the assembled cowpea panel. This 
panel was grouped into six: high yielding accessions, drought tolerant, susceptible, late 
61 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
maturing and medium maturing. This has created an opportunity for population development 
through introgression of new alleles from different backgrounds into adapted local varieties. 
The SNP markers used in this study could be utilized to analyse and group other new set of 
collections in the future. 
 
 
 
 
 
 
 
 
 
 
 
 
 
62 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
CHAPTER FOUR 
4.0 IDENTIFICATION OF DROUGHT TOLERANCE AMONG DIVERSE COWPEA 
GERMPLASM 
 
4.1 Introduction 
Parental line selection in cowpea improvement is a crucial factor in population development 
or hybridisation process. For drought tolerance, earliness and other physiological traits such 
as chlorophyll content, stomatal conductance, delayed leaf senescence, stem greenness, and 
root architecture are essential features for parental line selection. Seed size, colour, and shape 
are very important traits in addition to the objective of yield improvement under drought 
stress conditions. For efficient drought tolerance screening, it is advisable to explore both 
empirical and physiological approaches either through gauging canopy temperature or root 
traits because there is a direct relationship between water lost through transpiration and 
absorption by the root system. Numerous screening methods for identification of drought-
tolerant cowpea genotypes have been used. These methods include pot evaluations, wooden 
boxes, pin boards and field evaluations (Watanabe et al., 1997; Mai-Kodomi et al., 1999a; 
Matsui and Singh, 2003; Muchero et al., 2008; Agbicodo et al., 2009).  
There is little information about the level of drought tolerance in the locally adapted, farmers 
preferred cowpea varieties in South Sudan despite research by International Institute of 
Tropical Agriculture (IITA), University of California, Riverside (UCR) and other National 
Agricultural Research Organisations (NARO) in Africa. Genotype selection for drought 
tolerance is very important for world food production, primarily in arid regions with erratic 
rain distribution. The objective of this work was to screen 49 cowpea genotypes for drought 
tolerance under rainout shelter, identify and select drought tolerant cowpea genotypes for 
genetic improvement.  
  
63 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
4.2 Materials and Methods  
4.2.1 Genetic Materials and Experimental Site 
Forty-nine cowpea genotypes were selected after a preliminary drought tolerance screening 
of 106 genotypes under a rainout shelter at Crop Science Farm of the University of Ghana in 
May 2016. These screened genotypes were collected from farmers in South Sudan, and five 
research Institutes in West African countries of Togo, Ghana, Niger, Burkina Faso and 
Nigeria (IITA). The experiment was laid out in lattice square design (7×7) replicated twice 
with accessions and watering regimes as treatments. Four healthy looking seeds were planted 
in 10 litre grafting plastic pots with a diameter of 40 cm. Seedlings were thinned to two 
plants per pot eight days after planting. Each pot was filled with topsoil known as Adenta 
series after sieving. The soil series was analysed and found to have pH of (4.47−4.71) and 
30−35 % clay, (0.5 %) loam, (5−10 %) sand and (60−62.5 %) organic matter, available P 
(3.26 mg/kg), (Department of Crop Science, University of Ghana). The pots were watered at 
field capacity two days prior to planting. Moisture stress was applied as described by 
Muchero et al. (2008).  
The plants were watered to field capacity (moisture content 30%) until the reproductive stage 
was attained (thirty days from germination) and then water was withdrawn for three weeks 
for drought response measurements. During the period of stress, day and night temperatures 
were recorded as 22.3−28.6°C and 16.3−21.4°C, respectively using Thermo-hygrometer 
device. The soil moisture content during the water stress was monitored at 20 cm depth using 
Fieldscout TDR 150 once in a week for the second and third week of imposed drought stress. 
After the period of imposed drought stress, plants were watered twice a week for two weeks 
and the recovery scored (regrowth and stem greenness). In total, 11 variables were recorded 
after stressing the plants to assess drought tolerance of the 49 accessions as described in 
Table 4.1.  
64 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
Table 4.1: Eleven variables used to categorise the drought tolerance of the 49 cowpea 
accessions  
Variable Identifier   Description 
LWI 1 Leaf wilting index after the first week of stress 
LWI 2 Leaf wilting index after the second week of stress 
LWI 3 Leaf wilting index after the third week of stress 
IB 2 International Board on Plant Genetic Resources scale after the second week of drought stress 
IB 3 International Board on Plant Genetic Resources scale after the third week of drought stress 
IB 4 International Board on Plant Genetic Resources scale after the fourth week of drought stress 
MAIK Mai-Kodomi et al. (1999) scale after the second, third and fourth week of drought stress 
SMC1 Moisture content after the second week of drought stress 
SMC2 Moisture content after the third week of drought stress 
STG Stem greenness after 2 weeks of re-watering 
Re-growth Resumption of growth after 2 weeks of re-watering 
 
4.2.2 Data Collection  
Data of both qualitative and quantitative traits were recorded using the cowpea descriptors 
developed by the International Board for Plant Genetic Resources (IBPGR, 1983).  
i. Delayed leaf senescence (DLS) as visual scoring on a scale of 1−5, where 1 = totally 
green and turgid and 5 = completely yellow to brown almost dead.  
ii. Stem greenness was scored using a scale of 1-5, where 1 was brown and 5 completely 
green.  
iii. Re-growth was scored using three categories as: 1 with no recovery, 3 recovery from 
axillary buds, and 5 recovery from the apical stem.  
iv. Leaf Wilting Index (LWI) was calculated using the ratio of leaves showing wilting 
signs and a total number of leaves per plant.  
v. Drought stress index (DSI) as a visual score on a scale of 1−9, where 1 low 
susceptibility, 5 medium susceptibility, and 9 highly susceptible in order to identify 
genotypes with ‘Type 1’ and ‘Type 2’ drought tolerance. 
65 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
vi. Soil moisture level was measured using the Fieldscout TDR 150 operation 
instrument. 
vii. Days to first flowering was recorded as the number of days from planting to a stage 
when single plants in a plastic pot began to flower.  
viii. Days to 95% maturity was scored as the number of days when the majority of the 
pods had attained physiological maturity.  
ix. At flowering, plant height was measured in cm on randomly selected plants in each 
plastic pot, taking measurements from the base of the plant to the last node on the 
main stem.  
x. Number of seeds and locules per pod was recorded as an average number of locules 
and seeds from 10 pods,  
xi. whereas the number of pods per plant was recorded as the mean of matured pods 
from randomly selected plants.  
xii. Number of branches per plant was counted as the branches whose origin is in the leaf 
axils on the main stem recorded at the 8th week after sowing.  
xiii. Pod lengths from randomly selected plants in each plastic pot were measured in cm.  
xiv. Grain yield adjusted to 12% moisture content was computed from the grain weight of 
harvested pods per plastic pot.  
4.2.3 Data Analysis 
Analyses of data combined across two contrasting environments (well-watered and drought- 
stressed) were performed using GenStat 18th Edition considering all effects as random except 
genotypes according to Vargas et al. (2014). META-R software was used to compute 
phenotypic correlation. IBM SPPS Statistics software version 22nd was used to compute 
cluster and dendrogram. 
66 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
4.3 Results  
4.3.1 Leaf Wilting Index (LWI), Levels and Types of Drought Tolerance 
Three weeks of drought stress induced highly significant differences among the genotypes 
for leaf wilt indices (Table 4.2). Of the 49 genotypes evaluated, 22.5% showed no sign of 
wilting in the first week while 77.6% showed signs of leaf wilting with GH7228 having the 
highest wilting ratio (0.69) and IT99K−573−11 having the lowest (0.13). Increase in LWI 
were observed from the second to the third week. The highest LWI was 0.96 (GH7875) and 1 
for GH 7828 and Rec105, respectively. The lowest values were 0, 0.19, 0.26 and 0.33. The 
highest values were observed in GH7828 and Rec105 and the lowest values were registered 
in Mouride, IT93K−503−1, Dan lla, KVX−404−8−1 and IT97K−499−35 in the second and 
third week (Table 4.2). Susceptible genotypes in the collection accounted for 14.29%. Those 
that showed ‘Type 1’ drought tolerance accounted for 24.49% and these are plants that cease 
growth but preserve moisture and keep all the leaves and growing tips alive for long period 
of time. Whereas, ‘Type 2’ accounted for 62.22% and they assemble moisture from the lower 
leaves to the growing tips remain alive for longer period of time compare to ‘Type 1”.  
(Table 4.3). 
T able 4. 2:  Mean squares of 10 measured traits of 49 cowpea genotypes 
Traits F Sig. 
Mean of aborted pods 55.61  2.46 0.05 
Leaf wilting index 1 0.04  0.63 0.64 
Leaf wilting index 2 0.02  0.35 0.84 
Leaf wilting index 3 0.03  0.76 0.05 
Soil moisture content 1 3.16  1.31 0.04 
Soil moisture content 2 2.03  0.55 0.05 
Drought stress index 4.85  1.23 0.00 
Stem greenness  3.53  0.89 0.04 
Re-growth 1.79  0.75 0.06 
Delayed leaf senescence  2.36  2.29 0.04 
F: Computed F value and Sig. Level of significance  
 
  
67 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
                    Table 4. 3:  Drought tolerance scores* and moisture levels in 49 cowpea genotypes 
Genotypes     Drought tolerance Leaf wilting      Moisture level 
type Re- Aborte
 DLS DSI ** STG growth d pods LW1 LW2 LW3 SMC2 SMC3 
AGRAC116 2 5 2 1 3 6.5 0.60 0.62 0.78 38.00 32.35 
AGRAC216 2 5 1 5 5 5.0 0.53 0.73 0.85 37.50 35.75 
AGRAC316 3 1 2 5 3 5.0 0.23 0.40 0.65 36.75 34.65 
Amos V 3 5 2 1 3 2.5 0.34 0.77 0.84 37.50 35.40 
Apagu 1A 4 5 2 1 3 12.5 0.36 0.70 0.87 37.75 34.35 
Apagu 1B 2 5 S 1 1 10.5 0.29 0.76 0.82 34.75 30.65 
Apagu 2A 2 1 2 5 3 6.0 0.32 0.52 0.71 38.50 36.90 
Beledi A 2 5 S 1 5 9.0 0.57 0.78 0.93 38.50 35.55 
Beledi C 3 5 S 1 5 5.0 0.49 0.62 0.88 37.50 36.15 
BK 4 5 S 1 5 12.5 0.40 0.56 0.81 37.25 34.80 
Dan lla 1 1 2 5 5 0.5 0.00 0.26 0.47 37.50 36.05 
GH7228 5 5 2 1 1 8.5 0.69 0.74 1.00 38.50 37.10 
GH7875 4 5 1 1 3 10.0 0.48 0.74 0.96 36.50 34.35 
Gorom 2 1 2 5 5 0.0 0.15 0.47 0.40 38.75 36.95 
IT08K125100 3 1 2 5 5 0.5 0.58 0.62 0.95 36.75 35.25 
IT08K15024 4 5 2 1 5 2.5 0.19 0.62 0.76 38.25 35.15 
IT08K18724 5 1 1 5 1 0.0 0.45 0.66 0.81 36.75 34.75 
IT10K8177 3 1 1 5 5 0.0 0.00 0.60 0.73 37.25 35.90 
IT93K5031 1 1 2 5 3 0.5 0.00 0.19 0.50 38.25 34.95 
IT96D604 2 5 2 5 3 0.0 0.17 0.67 0.88 38.00 36.85 
IT97K49935 2 1 2 5 5 0.0 0.00 0.37 0.71 37.50 36.00 
IT97K49938 3 1 2 5 5 0.0 0.00 0.53 0.59 37.25 35.35 
IT97K56818 2 1 1 5 3 1.0 0.00 0.49 0.55 38.75 37.00 
IT98K10921 2 1 1 5 5 0.0 0.18 0.55 0.74 37.75 36.05 
IT98K2058 2 1 1 5 5 0.0 0.15 0.62 0.77 39.00 37.00 
IT98K5061 2 5 2 1 3 0.0 0.00 0.42 0.79 37.50 36.70 
IT99K57311 3 1 2 5 3 0.5 0.13 0.64 0.84 37.75 36.00 
IT99K57321 2 1 1 5 5 0.0 0.24 0.69 0.93 36.75 34.70 
KNI 3 1 2 5 1 0.0 0.28 0.42 0.59 38.00 36.80 
KVX303096G 2 5 1 5 5 0.0 0.18 0.64 0.81 39.25 37.85 
KVX396452D 2 1 2 5 5 0.0 0.00 0.40 0.53 39.00 36.05 
KVX40481 2 1 1 5 5 0.0 0.19 0.33 0.40 38.70 33.80 
KVX745118 3 1 1 5 5 0.5 0.00 0.59 0.81 38.00 36.60 
Laduni1B 4 5 S 1 5 4.5 0.67 0.81 0.95 38.25 35.75 
Laduni2B 4 5 2 5 3 14 0.46 0.71 0.95 38.00 35.40 
Mangala Mangala A 2 5 2 5 3 8.0 0.20 0.77 0.83 39.00 34.00 
Mangala Mangala B 2 1 2 5 3 4.5 0.19 0.43 0.71 37.50 34.60 
Mouride 1 1 2 5 5 1.0 0.00 0.00 0.28 36.10 33.95 
Padi_tuya 2 5 2 5 3 6.5 0.18 0.79 0.83 37.50 35.80 
Pobe local 1 5 2 5 5 0.0 0.00 0.47 0.82 37.75 35.40 
Rec016 5 5 2 1 1 6.5 0.26 0.65 0.87 37.00 33.20 
Rec041 4 5 1 1 1 1.5 0.22 0.54 0.73 38.65 34.60 
Rec105 3 5 2 1 5 4.0 0.31 0.71 1.00 38.70 36.85 
Songotra 2 5 2 5 3 4.5 0.44 0.52 0.71 38.75 35.35 
Titinwa A 3 5 S 1 5 10.5 0.33 0.65 0.95 38.75 36.25 
Titinwa B 3 5 S 1 5 9.0 0.53 0.82 0.94 34.00 32.40 
Titinwa C 3 5 2 1 3 8.5 0.50 0.82 0.93 38.00 34.25 
WAC101 3 5 2 1 5 11.5 0.37 0.83 0.93 37.75 34.10 
WAC121 4 5 2 1 1 11.0 0.41 0.79 0.95 37.25 34.45 
*Planted on September 02 2016 and stressed at reproductive stage October 02, 2016 
**Type 1- total plant tolerance, Type 2-moisture mobilisation from lower leaves to tips, S- Drought susceptible, 
DLS: Delayed leaf senescence, DSI: Score of drought stress, STG: Stem greenness, LWI: Leaf wilting index 
and VWC: Volumetric water content  
 
68 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
4.3.2 Wilting Scales (Delayed Leaf Senescence) 
Wilting scored according to IB scale showed a highly significant difference (P<0.001) among 
the accessions (Table 4.2). Rec016, IT08K−187−24, and GH7228 had a score of 5, indicating 
that these genotypes almost died. IT93K−503−1, Mouride, Pobe local, Dan lla with score of 
1 were totally green with turgescent leaves. Whereas Titinwa A showed green to yellow and 
wilted leaves (3), Laduni 1B had yellow-green and severely wilted leaves (4) and Beledi A, 
Apagu 1B and Beledi C (2) had green slightly wilted leaves.    
4.3.3 Score of Drought Stress Index (DSI) 
Scores of drought stress index showed highly significant differences (P<0.001) among the 
evaluated genotypes (Table 4.2). Genotypes AGRAC−116, Apagu 1B, IT98K−506−1, 
Rec016, Titinwa A, Titinwa B, WAC 121 and others accessions had a score of 5 indicating 
that these genotypes were completely dead after being stressed for three weeks. Dan lla, 
IT99K−573−21, Mouride, IT93K−503−1 and many others had green turgid leaves with a 
score of 1. 
4.3.4 Soil Moisture Content (SMC) 
There were significant differences (P<0.05) among the genotypes relating to measured 
moisture and depletion in the second and third week of water stress. The highest soil 
moisture level in the second week of stress was recorded in Rec105 (40.58) and Titinwa B 
had the lowest moisture level (34.00). Whereas, in the third week, Rec105 continued to have 
the highest moisture level and Apagu 1B had the lowest (32.00). 
4.3.5 Re-growth and Stem Greenness (STG) 
Both stem greenness (STG) and re-growth in the assessed genotypes showed highly 
significant differences at a probability level of 0.001 (Table 4.2). For regrowth, a score of 5 
was observed in AGRAC−216, IT08K−150−24, IT97K−499−35, KNI, Pobe local, 
69 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
IT97K−499−38, IT93K−503−1, IT98K−109−21, KVX−303−096G, Mouride, Dan lla, and 
many others indicating that these genotypes had apical meristem recovery. Apagu 1B, 
GH7828, Rec016, Reco41, and WAC121 had a score of 1 indicating no regrowth after the 
resumption of watering for two weeks. Apagu 1A, AGRAC−316, Amos V, Beledi A, Beledi 
C, GH7875, Mangala Mangala A and other genotypes with score 3 had auxiliary buds 
regrowth. For STG, almost forty-three percent of the genotypes including Apagu 1A, Apagu 
1B, and AGRAC−116 had completely yellow stems. AGRAC−316, Padi-tuya, Dan lla, 
Mouride, IT97K−499−35, IT93K−503−1 had completely green stem after alleviation of 
drought stress (Table 4.3). 
Table 4. 4: Descriptive statistic for quantitative parameters measured under drought stress     
condition 
Parameters  Mean ± SE Range CV (%) 
Plant Height (cm) 27.17 ±1.33 7.7 ± 82.5 44.77 
Number of days to the first flower 37.76 ± 0.38 30.0 ± 49.0 10.06 
Number of days to 95% maturity 62.86 ± 0.48 50.0 ± 76.0 8.40 
Number of pods per plant 16.41 ± 0.63 3.3 ± 44.0 31.82 
Pod length (cm) 10.29 ± 0.32 4.7 ± 27.0 23.93 
Number of locules per pod 10.57 ± 0.29 3.9 ± 17.0 19.49 
Number of seeds per pod 7.81 ± 0.25 2.6 ± 15.0 27.58 
Yield (g) 2.31 ± 0.13 0.30 ± 6.30 47.48 
SD: Standard deviation, CV%: Coefficient of variation  
 
Significant differences in all the traits were recorded with a mean of 27.17 cm plant height 
under stressed conditions and 54.10 cm under well-watered conditions. The number of pods 
per plant was equally affected by the two treatments. Plants under optimum condition 
produced a higher number of pods (24.78) versus 16.41 for those under water stressed 
conditions. Yield under stressed conditions was 50% (2.31 g) lower than yield under well-
watered conditions (5.11 g) (Table 4.5). Combined analysis of variance revealed significant 
differences for the interaction between genotypes and environment at P<0.001 level of 
significant for plant height, number of pods per plant, number of seeds per pod and yield. 
Whereas P<0.05 probability level was recorded for pod length and the remaining traits 
70 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
(number of days for the first flower, number of days to 95% maturity and number of locules 
per pod) showed no significant interaction with the environment. The 49 cowpea accessions 
were grouped into two distinct clusters and five sub-clusters based on the Euclidean distance 
using Ward’s method (Fig. 4.1). Accessions from South Sudan are found across sub-clusters 
1, 2, 4 and 5. Sub-clusters 2 and 3 predominantly contain accessions introduced from West 
African countries (Table 4.6). 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
71 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
Table 4. 5:  Mean of traits of 49 cowpea genotypes evaluated across two environments 
Env Height (cm) First flow 95%Maturity Number of pods Pod length (cm) Number of locules per pod Number of seeds Yield (g) 
Stressed Mean 27.17 38 63 16.41 10.29 10.57 7.81 2.31 
Std. Deviation 13.13 3.78 4.77 6.23 3.15 2.86 2.47 1.24 
Watered Mean 54.10 37 64 24.78 11.68 11.66 9.12 8.03 
Std. Deviation 19.84 2.81 4.41 10.23 2.52 3.03 2.66 5.11 
Across Mean 40.63 38 63 20.60 10.99 11.12 8.47 5.17 
Std. Deviation 21.53 3.33 4.60 9.43 2.93 2.99 2.64 4.69 
CV%  47.79 7.90 7.87 19.61 20.21 17.83 23.84 62.76 
LSD  3.82 1.84 1.43 2.62 3.28 2.83 2.96 4.61 
Height (Plant height), First flow (number of days to first flower), 95%maturity (number of days to 95% maturity), Number of pods (number of pods per 
plant), Number of seeds (number of seeds per pod) 
 
 
 
 
 
 
 
 
 
 
 
 
72 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
Table 4. 6:  Distribution of 49 cowpea accessions in five clusters according to quantitative 
traits  
Sub-cluster Number of Genotypes Attributes   Genotypes 
1 10 Medium maturing  Rec041, WAC101, Rec105, GH7875, 
GH7228, Laduni 2B, Laduni 1B, Amos 
V, Rec016 and AGRAC−316 
2 11 High yielding  KVX396452D, Beledi A, IT98K5061, 
KVX40481, Pobe local, WAC121, 
IT08K15024, IT08K125100, Songotra, 
IT08K18724, BK 
3 17 Drought tolerant Mouride, Gorom, KVX303096G, Dan 
lla, KVX745118, IT99K57321, 
IT93K5031, IT96D604, AGRAC116, 
KNI, AGRAC216, IT97K56818, Padi-
tuya, IT98K10921, IT97K49935, 
IT97K49938, IT98K2058 
4 6 Susceptible to drought Titinwa C, Beledi C, Titinwa A, Apagu 
1A, Titinwa B, Apagu 1B,   
5 5 Late maturing  Mangala Mangala B, IT99K57311, 
Mangala Mangala A, IT10K8177, 
Apagu 2A, 
 
 
Cluster 3 had the largest number of genotypes with 17 members representing 34.7% of 
the genotypes. In the second position was cluster 2 whose membership accounts for 
22.5% of the genotypes. The third was cluster one with 10 members representing 20.4% 
of the genotypes. Cluster four predominately contains accessions from South Sudan and 
constituted 12.2% of the total accessions. Cluster 5, with five accessions, represented 
10.2% of the 49 genotypes studied (Table 4.7). Within cluster three, Dan lla stood out as 
a unique genotype. The dendrogram revealed that the most diverse genotypes were 
Apagu 2A and Mangala Mangala B (Fig. 4.1). 
Cluster analysis using Ward’s method of minimum variance exhibited a distinct pattern 
of group formation (Table 4.7). The shortest number of days characterised the genotypes 
in cluster IV to first flower, and the highest number of seeds per pod, number of pods per 
plant and grain yield. cluster V showed a longer number of days to first flower and the 
73 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
highest number of seeds per pod, while members of cluster III registered the lowest grain 
yields.  
Table 4. 7: Mean values of cowpea agronomic traits measured under drought stress and well-
watered conditions  
 Mean 
Parameters Sub-cluster Sub-cluster Sub-cluster Sub-cluster Sub-cluster 
1 2 3 4 5 
Plant height (cm) 43.94 39.33 39.02 39.45 43.84 
Number of days to the first 
flower 36.00 37.00 39.00 36.00 40.00 
Number of days to 95% maturity  63.00 63.00 64.00 63.00 64.00 
Number of pods per plant 20.71 18.26 22.27 23.48 16.37 
Pod length (cm) 11.44 11.71 10.15 10.45 11.99 
Number of locules per pod  11.94 11.44 9.97 11.15 12.60 
Number of seeds per pod 8.66 8.76 7.61 9.02 9.67 
Grain yield (g)  5.04 5.79 4.75 5.68 4.85 
 
4.3.6 Correlation between Leaf Wilting Index and Drought Tolerance Traits 
Correlation analysis of drought tolerance related traits revealed significant differences at a 
significance level of 0.05 for various traits (Table 4.8). Drought stress index (DSI) was found 
to be strongly correlated with LWI and STG, but negatively correlated with regrowth. LWI 
was strongly but negatively correlated with regrowth and STG and regrowth positively 
correlates with LWI. However, stem greenness positively associates DLS and regrowth and 
negatively correlates with LWI 1, LWI 2 and LWI 3.   
Table 4. 8: Associations among 10 drought related traits 
DLS 1 -         
DSI 2 0.37 -        
LWI 1 3 0.47 0.50 -       
LWI 2 4 0.47 0.52 0.59 -      
LWI 3 5 0.47 0.57 0.58 0.87 -     
Regrowth 6 -0.47 -0.35 -0.37 -0.32 -0.29 -    
STG 7 0.52 -0.60 -0.50 -0.44 -0.44 0.41 -   
SMC2 8 -0.07 -0.03 -0.12 -0.11 -0.07 0.33 0.17   
SMC3 9 -0.03 -0.08 -0.25 -0.16 -0.09 0.34 0.27 0.74 - 
   1 2 3 4 5 6 7 8 9 
All the bold values are significant at 0.01probability level 
Note: Leaf wilting indices (LWI), Delayed leaf senescence (DLS), Drought stress index (DSI), soil moisture 
content (SMC), and stem greenness (STG)  
74 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
. 
I 
II 
 
Fig. 4. 1 Dendrogram resulting from the analysis of 49 cowpea accessions grouped into I susceptible and II 
drought tolerant genotypes obtained by Ward method using the overall distance of Euclidean 
75 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
4.3.7 Phenotypic Correlation  
Phenotypic correlation analysis showed a positive and strong correlation between plant height 
and number of seeds per pod (r = 0.30) and grain yield (r = 0.31) at 0.05 significance level. 
Whereas, number of days to 95% maturity was strongly but negatively correlated with pod 
length (r = −0.46) and number of locules per pod (r = −0.39) at P<0.05 and (r = −0.36) 
number of seeds per pod at P<0.01 level of significance. Likewise, strong and positive 
association was found between pod length and the number of locules per pod (r=0.73) and 
number of seeds per pod (r = 0.67) at P<0.01 and with seed yield (r = 0.33) at P<0.05. 
Number of locules per pod had a strong association with number of seeds per pod (r = 0.87) at 
P<0.01 and seed yield (r = 0.21) at P<0.05. Positive and strong correlation was found between 
number of seeds per pod and grain yield. Highly significant and positive correlation was 
observed between number of pods per plant and grain yield (r = 0.29) at a significance level of 
P<0.01 (Table 4.9). 
 
