COMBINING ABILITY AND YIELD PERFORMANCE STUDIES OF TROPICAL MAIZE (Zea mays L.) GERMPLASM UNDER DROUGHT STRESS AND WELL- WATERED ENVIRONMENTS BY LAOUALI, MAHAMANE NASSER (10325631) THIS THESIS IS SUBMITTED TO THE UNIVERSITY OF GHANA, LEGON IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE AWARD OF DOCTOR OF PHILOSOPHY PLANT BREEDING DEGREE WEST AFRICA CENTRE FOR CROP IMPROVEMENT COLLEGE OF BASIC AND APPLIED SCIENCES UNIVERSITY OF GHANA LEGON DECEMBER, 2014 University of Ghana http://ugspace.ug.edu.gh i DECLARATION I hereby declare that except for references to works of other researchers, which have been duly cited, this work is my original research and that neither part nor whole has been presented elsewhere for the award of a degree. .................................................. Laouali Mahamane Nasser Student .................................................. Prof. Vernon Gracen Supervisor .................................................. Prof Essie T. Blay Supervisor .................................................. Dr Badu-Apraku Supervisor University of Ghana http://ugspace.ug.edu.gh ii ABSTRACT Maize is increasingly consumed by Nigerien households. Limited studies have been conducted on the production and productivity of this crop in Niger and very few cultivars (mostly open pollinated varieties) are available to farmers. Knowledge and understanding of farmers’ preferences for new varieties would be invaluable in designing a successful maize breeding program. Maize cultivars grown in Niger are derived mainly from tropical yellow-grained extra- early and early-maturing lowland germplasm sources. Lack of availability of appropriate early and extra-early testers as well as information on the combining ability, the mode of gene action, and the heterotic patterns of the available early and extra-early inbreds impede the successful development and adoption of maize hybrids in West Africa. Three activities were conducted in the present study. The first was a Rapid Rural Appraisal to assess the appropriateness of maize crop management practices used by farmers in Niger, to identify major factors limiting maize production and productivity, and to identify traits preferred by farmers. The study revealed that the production of maize in the country is carried out using few inputs and poor crop management practices. The current level of production is very low due to many factors including drought, low soil fertility, poor crop management and lack of availability of improved varieties. Yellow maize cultivars that combine earliness and high yield potential are desired by farmers. The second activity consisted of two parts including an experiment to determine the right time to impose water stress in maize drought trials in Niger and a diallel study to explore the combining ability of grain yield and other traits of early maturing, yellow-grained maize inbreds under drought and well-watered conditions. The objectives of the this activity were to test the efficiency of the CIMMYT/IITA screening method for drought tolerance in Niger, examine the University of Ghana http://ugspace.ug.edu.gh iii combining ability and the mode of gene action for drought tolerance among 15 early maturing inbred lines, classify the inbreds into contrasting heterotic groups using two different classification methods i.e. the SCA effects of grain yield and the HSGCA, and assess the yield performance and stability of hybrids under contrasting environments. The findings of the activity revealed that the drought screening method that involved (i) stopping irrigation 3 to 2 weeks before anthesis, (ii) applying a “rescue irrigation” 13 to 15 days after imposing drought stress, and (iii) resuming the normal irrigation 10 to 12 days later until crop maturity was the most appropriate for maize drought trials in Niger. A moderate yield reduction under drought of 36.84% of the yield under well-watered environment was observed. Significant GCA and SCA effects for grain yield and most measured traits were detected with predominance of GCA effects over SCA effects, indicating that most traits were controlled predominantly by additive gene action. Two heterotic groups were identified based on the classification with the SCA_PY method. Inbreds ENT 13, TZEI1 124 and TZEI 129 displayed the highest GCA values in well- watered environments and were the best parents that are likely to contribute favorable alleles to progeny. Inbred ENT 13 was the best combiner under both drought and well-watered conditions suggesting its inclusion in population imporvement and in the development of hybrids and synthetics. Significant differences were observed among the single cross hybrids under the two water treatments. The single cross hybrids TZEI 124 x ENT 13; TZEI 160 x TZEI 157; TZEI 182 x ENT 13 and TZEI 167 x ENT 13 were the highest yielding hybrids across test environments and out-yielded all the checks. In the third activity, three related single cross (RSC) parents, two single cross checks and 103 modified single crosses (MSCs) derived from crosses of the three RSCs to 36 inbreds of varying maturity groups were evaluated under drought and well-watered conditions to assess the yield University of Ghana http://ugspace.ug.edu.gh iv performance of MSCs and examine the relationship between grain yield of MSCs and their maturity groups. Results revealed a moderate but adequate level of moisture stress (62.3% yield reduction). Significant differences were observed among the MSCs under the two moisture regimes. Under drought stress, the 3 highest yielding MSCs (TZEEI 76 x TZEEI 82) xCMK ENT 23; (TZEEI 76 x TZEEI 82) x CMZ ENT 8; and (TZEEI 95 x TZEEI 79) x CMK ENT 8 outyielded the best single cross check (ENT 13 x CamInbgp.1-17) by 91%, 67%, and 66%, respectively. Under well-watered conditions, the 3 top yielding MSCs (TZEEI 76 x TZEEI 82) x TZEI 124; (TZEEI 78 x TZEEI 79) x ENT 13 and (TZEEI 95 x TZEEI 79) x CMZ ENT 5 outyielded the best single cross check TZEI 124 x ENT13 by 19.3%, 17.8% and 17% respectively. Modified single cross hybrids developed from early and intermediate maturing inbreds yielded higher than those developed from extra-early inbreds indicating that extra-early germplasm is less advantageous in terms of grain yield potential particularly under optimal growing conditions. University of Ghana http://ugspace.ug.edu.gh v ACKNOWLEDGEMENTS Grace is to ALLAH the ALMIGHTY who gave me strength to complete this work. The two years training in plant breeding that I received at the West African Center for Crop Improvement (WACCI) of the University of Ghana-Legon and the three years of my thesis field work in Niger were made possible by the funding from the Alliance for a Green Revolution in Africa (AGRA). I gratefully appreciate the support of AGRA for my PhD training. I hereby express my sincere gratitude to my supervisory committee at WACCI composed of Prof Vernon Gracen, Prof Essie T. Blay and late Dr Charles The. I also express my sincere thanks to my in-country supervisor Dr B. Badu-Apraku for the practical training, advice and support he provided to ensure the success of this thesis research. I seize this opportunity to thank the University of Ghana-Legon that hosted through WACCI our training in Plant Breeding. I express my sincere gratitude to WACCI Directors, WACCI administration, Staff, and students for the training and support I received during the five years. I express my heartfelt gratitude to WACCI lecturers and all lecturers from different countries and institutions that contributed to the success of this work through the intensive and high quality training they gave us at WACCI. I’m very grateful to Cornel University that contributed to our training and also for allowing us to have access to the Cornel Library. The use of maize germplasm from the International Institute for Tropical Agriculture (IITA), CIMMYT, and the Maize Breeding Program of WACCI is sincerely acknowledged. I acknowledge the support and the use of research facilities of my research institution INRAN. Finally, I express my gratitude to the external examiners Dr J.M. Ngeve and Dr J. Manu-Aduening and the internal examiner Prof Essie Blay for the corrections they made to improve this document. University of Ghana http://ugspace.ug.edu.gh vi DEDICATION To African resource poor farmers, those who are struggling to feed Africa despite the lack of resources and their limited access to modern farming technologies University of Ghana http://ugspace.ug.edu.gh vii TABLE OF CONTENTS DECLARATION…………………………………………………………………………………………… I ABSTRACT ……………………………………………………………………………………………….. II ACKNOWLEDGEMENTS………………………………………………………………………………… V DEDICATION………………………………………..…………………………………………………….. VI TABLE OF CONTENTS………………………………..………………………………………………….. VII LIST OF TABLES………………………………………………………………………………………….. XI LIST OF FIGURES………………………………………………………………………………………….. XIII LIST OF ABBREVIATIONS……………………………………………………………………………… XIV CHAPTER ONE…………………………………………………………………………………........................ 1 1.0. GENERAL INTRODUCTION…………….………………………………………………………………… 1 CHAPTER TWO…………………………………………………………………………………………… 6 2.0. LITERATURE REVIEW…………………………………………………………………………………… 6 2.1. Inbreeding depression and heterosis……………………………………………………………… 6 2.2. Inbreeding depression and heterosis in hybrid maize development…………………………………… 6 2.3. Genetic basis of inbreeding depression and heterosis…………………………………………….. 7 2.4. Combining ability……………………………………………………………………………………… 8 2.5. Heterotic groups and heterotic patterns………………………………………………………………… 9 2.6. Establishment of heterotic groups and heterotic patterns in maize………………………………. 9 2.7. Estimates of variance components……………………………………………………………………… 11 2.7.1. Approaches using first and second degree statistics……………………………………………. 11 2.7.2. Molecular approaches………………………………………………………………………………… 14 2.8. Drought and maize production………………………………………………………………………… 14 2.9. Breeding maize for drought-prone environments………………………………………………….. 15 2.10. Strategies for successful improvement of drought tolerance in maize………………………….. 16 2.10.1. Drought management methods………………………………………………………………… 16 University of Ghana http://ugspace.ug.edu.gh viii 2.10.2. Type of drought tolerance to breed for in maize…………………………………………………. 17 2.10.3. Selection based on yield and secondary traits under drought………………….…………….. 18 2.10.4. Selection of germplasm for drought tolerance improvement in maize………………………….. 18 2.10.5. Selection of breeding schemes…………………………………………………………………. 19 2.11. Characteristics of an ideal drought tolerant genotype……………………………………………… 20 2.12. Molecular approaches for drought tolerance improvement in maize……………………………. 20 2.13. Genotypes by environment interactions in drought experiments……………………………….. 21 2.14. Participatory plant breeding…………………………………………………………………………… 22 2.15. The use of participatory methods to assess farmers’ knowledge on crop production……………. 23 CHAPTER THREE…………………………………………………………………………………………. 24 3.0. Maize production in Niger: crop management practices and production constraints………………. 24 3.1. Introduction……………………………………………………………………………………… 24 3.2. Research methodology………………………………………………………………………………… 25 3.2.1. Sites of the study…………………………………………………………………………………… 25 3.2.2. RRA team and farmers’ selection…………………………………………………………………… 27 3.2.3. Data collection……………………………………………………………………………………… 28 3.2.4. Data analysis……………………………………………………………………………………… 30 3.3. Results…………………………………………………………………………………………… 30 3.3.1. Importance of maize among the cereal crops produced in Niger……………………………….. 30 3.3.2. Utilization of maize produced in Niger………………………………………………………. 31 3.3.3. Farming practices used by maize growers in Niger…………………………………………… 31 3.3.4. Maize traits preferred by farmers………….…………………………………………………. 35 3.3.5. Major factors limiting maize production in Niger………………………………………………. 35 3.3.6. Availability of production facilities for increasing maize production in the visited sites…..… 37 3.4. Discussion………………………………………………………………………………………. 37 3.5. Conclusion and breeding implications………………………………………………………………… 42 University of Ghana http://ugspace.ug.edu.gh ix CHAPTER FOUR………………………………………………………………………………………….. 44 4.0. Combining ability for grain yield and other agronomic traits of early maturing maize inbreds under drought and well-watered conditions……………………………...…………………………………… 44 4.1. Introduction……………………………………………………………………………………… 44 4.2. Materials and methods……………………………………………………………………………… 47 4.2.1. Germplasm and mating design……………………………………………………………………… 47 4.2.2. Stress management……………………………………………………………………………… 48 4.2.3. Experimental design and evaluation sites………………………………………………………… 49 4.2.4. Field practices and stress management……………………………………………………………… 50 4.2.5. Data collection…………………………………………………………………………………. 52 4.2.6. Statistical analysis……………………………………………………………………………… 53 4.3. Results……………………………………………………………………………………………. 57 4.3.1. Identification of the time to impose water stress in maize drought studies in Niger……………. 57 4.3.2. Performance of single cross hybrids, inbred parents and checks under drought and well- watered conditions…………………………………………………………………………… 57 4.3.3. Correlations, selection indices and identification of productive single-cross hybrids for production under drought stress conditions………………………………………………….. 66 4.3.4. Genetic analysis of the diallel crosses evaluated under drought and well-watered conditions 68 4.3.5. Combining ability effects of measured traits…………………………………………………. 72 4.3.6. Mode of gene action controlling measured traits…………………………………………….. 74 4.3.7. Heterotic grouping of the early maturing inbreds using SCA_PY and HSGCA methods..….. 76 4.3.8. Stability analysis of GCA effects of early maturing inbreds across test environments……… 84 4.3.9. GGE biplot analysis of grain yield performance and stability of early maturing single cross hybrids……………………………………………………………………………………….... 88 4.4. Discussion……………………………………………………………………………………….. 92 4.5. Conclusion………………………………………………………………………………………… 100 University of Ghana http://ugspace.ug.edu.gh x CHAPTER FIVE…………………………………………………………………………………………… 102 5.0. Assessment of yield performance of modified single crosses with varying maturity under drought and well-watered conditions………………………………………………………………………..….. 102 5.1. Introduction……………………………………………………………………………………… 102 5.2. Materials and methods……………………………………………………………………………… 104 5.2.1. Genetic materials……………………………………………………………………………… 104 5.2.2. Experimental design and evaluation sites…………………………………………………………… 107 5.2.3. Stress management……………………………………………………………………………… 108 5.2.4. Field management practices………………………………………………………………………… 109 5.2.5. Data collection………………………………………………………………………………… 109 5.2.6. Statistical analysis……………………………………………………………………………… 110 5.3. Results…………………………………………………………………………………………… 111 5.3.1. Performance of related single crosses and inbred parents under drought and well-watered conditions……………………………………………………………………………………… 111 5.3.2. Performance of modified single crosses under drought and well-watered conditions……….. 119 5.3.3. Relationship between yield performance and maturity group of modified single crosses…… 122 5.4. Discussion……………………………………………………………………………………….. 125 5.5. Conclusions………………………………………………………………………………………. 128 CHAPTER SIX…………………………………………………………………………………………….. 130 6.0. Research Overview ……..……………………………………………………………………………… 130 6.1. Introduction……………………………………………………………………………………… 130 6.2. Major findings and their implications………………………………………………………………. 131 6.3. Limits of the study and recommendations for future research…………………………………… 139 REFERENCES……………………………………………………………………………………………….. 142 APPENDICES………………………………………………………………………………………………. 153 University of Ghana http://ugspace.ug.edu.gh xi LIST OF TABLES Table 3.1. Characteristics of the sites used for the RRA study 27 Table 3.2. Objectives of the RRA study and RRA tools used to achieve them 29 Table 4.1. Origin, pedigree and reaction to abiotic stresses of the inbred lines used in the diallel study 48 Table 4.2. Grain yield of the drought experiment conducted to identify the right time to impose water stress in maize drought trials in Niger 57 Table 4.3. Mean squares from the combined analysis of variance of grain yield and other agronomic traits across environments 61 Table 4.4. Proportions of the total variance attributable to the sources of variance for grain yield of maize hybrids evaluated in seven contrasting environments in Niger from 2012 to 2014 62 Table 4.5. Grain yield and other agronomic traits of the 5 highest yielding hybrids, the 5 lowest yielding hybrids and the 5 checks evaluated under drought stress and WWC 63 Table 4.6. Grain yield and other agronomic traits of the 15 inbreds evaluated under drought stress and well-watered conditions at Maradi during the dry season of 2013/2014 65 Table 4.7. Pearson correlation coefficients between pairs of traits across 105 single cross hybrids and 5 checks evaluated under drought stress and well-watered conditions in Niger 67 Table 4.8. Combined analysis of variance of grain yield and other traits of early maturing single cross hybrids evaluated across well-watered environments in Niger 69 Table 4.9. Combined analysis of grain yield and other traits of early maturing single cross hybrids evaluated across stress environments in Niger 70 Table 4.10. Combined analysis of variance of grain yield and other traits of early maturing single cross hybrids evaluated across the seven test environments in Niger 71 Table 4.11. GCA effects of grain yield and other agronomic traits of early maturing yellow inbred lines evaluated under drought and well-watered conditions in Niger 73 Table 4.12. SCA effects of grain yield of the 105 single crosses evaluated under WW environments in Niger, from 2012 to 2014 78 Table 4.13. Classification of the 15 inbreds into heterotic groups based on SCA effects for grain yield across environrnents 79 Table 4.14. HSGCA values for the 15 inbreds used in the diallel study 80 Table 4.15. Classification of the 15 inbreds into heterotic groups based on HSGCA effects for grain yield under well-watered environments 81 University of Ghana http://ugspace.ug.edu.gh xii Table 4.16. Number of hybrids of the three yield groups identified by each of the two classification methods 82 Table 4.17. Average grain yield for intra-group and inter-group crosses and average heterosis for the different heterotic groups identified by the HSGCA method 82 Table 4.18. Average grain yield for intra-group and inter-group crosses and average heterosis for the different heterotic groups identified by the SCA_PY method 83 Table 4.19. Heterotic grouping based on the application of the classification procedure described by Fan et al. (2009) on the SCA effects of crosses under well-watered conditions 83 Table 5.1. Characteristics of extra-early maturing maize inbred lines used in the study 104 Table 5.2. Related single crosses developed from intra-group crosses plus the two single cross and local OPV checks used in the study 104 Table 5.3. Selected inbred lines of varying maturity period used in the development of the modified single cross hybrids 105 Table 5.4. Grain yield and other agronomic traits of the related single crosses and checks evaluated under drought stress and well-watered conditions in Niger in 2012 and 2013 113 Table 5.5. Mean squares from the combined analysis of variance for grain yield and other agronomic traits of the related single crosses and checks evaluated across six environments in Niger 115 Table 5.6. Grain yield and other agronomic traits of the extra-early maturing inbreds evaluated under drought stress and well-watered conditions at six contrasting environment in Niger 118 Table 5.7. Grain yield and other agronomic traits of the 5 highest yielding MSCs, the 5 lowest yielding MSCs and 5 checks evaluated under drought stress and WWC at INRAN Maradi 121 Table 5.8. Grain yield and other agronomic traits of the three maturity groups of the modified single crosses evaluated under drought stress and WWC at INRAN Maradi 124 University of Ghana http://ugspace.ug.edu.gh xiii LIST OF FIGURES Figure 3.1. Sites of the rapid rural appraisal study on maize production in Niger 26 Figure 3.2. Simulation of planting at Koderawa and fertilizer application at Jiratawa 29 Figure 3.3. Percentage of farmers applying different plants population densities across study sites 32 Figure 3.4. Plant population density adopted by farmers in the visited villages 33 Figure 3.5. Percentage of farmers applying different level of mineral fertilizers across the sites 34 Figure 3.6. Quantity of mineral fertilizers applied by farmers in the villages surveyed 34 Figure 3.7. Maize stemborer (Sesamina calamistis) attack in maize field at Jiratawa 36 Figure 3.8 The ring-necked parakeet (Psittacula krameri) attack on maize cob at Jiratawa 36 Figure 4.1. Variation of grain yield depending on the water stress intensity in the drought experiment 58 Figure 4.2. Grain yield reduction depending on the stress intensity from the drought experiment 59 Figure 4.3. Proportion of additive and non-additive genetic variance for grain yield and other traits of the 105 single cross hybrids under drought 74 Figure 4.4. Proportion of additive and non-additive genetic variance for grain yield and other traits of the 105 single cross hybrids under well-watered environments 75 Figure 4.5. Proportion of additive and non-additive genetic variance for grain yield and other traits of the 105 single cross hybrids across test environments 75 Figure 4.6. An entry/tester genotype main effect plus genotype by environment interaction biplot for GCA effects of grain yield of 15 inbreds across test environments (E1through E7) 85 Figure 4.7. A ‘‘which won where” genotype plus genotype x environment interaction biplot of GCA effects for grain yield of the 15 early maturing maize inbreds across seven environments 87 Figure 4.8. GGE biplot for grain yield performance and stability of 23 selected single cross hybrids and two checks across seven environments (E1 through E7) 89 Figure 4.9. A ‘‘which won where” GGE biplot of grain yield of 23 early maturing maize hybrids and two checks evaluated across seven environments 91 Figure 5.1. Grain yield (Kg ha-1) of the related single crosses and checks evaluated across six contrasting environments in Niger in 2012 and 2013 116 University of Ghana http://ugspace.ug.edu.gh xiv LISTE OF ABREVIATIONS AMMI: Additive Main effects and Multiplicative Interaction ANOVA: Analysis of Variance ATA: average tester coordinate abscissa ATO: average tester coordinate ordinate CIMMYT: International Maize and Wheat Improvement Center FAOSTAT: Statistical Database of the Food and Agriculture of the United Nations FESA: Ferme Semenciere Amate (Amate Seed Company) FGD: Focus Group Discussion GCA: General Combining Ability GGE: Genotypes + Genotypes by environments interactions GEI: Genotypes by environments interactions HGCAMT: Heterotic grouping based on GCA of Multiple Traits HSGCA: Heterotic groups’ Specific and General Combining Ability IITA: International Institute of Tropical Agriculture INRAN: Institut National de la Recherche Agronomique du Niger (National Agronomic Research Institute of Niger) IRAT: Tropical Agriculture Research Institute (Niger) JAICAF: Japan Association for International Collaboration of Agriculture and Forestry MSC: Modified Single Cross OPV: Open Pollinated Varities PC: Principal Component PPCR: Pilot Programme for Climate Resilience in Niger PRA: Participatory Rural Appraisal QTL: Quantitative Trait Loci RCBD: Randomized complete blocks design RRA: Rapid Rural Appraisal RSC: Related Single Cross SCA: Specific Combining Ability SVP: Singular Value Partitioning WCA: West and Central Africa WSC: Water Stress Conditions WWC: Well-watered Conditions University of Ghana http://ugspace.ug.edu.gh 1 CHAPTER ONE 1.0. GENERAL INTRODUCTION Maize (Zea mays L.) is an important cereal crop grown worldwide for various purposes. It occupies the second position after wheat in terms of area of production but represents the most important cereal crop in terms of quantity produced worldwide (FAOSTAT, 2015). The USA is the leading producer of maize in the world accounting for about 37% of the total production, followed by China, Brazil, and Mexico. The whole continent of Africa contributes only 7% of the global production (FAOSTAT, 2015). Maize is mostly used as an animal feed in the developed countries but it is an important staple crop in developing countries. It is consumed as food at various developmental stages from baby corn to matured grain. It is estimated that 65 percent of the total world maize production is used for feed, about 15 percent for food and the remaining for various types of industrial uses (Abbassian, 2008). Maize production worldwide is constrained by biotic and abiotic stresses. The two types of stresses are responsible for yield reduction of 22% and 43% respectively (Gibbon et al., 2007). Drought is the most important abiotic stress constraining and destabilizing maize grain production in both temperate and tropical environments (Edmeades, 2008; Ribaut et al., 2009). Yield losses due to drought average 15 percent annually in sub-Saharan Africa (Badu-Apraku and Lum, 2010; Badu-Apraku et al., 2011a) even in areas where total rainfall is reasonably high (Fischer et al., 1982). Total yield losses may be observed when severe drought occurs at the reproductive stages of plant development. Besides drought low soil fertility, particularly low soil nitrogen (Gibbon et al., 2007) and Striga infestation constitute the major constraints to maize production in West and Central Africa (Badu-Apraku et al., 2013a; Akinwale et al., 2014). Striga hermonthica can cause yield reduction of 20 to 80% and even complete yield losses under University of Ghana http://ugspace.ug.edu.gh 2 severe infestation (Badu-Apraku and Akinwale, 2011) whereas annual yield losses due to low-N are estimated to vary from 10 to 50% (Wolfe et al., 1988). Maize is ranked fourth among the cereals grown in Niger i.e. millet, sorghum, rice and maize. There are three types of maize production in the country. The most important production is done under irrigation in areas where irrigation facilities are available e.g. public irrigation schemes of Jiratawa and Konni. The second type is maize production under rain fed conditions. This production is mostly restricted to the region of Gaya where the annual rainfall ranges from 650 mm to 900 mm. The third type of production is the "decrue maize" where maize is cultivated along seasonal flowing rivers (Komadougou yobe, goulbi Maradi etc.) that inundate maize fields during the rainy season (JAICAF, 2009). Maize is planted after water withdrawal from the fields. The high water holding capacity of the soils allows the production of extra early varieties without any additional irrigation. Maize is often produced in association with legume and vegetable crops (Habsatou et al., 2010; Mossi et al., 2010). Extra-early and early maturing yellow varieties are mostly grown by farmers and used for green maize consumption. The overall climatic conditions in Niger, however, are not suitable for rain-fed maize production. Niger has semi-arid tropical climate with two seasons: a long dry season (from October to May ) and a short rainy season (from June to September) (PPCR, 2010). The climate is characterized by high variability especially in terms of rainfall which is not only irregularly distributed throughout the year, but also erratically dispersed over the country (PPCR, 2010). From north to south, four main agro-climatic zones are recognized in the country: ­ The Saharan zone, which covers 77% of the country has an annual precipitation hardly reaching 150 mm. Irrigated farming is practiced in this zone. ­ The Sahelo-Saharan zone covers 12% of Niger’s total area. The annual precipitation is between 150 mm and 300 mm. University of Ghana http://ugspace.ug.edu.gh 3 ­ The Sahelian zone, which covers 10% of Niger’s total area and receives 400 to 600 mm of annual precipitation. It is conducive to agro-pastoralism. ­ The Sahelo-Sudanian zone, which covers about 1% of Niger’s total surface area and receives 600 to 800 mm of annual precipitation. It is conducive to food production and animal husbandry (PPCR, 2010). Maize production in Niger is constrained by low soil fertility, drought, poor crop management and the lack of availability of improved varieties. However, drought remains the most important limiting factor that dissuades farmers from producing maize under rainfed conditions. The national maize production remains low and averages 4,000 tons annually. The grain yield is around 0.8 t ha-1, being the lowest in West Africa. This situation is mostly due to the aforementioned complex limiting factors. To meet the increasing demand of maize grain in the country, around 46,000 tons of maize grain is imported each year from neighboring countries (Nigeria, Benin, Togo, Burkina, and Mali). Maize is an orphan cereal crop in Niger which has not been extensively studied by the research community unlike millet and sorghum. The research conducted so far on maize in the country was limited to: (i) the improvement of a local OPV P3 Kollo for earliness and resistance to maize streak virus, (ii) conduct of regional trials, and (iii) identification of appropriate planting dates (Jika, 1991). The other available information on maize is the basic crop statistics (area cultivated, average grain yield etc.). There is therefore a need to conduct more studies in maize production areas of the country to develop a better understanding of the maize production system. This will allow setting appropriate maize breeding priorities for the national maize breeding program. A Rapid Rural Appraisal study was conducted in different maize production sites of the country for this purpose. University of Ghana http://ugspace.ug.edu.gh 4 To boost the national maize production from its current level, both productions under irrigation and rain-fed conditions should be increased. This requires an increased availability and accessibility to high yielding and stable maize varieties by Nigerien farmers. These varieties should also carry drought tolerance genes for a better production of maize under rainfed conditions where drought remains an important threat. Currently limited genetic variability is available to maize breeders in the country to develop such superior maize genotypes. It is therefore important to introduce new maize germplasm into the country to increase the genetic variability which can be exploited to enable farmers to have access to a wide range of improved maize genotypes such as single crosses and modified single crosses. These are expected to perform higher than the old and low yielding OPVs currently grown by farmers in the country. The international centres for maize improvement IITA and CIMMYT have developed a wide range of maize germplasm which is adapted to the climatic conditions of sub-Saharan African countries such as Niger. A large number of their inbred lines were developed under stress conditions such as Striga, drought and low nitrogen. IITA and CIMMYT early and extra- early maize inbreds represent an important source of maize genetic diversity for the national maize breeding program in Niger. However, information on the combining abilities and the heterotic patterns of these lines is limited. Knowledge and understanding of the heterotic patterns of the inbreds will be very useful in developing high yielding and stable maize hybrids to enhance the national maize production in Niger. The overall objective of this study is to exploit the genetic diversity existing in IITA and CIMMYT maize inbreds to develop high yielding and stable maize single cross and modified single cross hybrids across contrasting environments in Niger. University of Ghana http://ugspace.ug.edu.gh 5 The specific objectives of this study are to: ­ identify maize production constraints and maize agronomic traits desired by farmers; ­ estimate the combining ability and the mode of gene action for yield and drought tolerance of early maturing inbred lines; ­ evaluate the agronomic performance of F1 single-cross hybrids under drought and well- watered conditions and estimate their yield stability across varying environments; ­ classify inbred lines into heterotic groups; ­ assess the yield performance of modified single cross hybrids under drought and well- watered conditions and identify the relationship between their grain yield and maturity. University of Ghana http://ugspace.ug.edu.gh 6 CHAPTER TWO 2.0. LITERATURE REVIEW 2.1. Inbreeding depression and heterosis Inbreeding depression and heterosis are two important genetic phenomena that are observed in the reproduction of species with out-crossing mating system. Inbreeding depression refers to the overall reduction in fitness, fertility, and survival of individuals in a population brought about by breeding genetically related individuals (Charlesworth and Willis, 2009). Inbreeding increases homozygosity, hence, it leads to the reduction of the within families genetic variance and the increase of the among families variance (Moorad and Wade, 2005). It causes a decline in mean fitness of individuals owing to the manifestation of recessive deleterious alleles, the reduction in frequency of favorable overdominant heterozygous loci and the disruption of advantageous gene interactions (Moorad and Wade, 2005). On the other hand, heterosis as defined by George H. Shull (Shull, 1948) refers to the phenomenon in which the hybrid progeny exhibits superior phenotypic characteristics compared to the mean of the two parents or to the best parent (Springer and Stupar, 2007). 2.2. Inbreeding depression and heterosis in hybrid maize development Heterosis is responsible for the success of the maize hybrid seed industry that uses the inbred/hybrid production method (Duvick, 1999). The inbreeding depression per se is not a desirable phenomenon in inbred line development. It however, indicates that there exists genetic variation in traits controlling fitness, survival and fertility in natural populations (Charlesworth and Willis, 2009). It plays an important role (Charlesworth and Willis, 2009) in the control of the stability and uniformity of inbred lines. In the development of maize inbred lines, a good level of University of Ghana http://ugspace.ug.edu.gh 7 stability is reached when the reduction in plant vigor slows down and plants become uniform from one generation of selfing to the next. Heterosis has been extensively exploited to develop commercial hybrids. Traits like grain yield, plant height and number of days to flowering have been significantly improved by exploiting heterosis (Duvick, 1999). However, heterosis per se is a desirable phenomenon only when it is associated with high performance of the hybrid for the trait under improvement. A low yielding hybrid exhibits a high but useless heterosis for yield when its two parents are also very low yielding (Duvick, 1999). Furthermore, high heterosis for grain yield alone cannot explain the dramatic increase in yield that has been achieved in modern commercial hybrids grown in developed countries. The yield improvement is also due to the adoption of efficient crop management practices by farmers and also to significant improvement of many important defensive traits that may not show any heterosis (Duvick, 1999). 2.3. Genetic basis of inbreeding depression and heterosis Inbreeding depression and heterosis have been under investigation since Darwin’s era in 1876 (Birchler et al., 2010). Current advances in genetic studies have confirmed that both are genetic phenomena (Charlesworth and Willis, 2009). However the genetic causes underlying these phenomena are still not fully understood (Tang et al., 2010). Shull (1948) acknowledged that no single cause or mechanism can be properly assumed to apply to all the cases. Several hypotheses that are not mutually exclusive have been proposed to explain heterosis (Shull, 1948) but they have proven difficult to test. The most prominent of them are the dominance hypothesis, the overdominance hypothesis, and the epistasis hypothesis (Reif et al., 2005). The dominance hypothesis attributes the superiority of hybrids to the suppression of detrimental alleles that are University of Ghana http://ugspace.ug.edu.gh 8 largely recessive from one parent by dominant alleles from the other. The overdominance hypothesis attributes heterosis to the superior fitness of heterozygous genotypes over homozygous genotypes. The epistasis hypothesis explains heterosis by the presence and the magnitude of favorable non allelic interactions between many loci. Charlesworth and Willis (2009) indicated that classical genetic studies and modern molecular evolutionary approaches suggest that inbreeding depression and heterosis are predominantly caused by the presence of recessive deleterious mutations in populations, probably with a contribution from overdominance at few loci. However other QTL studies revealed that epistasis also plays a much more significant role in the manifestation of heterosis than previously believed (Melchinger et al., 2007a; Melchinger et al., 2007b). Despite the fact that the genetic dissection of heterosis is not fully resolved, Hallauer et al. (2010) supported the idea that significant improvement of maize cultivars can still be made by exploiting this phenomenon. 2.4. Combining ability Combining ability refers to the capacity of an individual to contribute superior fitness related traits to hybrid progeny. It is very useful in mating designs to investigate the performance of lines in hybrid combinations (Griffing, 1956). In 1941 Sprague and Tatum introduced the concepts of general combining ability (GCA) and specific combining ability (SCA) (Hallauer et al., 2010). They defined the two terms respectively as "the average performance of a line in hybrid combinations" and "the cases where certain combinations do relatively better or worse than would be expected on the basis of the average performance of the lines involved" (Griffing, 1956). The concepts of GCA and SCA are very useful in plant breeding and have particular significance to the diallel mating design (Hallauer et al., 2010). They have been extensively used by breeders in determining heterotic patterns in maize germplasms. University of Ghana http://ugspace.ug.edu.gh 9 2.5. Heterotic groups and heterotic patterns Heterotic groups and patterns are two fundamental concepts exploited in maize hybrid breeding theory and practice (Reif et al., 2005). The concept of heterotic group was defined by Melchinger and Gumber (1998) "as a group of related or unrelated genotypes from the same or different populations, which display similar combining ability and heterotic response when crossed with genotypes from other genetically distinct germplasm groups". On the other hand, the concept of heterotic patterns designates "a specific pair of two heterotic groups which express high heterosis and consequently high hybrid performance in their crosses" (Melchinger and Gumber, 1998). The identification of heterotic groups and patterns is of utmost importance for any hybrid breeding program. Sprague (cited in Reif et al., 2005) stated that:" the single most important element of a breeding program is the recognition and utilization of heterotic patterns. This recognition both simplifies and increases the efficiency of all subsequent operations". Experimental results reviewed by Melchinger and Gumber (1998) revealed the superiority of inter-group over intra-group hybrids and hence demonstrated the necessity of the development of distinct heterotic groups in a breeding program. 2.6. Development of heterotic groups and heterotic patterns in maize Maize heterotic patterns were developed by breeders and are not the result of historical or geographical contingencies (Tracy and Chandler, 2004). The fundamental principle underlying the development of heterotic patterns is to cross in a diallel or testcross fashion a large number of inbreds and identify those showing the highest hybrid performance (Melchinger and Gumber, 1998). Further improvement of the heterotic pattern is done by the adoption of a breeding and selection technique designated as reciprocal recurrent selection (Reif et al., 2005). In the U.S. the first attempt at the identification of maize heterotic patterns was made when maize populations University of Ghana http://ugspace.ug.edu.gh 10 from the Reid Yellow Dent group were found to combine well with the Lancaster Sure Crop group. The Dent x Flint heterotic patterns has since then been extensively exploited to generate outstanding maize hybrids in the U.S., in Central and Western Europe, in Japan (Reif et al., 2005; Fan et al., 2009) in China, in Argentina, and in South Africa. Several approaches have been used to assign germplasms into distinct heterotic groups. The classical SCA method relates the heterosis observed in crosses with the origin of the parents involved in the crosses and/or the hybrid yield information (Melchinger and Gumber, 1998; Fan et al., 2009). Another method employs various molecular markers to compute genetic similarity or genetic distance to assign maize lines to different heterotic groups (Menkir et al., 2004; Aguiar et al., 2008; Fan et al., 2009; Badu-Apraku et al., 2013a; Akinwale et al., 2014). Results were not always consistent with the ones of the classical approach which therefore is still used as a major method for maize heterotic group classification. Fan et al., (2009) proposed another alternative method that uses both GCA and SCA to assign inbred lines to known maize heterotic groups. The method designated as heterotic group’s specific and general combining ability (HSGCA) is reported to explain more variation in maize hybrid yield and produce more predictable yield than the previous two methods (Fan et al., 2009; Akinwale et al., 2014). Badu- Apraku et al., (2013) proposed another method designated as heterotic grouping based on GCA of multiple traits (HGCAMT) which could be more effective in classifying inbred lines into appropriate heterotic groups. The advantage of the HGCAMT method particularly in breeding for stress environments is that it uses GCA of multiple traits rather than GCA of yield alone which has low heritability in those environments. University of Ghana http://ugspace.ug.edu.gh 11 2.7. Estimates of variance components The type of gene action involved in the expression of quantitative traits has been under investigation since the era of earlier geneticists (Hallauer and Miranda, 1981). Knowledge of the mode of gene action is fundamental in making solid inferences on breeding populations. It informs breeders of the appropriate selection method to adopt for rapid progress in their breeding programs. Several approaches have been proposed for estimating variance components (Reif et al., 2005). These include: 2.7.1. Approaches using first and second degree statistics Various mating designs based on first and second-degree statistics have been used to estimate variance components mostly attributable to additive and dominance effects (Reif et al., 2005). Most of these designs assume a normal Mendelian diploid inheritance, no maternal effects, linkage equilibrium, non-inbred relatives, randomly selected parents and relatives, arbitrary allelic frequencies, and no epistasis. Hallauer et al., (2010) reviewed in detail the following designs that are extensively used to produce progenies for evaluation: - Bi-parental Progenies: It is one of the simplest mating designs for estimation of genetic variance in a reference population. Pairs of individual plants randomly taken from a population are crossed and progenies are evaluated in replicated trials. The information provided by bi- parental progenies gives an indication as to whether indeed sufficient genetic variability is present in the population to allow significant gains from the selection program. ­ Parent–Offspring Regressions: this method regresses offspring or selfed progeny measurements on those of one parent or both parents (mid parent-offspring regression). The simplest method consists of making measurements on individual plants in a reference population. University of Ghana http://ugspace.ug.edu.gh 12 The same traits are measured on the offspring derived from each parent to determine the degree of association between the traits measured in the parents and in their respective offspring. This method allows a quick estimate of additive genetic component and the heritability of the trait of interest. ­ North Carolina design I: It is adequate only for estimating genetic components of variance for a reference non inbred population. Male and female individuals are randomly selected in the reference population. Each male is crossed to a different set of females to produce progenies for evaluation in replicated trials. It is cost effective in producing a large number of progenies. This design provides estimates of additive genetic variance. It does not give a direct estimate of dominance variance. Since the males used for crossing are non-inbreds, specific combining ability (SCA) effects cannot be worked out. ­ North Carolina design II: From a reference non inbred population males and females individuals are randomly selected and each male is crossed to each female to produce progenies for evaluation in replicated trials. Design II offers the following advantages over diallel designs: (i) more parents can be included, (ii) two independent estimates of additive variance are available, and (iii) an estimate of dominance variance is determined directly from the mean squares. However it is more difficult to apply in open-pollinated species with only one inflorescence since multiple crosses on female plants are not possible. ­ North Carolina design III: Randomly selected F2 individuals derived from crosses between two inbreds are backcrossed to the inbred parents. Design III provides good estimates of additive variance and dominance variance for F2 populations assuming absence of linkage and epistasis. This design has an advantage over the design II because it can measure levels of dominance and the estimation of dominance is not subjected to any assumption regarding allele University of Ghana http://ugspace.ug.edu.gh 13 frequencies. It can also determine the effects of linkages on the estimates of additive variance and dominance variance, and average level of dominance. ­ Diallel designs: They are the most extensively used designs in maize breeding particularly in the public sector. Diallel cross refers to all possible crosses among n parents ranging from inbred lines (Kempthorne, 1955) to broad genetic base varieties (for Gardner and Eberhart’s analysis) (Hallauer et al., 2010). It is very powerful if correctly analyzed and interpreted. Griffing (1956) described four diallel methods depending on whether or not parental inbreds or the reciprocals or both are included: (i) method 1: full diallel in which parents and reciprocal crosses are involved along with F1, (ii) method 2: half diallel with parents and without reciprocal crosses, (iii) method 3: full diallel without inclusion of parents, and (iv) method 4: half diallel without parents and reciprocal crosses. The diallel analysis takes into consideration two models i.e. fixed model (model 1) and random model (model 2). The fixed model suggests that parents of the progenies are the genotypes for which the inference will be drawn. On the other hand the random model suggests that parents are a sample of genotypes representing a reference population. In the fixed model estimation of GCA and SCA is more appropriate and more informative than estimation of components of variance while for the random model analysis, estimation of the components of variance is of prime interest. One major disadvantage of diallel methods is the limitation of the number of parents to be included due to the large number of crosses needed as the number of parents increases. ­ Line x tester: The concept of line x tester analysis was developed by Kempthorne in 1957 (Sharma, 2006). It is an extension of the topcross design in the sense that in the case of top crosses only one tester is used, while in the case of line x tester crosses several testers are used (Nduwumuremyi et al., 2013). Each tester is mated to each of the lines which usually are more in University of Ghana http://ugspace.ug.edu.gh 14 number. The potential of the lines in hybrid combination could therefore be assessed from the evaluation of the F1 progeny. Line x tester is the simplest mating design that provides both full- sibs and half-sibs simultaneously as opposed to top cross which provides only half-sibs (Nduwumuremyi et al., 2013). It provides SCA of crosses and better estimates of GCA of testers than of lines (Nduwumuremyi et al., 2013). Additive and dominance variance components can also be obtained. An important advantage of line x tester design over diallel methods is that it allows the evaluation of a much higher number of lines. 2.7.2. Molecular approaches The advent of molecular markers in the late 1980’s offers another alternative way of determining the type of gene action involved in the expression of quantitative traits (Reif et al., 2005). QTL studies have been extensively conducted to decipher the genetic basis of heterosis. Evidence from modern molecular studies of heterosis suggested an important contribution of dominance (Charlesworth and Willis, 2009). However the proportion of overdominance, dominance, epistasis effects as well as their interactions greatly depends on the genetic materials used in the study, the number and types of molecular markers, and the model used for the analysis (Reif et al., 2005; Swanson-Wagner et al., 2006; Ma et al., 2007; Melchinger et al., 2007a; Li et al., 2008; Tang et al., 2010). 2.8. Drought and maize production Drought is a meteorological term commonly defined as a period without significant rainfall (Jaleel et al., 2009). It is the most important threat to maize production in both temperate and tropical environments (Edmeades, 2008; Ribaut et al., 2009). Yield losses due to drought average 15 percent annually in sub-Saharan Africa (Badu-Apraku and Lum, 2010; Badu-Apraku University of Ghana http://ugspace.ug.edu.gh 15 et al., 2011a) even in areas where total rainfall is reasonably high (Fischer et al., 1982). Grain yield losses could be higher if drought occurs at the flowering and grain filling periods (Denmead and Shaw, 1960; NeSmith and Ritchie, 1992). Total yield losses may even be observed when severe drought occurs at the reproductive stages. Furthermore, in drought prone environments, the high probability of drought occurrence dissuades farmers from applying adequate levels of fertilizers and other inputs. Drought is therefore probably responsible for a much higher economic loss than indicated (Fischer et al., 1982). Drought may occur at the beginning, the middle, or the end of the rainy season. In addition, drought duration and intensity are also totally unpredictable (Ribaut et al., 2009). In the tropics, periodic drought caused by irregular rainfall distribution and accentuated by low water-holding capacity of the soils is the major cause of significant reductions in maize yields (Wolf et al., 1974). 2.9. Breeding maize for drought prone environments Several strategies have been proposed to mitigate drought damage on crop production in drought-prone environments. These include (i) drought escape, primarily by avoiding the coincidence of stress with crops sensitive stages (ii) increasing the crop’s water use efficiency and (iii) increasing the drought tolerance per se in plants (Ribaut et al., 2009). Earliness is the drought escape mechanism mostly preferred by farmers especially those dealing with frequent late season drought (Defoer et al., 1997; Abebe et al., 2005; Obaa et al., 2005; Ouma et al., 2010; Sibiya et al., 2013). However, when the occurrence and the severity of drought stress are completely unpredictable, drought avoidance by earliness alone or by planting date is not an appropriate solution (Fischer et al., 1982). Therefore any maize genotypes to be deployed in drought prone environments should carry drought tolerance genes that will confer the ability to survive and to produce acceptable levels of grain yield under water limiting conditions. University of Ghana http://ugspace.ug.edu.gh 16 2.10. Strategies for successful improvement of drought tolerance in maize Fischer et al. (1982) reported two hypotheses described by Blum (1979) as the first attempts for improving drought tolerance in maize: (i) a superior variety with high yield potential under optimum growing conditions will also perform relatively well under sub-potential levels, (ii) breeding for yield performance under stress environments maintains that potential yield. Neither of these approaches succeeded in substantially improving drought tolerance in maize. Indeed many reports confirm that selection for yield in non-stressed environments is generally less effective in identifying individuals which will perform well in low-yielding stressed environments (Fischer et al., 1982; Bolaños et al., 1993; Bänziger et al., 2000; Ribaut et al., 2009). The alternative approach to breeding maize for drought tolerance is to develop selection methods that can increase or at least maintain potential yield and at the same time improve drought tolerance traits (Fischer et al., 1982). This approach relies on (i) the development of efficient stress management methods that will effectively discriminate and elicit true differences among individuals screened for drought tolerance, (ii) the identification of useful secondary traits that have high correlation with yield under drought stress, (iii) the careful selection of evaluation sites representative of the most prevalent drought type in the target area, and (iv) the appropriate choice of germplasm and breeding schemes (Bänziger et al., 2000; Badu-Apraku et al., 2004; Ribaut et al., 2009). 2.10.1. Drought management methods The unpredictability of drought occurrence and severity coupled with soil variability and high genotype x environment interactions inhibits a reliable detection of treatment differences in drought experiments conducted under natural conditions. Therefore the use of uniform soils and appropriate stress management methods along with rigorous experimental designs and University of Ghana http://ugspace.ug.edu.gh 17 appropriate statistical analyses are the alternative to guaranteeing good selection progress (Ribaut et al., 2009). CIMMYT and IITA maize breeding programs have developed and extensively used stress management methods that simulate a clearly defined stress that is relevant in farmers’ fields (Bänziger et al., 2000; Badu-Apraku et al., 2004). For the drought stress, the method consists of conducting experiments partly or entirely during the dry season and managing the stress through irrigation while taking into consideration the timing, intensity, and uniformity of the stress (Bänziger et al., 2000). Experiment blocks are managed with one stress level only and are far enough apart to prevent border effects. Irrigation is stopped so that drought stress is sufficiently intense at the targeted critical growth stage (Bänziger et al., 2000). 2.10.2. Type of drought tolerance to breed for in maize With respect to drought tolerance per se in maize, three primary breeding targets can be distinguished i.e. drought tolerance at flowering time, drought tolerance during grain filling, and yield stability across a range of environments (Ribaut et al., 2009). Greenhouse experiments, field trials and historical analysis of on-farm data revealed that drought occurring around flowering time has a major effect on maize grain yield while stress during the grain filling and the vegetative stages may have less effect on yield (Fischer et al., 1982; Bänziger et al., 2000; Kamara et al., 2003; Ribaut et al., 2009; Badu-Apraku et al., 2010). Therefore Breeding for drought tolerance during flowering and/or post-flowering has the best chance of affecting maize production, provided those types of drought stress are the most relevant in the target environments (Bänziger et al., 2000). For drought stress at flowering stage, irrigation is designed so that drought at flowering is severe enough to delay silking and cause ear abortion while in the case of drought stress during grain filling, irrigation is designed so that drought develops directly after flowering and accelerates leaf senescence (Bänziger et al., 2000). University of Ghana http://ugspace.ug.edu.gh 18 2.10.3. Selection based on yield and secondary traits under drought Secondary traits are plant characters other than grain yield that provide additional information on how the plant performs under a given environment (Lafitte et al., 2003). The low heritability of maize grain yield under drought necessitates their use. An ideal secondary trait should be genetically variable, genetically correlated with grain yield in the target environment, should have a high level of heritability, be simple and cheap, non-destructive and fast to assay, be stable throughout the measurement period, and not associated with any yield penalty under non- stress conditions (Edmeades et al., 1997; Lafitte et al., 2003; Badu-Apraku et al., 2012). In maize breeding for drought tolerance, the secondary traits that have been used include anthesis- silking interval (ASI), number of ears per plant (EPP), the number of kernels per ear, plant and ear height, leaf area, extent of leaf rolling, osmotic adjustment, leaf chlorophyll concentration, stomatal conductance, canopy temperature, hydraulic conductivity, abscissic acid concentration, ear growth and barrenness, stay-green characteristic, erect upper leaves, tassel size, and root architecture (Edmeades et al., 1997; Bänziger et al., 2000; Ribaut et al., 2009; Badu-Apraku et al., 2012). ASI and EPP were identified by Edmeades et al. (1997) as the most reliable secondary traits among a number of secondary traits ranked on the basis of heritability, association with grain yield, cost, genetic variability, and time required for recording. 2.10.4. Selection of germplasm for drought tolerance improvement in maize Breeding for drought tolerance is very challenging because of the complex nature of the stress tolerance due to factors including the largely polygenic nature of the tolerance, the typically low frequency of tolerance alleles in most maize germplasm, and the difficulties commonly encountered in field evaluation (Bänziger et al., 2000). The selection of appropriate germplasm and efficient breeding methods is critical to improving the drought tolerance of maize University of Ghana http://ugspace.ug.edu.gh 19 populations and hybrids (Bänziger et al., 2000; Ribaut et al., 2009). CIMMYT evaluated a wide range of landraces and adapted elite germplasm for drought tolerance and found that improving the latter ones is probably nearly always better than working with landraces (Bänziger et al., 2000). However, there are other alternatives such as (i) creation of synthetic populations from local landraces and improved adapted varieties, (ii) improvment of non-adapted but drought tolerant populations for local adaptation, (iii) development of new breeding population by introgressing locally adapted germplasm with drought tolerant germplasm (Bänziger et al., 2000). 2.10.5. Selection of breeding schemes Some important factors that might be taken into consideration when designing a breeding program for drought tolerance are (i) the type of genotypes to produce e.g. OPVs, synthetics, hybrids, (ii) the most important traits to be preserved in the new product e.g. early maturity, grain characteristics, and (iii) strategy for developing and deploying the products (Bänziger et al., 2000). Maize plants offer a wide array of options with respect to breeding methodologies. Among the various selection methods, recurrent selection remains the breeding procedure extensively used by CIMMYT and IITA to improve populations’ performance under drought stress and maintain genetic variability to ensure continued progress from selection (Bolaños and Edmeades, 1993; Bolaños et al., 1993; Bänziger et al., 2000; Monneveux et al., 2008; Ribaut et al., 2009; Badu-Apraku et al., 2013b). The resultant populations improved for drought tolerance are also useful sources for extracting drought tolerant inbred lines for hybrid development. However, in hybrid development for drought-prone environments, the relationship between the performance of inbred lines and their hybrids is an important issue to consider (Bänziger et al., 2000). Many reports have found significant dosage effect under drought suggesting the need for University of Ghana http://ugspace.ug.edu.gh 20 including drought tolerant parents on both sides of the hybrid to achieve acceptable drought tolerance where the stress is severe (Bänziger et al., 2000; Betràn et al., 2003a; Ribaut et al., 2009; Badu-Apraku et al., 2011b). 2.11. Characteristics of an ideal drought tolerant genotype Drought tolerance in an agricultural sense refers to the ability of a crop plant to produce its economic product with limited available water; thus, the mere survival with no grain yield is of little importance (Fischer et al., 1982; Bänziger et al., 2000; Cattivelli et al., 2008; Ribaut et al., 2009). The expression of tolerance depends greatly on both the developmental stage during which water stress is imposed and the intensity and duration of the stress (Ribaut et al., 2009). Ideal drought tolerant cultivars are characterized by increased production under drought (Bänziger et al., 2000) and should display no yield penalty in non stress environments. These genotypes are expected to display reduced ear and kernel abortion, shorter anthesis-silking interval, greater ear growth rate at flowering, increased harvest index, reduced leaf area (especially in the upper part of the plant), short and thick stems, small tassels, erect leaves and delayed senescence (Bänziger et al., 2000; Ribaut et al., 2009). Other traits difficult to measure include root biomass, root system depth and root lateral ramifications (Jaleel et al., 2009). 2.12. Molecular approaches for drought tolerance improvement in maize Extensive studies of drought tolerance traits have been carried out in maize over the last decade, yielding a large number of QTLs involved in the determination of morphological traits, yield components, flowering traits and plant height (Bänziger et al., 2000; Messmer et al., 2009; Ribaut et al., 2009; Tsonev et al., 2009; Setter et al., 2011). Several microarray-based studies, with inconsistent results however, have also been carried out on drought stressed maize leading University of Ghana http://ugspace.ug.edu.gh 21 to the putative identification of multiple genes, including transcription factors and regulatory proteins as being part of the drought response network (Ribaut et al., 2009). The candidate gene approaches have also been used revealing many drought tolerance candidate genes. However, few of them have been validated either via reverse genetics or transgenic approaches (Ribaut et al., 2009). Despite the documentation of numerous QTLs involved in the expression of drought tolerance in maize, their exploitation in public breeding programs has to date been limited, although it is not the case in the private sector (Ribaut et al., 2009). A better understanding of the genetic basis of maize response to drought brought about by molecular approaches, together with accurate phenotyping methods will probably play an important role in the identification, the transfer and the selection of key genes and alleles to build genotypes with much improved tolerance to drought. 2.13. Genotypes by environment interactions in drought experiments Genotype by environment (G x E) interactions are common under drought and make breeding progress difficult (Bänziger et al., 2004). G x E refers to changes in order, ranking, and relative values among genotypes evaluated in multi-locational and multi-year trials (Hallauer et al., 2010). In such trials, G x E are often found to account for a greater degree of variance than G alone (Chapman et al., 1997). G x E has been a research focus among biometricians, quantitative geneticists and breeders with the notion that they are undesirable or confounding factors in the sense that they complicate the selection and/or recommendation of materials (Annicchiarico, 1997; Yan and Tinker, 2006). G x E are due mostly to macroenvironmental effects (Hallauer et al., 2010) particularly the variation in weather conditions between and within seasons and soil properties (Derera, 2005). However, G x E study is of special interest for breeding programs since it may lead to the subdivision of a target region into different sub- University of Ghana http://ugspace.ug.edu.gh 22 regions, relatively uniform with respect to genotypic responses and alternatively provide guidelines for the choice of crucial test sites (Annicchiarico, 1997). Better approaches to analyze G x E data are now available such as the GGE biplot method (Yan and Tinker, 2006) and the AMMI analysis (Annicchiarico, 1997). These methods can help breeders to identify environments that are similar in the way in which they discriminate among genotypes and also determine genotypes that have similar patterns of response across a range of environments (Yan and Tinker, 2006). 2.14. Participatory plant breeding The ultimate aim of any breeding program is the high adoption of the cultivars developed for farmers; otherwise all the intensive research work and research funds invested are a waste. In the developing countries many attempts to develop the agriculture through research projects failed to increase the adoption rate of the research technologies by the target populations. A likely explanation for this low level of technology adoption is that stakeholders (e.g. farmers, consumers) were not involved enough in the research to adopt the findings or the findings were not within the capacity of their implementation (Njoku, 2012). Participatory approaches in plant breeding have been proposed as an alternative to address these issues. When participatory techniques are appropriately employed in plant breeding they can have an impact by quickly and cost-effectively producing much improved crop varieties and by accelerating their dissemination and their rate of adoption by farmers (Abebe et al., 2005; Witcombe et al., 2005). Therefore, it is of fundamental importance that farmers be involved in the variety selection and testing process, and that researchers pay careful attention to maize traits that are of most importance to farmers (Bänziger et al., 2000). University of Ghana http://ugspace.ug.edu.gh 23 2.15. The use of participatory methods to assess farmers’ knowledge on crop production Information gathering is a necessary prerequisite for designing successful programs and project activities to help rural communities (Freudenberger, 1999). Participatory and qualitative methods such as RRA (Rapid Rural Appraisal) and PRA (Participatory Rural Appraisal) have proven to be very useful tools in gathering information in rural areas (Bhandari, 2003). This information enables breeders to customize their varieties and refine their approaches according to the needs and circumstances of the farmers (Freudenberger, 1999). PRA is a continuous process and a more participatory approach that actively involves the communities’ members, facilitates their analytical capabilities and empowers them to plan and undertake sustainable action (Freudenberger, 1999). On the other hand RRA is rather more extractive (Campbell, 2001) and less participatory (Freudenberger, 1999). RRA is relevant when information is needed quickly and decisions are preempted by the passage of time (Chambers, 1981). It is a short time study conducted in a limited number of representative sites to collect rough but accurate baseline information (Campbell, 2001). RRA is most important in the very early phases of project planning where basic information is needed to define project’s priorities and project’s approaches to be adopted to address the target population’s concerns. University of Ghana http://ugspace.ug.edu.gh 24 CHAPTER THREE 3.0. Maize production in Niger: crop management practices and production constraints 3.1. Introduction Maize production in Niger has not been well documented compared to millet or sorghum production. Data on maize production in Niger is not exhaustive and has not been regularly updated. The data available is mostly on basic crop statistics e.g. area cultivated, grain yield, price and quantity imported. According to the National Statistics Institute (INS-Niger, 2013) the area devoted to maize cultivation in the country averages 7,400 hectares annually. The grain yield per hectare is around 0.8 ton, being the lowest in West Africa. The national production of maize remains low and averages 4,000 tons annually. However, it should be noted that the data on the national production does not take into consideration all the maize produced for green consumption, yet the green maize probably represents the most important part of the national production. Recently the government of Niger decided to boost the national agricultural production through a new policy called ‘‘3N Initiative: Nigerien feed Nigerien’’ which basically aims at developing irrigation facilities and facilitating access to agricultural inputs and other modern farming technologies to farmers (HCI-3N, 2012). Maize was identified as a strategic crop in this new initiative particularly in areas where irrigation facilities are available. The national demand of high quality seeds of improved maize varieties is increasing and the agronomic research institutions are expected to play an important role in this new initiative through the development of new improved maize cultivars. In order to accomplish this important task and meet expectations, maize breeders in the country need to understand the maize production system. This includes gathering information on University of Ghana http://ugspace.ug.edu.gh 25 the types and characteristics of maize varieties that farmers prefer, the major factors limiting maize production, and the farming practices used by maize growers e.g. fertilizer application, crop association, pesticide usage, level of adoption of new cultivars and new farming technologies. The present study is a rapid rural appraisal conducted in eight maize production sites located in three different agro-ecological zones in Niger as an attempt at accessing to such important information on maize production in the country. The main objectif of the study was to develop a better understanding of the maize production system in the country. The specific objectives were to (i) assess the importance of maize among the different crops produced in Niger, (ii) determine the different uses of maize crop in the country, (iii) assess the efficiency of farming practices used by maize growers in Niger, (iv) identify maize varietal preferences of farmers, (v) identify major factors that limit maize production in Niger, and (vi) determine the production facilities that could allow enhancing maize production in the visited sites. 