 
 
 
 
 
 
 
 
 
 
 
76 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
Table 4. 9: Phenotypic correlation coefficients between morphological, grain yield and sub- yield components of cowpea genotypes measured 
under drought condition  
Traits Plant height (cm) First flower Maturity Number pods Pod length (cm) Number of locules Seeds per pod Yield (g) 
Plant height (cm) 1        
First flower -0.05 1       
Maturity  -0.14 0.04 1      
Number pods -0.06 0.03 -0.14 1     
Pod length (cm) 0.18 0.06 -0.46* 0.01 1    
Number of locules 0.21 -0.17 -0.39* -0.16 0.73** 1   
Seeds per pod 0.31* -0.13 -0.36** -0.13 0.67** 0.87** 1  
Yield (g) 0.30* -0.06 -0.13 0. 25** 0.33* 0.21* 0.29** 1 
* Significant at 0.05 probability level and ** Significant at 0.01probability level 
 
 
 
 
 
 
 
 
 
 
77 
 
University of Ghana htt p://ugspace.ug.edu.gh
 
4.4 Discussion  
The genetic variability amongst the 49 accessions in responsiveness to water deficit became 
pronounced as the duration of induced stress continued. Varietal differences were apparent in 
the second week of the induced stress. Apagu 1B, GH7828, Beledi A, Beledi C, BK, Laduni 
1B, Titinwa A, and Titinwa B completely wilted. Whereas, the remaining 42 accessions 
exhibited either Type 1 or 2 of drought tolerance. Cowpea germplasms with ‘Type 1’ drought 
tolerance in this study are AGRAC−216, GH7875, IT08K−187−24, IT98K−503−1, 
IT97K−568−18, IT98K−205−8, IT99K−573−21, KVX−3030−96G, KVX−404−81, 
KVX−745−118, and Rec041. They stopped growing after the induction of drought stress and 
kept their uniformity and turgidity, but this declined with an increase in the duration of the 
stress, resulting in drying of the plant. These germplasms are known as dehydration avoiders, 
having strong stomatal sensitivity and reduced rate of growth. They could be utilised in the 
development of early or medium maturing varieties. These findings corroborate with the 
results obtained by Mai-Kodomi et al. (1999a) and Agbicodo et al. (2009) who used 
genotypes possessing ‘Type 1’ drought tolerance in development of early and medium 
maturing cowpea varieties.  
AGRAC−116, Dan lla, Mouride, IT93K−503−1, and the remaining 23 accessions showed 
Type 2 drought tolerance. They stayed green longer with continuous slow growth even when 
the severity of drought stress intensified. The mechanism of staying green and remaining 
active for a lengthy period of time could be due to the ability of these genotypes to drop older 
leaves. Members with Type 2 drought tolerance have a combination of three physiological 
mechanisms: selective moisture mobilisation (hygroscopic water), stomatal regulation and 
osmotic pressure regulation. They could be used to bred varieties that may possess the ability 
to withstand drought stress at both intermittent and terminal levels. Agbicodo et al. (2009) 
and Muchero at al. (2013) reported similar findings. These results confirm the findings of 
78 
 
University of Ghana htt p://ugspace.ug.edu.gh
 
Mai-Kodomi et al. (1999b) study on Type 1 and 2 of drought tolerance in cowpea. Both type 
1 and 2 enhances the crops’ ability to cope with prolonged drought common in Semi-arid 
Africa.  
The phenotypic correlation coefficients between leaf wilting index, wilting scale (IB and 
MAIK), soil moisture content, stem greenness and regrowth are proxy traits. This agrees with 
Agbicodo et al. (2009) who used those traits as indicators of tolerance to drought stress.  In 
addition, Pungulani et al. (2013) reported that leaf wilting index could fast-track breeders 
screening for drought tolerant genotypes. They further, reported the commonly used 
approaches and traits such as delayed leaf senescence and stem greenness have their 
limitations. The visual score of qualitative traits requires experience, LWI is easy to score 
since it is a ratio of wilted to the total number of leaves per plant.  However, LWI should only 
be applied to assess crop plants that will is a good indicator of drought stress. In scenario 
some genotypes possess the ‘Type 2’ adaptability gene LWI would classify those genotypes 
susceptible. From the present study, the highest contributing traits to grain yield were number 
of seeds per pod, number of pods per plant, number of locules per pod and pod length. This 
confirms the findings of Jackai (1995) and Aremu et al. (2007) who reported that yield 
components were the main contributors to high yield. Number of days to 95% maturity was 
strongly but negatively correlated with pod length and the number of locules per pod at 
P<0.05 and number of seeds per pod at P<0.01 level of significance. This confirms the results 
of Aliyu and Makinde (2016). Likewise, strong and positive association was observed 
between pod length and the number of locules per pod and number of seeds per pod at 
P<0.01 and P<0.05, respectively, corroborating with findings of Aliyu and Makinde (2016). 
Highly significant and positive correlation was observed between number of pods per plant 
and grain yield at a probability level of P<0.01. Similarly, a positive and strong correlation 
was observed between number of seeds per pod and grain yield confirming de Almeida et al. 
79 
 
University of Ghana htt p://ugspace.ug.edu.gh
 
(2014). An increase in the number of pods per plant could result in an increased number of 
seeds per pod per plant and yield. Therefore, seed yield could be improved indirectly through 
selection for increased number of pods per plant.  
Agglomerative hierarchical clustering with Euclidean distance matrix using Ward’s Linkage 
method showed two distinct clusters with five sub-clusters. Cluster I members were 
susceptible to drought stress, early maturing and high yielding. Sub-cluster I members were 
medium maturing accession (60−75 days). Accessions in sub-cluster II are high yielding, and 
those in sub-cluster III were susceptible to drought stress. However, cluster II was made up of 
sub-cluster IV drought tolerant accessions, and sub-cluster V were late maturing accessions 
(take 90−120 days to mature). The dendrogram grouped the 49 cowpea accessions according 
to their responsiveness. The dissimilarity observed between a pair of varieties on the basis of 
a multivariate scale is useful in determination of which trait or traits are responsible for the 
dissimilarity and the relative contribution of these traits to total genetic variability (Tuhina-
Khatun et al., 2015). 
Fukai and Cooper, (1995) reported that cluster analysis associates genotype more on the basis 
of relatedness among individuals. Satish et al. (2009) also reported similar findings stating 
that geographical distribution of accessions is not the sole reason that constitutes 
morphological and genetic diversity. Several other factors may be considered as causes of the 
genetic divergence. These factors include natural or artificial selection, exchange of breeding 
materials, genetic drift and environmental variations. Therefore, identification and selection 
of parental lines should focus more on genetic potentials rather than geographical divergence.  
In the present study, sub-cluster IV was more divergent than the other sub-clusters. 
Genotypes in sub-cluster IV recorded the shortest number of days to first flowering as well as 
some yield contributing components. It is worth mentioning that breeding objectives such as 
80 
 
University of Ghana htt p://ugspace.ug.edu.gh
 
selection and hybridisation should take into consideration the genetic potentials of each 
cluster as well as individual genotypes within a cluster (Chahal and Gasal, 2002).  
4.5 Conclusions  
These materials AGRAC−216, GH7875, IT96D604 and Rec041had ‘Type 1’ and ‘Mouride, 
Padi-tuya, IT93K−503−1, Dan lla and AGRAC−116 were possessing Type 2’ drought 
tolerance. However, these three entries, IT93K−503−1, Dan lla and Mouride were identified 
as potential parents for drought tolerance. They showed high level of delayed leaf 
senescence, stem greenness and were having ‘Type 2’ drought tolerance. However, 
parameters such as delayed leaf senescence, leaf wilting index, soil moisture content, drought 
stress index, and stem greenness are good traits for selecting parents. Emphasis should be on 
leaf wilting index when identifying drought tolerant genotypes in cowpea. It is easy to score 
and no expertise required in selection of drought tolerant genotypes. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
81 
 
University of Ghana htt p://ugspace.ug.edu.gh
 
CHAPTER FIVE 
5.0 COMBINING ABILITY FOR DROUGHT TOLERANCE AND EARLINESS IN 
COWPEA 
 
5.1 Introduction 
Drought stress has an enormous effect on cowpea production and productivity, threatening 
nutrition and food security in SSA. It was observed that drought stress reduces water-use 
efficiency, disrupts photosynthetic pathways and activities, and it impairs growth and 
development. Hence, selection of accessions to be used as parental lines in hybridisation is 
one of the most important decisions faced by plant breeders. The decision has to be logical 
and careful though-through, for the reason that a population with narrow genetic divergence 
may lead to waste of money and time. Individual performance (tolerance or resistance), wide 
adaptability and yield stability have constituted the base for selection (Bertan et al., 2007). 
However, several mechanisms are involved in plants ability to cope with stress. Henceforth, it 
is important to distinguish between genetic and non-genetic aspects of tolerance to drought 
stress. Knowledge of the genetic basis of drought tolerance could be useful in developing 
novel varieties of cowpea with reasonably good yield even under drought conditions. 
Combining good seed yield with drought tolerance requires complementary gene action from 
parents into a common background. This will prepare the way for selection that could lead to 
some progenies with a good compromise of seed yield and drought tolerance.  
The specific objectives of this study were to:  
(i) estimate combining ability for drought tolerance and earliness, and  
(ii) estimate heritability of earliness and drought tolerance in the progenies. 
 
 
 
82 
 
University of Ghana htt p://ugspace.ug.edu.gh
 
5.2 Materials and Methods   
 
5.2.1 Experimental Site 
The experimental site was the University farm that is located on the University of Ghana 
campus. The soil belongs to the Adenta series, with pH of (4.47−4.71) and 30−35 % clay, 
(0.5 %) loam, coarse silt, (5−10 %) sand and (60−62.5 %) organic matter, available P (3.26 
mg/kg), (Department of Soil Science, University of Ghana).   
 
5.2.2 Combining Ability Study for Drought-stress and Earliness  
 
The genetic material used included five locally adapted and farmers' preferred varieties, 
Titinwa A, Laduni 1B, Beledi A, Apagu 1B, and Beledi C and three drought-tolerant (Dan 
lla, Mouride, and IT93K−503−1) varieties selected from germplasm screening in the same 
location. The five adapted and farmers preferred varieties were crossed with each of the three 
drought tolerant genotypes in a 5 x 3 North Carolina II mating design to generate 15 F1s.   
Then the F1 s seeds generated from the first cross were planted in the field so as to generate 
enough seeds for the evaluation.  Twenty-five entries, made up of the 15 F1s, the 8 parents 
and two checks Asontem from Ghana (early maturing) and AGRAC−216 (high yielding and 
drought tolerant) from South Sudan (Table 5.1), were planted in pruning bags with a diameter 
of 20 cm under a rainout shelter from January to May 2018.  
Seeds from the selfed F1 generation plants were used for this study.  The 25 cowpea entries 
were evaluated in randomized complete block design in a split-plot design fashion. The 
treatments were replicated four times, under two watering regimes on the 12/01/2018. These 
regimes are designated as drought stressed (DS) and well-watered (WW). The soil moisture 
content during the water stress was monitored at 20 cm depth using Fieldscout (TDR 150) 
once in a week for the second and third week of imposed drought stress. After the period of 
83 
 
University of Ghana htt p://ugspace.ug.edu.gh
 
imposed drought stress, plants were watered twice a week for two weeks and the recovery 
scored (regrowth and stem greenness). 
Table 5.1: Farmers preferred drought tolerant cowpea genotypes used as parental lines in 
North Carolina II mating design  
Parental line Days to Maturity Growth Habit Botanical Name Origin 
Drought tolerant (Male) 
Dan lla 73 Semi-prostrate V. unguiculata Niger 
Mouride 71 Semi-erect V. unguiculata Burkina Faso 
IT93K−503−1 70 Semi-erect V. unguiculata IITA/Nigeria 
Farmers’ preferred (Female) 
Beledi (C) 60 Semi-prostrate V. unguiculata South Sudan 
Laduni (1B) 63 Semi-prostrate V. unguiculata South Sudan 
Titinwa (A) 62 Semi-prostrate V. unguiculata South Sudan 
Apagu (1B) 61 Erect V. unguiculata South Sudan 
Beledi (A) 65 Semi- prostrate V. unguiculata South Sudan 
 
 
5.2.3 Data Collection 
 
The following variables were recorded. 
1. Plant height at 20 days after planting was measured using a meter ruler from the 
ground level to the base to the tip of the plant meristem expressed in cm. 
2. Number of days to first flower: recorded as the number of days from sowing to stage 
when plants had open flowers. 
3. Number of days to 50% flowering: recorded as the number of days from sowing to 
stage when 50% of plants have flowered. 
4. The number of days to first mature pods: recorded as the number of days from sowing 
to stage when 50% of plants had mature pods. 
5. Pod length (cm): computed as the mean of 10 fully mature pods randomly selected. 
6. Number of locules per pod: mean number of the 10 pods counted. 
7. Days to 95% maturity: was recorded as the number of days from seedling emergence 
to harvest of each genotype per plot (duration). 
84 
 
University of Ghana htt p://ugspace.ug.edu.gh
 
8. Number of main branches: computed as the number of branches whose origin is in the 
leaf axils on the main stem; recorded in the 8th week after sowing. Mean of 10 
randomly selected plants computed was recorded for each experimental unit. 
9. Leaf chlorophyll content: SPAD Meter was used to measure chlorophyll content at 
30, 45 and 60DAP. 
10.  Number of pods per plant: mean number of mature pods from 10 randomly selected 
plants.   
11.  100−seed weight (g): determined by randomly counting 100 seeds from a bulked 
seed and weighed using a digital weighing scale. 
12.  Seed yield (g): determined by measuring bulked seed using a digital weighing-scale 
13.  Delay leaf senescence (DLS): visual scored on a scale of 1−5, where 1 = totally green 
and turgescent and 5 = completely yellow to brown almost dead.  
14.  Drought stress index: visual scored on a scale of 1− 9, where 1 low susceptibility, 5 
medium susceptibility, and 9 highly susceptible. 
 
5.2.4 Data Analysis  
 
The model for split-plot design used was as follows: 
𝑋𝑖𝑗𝑘 = 𝑌 … 𝑀𝑖 + 𝐵𝑗   +  𝑑𝑖𝑗 + 𝑆𝑘 + (𝑀𝑆)𝑗𝑘 + 𝑒𝑖𝑗𝑘…………………………………………… (1) 
Where 𝑋𝑖𝑗𝑘 = mean observation, Y= the experimental mean, 𝑀𝑖 =the main plot factor effect, 
𝐵𝑗   = replication or block effect, 𝑑𝑖𝑗 = mean plot error equally known as error a, 𝑆𝑘 =the 
subplot factor effect, (𝑀𝑆)𝑗𝑘 =the mean plot and subplot treatment interaction effect, 𝑒𝑖𝑗𝑘 
=the subplot error also known as error b.  𝑖= a particular main plot factor, 𝑗= a particular 
number of replication or block, 𝑘= a particular subplot factor.  
Drought intensity index (DII) was quantified according to Fischer and Maurer (1978). 
𝑋
DII = 1 − 𝑆⁄𝑋   𝑝 
85 
 
University of Ghana htt p://ugspace.ug.edu.gh
 
Where Xs and Xp were the mean seed yield of all the genotypes under well-watered and 
drought stressed conditions respectively. 
Data collected were subjected to analysis of variance using the GenStat 18th edition of VSN 
international. The means of 15 F2 populations, two checks, and eight parents were compared 
in analysis of variance (ANOVA) by means of a linear model and AGD−R version 3 
Software. Combined analysis across the contrasting watering regime was carried out. The 
phenotypic variation was partitioned into males and females and their interactions. The 
individual responsiveness k of the 15 F2s resulting from the hybridization between male i and 
female j is modeled as:   
Y  = µ + r + m + f  + (mf) + e  
ijk k i j  ij ijk …………………….…………………………...… (2)
Where: 
Y  is the progeny observed value of the ith male crossed with jth female in the kth 
ijk
replication 
µ is the overall population mean 
m is general combining ability of the ith males 
i 
f  is the general combining ability of the jth females 
j
(mf)  is the specific combining ability of the ith X jth cross 
ij
r  is the replication effect 
k
e is the experimental error 
ijk 
 
Table 5.2: Format of analysis of variance for North Carolina mating design II 
Mean Square Expected mean Genetic variance 
Source of Variation D.F 
(MS) squares Components (σ2) 
Replication r−1    
GCA Male  m−1 MSm σ 2 2e  + rσfm  + rfσ 2 σ 2m m =(MSm-MSmf)/rf 
GCA   f−1 MS  σ 2Female f e  + rσ 2fm  + rmσ 2 2f  σf =(MSf-MSmf)/rm 
SCA=M x F (m−1) (f−1) MS 2mf σe  + rσ 2fm  σ 2fm =(MSmf-MSe)/r 
Error (e) mf(r−1) MS 2e σe  σ 2e =MSe 
Total rmf−1    
r = number of replications; m = number of males lines; f = number of females lines, σ2e =environmental variance, σ2f = 
variation between females, σ2m = variation between males, σ2fm = variation due to interaction between females and males  
 
Source: Kearsey and Pooni, (1996) 
 
 
86 
 
University of Ghana htt p://ugspace.ug.edu.gh
 
 
a) The additive (σ2A) and dominance (σ2D) variance components were estimated as follows:  
2 2 2 2
     σ = 4σ and       σ = 4σ  
A m D mf
b) The general combining ability for both female and male was computed as follows:  
GCA  = X  – µ and GCA  = X  – µ 
f f m m
Where:  
X  and X = Mean of male and female parents, respectively 
f m 
GCA  and GCA = General combining ability of male and female parents, respectively 
m f
µ = Overall mean of crosses in the trial. 
c) Specific combining ability was computed using the formula proposed: 
SCA = X − E (X ) = X  – [GCA  + GCA + µ] 
X X X f m
Where: 
X = Observed mean value of the cross  
X 
E (X ) = Expected mean value of the cross based on the 2 GCAs of its parents 
X
SCA  = specific combining ability of the cross x 
x
Ozimati et al. (2014) estimation formulae based on genotype mean was used to compute 
broad and narrow sense coefficient of genetic determination (BSCGD and NSCGD). 
 
BSCGD =               σ2𝐺𝐶𝐴𝑓 + σ2𝐺𝐶𝐴𝑚 +  σ2𝑆𝐶𝐴𝑓𝑚𝑓 
                          σ2𝐺𝐶𝐴𝑓 + σ2𝐺𝐶𝐴𝑚 + σ2𝑆𝐶𝐴𝑓𝑚 + σ2𝑒/𝑟 
 
                                       σ2𝐺𝐶𝐴𝑓 + σ2𝐺𝐶𝐴𝑚  
NSCGD =        σ2𝐺𝐶𝐴𝑓 + σ2𝐺𝐶𝐴𝑚 +  σ2𝑆𝐶𝐴𝑓𝑚 + σ2𝑒/ 𝑟 
 
                                            σ2𝐺𝐶𝐴𝑓 + σ2𝐺𝐶𝐴𝑚  
Bakers ratio =        σ2𝐺𝐶𝐴𝑓 + σ2𝐺𝐶𝐴𝑚 +  σ2𝑆𝐶𝐴𝑓𝑚  
 
 
 
 
 
87 
 
University of Ghana htt p://ugspace.ug.edu.gh
 
5.3 Results  
 
5.3.1 Response of F2 Populations and Parental Lines to Drought Stress 
The combined analysis of variance for the number of pods per plant, number of seeds per 
pod, 100-seed weight and seed yield for the parental lines, checks, and F1 populations showed 
significant differences at P<0.05, P<0.01 and P<0.001, respectively. The watering regimes 
exhibited significant mean squares for yield and yield component (Table 5.3).   
Table 5.3: Mean squares for responses of 15 F2 populations, eight parental lines and two 
checks subjected to drought stress and well-watered conditions at reproductive 
stage 
Source of Variation D.F NPP NSP 100 SWT Seed Yield 
Replications 3 2.5 9.7 2.8 357.7 
Watering Regime 1 75.2* 115.0** 202.2* 104868.3*** 
Main Plot Error 3 7.2 2.3 28.7 800.9 
Genotypes 24 10.3*** 8.5*** 17.5*** 463.8** 
Watering Regime × Genotypes 24 6.2 ns 5.1* 5.8 ns 417.1* 
Sub Plot Error 144 4.5 2.9 4.7 282.5 
Total 199     
*, **, *** significant at 0.05, 0.01 and 0.001 probability levels respectively; DF, degree of freedom; NPP, number of pods 
per plant, NSP, number of seeds per plant; 100 SWT, 100 seed weight; ns = non-significant at P > 0.05 
 
The estimated mean squares for delayed leaf senescence (DLS) was highly significant for 
both GCA and SCA effects. The values of GCAf and GCAm were bigger than SCAf.m and this 
implies that additive gene effects were more important than non-additive effects for DLS 
(Table 5.4).  
 
Table 5.4: Estimated mean squares for GCA and SCA for delayed leaf senescence 
Source D.F S. S M.S F-value 
GCA male 2 30.40 15.20 8.87** 
GCA female 4 27.73 6.93 4.04** 
SCA (male. female) 8 100.27 12.53 7 .31** 
Error 42 72.00 1.71 
**Significant at the ≤ 0.01 level of probability 
 
 
88 
 
University of Ghana htt p://ugspace.ug.edu.gh
 
The mean squares of the 15 F2 populations, 8 parents and 2 checks showed a significant 
difference (P<0.01) for the number of pods per plant, seeds per pod, pod length and 100-seed 
weight at and grain yield. Chlorophyll content at 30DAP, 45DAP and 60DAP were 
significantly different (Table 5.5). Mean performance of the crosses for number of pods per 
plant and number of seeds per pod were higher than the average of the parents expects for 
100-seed weight and seed yield (Table 5.6). 
At the commencement of drought stress conditions, TA×I registered the highest normalised 
vegetation index value of 50.2 whereas, BA×D had the lowest index value of 37.23. At 
45DAP, Asontem had the highest value for chlorophyll content (58.65) and BC×M had the 
lowest value of 35.5. At 60DAP, TA had the lowest chlorophyll content (34.9), whereas 
IT93K−503−1 had the highest concentration of 60.53 (Fig. 5.1). 
70
60
50
40
30
20
10
0
NDVI30 NDVI45 NDVI60
 
Fig 5. 1 Chlorophyll content of the genotypes under drought stress condition and re-watering at 30, 45 and 60 
days after planting  
 
 
 
89 
 
University of Ghana htt p://ugspace.ug.edu.gh
 
Table 5.5: Mean squares for 15F2 cowpea population, eight parents and two checks subjected to drought stress at reproductive stage 
Source of Variation  D.F NPP NSP Pod length  100−seed weight  Seed Yield NDVI 30 NDVI 45 NDVI 60 
Replication  3 4.54 7.77 2.30 12.25 357.7 10.59 10.94 47.02 
Genotype 24 8.83** 7.30** 15.58** 12.90 ** 463.8** 43.45* 54.32* 131.33* 
Error 72 3.52 3.24 3.91 4.87 293 23.05 31.73 72.95 
Total 99             
*Significant at P < 0.05, **Significant at P < 0.01, ns: non-significant, NPP: number of pods per plant, NSP: number of seeds per pod and NDVI: SPAD chlorophyll meter 
reading at 30, 45, 60DAP 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
90 
 
University of Ghana htt p://ugspace.ug.edu.gh
 
Table 5.6: Means of yield and yield components in F2 cowpea population and their eight parents evaluated under drought-stress and well-
watered conditions 
 Watering   Regimes                          DS WW DS WW DS WW DS WW 
 Population  Genotype NPP NPP NSP NSP 100−seed 100−seed Seed Yield (g) Seed Yield (g) 
   A1B×D 5.75 6.75 8.75 10.30 13.30 14.98 21.30 69.80 
   A1B×I 7.00 7.25 9.53 9.90 11.70 15 28.00 77.5 
   A1B×M 6.00 7.00 9.58 10.8 13.75 15.5 23.30 75.00 
   BA×D 4.00 6.00 8.60 8.18 13.10 15.7 31.80 55.00 
   BA×I 4.75 5.75 6.68 8.90 14.50 17.45 22.30 57.50 
 F2  BA×M 5.75 4.25 5.63 9.03 12.80 17.65 23.40 55.00 
   BC×D 3.50 5.75 6.98 10.35 14.73 16.13 24.80 72.50 
   BC×I 4.25 7.25 6.45 10.23 13.40 16.93 23.50 87.20 
   BC×M 5.25 6.50 8.88 10.33 14.93 16.5 24.80 77.50 
   L1B×D 7.75 6.50 8.38 10.85 13.58 17.9 26.30 82.50 
   L1B×I 4.25 9.50 8.73 7.80 13.73 15.43 23.40 64.80 
   L1B×M 4.25 5.75 9.03 12.00 11.65 17.4 15.90 75.00 
   TA×D 7.25 7.50 7.85 9.08 13.93 18.38 30.50 62.50 
   TA×I 8.50 8.25 6.40 9.50 16.00 17.88 30.70 72.50 
   TA×M 6.00 5.75 7.08 8.28 15.00 17.2 18.00 35.00 
 Mean   5.62 6.65 7.90 9.70 13.74 16.67 24.53 67.95 
   A1B 6.53 7.50 8.83 9.20 14.38 14.13 32.50 57.50 
   BA 4.00 8.75 8.75 9.10 13.35 15.90 20.30 67.50 
 Parents  BC 5.50 8.00 7.68 11.35 12.68 13.45 27.40 78.30 
   Dan lla 5.75 6.38 7.80 7.10 17.10 16.80 23.70 80.00 
   IT93K−503−1 4.95 5.75 6.40 8.63 16.65 16.23 27.30 70.50 
   Mouride 4.50 6.50 6.19 7.90 17.63 17.90 30.50 62.50 
   L1B 6.00 8.25 10.85 11.10 16.10 15.70 28.70 99.80 
   TA 3.75 6.50 6.53 10.68 15.00 16.83 22.30 67.50 
 Mean   5.18 7.14 7.88 9.42 15.18 15.96 26.36 72.39 
 LSD (5%)   2.75 3.38 2.57 2.23 3.20 3.09 9.90 32.83 
 Geno. Probability    * ns ** *** ** ns * ns 
*Significant at P < 0.05, **Significant at P< 0.01, ***Significant at P< 0.001, NPP: Number of pods per plant, NSP: Number of seeds per pod and 100−seed: 100−seed 
weight, DS: Drought Stress and WW: Well-Watered conditions 
 
91 
 
University of Ghana htt p://ugspace.ug.edu.gh
 
Mean squares of the F1 crosses were significantly different under drought stress and well-
watered conditions (Table 5.7) for number of pods per plant, number of seeds per pod, 
100−seed weight, seed yield and chlorophyll content at different number of days after 
planting.  
Under drought stress condition, mean squares for male parental lines were significantly 
different at P<0.05 for seed yield and chlorophyll content at 30DAP and were highly 
significantly different (P<0.01). Whereas, female parents GCA were highly significant at 
P<0.001 for number of pods per plant, 0.05% for number of seeds per pod and chlorophyll 
content at 30 and 45DAP (Table 5.7).  However, the Baker’s ratio and broad sense coefficient 
of genetic determination (BSCGD) were high for number of pods per plant, the number of 
seeds per pod, 100-seed weight and seed yield. Hence narrow sense coefficient of genetic 
determination (NSCGD) for these traits was moderate to low (<40%) (Table 5.7). 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
92 
 
University of Ghana htt p://ugspace.ug.edu.gh
 
 
Table 5.7: Mean squares, variance components for the responses of 15 F2 population evaluated under well-watered and drought stress condition 
                                                                                   Drought-Stressed (DS) Condition 
Source of Variation D.F NPP NSP             100−Seed Weigh (g)t Seed Yield (g)     NDVI30    NDVI45 NDVI60     
GCA Male 2 1.80 1.81 0.29 188.21** 141.69** 5.54 50.04 
GCA Female 4 14.04*** 12.85* 9.67** 37.39 32.29* 78.16* 65.55 
SCA (Male. Female) 8 6.84** 4.30 5.01* 73.93* 17.00 9.08 36.62 
Genotypes 14 8.18** 6.39 5.67** 79.81* 39.18* 28.31 46.80 
Error 42 3.18 3.59 5.07 60.13 20.63 35.01 52.57 
Genotype Variance   0.21 0.33 0.10 0.94 3.53 3.06 1.62 
Additive Variance   0.85 1.33 0.42 3.75 14.12 12.24 6.48 
Dominance Variance   3.66 0.70 0.00 13.80 0.00 0.00 0.00 
Environmental Variance   12.74 14.37 20.26 240.50 82.52 140.04 210.26 
The coefficient of genetic determination on entry mean         
Bakers ratio   0.70 0.77 0.67 0.75 0.91 0.90 0.76 
BSCGD   0.86 0.84 0.75 0.83 0.90 0.73 0.74 
NSCGD   0.45 0.45 0.40 0.44 0.47 0.41 0.41 
                                                                                     Well-Watered (WW) Condition 
Source of Variation D.F NPP NSP             100−Seed Weight Seed Yield (g) NDVI30 NDVI45 NDVI60 
GCA Male 2 17.92* 3.40              0.54 356.55 81.48** 9.86 40.66 
GCA Female 4 13.85* 7.77*            11.25* 1419.94** 39.59* 100.86*** 38.04 
SCA (Male. Female) 8 2.63 4.67*            3.33 441.34 13.46 34.46** 12.94 
Genotypes 14 8.02 5.37              5.19 708.83* 30.64 49.92 24.07 
Error 42 5.92 2.38              4.88 451.32 15.21 11.43 33.63 
Genotype Variance   0.86 0.11 0.30 42.56 2.73 2.46 1.77 
Additive Variance   3.43 0.45 1.19 170.22 10.94 9.84 7.08 
Dominance Variance   0.00 2.28 0.00 0.00 0.00 23.03 0.00 
Environmental Variance   23.67 9.53 19.51 1805.30 60.85 45.72 134.51 
Bakers ratio   0.93 0.71 0.78 0.80 0.90 0.93 0.73 
BSCGD   0.85 0.87 0.76 0.83 0.90 0.73 0.75 
NSCGD   0.79 0.61 0.59 0.67 0.81 0.71 0.63 
*Significant at P < 0.05, **Significant at P < 0.01, ***Significant at P < 0.001, ns: non-significant, NPP, number of pods per plant, NSP, number of seeds per pod and NDVI: 
SPAD chlorophyll meter reading at 30, 45 60 DAP, BSCGD, Broad sense coefficient of genetic determination; NSCGD, Narrow sense coefficient of genetic determination
93 
 
University of Ghana htt p://ugspace.ug.edu.gh
 
5.3.2 Estimates of General and Specific Combining Ability Effects 
Female parents Beledi A and Titinwa A showed desirable significant positive GCA for seed 
yield (Table 5.8). Beledi C and Titinwa A had desirable GCA for 100−seed weight. Whereas, 
parental lines Apagu 1B and Laduni 1B showed desirable, significant GCA for number of 
seeds per pod. Apagu 1B had significant positive GCA for normalised difference vegetation 
index (NDVI) at 45 and 60DAP while Titinwa A had a similar effect at 45DAP under 
drought-stressed condition. The parent Laduni 1B showed significant negative GCA effect 
for grain yield, 100−seed weight, and NDVI at 45DAP. Male parents, Dan lla and 
IT93K−503−1, had desirable positive GCA for seed yield, whereas Mouride showed similar 
trend but negative GCA effect (Table 5.8). Additionally, IT93K−503−1 had significant 
positive GCA effect for NDVI at 30DAP.  
Beledi A × Dan lla showed significant positive SCA effect for seed yield and number of 
seeds per pod, whereas Laduni 1B × Dan lla had significantly positive SCA effect for grain 
yield and number of pods per plant. The cross Apagu 1B × IT93K−503−1 had significant 
positive SCA effect for seed yield and a negative SCA effect for 100−seed weight (Table 
5.9). Titinwa A × IT93K−503−1 had significantly positive SCA effect for grain yield, the 
number of pods per plant, 100−seed weight and drought stress index. The cross Apagu 1B × 
Mouride had significantly positive SCA effect for 100−seed weight and seed yield as well as 
a desirable negative SCA effect for drought stress index. While Beledi C × Mouride showed 
desirable, significant positive SCA effect for grain yield, the number of pods per plant and 
number of seeds per pod (Table 5.9). Four crosses, Beledi A × IT93K−503−1, Apagu 1B × 
Dan lla, Laduni 1B × Mouride and Titinwa A × Mouride, had desirable negative SCA effect 
for seed yield, while Beledi C × IT93K−503−1 Beledi C × Dan lla, Titinwa A × Dan lla, 
Laduni 1B × IT93K−503−1 and Beledi A × Mouride showed no desired SCA effect for seed 
yield and some of the yield components measured (Table 5.9). 
94 
 