3.2. Research methodology 3.2.1. Sites of the study The RRA study was conducted in three districts (Figure 1) corresponding to three different agro- ecological zones and two different types of maize production systems: ­ District of Tchirozérine (region of Agadez). It is located in the Saharan zone which represents about 77% of the country. This zone has an annual precipitation hardly reaching 150 mm. Maize is produced under irrigation. University of Ghana http://ugspace.ug.edu.gh 26 ­ District of Madarounfa (Region of Maradi). It is located in the Sahelian zone which covers about 10% of Niger’s total area. This zone has an annual precipitation of 400 to 500 mm. Maize is mostly produced under irrigation but also under rain-fed conditions. ­ District of Gaya (region of Dosso). It is located in the Sahelo-Sudanian zone which covers about 1% of Niger’s total surface area and receives 700 to 900 mm of precipitation annually. Maize production is carried out under rain-fed conditions. Figure 3.1. Sites of the rapid rural appraisal study on maize production in Niger In each of the districts of Madarounfa and Gaya three villages that are maize production sites were chosen for the study. Initially, the district of Tchirozerine was not among the proposed sites to visit during the study. An opportinuty offerd by a mission conducted by the sorghum breeding unit of INRAN in the district was seized to carry out the RRA study in two villages of the Bagzan Mountain located in that district. However some of the RRA tools (e.g. simulation of University of Ghana http://ugspace.ug.edu.gh 27 field activities by farmers) could not be applied in these two villages for lack of time. In total eight villages were used for the study (Table 3.1.). Table 3.1. Characteristics of the sites used for the RRA study. 3.2.2. RRA team and farmers’ selection The RRA team was composed of the PhD Student (myself), a female agronomist, an agricultural development technician, the agricultural extension officer in charge of the village and a representative of farmers’ organization in the village. In each of the eight villages a stratified random sampling procedure was used to select a group of 15 to 20 farmers that were representatives of the maize farmers’ population in the village. The local authorities, the farmers’ association and the agricultural extension officer were involved in the process of selection of participating farmers. The strata considered were gender, age, social status, and maize production capacity. The selection was done in such a way that elements of all the strata were represented in Districts Villages GPS Coordinates Average rainfall (mm) Length of the rainy season (months) Type of land Altitude (m) Type of production Tchirozérine Emaléoulé N 17°42’ E 8° 46’ 100-150 2-3 high land 1500 irrigated maize Bagzan n’Amass N 17°42’ E 8° 45’ 100-150 2-3 high land 1500 irrigated maize Djiratawa N 13°24’ E 7°8’ 400 -500 3-4 Valley land 350 irrigated maize Madarounfa Riadi N 13°25’ E 7° 8’21’ 400 -500 3-4 Valley land 350 irrigated maize Koderawa N 13° 22’ E 7° 9’ 400 -500 3-4 Valley land 350 irrigated maize Yelou N 12° 15’ E 3° 34’ 600 -800 5-6 Valley land 185 rainfed maize Gaya Guidan Gaba N 12° 10’ E 3° 27’ 500 -700 5-6 Valley land 190 rainfed maize Bengou N 11° 59’ E 3° 35’ 600 -800 5-6 Valley land 185 rainfed maize University of Ghana http://ugspace.ug.edu.gh 28 the group. However, for social consideration and also to avoid gender bias, female groups and male groups were kept separated. 3.2.3. Data collection Local authorities and agricultural extension officers were actively involved to facilitate the creation of good rapport with farmers and to sensitize them to be committed to the survey. Different RRA tools including (1) focus group discussion (FGD), (2) matrix ranking, (3) transect walks for field observations, and (4) simulation of field activities by farmers were used to obtain information (Figure 3.2.). The group discussion was guided by a checklist (Appendix 3.1.). The agricultural extension officer intervened as a facilitator and ensured that each farmer participated in the discussion and brought his own opinion on the matter. The topics for the discussions were the importance of maize among the crops produced by farmers, the different uses of maize, the type of production, the level of the production, and the different challenges facing maize growers. The farmers were asked to list and rank the major constraints to maize production and the key traits they pay attention to in adopting new maize cultivars. At the end of the discussion, except in the district of Tchirozérine, farmers were asked to simulate field operations such as planting and fertilizer application as practicised in their own fields. The purpose was to have an idea of the plant density and the quantity of fertilizers used by farmers. After the group discussions a transect walk was carried out for field observations with some of the farmers. The objectives of the group discussions and the different RRA tools used to achieve them are presented in Table 3.2. University of Ghana http://ugspace.ug.edu.gh 29 Table 3.2. Objectives of the RRA study and RRA tools used to achieve them PRA tools Objectives FGD with Checklist Matrix ranking Transect walk Simulation of field activities Assess the importance of maize among the different crops produced in Niger   Determine the different uses of maize crop in the country  Assess the efficiency of the farming practices used by the maize growers in Niger    Identify farmer preferences for maize cultivars   Identify major factors limiting maize production in Niger    Determine the production facilities that could allow enhancing maize production in the visited sites   Figure 3.2. Simulation of planting at Koderawa (left) and fertilizer application at Jiratawa (right) University of Ghana http://ugspace.ug.edu.gh 30 3.2.4. Data analysis It should be noted that unlike quantitative methods (e.g. formal questionnaire) which generate information that can be captured numerically, qualitative methods such as RRA and PRA generally do not generate numbers for statistical analysis (Freudenberger, 1999). Their main focus is to explore meanings, processes, reasons and explanations. This is then captured in text or diagrams, but generally not in numbers (Freudenberger, 1999). Therefore most of the results of this study are narrative reports and not supported with statistically analyzed data. They are however worth reporting as an attempt to provide roughly accurate and useful information on the maize production system in Niger. Data on plant population densities and amount of mineral fertilizers applied by farmers obtained from the simulation of field activities is analyzed using Microsoft Excel 2007. Means and standard deviations are computed for a rough comparison of field management practices among farmers from different sites. 3.3. Results 3.3.1. Importance of maize among the cereal crops produced in Niger The RRA study revealed that maize production under rain-fed conditions has significantly reduced in the district of Madarounfa and even in the district of Gaya. In the district of Tchirozerine maize is produced solely under irrigation. Production under rain-fed conditions in the first two districts was supplanted by crops such as pearl millet, cowpea, sorghum and rice. However, maize is the first cereal crop grown under irrigation in all the three districts even though it is competing for the first place with wheat under irrigation in the district of Tchirozerine. The average land holding per farmer hardly exceeds a hectare. Maize is mostly University of Ghana http://ugspace.ug.edu.gh 31 associated with legume crops for the production under irrigation in the district of Madarounfa. Maize sole crop production is found in the district of Gaya under rain-fed conditions and in the district of Tchirozerine under irrigation during the hot period (June to September). 3.3.2. Utilization of maize produced in Niger Maize produced in Niger cannot meet the total national demand; it is therefore used for internal consumption. Maize grain is basically used for preparing traditional dishes such as the famous Tuwo with okra sauce. A large part of the national production is consumed as green maize either roasted or boiled while the green stalks are used as fodder for cattle and sheep. Apart from the two villages of the district of Tchirozerine where maize is produced for farmers’ consumption, the majority of farmers interviewed were producing maize as a cash crop. In the villages of the district of Madarounfa, farmers reported that the beginning of the rainy season coincides with the period when farmers are the most vulnerable and when they also need to buy inputs for the new season. The green maize ears produced are therefore sold in the neighboring cities in order to raise cash to buy inputs and to meet other family needs. In the case of the majority of female farmers of the two districts of Madarounfa and Gaya, maize is not sold as green ears but kept until complete maturity. The production is then conserved for use during special traditional celebrations or in cases of family emergency. 3.3.3. Farming practices used by maize growers in Niger The RRA findings revealed that crop management practices tend to be similar among farmers from the same locality. However, the ability to cultivate and manage a maize field according to what is recommended by the agronomic research depends on the awareness of farmers, the training they had received, and their financial resources. Few of the farmers University of Ghana http://ugspace.ug.edu.gh 32 interviewed were aware of good crop management practices but could not adopt them for lack of financial resources. The RRA study also established that farmers from the eight villages practice land preparation with animal drawn ploughs before sowing but only a few of them apply manure because of lack of availability. Plants are inappropriately thinned to three per hill in all the villages. In Yelou and Bengou the plants were even thinned to four plants per hill. The weeding was done manually three or four times during plant growth and development depending on weed infestation and the ability of farmers to afford more weed control. Despite the thinning of plants to three or four per stand, the final plant population densitiy used by farmers remains low (around 44,000 plants/ha). Results from the simulation of planting by farmers showed that 75% of them used plant population density below the recommended level (62,500 plants/ha) (Figure 3.3). Plants population density (PPD) Very low: PPD <=35000 plants/ha; Low: 35000 65000 plts/ha Figure 3. 3. Percentage of farmers applying different plants population densities across study sites University of Ghana http://ugspace.ug.edu.gh 33 Results of the survey in the districts of Madarounfa and Gaya showed that farmers from the villages of Jiratawa and Yelou tended to use plant densities close to the recommended level. The lowest plant densities were observed at Guidan Gaba and Bengou (Figure 3.4). Figure 3.4. Plant population density adopted by farmers in the villages surveyed The majority of farmers interviewed were aware of the right type of mineral fertilizers to apply. However, few of them were aware of the quantity of fertilizer recommended per hectare, the right time to apply the fertilizer and the appropriate method to use for the application. The common practice used by farmers is the placement of the dose of fertilizer on the ground at the base of the maize stalks. Results across the visited sites revealed a low application of mineral fertilizers by farmers particularly the compound fertilizers NPK. Ninety percent and 65% of farmers across the villages applied NPK and Urea fertilizers respectively far below the recommended rates (Figure 3.5). 0 10000 20000 30000 40000 50000 60000 Jiratawa Koderawa Riadi Yelou Bengou Guidan Gaba 54242 37458 39062 53325 35986 33015 Mean Std Dev Recommended level Population density (plants/ha Villages surveyed during the RRA University of Ghana http://ugspace.ug.edu.gh 34 0 20 40 60 80 100 120 140 160 180 200 Jiratawa Koderawa Riadi Yelou Bengou Guidan G. NPK Urea Quantity of fertilizers (Kg/ha) Recommended NPK level Recommended Urea level Villages surveyed during the RRA Figure 3.5. Percentage of farmers applying different level of mineral fertilizers across the sites of the study Like for plant population densities, farmers from Jiratawa and Yelou also tended to apply the highest levels of mineral fertilizers (Figure 3.6.) Figure 3.6. Quantity of mineral fertilizers applied by farmers in the villages surveyed 1 1 Levels of NPK fertilizers Very low: NPK <= 100 kg/ha; Low: 100 kg/ha < NPK <= 200 kg/ha: Recommended: 200 kg/ha 250 kg/ha 2 Levels of Urea fertilizers Very low: Urea <= 50 kg/ha; Low: 50 kg/ha < Urea <= 100 kg/ha Recommended: 100 kg/ha 150 kg/ha 2 University of Ghana http://ugspace.ug.edu.gh 35 3.3.4. Maize traits preferred by farmers There was a common agreement among the farmers interviewed that earliness and yield are the most important traits considered in the adoption of new maize cultivars. The second most important criterion used in varietal selection is the grain color. The yellow endosperm is preferred by the majority of the farmers. The maize yellow endosperm grain is believed to be more delicious than the white endosperm grain type. Farmers also indicated that the yellow grain endosperm provides more flour for preparing the popular local meal "tuwo massara" than the white grain type. However, in the district of Tchirozérine, farmers preferred the white endosperm. It seemed that farmers did not believe much in traits conferring tolerance to abiotic stresses. Apart from the earliness, none of these traits was ranked among the criteria used in varietal selection. 3.3.5. Major factors limiting maize production in Niger Low soil fertility and the lack of availability of fertilizers at affordable prices constituted the most important constraints to the production of maize under irrigation as well as under rain-fed conditions. The second most important constraint is the lack of availability of good quality seeds of improved varieties. Farmers reported that seed saving and seed exchanges among farmers remain the major sources of maize seeds each growing season. In the district of Madarounfa farmers mentioned the African stem borer (Sesamina calamistis) (Figure 3.7) attack during the dry and cold seasons and bird damage as another major production constraint. Since 2008, the ring-necked parakeet (Psittacula krameri) (Figure 3.8) is reported by farmers to have become a serious constraint to green maize production. In the village of Bengou farmers added that maize production was also challenged by flood and field devastation by grazing animals. Drought was not identified among the constraints to maize production listed by farmers. However, when the University of Ghana http://ugspace.ug.edu.gh 36 issue of drought was discussed by farmers, it came out that the significant changes observed in the rainfall distribution pattern represents the most important cause of the reduction of maize cultivation under rain-fed conditions in Niger. When drought became frequent, farmers preferred to grow other crops which are more tolerant to drought such as pearl millet, cowpea and sorghum. Only very few of them still grow extra-early maize cultivars on very limited areas. Figure 3.7. Maize stem borer (Sesamina calamistis) (left) attack in maize field at Jiratawa (right) Figure 3.8 The ring-necked parakeet (Psittacula krameri) (left) attack on maize cob at Jiratawa (right) P. krameri image source: (Sevcik, 2014) University of Ghana http://ugspace.ug.edu.gh 37 3.3.6. Availability of production facilities for increasing maize production in the visited sites The transect walk carried out in the study sites and the group discussion with farmers revealed that maize grain yield and maize productivity are significantly higher in areas where irrigation facilities are available than in areas where farmers rely only on the rainfall for maize production. This suggest that the expansion of maize production in Niger will depend largely on the development of irrigation facilities. The Republic of Niger has an important potential of underground water and the authorities are currently investing in irrigation development. The irrigation facilities developed vary from the classic animal drawn water at Emaléoulé in the district of Tchirozerine, to artesian boreholes connected with water reservoir for flood or drip irrigation at Yélou in the district of Gaya, and to electric powered boreholes at Jiratawa in the district of Madarounfa. 3.4. Discussion The RRA study findings suggest that maize production in Niger is carried out under poor crop management practices. This is certainly one important reason for the low grain yield recorded in the country. These findings should be taken into consideration by the policy-makers whose ambition is to boost the national production by facilitating farmers’ access to seeds of improved varieties and other agricultural inputs. This approach should be accompanied by the creation of awareness and training of farmers on appropriate and effective crop management practices. The incredible increase in maize grain yield recorded in the USA from 1930 to date is not only due to the wide adoption of maize hybrids but also to changes in crop management practices such as increased nitrogen fertilizer application and higher plant densities (Russell, 1991; Duvick, 1999; Duvick, 2005b). University of Ghana http://ugspace.ug.edu.gh 38 Despite farmers’ knowledge of the importance of organic and mineral fertilizers in maize production, the quantity of fertilizers applied by farmers remains very low. This is essentially due to the inability of the majority of farmers to afford enough fertilizers particularly at the beginning of the rainy season which coincides with the period where farmers’ food stock is low and, thus, are the most vulnerable. Mineral fertilizer prices on the market at that period in Niger are even higher than the main staple food crops such as millet, maize and sorghum. The issue of low application of mineral fertilizers because of their high prices by farmers has been reported by many authors as one of the major factors contributing to low crop productivity in Africa (Heisey and Mwangi, 1996; Gruhn et al., 2000; Morris et al., 2007; Denning, 2009; FARA, 2009; Druilhe and Barreiro-Hurlé, 2012). In addition to the low use of mineral fertilizers, inappropriate application consisting of placing the fertilizer dose on the ground near the plant stalks contributes to an inefficient nutrient uptake by plants. A study on nitrogen fertilizers conducted in Niger revealed that only 20 percent of the applied nitrogen is found in the harvested crop (Christianson and Vlek, 1991). However the low amount of mineral fertilizers used by farmers is justified where soil moisture is a limiting factor. In environments exposed to frequent drought, high nitrogen application may actually diminish yield (Christianson and Vlek, 1991; Morris et al., 2007). The study revealed the use of plant densities below the level of 62,500 plants/ha recommended by the National Agronomic Research Institute (INRAN). Optimum plant densities for maximum maize grain yield depends on water availability, soil fertility, maturity period, planting date and row spacing (Sangoi, 2000). Given the low soil fertility and harsh climatic conditions in Niger, the adoption by farmers of low population densities under rain-fed conditions may be a reasonable practice. Maize grain yield has been reported to be dramatically University of Ghana http://ugspace.ug.edu.gh 39 reduced under high plant densities when water and nitrogen supply are limiting factors (Keating and Wafula, 1990; Sangoi, 2000; Fasoula and Tollenaar, 2005). However, there is no single recommendation for plant population densities for all conditions (Sangoi, 2000). It is likely that early maturing genotypes (which are the types grown in Niger) may require higher plant densities for maximum yield than the late maturing genotypes (Tollenaar, 1992; Sangoi et al., 2002). A study in pearl millet conducted in Niger revealed that under drought conditions and four different levels of N fertilizer, yield reduction under relatively high planting density i.e. 60,000 plants/ha was not significantly different from that at the low planting density i.e.10,000 plts/ha (Christianson and Vlek, 1991) suggesting that low plant density does not guarantee a better yield in water limiting environments. The higher plant densities observed in the villages of Jiratawa and Yelou may be explained by the collaboration and the training received by farmers from the two main agronomic research institutions in the country i.e. INRAN and ICRISAT. Low soil fertility is a serious concern to maize production at all the study sites as to any crop production in the whole country. Many reports indicate that lack of moisture is a major constraint to crop production in West Africa but that poor soil fertility is a more serious concern in the long run (Bationo and Mokwunye, 1991; Gruhn et al., 2000). In general, soils in Niger have low organic matter content and are naturally deficient in phosphorus and nitrogen. In addition to these, continuous cropping over decades with no measures in place to regenerate the soil productivity has worsened the soil fertility status and the level of crop production in the country as in many African countries (Keating and Wafula, 1990; Smithson and Giller, 2002; FARA, 2009). Another important challenge to maize production in the country is the lack of availability of improved varieties. Maize has not been considered as a priority crop in the breeding programs University of Ghana http://ugspace.ug.edu.gh 40 of INRAN. Up to now very few early maturing yellow-grained OPVs (P3Kollo, CET) have been released by the institute. Some OPVs (e.g. MAKA, EV 84-22 RS) have been introduced in the country through research collaboration with regional and international research institutions (MA-Niger, 2012). However P3 Kollo is still the most widely grown maize variety because of its shorter cycle (75 to 85 days) even though its yield potential rarely exceeds 4 tons/ ha under optimal growing conditions. Stemborer attack and bird damage were also reported by farmers to be a big threat to maize production particularly in the villages of the district of Madarounfa. The African stemborer Sesamina calamistis attack is widely reported at these sites. A similar stemborer species Busseola fusca has not been reported in Niger but it is present in almost all the neighboring countries as well as in Central, Eastern and Southern Africa (Harris and Nwanze, 1992). Sesamina infestation is observed on maize mostly during the dry and cold seasons i.e. from October to February. Signs of the attack appear 2-3 weeks after planting and increase in severity until the harvest; plant growth is therefore highly affected (Kouamé, 1995; Kalule et al., 1997; Oigiangbe et al., 1997). However the stemborer populations and the degree of infestation vary over seasons and from one site to another (VandenBerg, 2012). Bird damage is mostly caused by the ring-necked parakeet Psittacula krameri krameri whose populations are found from Senegal to Uganda and in South Sudan (Butler, 2003; Ahmad et al., 2011). The Asian sub species P.k. parvirostris was reported to reduce maize crop yields by up to 81% (Reddy and Long cited in Butler, 2003). In Niger P.k. krameri damage was observed on forest trees but was recently reported on mango, guava, and maize. Ahmad et al. (2011) reported that the damage on maize caused by P. krameri is highest at the maturity stage (23.8%) and minimum (7.60%) at the germination stage. University of Ghana http://ugspace.ug.edu.gh 41 Niger is one of the Sahelian countries whose agriculture is frequently challenged by drought. Yet farmers at the study sites did not rank drought among the major constraints to maize production. Farmers mentioned problems for which they think there may be a solution. They believe that drought is not among such problems. It is a phenomenon that cannot be under the control of men. Farmers don’t believe much in drought and low nitrogen tolerance in maize particularly for the type of severe drought they are experiencing in Niger. Their perception is that maize is an exigent crop that should be grown only when well-watered conditions and good level of soil fertility are guaranteed. Therefore, when moisture becomes a limiting factor farmers prefer to grow crops that are more tolerant to drought like pearl millet, sorghum or cowpea. However, considering the harsh climatic conditions of maize growing areas in Niger, there is no doubt that drought tolerant maize genotypes are more appropriate even where irrigation facilities are available. Indeed water availability for irrigation varies even in those areas. For instance the frequency of irrigation in the large public irrigation schemes of Konni and Jiratawa depends on the water level in the dam of Cerasa for Konni and on the availability of electricity for running the boreholes at Jiratawa. Earliness and yield are the most important traits in farmers’ criteria for the selection of maize cultivars at all the visited sites. The two negatively correlated traits are reported to be of major importance to farmers by many authors (Defoer et al., 1997; Abebe et al., 2005; Obaa et al., 2005; Ouma et al., 2010; Sibiya et al., 2013). Early maturing varieties are needed by farmers for different reasons. Obaa et al. (2005) reported that varieties with shorter cycle can escape drought and ensure early and quick provision of cash and food to households for hunger alleviation. They may also be planted twice in environments where rainfall has a bimodal distribution. The next important trait mentioned by farmers is the endosperm color. Yellow- University of Ghana http://ugspace.ug.edu.gh 42 grained maize is desired by farmers for green consumption. It is also reported to have a better taste in local meals whereas white endosperm maize is reported to be preferred in many sub Saharan countries (Sibiya et al., 2013). Traits conferring tolerance to biotic stresses (pests and diseases) and abiotic stresses were not mentioned by farmers probably because genotypes carrying such traits are not available for them to see the difference. Taking into consideration farmers’ preferred traits in a breeding program may increase the chance of adoption of newly released varieties. However, it is difficult to predict farmers’ decisions on whether to adopt or not to adopt a new variety (Doss et al., 2003). The best option is to give farmers choice and let them test the new materials. In the villages of Emaléoulé and Bagzan n’Amass farmers grow a recently introduced white grained and intermediate maturing variety. They preferred it over their local extra-early yellow grained variety for its higher grain yield. A similar case showing farmers’ decision to tolerate late maturing variety when it is associated with higher grain yield and other desirable traits was reported by Tiwari et al. (2009).. 3.5. Conclusions and breeding implications The study was conducted to understand maize crop management practices in Niger, identify maize characteristics desired by farmers and the major constraints to maize production in the country. Some important findings that can have practical implications for the national breeding program and policy were revealed by the study. Maize production in Niger is carried out under poor crop management practices particularly low fertilizer application and low plant population densities. This suggests that more effort is needed in training and creating awareness among farmers on appropriate and effective maize field management practices. This also calls for the government of Niger to implement a national input subsidy program to help farmers increase the rates of fertilizer application for a better response of their crops. Early maturing yellow-grained University of Ghana http://ugspace.ug.edu.gh 43 cultivars with acceptable grain yield are generally preferred by farmers. This suggests that the national breeding program should take into consideration those traits in the development of new maize varieties. Maize production in Niger is challenged mainly by low soil fertility and water limiting conditions suggesting that new maize cultivars should carry low nitrogen and drought tolerance genes. Biotic stresses such as stemborer (Sesamina calamistis) and bird (Psittacula krameri) damage are also serious concerns in some areas. Therefore appropriate control measures for stemborer attack should be taken for the dry season maize production. Also, careful monitoring of P. krameri populations is necessary in order to take appropriate control measures that can prevent this bird from causing serious yield reduction. University of Ghana http://ugspace.ug.edu.gh 44 CHAPTER FOUR 4.0. Combining ability for grain yield and other agronomic traits of early maturing maize inbreds under drought and well-watered conditions 4.1. Introduction Maize is a cereal crop that underpins the household and national food security in many African countries (Gibbon et al., 2007). Its production is frequently challenged by drought. Significant yield losses in maize due to drought are expected to increase with global climate change as temperature rises and rainfall becomes uncertain in key traditional production areas (Campos et al., 2004; Ribaut et al., 2009). In the savannah agro-ecology of West and Central Africa (WCA), all the countries experience drought during the growth cycle of maize (Badu- Apraku et al., 2004). Yield losses due to drought average 15 percent annually in the sub-region (Badu-Apraku and Lum, 2010; Badu-Apraku et al., 2011a) even in areas where total rainfall is reasonably high (Fischer et al., 1982). Drought is probably responsible for a much higher economic loss than indicated because in drought-prone environments, the high probability of drought occurrence dissuades farmers from applying adequate levels of fertilizers and using high planting densities (Fischer et al., 1982; Pswarayi and Vivek, 2008). Adapting crop phenology to rainfall patterns through drought escape mechanisms is one strategy adopted to achieve yield stability in drought-prone environments (Fischer et al., 1982; Bänziger et al., 2004). Fischer et al. (1981) indicated that the most effective means of reducing the effects of drought on maize would be to escape periods of low moisture availability through the manipulation of genotype maturity and planting date. It is obvious that early maturing genotypes are more suited to areas with a short growing period and those facing recurrent terminal drought as they can complete their growing cycle before the occurrence of the drought. University of Ghana http://ugspace.ug.edu.gh 45 They are also preferred by farmers as they provide an early harvest to bridge the "hunger period" before the harvest of full-season crops. They provide flexibility in planting dates which enables multiple plantings in a season to spread risk of losing a single crop due to drought. Lastly, they are suitable for late plantings when there is a delay in the onset of rainfall (Pswarayi and Vivek, 2008; Badu-Apraku et al., 2013b). However, unlike low soil fertility or mineral toxicity, drought stress is strongly dependent upon random weather conditions (Campos et al., 2004). Therefore, the manipulation of the maturity cycle or the planting dates of maize cultivars alone is not an appropriate solution in such environments. The use of genes that confer drought tolerance to improve early and extra-early maturing maize cultivars is the most appropriate alternative to improve productivity and ensure stable production of maize crop in WCA (Campos et al., 2004; Badu-Apraku et al., 2013a). The previous Rapid Rural Appraisal study conducted in eight maize production sites in Niger revealed that the low maize production observed in the country is due to multiple factors such as drought and lack of availability of high yielding and stable maize genotypes. It also revealed that early maturing maize genotypes of yellow type are preferred by farmers. It is therefore important to firstly develop high yielding and stable maize varieties of yellow type that could also display less yield reduction under drought and, secondly, facilitate their accessibility to Nigerien farmers. The present study is an attempt at achieving the first objective and consists of (i) introducing early maturing and drought tolerant yellow maize inbreds in the country, (ii) developing an efficient method for screening maize genotypes under drought, (iii) developing maize single cross hybrids from the introduced inbreds and evaluating their performance under drought and optimum growing conditions, and (i) exploring the combining ability and heterotic pattern of these inbreds for future studies. University of Ghana http://ugspace.ug.edu.gh 46 The international centers for maize improvement IITA and CIMMYT have developed a large number of early maturing maize inbreds that are adapted to the climatic conditions of sub- Saharan African countries. However, information on the combining abilities and the heterotic patterns of these lines is limited (Pswarayi and Vivek, 2008; Badu-Apraku et al., 2011c; Badu- Apraku et al., 2013a; Akinwale et al., 2014). Knowledge and understanding of the heterotic patterns of these inbreds will be very useful in developing high yielding and stable hybrids (Badu-Apraku et al., 2011c). The efficiency of inbred classification into distinct heterotic group depends greatly on the classification method used (Fan et al., 2009). A good heterotic group classification method identifies heterotic groups for which inter-heterotic group crosses produce hybrids superior to the within-group crosses (Fan et al., 2009). Different methods have been utilized for classifying inbreds into heterotic groups. The most widely used classification method is the SCA_PY method which uses specific combining ability with some line-pedigree information and/or field hybrid-yield information to assign a maize line to a heterotic group (Menkir et al., 2004; Fan et al., 2009; Badu-Apraku et al., 2013c). Another method employs various molecular markers to compute genetic similarity or genetic distance to assign maize lines to different heterotic groups (Menkir et al., 2004; Aguiar et al., 2008; Fan et al., 2009; Badu-Apraku et al., 2013a; Akinwale et al., 2014). Fan et al. (2009) introduced the HSGCA method defined as heterotic group’s specific and general combining ability that combines both GCA and SCA to assign inbred lines to known heterotic groups. Recently another method designated as heterotic grouping based on GCA of multiple traits (HGCAMT) has been proposed by Badu-Apraku et al. (2013c) to classify extra-early inbreds into heterotic groups. University of Ghana http://ugspace.ug.edu.gh 47 The objectives of the present study are to (i) test the efficiency of CIMMYT/IITA screening method (with some modifications) for drought tolerance in Niger, (ii) examine the combining ability and the mode of gene action modulating drought tolerance in 15 early maturing yellow- grained inbreds, (iii) classify the inbreds into heterotic groups using the SCA_PY and HSGCA methods, and (iv) assess the yield performance and stability of the inbreds and hybrids under contrasting environments. 4.2. Materials and methods 4.2.1. Germplasm and mating design Fifteen early maturing tropical yellow-grained maize inbred lines from IITA and CIMMYT were used in this study. Most of these inbreds were developed under environmental stresses such as drought, low N and Striga but information about their combining ability is limited. The characteristics of these inbreds are given in Table 4.1. The fifteen inbred lines were crossed in a 15 x 15 diallel at the National Agronomic Research Institute (INRAN) of Maradi (Niger) to generate 105 single cross hybrids. The diallel crosses and the reciprocals were bulked to obtain enough seed for the field evaluations. It was assumed that no maternal effects on drought tolerance in maize occur. Four early maturing single crosses from IITA (TZEI 9 x TZEI 16, TZEQI 82 x TZEQI 93, TZEI 16 x TZEI 8, and TZEI 23 x TZEI 13) and an open-pollinated variety (OPV) P3 Kollo from INRAN, Niger were used as checks for the evaluations. University of Ghana http://ugspace.ug.edu.gh 48 Table 4.1. Origin, pedigree and reaction to abiotic stresses of 15 inbred lines used in the diallel study No Entries Origin Pedigree Reaction to : Striga Drought 1 ENT 4 CIMMYT [[KILIMA ST94A]-30/MSV-03-2-10-B-1-B-B-xP84c1 F26- 2-2-6-B-3-B] F17-1-2-1-1 x P43C9-1-1-1-1-1-BBBB-1-xP Resistant Tolerant 2 ENT 13 CIMMYT [M37W/ZM607#bF37sr-2-3sr-6-…1-B x CML486]-1-1 Susceptible Tolerant 3 ENT 15 CIMMYT CLA149 Tolerant Tolerant 4 ENT 17 CIMMYT [(87036/87923)-X-800-3-1-X-1-B-B-1-1-1 Susceptible Tolerant 5 TZEI 14 IITA TZE Comp5-Y C6 S6 Inbred 21 Tolerant Tolerant 6 TZEI 16 IITA TZE Comp5-Y C6 S6 Inbred 31 Resistant Tolerant 7 TZEI 129 IITA TZE-Y Pop STR Co S6 Inbred 16-1-3 Tolerant Susceptible 8 TZEI 23 IITA TZE-Y Pop STR C0 S6 Inbred 62-2-3 Tolerant Tolerant 9 TZEI 124 IITA TZE-Y Pop STR Co S6 Inbred 3-1-3 Resistant Susceptible 10 TZEI 135 IITA TZE-Y Pop STR Co S6 Inbred 17-2-3 Tolerant Tolerant 11 TZEI 157 IITA TZE-Y Pop STR Co S6 Inbred 102-1-2 Tolerant Tolerant 12 TZEI 161 IITA TZE-Y Pop STR Co S6 Inbred 103-2-3 Resistant Tolerant 13 TZEI 167 IITA TZE Comp5-Y C6 S6 Inbred 13 Resistant Susceptible 14 TZEI 160 IITA TZE-Y Pop STR Co S6 Inbred 102-2-3 Resistant Tolerant 15 TZEI 182 IITA TZE-Y Pop STR Co S6 Inbred 152-2-2 Susceptible Tolerant 4.2.2. Stress management To determine the right time to apply the last irrigation in drought trials in Niger, the drought screening method described by Bänziger et al., (2000) and Badu-Apraku et al. (2004) was used in an experiment conducted during the dry season 2011/2012 at SEHA (the hydro- agricultural experimental station of INRAN Maradi). The local OPV check P3 Kollo and a bulk of F2 seeds from an IITA DT hybrid (TZEE 95 x TZEE 82) were used in the experiment. The two genotypes were sown at five planting dates at five days intervals. For each planting date a block of 10 rows was established for each genotype. An adjacent block of 10 rows of each University of Ghana http://ugspace.ug.edu.gh 49 genotype was also established with plants grown under well-watered conditions until maturity. All five blocks were irrigated twice a week at the same time. The last irrigation for all five blocks was carried out when plants in the first block were at about one week to anthesis; that is when 50% of plants in that block had hard and swollen flag leaves, an indication of an imminent tassel emergence. A “rescue irrigation” was applied to all the blocks 10 days later. The normal irrigation was resumed 10 days after the “rescue irrigation” until the maturity of plants. The right time to stop irrigation was determined as the one for which about 80 to 85% of yield reduction was observed compared to the yield of the adjacent well-watered block. 4.2.3. Experimental design and evaluation sites The 105 single crosses and the 5 checks were evaluated under managed drought and well- watered conditions using a 10 x 11 alpha lattice design with 2 replications at the INRAN Research Stations at Maradi (N 13º26’ E 7º06’; 350 m altitude; 490 mm average annual rainfall), Konni (N 13º47’ E 5º14’; 260 m altitude; 530 mm average annual rainfall), and at FESA Seed Company (N 13º 39’ E 7º 04’; 382 m altitude; 500 mm average annual rainfall). The evaluation was done under managed drought and well-watered conditions during the dry season 2012/2013 (October, 2012 to March, 2013) at INRAN Stations at Maradi and Konni. During the rainy season of 2013, the evaluation was done under well-watered conditions at FESA and INRAN Maradi. The second evaluation under managed drought was done during the dry season 2013/2014 (October, 2013 to March, 2014) at INRAN Maradi. In summary, the evaluation was done at the following seven environments (four well-watered environments and three stressed environments): University of Ghana http://ugspace.ug.edu.gh 50 1. MDS12WWC: Maradi (M), cold and dry season 2012/2013 i.e. October, 2012 to March, 2013 (DS12), under well-watered conditions (WWC); 2. KDS12WWC: Konni (K), cold and dry season 2012/2013 i.e. October, 2012 to March, 2013 (DS12), under well-watered conditions (WWC); 3. MRS13WWC: Maradi (M), rainy season 2013 (RS13), under well-watered conditions (WWC); 4. FRS13WWC: FESA Seed Co (F), rainy season 2013 (RS13), under well-watered conditions (WWC); 5. MDS12WSC: Maradi (M), cold and dry season 2012/2013 i.e. October, 2012 to March, 2013 (DS12), under water stress conditions (WSC); 6. KDS12WSC: Konni (K), cold and dry season 2012/2013 i.e. October, 2012 to March, 2013 (DS12), under water stress conditions (WSC); 7. MDS13WSC: Maradi (M), cold and dry season 2013/2014 i.e. October, 2013 to March, 2014 (DS13), under water stress conditions (WSC); It is important to note that during the first part of the dry season in Niger (November - January) the cold weather conditions prolong the maturity cycle of maize plants by about 20 days compared to the maturity cycle of the rainy season. 4.2.4. Field practices and stress management Except for the water treatment, all the field management practices were similar in both the stressed and the well-watered blocks. Each entry was planted in a single row of 5m length with spacing of 40 cm within rows and 75 cm between rows. Three seeds were sown per hill at University of Ghana http://ugspace.ug.edu.gh 51 planting and the seedlings thinned to two plants per hill about two weeks after emergence to give a final population density of 66 000 plants ha–1. A compound fertilizer (NPK-15-15-15) was applied at the rate of 52 kg N ha–1, 52 kg P ha–1, and 52 kg K ha–1 as follows: at planting 2 grams of NPK fertilizer was placed in each hill with the seeds; the remaining NPK fertilizer (9 grams of NPK /hill) was applied two weeks after planting. Urea was top-dressed at the rate of 46 kg of N ha-1 20 days later. Weed control was done manually. The time to impose water stress was determined from the drought screening experiment described above. From planting to water stress imposition plants were irrigated twice a week in all the blocks. Irrigation continued in the well-watered blocks until crop maturity. In the stressed blocks, the water stress was imposed by stopping the irrigation two weeks before flowering. It is important to note that in Niger, during the dry season (October to May) the weather conditions are so harsh that plants could not have survived if the stress had been maintained until the end of the experiment, particularly at Maradi. Therefore the stress intensity was monitored and a “rescue irrigation” was applied when plants started showing signs of severe stress very early in the morning. In 2012, the water stress was imposed at 48 days after planting at Maradi. A“rescue irrigation” was applied 12 days later and the normal irrigation was resumed 10 days after the “rescue irrigation”. In 2012 the stress was imposed at Konni 44 days after planting. A “rescue irrigation” was applied 17 days later and the normal irrigation resumed 11 days after the “rescue irrigation”. In 2014 the stress was imposed at Maradi 50 days after planting. A “rescue irrigation” was applied 17 days later and the normal irrigation resumed 10 days after the “rescue irrigation”. University of Ghana http://ugspace.ug.edu.gh 52 4.2.5. Data collection Data for grain yield and other agronomic traits were recorded as follows: DYA (days to anthesis) is the number of days from planting to when 50% of the plants had shed pollen; DYS (days to silking) is the number of days from planting to when 50% of the plants had emerged silks; ASI (anthesis-silking interval) is the number of days between DYA and DYS; PLTH (Plant height) measured as the distance from the base of the plant to the first tassel branch was computed as the average height of 10 plants randomly chosen in the row; EHT (ear height) measured as the distance from the base of the plant to the node bearing the upper ear and was computed as the average ear height of 10 plants randomly chosen in the row; PASP (plant aspect) estimated based on the overall plant appeal on a scale of 1 to 5, where 1 = excellent plant type and 5 = poor plant type; HUSK (husk cover) rated on a scale of 1 to 5, where 1 = husks tightly arranged and extended beyond the ear tip and 5 = open tip cover. NPHARV (number of plants harvested) the number of plants harvested per plot EHARV (number of ears harvested) the number of ears bearing at least one developed kernel in the plot; EPP (number of ears per plant) estimated by dividing the total number of ears per plot by the number of plants harvested; EASP (ear aspect) estimated based on absence of disease and insect damage, ear size, uniformity of ears, and grain filling. It was scored on a scale of 1 to 5, where 1 = clean, uniform, large, and well-filled ears and 5 = ears with undesirable features. University of Ghana http://ugspace.ug.edu.gh 53 ADJGRY: grain yield adjusted to 15% moisture content. It was computed using the formula: Wf =Wi * ((100-MCi)/(100-MCf)) where Wf is the final grain weight; Wi is the initial grain weight; MCi is the initial moisture content and MCf is the final moisture content (15%); YLD_HA (grain yield adjusted to 15% moisture and converted to kg ha-1). This was computed from the shelled grain weight and grain moisture at harvest. SGREEN1 (stay-green characteristic) scored for the water stressed blocks at 90 days after planting on a scale of 1 to 9, where 1 = almost all leaves green and 9 = virtually all leaves dead; SGREEN2 (stay-green characteristic) scored for the water stressed blocks at 100 days after planting on a scale of 1 to 9, where 1 = almost all leaves green and 9 = virtually all leaves dead. 4.2.6. Statistical analysis Analysis of variance (ANOVA) was performed on plot means for grain yield and other agronomic characters for each and across environments using PROC GLM procedure of SAS software, version 9.2 TS2M0 (SAS Institute, 2002). In the combined ANOVA, the combination of location, cropping-season, and water treatment was considered as an environment. The entries (105 single-cross hybrids and 5 checks) were considered as fixed factors while environments, replications and blocks were considered as random factors. The statistical model used for the combined analysis is as follows: Yijkl = μ + αl + bkl + vij + (αv)ijl + eijkl Yijkl = observed trait value from each experimental unit (i and j = parents; k =replication, l = location); µ = population mean; αl = environment effect; bkl = block or replication within environment effect; vij = F1 hybrid effect; (αv)ijl = interaction between F1 hybrids and environments; eiikl = residual effect (Zhang and Kang, 1997). University of Ghana http://ugspace.ug.edu.gh 54 To identify productive single-cross hybrids under stress conditions, a correlation matrix displaying relationship between grain yield and secondary traits was computed. Then a selection index similar to that used by Badu-Apraku et al. (2011d) was computed using standardized data for selected variables that were highly and significantly correlated to grain yield: I = [(2 × Yield) + PLTH – ASI – PASP – EASP – SG2] PLTH = plant height, ASI = anthesis-silking interval; EASP = ear aspect; PASP = plant aspect; SG2 = stay green characteristic 2 Genetic analysis of the GCA effects of the parents and SCA effects of the crosses in each environment and across environments was performed using the Griffing’s method 4 and model 1 (Griffing, 1956) and the DIALLE-SAS05 program (Zhang et al., 2005). Only the crosses (excluding the checks) were used in the diallel analysis. The statistical model used for the combined diallel analysis across environments is the following: Yijk = μ + Ee + gi + gj + sij + gEeg + sEes + εijk - Yijk : the observed measurement for the ijth cross grown in the kth environment - μ: grand mean - Ee: effect of environment - gi & gj: GCA effects; (i and j = parents; i=1…15; j=1…15; k =replications) - sij: SCA effects; - gEeg: interaction effect between GCA and the environment; - sEes: interaction effect between SCA and the environment, - εijk: error term associated with the ijth cross evaluated in the kth replication and the Ee environment. University of Ghana http://ugspace.ug.edu.gh 55 GCA and SCA effects for the measured traits were computed from the mean values adjusted for the block effects for each environment and across environments: GCA = Line mean (X.j) – Overall mean (X..) SCA = Cross mean (Xij) – Line mean (X .j) – Tester mean (X i.) + Overall mean (X..) i and j = parents, i= 1 to15; j= 1 to 15 (Fan et al., 2009) The relative importance of general combining ability and specific combining ability on the performance of the F1 hybrids was estimated as the ratio: where σ2GCA is the quadratic form (analogous to a variance component but referring to a fixed effect) derived from the mean square of GCA effect and σ2SCA is the quadratic form of SCA effects since the total genetic variation among single-cross progeny is equal to twice the GCA component plus the SCA component (Hung and Holland, 2012). Two different classification methods i.e. SCA_PY method and HSGCA method were used to assign the 15 early maturing inbred lines used in the diallel study. The grouping efficiencies of the two classification methods were also compared. ­ SCA_PY method: the classical method based on the use of specific combining ability with some line-pedigree information and/or field hybrid-yield information (Fan et al., 2009). The procedure used is the one described by Menkir et al., (2004). ­ HSGCA method: the heterotic group’s specific and general combining ability proposed by Fan et al., (2009) and computed as follows: University of Ghana http://ugspace.ug.edu.gh 56 HSGCA = Cross mean X ij – Tester mean (Xi.) = GCA + SCA ­ where X ij is the mean yield of the cross between ith tester and jth line; ­ X j. is the mean yield of the ith tester; ­ X.j is the mean yield of jth line. The computed HSGCA values were subjected to the three classification steps described by Fan et al., (2009). The GGE biplot analysis (Yan and Tinker, 2006) was used on the grain yield means adjusted for block effects to obtain information on yield performance and stability of the single-cross hybrids across test environments. It was also applied to the estimated GCA effects of the inbred parents across environments for stability analysis. The GGE biplot model used is: Yij- βj = λ1ξi1ηj1 + λ2ξi2ηj2 + εij - Yij is the genetic value (yield) of the combination (hybrid) between Entry i (hybrid) and Tester j (environment) for the trait of interest; - βj is the mean of all combinations involving Tester j; - λ1, and λ2 are the singular values for PC1 and PC2; - ξi1 and ξi2 are the PC1 and PC2 eigenvectors, respectively, for Entry i; - ηj1 and ηj2 are the PC1 and PC2 eigenvectors, respectively, for Tester j; - εij is the residual of the model associated with the combination of Entry i and Tester j. University of Ghana http://ugspace.ug.edu.gh 57 4.3. Results 4.3.1. Identification of the time to impose water stress in maize drought studies in Niger In the drought experiment conducted to identify the right time to impose water stress for maize genotypes evaluated under managed drought in Niger, data from two blocks corresponding to the 4th and the 5th planting dates for P3 Kollo was discarded because of high soil heterogeneity observed in the field and termite damage. Data for the number of days to anthesis was not recorded because more than 50 percent of plants had tassel blast and did not shed pollen. Therefore it was not possible to compute the anthesis - silking interval. The highest grain yields i.e. 3,627 Kg ha-1 and 2,369 Kg ha-1 displayed respectively by P3 Kollo and the F2 of the cross TZEEI 95 x TZEEI 82 were obtained from the well-watered blocks. The lowest grain yields were observed in the first and second planting dates for which irrigation was stopped 7 and 12 days before flowering, respectively (Table 4.2 and Figure 4.1). Table 4.2. Grain yield of the drought experiment conducted to identify the right time to impose moisture stress in maize drought trials in Niger. Genotypes Grain Yield in Kg ha-1 adjusted to 15% moisture content WWC* WSC* 1 WSC 2 WSC 3 WSC 4 WSC 5 (TZEEI 95 x TZEEI 82) F2 2,369 322 397 800 753 1097 P3 Kollo 3,627 462 587 922 - - *WWC: Well-Watered Conditions; *WSC: Water Stress Conditions; WSC 1: water stress imposed 1week before anthesis; WSC 2: water stress imposed 12 days before anthesis; WSC 3: water stress imposed 17 days before anthesis; WSC 4: water stress imposed 22 days before anthesis; WSC 5: water stress imposed 27 days before anthesis University of Ghana http://ugspace.ug.edu.gh 58 Figure 4.1. Variation in grain yield depending on the water stress intensity in the drought experiment conducted to identify the right time to impose water stress in maize drought trials in Niger. WWC: well- watered conditions; WSC: water stress conditions; PD: planting date; 1st PD: water stress imposed 1week before anthesis; 2nd PD: water stress imposed 12 days before anthesis; 3rd PD: water stress imposed 17 days before anthesis; 4th PD: water stress imposed 22 days before anthesis; 5th PD: water stress imposed 27 days before anthesis CIMMYT recommends to impose water stress by stopping irrigation when one can predict that the stress will be severe enough to bring the yield level of the stressed blocks to about 15-20% of the yield of the well-watered blocks (Bänziger et al., 2000). Such stress level was observed in this experiment in the case of the first and the second planting dates for which irrigation was stopped 7 and 12 days before anthesis respectively (Figure 4.2). However for the first and the second planting dates, the stress intensity observed in the field was too severe and a large number of plants collapsed even before the anthesis. It is therefore advisable to target a moderate stress level for future maize drought trials in Niger. This could be obtained if water stress is imposed 3 to 2 weeks before anthesis. A “rescue irrigation” should be applied 13 to15 days later. Then the normal irrigation should be resumed 10 to 12 days after the “rescue irrigation” and continued until plant maturity. University of Ghana http://ugspace.ug.edu.gh 59 Figure 4.2. Grain yield reduction depending on the stress intensity from the drought experiment conducted to identify the right time to impose water stress in maize drought trials in Niger. PD: planting date; 1st PD: water stress imposed 1week before anthesis; 2nd PD: water stress imposed 12 days before anthesis; 3rd PD: water stress imposed 17 days before anthesis; 4th PD: water stress imposed 22 days before anthesis; 5th PD: water stress imposed 27 days before anthesis. 4.3.2. Performance of single cross hybrids, inbred parents and checks under drought and well-watered conditions The analysis of variance of hybrids evaluated under drought stress showed significant genotype (G), environments (E), and genotype x environment interaction (GEI) mean squares for all the measured traits except the E for ear aspect and the GEI for the number of days to silking, plant and ear height, plant aspect, and number of ears per plant (Appendix 4.1). The ANOVA across the well-watered environments also revealed significant mean squares for all the traits measured for genotypes, environments, and GEI (Appendix 4.1). Similarly the combined analysis of variance across the seven environments revealed that genotypes, environments and GE interactions were significantly different for all the traits measured (Table 4.3). The test University of Ghana http://ugspace.ug.edu.gh 60 environments contributed to about 65.6% of the total variation in grain yield; while the genotypes accounted for about 9.5% and GE sources of variation, about 13.6 % (Table 4.4). Mean grain yield of the single crosses under drought stress was 1684 kg ha-1 representing 36.84% of the well-watered yield which was 4571 kg ha-1 (Appendix 4.1). Under drought stress, grain yield ranged from 594 kg ha-1 for ENT 15 x ENT 4 to 2753 kg ha-1 for TZEI 182 x ENT 13 (Table 4.5). Under drought, eight single-cross hybrids out- yielded the best check, TZEQI 82 x TZEQI 93 (2309 kg ha-1) (Appendix 4.1). The top five hybrids in terms of grain yield under drought stress, TZEI 182 x ENT 13 (2753 kg ha-1), TZEI 160 x TZEI 157 (2570 kg ha-1), ENT 13 x TZEI 157 (2560 kg ha-1), TZEI 182 x TZEI 161 (2401 kg ha-1), and TZEI 129 x ENT 17 (2399 kg ha-1) out- yielded the best drought-tolerant single cross check by 19.22%, 11.30%, 10.83%, 3.95% and 3.87%, respectively (Table 4.5). Across well-watered conditions, grain yield ranged from 1995 kg ha-1 for TZEI 14 x TZEI 16 to 6584 kg ha-1 for TZEI 124 x ENT 13 (Table 4.5). The four highest yielding hybrids were TZEI 124 x ENT 13 (6584 kg ha-1), TZEI 124 x TZEI 23 (5806 kg ha-1), TZEI 167 x ENT 13 (5711 kg ha-1), and TZEI 160 x TZEI 157 (5708 kg ha-1). The highest yielding hybrid, TZEI 124 x ENT 13 out-yielded the 5 checks by 23 to 42% (Table 4.5). The intensity of the induced water stress estimated as the ratio of the average grain yield from the water stress environments to that of the well-watered environments was 36.84% falling outside the limits of 15-20% or 20-30% proposed by Bänziger et al. (2000) and Bolaños and Edmeades (1996), respectively. University of Ghana http://ugspace.ug.edu.gh 61 Table 4.3. Mean squares from the combined analysis of variance of grain yield and other agronomic traits across environments Source of variation DF Yield (kg ha-1) DYA DYS ASI PLTH (cm) EHT (cm) PASP (1-5) EASP (1-5) EPP HUSK (1-5) ENV 6 601167066** 25443** 34902** 2002.94** 81785** 38742** 23.18** 92.03** 6.59** 11.52** Rep(ENV) 7 7324402** 22.92** 54.74** 58.23** 1312.90** 762.23** 1.32** 3.34** 0.08* 0.77** Block(Rep*env) 140 1849399** 6.66** 14.52** 6.64ns 351.00** 129.53* 1.15** 0.67** 0.03ns 0.21ns Entry 109 4783030** 67.64** 91.98** 15.59** 1791.62** 720.26** 2.21** 1.79** 0.07** 1.32** ENV*Entry 654 1145913** 5.30** 9.67** 7.00** 209.95** 86.00* 0.58** 0.41** 0.04** 0.21** Residual 623 664552 3.41 6.79 4.51 161.42 72.71 0.44 0.26 0.03 0.15 YIELD: Grain yield; DYA: days to anthesis; DYS: days to silk; ASI: anthesis silking interval; PLTH: plant height; EHTH: ear height; PASP: plant aspect; EASP: ear aspect; EPP: ears per plant; HUSK: husk cover; ENV: environment; Rep: replication; DF: degree of freedom; *Significant at the 0.05 probability level **Significant at the 0.01 probability level ns: not significant University of Ghana http://ugspace.ug.edu.gh 62 Table 4.4. Proportions of the total variance attributable to the sources of variation for grain yield of maize hybrids evaluated in 7 contrasting environments in Niger from 2012 to 2014 Source of variation DF SS MS % of total variation Total 1539 5494758680 ENV 6 3607002393 601167066 ** 65.64% Rep(ENV) 7 43946412 7324402 ** Block(Rep*env) 140 36987971 1849399 ** Entry 109 521350315 4783030 ** 9.49% ENV*Entry 654 749427354 1145913 ** 13.64% Residual 623 493762429 664552 Model 916 5000996250 6282659 ENV: environment; Rep: replication; DF: degree of freedom; SS: sum of squares; MS: mean squares **Significant at the 0.01 probability level University of Ghana http://ugspace.ug.edu.gh 63 Table 4.5 Grain yield and other agronomic traits of the 5 highest yielding hybrids, the 5 lowest yielding hybrids and the 5 checks evaluated under drought stress and well-watered environments in Niger from 2012 to 2014 (see complete table in Appendix 4.1) Single cross hybrids Yield Kg ha -1 DYS ASI PLTH PASP SGC2 WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WSC TZEI 124 x ENT 13 ¥ 6584 1993 58 77 1 9 194 164 1.1 2.7 6.2 TZEI 124 x TZEI 23 ¥ 5806 1897 56 70 1 6 170 151 2.5 3.2 7.5 TZEI 167 x ENT 13 ¥ 5711 2297 61 77 2 5 181 141 2.4 2.8 6.2 TZEI 160 x TZEI 157 ¥ ‡ 5708 2570 57 72 1 5 182 157 2.8 3.0 6.7 TZEI 16 x TZEI 161 ¥ 5652 2125 59 76 1 7 175 142 2.9 2.7 6.7 ENT 13 x TZEI 157 ‡ 5454 2560 60 75 1 6 195 161 2.1 2.7 5.5 TZEI 182 x ENT 13 ‡ 5416 2753 58 72 1 5 189 158 2.6 2.7 6.0 ENT 15 x TZEI 124 £ 5290 736 61 83 3 12 189 156 2.5 4.2 7.7 TZEI 129 x ENT 17 ‡ 5141 2399 60 75 1 4 191 156 3.0 3.0 6.2 TZEI 124 x ENT 4 £ 5097 649 60 82 2 14 182 154 2.5 4.2 7.3 TZEI 23 x TZEI 13 Check 5063 1992 59 76 1 4 168 134 2.9 3.3 6.5 TZEI 9 x TZEI 16 Check 4981 1835 59 76 1 7 164 144 2.5 3.5 7.0 ENT 17 x TZEI 124 £ 4859 721 60 84 1 10 190 142 2.8 4.5 7.8 P3 Kollo Check 4690 1439 55 70 1 6 172 156 3.0 3.8 6.8 TZEI 16 x TZEI 8 Check 4423 2008 59 76 1 7 156 135 2.8 3.2 7.3 TZEI 182 x TZEI 161‡ 4045 2401 58 71 2 4 166 145 3.4 3.2 4.7 TZEQI 82 x TZEQI 93 Check 3792 2309 65 78 2 4 175 147 2.9 2.3 5.2 TZEI 160 x TZEI 161† 2533 1043 58 77 -1 6 131 111 4.0 4.2 7.3 TZEI 182 x TZEI 157 † 2397 1298 59 74 1 7 150 141 4.5 4.5 5.5 TZEI 167 x TZEI 14 † 2248 1183 66 80 2 4 151 131 3.8 4.2 5.7 ENT 15 x ENT 4 † £ 2021 594 66 84 2 7 167 131 4.4 4.3 7.2 TZEI 14 x TZEI 16 † £ 1995 819 67 85 2 4 150 119 4.1 4.2 6.2 Overall mean 4571 1684 60 76 1 7 174 145 3.0 3.5 6.5 Lsd 865 775 1.7 3.9 1.3 3.2 12 15 0.5 0.9 1.1 Genotypes ** ** ** ** ** ** ** ** ** ** ** Environments ** ** ** ** ** ** ** ** ** * ** G x E ** * ** ns ** ** * ns ** ns * ¥: top yielding hybrids under WWC; †: lowest yielding hybrids under WWC; WWC: well-watered conditions; WSC: water stress conditions; ‡: top yielding hybrids under WSC; £: lowest yielding hybrids under WSC; YIELD: Grain yield; DYS: days to silk; ASI: anthesis silking interval; PLTH: plant height; PASP: plant aspect; SGC2: stay-green 2; *Significant at the 0.05 probability level; **Significant at the 0.01 probability level University of Ghana http://ugspace.ug.edu.gh 64 The evaluation of the parental inbred lines under water stress and well watered conditions at Konni during the dry season 2013/2014 was lost due to a failure of the irrigation system. Therefore, only the results from the evaluation at Maradi were considered for the analysis. The analysis of variance of the grain yield and other agronomic traits revealed significant differences among the inbreds for grain yield and all the measured traits except for ear aspect under drought and the number of ears per plant under well watered conditions (Table 4.6). The average grain yield of the inbreds under drought stress was 670 kg ha-1 representing 36.88% of the well watered yield which was 1,818 kg ha-1. Under drought stress, grain yield ranged from 27 kg ha-1 to 1605 kg ha-1 (Table 4.6). The five top yielding inbreds under drought stress were ENT 13, ENT 17, TZEI 167, TZEI 161 and TZEI 135 (Table 4.6). Inbreds TZEI 124 and TZEI 16 were the most sensitive to the water stress. They yielded less than 100 kg ha-1 under drought. Under well-watered conditions, the grain yield ranged from 748 kg ha-1 for TZEI 157 to 3,723 kg ha-1 for TZEI 167. Inbreds TZEI 167, TZEI 135, ENT 13, ENT 15 and ENT 17 were the highest yielding with grain yields of 3723 kg ha-1; 3161 kg ha-1; 2191 kg ha-1; 2130 kg ha-1; and 2102 kg ha-1 respectively (Table 4.6). Across water regimes, TZEI 167, TZEI 135 and ENT 13 were the three top yielding inbreds while TZEI 16 and TZEI 157 were the poorest. University of Ghana http://ugspace.ug.edu.gh 65 Table 4.6 Grain yield and other agronomic traits of the 15 inbreds evaluated under drought stress and well-watered conditions at Maradi during the dry season of 2013/2014 Entry YIELD Kg ha-1 DYA DYS ASI PLHT EHT PASP EASP EPP wwc wsc awr wwc wsc wwc wsc wwc wsc wwc wsc wwc wsc wwc wsc wwc wsc wwc wsc TZEI 160 1588 475 1031 73 76 71 80 -1.5 4 105 107 58 69 3.5 3.5 4.2 4.3 1.3 0.5 TZEI 129 1781 303 1042 75 76 73 82 -1.5 5.5 168 154 90 99 2.5 3.5 3.6 4.4 1.1 0.4 ENT 15 2130 329 1230 76 76 78 89 2.5 13 145 122 70 78 2.5 4.5 3.3 3.8 1.0 0.4 ENT 17 2102 1265 1683 78 79 78 81 -0.5 2 128 123 73 76 2.5 3 3.6 3.9 1.0 0.6 TZEI 182 1036 637 836 71 71 72 79 1 8.5 108 112 55 61 4.5 3.5 4.6 4.6 1.1 0.5 TZEI 167 3723 1243 2483 72 74 71 76 -1 2.5 155 143 80 93 1.0 2 2.7 4.1 1.0 0.7 TZEI 14 1355 862 1109 80 82 82 88 2 5.5 103 116 50 70 3.0 1.5 4.2 4.6 1.0 0.8 TZEI 124 1513 76 794 72 72 75 89 2.5 17 145 140 60 74 3.0 4.5 3.3 3.3 0.9 0.3 TZEI 16 1292 27 659 76 78 79 85 2.5 7.5 145 133 80 82 3.0 3 4.3 2.5 1.1 0.1 TZEI 23 1565 714 1140 70 71 69 72 -0.5 1.5 100 115 53 64 3.5 3 3.8 3.9 1.0 0.6 ENT 4 1522 341 931 77 78 80 92 3 13.5 135 121 70 76 3.0 4.5 3.5 3.0 1.4 0.3 TZEI 161 1564 1009 1287 73 72 72 74 -1 2 123 139 68 88 3.0 2.5 3.8 4.2 1.0 0.8 TZEI 135 3161 868 2015 70 71 72 80 1.5 8.5 165 138 85 83 1.5 2.5 2.6 3.3 1.1 0.5 ENT 13 2191 1605 1898 78 78 79 80 1.5 2.5 150 144 80 92 2.0 2 3.4 3.7 1.1 0.9 TZEI 157 748 302 525 76 75 79 81 3 6.5 130 134 68 87 4.0 3.5 4.8 5.0 1.0 0.6 mean 1818 670 1244 74 75 75 82 0.9 7 134 129 69 80 2.8 3.1 3.7 3.9 1.1 0.5 lsd 1070 768 1198 1.6 4.7 2.6 4.8 2.2 4.2 21 21 18 12 1.1 1.7 1.1 1.6 0.5 0.3 Rep ns ns ns ns ns ns ns ns ns ** ns ** * ns ns ns ns ns ns Entry ** * ** ** ** ** ** ** ** ** ** ** ** ** * * ns ns ** WWC: well-watered conditions; WSC: water stress conditions; AWR: across water regimes; Rep: replication; lsd: least significant difference; YIELD: Grain yield; DYA: days to anthesis; DYS: days to silk; ASI: anthesis silking interval; PLTH: plant height; EHT: ear height; PASP: plant aspect; EASP: ear aspect; EPP: number of ears per plant; *Significant at the 0.05 probability level; **Significant at the 0.01 probability level ; ns: not significant University of Ghana http://ugspace.ug.edu.gh 66 4.3.3. Correlations, selection indices and identification of productive single-cross hybrids for production under drought stress conditions The Pearson correlation analysis revealed significant correlations between the measured traits and the grain yield of the hybrids under drought and well-watered conditions except for ear height and husk cover under well-watered conditions (Table 4.7). Positive correlations were found between plant height, number of ears per plant and grain yield under the two water treatments. In contrast, plant and ear aspects were strongly and negatively correlated to grain yield under the two water regimes. Under drought stress, negative and relatively strong correlations were found between the grain yield and the number of days to anthesis, the number of days to silk, anthesis to silking interval, plant and ear aspects, and stay-green characteristic 2 (Table 4.7). However traits such as the stay- green characteristic 1 and husk cover showed negative but weak associations with grain yield under drought (Table 4.7). Grain yield and other traits displaying strong correlation with yield under drought such as ASI, ear aspect, plant aspect, stay- green characteristic 2 and number of ears per plant were used to compute the selection indices to identify outstanding hybrids under drought conditions. Based on the selection indices under drought stress, significant differences were detected among the single cross hybrids (Appendix 4.1). Three single crosses TZEI 182 x ENT 13, ENT 13 x TZEI 157 and TZEI 182 x TZEI 161 out- performed the best check TZEQI 82 x TZEQI 93 based on the selection index. The best five hybrids under stress based on the selection index (Appendix 4.1) were TZEI 182 x ENT 13 (2753 kg ha-1), ENT 13 x TZEI 157 (2560 kg ha-1), TZEI 182 x TZEI 161 (2401 kg ha-1), TZEQI 82 x TZEQI 93 Check (2309 kg ha-1), and TZEI 160 x TZEI 157 (2570 kg ha-1). University of Ghana http://ugspace.ug.edu.gh 67 Table 4.7. Pearson correlation coefficients between pairs of traits across 105 single cross hybrids and 5 checks evaluated under drought stress (lower diagonal) and well-watered conditions (upper diagonal) in Niger from 2012 to 2014 YIELD DYA DYS ASI PLTH EHT PASP EPP EASP HUSK SG1 SG2 YIELD -0.26 ** -0.30** -0.24** 0.29** -0.03ns -0.69** 0.48** -0.78** -0.03ns - - DYA -0.49 ** 0.99** -0.15** 0.52** 0.72** 0.18** -0.17** 0.04ns 0.02ns - - DYS -0.67 ** 0.77** 0.01ns 0.52** 0.73** 0.20** -0.20** 0.08** 0.03ns - - ASI -0.50 ** 0.08* 0.69** -0.06ns -0.03ns 0.14** -0.21** 0.22** 0.07* - - PLTH 0.45 ** -0.33** -0.37** -0.21** 0.82** -0.29** -0.04ns -0.48** 0.00ns - - EHT 0.29 ** -0.09* -0.10* -0.04ns 0.77** -0.03ns -0.27** -0.15** 0.14** - - PASP -0.70 ** 0.29** 0.47** 0.41** -0.43** -0.26** -0.23** 0.62** 0.07* - - EPP 0.25 ** -0.01ns -0.05 ns -0.07 ns -0.06 ns 0.00 ns -0.18** -0.30** -0.19** - - EASP -0.72 ** 0.31** 0.47** 0.38** -0.57** -0.38** 0.69** -0.14** 0.04ns - - HUSK -0.10 ** 0.17** 0.15** 0.04 ns 0.04 ns 0.17** 0.08 ns -0.17** -0.06 ns - - SG1 -0.29 ** 0.11** 0.26** 0.28** -0.16** 0.04ns 0.26** -0.08* 0.29** 0.09* - SG2 -0.43 ** 0.25** 0.35** 0.26** -0.08* 0.06ns 0.36** -0.22** 0.32** 0.17** 0.44** YIELD: Grain yield; ASI: anthesis-silking interval; PASP: plant aspect; HUSK: husk cover; SG1: stay-green 1; DYA: days to anthesis; PLTH: plant height; EPP: ear per plant; SG2: stay-green 2; DYS: days to silk EHT: ear height EASP: ear aspect University of Ghana http://ugspace.ug.edu.gh 68 4.3.4. Genetic analysis of the diallel crosses evaluated under drought and well- watered conditions The analysis of variance of the single crosses revealed significant mean squares for genotypes (G), environments (E) and genotypes by environments interactions (GEI) for all the traits under well-watered conditions and across contrasting environments (Tables 4.8 and 4.10) except ear height for GEI across environments. Significant mean squares were also found for G and E under drought conditions except ear aspect for environments (Table 4.9). However the GEI mean squares under drought were not significant