University of Ghana htt p://ugspace.ug.edu.gh
 
Table 5.8: Estimates of general combining ability effects of yield and yield components for both male and female 
                                   Well-Watered Condition                                              Drought Stress Condition 
Donor Parents  NPP NSP 100−SWT Seed Yield NDVI 30 NDVI 45 NDVI 60 NPP NSP 100−SWT Seed Yield 
Mouride  -0.67* 0.39 0.18 -4.45*** -1.34* -0.24 -1.74* -0.3 0.14 -0.11 -3.46*** 
IT93K−503−1 1.08* -0.43 -0.13          3.95** 3.06*** 0.60 1.36* 0.01 -0.35 0.13 1.06* 
Dan lla -0.42 0.05 -0.05     0.50 -1.73* -0.37 0.38 0.3 0.21 -0.01 2.4** 
                                  Well-Watered Condition                                                Drought Stress Condition 
      
Recurrent Parents     NPP       NSP        100−SWT     Seed Yield NDVI 30 NDVI 45 NDVI 60       NPP NSP   100−SWT    Seed Yield 
Apagu 1B 0.48 0.63 -1.51* 6.13** 0.04 2.06** 3.13** 0.50 1.38** -0.82* -0.33 
Laduni 1B 0.73* 0.52 0.24 6.13** 0.16 -1.57* 0.62 -0.33 0.81* -0.76* -2.66** 
Beledi A -1.85** -1 0.27 -12.12*** -1.60 -1.66* -1.60* -0.25 -0.93* -0.27 1.28* 
Beledi C -0.02 0.6 -0.15 11.13*** -1.21 -2.24** -2.92** -1.42** -0.47 0.61 -0.19 
Titinwa A 0.65* -0.75 1.15* -11.28*** 2.59 3.40*** 0.77 1.50** -0.79* 1.24** 1.9* 
*Significant at P < 0.05, **Significant at P < 0.01, ***Significant at P < 0.001, NPP: Number of pods per plant, NSP: Number of seeds per pod, 100−SWT: 100−seed weight and 
NDVI: Normalised difference vegetation index 
 
 
 
 
 
 
 
 
 
95 
 
University of Ghana htt p://ugspace.ug.edu.gh
 
Table 5.9: Estimates of specific combining ability effects for crosses under well-watered and drought stress condition  
           Drought-Stressed Condition                        Well-Watered Condition  
Population  Genotype Yield NPP NSP 100SDWT DS Yield NPP NSP 100WT 
 Apagu 1B × Dan lla -5.28*** -0.80 -0.74 0.40 0.53 -4.83** 0.17 -0.08 -0.13 
 Beledi A × Dan lla 3.56** 0.20 1.42* -0.35 0.20 -1.33 -0.25 -0.58 -1.18* 
 Beledi C × Dan lla -1.97 -1.13 -0.67 0.39 -0.13 -7.08*** -0.33 0.00 -0.34 
 Laduni 1B × Dan lla 2.00* 2.03** -0.54 0.61* -0.47 7.92*** -0.33 0.58 1.04* 
 Titinwa A × Dan lla 1.69 -0.30 0.53 -1.04** -0.13 5.33*** 0.75 0.07 0.61 
 Apagu 1B × IT93K−503−1 2.76* 0.75 0.59 -1.34** 0.53 -0.53 -0.83 0.00 -0.03 
F2 Beledi A × IT93K−503−1 -4.60** -0.75 0.05 0.91* -0.80* -2.28* 0.00 0.63 0.65 
 Beledi C × IT93K−503−1 -1.91 -0.08 -0.64 -1.08** -0.13 4.22** -0.33 0.36 0.54 
 Laduni 1B × IT93K−503−1 0.49 -1.17 0.36 0.62* -0.47 -13.28*** 1.17* -1.98* -1.35* 
 Titinwa A × IT93K−503−1 3.26** 1.25* -0.36 0.90* 0.87* 11.88*** 0.00 0.98 0.19 
 Apagu 1B × Mouride 2.52* 0.05 0.16 0.95* -1.07** 5.37*** 0.67 0.08 0.16 
 Beledi A × Mouride 1.04 0.55 -1.48* -0.55 0.60 3.62** 0.25 -0.06 0.53 
 Beledi C × Mouride 3.88** 1.22* 1.31* 0.69* 0.27 2.87* 0.67 -0.36 -0.20 
 Laduni 1B × Mouride -2.49* -0.87 0.18 -1.22** 0.93* 5.37*** -0.83 1.40* 0.31 
Titinwa A × Mouride -4.95** -0.95 -0.17 0.14 -0.73* -17.22*** -0.75 -1.06* -0.80 
Standard Error of the Mean 3.14 0.96 0.76 0.82 0.60 7.67 0.59 0.79 0.67 
*Significant at P < 0.05, **Significant at P < 0.01, ***Significant at P < 0.001, NPP: Number of pods per plant, NSP: Number of seeds per pod, 100SWT: 100−seed weight 
 
 
96 
 
 University of Ghana http://ugspace.ug.edu.gh
 
5.3.3 Computation of Earliness in Cowpea 
Mean square value of the 15 F2 population, 8 parents and 2 checks showed significant 
(P<0.01) differences among the genotypes for number of days to first flower, number of days 
to 50% flowering, number of days to first mature pod, number of days to 95% maturity, 
number of seeds per pod, pod length and 100−seed weight and chlorophyll content at 30 
DAP, 45 DAP and 60 DAP, whereas number of pods per plant showed no significant 
difference among the 25 entries (Table 5.10).  
The estimates of GCA effects of the parents revealed that Mouride, Apagu 1B, and Beledi C 
exhibited negative and significant values for number of days to the first flower, number of 
days to 50% flowering, number of days to first mature pod and number of days to 95% 
maturity. While Laduni 1B exhibits significant negative GCA effects for the number of days 
to first flower and number of days to 50% flowering, Titinwa A showed negative, but 
significant GCA values for number of days to first flower and number of days to 95% 
maturity (Table 5.11). 
Negative estimates of SCA effects for number of days to first flower was observed in nine of 
the fifteen crosses (Table 5.13). However, significant and negative SCA effects estimates for 
the trait is observed in Apagu 1B × Dan lla, Laduni 1B × Dan lla, Beledi A × IT93K−503−1, 
Beledi C × IT93K−503−1, Beledi A × Mouride, Laduni 1B × Mouride and Titinwa A × 
Mouride (Table 5.11). 
The estimates of negative SCA for number of days to 50% flowering, seven of the crosses 
indicated significant negative SCA. These hybrids in this category were Titinwa A × Dan lla, 
Apagu 1B × IT93K−503−1, Beledi A × IT93K−503−1, Beledi A × Mouride and Laduni 1B × 
Mouride. For number of days to 50% flowering, estimates of SCA was observed in eight of 
the fifteen combiners. Five of the eight hybrids that exhibited significant negative SCA were 
97 
 
 University of Ghana http://ugspace.ug.edu.gh
Beledi C × Dan lla, Titinwa A × Dan lla, Beledi C× IT93K−503−1, Apagu 1B × Mouride and 
Laduni 1B × Mouride (Table 5.11). 
Negative estimates of SCA effect for the number of days to 95% maturity were observed in 
seven of the fifteen crosses. Five out of the seven crosses exhibited significant negative SCA 
for this trait. The hybrids exhibiting significant negative SCA effects were Titinwa A × Dan 
lla, Beledi C × Dan lla, Apagu 1B × IT93K−503−1, Titinwa A × IT93K−503−1 and Laduni 
1B × Mouride (Table 5.11). 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
98 
 
 University of Ghana http://ugspace.ug.edu.gh
 
Table 5.10: Mean squares for 15 F2 population, eight parents and two checks measured under well-watered condition   
Source of Variation D.F Firstflo 50%flo FMP MT NPP NSP PL 100sdwt NDVI30 NDVI45 NDVI60 
  
Replication  3 2.73 3.81 1.93 5.37 5.16 4.22 6.19 19.20 11.98 55.20 65.53 
Genotype 24 22.73** 36.36** 49.63** 51.52** 7.76ns 6.32** 7.33** 10.35* 33.75** 94.85** 90.88** 
Error 72 6.54 7.58 11.00 10.66 5.37 2.64 1.89 4.58 13.60 19.89 31.41 
*Significant at P < 0.05, **Significant at P < 0.01, ns: non-significant, Firstflo, number of days to first flower, NPP, number of pods per plant, NSP, number of seeds per pod, 
FMP, first mature pod, MT, number of days to 95% maturity, 50%flo, number of days to 50% flowering, PL, pod length, 100sdwt, 100-seed weight and NDVI: SPAD 
chlorophyll meter reading at 30, 45 60DAP 
 
 
 
 
 
 
 
 
 
 
 
 
99 
 
 University of Ghana http://ugspace.ug.edu.gh
Table 5.11: Estimates of general and specific combining ability among eight parental lines and their F2 for the number of days to the first flower, 
days to 50 % flowering, days to the first mature pod and 95% maturity in cowpea  
Number of days to the Number of days to 50% Number of days to first mature Number of days to 95% 
Parent GCA first flower flowering pod maturity 
Dan lla 0.48* 0.43* 0.87* 0.35 
IT93K−503−1 -0.07 0.03 -0.18 -0.10 
Mouride -0.42* -0.47* -0.68* -0.25* 
Apagu1B -0.97* -1.15** -0.43* -0.77* 
Beledi A 2.62** 3.02** 2.82** 2.57** 
Beledi C -1.13** -1.48** -1.77* -1.35** 
Laduni1B -0.47* -0.40* 0.48* 0.23* 
Titinwa A -0.05 0.02 -1.10** -0.68** 
F2 Population SCA     
Apagu1 B × Dan lla -0.48* 0.15* 0.38* 0.57* 
Beledi A × Dan lla 1.93** 1.73** 0.13 0.23 
Beledi C × Dan lla -0.32 -0.52* -0.78* -2.10** 
Laduni 1B × Dan lla -0.98** -0.35* 0.72* 1.82* 
Titinwa A × Dan lla -0.15 -1.02** -0.45* -0.52* 
Apagu 1 B × IT93K−503−1 0.32 -0.45* -0.07 -0.73* 
Beledi A × IT93K−503−1 -1.27** -1.37** -0.07 -0.32 
Beledi C × IT93K−503−1 -0.77** -0.12 -0.48** 2.35** 
Laduni 1B × IT93K−503−1 1.57** 1.55** 0.52* 1.27* 
Titinwa A × IT93K−503−1 0.15 0.38* 0.10 -2.57** 
Apagu 1 B × Mouride 0.17 0.30* -0.32* 0.17* 
Beledi A × Mouride -0.67* -0.37* -0.07 0.08 
Beledi C × Mouride 1.08** 0.63* 1.27* -0.25 
Laduni 1B × Mouride  -0.58* -1.20* -1.23** -3.08** 
Titinwa A × Mouride  -0.43* 0.63* 0.35* 3.08** 
*Significant at P < 0.05, **Significant at P < 0.01, 
 
100 
 
 University of Ghana http://ugspace.ug.edu.gh
The means square for performance and variance for all the parental lines and their progenies 
were significantly different for number of days to the first flower, number of days to 50% 
flowering, number of days to first mature pod and number of days to 95% maturity. The 
broad sense heritability of these traits was equally high (Table 5.12). 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
101 
 
 University of Ghana http://ugspace.ug.edu.gh
 
Table 5.12: Mean square for performance and genetic components for days to the first flower, days to 50 % flowering, maturity in 15 crosses of 
cowpea 
Source of variation  D.F Days to the first flower Days to 50% flowering  Days to first mature Days to 95% maturity 
pod 
Rep 3 21.17 30.14 48.55 68.17 
Genotypes 14 21.19* 23.33* 23.52* 33.60* 
Male 2 7.62 9.39 7.45* 7.20 
Female 4 51.51*** 54.93** 61.17* 97.79* 
Male × female 8 9.42 11.01 8.70 8.10* 
Error 42 7.18 7.51 6.84 4.69 
Additive Variance (D)  7.49    7.84 9.43 9.94 
Dominance Variance (H)  2.24 3.51 1.87 6.67 
Environmental Variance (E)  1.79 1.88 1.71 2.61 
Narrow-sense (HN) (%)  65 59 72 52 
*Significant at P <0.05, **Significant at P < 0.01, and ***Significant at P < 0.001 
 
 
 
 
 
 
 
 
102 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
5.4 Discussion  
 
Climate influences the amount of water available in the soil profile for agricultural 
production. Thus, creation of genetic variability for traits under selection is crucial (Falconer 
and Mackay, 1996). The computed drought intensity index (DII) value of 0.64 was adequate 
to explain the degree of reduced cowpea yield and its components. Miklas et al. (2006) 
reported that drought stress is considered severe when the intensity index value exceeds 0.7. 
Thus, the computed DII in the present study of 0.64 is adequate to distinguish drought 
susceptible and tolerant genotypes.  
The crosses and parental lines were significantly different for traits studied under induced 
drought stress at the reproductive stage. This difference demonstrates the level of genetic 
variability among the genotypes for drought tolerance. The mean performance of yield and 
yield components of the hybrids were higher than most of the parental lines except for 
100−seed weight. This difference indicates segregation as a result of gene introgression. 
Marame et al. (2009) reported that complementary gene action influences performance as a 
result of transgressive segregation which suggests polygenic inheritance of gene for drought 
tolerance. The best breeding approaches for polygenic traits are backcross, pedigree and 
recurrent selection methods. These methods allow increase of frequency of the desirable 
alleles.  
Mean number of pods per plant, number of seeds per pod, 100−seed weight and seed yield 
were higher under well-watered conditions than the mean of parental lines and crosses under 
drought stress conditions, indicating that drought stress significantly reduces yield and yield 
components of the progenies and their parents. Both additive and non-additive effects 
contributed to the heritability of these traits.  
 
103 
 
University of Ghana http://ugspace.ug.edu.gh
 
Combining ability analysis identifies the breeding value of parental lines to produce hybrid 
progenies (Sprague and Tatum, 1942; Griffings, 1956). The results of this study revealed that 
both GCA and SCA effects were significant for DLS trait, indicating that both additive and 
non-additive gene actions were important and contributed to the expression of DLS. 
The GCA mean squares for the traits seed yield, 100−seed weight, number of seeds per pod 
and number of pods per plant were greater than SCA across the two watering regimes. This 
implies that additive genetic effects were more important than non-additive genetic effects in 
the inheritance of these traits. Chiulele (2010) reported a predominance of non-additive gene 
action over additive genetic effects for seed yield and some yield components including 
number of pods per plant, number of seeds per pod, 100-seed weight, DLS and NDVI which 
could be used for indirect selection in cowpea improvement program under drought stress 
conditions.   
The high value of Baker’s ratio and narrow sense coefficient of genetic determination for the 
traits seed yield, 100−seed weight, and number of pods per plant, implies that there could be 
an equally high probability of hybrid performance from the parental lines GCA effects for 
these traits. This ratio indicates the breeding value (additive genetic variance) acts as the 
main determinant of response to selection (Falconer and Mackay, 1996). These traits could be 
improved by simple selection methods such as mass and pedigree selection which agrees with 
the report of Chiulele (2010). Furthermore, the non-additive genetic effects due to dominance 
and/or epistasis indicated preponderance of SCA effects and its contribution in the total 
genetic variation in seed yield, 100−seed yield and number of pods per plant. The existence 
of non-additive gene action hampers the process of early generation selection (Olajide and 
Ilori, 2017) and provides the avenue for hybridisation. 
The BSCGD was high for all the traits across the contrasting environments. The high 
estimate of BSCGD could have been caused by lower environmental variance, interaction 
104 
 
University of Ghana http://ugspace.ug.edu.gh
 
between genotypes and environment or higher additive gene action. Acquaah (2012) reported 
that environmental variance, interaction between genotype and environment and level of 
additive gene action contributed to high score of BSCGD. From the results of this study, 
NSCGD was moderate to low (<40%) under drought stress conditions. Narrow sense 
coefficient of genetic determination is a powerful statistical tool that could help breeders in 
the estimation of a gene transmitted to progenies. Low NSCGD suggests the presence of 
dominant gene action. This finding agrees with the results reported by Abney et al. (2001) 
who found that both narrow and broad sense heritability of quantitative traits are fundamental 
in studying a founder population.  
However, selection of crosses with favourable estimates of SCA genetic effects should be 
prioritised and the cross combinations should involve at least one parental line that has shown 
favourable GCA effect. The hybrids Titinwa A × IT93K−503−1, Apagu 1B × Mouride 
Laduni 1B × Dan lla and Beledi C × Mouride showed desirable, positive and significant SCA 
effects for seed yield and its components. This implies that the developed populations 
performed better than the prediction based on their parental GCA effects.  The mean 
performance of Beledi A × Dan lla, Titinwa A × IT93K−503−1, and Beledi C × Mouride 
under drought stress conditions was higher than the mean seed yield of all the crosses. 
Falconer (1989) attributed the superiority of such crosses over their parents to complementary 
and duplicate gene action. This sets the precedence for the development of desirable 
segregants that could be utilised in cowpea improvement programs.     
The best combiner for seed yield and number of pods per plant was IT93K−503−1. Its 
outstanding performance could be attributed to high seed yield and number of pods across the 
contrasting environments as well as positive GCA effects. These results suggest that 
IT93K−503−1 could be the most desirable donor parent for seed yield improvement in water-
105 
 
University of Ghana http://ugspace.ug.edu.gh
 
limited environments. Hence, number of pods per plant could be used as an important index 
for indirect selection. This finding confirms the study carried out by Muchero et al. (2013).  
5.5 Conclusions  
 
Having adequate knowledge about the heritability of the numerous traits can help breeders 
design an efficient crop improvement program where several traits are to be improved 
concurrently. The study assessed the environmental responsiveness of 15 F2s and 8 parents 
under well-watered conditions for earliness in cowpea. Both additive and non-additive gene 
effects were important in governing early maturity in cowpea. Improving genotypes for 
earliness should consider both types of gene effects. The gene action conditioning genotype 
responsiveness and adaptation to drought stress environments are present in the selected 
donor parents. Both additive and non-additive effects controlled the transmission of genes 
responsible for drought tolerance in cowpea. However, non-additive (dominance) gene 
actions were more important than additive gene actions as shown by the outstanding 
performance of the inbreds in specific combinations. Thus, cowpea improvement for drought 
tolerance could be done by selecting progenies with desirable and positive SCA effects and 
subsequent advancement to later generations. The crosses Titinwa A × IT93K−503−1, Apagu 
1B × Mouride, Laduni 1B × Dan lla, and Beledi C × Mouride showed desirable positive 
significant SCA value for seed yield and its components. The parent IT93K−503−1 could be 
recommended as the most desirable donor for yield improvement in water-limited 
environments and as a tester for new hybrid development. The F2 progenies advanced from 
the crosses Beledi A × Dan lla, Titinwa A × IT93K−503−1, Apagu 1B × Mouride, Laduni 1B 
× Dan lla and Beledi C × Mouride, Apagu 1B × IT93K−503−1, Laduni 1B × IT93K−503−1, 
Beledi A × IT93K−503−1, Laduni 1B × Mouride were promising combinations with 
significant and positive SCA effects for yield and yield components. These combinations 
should be subjected to further selection that would eventually result in release of improved 
106 
 
University of Ghana http://ugspace.ug.edu.gh
 
drought-tolerant cowpea variety (varieties). However, progeny testing should be done on the 
F2 population, where segregation is expected to be maximum, before carrying out the final 
selection. Based on the analysis of general combining ability for the four traits, Beledi C is 
recommended as a desirable source of gene conditioning earliness traits in cowpea. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
107 
 
University of Ghana http://ugspace.ug.edu.gh
 
CHAPTER SIX 
 
6.0 YIELD AND YIELD STABILITY OF COWPEA GENOTYPES  
6.1 Introduction  
Traits that are of economic importance such as drought tolerance and yield are polygenic in 
nature, and they are sensitive to environmental differences. Most breeding programs conduct 
experiments in a number of contrasting sites to assess polygenic traits such as drought 
tolerance and grain yield. In those trials, variations in the performance of genotypes across 
diverse environments are consistently noticed. This is known as genotype by environment 
interaction (GEI), useful for improving quantitative traits (Bernardo, 2002). GEI makes it 
difficult to identify genotypes that are more stable and high yielding (Yan and Hunt, 1999). 
Analysis of GEI effects permits grouping of genotypes for their responsiveness in dissimilar 
conditions indicating whether they are stable or adaptable to broader or specific locations. 
Stability refers to the stability of performance exhibited by genotypes across locations. Yan 
and Kang (2003) grouped stability into two types, the dynamic (agronomic) and static 
(biological) stability. Plant breeders are mostly interested in agronomic stability. Agronomic 
stability requires that the genotypes are stable over testing sites (Yan and Kang, 2003; Frutos 
et al., 2014). Adaptability is an organism adjustment to its environment. A high yielding 
genotype in a specific agro-ecological condition could be a poor yielder in other 
environments (Casanoves et al., 2005). Numerous statistical methods are available to analyse 
GEI. Some of the frequently used methods are combined analysis of variance (ANOVA), 
multivariate and stability analyses.  
Combined ANOVA is one of the often-used statistical approaches in the identification of GEI 
in trials across locations. The drawback associated with combined ANOVA is the assumption 
of homogeneity of variance among environments, which is a core requirement for 
determining genotype dissimilarities. However, this type of analysis permits determination of 
the variances of those constituents from diverse factors: genotype, environment, and GEI, but 
108 
 
University of Ghana http://ugspace.ug.edu.gh
 
it does not allow breeders to see the non-additive responses of the genotypes evaluated (Zobel 
et al., 1988; Gauch et al., 2008).  
Analysis of genotype stability offers an opportunity for assessing the environmental 
responsiveness of the genotypes to changes in the environment. Cox (1997) suggested an 
analysis of linear regression as a statistical approach for gauging stability. The analytical 
approaches proposed and revised by Finlay and Wilkinson (1963), Eberhart and Russell 
(1966), Francis and Kannenberg (1978), Lin (1985), and Crossa (1990) have been widely 
used.  
Genotypic stability analysis encompasses regression of the mean of genotypes on an 
environmental index (mean seed yield of all the entries evaluated across locations) known as 
stability index. Core biological drawback shows up when only a handful of genotypes with 
very high and low yields from different locations are involved in the stability analysis. 
Another problem with this approach is the hypothesis of a linear association between 
environmental means and GEI while the actual genotypic responses to the locations are 
multi-variate (Crossa, 1990). There are three main reasons for carrying out multivariate 
analysis when GEI effect is the core interest: (a) removal of "noise" in a data set 
(distinguishing systematic and non-systematic variation); (b) information summary; and (c) to 
reveal association in the data set (Crossa, 1990; and Gauch, 1992).  The multivariate analysis 
models are built on PCA, such as Site regression (SREG) and Additive main effects and 
multiplicative interaction (AMMI). There are both linear and bilinear models with the main 
effect of the environment or genotypes (additive component) and GEI (multiplicative 
component). 
AMMI analysis is an amalgamation of ANOVA, the core for analysing genotypes and 
environmental effects, and PCA of GEI (Zobel et al., 1988; Gauch et al., 2008). AMMI 
models depend on the number of principal components (PCs) used in the interaction between 
genotype and environment known as AMMI (1), AMMI (2) …AMMI (n) (Hongyu et al., 
109 
 
University of Ghana http://ugspace.ug.edu.gh
 
2014). This results in biplots a graphical representation of the interaction in biplot (Gabriel, 
1971) which allows (1) display of the interaction of environments (vectors) and genotypes 
(points) in the same graph, and (2) examination of GEI effects. The angles between the 
environments and genotypes represented in the biplot graph gives the level of interaction 
between genotype and environment. The distances from the origin of the graph to the average 
environment axis (AEC) indicate the level of interaction that the genotypes display across 
locations.  
Site regression analysis, also known as genotype main effect and genotype and GEI analysis, 
is both a linear and bilinear model that eliminates the effect of site and states the response 
simply as a function of the effect of genotypes and the GEI (Yang et al., 2010; and Crossa et 
al., 2015). This model is commended when the locations are the main cause of variation and 
it contributes to the genotypes and GEI effect taking into consideration the total disproportion 
of both environmental and genotypic contributions to the total variations (Casanoves et al., 
2005). Moreover, GGE is quite different from the AMMI model. The GGE analysis considers 
the GEI effects as crossover effects arising from pronounced variations in the performance of 
genotypes across locations (Yan et al., 2000).  
According to Yan et al. (2000), the use of GGE biplot graphical analysis has helped 
researchers to understand the behaviour and interactions of genotypes without the effect of 
the environment. Furthermore, they postulated that the first principal component (PC 1) 
usually accounts for responses of the genotypes that are comparative to the locations linked 
with the GEI without alteration of the assortment. Whereas, the second principal component 
(PC 2) offers evidence about cultivation sites that are not proportionate to the locations, 
showing attributes that are accountable to the G×E crossover interaction. The GGE biplot 
also permits identification and grouping of the environment into mega-environments, which 
entails those portions of the testing site of genotypes that show homogeneous agro-ecological 
conditions, where performances of some genotypes are consistent over the periods of 
110 
 
University of Ghana http://ugspace.ug.edu.gh
 
evaluation (Gauch and Zobel, 1988). In each designated mega-environment, genotype by 
location interaction effects are inadequate or insignificant (Yan and Hunt, 2000).  
The objectives of this study were, therefore, to 1) assess GEI effects using GGE and AMMI 
approaches for yield and yield components of drought-tolerant early maturing cowpea 
populations, and 2) determine the yield stability for seed yield of 25 cowpea genotypes grown 
under drought-stressed and well-watered conditions during off-seasons in three contrasting 
locations.  
6.2 Materials and Methods  
 
6.2.1 Population Development  
 
The genetic material used in the development of the genotypes for this study included five 
farmers' preferred varieties, Titinwa A, Laduni 1B, Beledi A, Apagu 1B, and Beledi C, and 
three drought-tolerant varieties, Dan lla, Mouride, and IT93K−503−1. The five early 
maturing farmers’ preferred varieties were crossed with each drought-tolerant material in a 5 
x 3 North Carolina II mating design to generate 15 early droughts tolerant first filial (F1s) 
generation. Beledi C, Laduni 1B, Beledi A, Apagu 1B and Titinwa A (recurrent parents) and 
three other varieties, Dan lla, Mouride, and IT93K−503−1 (donor parents), were crossed in 
September 2016 at WACCI farm. The 15 F1s were backcrossed to their donor parents three 
times. The crossing block was set up on the 5th and 13th September 2016 to synchronize the 
flowering pattern of the genetic materials. Each of the breeding materials (recurrent/ donor 
parent) was planted in a three-row plot of 4 m long with an inter-row spacing of 1 m and 
intra-row spacing of 1 m between hills. Two seeds were planted per hill and each plot had a 
total population density of 30 plants. Each male parent was planted in one row of 20 plants 
along with two rows of the female parent. The male parents were planted seven days earlier 
than the female parents to flower about the same time (nicking). Standard cultural practices 
of pesticide, and weeding were applied to all the plots to keep the field free of weeds and 
pests. Necessary precautions were taken to avoid the contamination of the parental lines at 
111 
 
University of Ghana http://ugspace.ug.edu.gh
 
the time of crossing. Hand emasculation of the receptive floral buds and pollination were 
carried out by skilled workers under the management of the researcher according to IITA 
guide.  Pollination commenced on 13th October 2016 and lasted for four weeks. After 
emasculation (cutting half of the upper petal of the female parent and removal of stamens), 
pollen was collected from a freshly opened male parent flowers and dusted on the stigma of 
the female parent. The crosses were harvested on 24th of November 2016.   
Backcross breeding method was employed to recover the recurrent parent's performance with 
an added drought tolerance attribute. Fifteen F1s and five recurrent parents were planted on 
the 22nd November 2016. Emasculation and pollination started on the 26th December 2016 for 
four weeks and harvest of the backcross one F1s (BC1F1s) was on the 17
th of January 2017. 
Subsequently, the BC1F1s seeds and donor (female parents) were plant on 18
th March 2017 to 
develop BC2F1s. Crossing started on the 24
th April 2017 for four weeks and BC2F1s seeds 
were harvested from a single plant on the 17 May 2017. The crossing block for developing 
BC3F1s was set up on 29
th and 30th June 2017. Crossing commenced on the 10th of August 
2017 for four weeks (9th of September 2017).  BC F th3 1s seeds were harvested on the 6  of 
October 2017. The BC3F1s were selfed twice from 16
th September 2017 to 10th January 2018 
to generate BC3F3 and ensure purity.  
6.2.2 Genetic Materials and Environment of the Test Locations 
 
The 15 BC3F3s, their eight parents and two checks (AGRAC−216 and Asontem) were 
evaluated during the offseason of 2018 in three testing sites: Coastal Savanna Agro-
ecological Zone (WACCI farm/Legon, Accra), Guinea Savanna Agro-ecological Zone 
(SARI/Nyankpala, Tamale), and Forest-Savanna Agro-ecological Zone (CSIR-Crops 
Research Institute CRI/, Fumesua, Kumasi) (Table 6.2).   
 