for grain yield, number of days to silking, plant and ear heights, and plant aspect (Table 4.9). Partitioning of the entries into components revealed that GCA and SCA mean squares were significant for all measured traits under drought, well-watered conditions and across research environments except the SCA for EPP under drought (Tables 4.8, 4.9 and 4.10). Similarly GCA by environment interactions were significant for all the traits under drought, well-watered conditions and across research environments except for plant aspect under drought. SCA by environment interactions were not significant for all measured traits under drought stress (Table 4.9) and for DYA, EHT, PASP and husk cover under well-watered conditions (Table 4.8). The proportions of GCA effects were larger than those of SCA effects across test environments. University of Ghana http://ugspace.ug.edu.gh 69 Table 4.8 Combined analysis of variance for grain yield and other traits of the early maturing single crosses evaluated across well-watered environments in Niger from 2012 to 2014 Source of variation DF Yield DYA DYS ASI PLTH EHT PASP EASP EPP HUSK ENV 3 120650005** 31206** 29741** 77** 49811** 74953** 8.2** 20.4** 6.26** 9.2** Rep(ENV) 4 7391247** 25** 34** 3ns 1325** 674** 0.7ns 1.4** 0.07* 0.4* Entry 104 5555332** 28** 41** 5** 1571** 490** 1.7** 1.6** 0.07** 0.9** ENV*Entry 312 1296562** 4** 6** 3** 183** 95** 0.5** 0.3** 0.04** 0.2** GCA 14 16217668** 149** 241** 28** 9041** 2379** 6.1** 6.6** 5.25** 3.2** SCA 90 3896747** 9** 10** 2** 408** 196** 1.0** 0.9** 0.06** 0.5** ENV*GCA 42 3049903** 11** 18** 8** 273** 254** 1.1** 0.6** 0.04** 0.5** ENV*SCA 270 1023820* 3ns 4* 2* 169* 70ns 0.4ns 0.2* 0.02** 0.1ns Residual 416 804728 3 3 1.7 139 64 0.3 0.2 0.01 0.1 YIELD: Grain yield ASI: anthesis-silking interval PLTH: plant height PASP: plant aspect HUSK: husk cover DYA: days to anthesis DYS: days to silk EHT: ear height EASP: ear aspect EPP: ear per plant *Significant at the 0.05 probability level; **Significant at the 0.01 probability level; ns: not significant. University of Ghana http://ugspace.ug.edu.gh 70 Table 4.9 Combined analysis of variance for grain yield and other traits of the early maturing single crosses evaluated across stress environments in Niger from 2012 to 2014 Source of variation DF Yield DYA DYS ASI PLTH EHT PASP EASP EPP SGREEN2 ENV 2 33974256** 1732** 4017** 676** 4418** 376* 2.5* 0.7ns 4.8** 106** Rep(ENV) 3 6374226** 17* 154** 170** 1633** 591** 3.8** 6.8** 0.1ns 4** Entry 104 1388241** 46** 65** 22** 594** 352** 1.4** 0.9** 0.1** 3** ENV*Entry 208 620636ns 7** 14ns 11* 207ns 87ns 0.7ns 0.5* 0.1* 1* GCA 14 4439140** 246** 342** 79** 2192** 1262** 2.7** 2.5** 0.3** 11** SCA 90 913656** 15** 22** 13** 346** 210** 1.2** 0.6** 0.1ns 1* ENV*GCA 28 1456065** 13** 27** 23** 323* 160* 0.9ns 0.9** 0.1** 2** ENV*SCA 180 490680ns 6ns 12ns 9ns 189 ns 76ns 0.6ns 0.4ns 0.1ns 1ns Residual 312 531587 5 13 9 206 91 0.6 0.4 0.1 1 YIELD: Grain yield ASI: anthesis-silking interval PLTH: plant height PASP: plant aspect SGREEN2: stay-green characteristic 2 DYA: days to anthesis DYS: days to silk EHT: ear height EASP: ear aspect EPP: ear per plant *Significant at the 0.05 probability level **Significant at the 0.01 probability level ns: not significant University of Ghana http://ugspace.ug.edu.gh 71 Table 4.10 Combined analysis of variance for grain yield and other traits of the early maturing single crosses evaluated across the seven test environments in Niger from 2012 to 2014 Source of variation DF Yield DYA DYS ASI PLTH EHT PASP EASP EPP HUSK ENV 6 575606929** 24342** 33463** 1945** 79606** 37725** 22** 89.5** 6.3** 10.7** Rep(ENV) 7 6955381** 21** 85** 75** 1457** 639** 2** 3.7** 0.1* 0.6** Entry 104 5009929** 67** 93** 16** 1853** 773** 2** 1.9** 0.1** 1.3** ENV*Entry 624 1177434** 5* 9* 7** 213** 88ns 1** 0.4** 0.0** 0.2** GCA 14 13026012** 375** 549** 74** 9911** 3500** 6** 5.9** 0.23** 5.9** SCA 90 3762982** 20** 22** 7** 599** 349** 2** 1.2** 0.05** 0.6** ENV*GCA 84 3282106** 13** 23** 17** 465** 204** 1** 1.1** 0.08** 0.5** ENV*SCA 540 850040* 4* 8ns 6* 174ns 70ns 0.5ns 0.3ns 0.04** 0.2* Residual 728 687668 4 7 5 168 75 0.5 0.3 0.03 0.2 YIELD: Grain yield ASI: anthesis-silking interval PLTH: plant height PASP: plant aspect HUSK: husk cover DYA: days to anthesis DYS: days to silk EHT: ear height EASP: ear aspect EPP: ear per plant *Significant at the 0.05 probability level **Significant at the 0.01 probability level ns: not significant University of Ghana http://ugspace.ug.edu.gh 72 4.3.5. Combining ability effects of measured traits Significant GCA effects for grain yield and other traits were observed for the early maturing yellow grained inbreds under drought, well-watered conditions and across test environments (Table 4.11). Under drought stress, GCA for grain yield ranged from -320 kg ha-1 for TZEI 124 to 472 kg ha-1 for ENT 13. Of the 15 inbreds, only ENT13, TZEI 182, and TZEI 161 showed significant positive GCA effects for grain yield under drought conditions (Table 4.11). In contrast, TZEI 124, ENT15, ENT4, TZEI 14, and TZEI 16 showed significant negative GCA effects for grain yield under stress conditions. Under well-watered conditions, GCA of grain yield ranged from -830 kg ha-1 for TZEI 14 to 693 kg ha-1 for ENT13. The highest significant positive GCA effects were observed for ENT 13, TZEI 124, and TZEI 129 (Table 4.11) while TZEI 14, TZEI 182 and TZEI 167 displayed the lowest significant negative GCA in well-watered environments. Only ENT13 and TZEI 129 had significant and positive GCA effects for grain yield across test environments. In addition, across test environments, four inbreds (TZEI 160, TZEI 182, TZEI 161 and TZEI 23) displayed significant negative GCA for the number of days to silking, four inbreds (TZEI 129, TZEI 124, ENT 13 and TZEI 135) showed significant positive GCA for plant height, while three inbreds (ENT 13, TZEI 124 and TZEI 157) had significant negative GCA for plant aspect (Table 4.11). University of Ghana http://ugspace.ug.edu.gh 73 Table 4.11. GCA effects of grain yield and other agronomic traits of early maturing maize inbred lines from the evaluation of the single crosses under drought and well-watered conditions in Niger from 2012 to 2014 Inbred Lines . Grain yield kg ha -1 . . DYS . . PLTH . . PASP . . EPP . SGC2 WWC WSC Across WWC WSC Across WWC WSC Across WWC WSC Across WWC WSC Across WSC TZEI 160 -49ns 114 ns 21ns -1.9 ns -2.4** -2.1** -12** -4.4** -8.6** 0.2* -0.07 ns 0.05 ns 0.04 ns 0.04 ns 0.04* -0.05 ns TZEI 129 517** 30 ns 308* -0.4 ns -0.1 ns -0.3 ns 14** 7.5** 11** -0.1* 0.03 ns -0.06 ns -0.01 ns 0.08* 0.03 ns -0.14 ns ENT 15 -109ns -286** -185ns 1.6 ns 2.2** 1.9* 5* 0.1ns 2.6ns 0.1ns 0.24 ** 0.16** -0.03 ns -0.07* -0.05** 1.15** ENT 17 37ns -65ns -7ns 0.7 ns 1.3* 1 ns 4ns -2.5ns 0.9ns 0.1ns -0.02 ns 0.03 ns 0.06** -0.02 ns 0.02 ns 0.23 ns TZEI 182 -446** 333** -112ns -1.2 ns -2.6** -1.8* -8** 0.3ns -4.4** 0.2** -0.20 * 0.05 ns -0.01 ns 0.07* 0.03 ns -0.86** TZEI 167 -270* -122ns -206ns 1.4 ns 1.3* 1.3 ns -5** -4.4** -4.9** -0.1ns 0.06 ns 0.00 ns -0.05* 0.00 ns -0.03 ns 0.03 ns TZEI 14 -830** -210* -564** 2.6 * 3.0** 2.8** -9** -8.3** -8.6** 0.2** 0.16 ns 0.20** -0.02 ns -0.06* -0.04* -0.15 ns TZEI 124 612** -320** 213ns -0.3 ns 1.4* 0.5ns 11** 9.7** 11** -0.6** 0.04 ns -0.33** -0.02 ns -0.12** -0.07** 0.92** TZEI 16 -121 ns -202* -156ns 2.0 ns 2.0** 2* 3ns -0.7ns 1.1ns 0.1ns -0.03 ns 0.01 ns 0.08** 0.03 ns 0.06** -0.04 ns ENT 4 -96 ns -253** -163ns 1.0 ns 2.1** 1.5 ns 3ns -0.4ns 1.3ns 0.1* 0.31** 0.21** -0.01 ns -0.04 ns -0.02 ns 0.63** TZEI 161 59 ns 222* 129ns -1.8 ns -2.4** -2 * -7** -0.4ns -4* 0.1* -0.01 ns 0.08 ns 0.00 ns -0.02 ns -0.01 ns -0.38* TZEI 23 -168 ns 151ns -32ns -2.5 * -3.3** -2.9 ** -18** -8.4** -14** 0.2** -0.06 ns 0.07 ns 0.01 ns 0.04 ns 0.02 ns -0.26 ns ENT 13 693** 472** 598** 0.0 ns -0.8 ns -0.3 ns 13** 6.2** 10** -0.4** -0.47** -0.45** -0.01 ns 0.07* 0.02 ns -0.74** TZEI 135 -0.1 ns -20ns -9ns -0.5 ns 0.1 ns -0.2 ns 5** 2.2ns 4* 0.1ns 0.13 ns 0.10 ns -0.01 ns 0.00 ns -0.01 ns -0.19 ns TZEI 157 172 ns 157ns 165ns -0.8 ns -1.9** -1.3 ns 3ns 3.7* 3.3ns -0.2* -0.10 ns -0.13* -0.02 ns 0.03 ns 0.00 ns -0.14 ns SE ± 43 34 49 0.4 0.3 0.3 0.7 0.6 0.6 0.02 0.03 0.02 0.03 0.01 0.01 0.06 PLTH: plant height; DYS: days to silk; EPP: ear per plant ; PASP: plant aspect; SGC2: stay-green 2 WWC: well-watered conditions; WSC: water stress conditions; Across: across the seven contrasting environments *Significant at the 0.05 probability level; **Significant at the 0.01 probability level; ns: not significant University of Ghana http://ugspace.ug.edu.gh 74 4.3.6. Mode of gene action controlling measured traits The relative importance of GCA and SCA effects was assessed by computing the ratio of GCA effects to the total genetic effects (twice the GCA effects plus the SCA effects). The closer the ratio to unity, the greater the predictability based on GCA alone (Baker, 1978). GCA effects were greater than the SCA effects for grain yield and all the measured traits under drought stress (Fig. 4.3), well-watered conditions (Fig. 4.4) and across test environments (Fig. 4.5). In all cases, GCA effects accounted for more than 80% of the genetic variation in the traits compared to SCA effects that accounted for less than 20%. This is an indication that additive gene action is more important in controlling the measured traits than non-additive gene action. Figure 4.3. Proportion of additive (lower bar) and non-additive (upper bar) genetic variance for grain yield and other traits of 105 single cross hybrids derived from 15 × 15 diallel crosses among 15 early maturing maize inbred lines and evaluated at 3 water stress environments 70% 75% 80% 85% 90% 95% 100% Non Additive Additive University of Ghana http://ugspace.ug.edu.gh 75 70% 72% 74% 76% 78% 80% 82% 84% 86% 88% 90% 92% 94% 96% 98% 100% Yield DYA DYS ASI PLTH EHT PASP EASP EPP HUSK Additive Non Additive 70% 72% 74% 76% 78% 80% 82% 84% 86% 88% 90% 92% 94% 96% 98% 100% Yield DYA DYS ASI PLTH EHT PASP EASP EPP HUSK Additive Non Additive Figure 4.4 Proportion of additive (lower bar) and non=additive (upper bar) genetic variance for grain yield and other traits of 105 single cross hybrids derived from 15 × 15 diallel crosses among 15 early maturing yellow- grained maize inbred lines and evaluated at 4 well-watered environments Figure 4.5 Proportion of additive (lower bar) and non-additive (upper bar) genetic variance for grain yield and other traits of 105 single cross hybrids derived from 15 × 15 diallel crosses among 15 early maturing yellow- grained maize inbred lines across test environments University of Ghana http://ugspace.ug.edu.gh 76 4.3.7. Heterotic grouping of the early maturing inbreds using SCA_PY and HSGCA methods In this study GCA effects for grain yield were largely predominant (more than 86%) over SCA effects. Under drought stress and across environments none of the positive SCA effects displayed by the single crosses was significant (tables not shown). Therefore no appropriate heterotic grouping of the inbreds could be expected from this data. However, under well-watered environments, 8 crosses exhibited positive and significant SCA effects for grain yield (Table 4.12). Therefore the classification of the inbreds into heterotic groups using the SCA_PY and the HSGCA methods was based on SCA values from the evaluation of the single crosses under well- watered environments. ­ Heterotic grouping based on SCA_PY method The procedure described by Menkir et al. (2004) was adopted to classify inbreds into heterotic groups based on SCA effects under well-watered conditions (Table 4.12) and hybrid yield information under well-watered conditions (Appendix 4.1). According to this procedure, two testers e.g. tester 1 and tester 2 are considered. For an inbred to be classified into a heterotic group, it should have a positive SCA for grain yield with one tester, a negative SCA with the other tester and the testcross of the line with the opposite tester should have a yield equal or superior to the yield of tester1 x tester 2 hybrid. The ideal situation for an efficient heterotic grouping is when the SCA effects are significant. The yield of the best hybrid check under well- watered conditions TZEI 23 x TZEI 13 (5063 kg ha-1) was considered as reference yield for the heterotic grouping of inbreds. Only hybrids showing a grain yield equal or superior to that of the check were used for the classification. The highest and significant SCA effect for grain yield University of Ghana http://ugspace.ug.edu.gh 77 (1143 kg ha-1) was displayed by the single cross TZEI 160 x TZEI 157 (table 4.12). The two inbreds TZEI 160 and TZEI 157 were considered as testers to classify the other inbreds. Two inbreds ENT 13 and TZEI 161 showed positive SCA when crossed to TZEI 157 and negative SCA when crossed to TZEI 160. Their hybrids with TZEI 157 outyielded the best check by 8 % and 11% (Tables 4.12). ENT 13 and TZEI 161 are thus opposite to TZEI 157; they were therefore classified in the group with TZEI 160. Inbreds ENT 4, TZEI 124 and TZEI 129 showed positive SCA when crossed to TZEI 160 and negative SCA when crossed to TZEI 157. Their hybrids with TZEI 160 outyielded the best check by 0.1%, 8% and 11% (Tables 4.12). ENT 4, TZEI 124 and TZEI 129 are thus opposite to TZEI 160; they were therefore classified in the group with TZEI 157. The SCA_PY method could classify 7 out of the fifteen inbreds used in the diallel (Tables 4.13). University of Ghana http://ugspace.ug.edu.gh 78 Table 4.12 SCA effects for grain yield (Kg ha-1) of the 105 single-cross hybrids evaluated under WWC and standard heterosis for grain yield of testcrosses involving the inbreds and the two testers TZEI 160 and TZEI 157 over the best check under WWC (TZEI 23 x TZEI 13) Inbreds TZEI 160 TZEI 129 ENT 15 ENT 17 TZEI 182 TZEI 167 TZEI 14 TZEI 124 TZEI 16 ENT 4 TZEI 161 TZEI 23 ENT 13 TZEI 135 TZEI 157 TZEI 160 561ns 322 ns 175 ns -267 ns -451 ns 173 ns 350 ns 484 ns 637 ns -2048 ns -1757 ns -24 ns 829* 1143** TZEI 129 -208 ns 17 ns 274 ns 710 ns 685 ns -741 ns -140 ns 616 ns -91 ns -116 ns -525 ns -670 ns -369 ns ENT 15 -455 ns 257 ns 163 ns 389 ns 217 ns -176 ns -2344 ns 305 ns 623 ns -308 ns 316 ns 901* ENT 17 266 ns 803* 170 ns -361 ns -537 ns 70 ns 297 ns 472 ns -653 ns -271 ns 8 ns TZEI 182 324 ns 886* -140 ns 173 ns 293 ns -139 ns -107 ns 599 ns -521 ns -1899 ns TZEI 167 -1223** 236 ns -715 ns -209 ns -261 ns -254 ns 718 ns -85 ns 243 ns TZEI 14 261 ns -1624 ns -188 ns 486 ns 139 ns 424 ns 209 ns -786 ns TZEI 124 60 ns 9 ns -37 ns 791 ns 709 ns -549 ns -804* TZEI 16 72 ns 1015* 751 ns -45 ns 52 ns 502 ns ENT 4 754 ns 1038* -100 ns -373 ns -275 ns TZEI 161 -1119** -469 ns 352 ns 824* TZEI 23 -506 ns -14 ns 59 ns ENT 13 -163 ns 19 ns TZEI 135 561 ns TZEI 157 SCA with TZEI 160 561 322 175 -267 -451 173 350 484 637 -2048 -1757 -24 829 1143 SCA with TZEI 157 1143 -369 901 8 -1899 243 -786 -804 502 -275 824 59 19 561 SHGY * with TZEI 160 11% -7% -7% -25% -25% -24% 8% -4% 0.1% -50% -49% 3% 6% 13% SHGY with TZEI 157 13% -3% 9% -5% -53% -7% -38% -10% 1% -14% 11% -8% 8% 5% *SHGY: standard heterosis for grain yield of testcrosses involving the inbreds and the two testers TZEI 160 and TZEI 157 over the best check under Well-Watered Conditions TZEI 23 x TZEI 13 (5063 kg ha-1) University of Ghana http://ugspace.ug.edu.gh 79 Table 4.13 Classification of the 15 inbreds into heterotic groups based on SCA effects and standard heterosis for grain yield of testcrosses involving the inbreds and the two testers TZEI 160 and TZEI 157 over the best check under well-watered conditions (TZEI 23 x TZEI 13) Groupe A Group B Not classified TZEI 160 TZEI 157 ENT 15 TZEI 167 TZEI 161 TZEI 129 TZEI 135 TZEI 14 ENT 13 TZEI 124 TZEI 16 TZEI 23 ENT 4 TZEI 182 ENT 17 ­ Heterotic grouping based on HSGCA method For the classification based on the HSGCA method, each of the 15 inbreds used for the diallel crosses was considered as a tester. The HSGCA values were computed based on the SCA effects of crosses (Table 4.12) and the GCA effects of inbreds (Table 4.11). The computed HSGCA values (Table 4.14) were subjected to the three classification steps described by Fan et al. (2009): Step 1: Place all inbred lines with negative HSGCA effects into the same heterotic groups as their tester. Step 2: If an inbred line was assigned to more than one heterotic group in Step 1, keep the line in the heterotic group if its HSGCA had the smallest value (or largest negative value) and remove it from other heterotic groups. Step 3: If a line had a positive HSGCA effect with all the testers, do not assign that line to any heterotic group because the line might belong to a heterotic group different from the testers used. The results showed four heterotic groups (Table 4.15). The first group was composed of TZEI 160, TZEI 161 and TZEI 23. The second group consisted of TZEI 157, TZEI 182, TZEI 124, TZEI 129 and TZEI 135. Inbreds TZEI 14, TZEI 16, and TZEI 167constituted the third group and ENT 15 and ENT 4 composed the fourth. Inbreds ENT 17 and ENT 13 were not assigned to any of the four groups. University of Ghana http://ugspace.ug.edu.gh 80 Table 4.14 HSGCA+ values for the 15 inbreds used in the diallel study computed based on the GCA effects for grain yield of inbreds and SCA effects of crosses under well-watered environments. No Inbred HSGCA1 HSGCA2 HSGCA3 HSGCA4 HSGCA5 HSGCA6 HSGCA7 HSGCA8 HSGCA9 HSGCA10 HSGCA11 HSGCA12 HSGCA13 HSGCA14 HSGCA15 1 TZEI 160 -49 511 272 125 -316 -500 124 300 435 587 -2097 -1807 -73 779 966 2 TZEI 129 1078 517 309 534 791 1227 1202 -224 377 1133 426 401 -8 -153 148 3 ENT 15 213 -318 -109 -564 148 53 279 108 -285 -2453 195 514 -417 207 792 4 ENT 17 211 53 -418 37 303 840 207 -324 -501 107 334 508 -617 -234 45 5 TZEI 182 -713 -172 -189 -180 -446 -122 440 -586 -273 -153 -585 -553 153 -967 -2345 6 TZEI 167 -720 441 -107 533 54 -270 -1493 -34 -984 -479 -530 -524 448 -354 -27 7 TZEI 14 -656 -145 -441 -660 57 -2053 -830 -569 -2454 -1018 -343 -691 -406 -621 -1616 8 TZEI 124 962 -129 829 251 473 848 873 612 672 622 575 1404 1321 64 -192 9 TZEI 16 363 -262 -298 -659 51 -836 -1746 -62 -121 -49 1022 629 -167 -69 381 10 ENT 4 541 520 -2440 -25 197 -305 -284 -86 -23 -96 659 942 -196 -468 -370 11 TZEI 161 -1988 -31 364 357 -79 -201 546 22 1202 814 59 -1060 -409 412 883 12 TZEI 23 -1925 -284 455 303 -275 -422 -29 623 582 869 -1287 -168 -674 -182 -109 13 ENT 13 669 167 384 39 1292 1411 1117 1401 647 592 224 187 693 530 712 14 TZEI 135 829 -670 316 -271 -521 -85 209 -549 52 -373 352 -14 -163 0 561 15 TZEI 157 1187 -198 1073 180 -1727 414 -615 -632 674 -103 996 231 191 732 172 + HSGCA1 through HSGCA15 were computed considering inbred 1 (TZEI 160) through inbred15 (TZEI 157) as testers. University of Ghana http://ugspace.ug.edu.gh 81 Table 4.15 Classification of the 15 inbreds into heterotic groups based on HSGCA effects of grain yield under well-watered environments Group 1 Group 2 Group 3 Group 4 TZEI 160 TZEI 157 TZEI 14 ENT 15 TZEI 161 TZEI 182 TZEI 16 ENT 4 TZEI 23 TZEI 124 TZEI 167 TZEI 129 TZEI 135 Grouping efficiency Grouping efficiency was defined by Fan et al. (2009) as the percentage of superior high-yielding hybrids obtained across the total number of inter-heterotic group crosses. However, it is still highly possible to observe superior hybrids from intra-group crosses (Fan et al., 2009). To compare the grouping efficiency of the two methods, the procedure described by Fan et al. (2009) was applied to the results of the present study. Hybrids developed from inbreds that were effectively classified by each of the two methods were used for the comparison. The procedure consisted of dividing the total number of hybrids for each method into two major groups i.e. inter-group and intra-group crosses. These two groups were subsequently divided into high yielding hybrids (Yield group 1) with a mean grain yield superior to that of the best check i.e. 5063 kg.ha-1; intermediate hybrids (Yield group 2) with a mean grain yield between 5063 and 3700 kg.ha-1 and low yielding hybrids (Yield group 3) with a mean grain yield < 3700 kg.ha-1. The best classification method is the one whose heterotic groups allowed inter-heterotic group crosses to produce more superior hybrids than the intra-group crosses (Fan et al., 2009). University of Ghana http://ugspace.ug.edu.gh 82 Of the 21 high yielding hybrids identified by the HSGCA method, only one hybrid was among the intra-group crosses. Three out of the 13 high yielding hybrids identified by the SCA_PY method were among the intra-group crosses (Table 4.16). The grouping efficiency of the HSGCA and the SCA_PY methods were therefore 95% and 77% respectively. Table 4.16 Number of hybrids with mean grain yield greater than 5063 kg.ha -1 (yield group 1), between 5063 and 3700 kg.ha -1 (yield group 2), and inferior to 3700 kg.ha -1 (yield group 3) out of the total number of hybrids identified by each of the two classification methods Yield Group Cross type HSGCA SCA_PY 1 Inter-group 20 10 Intra-group 1 3 2 Inter-group 39 2 Intra-group 7 5 3 Inter-group 2 0 Intra-group 9 1 Total 78 21 Comparison of average grain yield heterosis exhibited by crosses between opposing heterotic groups for each of the two methods revealed that the HSGCA method tended to identify heterotic groups displaying higher heterosis for grain yield than the SCA_PY method (Tables 4.17 and 4.18). In contrast, the SCA_PY method identified inter-group crosses that showed higher average grain yield (5449 kg ha-1) compared to the best average grain yield of inter-group crosses (5055 kg ha-1) identified by the HSGCA method (Tables 4.17 and 4.18). Table 4.17 Average Grain Yield (GY) (Kg.ha -1 ) for intra-group and inter-group crosses and average GY heterosis (%) showed by crosses from opposing heterotic groups for the HSGCA method Heterotic groups Group1 Group2 Group3 Group4 Group1 2824 4956 (12%) 4357 (54%) 5028 (78%) Group2 4427 4556 (3%) 5055 (14%) Group3 2569 4069 (58%) Group4 2021 University of Ghana http://ugspace.ug.edu.gh 83 Table 4.18 Average Grain Yield (GY) (Kg.ha -1 ) for intra-group and inter-group crosses and average GY heterosis (%) showed by crosses from opposing heterotic groups for the the SCA_PY method Heterotic groups Group A Group B Group A 4192 5449 (11%) Group B 4912 The SCA effects of crosses under well-watered conditions (Table 4.12) were also subjected to the three classification steps described by Fan et al., (2009). The results revealed a classification identical to that of the HSGCA method with additional information on the two inbreds (ENT 13 and ENT 17) that the HSGCA method failed to classify (Table 4.19). The SCA method here revealed that ENT 13 and ENT 17 belong to another heterotic group. The HSGCA method therefore has no advantage over the SCA method since the same classification could be obtained by either method. Table 4.19 Heterotic grouping based on the application of the classification procedure described by Fan et al. (2009) on the SCA effects of crosses under well-watered conditions Group 1 Group 2 Group 3 Group 4 Group 5 TZEI 160 TZEI 157 TZEI 14 ENT 15 ENT 13 TZEI 161 TZEI 182 TZEI 16 ENT 4 ENT 17 TZEI 23 TZEI 124 TZEI 167 TZEI 129 TZEI 135 University of Ghana http://ugspace.ug.edu.gh 84 4.3.8. Stability analysis of GCA effects of early maturing inbreds across test environments The GGE biplot of the GCA values of the grain yield of the fifteen early maturing yellow maize inbreds revealed that PC1 explained 50.9 % of the total variance while PC2 explained 26.4 %, thus both PCs accounted for 77.3 % of the total variation for the GCA effects of the inbreds across test environments (Figure 4.6). In the average-tester-coordination view of the GGE biplot (Figure 4.6), the average environment (represented by the small circle on the single-arrowed red horizontal line) has the average coordinates of all test environments, while the single-arrowed red line is the Average Tester coordinate Abscissa (or ATA) which points to higher GCA effect of grain yield across environments. The longer the absolute length of the projection of an inbred on to this line in the direction of the arrow, the higher is the GCA of the inbred. Therefore, inbreds ENT13, TZEI 129, TZEI 124 and TZEI 157 had the highest GCA values across test environments. The absolute length of the projection of an inbred on to the Average Tester coordinate Ordinate (ATO) (double-arrowed blue horizontal line passing through the origin of the biplot on the y-axis) is a measure of its stability. The shorter the projection, the more stable is the GCA of the inbred. Thus, the inbreds ENT 13, TZEI 157, TZEI 160, TZEI 167 and TZEI 14 were the most stable across test environments (Figure 4.6). Based on their high and stable GCA effects for grain yield across environments, inbreds ENT 13 and TZEI 157 are likely to contribute favorable alleles for grain yield to their progenies. University of Ghana http://ugspace.ug.edu.gh 85 Figure 4.6. An entry/tester genotype main effect plus genotype by environment interaction biplot for GCA effects of grain yield of 15 early maturing yellow inbreds across test environments (E1through E7) University of Ghana http://ugspace.ug.edu.gh 86 The polygon view or the "which-won-where" view of a biplot provides an opportunity to visualize the interaction patterns between entries and environments. It allows easy identification of genotypes that were outstanding in each environment (Yan and Tinker, 2006). In the polygon view (Figure 4.7) the perpendicular lines dividing the biplot into sectors are called equality lines, they facilitate visual comparison among the inbreds (Yan and Tinker, 2006). The winning inbred for each sector in terms of GCA effects is the one located on the respective vertex. In the present study, the seven environments fell into two to three sectors (Figure 4.7). ENT13 was the best inbred in terms of GCA effects for grain yield in environments E1(MDS12WWC), E2 (KDS12WWC), E4 (FRS13WWC) and E5 (MDS12WSC) while TZEI 124 was the best in environment E3(MRS13WWC) (Figure 4.7). The environments E6 (KDS12WSC) and E7 (MDS13WSC) were located on the equality line 2, and thus, fell into two sectors. Therefore inbreds TZEI 182 and ENT 13 had similar GCA effects and were the best inbreds in these two environments. Inbreds TZEI 14, TZEI 167, TZEI 16 and ENT 4 were the poorest in terms of GCA effects for grain yield in environments E1, E2, E4, E5, E6 and E7 while inbreds TZEI 182, TZEI 23, ENT 17 and TZEI 135 were the worst inbreds in environment E3 (Figure 4.7). University of Ghana http://ugspace.ug.edu.gh 87 Figure 4.7. A ‘‘which won where” genotype plus genotype x environment interaction biplot of GCA effect for grain yield of the 15 early maturing maize inbreds across seven environments (E1through E7) E2 University of Ghana http://ugspace.ug.edu.gh 88 4.3.9. GGE biplot analysis of grain yield performance and stability of early maturing single cross hybrids The biplot analysis could not display all the 105 single crosses and five checks in a readable pattern. Therefore only 25 hybrids derived from a selection of two checks, the 18 top yielding hybrids and the five lowest yielding hybrids across environments were used for the GGE biplot analysis. The GGE biplot analysis revealed that PC1 and PC2 together accounted for 82.9% of the total variance for grain yield across test environments (Figure 4.8). PC1 explained 76.2 % of the total variance while PC2 explained 6.7 %. The GGE biplot was based on genotype-metric preserving (SVP=1) which is more appropriate for comparing genotypes. The red single-arrowed line is the average environment abscissa (AEA) that points to higher mean yield across environments. Therefore, the five top yielding hybrids across test environments were in descending order: TZEI 124 x ENT 13, TZEI 167 x ENT 13, TZEI 160 x TZEI 157, TZEI 124 x TZEI 23 and TZEI 161 x TZEI 157. The lowest yielding hybrids were ENT 15 x ENT 4, TZEI 14 x TZEI 16 and TZEI 167 x TZEI 14 (Figure 4.8). The double-arrowed blue line is the average environment ordinate pointing to greater variability or poorer stability in either direction. Thus, hybrids TZEI 161 x TZEI 157, TZEI 167 x ENT 13 and TZEI 124 x ENT 13 were the best hybrids in terms of stability and yield performance across test environments (Figure 4.8). TZEI 124 x ENT 13 was the highest yielding hybrid across test environments but it was less stable than TZEI 161 x TZEI 157 and TZEI 167 x ENT 13. The single crosses TZEI 16 x TZEI 161 and TZEI 124 x TZEI 23 were the most unstable genotypes of the selection (Figure 4.8). University of Ghana http://ugspace.ug.edu.gh 89 Figure 4.8. GGE biplot for grain yield performance and stability of 23 selected single cross hybrids and two checks across seven environments (E1 through E7) University of Ghana http://ugspace.ug.edu.gh 90 The 23 hybrids and two checks were also compared using the “which-won-where” function of the GGE biplot which is very useful for determining the winning hybrids in one or more environments (Figure 4.9). The perpendicular lines to the sides of the polygon are equality lines between adjacent genotypes in the polygon; they facilitate visual comparison among the genotypes. The winning genotype for each sector of the polygon defined by the equality lines is the one located on the respective vertex. The seven environments lay between two sectors. The single cross hybrid TZEI 124 x ENT 13 (entry 1) was the best hybrid in terms of grain yield at four environments (E1, E4, E5 and E6) while TZEI 16 x TZEI 161 (entry 7) was the best hybrid at the well-watered environment (E2). The two hybrids displayed similar performance in the two environments, E3 and E7 located on the equality line (2) between entry 1 and entry 7 (Figure 4.9). The early maturing single cross hybrid TZEI 124 x ENT 13 (Appendix 6.1) was therefore the best genotype in terms of grain yield performance under well-watered environments as well as across the seven test environments. University of Ghana http://ugspace.ug.edu.gh 91 Figure 4.9. A ‘‘which won where” GGE biplot of grain yield of 23 selected early maturing single cross hybrids and two checks evaluated across seven environments (E1through E7) University of Ghana http://ugspace.ug.edu.gh 92 4.4. Discussion One of the strategies proposed by CIMMYT to identify the time to impose water stress in maize drought trials is to conduct an experiment in which a particular maize genotype is sown at different dates at five days intervals. All the blocks are irrigated at the same time. The last irrigation is applied before flowering when one can predict that it will result in an appropriate drought stress intensity for the 2nd planting date (Bänziger et al., 2000).The intensity of the induced water stress in maize drought trial can be estimated by comparing the grain yield under stress with that under well-watered conditions. Bolaños and Edmeades (1996) proposed that average yield under drought and/or low-N should be 20 to 30% of the expected average yield in the same location under optimal management while Bänziger et al. (2000) proposed a much more severe stress that brings the grain yield from the stressed blocks to about 15 to 20% of well-watered yields. However, this study revealed that such a level of yield reduction should not be targeted under harsh climatic conditions. The study revealed that due to the low water holding capacity of the soils at Maradi and Konni coupled with the low relative humidity and frequent windy weather conditions, most of the plants in the stressed blocks would not survive if such a yield reduction was targeted. The study revealed therefore that a moderate yield reduction of about 30-40% of the well-watered yield was a more appropriate target for drought experiments in Niger. This could be achieved by stopping the irrigation 2 to 3 weeks before anthesis, and thereafter, applying one “rescue irrigation” 13 to 15 days after imposing the stress, then resuming the normal irrigation 10 to 12 days later. A similar experiment was conducted by Badu-Apraku et al., (2004) at Soumis and Ferkessedougou in Côte d’Ivoire where they imposed the water stress about 2 weeks before flowering until maturity. Their results indicated that the stress intensity was too severe and they suggested that for such trials, a “rescue irrigation” should be applied University of Ghana http://ugspace.ug.edu.gh 93 once every 10 days during the period of managed drought, or when plants started showing signs of temporary wilting early in the day. The significant differences observed among genotypes for grain yield and the associated traits under water stress and well-watered environments indicated that adequate genetic variation existed among the hybrids which could allow good progress from selection under contrasting environments. The significant environmental variation for grain yield and most traits under water stress and well-watered environments indicated that the environments were different which is desirable for assessing the performance and stability of the single cross hybrids in contrasting environments. However, this calls for extensive evaluation of the hybrids in more environments over years in order to draw solid inferences about hybrids’performance and stability. The significant genotype x environment interactions for grain yield and associated traits under water stress and well-watered environments implied the need for extensive testing of cultivars in multiple environments over years before recommendations could be made. Similar findings have been reported for early maturing tropical maize genotypes by several researchers (Pswarayi and Vivek, 2008; Badu-Apraku et al., 2010; Badu-Apraku et al., 2011c; Badu-Apraku et al., 2011d; Badu-Apraku et al., 2013b; Badu-Apraku et al., 2014). An important objective of the present study was to examine the combining ability for grain yield of the fifteen early maturing inbreds under both drought stress and optimal growing conditions. The results revealed significant GCA and SCA effects for grain yield, flowering dates, ASI, and other agronomic traits under drought and well-watered conditions except the SCA for number of ears per plant under drought, indicating the importance of both additive and non-additive gene action in controlling these traits. The study revealed larger GCA mean squares compared to SCA mean squares for all measured traits under drought (> 80%) and under well- watered conditions (>90%). University of Ghana http://ugspace.ug.edu.gh 94 The preponderance of GCA mean squares over SCA mean squares implied that additive gene action was more important than non-additive gene action for most traits and that GCA is the major component accounting for the differences among the single cross hybrids. This suggested a greater contribution of inbred parents with high GCA effects to the performance of the hybrids across environments (Baker, 1978). However the prediction of hybrid performance based on GCA effects of the parents alone might not be accurate since SCA effects are also significant. The results of this study corroborate the findings of Betràn et al. (2003) and Badu-Apraku et al. (2011b) who reported ratios of additive to non-additive variances for grain yield of 84% and 63% under drought and 60% and 80% under well-watered conditions, respectively. The results are also in agreement with the findings of other researchers (Melani and Carena, 2005; Derera et al., 2008; Pswarayi and Vivek, 2008; Makumbi et al., 2011; Badu-Apraku and Oyekunle, 2012). However, these results are in disagreement with those of Badu-Apraku et al. (2011c) who reported a minor importance of GCA effects for grain yield and overdominant SCA effects in a diallel study involving nine early maturing yellow inbreds. The difference in these results could be attributed to the differences in the germplasm used and probably the method used to assess the relative contribution of GCA effects versus SCA effects. In this study the relative importance of general and specific combining ability was assessed by the ratio given by Griffing (1956) while it was assessed based on the GGE Biplot analysis by Badu-Apraku et al. (2011c). The significant GCA × Environment interaction mean squares for all traits under drought and well-watered environments indicated that GCA effects associated with parents were not consistent over environments. This justified the use of the GGE biplot analysis on the GCA estimates of the inbreds to identify those with high and stable GCA effect across environments. The lack of significant SCA × Environment interaction mean squares for grain yield and almost University of Ghana http://ugspace.ug.edu.gh 95 all other traits indicated that the hybrids expressed the traits consistently in the test environments. Of the 15 inbreds, only ENT13, TZEI 182, and TZEI 161 showed significant and positive GCA effects for grain yield under drought conditions suggesting that these lines are likely to contribute favorable alleles in a recurrent selection program for drought tolerance improvement and such lines could be used as parents to form a synthetic population that could be improved for tolerance to drought. This also suggests that these inbreds are tolerant to drought as confirmed by the high grain yield per se of inbreds ENT 13 and TZEI 161 under drought. ENT 13 and TZEI 161 could therefore be considered as the best parents under drought conditions. In contrast, TZEI 124, ENT15, ENT4, TZEI 14, and TZEI 16 showed significant negative GCA for grain yield under stress conditions suggesting that hybrids derived from these inbreds would suffer severe yield reduction in presence of water deficit. Under well-watered conditions, the highest, positive and significant GCA effects for grain yield were displayed by inbreds ENT 13, TZEI 124 and TZEI 129 suggesting a greater contribution of these inbreds to the yield performance of their progeny under optimal growing conditions. The CIMMYT line ENT 13 had the highest and also highly significant and positive GCA effect for grain yield under drought, well-watered conditions and across test environments. It specifically combined well with TZEI 124, TZEI167 and TZEI 157. It was also identified as the best and most stable inbred in terms of GCA effects for grain yield by the biplot analysis. Furthermore this inbred produced an acceptable grain yield under drought (1,605 kg ha-1) and well-watered environments (2,191 kg ha-1) and hybrids developed from this line averaged a grain yield of 5,214 kg ha-1 under well-watered conditions and 2,110 kg ha-1 under drought conditions. The inbred line ENT 13 is therefore the most attractive general combiner. This inbred could be a desirable parent for the development of hybrids and synthetics as well as for population improvement, since it may contribute favorable alleles in the synthesis of new varieties (Legesse et al., 2009). University of Ghana http://ugspace.ug.edu.gh 96 Inbred ENT 13 was also ranked among inbreds with superior GCA effects for grain yield under low nitrogen and optimal growing environments by Ifie (2013). TZEI 124 and TZEI 129 are also good combiners under well-watered conditions based on their high positive GCA effects for grain and high grain yield performance of their hybrids (> 5,000 kg ha-1) in optimal growing environments. However, these two inbreds are very susceptible to drought as confirmed by their low stability across environments by the GGE biplot analysis. Across test environments inbreds TZEI 160, TZEI 182, TZEI 161 and TZEI 23 displayed significant negative GCA effects for the number of days to silking suggesting their use as parental lines in breeding for earliness. Another important objective of the present study was to identify superior hybrids under drought stress, well-watered conditions and across environments. Relying on grain yield alone for the selection of the best hybrids under drought is not an appropriate solution as its heritability usually decreases under stressed condition, while the heritability of some secondary traits remains high and their genetic correlation with grain yield increases sharply under stress (Bolaños and Edmeades, 1996; Bänziger et al., 2000). This study revealed high and significant correlations between grain yield under drought and secondary traits such as flowering dates (DYA and DYS), ASI, PASP and EASP PLTH and EHT, EPP and SG1 and SG2 indicating that some of these traits could be included in an index for reliable selection of the best hybrids under drought. Badu-Apraku et al. (2011a) identified PASP, EASP, ASI and EPP as the most reliable traits for selection for yield under drought in a study conducted in tropical early maturing maize cultivars. Similarly, numerous studies have reported secondary traits such as EPP, ASI and stay green characteristic to possess strong correlations with grain yield under drought conditions and have been used to select for higher levels of tolerance to drought and low-N in maize (Bolaños and Edmeades, 1996; Bänziger et al., 2000; Betràn et al., 2003b; Ribaut et al., 2009). This justified the use of a base index that integrates grain yield under drought stress, ASI, EPP, stay University of Ghana http://ugspace.ug.edu.gh 97 green characteristic, PASP and EASP in this study. A similar index has been extensively used by CIMMYT and IITA researchers (Badu-Apraku et al., 2011a; Badu-Apraku et al., 2011d). Drought tolerant genotypes are expected to suffer less yield reduction under drought stress. They are expected to display better values for secondary traits that are strongly correlated to high grain yield under drought i.e. reduced ASI, increased EPP, excellent PASP and EASP and delayed leaf senescence (Bolaños and Edmeades, 1996; Bänziger et al., 2000; Betràn et al., 2003b; Ribaut et al., 2009). In the present study, the single cross hybrids TZEI 182 x ENT 13, TZEI 160 x TZEI 157, ENT 13 x TZEI 157, TZEI 182 x TZEI 161 and TZEI 129 x ENT 17 were ranked as the top five yielding hybrids under drought stress. The performance of the top four hybrids was confirmed by the selection index. All the four hybrids were obtained from combinations involving two drought tolerant inbreds suggesting a dosage effect of drought tolerance genes in these hybrids. This is consistent with the findings of several authors who reported significant dosage effect under drought suggesting the need for including drought tolerant parents on both sides of a hybrid to achieve acceptable drought tolerance where the stress is severe (Bänziger et al., 2000; Betràn et al., 2003a; Ribaut et al., 2009; Badu-Apraku et al., 2011b). Under well-watered environments the best single cross hybrids identified were TZEI 124 x ENT 13; TZEI 124 x TZEI 23; TZEI 167 x ENT 13; TZEI 160 x TZEI 157 and TZEI 16 x TZEI 161. Only the single cross hybrid, TZEI 160 x TZEI 157 was identified among the five top yielding hybrids under drought as well as under well-watered conditions. Three single cross hybrids, TZEI 160 x TZEI 157; ENT 13 x TZEI 157 and TZEI 182 x ENT 13 out- yielded the checks and showed above-average grain yield > 5,400 kg ha-1 across well-watered environments and > 2,500 kg ha-1 under moderate drought stress suggesting that they should be extensively tested in on-farm trials under rainfed conditions in Niger. The two highest yielding hybrids under optimal growing conditions, TZEI124 x ENT13 and TZEI 124 x TZEI 23 were University of Ghana http://ugspace.ug.edu.gh 98 ranked at the 33rd and 41st positions under drought probably because they had in common the drought susceptible inbred TZEI 124 as parent. This is confirmed by the results of the GGE biplot analysis (restricted to only 25 hybrids) that showed that the hybrid TZEI124 x ENT13 has low stability despite its overall high performance across test environments. Therefore the hybrids TZEI124 x ENT13 and TZEI 124 x TZEI 23 should be recommended for maize production under irrigation or under rainfed conditions in areas where complementary irrigation can be provided in case of drought. The identification of heterotic groups and patterns is of fundamental importance in a maize hybrid breeding program (Melchinger and Gumber, 1998; Hede et al., 1999). In the present study the efficiency of a recently introduced method (HGSCA) for classifying inbreds into heterotic groups was compared to the classical SCA_PY method which combined line pedigree or hybrid yield information. The development of the HGSCA was justified by the fact that SCA effects are greatly influenced by the interaction between two inbred lines and by the interaction between hybrids and environments, thus, different studies might assign the same inbred line to different heterotic groups (Fan et al., 2009). The classification based on the two methods revealed different trends in assigning inbreds into heterotic groups. The SCA_PY is a highly selective method and, because of that, many inbreds are not classified. For an inbred to be classified by the SCA_PY method, it must display a negative SCA with one tester, a positive SCA with the other tester and the testcross of this line with the opposite tester should yield equal to or better than the yield of the best hybrid check. The HSGCA method tends to classify all the inbreds, however, when applied to diallel data where no tester is included; it tends to define a large number of heterotic groups which are of no practical use in the breeding program. University of Ghana http://ugspace.ug.edu.gh 99 The HSGCA method defined heterotic groups that show higher heterosis for grain yield compared to the SCA_PY method. In contrast, the SCA_PY method identified inter-group crosses that show the highest average grain yield (which is more desirable in breeding) compared to the HSGCA method. In the present study, the HSGCA method offered no advantage over the SCA method since identical or better inbred classification could be obtained when SCA effects are submitted to the procedure described by Fan et al., (2009) for grouping inbreds. These results are in desagreement with the findings of Fan et al. (2009), Badu-Apraku et al. (2013c) and Akinwale et al. (2014) who found the HSGCA method to be more efficient in classifying inbreds than the SCA method. Models of the HSGCA and SCA methods (Appendix 4.2) show that while the SCA method targets specific combinations where hybrids performed relatively better or worse than would be expected based on the GCA effects of both line and tester, the HSGCA method focuses on combinations where hybrids performed relatively higher or lower than would be expected based on GCA effects of the tester alone (see Appendix 4.2). The HSGCA method is likely to be less efficient than the SCA method when appropriately used. However more studies are needed to investigate this issue. The two heterotic groups defined by the SCA_PY method will be retained in this study since their inter-group crosses displayed the highest average grain yield. The first group is composed of inbreds TZEI 160, ENT 13 and TZEI 161 while the second group includes TZEI 157, TZEI 129 TZEI 124 and ENT 4. These inbreds could be recombined according to heterotic classification to develop new inbred lines and new versions of the high yielding hybrids ENT 13 x TZEI 124 and TZEI 160 x TZEI 157. University of Ghana http://ugspace.ug.edu.gh 100 4.5. Conclusions This study was set out to explore the genetic diversity among IITA and CIMMYT early maturing tropical yellow maize inbreds for breeding for drought tolerance. Drought remains a serious constraint that frequently challenges maize production particularly in sub Saharan African countries. Genotypes possessing drought tolerant genes are therefore needed to prevent drastic yield losses in case of drought occurrence during maize cropping season. IITA and CIMMYT maize breeding programs have developed outstanding early maturing maize inbreds that carry drought tolerant genes; however, their potential in hybrid combinations still needs to be explored in combining ability studies. Early maturing single crosses developed from a 15 x 15 diallel were evaluated in Niger under managed drought and well-watered conditions to examine the combining ability and the mode of gene action for drought tolerance among the 15 inbred lines, to classify inbreds into heterotic groups using two classification methods i.e. the SCA effects of grain yield and the HSGCA method, and to assess the yield performance and stability of the hybrids under contrasting environments. The study revealed that appropriate yield reduction due to water stress in drought experiments in Niger could be achieved if irrigation is stopped three to two weeks before anthesis, thereafter, an application of a “rescue irrigation” two weeks later and finally the resumption of the normal irrigation 10 to 12 days later. Significant GCA and SCA effects for grain yield and most measured traits were found across test environments with a predominance of additive gene effects over non- additive gene effects suggesting that inbred lines with high GCA effects for grain yield and other traits are likely to contribute favorable alleles for a better yield performance of their progeny across environments. Seven early maturing yellow maize inbreds were classified into two heterotic groups under well-watered environments based on the SCA_PY method which could identify more superior inter-group crosses than the HSGCA method. Under well-watered environments, the best single cross hybrids were TZEI 124 University of Ghana http://ugspace.ug.edu.gh 101 x ENT 13 and TZEI 124 x TZEI 23 while the single cross hybrids TZEI 182 x ENT 13 and TZEI 160 x TZEI 157 were the best under drought. The CIMMYT drought tolerant line ENT 13 was the most attractive general combiner across environments that could be a desirable parent for the development of synthetics and hybrids as well as for inclusion in breeding populations for grain yield and drought tolerance improvement. Inbreds TZEI 124, TZEI 129 and ENT 13 were the most promising parental inbreds for the development of hybrids and synthetics in well-watered environments while TZEI 182, TZEI 161 and ENT 13 were the best parents under drought conditions. These inbreds are likely to contribute favorable alleles for grain yield to their progenies. They could be introgressed into populations to improve grain yield and to increase genetic diversity. University of Ghana http://ugspace.ug.edu.gh 102 CHAPTER FIVE 5.0. Assessment of yield performance of modified single crosses of varying maturity under drought and well-watered conditions 5.1. Introduction Maize (Zea mays L.) is an important staple and cash crop which is increasingly consumed by Nigerien households. The production of this crop in the country remains low due to multiple factors including recurrent drought, low soil fertility, lack of availability of improved cultivars and poor crop management practices. Famers grow Open Pollinated Varieties (OPVs) whose seeds are obtained from the market or through saving seeds or farmer-to-farmer seed exchanges. The Rapid Rural Appraisal study conducted in maize production sites in Niger revealed that drought is a major constraint which dissuades farmers from producing maize under rainfed conditions and that the preferred maize cultivars in the country are the yellow-endosperm that combines extra-early (70-84 days) or early maturity (85-94 days) and high yield potential. Earliness is thus an important trait desired by farmers in Niger including those who produce maize under irrigation. However, this trait has been reported to be negatively correlated with high yield potential (Pswarayi and Vivek, 2008). It is therefore important to develop maize genotypes of varying maturity periods that also carry drought tolerance genes and allow farmers to test their yield performance. This will give them the opportunity to compare the advantages of using one type over the other (Tiwari et al., 2009). In Niger farmers’ accessibility to high quality seeds of improved varieties is currently being facilitated by emerging private seed companies which are involved in commercial production of seeds of these varieties. For maize production, a promising approach to increase farmers’ maize production and productivity and raise the seed companies’ income is to introduce high yielding University of Ghana http://ugspace.ug.edu.gh 103 maize hybrid varieties to replace the old low yielding OPVs currently grown by farmers. The best option to significantly increase farmers’ maize production is to adopt maize single cross hybrids. However one of the main problems of single cross hybrids is the inbreeding depression usually associated with the inbred lines used as parents (Castellanos et al., 2009). The problem is more crucial in the extra-early and early maturing maize for which appropriate testers are still lacking (Pswarayi and Vivek, 2008; Badu-Apraku and Oyekunle, 2012; Badu-Apraku et al., 2013a; Badu-Apraku et al., 2013c). Modified single-cross (MSC) hybrids represent an attractive alternative to single cross hybrids (Castellanos et al., 2009). MSCs are developed from three to four inbred lines. Instead of crossing two inbreds as in conventional single cross hybrids, either one or both parents of the MSCs is a related single cross i.e. a cross derived from two genetically related lines (Hallauer et al., 2010). The use of related single crosses (RSCs) as female parents in commercial hybrid production overcomes the problem of low amount of hybrid seeds often obtained when inbred parents are used. The present study is conducted to address the issues raised on: (i) the necessity to develop superior and drought tolerant maize genotypes of varying maturity periods and give farmers the opportunity to compare their merits and demerits, and (ii) the necessity to find an alternative to the low yield potential of inbreds used in commercial hybrid seed production. The objectives of this study were to (i) identify and select three RSCs with high yield potential across contrasting environments, (ii) develop MSCs from the three selected RSCs and inbreds of varying maturity cycles, (iii) assess the yield performance of MSC hybrids under drought and well-watered conditions and (iv) examine the relationship between grain yield and maturity. University of Ghana http://ugspace.ug.edu.gh 104 5.2. Materials and methods 5.2.1. Genetic materials Seven extra-early yellow maize inbred lines developed from two broad-based maize streak resistant populations were used in the present study (Table 5.1). The seven inbreds had been classified into three opposing heterotic groups by Badu-Apraku and Oyekunle (2012) in a previous study (Table 5.1). Inbreds from the same heterotic groups were used to develop five RSCs. Two single-cross hybrids and the local OPV P3Kollo were used as checks in the evaluation of the five RSCs (Table 5.2). Three out of the five RSCs were selected as female parents and crossed to 36 inbred lines of varying maturity (Table 5.3) to develop 103 MSCs. The maturity of the 36 inbreds was determined based on the results of an evaluation of the inbreds at INRAN Maradi during the rainy season of 2012. Table 5.1 Characteristics of extra-early maturing maize inbred lines used in the study Inbreds Pedigree Reaction to drought Classification TZEEI 95 TZEE-Y Pop Co S6 Inbred 47-3-4 DT Heterotic group A TZEEI 79 TZEE-Y Pop Co S6 Inbred 47-2-4A DT TZEEI 78 TZEE-Y Pop Co S6 Inbred 44 DS TZEEI 63 TZEE-Y SR BC1×9450 STR S6 Inb 7B DT Heterotic group B TZEEI 58 TZEE-Y SR BC1×9450 STR S6 Inb 1A DT TZEEI 82 TZEF-Y SR BC1 ×9450 STR S6 Inb 10B DT Heterotic group C TZEEI 76 TZEF-Y SR BC1 × 9450 STR S6 Inb 8B DT Table 5.2. the related single crosses developed from intra-group crosses, the two single cross checks and the local OPV check used in the study Related single-crosses Checks TZEEI 95 x TZEEI 78 TZEEI 79 x TZEEI 63 TZEEI 95 x TZEEI 79 TZEEI 79 x TZEEI 76 TZEEI 78 x TZEEI 79 P3 Kollo TZEEI 63 x TZEEI 58 TZEEI 76 x TZEEI 82 University of Ghana http://ugspace.ug.edu.gh 105 Table 5.3. Selected inbred lines of varying maturity periods used in the development of the modified single cross hybrids No Inbreds Source Pedigree Number of days to anthesis Maturity group 1 TZEEI 63 IITA TZEE-Y SR BC1×9450 STR S6 Inb 7B 41 Extra-early 2 TZEEI 64 IITA TZEE-Y SR BC1 x 9450 STR S6 Inb 8A 41 Extra-early 3 TZEEI 87 IITA TZEF-Y POP STR COS6 Inb 47-24B 43 Extra-early 4 TZEEI 89 IITA TZEF-YSR BC1 X 9450 STR S6inb 13A 43 Extra-early 5 TZEEI 67 IITA TZEE-Y SR BC1 x 9450 STR S6 Inb 10B 43 Extra-early 6 TZEEI 58 IITA TZEE-Y SR BC1×9450 STR S6 Inb 1A 44 Extra-early 7 TZEEI 81 IITA TZEF-Y SR BC1 x 9450 STR S6 Inb 9A 46 Extra-early 8 TZEEI 96 IITA TZEE-Y Pop Co S6 Inbred 78 47 Extra-early 9 TZEI 182 IITA TZE-Y Pop STR Co S6 Inbred 152-2-2 46 Extra-early 10 TZEI 157 IITA TZE-Y Pop STR Co S6 Inbred 102-1-2 46 Extra-early 11 TZEI 23 IITA TZE-Y Pop STR C0 S6 Inbred 62-2-3 45 Extra-early 12 TZEI 129 IITA TZE-Y Pop STR Co S6 Inbred 16-1-3 48 Early 13 TZEI 158 IITA TZE-Y Pop STR Co S6 Inbred 102-2-2 48 Early 14 TZEI 124 IITA TZE-Y Pop STR Co S6 Inbred 3-1-3 51 Early 15 TZEI 13 IITA TZE Comp5-Y C6 S6 Inbred 12 53 Early 16 ENT 13 CIMMYT [M37W/ZM607#bF37sr-2-3sr-6-2-X]-8-2-X-1-BB-B-4c1…. 51 Early 17 CMK ENT 20 CIMMYT Kenya [[Ent52:92SEW1-2/[DMRESR-W]EarlySel-#L-2-1- 51 Early 18 TZEI 16 IITA TZE Comp5-Y C6 S6 Inbred 31 53 Early 19 CMZ ENT 2 CIMMYT Zimbabwe DTPYC9-F46-1-2-1-2-B-B-B 53 Early 20 TZd 164 IITA TZE-W POP STR 104 S5 88/160-1/3 52 Early 21 TZd 181 IITA TZE-W POP STR 108 S5 164/198-1/1-1/5-1/2-5/5 53 Early 22 Cam Inbgp1 17 IRAD Cameroon TzmSR/BW-SR 54 Early 23 CMZ ENT 5 CIMMYT Zimbabwe DTPYC9-F46-3-4-1-1-B-B-B-B 54 Early 24 9848 IITA Hi29 x RppSR 55 Intermediate University of Ghana http://ugspace.ug.edu.gh 106 25 CMK ENT 22 CIMMYT Kenya ZEWBc2F2-79-1-BBB-B-B 55 Intermediate 26 TZd 176 IITA TZE-Y POP STR 106 S5 17/194-2/2-2/2-1/4-1/5 56 Intermediate 27 TZd 177 IITA TZE-Y POP STR 106 S5 17/194-2/2-2/2-1/4-3/5 56 Intermediate 28 88069 IRAD Cameroon Gusau BulSR/TzB 56 Intermediate 29 CMZ ENT 8 CIMMYT Zimbabwe DTPYC9-F143-1-6-1-B-B-B 56 Intermediate 30 CMZ ENT 12 CIMMYT Zimbabwe DTPYC9-F38-4-3-1-2-B-B-B-B 56 Intermediate 31 CMK ENT 8 CIMMYT Kenya MAS[MSR/312]-117-2-2-1-B*9-B 56 Intermediate 32 CMK ENT 23 CIMMYT Kenya [Ent320:92SEW2-77/[DMRESR-W]EarlySel-#I-2-4-386]-B… 56 Intermediate 33 9450 IITA B73 57 Intermediate 34 CMZ ENT 3 CIMMYT Zimbabwe DTPYC9-F143-5-4-1-2-B-B-B-B 57 Intermediate 35 CMZ ENT 6 CIMMYT Zimbabwe DTPYC9-F102-4-5-1-1-B-B-B-B 58 Intermediate 36 CMK ENT 24 CIMMYT Kenya ([ZEWAc1F2-164-3-2-B-1-BBB/[TZMI703/CML176]-B….. 59 Intermediate University of Ghana http://ugspace.ug.edu.gh 107 5.2.2. Experimental design and evaluation sites The five RSCs and three checks were evaluated under managed drought and well- watered conditions using RCBD with 2 replications at 3 sites: (i) INRAN research station at Maradi (N 13º26’ E 7º06’; 350 m altitude; 490 mm average annual rainfall), (ii) INRAN research station at Konni (N 13º47’ E 5º14’; 260 m altitude; 530 mm average annual rainfall), and (iii) FESA Seed Company (N 13º 39’ E 7º 04’; 382 m altitude; 500 mm average annual rainfall). The evaluation was conducted under managed drought and well-watered conditions during the dry season of 2012 at INRAN Maradi and INRAN Konni. It was conducted under well-watered conditions during the rainy season of 2013 at FESA Seed Co. and INRAN Maradi. The combination of location, cropping season, and water treatment was considered as an environment. Thus, the evaluation of the RSCs and checks was carried out under the following six environments: 1. MDS12WWC: Maradi (M), dry season of 2012 (DS12), under well-watered conditions 2. KDS12WWC: Konni (K), dry season of 2012 (DS12), under well-watered conditions 3. MRS13WWC: Maradi (M), rainy season of 2013 (RS13), under well-watered conditions 4. FRS13WWC: FESA (F), rainy season of 2013 (RS13), under well-watered conditions 5. MDS12WSC: Maradi (M), dry season of 2012 (DS12), under water stress conditions 6. KDS12WSC: Konni (K), dry season of 2012 (DS12), under water stress conditions. The seven extra-early maturing inbreds used in the development of the RSCs were also evaluated in separate blocks under managed drought and well -watered conditions using RCBD with two replications in the six environments to assess the high parent heterosis for grain yield. University of Ghana http://ugspace.ug.edu.gh 108 Based on the results of the evaluations of RSCs, three were selected and crossed to 36 inbreds of varying maturity to obtain 103 MSCs. The 103 MSCs plus the three extra-early RSC parents and two early maturing single-cross hybrids used as checks were evaluated during the dry season of 2013/2014 at INRAN station at Maradi and INRAN station at Konni. The 108 genotypes were divided into three maturity groups, each comprising 36 genotypes. The first group consisted of 33 extra-early MSCs plus the three RSC parents; the second group contained 34 early MSCs plus the two early maturing single-cross checks and the third group was made up of 36 intermediate maturing MSCs (appendix 5.1). At each evaluation site, the three maturity groups were evaluated separately under managed drought and well-watered conditions using a 9 x 4 alpha lattice design with 2 replications. The combination of location and water treatment was considered as an environment. Thus, the evaluation of the MSCs and checks was conducted at the following four contrasting environments: 1. MDWSC: INRAN station at Maradi (MD) under water stress conditions (WSC), 2. KNWSC: INRAN station at Konni (KN) under water stress conditions (WSC), 3. MDWWC: INRAN station at Maradi (MD) under well-watered conditions (WWC), 4. KNWWC: INRAN station at Konni (KN) under well-watered conditions (WWC). 5.2.3. Stress management The RSCs and the MSCs were all evaluated using the same stress management conditions. In the well-watered blocks, plants were irrigated twice a week until crop maturity. In the stressed blocks, plants were irrigated twice a week from planting to about 2 weeks before anthesis; thereafter, irrigation was stopped and the stress intensity was monitored. A rescue irrigation was applied in the stressed blocks (12 to 15 days after stress application) when 50% of the plants had started showing signs of severe water stress i.e. leaf rolling early in the morning, leaf senescence, University of Ghana http://ugspace.ug.edu.gh 109 tassel blasting and wilting. Between 10 to 12 days later, the normal irrigation was continued until crop maturity. 5.2.4. Field management practices Except for the water treatments, all field management practices were similar for both the stressed and the well-watered blocks. Each entry was planted in a single row of 5 m length with spacing of 40 cm within rows and 75 cm between rows. Three seeds were planted per hill at planting and the seedlings thinned to two plants per hill about two weeks after emergence to give a final population density of 66 000 plants ha–1. A compound fertilizer (NPK-15-15-15) was applied at the rate of 52 kg N ha–1, 52 kg P ha–1, and 52 kg K ha–1. Urea was top-dressed at the rate of 46 kg of N ha-1 20 days later. Weeds were controlled manually. 5.2.5. Data collection Data for grain yield and other agronomic traits were recorded as follow: DYA: (days to anthesis) number of days from planting to when 50% of the plants had shed pollen; DYS: (days to silking) number of days from planting to when 50% of the plants had emerged silks; ASI: (anthesis-silking interval) number of days between DYA and DYS; PLTH: (Plant height) measured as the distance from the base of the plant to the position of the first tassel branch. PLTH was computed as the average height of 10 plants randomly chosen in the row; EHT: (ear height) measured as the distance from the base of the plant to the node bearing the upper ear. This was computed as the average ear height of 10 plants randomly chosen in the row; PASP: (plant aspect) estimated based on the overall plant appeal on a scale of 1 to 5, where 1 = excellent plant type and 5 = poor plant type; University of Ghana http://ugspace.ug.edu.gh 110 HUSK: (husk cover) rated on a scale of 1 to 5, where 1 = husks tightly arranged and extended beyond the ear tip and 5 = open tip cover. NPHARV: (number of plants harvested) the total number of plants in the plot at harvest; EHARV: (number of ears harvested) the number of ears bearing at least one developed kernel in the plot; EPP: (number of ears per plant) estimated by dividing the total number of ears per plot by the number of plants harvested; EASP: (ear aspect) estimated based on absence of disease and insect damage, ear size, uniformity of ears, and grain filling. It is scored on a scale of 1 to 5, where 1 = clean, uniform, large, and well-filled ears and 5 = ears with undesirable features. YLD_HA : grain yield adjusted to 15% moisture content and converted to kg ha-1; it is computed from the shelled grain weight at harvest. SGREEN 1 and 2 (stay-green characteristic) scored for the water-stressed blocks at 90 and 100 days after planting on a scale of 1 to 9, where 1 = almost all leaves green and 9 = virtually all leaves dead. 5.2.6. Statistical analysis Analysis of variance (ANOVA) was performed on plot means for grain yield and other agronomic characters for each environment and across environments using PROC GLM procedure of SAS software, version 9.2 TS2M0 (SAS Institute, 2002). Entries (103 MSCs, 3 RSCs and 2 checks) were considered as fixed factors while environments, replications and blocks were considered as random factors. The statistical model used for the combined analysis is as follows: Ylki = μ + αl + bkl + vi + (αv)il + elki Yikl = observed trait value from each experimental unit (l = environment, k =replication, i = genotypes); µ = population mean; University of Ghana http://ugspace.ug.edu.gh 111 αl = environment effect; bkl = replication within environment effect; vi = genotype effect; (αv)il = interaction between genotypes and environments; elki = residual effect (adapted from Zhang and Kang (1997)). The high-parent (HPH) heterosis values exhibited by the RSCs were computed for grain yield as: HPH= x 100 where, RSC = Mean grain yield of the RSC, HP = Mean grain yield of the best inbred parent 5.3. Results 5.3.1. Performance of related single crosses and inbred parents under drought and well-watered conditions The analysis of variance for the RSCs evaluated under drought stress showed significant differences among genotypes (G) and environments (E) for grain yield and most measured traits except PASP, EASP and SGREEN for G and E; ASI and EHT for G. In contrast for genotype by environment interactions (GEI), no significant differences were observed in the measured traits except the grain yield, DYA and EHT (Table 5.4). The ANOVA across the well-watered environments also revealed significant differences among G and E for most measured traits except PASP for E; ASI, PASP and EASP for G. Similarly significant differences were detected for GEI for grain yield and all other traits except DYA, ASI and EPP (Table 5.4). University of Ghana http://ugspace.ug.edu.gh 112 Mean grain yield of the RSCs under drought was 1,522 kg ha-1 representing 38.68% of the grain yield under well-watered conditions (3,934 kg ha-1). Under drought stress, grain yield for the RSCs ranged from 1,111 kg ha-1 for TZEEI 95 x TZEEI 78 to 1,950 kg ha-1 for TZEEI 95 x TZEEI 79 (Table 5.4). The 3 top yielding RSCs under drought stress were TZEEI 95 x TZEEI 79 (1,950 kg ha-1); TZEEI 76 x TZEEI 82 (1,794 kg ha-1) and TZEEI 78 x TZEEI 79 (1,602 kg ha-1). The mean grain yield of these three RSCs under drought stress respectively was 72.4%, 66.6% and 59.5% of the grain yield of the best extra-early single cross check (TZEEI 79 x TZEEI 76) which was 2,692 kg ha-1 (Table 5.4). Under well-watered conditions, grain yield of the RSCs ranged from 3,078 kg ha-1 for TZEEI 63 x TZEEI 58 to 4,482 kg ha-1 for TZEEI 78 x TZEEI 79 (Table 5.4). The 3 top yielding RSCs were TZEEI 78 x TZEEI 79 (4,482 kg ha-1); TZEEI 95 x TZEEI 79 (4,235 kg ha- 1) and TZEEI 76 x TZEEI 82 (4,004 kg ha-1) representing 93.7 %, 88.5% and 83.7% of the grain yield of the best extra-early single cross check (TZEEI 79 x TZEEI 76) which was 4,782 kg ha-1 (Table 5.4). The RSCs exhibited high parent heterosis of 9.7% to 92% for grain yield under drought stress and of -5.7% to 64% under well-watered conditions (Table 5.4). The best three RSCs i.e. TZEEI 95 x TZEEI 79, TZEEI 78 x TZEEI 79 and TZEEI 76 x TZEEI 82 displayed high parent heterosis for grain yield of 64%, 64% and 35.3% respectively under well-watered conditions and 92.1 %, 37.3% and 57.5% , respectively under drought (see table 5.4). University of Ghana http://ugspace.ug.edu.gh 113 Table 5.4 Grain yield and other agronomic traits of the related single crosses and checks evaluated under drought stress and well-watered conditions in Niger in 2012 and 2013 Entry YIELD kg ha-1 DYA ASI PLHT EHT PASP EASP EPP SGR1 SGR2 HPH HPH WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WSC WSC WWC WSC TZEEI 95 x TZEEI 78 3872 1111 59 67 1.5 4.7 177 160 82 48 2.8 2.3 2.3 3.7 1.2 0.7 2.2 7.3 41.7 57.9 TZEEI 95 x TZEEI 79 4235 1950 56 64 1.8 4.2 175 151 81 43 2.7 2.7 2.7 3.5 1.0 0.8 2.7 7.8 64.0 92.1 TZEEI 78 x TZEEI 79 4482 1602 60 67 2.5 5.2 188 167 97 53 2.5 3.2 2.8 3.7 1.0 0.8 1.3 7.0 64.0 37.3 TZEEI 63 x TZEEI 58 3078 1152 58 68 1.4 3.8 180 149 88 46 3.1 3.8 2.6 4.3 1.1 0.6 1.5 6.8 -5.7 9.7 TZEEI 76 x TZEEI 82 4004 1794 57 66 1.5 3.8 168 154 83 52 4.0 3.7 3.2 3.7 1.3 0.7 1.5 7.2 35.3 57.5 TZEEI 79 x TZEEI 63 (Chk) 4663 2374 56 65 2.1 3.2 173 151 85 47 2.9 2.8 2.8 3.4 1.0 0.9 1.7 6.3 42.9 126.1 TZEEI 79 x TZEEI 76 (Chk) 4782 2692 56 63 1.3 2.8 181 164 91 53 2.5 2.2 2.3 3.2 1.2 0.8 1.3 6.0 66.1 136.4 P3 Kollo Chk 3863 1608 54 62 1.4 6.3 172 163 87 50 2.4 2.7 2.7 4.1 1.2 0.7 1.3 5.7 Overall mean 4122 1786 57 65 1.7 4.25 177 157 87 49 2.9 2.9 2.7 3.7 1.1 0.7 1.7 6.7 Lsd 741 699 1.3 1.7 1.5 2 8.5 9.5 4.8 9.3 2.1 2.3 0.4 0.6 0.1 0.2 0.8 1.2 Entry * * ** * ns ns * * ** ns ns ns ns ns ** ** ns ns Environment ** ** ** ** ** ** ** ** ** ** ns ns * ns ** ** ns ns Entry x Environment * * ns * ns ns * ns ** ** ** ns * ns ns ns ns ns WWC: well-watered conditions; WSC: water stress conditions; Chk: check YIELD: Grain yield; DYA: days to anthesis; ASI: anthesis silking interval; PLTH (cm): plant height; EHT (cm): ear height; PASP (1-5): plant aspect; EASP (1-5): ear aspect; EPP: number of ears per plant; SGR 1 and 2: stay green characteristic at 90 and 100 days after planting on a scale of 1 to 9 ; HPH (%): high parent heterosis for grain yield; *Significant at the 0.05 probability