 
112 
 
University of Ghana http://ugspace.ug.edu.gh
 
6.2.3 Experimental Layout  
 
The genetic materials consisting of 25 entries arranged in 5 × 5 lattice square design with 
three replications were evaluated across three test locations (Nyankpala, Fumesua, and 
Legon) in Ghana. The trial was planted in two adjustment blocks spaced 50 m apart denoting 
two water regimes. The first block (Block 1) was designated as well-watered (WW) and the 
second (Block 2) as severe stress (SS). Plants planted in Block 2 were subjected to severe 
water stress-imposed by withdrawing irrigation water for three weeks at flowering stage. 
Then watering was resumed for two weeks, twice a week, in order to score recovery and 
regrowth parameters of the genotypes. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
  
113 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
Table 6.1: List of 15 BC3F3 cowpea genotypes, eight parents and two checks evaluated in three locations in 2018 
      Pedigree    
Genotype Female Male  Attribute Source 
BC×M Beledi C Mouride Developed for drought tolerance and earliness  Breeders material  
BC×D Beledi C Dan lla Developed for drought tolerance and earliness  Breeders material  
BC×IT Beledi C IT93K−503−1 Developed for drought tolerance and earliness  Breeders material 
BA×D Beledi A Dan lla Developed for drought tolerance and earliness  Breeders material 
BA×IT Beledi A IT93K−503−1 Developed for drought tolerance and earliness  Breeders material 
BA×M Beledi A Mouride Developed for drought tolerance and earliness  Breeders material 
TA×M Titinwa A Mouride Developed for drought tolerance and earliness  Breeders material 
TA×IT Titinwa A IT93K−503−1 Developed for drought tolerance and earliness  Breeders material 
TA×D Titinwa A Dan lla Developed for drought tolerance and earliness  Breeders material 
A1B×M Apagu 1B Mouride Developed for drought tolerance and earliness  Breeders material 
A1B×D Apagu 1B Dan lla Developed for drought tolerance and earliness  Breeders material 
A1B×I Apagu 1B IT93K−503−1 Developed for drought tolerance and earliness  Breeders material 
L1B×I Laduni 1B IT93K−503−1 Developed for drought tolerance and earliness  Breeders material 
L1B×D Laduni 1B Dan lla Developed for drought tolerance and earliness  Breeders material 
L1B×M Laduni 1B Mouride  Developed for drought tolerance and earliness  Breeders material 
Mouride   Drought tolerant Burkina Faso 
Dan lla   Drought tolerant Niger 
IT93K−503−1   Drought tolerant Nigeria  
Laduni 1B   Preferred and adaptable  South Sudan 
Titinwa A   Preferred and adaptable  South Sudan 
Apagu 1B   Preferred and adaptable  South Sudan 
Beledi C   Preferred and adaptable  South Sudan 
Beledi A   Preferred and adaptable  South Sudan 
AGRAC−216   High yielding (check) South Sudan  
Asontem  Early maturing (check) Ghana 
  
 
 
114 
 
University of Ghana http://ugspace.ug.edu.gh
 
6.2.4 Data Collection  
 
Data were collected on an individual plant basis in accordance with the International Plant 
Genetic Resources cowpea descriptors (1983). 
i. number of days to 50% flowering was recorded as the number of days from planting 
to a stage when 50% of plants in a plot had begun to flower, and 
ii.  number of days to 95% maturity was scored when 95% of the plants in a plot has 
attained maturity.  
At physiological maturity, individual plants were harvested separately, then the mean of each 
trait measured on individual plant was computed to determine yield and yield components:  
iii. number of pods per plant recorded as the mean of matured pods from 20 plants,  
iv. number of seeds per pod recorded an average number of seeds from 20 pods, 
whereas, 
v. pod length was measured in centimetre as the mean of 20 pod length.  
vi. 100−seed weight was computed from 10 randomly selected plants in each plot 
weighed and recorded as the weight of 100−seed in grams, and  
vii. seed yield harvested at 12% moisture content was computed from the weight of 
harvested pods per plot (kgha−1), and  
viii. harvest index was computed as ratio of seed yield to biological yield. 
115 
 
University of Ghana http://ugspace.ug.edu.gh
 
Table 6.2: Description of test sites for 25 cowpea genotypes evaluated across three diverse agro-ecologies in Ghana 
Agro-ecological Altitude Average Rainfall Temperature 
Environment Soil Type Geographic Location DII STI TOL 
Zone (masl) (mm) (°C) 
          Min Max Latitude Longitude   
Legon  Coastal Savanna  Adenta series 97 809 23.8 31.2 5º 38' N 0º 10 'E 0.31 0.69 675.47 
Fumesua  Forest Savanna  Fumesua series 286 1500 22.4 32.1 6°41′ N  1º 28′ W 0.29 0.39 523.10 
Nyankpala Guinea Savanna  Tolon series 171 1091 23.9 35.3 9°24′ N 0°59′ W 0.61 0.39 612.70 
Source: Climate-data.org, STI: Stress tolerance index, TOL: Drought tolerance index and DII: Drought intensity index 
 
 
 
 
 
 
 
 
 
 
 
116 
 
University of Ghana http://ugspace.ug.edu.gh
 
6.2.5 Data Analysis 
 
The data were analysed by using GenStat 18th edition, to compute mean performance and 
mean separation. IBM SPSS 22nd edition was used to generate the regression graphics and 
GEA-R for stability analysis using AMMI, and GGE biplot analysis models, respectively.    
Selection Index  
Harvest index was computed using Donald and Hamblin (1976) formula  
                                              Grain Yield 
Harvest index (%) =                                            × 100  
                                            Biological Yield 
 
The percent reduction due to moisture stress and drought susceptibility index was computed 
using the formula suggested by Fischer and Maurer (1978). 
 
Yield under non-stress –Yield under stress 
Percent reduction =                                                                                     × 100  
Yield under non-stress 
 
(1−Yd /Yp) 
Drought susceptibility index (DSI) =      
Ȳp   
 
                                             Mean grain yield of all genotypes under drought stress condition 
Drought index (DII) = 1−    
                                             Mean grain yield of all genotypes under well-watered condition  
 
Where, 
 
Yd= Grain yield of genotypes under severe moisture stress condition. 
 
Yp= Grain yield of genotypes under well-irrigated condition. 
 
 
 
 
 
117 
 
University of Ghana http://ugspace.ug.edu.gh
 
6.3 Results  
 
The environmental conditions of the three testing sites coupled with induced drought stress 
were significant enough to show the degree of drought severity on seed yield as illustrated in 
Fig. 6.1.  
1600.00
1355.70
1400.00
1200.00
969.78
1000.00
800.00
600.00
400.00
200.00
0.00
SY (kg/ha)
Well-Watered Drought-Stressed
 
 
Fig. 6. 1 Estimated means for the twenty-five genotypes selected under drought stress and well-watered    
conditions at (P<0.05) 
 
The computed drought intensity index (DII) revealed that Legon, Tamale, and Kumasi were 
significantly (P<0.05) different. Hence, this index could be used for selecting genotypes for 
drier environments by comparing their DII with performance under both well-watered and 
drought stress conditions (Table 6.2).  
As shown in Fig. 6.2, the comparison of genotypes in terms of seed yield under drought stress 
condition at Legon, Fumesua and Nyankpala revealed that the genotypes generally gave a 
higher yield in Fumesua than in Legon and Nyankpala. This was supported by the numerical 
measures of seed yield. The median seed yield in Nyankpala (300 kgha−1) was lower than 
both Legon (800 kgha−1) and Fumesua (1200 kgha−1). It was observed that the third quartile 
of Nyankpala (600 kgha−1) was lower than the seed yield median for Legon.   
118 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
Fig. 6. 2 Boxplot estimated means for the significant effects under drought stress (DS) conditions at (P<0.05) 
 
6.3.1 Crop Performance at Nyankpala  
The overall drought tolerance index (DII) at Nyankpala (Tamale) was 0.61, stress tolerance 
index (STI) 0.39 and stress tolerance (TOL) was 612.7 (Table 6.2). Genotype Dan lla, a 
donor parent, had the highest yield of 849.1 kgha−1 under drought stress condition (Table 
6.3). BC×M the second-best performer yielded 839.9 kgha−1 under drought stress condition. 
This genotype had stress tolerance index (STI) of 0.5, mean productivity index (AMP) 
1191.2, and geometric mean productivity (GMP) of 1138.2. Additionally, BC×M had TOL of 
702.5 and DII of 0.46. BA×D had the lowest yield of 129.3 kgha−1 among the families despite 
having STI value of 0.1, an AMP (589.2), GMP was 368.3 and TOL was 919.8 with DII of 
0.88 (Table 6.3).  The computed mean squares at Nyankpala showed that under drought stress 
condition, Dan lla had the heaviest 100−seed weight (16.1 g) and BA×D registered the 
lightest (5.4 g) 100−seed weight under drought stress condition (Table 6.4). Whereas, under 
well-watered condition, genotype Mouride, a non-recurrent parent, had the heaviest 100−seed 
119 
 
University of Ghana http://ugspace.ug.edu.gh
 
weight (17.4 g), while BC×D had the lightest seed weight (7.5 g) under well-watered 
condition (Table 6.4). 
Asontem, a check in this study, had the highest number of seeds per pod (12.4) and TA×M 
the lowest (7.5) under the well-watered conditions. IT93K−503−1 registered the highest seed 
count of 8 seeds per pod, Whereas TA×I and TA×M both had the lowest number of seeds per 
pods (4.1) under drought stressed conditions. Beledi C had the highest number of pods per 
plant (14.6) and AGRAC−216 had the lowest (6.5) pods per plant under well-watered 
conditions at Nyankpala.  Laduni 1B registered the highest value of 7.3 pods and BA×M had 
the lowest number of pods per plant (0.8) under drought stressed environment (Table 6.4). 
BA−I registered the longest number of days to reach 95% (91 days) and Laduni 1B had the 
shortest (73.8) days to attain 95% maturity under drought stressed condition. AGRAC−216 
took the longest number of days (80.0) to reach 95% maturity under a well-watered 
condition, while Laduni 1B had the shortest number of days (69.0) to attained 95% maturity 
under well-watered condition. The number of days to attain 50% flowering was another trait 
for computing earliness in cowpea. AGRAC−216 had the longest number of days to 50% 
flowering (58 days), whereas, Laduni 1B had the shortest (43 days) under drought stress 
condition. Mouride, took 56 days to 50% flowering and Titinwa A, a recurrent parent, had the 
shortest number of days (40 days) to attain 50% flowering under well-watered conditions. 
Under drought stress conditions, BC×M had the highest harvest index (HI) of 55.6% and 
AGRAC−216 registered the lowest value of 0.2. Titinwa A had the highest percentage of 
85.1% for HI and L1B×D had the lowest value of 9.2% under well-watered conditions (Table 
6.4). 
 
 
 
120 
 
University of Ghana http://ugspace.ug.edu.gh
 
Table 6.3: Seed yield performance of 25 genotypes and other agronomic traits measured under drought stressed and well-watered conditions at 
Nyankpala 
Attribute Genotype WW DS STI AMP GMP TOL DII 
High Yielding Ability Dan lla 1004.4 849.1 0.8 926.8 923.5 155.3 0.15 
 BC×M 1542.4 839.9 0.5 1191.2 1138.2 702.5 0.46 
Medium Yielding Ability BA×M 604.8 588.3 1.0 596.6 596.5 16.5 0.03 
 TA×I 695.4 567.8 0.8 631.6 628.4 127.6 0.18 
 BC×I 1121.7 545.3 0.5 833.5 782.1 576.4 0.51 
 A1B×M 898.9 513.1 0.6 706.0 679.1 385.8 0.43 
 A1B×I 881.8 472.5 0.5 677.2 645.5 409.3 0.46 
 IT93K−503−1 1425.1 459.7 0.3 942.4 809.4 965.4 0.68 
 TA×D 846.4 448.2 0.5 647.3 615.9 398.2 0.47 
 Mouride 1700.4 412.6 0.2 1056.5 837.6 1287.8 0.76 
 A1B×D 950.9 397.6 0.4 674.3 614.9 553.3 0.58 
 L1B×I 1052.9 394.3 0.4 723.6 644.3 658.6 0.63 
Low Yielding Ability L1B×M 688.8 361.1 0.5 525.0 498.7 327.7 0.48 
 Asontem 989.3 359.6 0.4 674.5 596.4 629.7 0.64 
 BC×D 780.5 351.7 0.5 566.1 523.9 428.8 0.55 
 BA×I 1052 309.1 0.3 680.6 570.2 742.9 0.71 
 TA×M 824.4 284.5 0.3 554.5 484.3 539.9 0.65 
 Laduni 1B 1161.7 260.3 0.2 711.0 549.9 901.4 0.78 
 Beledi C 800.3 246.4 0.3 523.4 444.1 553.9 0.69 
 Apagu 1B 630.2 227.1 0.4 428.7 378.3 403.1 0.64 
 L1B×D 1288.4 212.1 0.2 750.3 522.8 1076.3 0.84 
 Titinwa A 1352 201.5 0.1 776.8 521.9 1150.5 0.85 
 Beledi A 996.5 163.3 0.2 579.9 403.4 833.2 0.84 
 BA×D 1049.1 129.3 0.1 589.2 368.3 919.8 0.88 
  AGRAC−216 641 67 0.1 354.0 207.2 574 0.90 
Mean  999.2 386.5       
CV%  39.80 62.20      
Significance    ns *           
WW: well-watered condition, DS: Drought-stressed condition, STI: Stress tolerance index, AMP: Arithmetic means productivity, GMP: Geometric means productivity, TOL: 
Drought tolerance index, DII: Drought intensity index 
 
 
121 
 
University of Ghana http://ugspace.ug.edu.gh
 
Table 6.4: Means from analysis of variance measured under drought stress and well-watered conditions at Nyankpala 2018 
Env 1 (Nyankpala Well-Watered) Env 4 (Nyankpala Drought-Stressed) 
Genotype HI SY (kg/ha) 100swt (g) NSP NPP 95%mat 50%flo HI SY (kg/ha) 100swt (g) NSP NPP 95%mat 50%flo 
A1B×D 38.2 950.9 8.2 9.8 11.2 70.0 44.0. 38.9 397.6 7.5 5.5 3.5 76.0 49.0 
A1B×I 68.7 881.8 8.8 10.9 10.6 70.0 46.0 44.7 472.5 8.0 5.2 5.2 81.0 48.0 
A1B×M 21.5 898.9 12.3 9.3 9.9 71.0 47.0 27.8 513.1 10.4 5.3 2.6 82.0 44.0 
AGRAC−216 11.5 641.0 14.5 8.0 6.5 80.0 53.0 0.2 67.0 10.8 5.4 3.2 81.0 58.0 
Apagu 1B 47.9 630.2 8.7 11.0 7.7 70.0 46.0 26.3 227.1 8.4 6.1 5.8 75.0 46.0 
Asontem 41.3 989.3 12.3 12.4 7.3 73.0 50.0 20.4 359.6 12.1 6.4 5.4 81.0 53.0 
BA×D 20.3 1049.1 9.1 9.1 9.5 77.0 49.0 1.8 129.3 5.9 7.0 4.3 87.0 50.0 
BA×I 11.4 1052.0 11.1 8.5 7.9 76.0 45.0 13.0 309.1 11.1 5.3 1.9 91.0 54.0 
BA×M 18.1 604.8 10.7 7.7 7.9 76.0 47.0 20.8 588.3 9.4 4.6 0.8 88.0 53.0 
BC×D 69.5 780.5 7.5 10.7 10.8 72.0 44.0 14.3 351.7 8.4 5.1 2.6 82.0 45.0 
BC×I 57.0 1121.7 11.8 9.9 11.2 72.0 42.0 28.4 545.3 8.5 4.5 4.0 82.0 45.0 
BC×M 52.9 1542.4 10.7 10.1 13.3 72.0 43.0 55.6 839.9 9.4 6.0 2.2 82.0 45.0 
Beledi A 20.9 996.5 8.8 10.5 12.6 77.0 47.0 4.3 163.3 9.6 6.5 3.3 83.0 49.0 
Beledi C 43.6 800.3 7.8 12.1 14.6 73.0 42.0 14.4 246.4 8.0 6.6 4.0 85.0 45.0 
Dan lla 29.3 1004.4 16.7 10.8 10.8 77.0 54.0 49.6 849.1 16.1 7.6 6.5 65.0 55.0 
IT93K−503−1 69.3 1425.1 14.3 11.1 11.1 74.0 48.0 12.8 459.7 13.4 8.0 4.8 88.0 51.0 
L1B×D 9.2 1288.4 12.3 9.0 8.3 72.0 42.0 13.6 212.1 9.9 5.2 3.8 76.0 44.0 
L1B×I 51.1 1052.9 10.9 8.4 7.9 71.0 43.0 42.2 394.3 8.5 5.1 2.8 76.0 43.0 
L1B×M 27.1 688.8 12.0 8.8 7.9 72.0 43.0 14.6 361.1 11.4 4.8 3.5 76.0 45.0 
Laduni 1B 52.7 1161.7 11.2 12.2 10.6 69.0 42.0 12.6 260.3 11.5 5.3 7.3 74.0 43.0 
Mouride 44.7 1700.4 17.4 9.1 10.8 77.0 56.0 15.3 412.6 16.0 6.0 4.1 87.0 53.0 
TA×D 44.4 846.4 12.4 8.2 9.6 76.0 45.0 25.4 448.2 9.2 4.9 4.3 84.0 46.0 
TA×I 38.4 695.4 9.5 8.1 10.5 72.0 42.0 36.5 567.8 10.1 4.1 3.2 82.0 44.0 
TA×M 51.3 824.4 13.5 7.5 9.5 71.0 42.0 13.9 284.5 13.8 4.1 3.0 83.0 44.0 
Titinwa A 85.1 1352.0 12.7 9.9 13.6 74.0 40.0 13.6 201.5 10.1 4.5 3.6 83.0 44.0 
Mean 41.0 999.2 11.4 9.7 10.1 73.0 46.0 22.4 386.5 10.3 5.6 3.8 81.0 48.0 
CV% 64.9 39.8 11.4 13.5 26.4 2.9 6.4 72.2 62.2 22.2 28.9 43.9 12.6 6.4 
Lsd p<0.05 45.5 680.4 2.2 2.2 4.5 3.6 5.0 27.8 411.4 3.9 2.7 2.9 17.5 5.2 
Hi: Harvest index, SY: Seed yield, 100swt: 100-seed weight, NPP: Number of pods per plant, NSP: Number of seeds per pod, 95%mat: 95% maturity and 50%flo: 50% 
flowering 
 
 
122 
 
University of Ghana http://ugspace.ug.edu.gh
 
6.3.2 Crop Performance at Legon  
The overall DII of the test site Legon (Accra) was 0.31, STI was 0.69 and TOL was 675.47 
(Table 6.2). At Legon, genotype AGRAC−216 had the highest 100−seed weight of 21.9 g 
and 21.5 g under both drought-stressed and well-watered conditions, respectively. A1B×D 
had the lowest 100−seed weight (7.4 g) under drought stress and Beledi C had the lowest 
100−seed weight (8.9 g) under well-watered conditions (Table 6.5). 
Under drought stress condition, Mouride, a non-recurrent parent, yielded 1375.1 kgha−1 
followed by BC×M with the second highest seed yield producing 1365.3 kgha−1. Genotype 
Beledi A had the lowest seed yield (375.2 kgha−1) under drought stress condition. Whereas, 
under well-watered condition, IT93K−503−1 yielded 2559.0 kgha−1 and BC×D had the 
lowest seed yield of 796.0 kgha−1 (Table 6.4). Among the progenies evaluated, BC×M had 
STI of 1.3, AMP of 1203.3 and GMP of 1192.1. This genotype also had TOL of -324.2 and 
DII of -0.3.  BC×M had the highest seed yield (1365.2 kgha−1) while BA×D had the lowest 
seed yield (466.2 kgha−1) under drought stress condition. However, BA×D had STI value of 
0.3, AMP (10221.1), GMP was 857.7 and TOL of 1111.8 with DII of 0.77. Parent Beledi A 
had the worst performance under both conditions producing seed yield below the overall 
mean despite having drought tolerance index of 0.5 (Table 6.5).  
Laduni 1B had 16 seeds per pod under stressed condition and TA×I had the lowest number of 
seeds per pods (10.8) under drought-stress condition. Nevertheless, L1B×I had the highest 
number of seeds per pod (16.6) and TA×I had the lowest number of seeds per pod (10.4) 
under well-watered condition. Beledi C had the highest number of pods per plant, while 
Beledi A had the lowest number of pods per plant and under drought stressed condition. 
IT93K−503−1, a non-recurrent parent, had the highest number of pods per plant and Asontem 
had the lowest number of pods per plant under well-watered condition (Table 6.6). 
123 
 
University of Ghana http://ugspace.ug.edu.gh
 
One of the earliness traits, number of days to 95% maturity, revealed that Dan lla took the 
longest number of days to attain 95% maturity (72 days) and A1B×I registered 51 days for 
the trait under drought stress conditions. Under well-watered conditions, Dan lla also took 71 
days to attain 95% maturity and progeny TA×M took only 54.0 days to reach 95% maturity 
under optimum conditions. Dan lla took 49.0 and 49.0 days to reach 50% flowering under 
both drought-stressed and well-watered conditions, respectively. A1B×M had the shortest 
number of days to attain 50% flowering (35 and 34 days) under both drought-stressed and 
well-watered conditions. Under drought stress conditions, A1B×I had the highest HI, while 
Beledi A scored the lowest HI value. Asontem had the highest HI value and Mouride 
registered the lowest value under well-watered conditions (Table 6.6). 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
124 
 
University of Ghana http://ugspace.ug.edu.gh
 
Table 6.5: Seed yield performance of 25 genotypes and other agronomic traits measured under drought stressed and well-watered conditions at 
Legon 
Attribute Genotype WW DS STI AMP GMP TOL DII 
High Yielding Ability L1B×M 1814.0 1064.2 0.6 1439.1 1389.4 749.8 0.4 
 Beledi C 778.0 1283.7 1.7 1030.9 999.4 -505.7 -0.7 
 L1B×D 1368.0 959.1 0.7 1163.6 1145.4 408.9 0.3 
 Titinwa A 1357.0 985.0 0.7 1171.0 1156.1 372.0 0.3 
 TA×M 1344.0 1234.3 0.9 1289.2 1288.0 109.7 0.1 
 L1B×I 1312.0 1115.7 0.9 1213.9 1209.9 196.3 0.1 
 AGRAC−216 1265.0 989.0 0.8 1127.0 1118.5 276.0 0.2 
 Dan lla 1244.0 1096.8 0.9 1170.4 1168.1 147.2 0.1 
 A1B×M 1232.0 1163.7 0.9 1197.9 1197.4 68.3 0.1 
 A1B×I 1065.0 1039.2 1.0 1052.1 1052.0 25.8 0.0 
 BC×I 1064.0 1012.0 1.0 1038.0 1037.7 52.0 0.0 
 BC×M 1041.0 1365.2 1.3 1203.1 1192.1 -324.2 -0.3 
 Mouride 1022.0 1375.7 1.3 1198.9 1185.7 -353.7 -0.3 
 A1B×D 2222.0 897.5 0.4 1559.8 1412.2 1324.5 0.6 
Medium Yielding Ability Laduni 1B 1309.0 854.2 0.7 1081.6 1057.4 454.8 0.3 
 Apagu 1B 1164.0 839.7 0.7 1001.9 988.6 324.3 0.3 
 TA×I 885.0 610.7 0.7 747.9 735.2 274.3 0.3 
 TA×D 1311.0 657.8 0.5 984.4 928.6 653.2 0.5 
 Asontem 904.0 625.0 0.7 764.5 751.7 279.0 0.3 
Low Yielding Ability IT93K−503−1 2559.0 453.7 0.2 1506.4 1077.5 2105.3 0.8 
 BA×D 1578.0 466.2 0.3 1022.1 857.7 1111.8 0.7 
 BA×M 1313.0 534.3 0.4 923.7 837.6 778.7 0.6 
 BA×I 1335.0 567.1 0.4 951.1 870.1 767.9 0.6 
 Beledi A 802.0 375.2 0.5 588.6 548.6 426.8 0.5 
 BC×D 796.0 518.2 0.7 657.1 642.3 277.8 0.3 
Mean   1283.4 883.3           
CV%  33.70 33.43      
Significance     ***  ***           
WW: well-watered condition, DS: Drought-stressed condition, STI: Stress tolerance index, AMP: Arithmetic means productivity, GMP: Geometric means productivity, TOL: 
Drought tolerance index, DII: Drought intensity index 
 
 
 
 
125 
 
University of Ghana http://ugspace.ug.edu.gh
 
Table 6. 6: Means from analysis of variance measured under drought stress and well-watered conditions at Legon 2018 
Env 3 (Legon Well-Watered) Env 6 (Legon Drought-Stressed) 
Genotype HI SY (kg/ha) 100swt (g) NSP NPP 95%mat 50%flo HI SY (kg/ha) 100swt (g) NSP NPP 95%mat 50%flo 
A1B×D 22.7 2222.0 9.6 12.7 16.5 56.0 38.0 28.7 976 7.4 13.3 11.0 56.0 39.0 
A1B×I 17.4 1065.0 9.4 13.4 14.6 56.0 38.0 30.3 1028 9.0 14.9 9.3 51.0 38.0 
A1B×M 8.6 1232.0 13.3 12.9 16.8 56.0 34.0 17.0 1301 12.8 13.2 8.3 530 35.0 
AGRAC−216 9.1 1265.0 21.5 11.9 11.6 59.0 45.0 8.6 1847 21.9 11.9 4.8 57.0 44.0 
Apagu 1B 15.2 1164.0 8.6 14.3 20.7 56.0 38.0 25.8 1052 7.5 13.9 9.9 54.0 38.0 
Asontem 27.5 904.0 15.2 15.1 6.9 58.0 45.0 5.5 1648 14.9 11.7 4.3 56.0 44.0 
BA×D 13.9 1578.0 9.2 14.6 20.6 60.0 39.0 6.2 1279 10.4 11.7 4.5 59.0 40.0 
BA×I 9.1 1335.0 12.0 14.8 15.4 55.0 40.0 5.2 1274 11.3 14.0 4.4 59.0 39.0 
BA×M 8.3 1313.0 11.0 13.4 21.6 57.0 37.0 3.9 1171 10.2 13.2 4.9 61.0 38.0 
BC×D 9.3 796.0 9.9 12.1 13.7 58.0 40.0 17.6 818 8.0 13.6 5.8 55.0 40.0 
BC×I 7.7 1064.0 10.0 10.9 21.9 57.0 39.0 21.2 1491 10.1 14.0 8.8 55.0 40.0 
BC×M 8.8 1041.0 12.5 11.1 17.6 61.0 37.0 24.6 1052 10.5 12.9 12.0 54.0 39.0 
Beledi A 7.2 802.0 10.1 13.0 16.0 59.0 40.0 3.7 1057 10.5 12.4 3.3 59.0 41.0 
Beledi C 22.7 778.0 8.9 11.0 14.3 61.0 38.0 20.9 905 9.2 13.7 12.1 53.0 38.0 
Dan lla 26.1 1244.0 19.0 11.0 13.1 71.0 49.0 7.5 1495 17.2 12.3 6.2 72.0 49.0 
IT93K−503−1 18.5 2559.0 13.6 16.1 29.3 67.0 44.0 6.3 1739 14.9 12.5 2.9 69.0 47.0 
L1B×D 13.2 1368.0 10.0 16.4 18.4 55.0 37.0 12.9 1211 11.6 14.3 6.9 53.0 38.0 
L1B×I 12.2 1312.0 11.7 15.7 16.9 59.0 36.0 10.4 1378 8.3 16.2 9.7 54.0 36.0 
L1B×M 14.4 1814.0 11.9 16.6 17.2 58.0 38.0 10.7 1272 11.0 15.3 7.5 56.0 39.0 
Laduni 1B 18.6 1309.0 10.5 16.1 20.6 57.0 40.0 12.6 1454 11.2 16.0 5.7 59.0 38.0 
Mouride 6.3 1022.0 18.2 12.6 11.9 66.0 44.0 12.6 1464 17.0 12.3 7.8 63.0 48.0 
TA×D 16.3 1311.0 11.7 13.1 15.9 59.0 38.0 11.6 1126 11.3 10.8 6.9 55.0 37.0 
TA×I 9.2 885.0 10.5 10.4 17.1 58.0 38.0 14.6 1213 10.6 12.0 5.8 55.0 39.0 
TA×M 11.7 1344.0 13.3 13.8 17.5 54.0 35.0 18.1 1175 14.8 13.5 7.0 53.0 37.0 
Titinwa A 14.6 1357.0 11.0 13.6 20.2 55.0 37.0 28.1 1110 11.5 12.8 8.0 56.0 37.0 
Mean 13.9 1283.4 12.1 13.5 17.1 59.0 39.0 14.6 1262 11.7 13.3 7.1 57.0 50.0 
CV% 79.8 33.7 12.5 17.3 29.3 5.8 5.3 46.4 33.4 11.8 12.4 28.6 4.3 4.9 
Lsd p<0.05 19.04 740.2 2.6 4.0 8.5 5.9 3.5 11.6 505.6 2.4 2.8 3.5 4.2 3.3 
Hi: Harvest index, SY: Seed yield, 100swt: 100-seed weight, NPP: Number of pods per plant, NSP: Number of seeds per pod, 95% mat: 95% maturity and 50%flo: 50% 
flowering 
 
 
126 
 
University of Ghana http://ugspace.ug.edu.gh
 
6.3.3 Crop Performance at Fumesua  
The overall drought tolerance index of the test site at Fumesua (Kumasi) was 0.29, STI was 
0.71 and TOL was 523.1 (Table 6.2). Under drought stressed condition, genotype 
AGRAC−216 gave the highest yield (1847.2 kg ha-1) and BC×D gave the lowest yielder 
(818.0 kg ha-1). Among the progenies evaluated under drought stress, BC×I gave the best 
yield producing 1491 kgha−1 of seed yield, while BC×D had the lowest seed yield (818.0 kg 
ha-1). Under well-watered condition, Asontem had the highest yield of 2750.0 kgha−1 and 
A1B×D had the lowest yield of 1211.0 kgha−1 (Table 6.7). AGRAC−216, which gave the 
highest yield under drought stress, had stress STI value of 0.7, AMP of 2169.6, and GMP of 
2145.5. However, among the progenies, BC×I had TOL value of −57.6 and DII of 0.0 despite 
having STI value of 1.0, AMP of 1461.8, with GMP of 1461.5 and DII of 0.0 (Table 6.7).  
AGRAC−216 had the highest 100−seed weight of 20.1 and 20 g, respectively, under both 
drought-stress and well-watered conditions. A1B×D registered the lowest 100−seed weight 
(8.6 g) under drought stress, while Beledi C had the lightest 100-seed weight (7.4 g) under 
well-watered conditions (Table 6.8). 
Under drought stress conditions, Asontem gave the highest number of seeds per pod (17) 
compared to AGRAC−216 and BC×D with the lowest number of seeds (11.7) per pods. 
Apagu 1B had the highest number of seeds per pod (16) and Mouride registered the lowest 
number of seeds per pod (11) under well-watered condition. BC×I recorded the highest 
number of pods per plant (11.7) and Dan lla registered the lowest numbers of pods per plant 
(8.4) under drought stress environment. Whereas IT93K−503−1 had the highest number of 
pods per plant (39.4) and TA×D had the lowest pod number per plant (20.9) under well-
watered condition.  (Table 6. 8). 
In terms of maturity, Mouride took the longest number of days (64 days) to maturity and 
Titinwa A had the shortest number of days (53 days) to reach 95% maturity (extra early) 
127 
 
University of Ghana http://ugspace.ug.edu.gh
 
under drought stress condition. Dan lla took the longest number of days (71 days) to attain 
95% maturity and A1B×D had the shortest number of days (54.0 days) to reach 95% maturity 
under optimum condition. Under drought stress condition, Dan lla and Mouride had the 
longest number of days (53 days) to attain 50% flowering, whereas BC×M had the shortest 
number of days (42 days) to reach 50% flowering. Dan lla took 48.0 days to achieve 50% 
flowering whereas both TA×D and TA×M attained 50% flowering at 37.0 days under well-
watered condition. Under drought stress condition, both Asontem and IT93K−503−1 had the 
highest HI value (44%), while L1B×D had the lowest HI value (9%). Asontem had the 
highest HI (31%) and BA×D registered the lowest HI value (9%) under the well-watered 
condition (Table 6.8). 
 
 
 
 
 
 
 
 
 
 
 
 
 
128 
 
University of Ghana http://ugspace.ug.edu.gh
 
 Table 6. 7: Seed yield performance of 25 genotypes and other agronomic traits measured under drought stressed and well-watered conditions at 
Fumesua 
Attribute Genotypes WW DS STI AMP GMP TOL DII 
High Yielding Ability AGRAC−216 2492.0 1847.2 0.7 2169.6 2145.5 644.8 0.3 
 IT93K−503−1 2698.0 1739.2 0.6 2218.6 2166.2 958.8 0.4 
 Asontem 2750.0 1648.0 0.6 2199.0 2128.8 1102.0 0.4 
 Dan lla 2070.0 1494.8 0.7 1782.4 1759.0 575.2 0.3 
 BC×I 1433.0 1490.6 1.0 1461.8 1461.5 -57.6 0.0 
 Mouride 2073.0 1464.0 0.7 1768.5 1742.1 609.0 0.3 
 Laduni 1B 1833.0 1454.4 0.8 1643.7 1632.8 378.6 0.2 
 L1B×I 1711.0 1377.8 0.8 1544.4 1535.4 333.2 0.2 
 A1B×M 1884.0 1301.2 0.7 1592.6 1565.7 582.8 0.3 
 BA×D 1282.0 1279.2 1.0 1280.6 1280.6 2.8 0.0 
 BA×I 2504.0 1274.2 0.5 1889.1 1786.2 1229.8 0.5 
 L1B×M 1901.0 1272.2 0.7 1586.6 1555.1 628.8 0.3 
Medium Yielding Ability TA×I 1403.0 1213.4 0.9 1308.2 1304.8 189.6 0.1 
 L1B×D 1647.0 1211.4 0.7 1429.2 1412.5 435.6 0.3 
 TA×M 1798.0 1174.6 0.7 1486.3 1453.2 623.4 0.3 
 BA×M 1991.0 1171.4 0.6 1581.2 1527.2 819.6 0.4 
 TA×D 1387.0 1126.0 0.8 1256.5 1249.7 261.0 0.2 
 Titinwa A 1839.0 1110.0 0.6 1474.5 1428.7 729.0 0.4 
 Beledi A 1384.0 1056.6 0.8 1220.3 1209.3 327.4 0.2 
 BC×M 2028.0 1052.2 0.5 1540.1 1460.8 975.8 0.5 
 Apagu 1B 1296.0 1051.8 0.8 1173.9 1167.5 244.2 0.2 
 A1B×I 1303.0 1028.2 0.8 1165.6 1157.5 274.8 0.2 
Low Yielding Ability A1B×D 1211.0 976.2 0.8 1093.6 1087.3 234.8 0.2 
 Beledi C 1315.0 905.4 0.7 1110.2 1091.1 409.6 0.3 
 BC×D 1381.0 818.2 0.6 1099.6 1063.0 562.8 0.4 
Mean   1784.6 1261.5      
CV%  25.30 15.82      
Significance   *** ***      
WW: well-watered condition, DS: Drought-stressed condition, STI: Stress tolerance index, AMP: Arithmetic means productivity, GMP: Geometric means productivity, TOL: 
Drought tolerance index, DII: Drought intensity index 
 
 
129 
 
University of Ghana http://ugspace.ug.edu.gh
 
Table 6. 8: Means from analysis of variance measured under drought stress and well-watered conditions at Fumesua 2018 
Env 2 (Fumesua Well-Watered) Env 5 (Fumesua Drought-Stressed)  
Genotype HI SY (kg/ha) 100swt (g) NSP NPP 95%mat 50%flo HI SY (kg/ha) 100swt (g) NSP NPP 95%mat 50%flo 
A1B×D 16.8 1211.0 9.1 13.0 24.1 54.0 41.0 16.9 976.0 8.6 13.2 10.2 63.0 47.0 
A1B×I 20.9 1303.0 9.6 13.8 24.9 56.0 40.0 13.8 1028.0 9.7 13.5 9.5 60.0 46.0 
A1B×M 20.1 1884.0 11.0 13.7 24.7 56.0 38.0 24.8 1301.0 10.8 14.2 10.1 58.0 43.0 
AGRAC−216 19.9 2492.0 20.0 12.7 21.8 59.0 46.0 38.8 1847.0 20.1 11.7 9.6 54.0 43.0 
Apagu 1B 15.4 1296.0 9.3 16.0 23.0 56.0 42.0 30.9 1052.0 9.1 13.8 10.1 57.0 46.0 
Asontem 31.3 2750.0 14.6 14.6 30.6 58.0 48.0 9.7 1648.0 13.5 17.0 8.7 60.0 49.0 
BA×D 8.9 1282.0 9.9 13.0 26.0 60.0 42.0 12.7 1279.0 10.6 13.5 10.6 63.0 45.0 
BA×I 15.3 2504.0 11.5 12.7 37.3 55.0 41.0 10.8 1274.0 11.0 13.3 10.4 59.0 42.0 
BA×M 13.5 1991.0 10.8 13.8 32.3 57.0 40.0 22.4 1171.0 10.9 13.2 9.8 62.0 50.0 
BC×D 21.8 1381.0 9.7 12.8 26.0 59.0 39.0 16.1 818.0 8.8 11.7 9.6 61.0 46.0 
BC×I 13.3 1433.0 10.7 12.4 27.1 59.0 39.0 21.9 1491.0 11.4 13.6 11.5 59.0 42.0 
BC-M 18.2 2028.0 9.7 12.0 31.9 60.0 38.0 24.6 1052.0 9.6 12.4 10.6 58.0 46.0 
Beledi A 10.5 1384.0 11.0 12.8 26.4 59.0 42.0 23.7 1057.0 10.6 13.3 8.9 61.0 52.0 
Beledi C 12.4 1315.0 7.4 11.3 32.0 59.0 40.0 24.0 905.0 9.0 12.1 10.0 54.0 42.0 
Dan lla 18.7 2070.0 16.1 12.0 24.8 71.0 48.0 35.7 1495.0 16.2 13.2 8.4 63.0 53.0 
IT93K−503−1 30.0 2698.0 13.6 13.8 39.4 67.0 47.0 43.7 1739.0 15.4 13.3 10.4 61.0 48.0 
L1B×D 14.4 1647.0 10.8 14.8 25.9 55.0 38.0 8.7 1211.0 10.5 14.4 9.5 58.0 42.0 
L1B×I 24.9 1711.0 10.8 15.5 23.0 59.0 37.0 33.9 1378.0 10.8 15.8 9.7 58.0 46.0 
L1B×M 14.0 1901.0 11.1 14.5 25.9 58.0 38.0 13.4 1272.0 11.1 14.9 9.2 60.0 48.0 
Laduni 1B 24.4 1833.0 11.6 15.7 26.7 57.0 41.0 24.8 1454.0 13.3 13.4 9.8 52.0 43.0 
Mouride 24.4 2073.0 15.8 11.0 24.7 66.0 49.0 18.9 1464.0 16.0 12.0 9.2 64.0 53.0 
TA×D 13.5 1387.0 11.9 13.3 20.9 59.0 37.0 15.3 1126.0 10.8 13.5 9.4 63.0 47.0 
TA×I 15.2 1403.0 11.1 12.7 25.3 58.0 38.0 19.3 1213.0 11.1 13.6 9.6 60.0 45.0 
TA×M 17.9 1798.0 10.6 13.5 23.9 55.0 37.0 24.2 1175.0 10.7 13.5 9.6 59.0 44.0 
Titinwa A 17.8 1839.0 10.9 13.3 30.7 55.0 39.0 24.2 1110.0 11.5 12.5 9.3 53.0 42.0 
Mean 18.1 1784.6 11.5 13.4 27.2 59.0 41.0 22.1 1261.4 11.7 13.5 9.7 59.0 46.0 
CV% 43.7 25.3 6.9 1.9 18.3 5.9 4.0 53.0 15.8 9.9 7.4 8.3 7.6 10.2 
Lsd p<0.05 13.6 773.1 1.4 8.2 8.5 5.9 2.8 20.1 341.7 2.0 1.7 1.4 7.7 8.04 
Hi: Harvest index, SY: Seed yield, 100swt: 100-seed weight, NPP: Number of pods per plant, NSP: Number of seeds per pod, 95%mat: 95% maturity and 50%flo: 50% 
flowering
130 
 
University of Ghana http://ugspace.ug.edu.gh
 
6.3.4 Combined Analysis of Variance  
Combined analysis of variance across the three test locations revealed highly significant 
(P<0.001) differences among the genotypes for all measured traits under both watering 
regimes (Table 6.9). Mean seed yield, 100−seed weight, number of pods per plant, number of 
seeds per pod, pod length, number of days to 95% maturity, number of days to 50% 
flowering and harvest index were significantly different among the cowpea genotypes under 
each watering regime. Environmental differences were significant (P<0.001) for all the 
measured traits. Watering regime was the main source of variation among the genotypes for 
seed yield and most of its components measured in this study showing highly significant 
(P<0.001) differences among the genotypes (Table 6.9 and Fig. 6.3). 
Seed yield in kg/ha
2000
1784.56
1800
1600
1400 1261.44 1283.36 1261.44
1200 999.172
1000
800
600
386.456
400
200
0
Fumesua Legon Nyankpala
Well-Watered Drought-Stressed
  
Fig. 6. 3 Estimated mean squares of seed yield across sites under drought stress (DS) and well-watered (WW) 
conditions 
 
 
 
 
 
 
131 
 
University of Ghana http://ugspace.ug.edu.gh
 
The structural equation modelling (SEM) and the extension of several multivariate analyses 
explained the relationships among variables. Number of days to 50% flowering and 95% 
maturity are exogenous variables and tend to be negatively related with number of pods per 
plant similar to the association between HI and number of seeds per pod. HI was positively 
associated with seed yield similar to the relationship between number of pods per plant and 
number of seeds per pod. The result also revealed that number of seeds per pod and number 
pods per plant had a positive influence on seed yield (Fig. 6.4).   
 
Fig. 6. 4 Path diagram cause (traits studied) and effect (seed yield) relationship of 25 cowpea genotypes 
 
 
 
132 
 
University of Ghana http://ugspace.ug.edu.gh
 
Table 6.9: Mean squares from the combined analysis of variance of seed yield and yield components measured under drought stress and well-
watered conditions 
Source of DF Seed Yield 100SWT (g) NSP Pod length NPP 95%mat 50% HI (kgha−1) 
Variation  (kgha−1) (cm) flowering 
                                                  Drought−Stressed Condition 
Bloc 4 81346 8.0 4.6 11.3 14.6 119.6 109.6 314 
Rep 2 215751 1.0 0.0 0.5 0.7 390.3 133.0 578.5 
Geno 24 199113*** 65.3*** 4.2** 26.82*** 9.3*** 69.0* 91.9*** 440.7*** 
Env 2 14445667*** 47.9*** 1527.6*** 652.0*** 659.5*** 13279.2*** 1311.4*** 1489.1*** 
Geno × Env 48 172397*** 6.6*** 4.1*** 4.2*** 8.9*** 44.0* 18.0* 250.6** 
Error 144 69975.0 2.9 2.1 1.0 2.7 43.4 13.6 166.5 
 Total 2 24 2 35618.0 10.9 16.4 10.5 10.8 171.1  37.3 2 32 
               Well−Watered Condition 
Bloc 4 332390 23.5 1.4 8.1 74.1 51.2 62.5 281.2 
Rep 2 387552 0.6 19.6 2.3 21.9 6.1 114.5 2258.2 
Geno 24 699491*** 63.3*** 10.4*** 30.4*** 67.6*** 88.3*** 106.0*** 502.9* 
Env 2 11858298*** 10.2*** 339.7*** 223.1*** 5550.5*** 5516.4*** 845.4*** 15923.7*** 
Geno × Env 48 361764*** 3.8*** 5.5*** 2.3*** 39.0*** 14.0* 6.9* 319.6* 
Error 144 183543 1.6 2.7 1.2 19.5 9.5 6.7 320.5 
Total 224 385732 9.1 7.3 6.7 79.2 68.8 26.8 495.8 
*, **, *** Significantly different at 0.05, 0.01 and 0.001 levels of probability, respectively, 100SWT: 100−seed weight, NSP: Number of seeds per pod, NPP: Number of 
pods per plant, 95% mat: Number of days to 95% maturity, 50% flowering: Number of days to 50% flowering and HI: Harvest index 
 
 
 
 
 
 
 
133 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
Fig. 6. 5 Estimated means of overall seed yield of 25 cowpea genotypes at (P<0.05) 
NB: Geno (genotypes) and seed yield (kgha-1) 
In the combined analysis of variance across stress and non-stressed environments, genotype, 
environment (individual water regimes considered as environment) and genotype by 
environment interactions were significant (P<0.001) sources of variation among genotypes 
for seed yield and yield components (Table 6.10). Significant differences were found for 
genotype by environment interaction for HI and number of pods per plant. Genotype and 
genotype by environment interaction were not significantly (P<0.05) different for the number 
of days to 95% maturity and the number of days to 50% flowering.   However, Apagu 1B, 
Beledi C × Dan lla and Beledi A had the lowest yield. Whereas, Beledi ×Mouride, Dan lla, 
IT93K−503−1 and Mouride were reported to be high yielding across the testing sites (Fig. 
6.5). 
 
 
 
134 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
Table 6.10: Mean squares from the combined analysis of variance for seed yield and yield components of 25 cowpea genotypes measured across 
three test sites under drought stress and well-watered conditions 
Source of Variation  D.F Seed Yield (kgha−1) 100SWT (g) NSP Pod length (cm) NPP 95%mat 50% flowering HI 
Bloc 4 301939 25.8 4.6 117.9 42.8 109.96 192.4 359.8 
Rep 2 67698 2.1 5.6 22.7 16.4 167.8 224.7 487.1 
Geno 24 607686*** 120.955*** 10.7*** 55.4** 34.7*** 145.03*** 187*** 750.8*** 
Env 5 16418495*** 23.016*** 827.5*** 706.4*** 5190.6*** 7915.1*** 1057.5*** 7449.5*** 
Geno × Env 120 270609*** 4.789*** 3.9*** 25.8* 30.5*** 36.53ns 11.1ns 273.1ns 
Pooled Error 294 129843 2.5 2.4 31.8 13.7 30.31 10.0 255.1 
Total 449 375650 9.9 12.5 39.7 77.2 127.23 34.0 368.5 
*, **, *** Significantly different at 0.05, 0.01 and 0.001 levels of probability, respectively, 100SWT: 100−seed weight, NSP:  Number of seeds per pod, NPP: Number of 
pods per plant, 95% mat: Number of days to 95% maturity, 50% flowering: Number of days to 50% flowering and HI: Harvest index 
 
 
 
 
 
 
 
 
 
 
135 
 
University of Ghana http://ugspace.ug.edu.gh
 
6.3.5 AMMI Stability Analysis 
The seed yield AMMI analysis showed significant (P<0.001) differences for genotype, 
environment, and genotype by environment interaction. Environment, genotype, and GEI 
accounted for 63.8, 10.8% and 25.4% of the total sum of squares, respectively. A high 
proportion of the sum of squares of the AMMI model was due to environment and GEI 
effects. Additionally, the model divided the GEI sum of squares into interaction principal 
component axes (IPCA) and residual. The mean squares of the first three IPCA were 
significantly different and accounted for 83.5% of the total variation (Table 6.11).  
The graphical representation of the AMMI analysis in the biplot revealed that genotype 
environmental responsiveness is enormous with each environment influencing a set of 
genotypes (Fig..6.6). The AMMI analysis for the first singular axis captured the highest 
percentage (38.9%) of the variability. The values of PC1 component showed that genotypes 
designated as 4 (AGRAC−216), 6 (Asontem), 8 (BA×I) and 16 (IT93K−503−1) were the best 
and equally productive under both well-watered and drought stressed conditions. The most 
stable genotypes were IT93K−503−1, Beledi C × Mouride and Mouride producing seed yield 
above average trial mean (995.54 kgha−1). 
 
 
 
 
 
 
 
136 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
Table 6. 11: AMMI analysis of variance for seed yield across locations  
Source of Variation D.F SS MS Sum of Squares Explained  
    %Total %G × E %G × E Cumulative 
 
Env 5 82092476.1 16418495.2*** 63.8     
Geno 24 13879110.0 578296.3*** 10.8    
Env × Geno 120 32672984.1 272274.9*** 25.4     
IPCA 1 28 12491716.6 446132.7*** 39.8 39.8 39.8  
IPCA 2 26 9790907.4 376573.4*** 31.2 31.2 71.0   
IPCA 3 24 3930368.2 163765.3 12.5 12.5 83.5  
IPCA 4 22 3225490.7 146613.2 10.3 10.3 93.8   
IPCA 5 20 1955458.4 97772.9 6.2 6.2 100.0  
Residuals 300 40022335.3 133407.8      
*** Significantly different 0.001 level of probability  
 
 
 
 
 
 
 
 
 
 
 
 
137 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
Fig. 6. 6 AMMI 1biplot for seed yield of 25 cowpea genotypes denoted by numbers and six environments using 
genotypic and environmental scores  
 
ENVIRONMENTS: ENVLN: Legon well-watered, ENVLS: Legon drought stressed, ENVKS: Fumesua well-
watered, ENVKT: Fumesua drought stressed, ENVTN: Nyankpala well-watered and ENVTS: Nyankpala 
drought stressed 
GENOTYPES:   Dan lla (15), Asontem (6), AGRAC−216 (4), IT93K−503−1 (16), Mouride (21), BC×M (12), 
Beledi C (14), BC×D (10), BA×I (8), BA×D (7), Laduni 1B (12), Titinwa A (25), TA×M (24), L1B×I (18), 
L1B×M (19), L1B×D (17), BA×M (9), BC×I (10), Beledi A (13), TA×D (22), Apagu 1B (5), TA×I (23), 
A1B×M (3), A1B×D (1) and A1B×I (2) 
 
 
 
 
 
 
 
 
 
 
 
 
138 
 
University of Ghana http://ugspace.ug.edu.gh
 
6.3.6 GGE Biplot Analysis 
The first two principal components of the biplot explain 62.6% of the variability among the 
genotypes across the test environments. There were a few of the genotypes designated by 
numbers on the crests of the polygon. Genotypes at the vertex of the polygon were 16 
(IT93K−503−1), 21 (Mouride), 12 (BC×M), 14 (Beledi C), 10 (BC×D), 8 (BA×I), and 7 
(BA×D). The majority of the genotypes were found within the polygon (Fig. 6.7).  
Fig. 6. 7 Polygon view of genotype by environment interaction of 25 cowpea genotypes 
 
Genotypes on the peak had a long detachment from the origin of the biplot. Genotype 16 
(IT93K−503−1) in ENVLN (Legon well-watered) is a vertex genotype with the highest seed 
yield. Environments Nyankpala drought stress (ENVTS), Fumesua drought stress (ENVKS), 
139 
 
University of Ghana http://ugspace.ug.edu.gh
 
Fumesua well-watered (ENVKN) and Nyankpala well-watered (ENVTN) were laid on the 
same portion of the graph, suggesting that these environments do not differ significantly in 
terms of seed yield. Genotypes 12 (BC×M) and 21 (Mouride) were the only two genotypes in 
the environment ENVTN (Fig..6.7). Genotypes 25 (Titinwa A), 24 (TA×I), and 18 (L1B×I) 
were less responsive and low yielding compared to the vertex genotypes. The best genotypes 
were 16 (IT93K−503−1), 21 (Mouride), and 12 (BC×M) (Fig. 6.7).  
The discrimination pattern of the genotypes and environmental representativeness showed 
that ENVTS and ENVTN were more representative environments for testing cowpea 
genotypes and likewise ENVKS and ENVKT. Environments with long vectors, ENVLS, 
ENVLN, ENVTN, and ENVKN, were more discriminating than those with short vectors, 
indicating their ability for genetic discrimination (Fig..6.8).  
140 
 
University of Ghana http://ugspace.ug.edu.gh
 
Fig. 6. 8 Discrimination power and representativeness of test sites 
 
The rationale behind the visual display of mean vs stability in GGE biplot analysis is to help 
breeders compare genotypes based on performance and stability across testing sites. The 
equality lines which connected lines between genotypes also aid in grouping the genotypes 
into a specific environment based on their responsiveness. Among the 25 cowpea genotypes 
evaluated across six environments, genotype 6 (Asontem) was the most stable and high 
yielding followed by genotypes 4 (AGRAC−216) and 8 (BA×I). However, genotype 16 
(IT93K−503−1) was high yielding but unstable across test environments. Genotype 20 
(Laduni 1B) was found to be a low responsive genotype across the test environments, but 
suitable only for the unfavourable environments (Fig..6.9).  
141 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
Fig. 6. 9 Mean vs Stability view of 25 cowpea genotype main effect and GEI effect 
 
 
 
 
 
 
 
 
 
.
142 
 
University of Ghana http://ugspace.ug.edu.gh
 
6.3.7 Adaptability and Yield Stability Analyses 
With respect to seed yield, genotypes BC×M, Dan lla, TA×M, L1B×I, A1B×D Mouride and 
A1B×M had good performance and were stable across the six test environments. They had 
higher mean seed yield; their bi values were less than one and had non-significant s
2di values 
(Table 6.8). Mouride, IT93K−503−1, L1B×D, L1B×M, Laduni 1B, Titinwa A, 
AGRAC−216, Asontem and BA×I had higher mean seed yield and non-significant s2di, but 
their bi values were either equal or more than (bi≤1.0). Based on Wricke's Ecovalence 
analysis of adaptability and stability, Laduni 1B was found to be stable, and AGRAC−216 
and BA×I were adaptable (Table 6.8). Laduni 1B had the lowest value, followed by L1B×D, 
and L1B×I (99150.2). Asontem registered the highest Ecovalence value of 1174065. 
Genotypes that are high yielding across sites with low coefficients of variability (CVi) are 
considered to be stable. Accordingly, A1B×M, BC×M, Dan lla, and L1B×I had CVi values 
lower than 40% and seed yield of more than 1000 kgha−1, whereas BA×I had CVi value of 
68.7% with seed yield of 1180.6 kgha−1 (Table 6.12). 
 