level; **Significant at the 0.01 probability level ; ns: not significant University of Ghana http://ugspace.ug.edu.gh 114 Across the six test environments, the combined analysis of variance revealed significant mean squares for G, E and GEI for grain yield and all the measured traits except EASP for G, PASP for E and ASI for G and GEI (Table 5.5). The 3 top yielding RSCs (TZEEI 78 x TZEEI 79; TZEEI 95 x TZEEI 79; and TZEEI 76 x TZEEI 82) identified under well-watered conditions were also the best RSCs across the six test environments (Figure 5.1). They were therefore selected to be used as female parents in the development of the MSC hybrids. University of Ghana http://ugspace.ug.edu.gh 115 Table 5.5 Mean squares from the combined analysis of variance for grain yield and other agronomic traits of the related single crosses and checks evaluated across six environments in Niger in 2012 and 2013 Source of variation DF Yield (kg ha-1) DYA DYS ASI PLTH (cm) EHT (cm) PASP (1-5) EASP (1-5) EPP Environments 5 65923555** 2495** 2654** 132** 4304** 26390** 3.8ns 7.83** 1.45** Rep (ENV) 12 1587928* 5.6* 12** 10** 360** 186** 2.3** 1.28** 0.03ns Genotypes 7 4662126** 65** 77** 4ns 578** 366** 4.4* 1.08ns 0.09* G x E interactions 35 1082532* 4.6* 9** 4ns 171* 89** 1.7** 0.54** 0.04** Residual 84 663742 2.5 4.4 3.3 94 34 0.6 0.3 0.02 *Significant at the 0.05 probability level **Significant at the 0.01 probability level ns: not significant University of Ghana http://ugspace.ug.edu.gh 116 0 500 1000 1500 2000 2500 3000 3500 4000 4500 1 2 3 4 5 6 7 8 3115 3473 3358 2436 3267 3900 4086 3111 Genotypes 1. TZEEI 95 x TZEEI 78 2. TZEEI 95 x TZEEI 79 * 3. TZEEI 78 x TZEEI 79* 4. TZEEI 63 x TZEEI 58 5. TZEEI 76 x TZEEI 82* 6. TZEEI 79 x TZEEI 63 Check 7. TZEEI 79 x TZEEI 76 Check 8. P3 Kolo Check * Best RSCs selected Grain Yield (kg ha -1) Related singles crosses Checks Figure 5.1 Grain yield (Kg ha-1) of the related single crosses and checks evaluated across six contrasting environments in Niger in 2012 and 2013 University of Ghana http://ugspace.ug.edu.gh 117 The analysis of variance for the extra-early inbred parents (used in the development of the five RSCs) evaluated under drought and well-watered conditions revealed no significant differences among the inbreds for most traits except grain yield, flowering dates, PLTH and EHT (Table 5.6). Mean grain yield of the inbreds ranged from 591 Kg ha-1 for TZEEI 78 to 1,139 Kg ha-1 for TZEEI 76 and from 2,192 Kg ha-1 for TZEEI 95 to 3,263 Kg ha-1 for TZEEI 63 under drought stress and well-watered conditions, respectively (Table 5.6). The average grain yield of the inbreds was 923 Kg ha-1 and 2,692 Kg ha-1 under drought stress and well-watered conditions, respectively, representing 60.6% and 68.4% of the average grain yield of their RSCs under similar conditions (1,522 Kg ha-1 and 3,934 Kg ha-1). University of Ghana http://ugspace.ug.edu.gh 118 Table 5.6 Grain yield and other agronomic traits of the extra-early maturing inbreds evaluated under drought stress and well-watered conditions in six contrasting environments in Niger, 2012 and 2013 Entry YIELD kg ha-1 DYA DYS ASI PLHT EHT SGREEN WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WSC TZEEI 78 2733 591 73 76 77 81 3.5 4.5 162 153 93 101 7.0 TZEEI 79 2582 809 68 71 71 77 3.0 6.0 142 135 75 89 7.0 TZEEI 95 2192 1015 66 67 67 74 0.5 7.0 108 118 40 49 6.0 TZEEI 76 2880 1139 65 64 66 70 1.0 5.5 145 149 78 83 6.0 TZEEI 82 2959 970 66 68 67 76 1.0 8.0 149 164 80 109 6.5 TZEEI 63 3263 1050 65 64 66 70 1.0 5.5 152 149 78 83 6.0 TZEEI 58 2239 886 66 67 67 74 0.5 7.0 108 118 40 49 6.0 Overall mean 2692 923 67 69 69 75 1.8 6.2 141 143 73 86 6.5 Lsd 343 529 2.6 6.3 3 11 3.6 7.7 17.2 44.0 15.5 26.9 1.7 Entry ** ** ** * ** ns ns ns ** ns ** * ns Environment ns ns ns ns ns ns ns ** ns ns ns ns ns Entry x Environment ns ns ns ns ns ns ns ** ns ns ns ns ns WWC: well-watered conditions; WSC: water stress conditions; YIELD: Grain yield; DYA: days to anthesis; DYS: days to silk; ASI: anthesis silking interval; PLTH (cm): plant height; EHT (cm): ear height; SGREEN: stay green characteristic at 90 days after planting on a scale of 1 to 9 ; *Significant at the 0.05 probability level; **Significant at the 0.01 probability level ; ns: not significant University of Ghana http://ugspace.ug.edu.gh 119 5.3.2. Performance of modified single crosses under drought and well-watered conditions Due to failure of the irrigation system at Konni, the trial involving the evaluation of the MSCs at INRAN station at Konni in 2013 was lost. Therefore, only the results of the trial from the Maradi Station are reported in this study. Significant differences due to genotypes were observed for grain yield and all other traits (Table 5.7 and Appendix 5.1) under both drought and well-watered conditions. Under drought stress the average grain yield of the MSCs was 1,429 Kg ha -1 representing 37.7% of the grain yield obtained under well-watered conditions (3,788 Kg ha- 1). The grain yield of the MSCs ranged from 377 Kg ha-1 to 3383 Kg ha-1 under drought and from 1765 Kg ha-1 to 5399 Kg ha-1 under well-watered conditions (Table 5.7). Under drought stress, 27 MSCs out-yielded the best single cross check ENT 13 x CamInbgp.1-17 (1770 Kg ha-1); 72 MSCs yielded higher than their respective RSC parents while 56 MSCs displayed high parent heterosis for grain yield superior to 50% (Appendix 5.1). The 5 top yielding MSCs under stress were (TZEEI 76 x TZEEI 82) x CMK ENT 23 (3,383 Kg ha-1); (TZEEI 76 x TZEEI 82) x CMZ ENT 8 (2,959 Kg ha-1); (TZEEI 95 x TZEEI 79) x CMK ENT 8 (2,937 Kg ha-1); (TZEEI 78 x TZEEI 79) x CMK ENT 23 (2,778 Kg ha-1) and (TZEEI 78 x TZEEI 79) x CMK ENT 8 (2,754 Kg ha-1) (Table 5.7). They out-yielded the best single cross check ENT 13 x CamInbgp.1-17 by 91%, 67%, 66%, 57% and 56%, respectively. Under well-watered conditions 18 MSCs out-yielded the best single cross check TZEI 124 x ENT 13 (4,524 Kg ha-1); 85 MSCs yielded higher than their respective RSC parents and 33 MSCs displayed high parent heterosis for grain yield superior to 50% (Appendix 5.1). The 5 top yielding MSCs under well-watered conditions were (TZEEI 76 x TZEEI 82) x TZEI 124 (5,399 Kg ha-1); (TZEEI 78 x TZEEI 79) x ENT 13 (5,328 Kg ha-1); (TZEEI 95 x TZEEI 79) x CMZ University of Ghana http://ugspace.ug.edu.gh 120 ENT 5 (5,293 Kg ha-1); (TZEEI 76 x TZEEI 82) x TZd 164 (5,272 Kg ha-1) and (TZEEI 76 x TZEEI 82) x TZEI 129 (5,202 Kg ha-1) (Table 5.7). They out-yielded the best single cross check TZEI 124 x ENT 13 by 19.3%, 17.8%, 17%, 16.5% and 15%, respectively. University of Ghana http://ugspace.ug.edu.gh 121 Table 5.7 Grain yield and other agronomic traits of the 5 highest yielding MSCs, the 5 lowest yielding MSCs and 5 checks evaluated under drought stress and well-watered conditions in Maradi in 2014. Entry Extra-early YIELD DYA ASI PLTH EHT PASP EPP SGREEN Early Intermediate WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC (TZEEI 76 x TZEEI 82) x TZEI 124 5399 1834 66 68 2.0 6.0 173 164 90 98 2.0 2.5 1.0 6.5 (TZEEI 78 x TZEEI 79) x ENT 13 5328 2376 70 71 0.5 3.0 201 182 130 123 1.5 3.0 1.0 6.0 (TZEEI 95 x TZEEI 79) x CMZ ENT 5 5293 1016 66 69 1.0 7.0 177 145 90 96 2.0 3.5 1.0 7.5 (TZEEI 76 x TZEEI 82) x TZd 164 5272 1440 68 72 -0.5 2.0 171 139 93 92 2.0 3.0 1.2 7.0 (TZEEI 76 x TZEEI 82) x TZEI 129 5202 2044 67 70 1.5 6.0 182 153 110 96 2.0 2.5 1.0 6.0 (TZEEI 76 x TZEEI 82) x CMK ENT 23 4092 3383 71 71 -0.5 2.0 163 169 83 109 3.0 2.5 1.1 6.5 (TZEEI 76 x TZEEI 82) x CMZ ENT 8 4350 2959 68 70 0.5 3.0 158 161 78 100 3.0 2.0 0.9 6.5 (TZEEI 95 x TZEEI 79) x CMK ENT 8 3108 2937 68 72 0.5 1.0 135 141 78 86 3.5 2.5 1.0 8.0 (TZEEI 78 x TZEEI 79) x CMK ENT 23 3109 2778 70 72 1.5 8.5 154 168 83 103 3.5 2.5 1.0 7.0 (TZEEI 78 x TZEEI 79) x CMK ENT 8 4662 2754 67 70 0.0 2.0 159 176 88 104 3.0 2.5 1.0 7.0 (TZEEI 76 x TZEEI 82) x TZEEI 63 1984 1002 66 69 1.0 5.5 139 136 80 88 4.5 3.5 1.5 5.5 (TZEEI 95 x TZEEI 79) x TZEEI 96 1835 744 68 72 2.0 2.5 144 126 80 87 4.5 3.5 0.7 7.0 (TZEEI 95 x TZEEI 79) x CMK ENT 20 4616 499 68 75 2.0 11.0 160 133 80 79 2.5 4.0 1.0 7.5 (TZEEI 76 x TZEEI 82) x TZEEI 89 3058 419 65 69 2.0 6.5 133 122 65 80 4.5 4.5 1.0 6.0 (TZEEI 78 x TZEEI 79) x TZEEI 87 2094 396 68 71 3.0 7.0 163 125 88 78 4.5 4.0 1.2 7.5 (TZEEI 78 x TZEEI 79) x TZEEI 96 1765 377 71 73 2.5 9.5 138 138 85 92 5.0 4.5 0.9 8.0 TZEI 124 x ENT 13 check 4524 1498 75 81 2.5 4.5 177 154 100 103 2.5 3.0 1.0 5.5 ENT 13 x CamInbgp.1-17 Check 4081 1770 70 70 1.5 11.0 178 153 108 92 3.0 2.5 1.0 6.0 TZEEI 78 x TZEEI 79 Check 3571 1053 66 69 1.5 2.5 167 159 85 95 3.0 2.0 1.3 7.0 TZEEI 76 x TZEEI 82 Check 3166 534 65 69 1.0 4.0 132 134 75 88 3.5 4.0 1.1 5.5 TZEEI 95 x TZEEI 79 Check 2353 821 69 71 4.5 6.5 153 136 80 91 3.5 4.0 0.8 8.0 Overall mean 3788 1429 67 71 1.9 5.7 158 142 83 90 2.9 3.2 1 6.8 Lsd 1538 918 2.2 4.2 2.3 5.5 20 23 20 17 1.3 1.2 0.2 1.7 Rep ns ns ns ** ** ns * ns ns ns ns ns * ** Block(Rep) ** ** * ** ns ns ** ** ** ** * ** ns ** Entry ** ** ** ** ** ** ** ** ** ** ** * * ** WWC: well-watered conditions; WSC: water stress conditions; YIELD (Kg ha -1 ): Grain yield; DYA: days to anthesis; ASI: anthesis silking interval; PLTH (cm): plant height; EHT (cm): ear height; PASP (1-5): plant aspect; EPP: number of ears per plant; SGREEN: stay green characteristics (1- 9) ; *Significant at the 0.05 probability level; **Significant at the 0.01 probability level ; ns: not significant University of Ghana http://ugspace.ug.edu.gh 122 5.3.3. Relationship between yield performance and maturity group of modified single crosses The evaluation of the MSCs was carried out during the cold and dry season in Maradi (November, 2013 to February, 2014). It is important to note that during that period, the low night and morning temperatures influenced the maturity cycles of plants by increasing the days to maturity by about 20 days compared to that observed during the main rainy season (June to September). The classification of the MSCs into different maturity groups i.e. extra-early, early and intermediate was based on the days to maturity of the inbreds used as male parents of the MSCs. From the evaluation of the MSCs under drought and well-watered conditions during the cold and dry season (2013/2014) in Maradi, the maturity groups of the MSCs as indicated by the number of days to anthesis for 50% of the plants in the plots was not consistent with the classification based on prediction from the maturity groups of the male inbred parents. Some MSCs classified in the intermediate group were among the 10 MSCs that were the first to flower while others from the extra-early group were among the 15 MSCs that flowered late. The number of days to anthesis varied from 67 to 85 under drought and from 62 to 75 under well-watered conditions. All the MSCs flowered earlier than the early maturing single cross check, TZEI 124 x ENT 13 (Appendix 5.1). Out of the top 30 MSCs that flowered early, 22 were from the extra-early group, 3 from the early group and 5 from the intermediate group. Of the 30 MSCs that flowered late, 3 were from the extra-early group, 11 from the early group and 16 from the intermediate group. University of Ghana http://ugspace.ug.edu.gh 123 The top five MSCs under well-watered conditions, (TZEEI 76 x TZEEI 82) x TZEI 124; (TZEEI 78 x TZEEI 79) x ENT 13; (TZEEI 95 x TZEEI 79)x CMZ ENT5; (TZEEI 76 x TZEEI 82) x TZd 164 and (TZEEI 76 x TZEEI 82)x TZEI 129 were all from the early group. Of the 20 top yielding MSCs under well-watered conditions none was from the extra-early group while 12 were from the early group and 8 from the intermediate group (Appendix 5.1). Under drought stress, the 5 top yielding MSCs (TZEEI 76 x TZEEI 82) x CMK ENT 23; (TZEEI 76 x TZEEI 82) x CMZ ENT8; (TZEEI 95 x TZEEI 79) x CMK ENT8; (TZEEI 78 x TZEEI 79) x CMK ENT 23 and (TZEEI 78 x TZEEI 79)x CMK ENT 8 were all from the intermediate maturing group. Of the 20 top yielding MSCs under drought stress five were from the extra-early group, three from the early group and 12 from the intermediate group (Appendix 5.1). Across water regimes, four out of the 30 top yielding MSCs were from the extra-early group, 11 from the early group and 15 from the intermediate group. In contrast, 16 out of the 30 lowest yielding MSCs were from the extra-early group, 9 from the early group and 5 from the intermediate group. The analysis of variance for grain yield and other traits revealed significant differences among the maturity groups for grain yield and most other traits under drought and well-watered conditions except for EHT and PASP under drought, as well as ASI and EPP under well-watered conditions (Table 5.8). Under drought and well-watered conditions the intermediate and early groups yielded significantly higher than the extra-early group while the extra-early group flowered earlier than the other groups under both water regimes. University of Ghana http://ugspace.ug.edu.gh 124 Table 5.8 Grain yield and other agronomic traits of the three maturity groups of the modified single crosses evaluated under drought stress and well- watered conditions in Maradi in 2014 Maturity YIELD DYA DYS ASI PLTH EHT PASP EPP group WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC Extra-early 3279 1285 66 69 68 74 2.0 4.9 152 138 80 89 3.4 3.3 1.0 0.7 Early 4141 1281 68 72 70 79 2.0 6.7 165 141 88 90 2.7 3.3 1.0 0.7 Intermediate 3944 1722 68 72 70 78 1.8 5.5 157 149 81 91 2.8 3.1 1.0 0.7 Overall mean 3788 1429 67 71 69 77 1.9 5.6 158 142 83 90 3 3.2 1 0.7 Lsd 305 244 0.8 1.16 0.81 1.5 0.47 1.12 4.8 5.9 5 4.2 0.26 1.9 0.04 0.06 Rep ns ns ns * ns ns ** ns ns ns ns ns ns ns ns ns Maturity group ** ** ** ** ** ** ns ** ** ** ** ns ** ns ns ** WWC: well-watered conditions; WSC: water stress conditions; YIELD (Kg ha -1 ): Grain yield; DYA: days to anthesis; DYS: days to silking; ASI: anthesis silking interval; PLTH (cm): plant height; EHT (cm): ear height; PASP (1-5): plant aspect; EPP: number of ears per plant; *Significant at the 0.05 probability level; **Significant at the 0.01 probability level ; ns: not significant University of Ghana http://ugspace.ug.edu.gh 125 5.4. Discussion In hybrid maize seed production, yield of the female parent is a major factor determining the seed production cost. One strategy to reduce the cost of hybrid seed production is to use RSCs which are the F1 between two highly related inbred lines (Lee et al., 2006). An important objective of this study was to identify RSCs with acceptable grain yield that could be used as female parents in hybrid seed production. The study revealed a significant increase in grain yield of the RSCs (grain yield ranging from 3078 kg ha-1 to 4482 kg ha-1) compared to their parental inbred lines (grain yield ranged from 2239 kg ha-1 to 2733 kg ha-1). These results suggest that the RSCs would be better female parents in hybrid seed production because of their higher grain yield potential compared to inbred lines. The higher performance of the RSCs is also confirmed by the high parent heterosis for grain yield (35% to 64%) exhibited over their inbred parents. Similar results were reported by Castellanos et al. (2009) who found an average grain yield of 5000 kg ha-1 to 8700 kg ha-1 for the RSCs compared to an average grain yield of 2700 kg ha-1 to 5700 kg ha-1 for the sister lines and an average heterosis for grain yield of 37% displayed by the RSCs over their inbred parents. The performance of the RSCs is in general more stable or predictable than that of the unrelated parental inbred lines (Lee et al., 2006). Average grain yields of 3800 kg ha-1 to 5400 kg ha-1 have been reported for intermediate and late maturing inbred lines (Duvick, 1999; Lee et al., 2006; Castellanos et al., 2009). In the present study, the relatively low grain yield of the extra- early inbreds (2239 to 2733 kg ha-1) and their corresponding RSCs (3078 to 4482 kg ha-1) is explained by the fact that extra-early maize genotypes are inherently lower yielding than early, intermediate and late maturing genotypes. For instance, Badu-Apraku (unpublished) reported an average grain yield of 2955 kg ha -1 for extra-early varieties across seven locations compared to an average grain yield of 4463 kg ha-1 for early varieties across six locations in a regional trial in University of Ghana http://ugspace.ug.edu.gh 126 2009 (Badu-Apraku and Oyekunle, 2012). Similarly a study on the productivity and optimum planting periods for five varieties representing five maturity groups in the Guinea savanna of Ghana conducted by Sallah et al. (1995) revealed an average grain yield of 3890, 5250, 5800, 5830 and 5880 kg ha-1 for the extra-early, early, normal endosperm intermediate, Quality Protein intermediate and the late maturing varieties, respectively. Another important objective of the present study was to determine the yield performance of the MSCs under drought and well water conditions. The significant differences observed among the MSCs for grain yield and most measured traits under water stress and well-watered conditions indicated that adequate genetic variation existed among the genotypes which could allow good progress from selection under the contrasting environments. The yields of MSCs under drought (37.7% of the well-watered yield) were slightly above the limits recommended by Bolaños and Edmeades (1996) who reported that the average yields under drought and low soil nitrogen conditions have to be between 20% and 30% of what the average yield would have been in the same location under optimum management. Modified single crosses have been reported to perform as well as their corresponding single cross hybrids (Fuhe et al., 1995; Lee et al., 2006; Castellanos et al., 2009). This is because a similar set of genes are associated with yield in both types of hybrids (Castellanos et al., 2009). In most cases MSCs display no ‘‘yield penalty’’ compared to their corresponding single crosses particularly when attention is paid to the degree of relatedness of the sister lines used (Fuhe et al., 1995). In the present study 27 and 18 MSCs out-yielded the best single cross checks of this study under well-watered and drought stress. Furthermore 56 MSCs under drought and 33 MSCs under well-watered conditions out-yielded their respective RSC parents by at least 50% confirming the hypothesis that MSCs may represent an attractive alternative to single cross hybrids in a situation University of Ghana http://ugspace.ug.edu.gh 127 where low grain yield of inbreds constitutes a major constraint to commercial maize single cross hybrid production. These results are in agreement with the findings of Lee et al., (2006) and Castellanos et al., (2009). The highest yielding MSCs hybrids under well-watered conditions were obtained from crosses with the early maturing inbreds TZEI 124 and ENT 13 confirming the results of the previous diallel study conducted in the same environment which revealed the outstanding combining ability of these two inbred lines under optimal growing conditions. The five highest yielding MSC hybrids under water stress conditions were obtained from crosses with the intermediate maturing CIMMYT lines CMK ENT 23 and CMZ ENT 8 suggesting that these lines carry genes conferring tolerance to drought. The present study also sought to assess the relationships between the yield performance of the MSCs and their maturity. The higher and significant grain yield displayed by MSCs from the intermediate and early maturity groups over MSCs from the extra-early group under both water regimes suggests that hybrids from parents of varying maturity groups may show high level of grain yield heterosis compared to within-maturity group crosses. These results are in disagreement with the findings of IITA and CIMMYT maize breeders who reported a failure of their outstanding intermediate to late maturing testers to discriminate effectively between early and extra-early maturing inbreds (Pswarayi and Vivek, 2008; Badu-Apraku et al., 2013a). The differences in the findings of this study and those of CIMMYT and IITA workers may be attributed to the differences in the germplasm used. University of Ghana http://ugspace.ug.edu.gh 128 5.5. Conclusions The present study was conducted to develop and select related single crosses with higher grain yield potential; to develop modified single cross hybrids from crosses between the selected RSCs and inbred lines from varying maturity groups, and to assess the yield performance of the MSCs under drought and well-water conditions as well as the relationships between the yield performance of the MSCs and their maturity. Results of the study revealed an average grain yield of 3,934 kg ha-1for the RSCs versus 2,692 kg ha-1 for their parental inbred lines. Furthermore the selected RSCs displayed high parent heterosis for grain yield of 57% to 92% and 35% to 64% respectively under drought and well-watered conditions suggesting that they would be good female parents for hybrid seed production. The study revealed that a moderate level of moisture stress (37.7% yield reduction) was achieved in the evaluation of the MSCs. Under drought stress, the 3 top yielding MSCs i.e. (TZEEI 76 x TZEEI 82) x CMK ENT 23; (TZEEI 76 x TZEEI 82) x CMZ ENT 8; and (TZEEI 95 x TZEEI 79) x CMK ENT 8 out-yielded the best single cross check (ENT 13 x CamInbgp.1-17) by 91%, 67%, and 66% respectively. Under well-watered conditions, the 3 highest yielding MSCs i.e. (TZEEI 76 x TZEEI 82) x TZEI 124; (TZEEI 78 x TZEEI 79) x ENT 13 and (TZEEI 95 x TZEEI 79) x CMZ ENT 5 out-yielded the best single cross check TZEI 124 x ENT13 by 19.3%, 17.8% and 17% respectively. Overall, 27 and 18 MSCs out-yielded the best single cross checks and displayed high parent heterosis for grain yield of 72% and 32% under drought and well-watered conditions, respectively, indicating that MSCs could be a good alternative to single cross hybrids where farmers could not afford maize hybrid seeds because of the high prices resulting from the lower grain yield of the inbred parents. MSCs developed from crosses with early and intermediate maturing inbreds yielded significantly higher than those developed from extra-early inbreds University of Ghana http://ugspace.ug.edu.gh 129 suggesting that MSC hybrids from the early and intermediate inbreds may be recommended for maize production in environments where complementary irrigation can be provided in case of drought. It is important to note, however, that in this study, results from the evaluation of the MSCs were from only one site in 2014. Therefore, no strong conclusions could be made about the performance of the MSCs under drought and well-water environments. It would be necessary to evaluate the MSCs at more sites to validate the results of the present study. University of Ghana http://ugspace.ug.edu.gh 130 CHAPTER SIX 6.0. Research Overview 6.1. Introduction Maize (Zea mays L.) is an important staple food crop in Niger. However, the current level of the production of maize grain is far below the national demand. Common crop production constraints i.e. drought, low soil fertility, poor crop management practices and lack of availability of improved varieties limit maize production in the country. Limited studies have been conducted on the production of this crop in Niger and very few cultivars (mostly open pollinated varieties) are available to farmers. Recently, maize has been included among the strategic crops whose production in multiple cropping systems is being promoted by Nigerien authorities through a national policy towards achieving food security in the country. More money is now being invested in purchasing maize seeds and developing small scale irrigation facilities and large public irrigation schemes to promote maize production in the country. In order to boost the national production from its current level, greater understanding of maize growing environments is needed and new maize improved varieties should be developed and/or introduced and tested in the different maize production areas of the country. Three major activities were conducted. The first activity was a Rapid Rural Appraisal to assess the performance of maize crop management practices used by farmers in Niger, identify maize cultivars’ traits preferred by farmers, and identify major factors that limit maize production in Niger. The second activity included: (i) a drought experiment to test the efficiency of the CIMMYT/IITA screening methodology for maize drought trials in Niger, and (ii) a diallel study involving fifteen early maturing yellow maize inbreds. The objectives of the diallel study was to estimate the combining ability and the mode of gene action for drought tolerance among the fifteen University of Ghana http://ugspace.ug.edu.gh 131 early maturing inbred lines, classify the inbreds into contrasting heterotic groups using two different classification methods, and assess the yield performance and stability of the diallel single cross hybrids under drought and well-watered conditions. The third activity was a study of the agronomic performance of modified single cross hybrids developed from crosses between three extra-early maturing related single crosses (used as female parents) and thirty-six inbreds of varying maturity periods. The study sought to assess the yield performance of the modified single crosses under drought and well-watered conditions and examine the relationship between the grain yield and the maturity of the modified single crosses. 6.2. Major findings and their implications The main findings of the thesis are activity specific and were discussed within the corresponding chapters. The following paragraphs aim at synthesizing these findings according to the different objectives targeted under each of the three activities conducted as well as providing their implications in the field of plant breeding and agriculture.  Maize production in Niger is performed under poor crop management practices. The study revealed common practices among farmers consisting of land preparation with animal drawn ploughs before sowing but rare application of manure, sowing at low planting densities, thinning to three or four plants per hill, low application of mineral fertilizers and practice of fertilizer application on the ground near the stalks. The poor crop management practices added to the low fertility of soils in Niger and the lack of availability of maize improved varieties worsen the conditions of maize production in the country. This could explain the low maize grain yields recorded in the country. The implications of these findings are that the current national policy for boosting crop production in the country University of Ghana http://ugspace.ug.edu.gh 132 which consists of facilitating farmers’ access to fertilizers, improved varieties and other agricultural inputs is likely not to be sufficient to significantly increase maize production in the country. Therefore, in addition to this strategy, training and increasing farmers’ awareness about appropriate crop management practices will undoubtedly contribute to boost the national maize production from its current level.  Yellow maize varieties combining earliness and high yield potential are predominantly preferred by maize growers in Niger. Farmers’ preference for extra-early and early maturing maize germplasm is justified because such germplasm could complete their growth cycle in environments where the beginning and the end of the rainy season are frequently challenged by drought. Several authors reported earliness and high yield potential to be of major importance to farmers (Defoer et al., 1997; Abebe et al., 2005; Obaa et al., 2005; Ouma et al., 2010; Sibiya et al., 2013). Farmers believe that yellow endosperm provides more flour and higher amounts of a locally consumed meal "Tuwo massara" than the white type. This should be investigated by researchers from food technology and consumer sciences. Earliness, yield and endosperm color are therefore traits to be taken into consideration by the maize breeding program in Niger to increase the chance of adoption of new maize cultivars by farmers. However extra-early and early maturing germplasm has been reported to yield lower than intermediate and late maturing germplasm (Sallah et al., 1995, Pswarayi and Vivek, 2008, Badu-Apraku and Oyekunle, 2012). Therefore intermediate germplasm should also be introduced and tested in Niger particularly in areas where irrigation facilities are available. A study from Tiwari et al, (2009) revealed that farmers could be convinced by the yield advantage of late germplasm and could therefore trade off earliness for yield under certain circumstances. University of Ghana http://ugspace.ug.edu.gh 133  Low soil fertility, recurrent drought and the lack of availability of improved varieties are major factors limiting maize production in Niger. Drought and low soil fertility are the most important threat to sustainable crop production and thus the most important obstacles to achieving food security in African countries. Drought challenges maize grain production in both temperate and tropical environments (Edmeades, 2008; Ribaut et al., 2009) and its effects are expected to be exacerbated due to global climate change (Witcombe et al., 2008; IAEA, 2013). Low soil fertility (Low phosphorus and nitrogen availabilities) is believed to become in the long run a more serious concern than the lack of moisture in West Africa (Bationo and Mokwunye, 1991; Gruhn et al., 2000) because of soil deficiency coupled with continuous cropping of the land over decades with no restitution. In Niger the majority of farmers are aware of the problem but seem helpless to change the situation. Therefore drought and low soil fertility have to be important priorities in the maize breeding program in Niger. The new cultivars should display improved productivity in low input systems and decreased input requirements in high input systems (IAEA, 2013). It is also important to develop and promote integrated crop, soil and nutrient management practices that could help farmers achieve sustainable crop production (IAEA, 2013).  An appropriate stress management procedure for maize drought trials in Niger consists of (i) stopping the irrigation 3 to 2 weeks before anthesis, (ii) applying a “rescue irrigation” 13 to 15 days after the stress imposition, and (iii) resuming the normal irrigation 10 to 12 days later until crop maturity. An important challenge to breeding maize for drought tolerance is the development of an efficient field screening methodology that will correctly discriminate and elicit true differences among genotypes (Badu-Apraku et al., 2004). A good drought screening method should bring the grain yield University of Ghana http://ugspace.ug.edu.gh 134 from the stressed blocks to about 15-20% or 20-30% of well-watered yields (Bolaños and Edmeades, 1996; Bänziger et al., 2000). In drought experiments conducted in Sahelian countries like Niger, frequent windy weather conditions and low relative humidity coupled with low water holding capacity of the soils lead to a rapid development of the stress and a rapid collapse of plants. A moderate stress that brings the grain yield from the stressed blocks to about 30-40% of well-watered yields should therefore be targeted in such environments to prevent the loss of the trial. This includes application of rescue irrigations when necessary as suggested by Badu-Apraku et al. (2004). Careful monitoring of the stress intensity is therefore necessary and the rescue irrigations should be applied when about 50% of plants in the stressed blocks present signs of severe stress i.e. leaf rolling, leaf senescence, tassel blast and stem wilting early in the morning.  An important requirement for a commercially successful maize hybrid program is the availability of information on the combining ability and heterotic patterns of the inbreds in the program (Badu-Apraku et al., 2013a). In the present study, significant GCA and SCA effects for grain yield and most measured traits were found under drought, well- watered conditions and across test environments with a large predominance (> 80%) of GCA effects over SCA effects. This suggests that grain yield and the other traits are predominantly controlled more by additive gene action than non additive gene action. The breeding implication of the significant GCA and SCA effects for grain yield and other traits is that appreciable breeding progress could be made using hybridization, backcrossing, and recurrent selection for the development of hybrids and synthetic varieties as well as in population improvement (Ifie, 2013). Inbreds with favorable GCA effects for grain yield and other traits are likely to transmit their characteristics to the progeny and could be useful University of Ghana http://ugspace.ug.edu.gh 135 in a breeding program. Such inbreds could be used as parents to form a synthetic population that could be improved for stress environments (Makumbi et al., 2011; Badu- Apraku et al., 2012). Significant GCA by environments effects were also found in the study suggesting that the combining abilities of the inbreds were not consistent across environments. A stability analysis of the GCA effects of the inbreds was therefore carried out to identify those with high and stable GCA effect across environments. The CIMMYT line ENT 13 had the highest and most stable GCA effect under drought and well-watered conditions. It was the most attractive general combiner across environments, which is likely to contribute desirable alleles to progeny. Based on the display of high GCA effects, inbreds TZEI 124, TZEI 129 and ENT 13 were the most promising parents under well- watered environments while TZEI 182, TZEI 161 and ENT 13 were the best parents under drought conditions. These inbreds could be used to form synthetics or varietal hybrids or crossed accordingly to heterotic classification to develop source populations from which superior inbred lines could be extracted. Furthermore, favorable alleles from inbreds ENT 13, TZEI 182 and TZEI 161 could be introgressed into breeding populations of national maize programs for drought tolerance improvement (Ifie, 2013).  The heterotic grouping of inbreds based on the SCA_PY method was more useful than based on the HSGCA method in the sense that the groups defined by the SCA_PY method displayed the highest average grain yield for inter-groups crosses compared to those defined by the HSGCA method. Sprague cited by Reif et al., (2005) stated: "The single most important element of a breeding program is the recognition and utilization of heterotic pattern. This recognition both simplifies and increases the efficiency of all subsequent operations". The efficiency of the classification of inbreds into heterotic University of Ghana http://ugspace.ug.edu.gh 136 groups greatly depends on the classification methods used (Fan et al., 2009). These authors proposed the use of the percentage of high yielding hybrids obtained across the total number of inter-heterotic group crosses as a means of assessing the grouping efficiency of a classification method. In this study the grouping efficiency was 95% for the HSGCA method versus 77% for the SCA_PY method. The HSGCA method was also more efficient in identifying heterotic groups that display the highest heterosis for grain yield compared to the SCA_PY method. In contrast the SCA_PY method was more efficient in identifying inter-group crosses that showed the highest average grain yield (5449 kg ha-1) compared to the HSGCA method (5055 kg ha-1). The heterotic grouping based on the SCA_PY would therefore be more useful for the future activities of the breeding program. The inbreds (TZEI 160, ENT 13 and TZEI 161 for the first group and TZEI 157, TZEI 129 TZEI 124 and ENT 4 for the second group) constituting the two heterotic groups defined by the SCA_PY method could be recombined according to heterotic classification to develop new inbred lines and new versions of the high yielding hybrids ENT 13 x TZEI 124 and TZEI 160 x TZEI 157.  