 
 
 
 
 
 
 
 
 
 
143 
 
University of Ghana http://ugspace.ug.edu.gh
 
 
Table 6.12: Means estimates of adaptability and phenotypic stability for 25 cowpea genotypes evaluated across three locations under drought 
 stressed and well-watere d conditions 
Eberhart & Russell Wricke's Ecovalence Francis and Kannenberg 
Genotype Code Mean Rank bi S2di Rank R2 Wi Sd CV (%) Rank 
A1B×D 1 1168.6 abcd 9 0.7 210679.9 2 0.4 1083814 569.5 48.7 16 
BC×D 10 766.6 e 25 0.7 -30663.2 24 0.9 140384.5 345.8 45.1 12 
BC×I 11 1090 bcde 16 0.6 -6174.8 14 0.7 348026.3 316.7 29.1 1 
BC×M 12 1346.4 ab 2 0.6 126614.3 4 0.4 812650.8 476.7 35.4 4 
Beledi A 13 814.7 de 24 0.9 22510.1 11 0.8 271734 477.5 58.6 21 
Beledi C 14 904.4 cde 22 0.7 61269.2 6 0.6 506372.2 440.4 48.7 15 
Dan lla 15 1288 abc 4 0.9 -9956.2 16 0.9 134438.9 467 36.3 6 
IT93K−503−1 16 1526.1 a 1 1.8 223245.7 1 0.8 1747226 956.9 62.7 22 
L1B×D 17 1100.4 bcde 15 1.1 -21834.3 22 0.9 86270 512.9 46.6 13 
L1B×I 18 1139.3 abcde 11 0.9 -20576.6 21 0.9 99150.2 440.4 38.7 7 
L1B×M 19 1151.2 abcde 10 1.2 35789.1 8 0.8 342413 600.1 52.1 18 
A1B−I 2 1021.6 bcde 17 0.7 -17817.2 19 0.9 187867.3 363 35.5 5 
Laduni 1B 20 1125.8 bcde 13 1.1 -36059.3 25 1 28299.8 497.6 44.2 11 
Mouride 21 1342.7 ab 3 1 123347.9 5 0.6 663875.9 583.5 43.5 10 
TA×D 22 961.7 bcde 20 0.8 -17129.1 18 0.9 157783.4 389.1 40.5 8 
TA×I 23 919.9 cde 21 0.6 -13501.1 17 0.8 283249 323.5 35.2 3 
TA×M 24 1132.8 abcde 12 0.9 -8580.7 15 0.9 138622.7 472.6 41.7 9 
Titinwa A 25 1117.8 bcde 14 1.1 -5730.8 13 0.9 168855.9 561.8 50.3 17 
A1B×M 3 1177.4 abcde 8 0.7 -25143 23 0.9 148802.4 361.9 30.7 2 
AGRAC−216 4 1240.7 abc 5 1.7 34643.4 10 0.9 836004.3 831.4 67 23 
Apagu 1B 5 843.4 de 23 0.8 -17842.1 20 0.9 140399.9 402.1 47.7 14 
Asontem 6 1183.3 abcd 6 1.6 139600.7 3 0.8 1174065 856.9 72.4 25 
BA×D 7 976.9 bcde 19 1 35441.5 9 0.8 312233.8 541.9 55.5 19 
BA×I 8 1180.6 abcd 7 1.7 9740.1 12 0.9 708310.3 810.8 68.7 24 
BA×M 9 996.7 bcde 1 8 1 .1  46530.1 7  0 .8 367308.6  581.3  58.4  20 
Overall mean 1100.4 4435526 
Means followed by a similar letter belongs to the same class (P<0.05), bi: regression coefficient R2: regression coefficient of determination   Wi: Ecovalence    
Sd: deviation from the regression, CVi: Francis and Kannenberg's Coefficient of variability  
 
 
 
144 
 
University of Ghana http://ugspace.ug.edu.gh
 
The mean AMP (Arithmetic means productivity) for seed yield across six environments was 
1104.0 kgha−1. Genotype IT 93K−503−1 had the highest AMP value of 1554.0 kgha−1, 
whereas BC×D had the lowest value of 748 kgha−1. A similar trend was observed in GMP for 
these genotypes. IT93K−503−1 had the highest GMP value (1398.8 kgha−1) and BC×D 
scored the lowest value (722.9 kgha−1). AGRAC−216 had the highest GMP (GMP: 
Geometric means productivity) value (18.0) for 100−seed weight, while A1B×D had the 
lowest (8.3). L1B×D had the highest 100−seed weight percent reduction index of 19.9 and 
Beledi C had the lowest value of −8.3 (Table 6.13). L1B×I had the highest GMP (12.9) for 
the number of seeds per pod, whereas TA×I had the lowest GMP value of 10.1. Percent 
reduction for HI revealed that Laduni 1B had the highest index of 20.3, while BC×I had the 
lowest value of 2.7. Beledi C had the highest GMP of 13.4 for number of pods per plant and 
AGRAC−216 had the lowest value (8.9). The highest percent reduction (60.7%) seed yield 
was observed in IT93K−503−1 and TA×I had the lowest value (18.0%) (Table 6.13).   
 
 
 
 
 
 
 
 
145 
 
University of Ghana http://ugspace.ug.edu.gh
 
Table 6.13: Overall mean performances of 25 cowpea genotypes evaluated across sites under drought stressed and well-watered conditions 
Genotype NPP       NSP       100−seed yield (g) Seed yield kgha−1   
  WW DS GMP PR% WW DS GMP PR% WW DS GMP PR% WW DS AMP GMP PR% STI 
Parents                   
Apagu 1B 17.0 8.5 12.1 50.0 13.6 11.1 12.3 18.4 8.9 8.5 8.7 4.5 1007 685.6 846 830.9 31.9 0.3 
Laduni 1B 19.4 7.6 12.1 60.8 14.7 11.7 13.1 20.4 11.1 11.9 11.5 -7.2 1444 868.8 1156 1120.1 39.8 0.6 
Mouride 15.5 7.0 10.4 54.8 11.0 10.2 10.6 7.3 17.1 16.3 16.7 4.7 1604 1078 1341 1315 32.8 0.9 
Beledi A 17.9 5.3 9.7 70.4 12.1 10.9 11.5 9.9 10.1 10.1 10.1 0.0 1044 548 796 756.4 47.5 0.3 
Beledi C 20.7 8.7 13.4 58.0 11.5 10.7 11.1 7.0 8.2 8.9 8.5 -8.5 999 795.4 897 891.4 20.4 0.4 
Dan lla 16.4 7.0 10.7 57.3 11.3 10.9 11.1 3.5 17.3 16.6 17 4.0 1435 1112.2 1274 1263.3 22.5 0.8 
IT93K−503−1 26.7 6.0 12.7 77.5 13.8 11.2 12.4 18.8 13.9 14.7 14.3 -5.8 2232 876.5 1554 1398.7 60.7 1 
Titinwa A 21.5 7.0 12.3 67.4 12.2 10.0 11.0 18.0 11.4 11 11.2 3.5 1487 774.7 1131 1073.3 47.9 0.6 
Mean 19.4 7.1 11.7 63.4 12.5 10.8 11.6 13.6 12.3 12.2 12.3 0.8 1406.5 842.4 1124 1081.1 40.1 0.6 
Crosses                   
A1B×D 17 8.4 11.9 50.6 11.9 10.9 11.4 8.4 9 7.6 8.3 15.6 1472 807.9 1140 1090.5 45.1 0.7 
A1B×I 16.5 7.9 11.4 52.1 12.6 11.1 11.8 11.9 9.3 9 9.1 3.2 1068 826.9 947 939.7 22.6 0.5 
A1B×M 16.8 7 10.8 58.3 12.1 10.8 11.4 10.7 12.1 11.4 11.8 5.8 1374 986.8 1180 1164.4 28.2 0.8 
BA×D 19.1 6.5 11.2 66.0 12.2 10.9 11.5 10.7 9.5 8.8 9.1 7.4 1303 656 980 924.5 49.7 0.5 
BA×I 20.4 5.8 10.9 71.6 11.8 11 11.4 6.8 11.4 11.1 11.3 2.6 1622 748.4 1185 1101.8 53.9 0.7 
BA×M 20.4 5.2 10.3 74.5 11.8 10.4 11.1 11.9 11 10.2 10.6 7.3 1306 774.1 1040 1005.5 40.7 0.6 
BC×D 17.1 5.9 10 65.5 11.7 10.1 10.8 13.7 8.7 8.4 8.6 3.4 940 556 748 722.9 40.9 0.3 
BC×I 20 8 12.7 60.0 11 10.7 10.8 2.7 10.6 9.9 10.3 6.6 1202 1017.8 1110 1106.1 15.3 0.7 
BC×M 21 8.2 13.1 61.0 11.1 10.3 10.7 7.2 10.9 10 10.4 8.3 1553 1062.7 1308 1284.7 31.6 1 
L1B×D 17.7 6.9 11.1 61.0 13.5 11.5 12.4 14.8 11.1 10.5 10.8 5.4 1474 830.6 1152 1106.5 43.6 0.7 
L1B×I 15.8 7.3 10.7 53.8 13.3 12.5 12.9 6.0 11.4 9.1 10.2 20.2 1365 955.1 1160 1141.8 30.0 0.8 
L1B×M 16.7 6.6 10.5 60.5 13.2 11.6 12.4 12.1 11.7 11 11.4 6.0 1438 897.2 1168 1135.9 37.6 0.8 
TA×D 15.5 6.8 10.2 56.1 11.4 9.6 10.4 15.8 11.8 10.7 11.2 9.3 1168 702.3 935 905.7 39.9 0.5 
TA×I 17.6 6.2 10.4 64.8 10.4 9.9 10.1 4.8 10.2 10.5 10.3 -2.9 976 800.5 888 883.9 18.0 0.5 
TA×M 17 6.4 10.4 62.4 11.6 10.4 10.9 10.3 12.3 13.1 12.7 -6.5 1304 895 1100 1080.3 31.4 0.7 
Mean 17.9 6.9 11 61.5 12 10.8 11.3 10.0 10.7 10.1 10.4 5.6 1304.3 834.5 1069 1039.6 36.0 0.6 
Checks                   
AGRAC−216 13.6 5.8 8.9 57.4 11 9.6 10.3 12.7 18.7 17.6 18.2 5.9 1521 963.4 1242 1210.5 36.7 0.6 
Asontem 14.8 6.2 9.6 58.1 14.1 11.6 12.8 17.7 14.2 13.7 13.9 3.5 1554 874 1214 1165.4 43.8 0.6 
Mean 14.2 6 9.2 57.7 12.5 10.6 11.5 15.2 16.5 15.6 16.1 5.5 1537.5 918.7 1228 1188 40.2 0.6 
Overall mean 18 6.9 11.1 61.7 12.2 10.8 11.5 11.5 11.8 11.4 11.6 3.4 1362.2 846.1 1104.0 1068.9 37.9 0.6 
WW: Well-watered condition, DS: Drought-stressed condition, NPP: Number of pods per plant, NSP: Number of seeds per pod, GMP: Geometric productivity mean, 
Arithmetic mean productivity, STI: Stress tolerance index and PR%: Percent reduction 
 
146 
 
University of Ghana http://ugspace.ug.edu.gh
 
6.4 Discussion  
The analyses of multi-location trials conducted in Nyankpala, Fumesua and Legon revealed 
that genotype, environment, and genotype by environment interactions were significant thus 
influencing the performance of the studied traits. This result corroborates with other reports 
(Temesgen et al., 2015, Sharifi et al., 2017 and Dia et al., 2018), who conducted trials under 
contrasting condition and reported effects of environment and GEI on their performance. 
Drought and high-temperature stresses were pronounced during the field evaluation in 
Nyankpala and seed yield was reduced by 38.7%. This finding agrees with results of a study 
conducted by Devasirvatham and Tan (2018) who reported that drought and heat stresses 
could reduce seed yield of field crops by 50%. Although the results obtained through boxplot 
analysis revealed the impact of drought stress on the genotypes evaluated in Nyankpala 
evidently high, the site could still be a good testing environment for drought tolerance study.  
Equally, it may be concluded that Legon and Fumesua could be good test sites performance 
study for optimum growing conditions. 
The path analysis aspect of the study revealed a total positive association of r = 0.91 between 
number of pods per plant, number of seeds per pod, harvest index and seed yield. This 
association implies that an increment of an independent variable by a unit will increase the 
other dependent variable (seed yield). Thus, high seed yield could be obtained through 
selection for increased HI, number of pods per plant, number of seeds per pod. These results 
agree with the findings of Sousa et al. (2015) who used yield components for indirect 
selection. 
Site-specific environmental influences on the performance of the traits were also observed. 
A1B×D, the highest yielding genotype under an optimum condition at Legon in the Coastal 
Savanna agro-ecological zone, outperformed the two checks and parent IT93K−503−1. 
However, it gave an appreciable yield above the average under drought stress condition. This 
147 
 
University of Ghana http://ugspace.ug.edu.gh
 
implies that the environment had an impact on yielding ability and maturity and drought 
tolerance of the genotypes evaluated in this study. Thus, this genotype should be tested across 
multiple sites to confirm the consistency of its yield performance.  
Environmental responsiveness of the genotypes showed their levels of tolerance using 
different indices. The same pattern of variation was observed at Nyankpala in the Guinea 
Savanna agro-ecological zone, where BC×M out yielded all the other genotypes with the 
exception of IT93K−503−1. This performance could be a result of genotype, environment, 
and genotype by environment interaction effects. Genotypes evaluated at Fumesua in the 
Forest agro-ecological zone registered the highest seed yield under both drought stress and 
well-watered conditions, suggesting that Fumesua is an ideal environment for growing 
cowpea. 
This trend in the performance of traits across the test sites suggests genotypic differences 
which influenced the expression of the traits. The variation exhibited by the genotypes 
revealed a significant amount of variability that could be exploited to improve early maturing 
cowpea genotypes with higher levels of tolerance to drought for farmers. Some of the 
identified early maturing genotypes, especially those combining earliness with drought 
tolerance, could be advanced through further yield and on-farm trials for production in 
drought-prone areas.  
Ten progenies (A1B×D, A1B×I, A1B×M, BA×D, BA×M, BA×I, L1B×D, TA×M, TA×D, 
and TA×M) attained 95% maturity at 54 and 57 days in two environments (ENVKS and 
ENVLS) indicating that they are extra-early in maturity. Such genotypes are important for 
farmers to provide food during the period of hunger gap when planted early and also have the 
advantage of producing seed yield under terminal drought when planted late in the season. 
This finding agrees with results obtained by Cláudio et al. (2003)  who classified cowpeas 
into those that matured in 60 days after planting as extra-early and those that matured 
148 
 
University of Ghana http://ugspace.ug.edu.gh
 
between 61 to 70 days after planting as early maturing genotypes. The medium-early varieties 
take between 71 to 80 days to mature and medium-late are those that matured in 81 to 90 
days after sowing. The late maturing varieties attained maturity in 91 days after planting.  
Farmers in drought-prone areas cultivate cowpea as an insurance crop, particularly if they 
expect rainfall too low for growing maize; they resort to growing crop varieties that are early 
and medium early maturing in 60 to 80 days. Of the fifteen progenies evaluated in this study, 
ten demonstrated earliness with good levels of tolerance to drought completing their life 
cycle in less than 60 days. Dias et al. (2009) reported that earliness can best be evaluated 
through the duration from seedling emergence to an indication of the first flower. Adeyanju 
and Ishiyaku (2007) and Hall (2012) reported that erect and short cycle cowpeas are good for 
the drought-prone Sahel zones of West Africa. Padi (2007) stated that earliness and seed size 
are important traits for cowpea adoption in the Savannas of West Africa.  
The link between time to 50% flowering and drought stress was also observed in this study. 
Drought stress delayed the number of days to 50% flowering and number of days to 95% 
maturity to great extent. This result agrees with the findings of Kazan and Lyons (2016) who 
observed that abiotic stresses such as drought stress influence flowering time.  
The use of the established biometrical tools to assess the adaptability of genotypes and their 
yield and yield components stability is crucial, particularly when the varieties to be evaluated 
are newly developed populations. In this study, the AMMI stability analysis identified some 
genotypes including AGRAC−216, BA×I, BC×I and Laduni 1B as adaptable and stable 
though some of these genotypes were low yielding. Environment accounted for 63.8% of the 
total variability in the experiments, while genotype accounted for only 10.8% and the 
genotype and environment interaction had slightly higher (24.4%) contribution to the total 
variation. This result showed that environment explained the highest portion of the observed 
variance and therefore largely influenced the performance of genotypes. Thus, more test sites 
149 
 
University of Ghana http://ugspace.ug.edu.gh
 
should be considered for logical inferences about the performance of these genotypes.  The 
results of this study agree with the findings of Rashidi et al. (2013) who reported slightly 
higher values for environment (81.2%). Genotypes G7 (BA×D), G9 (BA×M), G10 (BC×D), 
G13 (Beledi A), G22 (TA×D), G5 (Apagu 1B), G23 (TA×I), G14 (Beledi C) and G2 (A1B×I) 
showed specific adaptability to drought-stress environments. Test environments ENVLS, 
ENVKS, and ENVTS had seed yield less than the trial mean (995.6 kgha−1) suggesting that 
these genotypes demonstrated specific adaptability. They expressed good performance with 
high seed yield in a specific test environment(s).   
In this study, GGE biplot explained 62.5 % of the total yield variation in the test 
environments. However, IT93K−503−1 could be planted by farmers under conditions such as 
EVNLN (Legon well-watered), ENVKN (Fumesua well-watered), and ENVKS (Fumesua 
drought-stressed) and obtained high yields. Mouride and BC×M showed specific adaptability 
thus, these two genotypes could only yield to their potential at Nyankpala under well-watered 
(ENVTN) and Legon drought stressed (ENVLS) conditions. AGRAC−216 was relatively 
responsive across the testing locations, but had low grain yield compared to Dan lla, 
IT93K−503−1 and BC×M. Thus, testing this genotype across other contrasting sites could 
result into assigning it to a specific location. The remaining genotypes were not responsive 
and expressed low yield potential, far below the average trial mean yield (1104.0 kg ha-1). 
The poor performance of these genotypes may be attributed to GEI effects and their genetic 
backgrounds. 
In the polygon view of the discriminating environments, ENVLN (Legon well-watered), 
ENVTN (Nyankpala drought-stressed), ENVKS (Fumesua well-watered) and ENVLS 
(Legon drought-stressed) were strongly correlated and most discriminatory, suggesting that 
removing one of the locations as a test environment would not lead to any loss of 
information. This could cut down on resources that could be put to better use in other 
150 
 
University of Ghana http://ugspace.ug.edu.gh
 
locations (Meseka et al., 2016). This finding confirms the proposal of Yan and Tinker (2005) 
who postulated that test environments that are none discriminating provide no information on 
the genotypes and therefore should not be used as test environments. 
Seed yield of 25 cowpea varieties was affected linearly by the amount of water applied. 
Drought stress clearly reduced seed yield of the 25 genotypes, suggesting that inducing 
drought at the reproductive phase of the crop disrupts reproductive mechanisms including 
floral and pod development, number of ovules per locules and pod filling contributing to high 
seed yield. Although drought stress negatively affects the three phases, the degree of the 
effect is determined by the severity and duration of drought stress. In this study, a yield 
component, number of pods per plant was most severely affected by drought stress. On an 
average, drought stress reduced the number of pods per plant by 61%.  This high level of 
reduction could be attributed to the environmental responsiveness of the genotypes. These 
findings agreed with the results obtained by Pandey et al. (1984) who reported a similar trend 
of reduction of yield components in cowpea.  
6.5 Conclusions 
The genotype and environment main effects and genotype by environment interaction effects 
were significant for seed yield and other agronomic traits in 25 cowpea genotypes evaluated 
in this study. Five of the best ten progenies, in terms of seed yield across the environments, 
were A1B×D, BC×M, L1B×M, A1B×M, and BA×D. Some of these genotypes out yielded 
their parents as well as the checks. Additionally, seven of the 25 genotypes showed stability 
and adaptability, qualifying them as good performers across the test environments. Genotypes 
BC×M, L1B×I, TA×M and A1B×M could be further tested under drought stress to confirm 
their ability to tolerate drought stress and produce appreciable seed yield, which could 
provide the opportunity for their possible release to farmers. 
 
151 
 
University of Ghana http://ugspace.ug.edu.gh
 
CHAPTER SEVEN 
7.0 GENERAL CONCLUSIONS AND RECOMMENDATIONS 
7.1 General Conclusions  
A genetic divergence study was carried out in 106 cowpea germplasm collections using 
phenotypic and molecular approaches. There was high cophenetic correlation coefficient 
(CCC) of 0.76 resulting from nine discriminatory clusters using morphological data.  
However, molecular data indicated a higher level of genetic divergence among the assembled 
cowpea panel compared to morphological data, suggesting that a good opportunity exists for 
population development through introgression of new alleles from different backgrounds. 
The SNP markers used in this study could be utilized to analyse and group new cowpea 
collections in the future. The cowpea accessions from South Sudan were closely related to the 
materials from West Africa. This could reasonably be attributed to an early 1980's IITA’s 
germplasm expedition to South Sudan who collected some cowpea accessions from the 
Equatoria region of South Sudan. The evaluation of the 106 accessions revealed that Dan lla, 
Mouride, and IT93K−503−1 were drought tolerant with both Type 1 and 2 drought tolerance 
adaptabilities. Five farmers’ preferred accessions from South Sudan, Titinwa A, Beledi A, 
Beledi C, Laduni 1B, and Apagu 1B, were susceptible to drought stress. However, the 
progenies from the crosses between drought tolerant and farmers’ susceptible varieties were 
evaluated under well-watered and drought stressed conditions. Three outstanding 
combinations, Beledi A × Dan lla, Titinwa A × IT93K−503−1, and Beledi C × Mouride, with 
desirable, significant and positive SCA effects under drought stress were identified. The 
negative SCA effects detected in Titinwa A × Dan lla; Beledi A × IT93K−503−1, Laduni 1B 
× Mouride and Beledi C × Dan lla crosses suggested that they were early maturing.  
 
152 
 
University of Ghana http://ugspace.ug.edu.gh
 
Drought stress is highly heterogeneous in time and space and is unpredictable. This trait 
makes it hard to identify representative drought stress conditions. Environmental influence is 
associated with drought stress since phenotype is the ultimate expression of the interaction 
between genotype and environment. The unpredictable and variable forms in which drought 
stress manifests itself make the selection of individual genotypes and breeding for drought 
tolerance difficult. Multiple stresses occur, as was the case in Nyankpala (Tamale), where 
drought and heat stress were difficult to separate. The results of stability analysis revealed 
that BC×M, BC×I, L1B×I, TA×M and A1B×M were the highest yielding genotypes across 
the test environments. Determining useful selection indicators (traits) for yielding ability, 
drought tolerance and earliness in cowpea may be useful in varietal selection. These 
genotypes would be further assessed for earliness and drought tolerance and released to 
smallholder farmers growing cowpea in drought-prone areas. 
 
 
 
 
 
 
 
 
 
153 
 
University of Ghana http://ugspace.ug.edu.gh
 
7.2 Recommendations 
• Heat tolerance should be considered as an important component when breeding 
cowpea for tolerance to drought.  
• The four early-maturing and drought-tolerant genotypes exhibiting high and stable 
yields would be further evaluated in the farmers' fields to validate their performance 
for possible release.  
• A genome-wide association study on the cowpea populations used in this study would 
be useful to identify loci associated with drought tolerance.  
• Genomic data (SNPs) available from this study could be used for both background 
selection and diversity studies.  
 
 
 
 
 
 
 
 
 
 
 
 
 
154 
 
University of Ghana http://ugspace.ug.edu.gh
 
BIBLIOGRAPHY AND APPENDICES 
BIBLIOGRAPHY  
 
Abney, M., McPeek, M.S., and Ober, C. (2001). Broad and Narrow Heritabilities of 
Quantitative Traits in a Founder Population, American Journal of Human Genetics, 68, 
1302–1307. 
Acquaah, G. (2012). Principles of Plant Genetics and Breeding. Second Edition. John Wiley  
and Sons, LTD., Publication. 
Adewale, B.D., Adeigbe, O.O., and Aremu, C. (2011). Genetic distance and Diversity among 
some Cowpea (Vigna unguiculata L. Walp) genotypes. International Journal of 
Research in Plant Science, 1(2), 9–14. 
Adeyanju, A.O., and Ishiyaku, M. (2007). Genetic Study of Earliness in Cowpea (Vigna 
unguiculata L. Walp) Under Screen House Condition, International Journal of Plant 
Breeding and Genetics 1(1), 34–37. 
Adu-Dapaah, H., Singh, B.B., Chheda, H.R., and Fatokun, C.A. (1988). Heterosis and 
Inbreeding Depression in Cowpea. Tropical Grain Legume Bulletin, 35, 23−27. 
Afshari, M., Shekari., F, Azimkhani, R., Habibi, H and Fotokian, M.H. (2013). Effects of 
Foliar Application of Salicylic Acid on Growth and Physiological aAttributes of 
Cowpea under Water Stress Conditions, Iran Agricultural Research 32(1), 55–69. 
Agbicodo, E.M. (2009) Genetic Analysis of Abiotic and Biotic Resistance in Cowpea [ Vigna 
unguiculata ( L .) Walp .] PhD Thesis. ISBN 9789085854777 
Agbicodo, E.M., Fatokun, C.A., Muranaka, S., Visser, R.G.F., and Linden van der, C. G. 
(2009). Breeding Drought Tolerant Cowpea: Constraints, Accomplishments, and Future 
Prospects, Euphytica, 167, 353–370. 
Agyeman, K., Berchie, J.N., Tetteh Nartey, E., and Fordjour, J.K. (2014). Growth and Yield 
Performance of Improved Cowpea (Vigna unguiculata (L) Walp) Varieties in Ghana. 
Agricultural Science, 2(4), 44–52. 
Ahmed, F.E., and Suliman, A.S.H. (2010). Effect of Water Stress Applied at Different Stages 
of Growth on Seed Yield and wWater-use Efficiency of Cowpea.Agriculture and 
Biology of North America 1(4), 534−540. 
Ahsan, M. Z., Majidano, M.S., Bhutto, H., Soomro, A.W., Panhwar, F.H., Channa, A.R., and 
Sial, K.B. (2015). Genetic Variability, Coefficient of Variance, Heritability and Genetic 
advance of some Gossypium hirsutum L. Accessions, Journal of Agricultural Sciences 
7(2), 147–151.  
Ajeigbe, H. A., Ihedioha, D., and Chikoye, D. (2008). Variation in Physico-chemical 
Properties of Seed of Selected Improved Varieties of Cowpea as it Relates to Industrial 
Utilization of the Crop. African Journal of Biotechnology, 7(20), 3642–3647.  
 
155 
 
University of Ghana http://ugspace.ug.edu.gh
 
Ali, Z.B., Yao, K.N., Odeny, D.A., Kyalo, M., Skilton, R., and Eltahir, M. (2015). Assessing 
the Genetic Diversity of Cowpea (Vigna unguiculata L . Walp ) Accessions from Sudan 
using Simple Sequence Repeat (SSR) Markers. African Journal of Plant Science 9(7), 
293–305. 
Alidu, M.S, Akromah, R., and Atokple, I.D.K. (2013). Genetic Analysis of Vegetative-stage 
Drought Tolerance in Cowpea. Greener Journal of Agricultural Sciences, 3(6), 476–
491. View at Publisher · 
Aliyu, O.M., and Makinde, B.O. (2016). Phenotypic Analysis of Seed Yield and Yield 
Components in Cowpea (Vigna unguiculata L., Walp). Plant Breesing and 
Biotechnology, 4(2), 252–261. 
Alvarado, G., Lopez, M., Vargas, M., Pacheco, A., Rodriguez, F., Burguerio, J., and Crossa, 
J. (2015). META-R (Multi-Environment Trial Analysis with R for Windows), Version 
6.03. hdl:11529/10201, CIMMYT Research Data and Software Repository Network, 
V21. 
Andrews, S. (2010). FastQC: a Quality Control Tool for High Throughput Sequence Data. 
Available online at: http://www.bioinformatics.babraham.ac.uk/projects/fastqc 
Araus, J.L. (2002). Plant Breeding and Drought in C3 Cereals: What Should We Breed 
For?, Annals of Botany, 89(7), 925–940. 
Aremu, C.O, Adebayo, M.A, Ariyo, O.J, and Adewale, B.B. (2007). Classification of Genetic 
Diversity and Choice of Parents for Hybrisation in Cowpea. African Journal of 
Biotechnology, 6(20), 2333–2339. 
Aremu, C.O. (2011). Exploring Statistical Tools in Measuring Genetic Diversity for Crop 
Improvement. 339–348. ISBN: 9789530185-7, InTech 
Asare, A.T., Akrong, C.K., Gowda, B.S., Galyuon, I.K.A., and  Aboagye, L.M. (2011). 
Evaluation of Agro-morphological Diversity in Some Segregating Lines of Cowpea ( 
Vigna unguiculata L . Walp ), Journal of Ghana Science Association, 13(2), 1–11. 
Ashraf, M. (2010). Inducing Drought Tolerance in Plants: Recent Advances. Biotechnology 
Advances, 28, 169–183. 
Ashraf, M., Athar, H.R., Harris, P.J.C., and Kwon, T.R. (2008). Some Prospective Strategies 
for Improving Crop Salt Tolerance. Advance Agronomy, 97, 45–110. 
Ayo-Vaughan, M. A., Ariyo, O.J., Daniel, I.O., and Alake, C.O. (2013). Combining Ability 
and Genetic Components for Pod and Seed Traits in Cowpea Lines, Italian Journal of 
Agronomy, 8 (2), 10, 73–78. 
Babu, R.C., Nguyen, B.D., Chamarerk, V.P., Shanmugasundaram, P., Chezhian, P., 
Jeyaprakash, S.K., Ganesh, A., Palchamy, S., Sadasivam, S., Sarkarung, S., Wade, L.J., 
and Nguyen, H.T. (2003). Genetic Analysis of Drought Resistance in Rice by 
Molecular Markers, Crop Science, 43, 1457–1469. 
156 
 
University of Ghana http://ugspace.ug.edu.gh
 
Backiyarani, S., Nadarajan, N., Rajendran, C., and Shanthi, S. (2000). Genetic Divergence for 
Physiological Traits in Cowpea, (Vigna unguiculata. (L.) Walp). Legume Research, 
23(2), 114–117. 
Bahar, M., and Yildirim, M. (2010). Heat and Drought Resistances Criteria in Spring Bread 
Wheat: Drought Resistance Parameters. Scientific Research and Essays, 5(13), 1742–
1745. 
Bates, L.M., and  Hall, A.E. (1981). Stomatal Closure with Soil Water Depletion not 
Associated with Changes in Bulk Leaf Water Status, Oecologia, 50(1), 62–65. doi: 
10.1007/BF00378794. 
Baye, T.M., Abebe, T., and Wilke, R.A. (2011). Genotype by Environment Interaction and 
their Translational Implications. Personalised Medicine 8(1), 59–70. 
Becerra, V., Paredes, M., Ferreira, M.E., Gutierrez, E., and Diaz, L.M. (2017). Assessment of 
the Genetic Diversity and Population Structure in Temperate Japonica Rice Germplasm 
used in Breeding in Chile. Chilean Journal of Agriculture, 77(1), 15–26. 
Beebe, S.E., Rao, I.M, Blair, M.W., and Acosta-Gallegos, J.A. (2013). Phenotyping Common 
beans for Adaptation to Drought. Frontiers Physiology, 4, 1–20. 
Belko, N., Zaman-Allah, M., Diop, N.N., Cisse, N., Zombre, G., Ehlers, J.D., and Vadez, V. 
(2013). Restriction of Transpiration Rate Under High Vapour Pressure Deficit and Non-
limiting Water conditions is Important for Terminal Drought Tolerance in Cowpea. 
Plant Biology 15 304–316.  
Bernardo, R. (2002). Breeding for Quantitative Traits in Plants. Second Edition, University of 
Minnesota, Twin-Cities, ISBN 978-0-9720724-1-0. 
Bernier, J., Kumar, A., Serraj, R., Spaner, D., and Atlin, G. (2008). Review: Breeding Upland 
Rice for Drought Resistance. Journal of Science Food Agriculture, 88, 927–939. 
Bertan, I., de Carvalho, F.I.F, and de Oliveira, A.C. (2007). Parental Selection Strategies in 
Plant Breeding Programs. Journal of Crop Science and Biotechnology, 10(4), 211−222. 
Blum, A. (1988). Drought Resistance: Plant Breeding for Stress Environments. Boca Raton, 
FL: CRC Press  
Blum, A. (2011). Plant breeding for water limited environments. Springer-Verlag, New York. 
252.  
Brooks, K. (2013). Farming is the Key to Solving Youth Unemployment in Africa. 
IFPRI.http://www.ifpri.org/blog/farming-key-solving-youth-unemployment-
fricaByerlee, D. and Bernstein J. September 2013. Feed the Future Learning Agenda 
Literature. 
Bruce, W.B., Edmeades, G.O., and Barker, T.C. (2002). Molecular and Physiological 
Approaches to Maize Improvement for Drought Tolerance. Journal of Experimental 
Botany, 53 (366), 13–25. 
157 
 
University of Ghana http://ugspace.ug.edu.gh
 
Buerkle, C.A., and Lexer, C. (2008). Admixture as the Base for Genetic Mapping. Trends 
Ecology of Evolution, 23(12), 686–694. 
Casanoves, F., Baldessari, J., and Balzarini, M. (2005). Evaluation of Multi-environment 
Trial of Peanut Cultivars. Crop Science, 45(1), 15–26.  
Cattivelli, L., Rizza. F., Badeck, F.W., Mazzucotelli, E., Mastrangelo, A.M, Francia, E., 
Mare, C., Tondelli, A., Stanca, A.M. (2008). Drought tolerance improvement in crop  
plants: An integrative view from breeding to genomics, Field Crop Researrch. 105, 1–
14. 
Chahal, G.S,. and Gosal, S.S. (2002). Principles and Procedures of Plant Breeding: 
Biotechnology and Conventional A pproaches. Narosa Publishing House, New Delhi. 
Charles O.A., Lashermes, P., and Rouslot, P. (1997). Identification of RAPD Markers for 
Resistance to Coffee Berry Disease, Colletotrìchuin kahawae in Arabica Coffee. 
Euphytica, 97, 241–248. 
Chaves, M.M., Maroco, J.P., and Pereira, J.S. (2003). Understanding Plant Responses to 
Drought from Genes to the whole Plant. Functional Plant Biology, 30, 239–26. 
Chiulele, R.M. (2010). Breeding Cowpea (Vign unuiculata (L) Walp) for Improved Drought 
Tolerance in Mozambique. PhD Thesis (unpublished) submitted to African Centre for 
Crop Improvement, School of Agricultural Sciences and Agribusiness, Faculty of 
Science and Agriculture, University of KwaZulu-Natal, Republic of South Africa.  
Cláudio, F., Lúcia, R., Gomes, F., Rodrigues, F., and Filho, F. (2003). Genetic Control of 
Cowpea Seed Sizes, Scientia Agricola 60 (2), 315–318. 
Cobbinah, F.A., and Asante, I.K. (2011). Characterization , Evaluation And Selection of 
Cowpea ( Vigna unguiculata ( L .) Walp ) Accessions With Desirable Traits From Eight 
Regions of Ghana, Journal of Agriculural and Biological Science, 6(7), 21–32. 
Collard, B.C.Y., and Mackill, D.J. (2008). Marker-assisted Selection: An Approach for 
Precision Plant Breeding in the Twenty-first Century. Philosophical Transaction of 
Royal Society of Biology. 363 (1491), 557−572. doi:10.1098/rstb.2007.2170 
Condon, A.G., Richards, R.A., Rebetzke, G.J., and Farquhar, G.D. (2004). Breeding for High 
Water-use Efficiency, Journal of Experimental  Botany, 55, 2447–2460. 
Costa, L.D., Vedove, G.D., Gianquinto, G., Giovanardi, R., and Peressotti, A. (1997). Yield, 
water use efficiency and nitrogen uptake in potato: influence of drought stress, Potato 
Research. 40, 19–34. 
Coulibaly, S. Pasquet, R.S., Papa, R., and Gepts. P. (2002). AFLP Analysis of the Phenetic 
Organization and Genetic Diversity of (Vigna unguiculata L . Walp). Reveals Extensive 
Gene Flow between Wild and Domesticated Types, Theoritcal and Applied 
Genetics,104(1-2),  358–366. 
Cox, D.R. (1997). Introduction to Yates and Cochran (1938). The Analysis of Groups of 
Experiments. In: Kotz S., Johnson, N.L (eds). Breakthroughs in Statistics. Springer 
Series in Statistics (Perspectives in Statistics). Springer, New York, NY. 
158 
 
University of Ghana http://ugspace.ug.edu.gh
 
Craufurd, P. Q., Subedi, M., and Summerfield, R. J. (1997). Leaf Appearance in Cowpea: 
Effects of Temperature and Photoperiod, Crop Science, 37(131), 167–171. 
Crossa, J. (1990). Statistical Analyses of Multilocation Trials. Advances in Agronomy 44, 55–
85. 
Crossa, J., Vargas, M., Cossani, M., Alvardo, G., Burgueno, J., Mathews, K.L., and 
Reynolds, M.P.  (2015). Evaluation and Interpretation of Interactions. Advances in 
Agronomy 102 (2), 736–747.  
Cruz C.D., Regazzi, A.J., and Careiro, P.C.S. (2012). Biometric Models Applied to Genetic 
Improvement. 4th Edition. Viçosa, MG: UFV,  480. 
Cruz, C.D., and Reg azzi, A.J. (2001). Biometric Models Applied to Genetic Improvement. 
2nd Edition, UFV, Vicosa. 
Danecek, P., Auton, A., Abecasis, G., Cornelis, A. A., Banks, E., DePristo, M.A., Handsaker, 
R.E., Lunter, G., Marth, G.T., Sherry, S.T., McVean, G., Durbin, R.,  and 1000 
Genomes Project Analysis Group. (2011). The variant call format and VCFtools. 
Bioinformatics, 27 (15), 2156–2158. 
Danquah, E., and  Blay, E. (1999). Breeding for Stress Tolerance: Drought as a Case 
Study, Ghana Journal of Agricultural Science, 32, 229–236.  
Davies, W.J., and Zhang, J. (1991). Root Signals and the Regulation of Growth and 
Development of Plants in Drying Soil. Annual Review of Plant Physiology and Plant 
Molecular Biology, 42, 55−76. 
de Almeida, W.S., Ronald F.B.F., Teofilo, E.M., and de Magalhaes, B.C.H. (2014). 
Correlation and Path Analysis in Components of Grain Yield of Cowpea Genotypes’, 
Revista Ciencia Agronomica, 45(4),  726–736. 
de Souza, P.I., Egli, D., and Bruening, W. (1997). Water Stress during Seed Filling and Leaf 
Senescence in Soybean. Agronomy Journal, 89, 807–812. 
Derera, J. (2005). Genetic Effects and Associations between Grain Yield Potential, Stress 
Tolerance and Yield Stability in Southern African Maize (Zea mays L.). Base 
Germplasm. A PhD Thesis in Plant Breeding (unpublished), African Centre for Crop 
Improvement (ACCI), School of Biochemistry, Genetics, Microbiology and Plant 
Pathology, Faculty of Science and Agriculture, University of KwaZulu-Natal, Republic 
of South Africa.  
Devasirvatham, V., and Tan, D.K.Y. (2018). Impact of High Temperature and Drought 
Stresses on Chickpea Production. Agronomy, 8(8), 1–9. 
Dia, M., Wehner, T.C., Elmstrom, G.W., Gabert, A., Motes, J.E., Staub, J.E., and Widders, I. 
E. (2018). Genotype × Environment Interaction for Yield of Pickling Cucumber in 24 
U.S. Environments. Open Agriculture, 3(1), 1–16.  
Dias, F.T.C., da Silva,A.P.M., and  de Magalhães,C.H.C.B (2009). Genetic Divergence in 
Cowpea Genotypes with Upright Growth and Early Cycle. Crop Breeding and Applied 
Biotechnology. 253–259. 
159 
 
University of Ghana http://ugspace.ug.edu.gh
 
Donald, C.M., and Hamblin, J. (1976). The Biological Yield and Harvest Index of Cereals as 
Agronomic and Plant Breeding Criteria. Advances in Agronomy, 28, 361–405. 
Doumbia, I.Z., Akromah, R., and Asibuo, J.Y. (2014). Assessment of Cowpea Germplasms 
from Ghana and Mali Using Simple Sequence Repeat (SSR ) Markers, International 
Journal of Agriculture and Forestry, 4(2), 118–123.  
Earl, D.A., and von Holdt, B.M. (2012). STRUCTURE HARVESTER: A website and 
Program for Visualizing STRUCTURE Output and Implementing the Evanno Method. 
Conservation Genetics Resources, 4(2), 359–361. https://doi.org/10.1007/s12686-011-
9548-7 
Earl, H.J., and Davis, R.F. (2003). Effect of Drought Stress on Leaf and Whole Canopy 
Radiation Use Efficiency and Yield of Maize. Agronomy Journal. 95, 688– 696. 
Eberhart, S.A., and Russell, W. A. (1966). Stability Parameters for Comparing Varieties. 
Crop Science, 6(1), 36–40.  
El Naim, A.M., El, Jabereldar, A.A., Ahmed, S.E., Ismaeil, F. M., and Ibrahim, E.A. (2012). 
Determination of Suitable Variety and Plants per Stand of Cowpea ( Vigna Unguiculata 
L . Walp) in the Sandy, Advances in Life Sciences 2(1), 1–5.  
Ellis, R.H., Lawn, R.J., Summerfield, R.J., Qi, A., Robert, E.H., Chay, P.M., Brouwer, J.B., 
Rose, J.L., Yeates, S.J. and Sandover, S. (1994).   Towards the Reliable Predication of 
Time to Flowering in Six Annual Crops iv. Cultivated and Wild Mungbean. 
Experimental Agriculture, 30, 31–41.  
Evanno, G., Regnaut, S., and Goudet, J. (2005). Detecting the Number of Clusters of 
Individuals using the Software STRUCTURE: A Simulation Study. Molecular Ecology, 
14(8), 2611–2620.  https://doi.org/10.1111/j.1365-294X.2005.02553.x 
Fahad, S., Bajwa, A.A., Nazir, U, Anjum, S.A., Farooq, A., Zohaib, A., Sadia, S., Nasim, W., 
Adkins, S., Saud, S., Ishan, M.Z., Alharby, H., Wu, C., Wang, D., and Huang, J. (2017). 
Crop Production under Drought and Heat Stress: Plant Responses and Management 
Options. Frontier in Plant Science, 8(1147), 1−17. doi: 10.3389/fpls.2017.01147 
Falconer, D.S. (1989). Introduction to Quantitative Genetics, 3rd Edition. London, New York. 
Falconer, D.S., and Mackay, T.F.C. (1996). Introduction to Quantitative Genetics, 4th Edition. 
London: Pearson Prentice Hall. 
Fang, Y., and Xiong, L. (2015). General Mechanisms of Drought Response and their 
Application in Drought Resistance Improvement in Plants. Cellular and Molecular Life 
Sciences, 72, 673−689. doi: 10.1007/s00018-014-1767-0 
FAOSTAT. (2016). Food and Agriculture Organization (FAO), sub-Saharan Africa Cowpea 
Production. Available at: http://faostat3.fao.org/browse/Q/QC/E (accessed 23 December 
2018). 
 
160 
 
University of Ghana http://ugspace.ug.edu.gh
 
Farooq, M., Gogoi, N., Barthakur, S., Baroowa, B., Bharadwaj, N., Alghamdi, S.S., and 
Siddique, K.H.M. (2017). Drought Stress in Grain Legumes during Reproduction and 
Grain Filling. Journal of Agronomy and Crop Science. 203, 81−102. 
https://doi.org/10.1111/jac.12169 
Farooq, M., Wahid, A., Kobayashi, N, Fujita, D and Basra, S.M.A. (2009). Plant Drought 
Stress : Effects, Mechanisms and Management. Agronomy for Sustainable 
Development, Springer verlag/EDP Sciences/INRA, 29(1), 185–212. 
Fathi, A., and Tari, D.B. (2016). Effect of Drought Stress and its Mechanism in 
Plants, International Journal of Life Sciences 10, (1), 1–6.  
Fatokun, C.A., and  Ng, Q. (2007). Outcrossing in Cowpea. Journal of Food, Agriculture and 
Environment Vol.5, 5(3 & 4), 334–338. 
Fatokun, C.A., Boukar, O., and Muranaka, S. (2012) Evaluation of Cowpea (Vigna 
unguiculata (L.) Walp.) Germplasm Lines for Ttolerance to Drought, Plant Genetic 
Resources: Characterisation and Utilisation, 10(3), 171–176. 
Fehr, W.R. (1993). Principles of Cultivar Development. Vol. 1. Macmillian Publishing 
Company, New York, 536. 
Fieldscout TDR 150 Soil Moisture Meter. Spectrum Technologies, Inc 
Finlay, K.W., and Wilkinson, G.N. (1963). The Analysis of Adaptation in a Plant-Breeding 
Programme The ability of some crop varieties to perform well over a wide range of 
environmental conditions has lcng been appreciated by the agronomist and plant 
breeder . In the cereal belts of southern Au. Australian Journal of Agrieultural 
Research, 14, 742–754. 
Fischer, R., and Maurer, R. (1978). Drought Resistance in Spring Wheat Cultivars. Grain 
Yield Responses, Australian Journal of Agricultural Research, 29(4), 897–912. 
Francis, T.R., and Kannenberg, L.W. (1978). Yield Stability Studies in Short Season Maize: 
A Descriptive Method for Grouping Genotypes. Canadian Journal of Plant Science 58 
(4), 1029–1034. 
Frutos, E., Purificación, M.,  Galindo, P., and Leiva, V. (2014). An Interactive Biplot 
Implementation in R for Modeling Genotype-by-Environment Interaction. Stochastic 
Environmental Research and Risk Assessment 28 (7), 1629–1641. 
Fukai, S., and Cooper, M. (1995) Development of Drought-Resistant Cultivars Using 
Physiomorphological Traits in Rice, Field Crops Research, 40(2), 67–86. 
Gabriel, K R. (1971). The Biplot Graphic Display of Matrices with Application to Principal 
Component Analysis. Biometrika, 58(3), 453–67. 
Garrison, E., and Marth, G. (2012). Haplotype-based Variant Detection from Short-read 
Sequencing. arXiv, 476 arXiv:1207.3907. 
Gauch, H.G. (1992). Statistical Analysis of Regional Trials: AMMI Analysis of Factorial 
Designs. Elsevier, Amsterdam, Netherlands.  
161 
 
University of Ghana http://ugspace.ug.edu.gh
 
Gauch, H.G., and  Zobel, R.W. (1988). Predictive and Postdictive Success of Statistical 
Analyses of Yield Trils. Theortical Applied Genetics, 76, 1–10.  
Gauch, H.G., Piepho, H.P., and Annichiarico, P. (2008). Statistical Analysis of Yield Trials 
by AMMI and GGE: Further Considerations. Crop Science, 48, 866−889. doi: 
10.2135/cropsci2007.09.0513 
Goodnight, C.J. (1995). Epistasis and the Increase in Additive Genetic Variance: Implications 
for Phase 1 of Wright’s Shifting-Balance Process, Evolution International Journal of 
Organic Evolution, 49(3), 502–511. 
Gosal, S.S., Wani, S.H., and  Kang, M.S. (2009). Biotechnology and Drought 
Tolerance, Journal of Crop Improvement, 23(1), 19–54. 
Griffings, B. (1956). Concept of General and Specific Combining Ability in Relation to 
Diallel Crossing System. Australian Journal of Biological Science, 9 463−493. 
Hall, A. E. (2012). Phenotyping Cowpeas for Adaptation to Drought uses of Cowpeas. 
frontiers in Physiology, 3, 1–8.  
Hall, A.E., Cisse, N., Thiaw, S., Elawad, H.O.A, Ehlers, J.D., Ismail, A.M., Fery, R.L., 
Roberts, P.A., Kitch, L.W., Murdock, L.L., Boukar, O., Phillips, R.D. and McWatters, 
K.H. (2003). Development of Cowpea Cultivars and Germplasm by the Bean/Cowpea 
CRSP. Field Crops Research, 82, 103–134.   
Hamidou, F., Zombre, G., Diouf, D., Diop, N.N., Guinko, S., and Braconnier, S. (2007). 
Physiological, Biochemical and Agromorphological Responses of Five Cowpea (Vigna 
unguiculataL.) Walp.), Genotypes to Water Deficit under Glasshouse Conditions. 
Biotechnology, Agronomy,  Society and Environment. 11, 225–234. 
Haruna, I.M., and Usman, A. (2013). Agronomic Efficiency of Cowpea Varieties (Vigna 
unguiculata (L) Walp) under Varying Phosphorus Rates in Lafia, Nasarawa State, 
Nigeria. Asian Journal of Crop Science, 5, 209–215. doi: 10.3923/ajcs.2013.209.215 
Hazel, L. N. (1943) The Genetic Basis for Constructing Selection Indexes, Genetics, 28 (6), 
476–490. 
Hinkossa, A., Gebeyehu, S., and Zeleke, H. (2013). Generation Mean Analysis and 
Heritability of Drought Resistance in Common Bean (Phaseolus vulgaris L.). African 
Journal of Agricultural Research, 8(15), 1319–1329. 
Hongyu, K., Garcian-Pen, M., de Araujo, L.B., and Dias, C.T.S. (2014). Statistical Analysis 
of Yield Trials by AMMI Analysis of Genotype × Environment Interaction. Biometrical 
Letters, 51(2), 89–102. 
Huynh, B., Close, T. J., Roberts, P.A., Hu, Z., Wanamaker, S., Lucas, M. R., and Ehlers, J.D. 
(2013). Gene Pools and the Genetic Architecture of Domesticated Cowpea, The Plant 
Genome 6(3) 1–8.  
IBM Corp. Released (2013). IBM SPSS Statistical for Windows, Version 22.0, IBM Corp 
Armonk, NY.  
162 
 
University of Ghana http://ugspace.ug.edu.gh
 
Ikram, M. (2004). Inheritance of Earliness in Cowpea (Vigna unguiculata (L) Walp). Journal 
of Agricultural Research, 42(1), 1−12. 
Ishiyaku, M.F., Singh, B.B. and Craufurd, P.Q. (2005). Inheritance of Time to Flowering in 
Cowpea ( Vigna unguiculata ( L .) Walp ). Euphytica 142(3), 291–300.  
Jackai, L.E.N. (1995). Integrated Pest Management of Borers of Cowpea and Beans. 
International Journal of Tropical Insect Science, 16(3–4), 237–250.  
https://doi.org/10.1017/S1742758400017240 
Jain, S., and Khare, D. (2002). Characterization of Mung Bean Varieties for Verification of 
Genetic Purity at Plant Level. JNKVV Research  Journal., 36 (1&2), 38–43. 
Jakobsson, M., Scholz, S.W., Scheet, P., et al. (20 co-authors). (2008). Genotype, Haplotype 
and Copy-number Variation in Worldwide Human Populations. Nature, 451, 998–1003. 
Johnson, H.W., Robinson, H.F., and Comstock, R.E. (1955). Estimates of genetic and 
Environmental Variability in Soybean. Agronomy Journal, 47(7), 314–318. 
Karampatakis, T., Geladari, A., Politi, L., Antachopoulos, C., Iosifidis, E., Tsiatsiou, O., 
Karyoti, A., Papanikolaou, V., Tsakris, A and Roilides, E. (2017). Cluster-
distinguishing Genotypic and Phenotypic Diversity of Carbapenem-resistant Gram-
negative Bacteria in Solid-organ Transplantation Patients: A Comparative Study. 
Journal of Medical Microniology, 66, 1158–1169.  
Karkannavar, J.C., Venugopal, R., and Goud, J. V. (1991). Inheritance and Linkage Studies 
in Cowpea [Vigna unguiculata (L.) Walp.]. Indian Journal of  Genetics., 51,  203–207. 
Kazan, K., and  Lyons, R. (2016). The Link between Flowering Time and Stress Tolerance. 
Journal of Experimental Botany 67(1), 47–60. 
Kearsey, M.J., and Pooni, H.S. (1996). Th e Genetical Analysis of Quantitative Traits. 
London, Chapman and Hall. 
Khan, H.R., Paull, J.G., Siddique, K.H.M., and Stoddard, F.L. (2010). Faba Bean Breeding 
for Drought Affected Environments: A Physiological and Agronomic Perspective.  
Field Crops Research, 115, 279–286. 
Krueger F. (2015). Trim Galore! http://www.bioinformatics.babraham.ac.uk 
/projects/trim_galore 
Lachyan, T.S., and Dalvi, S.S. (2013). Inheritance Study of Qualitative and Quantitative 
Traits in Cowpea (Vigna unguiculata (L) Walp). International Journal of Science and 
Research, 4(4), 2170 −2173 
Lafitte, H.R., Ismail, A., and Bennett, J. (2004). Abiotic Stress Tolerance in Rice for Asia: 
Progress and the Future, 1–17. 
Levitt, J. (1972). Responses of Plants to Environmental Stresses. New York, NY: Academic 
Press. 698. 
Lin, C.S. (1985). Stability Analysis: Where Do We Stand?l, (1). 
163 
 
University of Ghana http://ugspace.ug.edu.gh
 
Lorens, G.F., Bennett, J.M., and Loggale, L.B. (1987). Differences in Drought Resistance 
between Two Corn Hybrids. II. Component Analysis and Growth Rates1, Agronomy 
Journal 79 (5), 808–813. 
Ludlow, M., and Muchow, R. (1990). A Critical Evaluation of Traits for Improving Crop 
Yields in Water-Limited Environments, Advances in Agronomy 43, 107–153. 
Lynch, M., and Walsh, B. (1998). Genetic and Analysis of Quantitative Traits. Sinauer 
Associates, Sunderland, Mass. 
Mai-Kodomi, Y., Singh, B., Myers, O., Yopp, J., Gibson, P. and Terao, T. (1999a). Two 
Mechanisms of Drought Tolerance in Cowpea. The Indian Journal of Genetics and 
Plant Breeding, 59, 309–316. 
Mai-Kodomi, Y., Singh, B.B., Terao, T., Myers Jr, O., Yopp, J.H, and Gibson, P.J. (1999b). 
Inheritance of Drought Tolerance in Cowpea. The Indian Journal of Genetics and Plant 
Breeding, 59 (3), 317–332. 
Marame, F., Desalegne, L., Fininsa, C., and Sigvald, R. (2009). Genetic Analysis for Some 
Plant and Fruit Traits, and its Implication for a Breeding Program of Hot Pepper 
(Capsicum annuumvar.annuum L.), Hereditas, 146(4), 131–140. 
Maréchal, R., Mascherpa, J.M., Stainier, F., and Marechal, R. (1978). Combinaisons et noms 
nouveaux dans les genres Phaseolus, Minkelersia, Macroptilium, Ramirezella et 
Vigna, 27, 199. 
Mariani, L., and Ferrante, A. (2017). Agronomic Management for Enhancing Plant Tolerance 
to Abiotic Stresses Drought, Salinity, Hypoxia, and Lodging. Horticulturae; 3(52), 1–
18. 
Martinez, C., Pons, E., Prats, G., and Leon, J. (2003). Salicylic Acid Regulates Flowering 
Time and Links Defense Responses and Reproductive Development. Plant Journal, 36, 
209–217. 
Matsui, T., and Singh, B. (2003). Root Characteristics in Cowpea Related to Drought 
Tolerance at the Seedling Stage. Experimental Agriculture, 39, 29–38. 
McTavish, E.J, and Hills, D.M. (2014). A Genomic Approach for Distinguishing between 
Recent and Ancient Admixture as Applied to Cattle.  Journal of Heredity, 105(4), 
445−456. 
Meseka, S., Menkir, A., Olakojo, S., Jalloh, A., Coulibaly, N., Bossey, O. (2016). Yield 
Stability of Yellow Maize Hybrids in the Savannas of West Africa, Agronomy Journal, 
108, 1313– 1320. 
Mikheenko, A., Prjibelski, A., Saveliey, V., Antipoy, D., and Gurevich, A. (2018). Versatile 
Genome Assembly Evaluation with QUAST-LG. Bioinformatics, 34, 145−150. doi: 
10.1093/bioinformatics/bty266 
Miklas, P.N., Kelly, J.D., Beebe, S.E., and Blair, M.W. (2006).  Common Bean Breeding for 
Resistance Against Biotic and Abiotic Stresses: From Classical to MAS Breeding. 
Euphytica, 147, 105–131. 
164 
 
University of Ghana http://ugspace.ug.edu.gh
 
Morgante, M., and Salamini, F. (2003). From Plant Genomics to Breeding Practice, Current 
Opinion in Biotechnology, 14(2), 214–219. 
Muchero, W., Ehlers, J.D., and Roberts, P.A. (2008). Seedling Stage Drought-Induced 
Phenotypes and Drought-Responsive Genes in Diverse Cowpea Genotypes, Crop 
Science, 48(2), 541–552. 
Muchero, W., Roberts, P.A., Diop, N.N., Drabo, I., and  Cisse, N. (2013). Genetic 
Architecture of Delayed Senescence , Biomass , and Grain Yield under Drought Stress 
in Cowpea, PLOS ONE 8(7). 1–10. 
Nduwumuremyi, A, Tongoona, P and Habimana, S. (2013). Mating Designs: Helpful Tool 
for Quantitative Plant Breeding Analysis. Journal of Plant Breeding and Genetics 1, 
117–129.  
Ng, N.Q. (1995). Cowpea Vigna unguiculata (Leguminosea Papilionoidea). In Smartt J, 
Simmonds N (eds). Evolution of Crop Plants. Longman, London, 326–332. 
Ngalamu, T., Meseka, S., Tongun, N., Oliwai, M.O., and Newton W.O. (2017). Challenges 
and Opportunities of Developing Cowpea for a Food and Nutrition Secure South Sudan. 
Poster presentation at the Third Global Food Security Conference, Cape Town, South 
Africa. 3−6 December. 
Ngalamu, T., Odra, J., and Tongun, N. (2015). Baseline Survey Report on Farmerss Cowpea 
Preferrence in Greater Equatoria Region of South Sudan (Cowpea Breeding Project 
Report, Unpublished).  
Ngalamu, T., Odra, James., and Tongun, J. (2014). Cowpea Production Handbook. First 
Edition, AFRISTAR International Ltd. 1–43. 
Nguyen, H., and Blum, A. (2004). Physiology and Biotechnology Integration for Plant 
Breeding. doi:10.1201/9780203022030.Chapter 16 
Nkouannessi, M. (2005). The Genetic, Morphological and Physiological Evaluation of 
AfricanCowpea Genotypes. Master dissertation (unpublished), University of the Free 
State Bloemfontein, South Africa. 
Nwofia, G.E., and Emeka, G. (2014). Inheritance of Leaf Shape, Pod Shape, Pod Colour and 
Seed Coat Colour in Cowpea (Vigna unguiculata (L.) Walp ). World Journal of 
Agricultural Sciences, 10(4), 178–184. 
Ogbonnaya, C.I., Sarr, B., Brou, C., Diouf, O., Diop, N.N., and Roy-Macauley, H. (2003). 
Selection of Cowpea Genotypes in Hydroponics, Pots, and Field for Drought 
Tolerance, Crop Science, 43, 1114–1120. 
Ojomo, O.A. (1971). Inheritance of Flowering Date in Cowpea (Vigna unguiculata (L) 
Walp). Tropical Agriculture (Trinidad), 48, 277–282. 
Olajide, A.A., and Ilori, C.O. (2017). Effects of Drought  on Morphological Traits in Some 
Cowpea Genotypes by Evaluating Their Combining Abilities, Advances in Agriculture, 
2017, 1–10. 
165 
 
University of Ghana http://ugspace.ug.edu.gh
 
Omoigui, L.O., Ishiyaku, M.F., Kamara, A.Y., and Alabi, S.O. (2006). Genetic Variability 
and Heritability Studies of some Reproductive Traits in Cowpea ( Vigna unguiculate ( L 
.) Walp), African Journal of Biotechnology, 5(13), 1191–1195. 
Othman, S.A., Singh, B.B., and Mukhtar, F.B. (2006). Studies on the Inheritance Pattern of 
Joints, Pods, and Flower Pigmentation in Cowpea [Vigna unguiculata (L.) Walp.]. 
African Journal of Biotechnology, 5(23), 2371–2376. 
Ouedraogo, N., Sanou, J., Kam, H., Hamidou, T., Adam, M., Gracen, V., and Danquah, E.Y. 
(2017). Farmers Perception on Impact of Drought and their Preference for Sorghum 
Cultivars in Burkina Faso. Agricultural Science Research Journal, 7(9), 277−284. 
Owusu, E.Y., Akromah, R., Denwar, N.N., Adjebeng-Danquah, J., Kusi, F., and Haruna, M. 
(2017). Inheritance of Early Maturity in Some Cowpea (Vigna unguiculata (L) Walp) 
Genotypes under Rain-fed Conditions in Northern Ghana.  Advances in Agriculture, 
2018, 1–10. https://doi.org/10.1155/2018/8930259 
Ozimati, A., Rubaihayo, P.R., Gibson, P., Edema, R., Kayondo, I.S, Ntare, B.R., and Okello, 
D.K. (2014). Inheritance of Resistance to Kernel Infection by Aspergillusflavus and 
Aflatoxin Accumulation in Groundnut. African Crop Science Journal, 2(1), 51–59. 
Pacheco, A., Vargas, M., Alvarado, G., Rodriguez, F., Crossa, J., and Burguerio, J. (2015). 
GEA-R (Genotype × Environment Analysis with R for Windows), Version 4.1. 
hdl:11529/10203, CIMMYT Research Data and Software Repository Network, V16. 
Padi, F. K. (2007). Early Generation Selection for High Yielding Cowpea Genotypes in 
Additive Series Intercropping Systems with Sorghum. Annals of Applied Biology 
151(3), 391–400. 
Padulosi, S., and Ng, N.Q. (1997). Origin, Taxonomy, and Morphology of Vigna unguiculata 
(L) Walp. In B.B. Singh, D.R. Mohan Raji and K.E. Dashiel. Advances in Cowpea 
Research. Ibadan, Nigeria, 1–12. 
Pandey, R.K., Herrera, W.A.T., Villegas, A.N., and Pendleton, J.W.  (1984). Drought 
Response of Grain Legumes Under Irrigation Gradient: III. Plant Growth1. Agronomy 
Journal 76 (4), 557–560. 
Passioura, J.B. (2002). Review: Environmental Biology and Crop Improvement, Functional 
Plant Biology, 29, 537–546. 
Patel, J. B., Patel, H., Sharma, S.C., and Acharya, S. (2016). Genetics of Seed Related 
Attributes in Cowpea [ Vigna unguiculata (L.) Walp.], Legume Research, 39(1), 1–6.  
https://doi.org/10.18805/lr.v39i1.8855 
Patil, H.E., and Navale, P.A. (2006). Combining Ability in Cowpea [Vigna unguiculata (L.) 
Walp.]. Legume Research, 29(4), 270–273. 
Perrier, X., Flori, A., and Bonnot, F (2003). Data Analysis Methods. In: Hamon, P., Seguin, 
M., Perrier, X, Glaszmann, J.C. Ed., Genetic Diversity of Cultivated Tropical Plants. 
Enfield, Science Publisher. Montpellier, 43−76. http://darwin.cirad.fr/darwin 
166 
 
University of Ghana http://ugspace.ug.edu.gh
 
Porebski, S., Bailey, L. G., and Baum, B.R. (1997). Modification of a Ctab Dna Extraction 
Protocol for Plants Containing. Plant Molecular Biology Reporter, 15(1), 8–15. 
Pottorff, M., Ehlers, J.D., Fatokun, C., Roberts, P. A., and Close, T.J. (2012). Leaf 
Morphology in Cowpea [Vigna unguiculata (L.) Walp]: QTL Analysis, Physical 
Mapping and Identifying a Candidate Gene using Synteny with Model Legume Species. 
BMC Genomics. https://doi.org/10.1186/1471-2164-13-234 
Pritchard, J.K., Stephens, M., and Donnelly, P. (2000). Inference of Population Structure 
Using Multi-locus Genotype Data. Genetics, 155(2), 945–959.  
Pungulani., L.L.M., James P.M, Williams., W. M., and Banda., M. (2013). Improvement of 
Leaf Wilting Scoring System in Cowpea (Vigna unguiculata (L) Walp.): From 
Qualitative Scale to Quantitative Index. Australian Journal of Crop Science, 
7(9):1262−1269. 
Puritz, J.B., Hollenbeck C.M., and Gold J.R. (2014). dDocent: a RADseq, Variant-calling 
Pipeline Designed for 528 Population Genomics of Non-model Organisms. Peer 
Journal, 2: e431; doi 10.7717/peerj.431. 
Rashidi, M., Farshadfar, E., and Jowkar, M.M. (2013). AMMI Analysis of Phenotypic 
Stability in Chickpea Genotypes over Stress and Non-Stress Environments. 
International Journal of Agriculture and Crop Sciences 5, 253–260. 
Rashwan, A.M.A. (2010). Estimation of some Genetic Parameters of Six Population of Two 
Cowpea. Asian Journal of Crop Science, 2(4), 267–267. 
Reiss, G.C., and Bailey, J.A. (1998). Striga Gesneriodes Parasitising Cowpea: Development 
of Infection Structures and Mechanisms of Penetration. Annals of Botany, 81, 431−440. 
Roberts, P.A., Mathews, W.C., and Ehlers, J.D. (2005). Nematode Resistant Cowpea Cover 
Crops in Tomato Production Systems. Agronomy Journal, 97, 1626–1635. 
Rodriguez, F., Alvarado, G., Pacheco, A., Crossa, J., and Burguerio, J. (2015). AGD-R 
(Analysis of Genetic Designs with R for Windows), Version 5.0. hdl:11529/10202, 
CIMMYT Research Data and Software Repository Network, V13. 
Rouf, R., Mainassara, M., Mir, R.R., Zaman-Allah, M., Sreenivasulu, N., Trethowan, R., and 
Varshney, R.K. (2012). Integrated Genomics, Physiology and Breeding Approaches for 
Improving Drought Tolerance in Crops. Theoretical and Applied Genetics, 125(4), 625–
645. https://doi.org/10.1007/s00122-012-1904-9 
Sabaghnia, N., Dehghani, H., and Sabaghpour, S.H. (2006). Nonparametric Methods for 
Interpreting Genotype × Environment Interaction of Lentil Genotypes, 46, 1100–1106. 
Salvi, S., and  Tuberosa, R. (2005). To Clone or not to Clone Plant QTLs: Present and Future 
Challenges. Trends Plant Science, 10, 297–304. 
Sarutayophat, T., Nualsri, C., Santipracha, Q., and Saereeprasert, V. (2007). Characterisation 
and Genetic Relatedness Among 37 Yardlong Bean and Cowpea Accessions Based on 
Morphological Characters and RAPD Analysis. Songklanakarin Journal of Science and 
Technology, 29(3), 591–560.  
167 
 
University of Ghana http://ugspace.ug.edu.gh
 
Satish, S., Raghavendra, M.P., and Raveesha, K.A. (2009). Antifungal Potentiality of some 
Plant Extracts against Fusarium sp, Archives of Phytopathology and Plant Protection, 
42(7), 618–625.  
Shahi, J.P., and Singh, I.S. (1985). Estimation of Genetic Variability for Grain Yield and its 
Components in a Random Mating Population of Maize, Crop Improvement, 12, 126–
129. 
Sharifi, P., Aminpanah, H., Erfani, R., Mohaddesi, A., and Abbasian, A. (2017). Evaluation 
of Genotype × Environment Interaction in Rice Based on AMMI Model in Iran. Rice 
Science 24 (3), 173–80. 
Shesshshayee, M.S., Bindumadhava, H., Shankar, A.G, Prasad, T.G, and Udayakumar, M. 
(2003). Breeding Strategies to Exploit Water-use efficiency for Crop Improvement. 
Journal of Plant Biology, 30, 253–268. 
Siddique, A.K., and Gupta, S.N. (1991). Genotypic and Phenotypic Variability for Seed Yield 
and other Traits in Cowpea [Vigna unguiculata (L.) Walp.]. International Journal of 
Tropical Agriculture. 9, 144–148. 
Simmonds, N.W., and Arthur, A.E. (2003). Crop Improvement |Plant Breeding Principles. 
Encyclopedia of Applied Sciences, 105–112. 
Sinclair, T., and Ludlow, M. (1986). Influence of Soil Water Supply on the Plant Water 
Balance of Four Tropical Grain Legumes, Functional Plant Biology, 13(3), 329–341. 
Singh, B.B., and Matsui, T. (2002). Cowpea Varieties for Drought Tolerance, in Challenges 
and Opportunities for Enhancing Sustainable Cowpea Production, eds Fatokun C.A., 
Tarawali, S.A, Singh, B.B, Kormawa, P.M, Tamo’ M, editors. (Ibadan: IITA) 287–300. 
Singh, B.B., and Ntare, B.R. (1985). Development of Improved Cowpea Varieties in Africa. 
In: Eds. S.R, Singh and K.O. Rachie, Cowpea Research, Production and Utilization. 
Wiley, New York, 106–115.   
Singh, B.B., Mai-kodomi, Y., and Terao, T. (1999). A simple Screening Method for Drought 
Tolerance in Cowpea. Indian Journal of Genetics, 59, 211–220. 
Singh, B.B., Raj, D.R.M., Dashiell, K.E. and Jackai., L.E.N. (1997) ‘Advances  in Cowpea 
Research Copublication of International Institue of Tropical Agriculture (IITA) and 
Japan International Research Center for Agricultural Sciences (JIRCAS), Ibadan, 
Nigeria,. Sayce Publishing, Devon, UK. 1−390’ 
Singh, R.K., and Chaudhary B.D. (1985). Biometrical Methods in Quantitative Genetic 
Analysis. New Delhi: Kalyani Publisher. 
Singh, S.P. (2007). Drought Resistance in the Race Durango Dry Bean Landraces and 
Cultivars. Agronomy Journal, 99, 1219–1225.  
Smith, H.F.(1936). A Discriminant Fuction for Plant Selection. Annual Eugenics, 7 240–250. 
168 
 
University of Ghana http://ugspace.ug.edu.gh
 
Sousa, C., Damasceno-Silva, K., Bastos, E., and Rocha, M. (2015). Selection of Cowpea 
Progenies with Enhanced Drought-tolerance Traits using Principal Component 
Analysis, Genetics and Molecular Research14 (4), 15981–15987.  
SPAD 502 Plus Chlorophyll Meter Basic Version, No Data Logger. Minolta, Japan. 
Sprague, G.F., and Tatum, L.A. (1942). General vs. Specific Combining Ability in Single 
Cross of Com. Journal of American Society of Agronomy. 34, 923–932. 
Taiz, L., and Zeiger, E. (2002). Plant Physiology, Third edition, Publisher: Sinauer 
Associates; Language: English ISBN: 0878938230 
Temesgen, T., Keneni, G., Sefera, T., and Jarso, M. (2015). Yield Stability and Relationships 
among Stability Parameters in faba bean (Vicia faba L.) Genotypes. Science Direct, CJ 
3(3), 258–268.  
Thi Lang, N and Chi Buu, B. (2008). Fine Mapping for drought Tolerance in Rice (Oryza 
sativa, L.). Omonrice, 16, 9–15. 
Thurling, N., and  Ratinam, M. (1987). Evaluation of Parent Selection Methods for Yield 
Improvement of Cowpea (Vigna unguiculata (L.) Walp.), 36, 913–926. 
Tuberosa, R. (2012). Phenotyping for Drought Tolerance of Crops in the Genomics Era; 
frontiers in Physiology, 8−33. 
Tuberosa, R., and Salvi, S. (2006). Genomics-based Approaches to Improve Drought 
Tolerance of Crops, 11(8). https://doi.org/10.1016/j.tplants.2006.06.003 
Tuhina-Khatun, M., Hanafi, M.M., Yusop, M.R., Wong, M.Y., Faezah, M.S., and Ferdous J. 
(2015) Genetic Variation, Heritability, and Diversity Analysis of Upland Rice (Oryza 
sativa L.) Genotypes Based on Quantitative Traits, BioMed Research International, 1–
8. 
 Turk, J.K., and Hall, A.E. (1980). Drought adaptation of Cowpea. III. Influence of Drought 
on Plant Growth and Relation with Seed Yield. Agronomy Journal, 72(3), 428–433. 
Uguru, M.I. (1996). Estimates of Variability and Genetic Gains in Cowpea (Vigna 
unguiculata L. Walp), Ghana Journal of Agricultural Science 29(2), 51–52. 
Vaillancourt, R.E., and Weeden, N.F. (1992). Chloroplast DNA Polymorphism Suggests 
Nigerian Center of Domestication for the Cowpea, Vigna unguiculata 
(Leguminosae), American Journal of Botany, 79(10), 1194–1199. 
Vargas, M., Glaz, B., Alvardo, G., Pietragalla, J., Morgounov, A., Zelenskiy, Y., and Crossa, 
J. (2014). Analysis and Interpretation of Interactions in Agricultural Research. 
Advances in Agronomy 102 (2), 748–762.  
Vernooy, R., Song, Y., Zhang, Z., Li, J., Liu, L., Martins, C., and Qin, T. (2013). 
Agroecology and Sustainable Food Systems Developing an Agricultural Biodiversity 
Policy for China Developing an Agricultural Biodiversity Policy for China. 
Agroecology and Sustainable Food Systems, 37, 1078–1095.  
169 
 
University of Ghana http://ugspace.ug.edu.gh
 
VSN International (2015). GenStat for Windows 18th Edition. VSN International, Hemel 
Hempstead, UK. Webpage: GenStat.co.uk 
Wamalwa, E.N., Muoma, J., Wekesa, C. (2016). Genetic Diversity of Cowpea (Vigna 
unguiculata L. Walp) Accessions in Kenya Gene Bank Based on Simple Sequence 
Repeat Markers. International Journal of Genomics, 24, 1–5. 
Wang, G., McGiffen, M.E, Ehlers, J.D., and Marchi, E.C.S. (2006). Competitive Ability of 
Cowpea Genotypes with Different Growth Habit. Weed Science, 54, 775–782. 
Watanabe, I., Hakoyama, S., Terao, T., and Singh, B. B. (1997). Evaluation Methods for 
Drought Tolerance of Cowpea. In: Advances in cowpea research. Singh, B.B., D.R. 
Mohan Raj, K.E. Dashiell and L.E.N. Jackai (eds). Co-publication of the International 
Institute of Tropical Agriculture (IITA) and Japan International Research Centre for 
Agricultural Sciences (JIRCAS). IITA, Ibadan, Nigeria. 141–146. 
Weiss, E.A. (1971). Castor, Sesameand Safflower, Leonard Hill Books, London, 311–355. 
Wilkinson, K.W., and Finlay G.N. (1963). The Analysis of Adaptation In A Plant-Breeding 
Programme The Ability of Some Crop Varieties to Perform Well over a Wide Range of 
Environ - Mental Conditions Has Lcng Been Appreciated by the Agronomist and Plant 
Breeder . In the Cereal Belts of Southern Australia. .Australian .Journal of Agrieultural. 
Research. 14, (1958), 742–54. 
XLSTAT (2017). Data Analysis and Statistical Solution for Microsoft Excel. Addinsoft, 
Paris, France. 
Xu, F., Guo, W., Xu, W., Wei, Y., and Wang, R..(2009). Leaf Morphology Correlates with 
Water and Light Availability: What Consequences for Simple and Compound Leaves.  
National Natural Science Foundation of China and Chinese Academy of Sciences, 
Progress in Natural Science, 19(12), 1789–1798.  
Xu, Y., and  Crouch, J.H. (2008). Marker-Assisted Selection in Plant Breeding: From 
Publications to Practice, Crop Science, 48, 391–407. 
Yan, W., and Hunt. L.A. (1999). Interpretation of Genotype × Environment Interaction for 
Winter Wheat Yield in Ontario. Crop Science., 41, 19–25. 
Yan, W., and Kang, M.S. (2003). GGE Biplot Analysis: A Graphical Tool for Breeders, 
Geneticists, and Agronomists. CRC Press. ISBN-13-978-0849313387 
Yan, W., and Tinker, N.A. (2005). An Integrated Biplot Analysis System for Displaying, 
Interpreting, and Exploring Genotype × Environment Interaction. Crop Science, 45 (3), 
1004–1016.  
Yan, W., Hunt, L.A., Sheng, Q., and Szlavnics, Z. (2000). Cultivar Evaluation and Mega-
Environment Investigation Based on the GGE Biplot. Crop Science, 40 (3), 597– 605. 
https://doi.org/10.2135/cropsci2000.403597x 
Yang, S., Vanderbeld, B., Wan, J., and Huang, Y. (2010). Narrowing Down the Targets: 
Towards Successful Genetic Engineering of Drought Toelrant Crops. Molecular Plant, 
3, 469–490.  
170 
 
University of Ghana http://ugspace.ug.edu.gh
 
Yilwat, M., Singh, B.B., and Ifenkwe, O.P. (2003). Because of Limited Conventional Genetic 
Studies, 18, 18–21 
Zobel, R.W., Wright, M.J., and Gauch, H.G. (1988). Statistical Analysis of a Yield Trial. 
Agronomy Journal, 80 (3), 388–393.  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
171 
 
University of Ghana http://ugspace.ug.edu.gh
 
APPENDICES 
Appendix III: Genetic clustering of the cowpea accessions 
Table 3.1 Distribution of 106 in nine clusters accessions of cowpea 
 
Cluster Number of Genotypes 
Genotypes  
1 35 AGRAC116, AGRAC216, AGRAC316, Apagu1A, Apagu1B,  
Apagu1C, Beledi A, Beledi B, Beledi C, GH5346B, WAC101,  
GH2347, GH5038, GH5346, GH6045, GH7220, GH7245,  
GH7875, IT08K1493, IT96D610, IT98D1399, Laduni1A,  
Rec017, Rec059, Laduni1B, Laduni2A, Laduni2B, Rec007,  
Rec014, Rec016, Titinwa A, Titinwa B, Titinwa C, WAC39 and  
Rec074         
2 39 Amos V, B, Blackeye, GH3701A, WAC 81, BLK19, WAC 32,  
GH2306, GH2307, GH3668, GH3689, GH5043, GH7228,  
IT08K15012, IT08K15024, IT08K18011, IT08K18724,  
IT93K4521, IT93K5031, IT97K49935, IT97K56818, IT98K5031,  
IT98K5061, IT99K1122, IT99K57211, IT99K57321, KNI,  
Kvx303096G, KVX3964525, Local 3, Mangala Mangala A, Pobe,  
Rec009, Rec021, Rec039, Rec041, Rec049, Rec105, Songotra  
3 1 Apagu 2A  
4 21 Dan lla, GH3701B, WAC19, WAC21, WAC115, WAC Q6,  
GH4524, IT08K125100, IT10K8177, IT97K56818, IT98K10921, 
 
IT98K2058, Kvx40481, Kvx74511, Mouride, Rec003, Rec046, 
 
Rec062, Rec083, Rec005, Rec108 
 
5 2 WAC91 and Gorom 
 
6 1 BLK452  
 
7 5 GH2200, GH2338, IT96D604, Padi_tuya, and Rec064  
 
8 1 GH4527 
9 1 Mangala Mangala B  
 
 
 
 
 
 
 
 
 
 
 
 
172 
 
University of Ghana http://ugspace.ug.edu.gh
 
Table 3.2.  Grouping of the genotypes into inferred clusters based on probabilities of 
association of greater than or equal to 35% to the 6 clusters. 
Genotypes Populations Inferred Cluster 
1 2 3 4 5 6 
Songotra 0.828 0.072 0.001 0.071 0.026 0.001 1 
Dan lla 0.902 0.084 0.000 0.000 0.010 0.003 1 
Padi_tuya 0.592 0.049 0.013 0.000 0.001 0.345 1 
Beledi A 0.588 0.067 0.009 0.001 0.000 0.335 1 
IT08K1493 0.712 0.000 0.286 0.001 0.001 0.001 1 
Titinwa A 0.001 0.000 0.001 0.001 0.040 0.957 6 
AGRAC116 0.306 0.002 0.321 0.053 0.003 0.316 Admixture 
AGRAC216 0.007 0.003 0.163 0.102 0.001 0.724 6 
Laduni_1A 0.046 0.002 0.929 0.002 0.020 0.001 3 
IT08K18724 0.408 0.125 0.000 0.000 0.055 0.411 6 
IT08K15024 0.001 0.178 0.000 0.000 0.820 0.000 5 
Apagu1A 0.838 0.032 0.000 0.000 0.129 0.001 1 
Laduni_1B 0.106 0.252 0.011 0.134 0.064 0.434 5 
Titinwa B 0.236 0.604 0.002 0.007 0.069 0.081 2 
Black eye 0.666 0.028 0.032 0.001 0.273 0.000 1 
Beledi C 0.001 0.001 0.004 0.001 0.002 0.991 6 
AGRAC316 0.001 0.207 0.000 0.791 0.000 0.001 4 
Laduni_2A 0.001 0.823 0.001 0.007 0.073 0.096 2 
Mangala_Mangala A 0.004 0.317 0.001 0.022 0.149 0.507 6 
IT08K18011 0.002 0.000 0.002 0.994 0.001 0.000 4 
IT08K15012 0.002 0.915 0.000 0.003 0.037 0.043 2 
Apagu_2A 0.001 0.902 0.000 0.001 0.095 0.001 2 
IT99K57211 0.928 0.000 0.069 0.001 0.001 0.001 1 
IT97K56818 0.785 0.001 0.204 0.003 0.000 0.007 1 
IT10K8177 0.001 0.001 0.995 0.002 0.000 0.001 3 
IT97K49935 0.689 0.144 0.022 0.006 0.134 0.005 1 
IT99K57321 0.722 0.001 0.000 0.080 0.196 0.000 1 
Local_3 0.159 0.050 0.007 0.177 0.366 0.242 5 
IT98K5061 0.885 0.007 0.001 0.100 0.005 0.002 1 
IT99K1122 0.681 0.001 0.172 0.001 0.139 0.006 1 
IT96D610 0.002 0.000 0.995 0.002 0.001 0.000 3 
IT98K5031 0.477 0.003 0.000 0.154 0.318 0.047 1 
Titinwa C 0.001 0.000 0.996 0.002 0.000 0.000 3 
IT98K2058 0.645 0.000 0.000 0.000 0.355 0.000 1 
IT93K5031 0.007 0.001 0.976 0.011 0.004 0.001 3 
IT96D604 0.182 0.673 0.000 0.000 0.144 0.000 2 
IT98K10921 0.000 0.001 0.000 0.000 0.996 0.002 5 
GH4527 0.302 0.001 0.001 0.379 0.310 0.008 4 
GH7228 0.997 0.001 0.000 0.001 0.001 0.000 1 
GH2338 0.615 0.097 0.007 0.001 0.248 0.032 1 
GH3689 0.538 0.293 0.000 0.000 0.168 0.000 1 
GH2347 0.003 0.213 0.000 0.781 0.001 0.001 4 
GH5038 0.207 0.019 0.393 0.089 0.003 0.289 3 
Mangala Mangala B 0.001 0.989 0.000 0.001 0.009 0.001 2 
GH4524 0.508 0.281 0.006 0.001 0.182 0.022 1 
GH2307 0.002 0.001 0.003 0.110 0.001 0.883 6 
GH3701_B 0.708 0.168 0.001 0.064 0.001 0.058 1 
GH2306 0.026 0.079 0.848 0.001 0.046 0.000 3 
GH5043 0.002 0.909 0.001 0.001 0.082 0.005 2 
GH7220 0.005 0.003 0.004 0.059 0.002 0.927 6 
GH3668 0.001 0.000 0.995 0.000 0.003 0.000 3 
GH7875 0.826 0.001 0.110 0.002 0.000 0.062 1 
GH3701_A 0.396 0.225 0.001 0.001 0.186 0.192 1 
GH6045 0.118 0.002 0.129 0.003 0.141 0.606 6 
GH7245 0.154 0.684 0.004 0.001 0.156 0.001 2 
GH5346 0.002 0.091 0.000 0.003 0.037 0.867 6 
173 
 
University of Ghana http://ugspace.ug.edu.gh
 
GH2200 0.198 0.214 0.167 0.013 0.311 0.098 Admixture 
GH5346_B 0.019 0.001 0.972 0.001 0.005 0.002 3 
Rec005 0.001 0.959 0.000 0.000 0.039 0.000 2 
IT97K56818 0.003 0.634 0.161 0.142 0.060 0.000 2 
Apagu_2B 0.068 0.001 0.910 0.019 0.001 0.001 3 
WAC91 0.002 0.000 0.003 0.990 0.001 0.004 4 
WAC_81 0.875 0.001 0.118 0.005 0.001 0.000 1 
Laduni_2B 0.002 0.977 0.001 0.001 0.018 0.001 2 
IT08K125100 0.005 0.442 0.318 0.164 0.070 0.001 2 
WAC101 0.007 0.166 0.001 0.064 0.642 0.121 5 
WAC19 0.522 0.199 0.001 0.003 0.152 0.123 1 
WAC21 0.796 0.062 0.001 0.001 0.139 0.001 1 
Apagu1C 0.029 0.000 0.967 0.001 0.000 0.002 3 
BLK19 0.006 0.001 0.889 0.030 0.000 0.074 3 
Apagu1B 0.024 0.274 0.004 0.147 0.483 0.068 5 
WAC115 0.001 0.003 0.001 0.001 0.001 0.993 6 
WAC_Q6 0.642 0.157 0.120 0.003 0.077 0.002 1 
BLK452 0.650 0.001 0.329 0.001 0.011 0.009 1 
IT93K4521 0.722 0.108 0.017 0.004 0.143 0.007 1 
WAC39 0.025 0.000 0.000 0.000 0.973 0.000 5 
WAC32 0.153 0.001 0.793 0.010 0.015 0.028 3 
Rec003 0.748 0.096 0.001 0.002 0.128 0.024 1 
Rec007 0.001 0.706 0.000 0.001 0.290 0.002 2 
Rec062 0.917 0.000 0.080 0.001 0.001 0.001 1 
Rec021 0.003 0.001 0.994 0.001 0.001 0.001 3 
IT98D1399 0.001 0.167 0.000 0.000 0.831 0.001 5 
Rec083 0.802 0.000 0.195 0.001 0.001 0.000 1 
Rec046 0.745 0.001 0.001 0.040 0.132 0.081 1 
Rec016 0.005 0.914 0.000 0.001 0.079 0.000 2 
Amos V 0.001 0.001 0.000 0.000 0.000 0.997 6 
Rec105 0.012 0.009 0.018 0.050 0.003 0.907 6 
Rec059 0.001 0.001 0.000 0.000 0.060 0.937 6 
Rec064 0.008 0.001 0.867 0.003 0.040 0.082 3 
Rec041 0.005 0.091 0.057 0.040 0.002 0.805 6 
Rec014 0.002 0.000 0.016 0.968 0.014 0.000 4 
Rec074 0.001 0.700 0.001 0.001 0.297 0.000 2 
Rec049 0.995 0.001 0.000 0.001 0.003 0.000 1 
Rec009 0.997 0.001 0.001 0.000 0.001 0.000 1 
Rec017 0.003 0.962 0.001 0.001 0.033 0.000 2 
Rec108 0.253 0.243 0.005 0.001 0.003 0.494 6 
Rec039 0.002 0.000 0.996 0.001 0.001 0.000 3 
Mouride 0.387 0.321 0.068 0.001 0.177 0.048 1 
Gorom 0.009 0.001 0.002 0.052 0.003 0.933 6 
Pobe 0.001 0.559 0.151 0.001 0.288 0.000 2 
KVX3964525 0.706 0.079 0.074 0.029 0.075 0.037 1 
Kvx74511 0.007 0.963 0.001 0.007 0.021 0.002 2 
KNI 0.025 0.655 0.000 0.000 0.319 0.000 2 
Kvx40481 0.001 0.001 0.001 0.001 0.687 0.309 5 
Kvx303096G 0.788 0.089 0.000 0.002 0.011 0.109 1 
Beledi B 0.096 0.458 0.002 0.081 0.215 0.148 2 
 
174