The purpose of multilocational trials (MLT) is to determine genotypes whose performance is consistent across environments and those that perform better only in specific environments (Yan and Tinker, 2006). The interaction between genotypes and environments is also useful as it provides guidelines for the choice of crucial test sites in a wide adaptation prospect (Annicchiarico, 1997). A requirement for successful MLTs is an adequate variability among test environments. In this study significant mean squares for grain yield and related traits were found among environments suggesting that each environment is different and unique. This implies that valid inferences about the broad or University of Ghana http://ugspace.ug.edu.gh 137 specific adaptation of the tested hybrids could be drawn from their evaluation in those environments. The significant mean squares for grain yield and other traits among the single cross hybrids under drought, well-watered conditions and across environments suggests that the hybrids displayed these traits differently. Therefore valid inferences on the agronomic performance of the hybrids could be made from this study. The least significant difference test used for mean separation combined with the stability test based on the GGE biplot analysis and the selection index under drought revealed that the single cross hybrids TZEI 160 x TZEI 157 and TZEI 182 x ENT 13 were the best hybrids under drought and also the most stable high yielding hybrids across environments. They could therefore be recommended to farmers who grow maize under rain-fed conditions in Niger. The single cross TZEI 124 x ENT 13 was the outstanding hybrid under well-watered environments. It could therefore be considered as a good candidate for commercial production under irrigation.  Modified single crosses (MSCs) could be an attractive alternative to single cross hybrids when the low grain yield of inbreds remains a challenge to minimizing the costs of hybrid seed production (Castellanos et al., 2009). An important advantage of MSCs in commercial maize hybrid seed production is the higher yield potential of related single crosses (RSCs) used as female parents compared to inbred lines. In this study the RSCs used to develop the MSCs, TZEEI 95 x TZEEI 79, TZEEI 78 x TZEEI 79 and TZEEI 76 x TZEEI 82 displayed high inbred parent heterosis for grain yield of 37.3% to 92.1% under drought and 35.3% to 64% under well-watered conditions suggesting that RSCs were better female parents for increasing hybrid seed production. Another advantage of MSCs is that they show no yield penalty compared to their corresponding single cross University of Ghana http://ugspace.ug.edu.gh 138 hybrids (Fuhe et al., 1995; Lee et al., 2006; Castellanos et al., 2009). In this study three MSCs i.e. (TZEEI 76 x TZEEI 82)x TZEI124; (TZEEI 78 x TZEEI 79)x ENT13 and (TZEEI 95 x TZEEI 79)x CMZENT5 out-yielded the best single cross check under well-watered conditions (TZEI 124 x ENT13) by 19.3%, 17.8% and 17% respectively. Three other MSCs i.e.(TZEEI 76 x TZEEI 82)x CMKENT23; (TZEEI 76 x TZEEI 82)x CMZENT8 and (TZEEI 95 x TZEEI 79) x CMKENT8 out-yielded the best single cross check under drought (ENT13 x CamInbgp.1-17) by 91%, 67% and 66% respectively. These results suggest that the MSCs above could be recommended to seed companies in Niger for a better commercial hybrid seed production.  Earliness is of major importance to maize growers in Niger even those for whom water availability is not a great concern. However earliness is often associated with low grain yield potential (Sallah et al., 1995; Pswarayi and Vivek, 2008; Badu-Apraku and Oyekunle, 2012). A strategy to increase the chances of adoption of new maize improved cultivars by farmers is to expose them to wide genotypes diversity and give them the opportunity to compare the advantages of one type over the other (Tiwari et al., 2009). One hundred and three MSCs of varying maturity were developed and evaluated under drought and well- watered conditions in this study. Significant differences among MSCs were found for grain yield and all traits. The grain yield of MSCs ranged from 377 Kg ha-1 to 3383 Kg ha-1 under drought and from 1765 Kg ha-1 to 5399 Kg ha-1 under well-watered conditions. Under the two water treatments the intermediate and early groups yielded significantly higher than the extra-early group. These results indicate adequate genetic variability that can be tested in participatory variety selection trials on farmers’ fields. This University of Ghana http://ugspace.ug.edu.gh 139 will increase the chance of replacing the old and low yielding farmers OPVs by higher yielding improved cultivars. 6.3. Limits of the study and recommendations for future research  In this study a rapid rural appraisal method was adopted to understand the maize production system in Niger. This method is a less participatory approach (Freudenberger, 1999) because the information gathering process is time-challenged (Chambers, 1981) and farmers’ participation is limited to providing only information that meets the researcher’s objectives. Future studies on maize production systems in Niger should adopt more participatory approaches that will emphasize farmers’ active participation in the diagnosis of maize production constraints, determination of maize breeding priorities and planning projects that will meet farmers’ expectations from the maize breeding program. Furthermore the RRA sites covered by this study might not be representative of all the maize production zones in Niger. Future studies should target the different maize production zones in the country particularly the sites along the Komadougou River and Chad Lake in Diffa region which has a high potential for maize production.  The screening methodology to identify drought tolerant maize genotypes that involves stopping the irrigation two weeks before flowering is likely to work only in environments characterized by shallow soils, low relative humidity and frequent windy weather conditions. An alternative approach for applying the water stress should be adopted if the drought trials are to be conducted in environments where the soil water retention capacity is much higher i.e. INRAN stations of Diffa, Tara and Bengou. The application of rescue irrigations to reduce the stress intensity is likely to disturb the effect of water deficit on University of Ghana http://ugspace.ug.edu.gh 140 plants particularly when soil variability is high. Traits like days to anthesis and silking, anthesis silking interval and grain yield might not be reliable. Therefore the performance of maize genotypes inferred from such drought trials where recue irrigations have been applied should be validated in different environments where maize drought trials require no application of rescue irrigation.  The heterotic grouping based on the HSGCA method was proposed by Fan et al., (2009) as a good alternative to SCA_PY method in assigning new inbreds to known heterotic groups. In diallel studies where no known appropriate tester is included, this method tended to define a large number of heterotic groups which are of no practical use in a breeding program. This indicates that the diallel is probably not the best design to adopt in assessing the grouping efficiency of the HSGCA method. Similarly the heterotic grouping based on the SCA_PY method was not done appropriately because of the minor importance of SCA effects compared to GCA effects in this particular set of inbreds. Most SCA effects of crosses used to assign inbreds into heterotic groups were not significant. The use of SCA_PY method for an efficient classification of inbreds is more appropriate in lines x tester designs where elite inbred testers are available. Given the current situation where such elite testers are limited in the extra-early and early maizegermplasm, the use of diallel and North Carolina designs to identify testers and attempt the heterotic grouping of inbreds is a valid approach even though these designs are not the best to identify elite single cross hybrids with high and significant SCA effects for grain yield (Duvick, 2005a; Mikel, 2008; Gracen, 2012). University of Ghana http://ugspace.ug.edu.gh 141  The evaluation of the modified single crosses under water stress and well watered conditions at Konni during the dry season 2013/2014 was lost due to a failure of the irrigation system. Only results from the evaluation of the MSCs under drought and well- watered conditions at the INRAN station at Maradi were considered for the analysis. Therefore no strong conclusions could be made about the performance of the MSCs under drought and well-water environments. It would therefore be necessary to evaluate the MSCs at more sites to validate the results of the present study. University of Ghana http://ugspace.ug.edu.gh 142 REFERENCES Abbassian, A. (2008) Maize: International Market Profile. Food and Agriculture Organization of the United Nations. http://siteresources.worldbank.org/INTAFRICA/Resources/257994- 1215457178567/Maize_Profile.pdf. Abebe, G., Assefa, T., Harrun, H., Mesfine, T. & Al-Tawaha, A. (2005) Participatory selection of drought tolerant maize varieties using mother and baby methodology: A case study in the semi arid zones of the central rift valley of Ethiopia. World Journal of Agricultural Sciences, 1, 22-27. Aguiar, C. G., Schuster, I., Júnior, A. T. A., Scapim, C. A. & Vieira, E. S. N. (2008) Heterotic groups in tropical maize germplasm by test crosses and simple sequence repeat markers Genetics and Molecular Research, 7, 1233-1244 Ahmad, S., Khan, H. A., Javed, M. & Rehman, K. U. (2011) Roost composition and damage assessment of rose-ringed parakeet (Psittacula krameri) on maize and sunflower in agro- ecosystem of central Punjab, Pakistan International Journal of Agriculture and Biology, 13, 731-736. Akinwale, R. O., Badu-Apraku, B., Fakorede, M. A. B. & Vroh-Bi, I. (2014) Heterotic grouping of tropical early-maturing maize inbred lines based on combining ability in Striga-infested and Striga-free environments and the use of SSR markers for genotyping. Field Crops Research 156, 48-62. Annicchiarico, P. (1997) Additive main effects and multiplicative interaction (AMMI) analysis of genotype-location interaction in variety trials repeated over years. Theor Appl Genet, 94, 1072-1077. Badu-Apraku, B. & Akinwale, R. O. (2011) Cultivar evaluation and trait analysis of tropical early maturing maize under Striga-infested and Striga-free environments. Field Crops Research, 121, 186-194. Badu-Apraku, B., Akinwale, R. O., Ajala, S. O., Menkir, A., Fakorede, M. A. B., et al. (2011a) Relationships among traits of tropical early maize cultivars in contrasting environments. Agronomy Journal 103, 3, 717-729 Badu-Apraku, B., Akinwale, R. O., Franco, J. & Oyekunle, M. (2012) Assessment of reliability of secondary traits in selecting for improved grain yield in drought and low-nitrogen environments. Crop Science, 52, 2050–2062. Badu-Apraku, B., Fakorede, M. A. B., Oyekunle, M. & Akinwale, R. O. (2011b) Selection of extra-early maize inbreds under low N and drought at flowering and grain-filling for hybrid production. 56, 17-21. Badu-Apraku, B., Fakorede, M. A. B., Oyekunle, M., Yallou, G. C., Obeng-Antwi, K., et al. (2014) Gains in grain yield of early maize cultivars developed during three breeding eras under multiple environments Crop Science, doi:10.2135/cropsci2013.11.0783. University of Ghana http://ugspace.ug.edu.gh 143 Badu-Apraku, B., Fontem, L. A., Akinwale, R. O. & Oyekunle, M. (2011c) Biplot analysis of diallel crosses of early maturing tropical yellow maize inbreds in stress and nonstress environments. Crop Science Society of America, 51, 173-188. Badu-Apraku, B. & Lum, A. F. (2010) The pattern of grain yield response of normal and quality protein maize cultivars in stress and nonstress environments. Agronomy Journal, 102, 2, 381–394. Badu-Apraku, B., Menkir, A., Ajala, S. O., Akinwale, R. O., Oyekunle, M., et al. (2010) Performance of tropical early-maturing maize cultivars in multiple stress environments. Canadian Journal of Plant Science, 90, 831-852. Badu-Apraku, B., Menkir, A., Fakorede, M. A. B., Kamara, A. Y. & Adam, A. (2004) Effects of drought screening methodology on genetic variances and covariances in Pool 16 DT maize population. Journal of Agricultural Science 142, 445-452. DOI: 10.1017/S0021859604004538 Badu-Apraku, B. & Oyekunle, M. (2012) Genetic analysis of grain yield and other traits of extra- early yellow maize inbreds and Hybrid performance under contrasting environments. Field Crops Research, 129, 99-110. Badu-Apraku, B., Oyekunle, M., Akinwale, R. O. & Aderounmu, M. (2013a) Combining ability and genetic diversity of extra-early white maize inbreds under stress and non stress environments. Crop Science, 53, 9-26 Badu-Apraku, B., Oyekunle, M., Akinwale, R. O. & Lum, A. F. (2011d) Combining ability of early-maturing white maize inbreds under stress and non-stress environments. Agronomy Journal, 103, 544-557 Badu-Apraku, B., Oyekunle, M., Menkir, A., Obeng-Antwi, K., Yallou, C. G., et al. (2013b) Comparative performance of early maturing maize cultivars developed in three eras under drought stress and well-watered environments in West Africa. Crop Science, doi: 10.2135/cropsci2012.11.0640. Badu-Apraku, B., Vroh, I., Oyekunle, M., Akinwale, R. O., Fakorede, M. A. B., et al. (2013c) Combining ability, heterotic patterns and genetic diversity of extra-early yellow inbreds under contrasting environments. Euphytica, DOI 10.1007/s10681-013-0876-4. Baker, R. J. (1978) Issues in diallel analysis. Crop Science, 18, 533-536. Bänziger, M., Edmeades, G. O., Beck, D. & Bellon, M. (2000) Breeding for Drought and Nitrogen Stress Tolerance in Maize: From Theory to Practice. Mexico, D.F.CIMMYT. Bänziger, M., Setimela, P. S., Hodson, D. & Vivek, B. (2004) Breeding for improved drought tolerance in maize adapted to southern Africa "New directions for a diverse planet". Proceedings of the 4th International Crop Science Congress, 26 Sep - 1 Oct 2004, . Brisbane, Australia. Bationo, A. & Mokwunye, A. U. (1991) Alleviating soil fertility constraints to increased crop production in West Africa: The experience of the Sahel. Fertilizer Research, 29, 95-115. University of Ghana http://ugspace.ug.edu.gh 144 Betràn, F. J., Beck, D., Bänziger, M. & Edmeades, G. O. (2003a) Genetic analysis of inbred and hybrid grain yield under stress and nonstress environments in tropical maize. Crop Science, 43, 807-817. Betràn, F. J., Beck, D., Bänziger, M. & Edmeades, G. O. (2003b) Secondary traits in parental inbreds and hybrids under stress and non-stress environments in tropical maize. Field Crops Research, 83, 51-65. Bhandari, B. (2003) ParticipatoryRural Appraisal (PRA). Institute for Global Environmental Strategies (IGES). Birchler, J. A., Yao, H., Chudalayandi, S., Vaiman, D. & Veitiac, R. A. (2010) Heterosis. The plant Cell, 22, 2105-2112. Blum, A. (1979) Genetic improvement of drought resistance in crop plants: a case for sorghum. pp. 429-445 IN MUSSELL, H. & STAPLES, R. C. (Eds.) Stress Physiology in Crop Plants. Wiley, Interscience, New York. Bolaños, J. & Edmeades, G. O. (1993) Eight cycles of selection for drought tolerance in lowland tropical maize. I. Responses in grain yield, biomass, and radiation utilization. Field Crops Research, 31, 233-252 Bolaños, J. & Edmeades, G. O. (1996) The importance of the anthesis-silking interval in breeding for drought tolerance in tropical maize. Field Crops Research, 48, 65-80. Bolaños, J., Edmeades, G. O. & Martinez, L. (1993) Eight cycles of selection for drought tolerance in lowland tropical maize. III. Responses in drought-adaptive physiological and morphological traits. Field Crops Research, 31, 269-286. Butler, C. J. (2003) Population biology of the introduced rose-ringed parakeet Psittacula krameri in the UK PhD Thesis. University of Oxford Department of Zoology. Edward Grey Institute of Field Ornithology Campbell, J. R. (2001) Participatory Rural Appraisal as Qualitative Research: Distinguishing methodological issues from participatory claims. Human Organization, 60. Campos, H., Cooper, M., Habben, J. E., Edmeades, G. O. & Schussler, J. R. (2004) Improving drought tolerance in maize: a view from industry. Field Crops Research, 90, 19-34. Castellanos, J. S., Córdova, H. S., Srinivasan, G., Quemé, J. L., Rincón, F., et al. (2009) Exploiting modified single crosses in maize (Zea mays L.) to facilitate hybrid use in developing countries. Maydica, 54, 77-90. Cattivelli, L., Rizza, F., Badeck, F., Mazzucotelli, E., Mastrangelo, A. M., et al. (2008) Drought tolerance improvement in crop plants: An integrated view from breeding to genomics. Field Crops Research, 105, 1-14. Chambers, R. (1981) Rapid Rural Appraisal: Rationale And Repertoire. Public Administration And Development, 1, 95-106. University of Ghana http://ugspace.ug.edu.gh 145 Chapman, S. C., Crossa, J. & Edmeades, G. O. (1997) Genotype by environment effects and selection for drought tolerance in tropical maize. I. Two mode pattern analysis of yield. Euphytica 95, 1-9. Charlesworth, D. & Willis, J. H. (2009) The genetics of inbreeding depression. Nature Reviews Genetics, 10, 783-796. Christianson, C. B. & Vlek, P. L. G. (1991) Alleviating soil fertility constraints to food production in West Africa: Efficiency of nitrogen fertilizers applied to food crops. Fertilizer Research, 29, 21-33. Defoer, T., Kamara, A. & DeGroote, H. (1997) Gender and variety selection: Farmers’ assessment of local maize varieties in southern Mali. African Crop Science Journal, 5, 65-76. Denmead, O. T. & Shaw, R. H. (1960) The effects of soil moisture stress at different stages of growth on the development and yield of corn. Agronomy Journal, 52, 272-274. Denning, G. K., P. Sanchez, P. Malik, A. Flor, R (2009) Input Subsidies to Improve Smallholder Maize Productivity in Malawi: Toward an African Green Revolution. PLoS Biol 7(1): e1000023. doi:10.1371/journal.pbio.1000023. 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. PhD. University of KwaZulu-Natal. African Centre for Crop Improvement (ACCI)/ School of Biochemistry, Genetics, Microbiology and Plant Pathology Derera, J., Tongoona, P., Vivek, B. S. & Laing, M. D. (2008) Gene action controlling grain yield and secondary traits in southern African maize hybrids under drought and non-drought environments Euphytica 162, 411-422. DOI 10.1007/s10681-007-9582-4 Doss, C. R., Mwangi, W., Verkuijl, H. & DeGroote, H. (2003) Adoption of maize and wheat technologies in Eastern Africa: A synthesis of the findings of 22 cases studies. . CIMMYT Economics Working Papers 03-06. Mexico, D.F.: CIMMYT. Druilhe, Z. & Barreiro-Hurlé, J. (2012) Fertilizer subsidies in sub-Saharan Africa. ESA Working paper No. 12-04. Rome, FAO. Duvick, D. N. (1999) Heterosis: Feeding people and protecting natural resources. IN COORS, J. G. & PANDLEY, S. (Eds.) The genetics and exploitation of heterosis in crops. Madison, WI., American Society of Agronomy, Inc., Crop Science Society of America, Inc., Soil Science Society of America, Inc. Duvick, D. N. (2005a) The contribution of breeding to yield advances in maize (Zea mays L.). Adv. Agron, 86, 83-145. Duvick, D. N. (2005b) Genetic progress in yield of united states maize (Zea mays L.). Maydica, 193-202, 50. Edmeades, G. O. (2008) Drought Tolerance in Maize: An Emerging Reality. A Feature In James,Clive. 2008. Global Status of Commercialized Biotech/GM Crops: 2008. ISAAA Brief No. 39. ISAAA: Ithaca, NY. University of Ghana http://ugspace.ug.edu.gh 146 Edmeades, G. O., Bolaños, J. & Chapman, S. C. (1997) Value of secondary traits in selecting for drought tolerance in tropical maize. pp. 222 - 234 IN EDMEADES, G. O., BÄNZIGER, M., MICKELSON, H. R. & PEÑA-VALDIVIA, C. B. (Eds.) Developing drought and low- N tolerant maize. CIMMYT, Mexico Fan, X. M., Zhang, Y. M., Yao, W. H., Chen, H. M., Tan, J., et al. (2009) Classifying maize inbred lines into heterotic groups using a factorial mating design. Agronomy Journal, 101, 106– 112. FAOSTAT (2015) Statistical Database of the Food and Agriculture of the United Nations. http://www.fao.org. FARA (2009) Patterns of change in maize production in Africa: Implications for maize policy development. Forum for Agricultural Research in Africa. Ministerial Policy Brief Series Number 3, December 2009. Fasoula, V. A. & Tollenaar, M. (2005) The impact of plant population density on crop yield and response to selection in maize. Maydica, 50, 39-48. Fischer, K. S., Johnson, E. C. & Edmeades, G. O. (1982) Breeding and selection for drought resistance in tropical maize. Symposium on Principles and Methods in Crop Improvement for Drought Resistance: With emphasis on Rice. Manila (Philippines) 4-8 May 1981, Mexico, DF (Mexico). Freudenberger, K. S. (1999) Rapid Rural Appraisal (RRA) and ParticipatoryRural Appraisal (PRA). A Manual for CRS Field Workers and Partners. Baltimore, Catholic Relief Services. Fuhe, L., Fengyin, L., Liangyue, J., Yianmin, H., Xilin, H., et al. (1995) Studies on methods for developing modified single crosses of maizeII. Combining ability of sister inbreds and sister crosses. Acta Agronomica Sinica 5. Gibbon, D., Dixon, J. & Flores, D. (2007) Beyond drought tolerant maize: Study of additional priorities in maize. Report to Generation Challenge Program. CIMMYT. Gracen, V. (2012) Elite inbred testers, heterotic groups and maize breeding. Cornell University & West Africa Center for Crop Improvement/UG-Legon. Griffing, B. (1956) Concept of general and specific combining abilitybin relation to diallel crossing systems. Australian Journal of Biological Sciences, 9, 4, 463 - 493. Gruhn, P., Goletti, F. & Yudelman, M. (2000) Food, Agriculture, and management, soil fertility, and sustainable agriculture: Current issues and future challenges. Food, Agriculture, and the Environment Discussion Paper 32. International Food Policy Research Institute. Washington, D.C. U.S.A. Habsatou, B., Saminou, E., Idrissa, Y. & Salifou, A. (2010) Diagnostic participatif rapide et planification des actions du périmètre de Djirataoua. Association Nigérienne pour l'Irrigation et le Drainage (ANID). University of Ghana http://ugspace.ug.edu.gh 147 Hallauer, A. R., Carena, M. J. & Miranda, J. B. F. (2010) Quantitative genetics in maize breeding, 3rd ed. Handbook of Plant Breeding Volume 6, New York, NY 10013, USA, Springer Science+Business Media, LLC. Hallauer, A. R. & Miranda, J. B. F. (1981) Quantitative genetics in maize breeding, 1st ed. Ames, Iowa State University Press. Harris, K. M. & Nwanze, K. F. (1992) Busseola fusca (Fuller), the African maize stalk borer: a handbook of information. Information Bulletin no. 33. Patancheru, A.P. 502 324, India: International Crops Research Institute for the Semi-Arid Tropics, and Wallingford, UK: CAB International. 92 pp., HCI-3N (2012) Initiative ‘‘ 3N’’ pour la sécurité alimentaire et nutritionnelle et le développement agricole durable : Plan d’investissement 2012-2015. Haut Commissariat à l'Initiative 3N, Volume 1. Hede, A. R., Srinnivasan, G., Stolen, O. & Vasal, S. K. (1999) Identification of heterotic pattern in tropical inbred maize lines using broad-based synthetic testers. Maydica, 44, 325-331. Heisey, P. W. & Mwangi, W. (1996) Fertilizer use and maize production in Sub-Saharan Africa. CIMMYT Economics Working Paper 96-01. Mexick, D.F.:CIMMYT. Hung, H. Y. & Holland, J. B. (2012) Diallel analysis of resistance to fusarium ear rot and fumonisin contamination in maize. Crop Science, 52, 2173-2181 IAEA (2013) Optimizing productivity of food crop genotypes in low nutrient soils.Vienna : International Atomic Energy Agency, 2013. IAEA-TECDOC series, ISSN 1011-4289. Ifie, B. E. (2013) Genetic analysis of Striga resistance and low soil nitrogen tolerance in early maturing maize (Zea mays L.) inbred lines. PhD. University of Ghana Legon. West Africa Centre for Crop Improvement/ School of Agriculture/ College of Agriculture and Consumer Sciences INS-Niger (2013) Annuaire statistique des cinquante ans d'indépendance du Niger. Institut National des Statistiques du Niger. http://www.stat-niger.org/statistique/. JAICAF (2009) Les céréales au Niger - accent sur le mil et le sorgho -. JAICAF L’Association pour la Collaboration Internationale en matière d’Agriculture et de Forêts du Japon. Jaleel, C. A., Manivannan, P., Wahid, A., Farooq, M., Somasundaram, R., et al. (2009) Drought stress in plants: a review on morphological characteristics and pigments composition. International Journal of Agriculture & Biology, 11, 100-105. Jika, N. (1991) Situation de la production du maïs et état actuel de la recherche pour le developpement de cette céréale au Niger. . Contribution à l'atelier conjoint des réseaux SAFGRAD, Niamey, 8-14 Mars 1991. INRAN. Kalule, T., Ogenga-Latigo, M. W. & Okoth, V. A. O. (1997) Seasonal fluctuations and damage of lepidopteran stemborers of maize in a major agroecozone of Uganda. African Crop Science Journal, 5, 4. University of Ghana http://ugspace.ug.edu.gh 148 Kamara, A. Y., Menkir, A., Badu-Apraku, B. & Ibikunle, O. (2003) The influence of drought stress on growth, yield and yield components of selected maize genotypes. Journal of Agricultural Science 141, 43-50. Keating, B. & Wafula, B. (1990) Simulation of plant density effects on maize yield as influenced by water and nitrogen limitations. http://www.bob-mccown.com/wp- content/uploads/2011/10/56_KeatingEtAl1990S... . Kempthorne, O. (1955) The theory of the diallel cross. Statistical Laboratory, Iowa State College, Ames, Iowa, http://www.genetics.org/cgi/reprintframed/41/4/451. Kouamé, K. L. (1995) Seasonal abundance of the two maize stemborers Sesamina calamistis and Eldana saccharina, and bionomics of the Sesamina egg parasite Telenomus busseolae. PhD Thesis. SIMON FRASER UNIVERSITY. Lafitte, H. R., Blum, A. & Atlin, G. (2003) Using secondary traits to help identify drought-tolerant genotypes pp. 37- 48 IN FISCHER, K. S., LAFITTE, R. H., FUKAI, S., ATLIN, G. & HARDY, B. (Eds.) Breeding rice for drought-prone environments IRRI, Los Baños Lee, E. A., Singh, A., Ash, M. J. & Good, B. (2006) Use of sister-lines and the performance of modified single-cross maize hybrids. Crop Science, 46, 312–320 Legesse, B. W., Pixley, K. V. & Botha, A. M. (2009) Combining ability and heterotic grouping of highland transition maize inbred lines. Maydica, 54, 1-9. Li, L., Lu, K., Chen, Z., Mu, T., Hu, Z., et al. (2008) Dominance, overdominance and epistasis condition the heterosis in two heterotic rice hybrids. Genetics, 180, 1725-1742 MA-Niger (2012) Catalogue National des Espèces et Variétés Végétales (CNEV). Ministere de L'agriculture, Republique du NIGER. Ma, X. Q., Tang, J. H., Teng, W. T., Yan, J. B., Meng, Y. J., et al. (2007) Epistatic interaction is an important genetic basis of grain yield and its components in maize. Mol Breeding 20, 41- 51. Makumbi, D., Betrán, J. F., Bänziger, M. & Ribaut, J. (2011) Combining ability, heterosis and genetic diversity in tropical maize (Zea mays L.) under stress and non-stress conditions. Euphytica. DOI 10.1007/s10681-010-0334-5 Melani, M. D. & Carena, M. J. (2005) Alternative maize heterotic patterns for the northern corn belt. Crop Science, 45, 2186-2194. Melchinger , A. E. & Gumber, R. K. (1998) Overview of heterosis and heterotic groups in agronomic crops. pp. 29-44 IN LAMKEY, K. R. & STAUB, J. E. (Eds.) Concepts and breeding of heterosis in crop plants. Madison, WI, CSSA. Melchinger, A. E., Piepho, H. P., Utz, H. F., Muminovic, J., Wegenast, T., et al. (2007a) Genetic basis of heterosis for growth-related traits in Arabidopsis Investigated by testcross progenies of near-isogenic lines reveals a significant role of epistasis. Genetics, 177, 1827- 1837 University of Ghana http://ugspace.ug.edu.gh 149 Melchinger, A. E., Utz, H. F., Piepho, H. P., Zeng, Z. B. & Schön, C. C. (2007b) The role of epistasis in the manifestation of heterosis: A systems-oriented approach. Genetics, 177, 1815-1825 Menkir, A., Melake-Berhan, A., The, A. C., Ingelbrecht , I. & Adepoju, A. (2004) Grouping of tropical mid-altitude maize inbred lines on the basis of yield data and molecular markers. Theor Appl Genet, 108, 1582-1590. Messmer, R., Fracheboud, Y., Bänziger, M., Vargas, M., Stamp, P., et al. (2009) Drought stress and tropical maize: QTL-by-environment interactions and stability of QTLs across environments for yield components and secondary traits. Theor Appl Genet, 119, 913-930. Mikel, M. A. (2008) Genetic diversity and improvement of contemporary proprietary North American dent corn. Crop Science, 48, 1686-1695. Monneveux, P., Sanchez, C. & Tiessen, A. (2008) Future progress in drought tolerance in maize needs new secondary traits and cross combinations. 146, 287-300. Moorad, J. A. & Wade, M. J. (2005) A genetic interpretation of the variation in inbreeding depression. Genetics Society of America, 170, 1373-1384. Morris, M., Kelly, V. A., Kopicki, R. J. & Byerlee, D. (2007) Fertilizer use in African agriculture: Lessons learned and good practice guidelines International Bank for Reconstruction and Development / World Bank 1818 H Street, NW Washington, DC Mossi, M. I., Saidou, A. M., Souleymane, I. A. & Haya, R. (2010) Diagnostic participatif rapide et planification des actions du périmètre de Galmi. Association Nigérienne pour l'Irrigation et le Drainage (ANID). Nduwumuremyi, A., Tongoona, P. & Habimana, S. (2013) Mating designs: Helpful tool for quantitative plant breeding analysis. Journal of Plant Breeding and Genetics, 03, 117-129. NeSmith, D. S. & Ritchie, J. T. (1992) Effects of soil water-deficit during tassel emergence on development and yield components of maize (Zea mays L.). Field Crops Research, 28, 251- 256. Njoku, D. N. (2012) Improving beta-carotene content in farmers’ preferred cassava cultivars in Nigeria. PhD. University of Ghana Legon. West Africa Centre for Crop Improvement/ School of Agriculture/ College of Agriculture and Consumer Sciences Obaa, B., Chanpacho, M. & Agea, J. G. (2005) Participatory farmers’ evaluation of maize varieties: A case study from Nebbi District, Uganda. African Crop Science Society, 7, 1389-1393. Oigiangbe, O. N., Ivbijaro, M. F., Ewete, F. K. & Mutsaers, H. J. W. (1997) Incidence and damage caused by maize stemborers on farmers' fields in south western nigeria. African Crop Science Journal, 5, 3, 295-302. Ouma, J. O., Odendo, M., Bett, C., De Groote, H., Mugo, S., et al. (2010) Participatory Farmer Evaluation of Stem borer Resistant Maize varieties in three maize growing ecologies of Kenya. Joint Conference: 3rd African Association of Agricultural Economists (AAAE) University of Ghana http://ugspace.ug.edu.gh 150 and 48th Agricultural Economists Association of South Africa (AEASA), Cape Town, South Africa, September 19-23, 2010. PPCR (2010) Strategic program for climate resilience Niger Pilot Programme for Climate Resilience (PPCR) Climate Investment Fund Washington, D.C. Pswarayi, A. & Vivek, B. (2008) Combining ability amongst CIMMYT’s early maturingmaize (Zea mays L.) germplasm under stress and non-stress conditions and identification of testers. Euphytica, 162, 353-362. Reif, J. C., Hallauer, A. R. & Melchinger, A. E. (2005) Heterosis and heterotic patterns in maize. Maydica, 50, 215-223. Ribaut, J., Betran, J., Monneveux, P. & Setter, T. (2009) Drought Tolerance in Maize. IN BENNETZEN, J. L. & HAKE, S. C. (Eds.) Handbook of Maize: Its Biology. Springer Science & Business Media. 331-344. Russell, W. A. (1991) Genetic improvement of maize yields. Adv. Agron, 46, 245-298. Sallah, P. Y. K., Twumasi-Afriyie, S. & Kasei, C. N. (1995) Grain productivities of four maturity groups of maize varieties in the Guinea savanna. pp. 173-178 IN BADU-APRAKU, B., AKORODA, M.O., OUEDRAOGO, M., QUIN, F.M (Ed.) Contributing to food self sufficiency: Maize Research and Development in West and Central Africa. Proceedings of a regional maize workshop. May 28–June 2, 1995. IITA-Cotonou, Benin Republic. Sangoi, L. (2000) Understanding plant density effects on maize growth and development: an important issue to maximize grain yield. Ciência Rural, Santa Maria, 31, 159-168. Sangoi, L., Gracietti, M. A., Rampazzo, C. & Bianchetti, P. (2002) Response of Brazilian maize hybrids from different eras to changes in plant density. Field Crops Research 79, 39-51. Setter, T. L., Yan, J., Warburton, M., Ribaut, J. M., Buckler, E. S., et al. (2011) Genetic association mapping identifies single nucleotide polymorphisms in genes that affect abscisic acid levels in maize floral tissues during drought. Journal of Experimental Botany, 62, 701-716. Sevcik, J. (2014) Rose-ringed Parakeet (Psittacula krameri). Retrieved on Januray 13th, 2014. http://www.naturephoto-cz.com/psittacula-krameri-photo_lat-1792.html. Sharma, J. R. (2006) Statistical and biometrical techniques in plant breeding, 1st Edition, New Delhi India, New Age International. Shull, G. H. (1948) What is ''heterosis''. Genetics, 33, 439-446. Sibiya, J., Tongoona, P., Derera, J. & Makanda, I. (2013) Farmers’ desired traits and selection criteria for maize varieties and their implications for maize breeding: A case study from a KwaZulu-Natal Province, South Africa. Journal of Agriculture and Rural Development in the Tropics and Subtropics, 114, 1, 39-49. Smithson, P. C. & Giller, K. E. (2002) Appropriate farm management practices for alleviating N and P deficiencies in low-nutrient soils of the tropics. Plant and Soil 245, 169-180. University of Ghana http://ugspace.ug.edu.gh 151 Springer, N. M. & Stupar, R. M. (2007) Allelic variation and heterosis in maize: How do two halves make more than a whole? Cold Spring Harbor Laboratory Press, 17, 264-275 Swanson-Wagner, R. A., Y., Y. J., Rhonda DeCook, R., Borsuk, L. A., Nettleton, D., et al. (2006) All possible modes of gene action are observed in a global comparison of gene expression in a maize hybrid and its inbred parents. PNAS 103, 6805-6810. Tang, J., Yan, J., Ma, X., Teng, W., Wu, W., et al. (2010) Dissection of the genetic basis of heterosis in an elite maize hybrid by QTL mapping in an immortalized F2 population. Theor Appl Genet, 120, 333-340. Tiwari, T. P., Virk, D. S. & Sinclair, F. L. (2009) Rapid gains in yield and adoption of new maize varieties for complex hillside environments through farmer participation. Improving options through participatory varietal selection (PVS). Field Crops Research, 111, 137– 143. Tollenaar, M. (1992) Is low plant density a stress in maize? . Maydica, 37, 2, 305-311. Tracy, W. F. & Chandler, M. A. (2004) The historical and biological basis of the concept of heterotic groups in corn belt dent maize. pp. IN LAMKEY, K. & LEE, M. (Eds.) Hallauer Plant Breeding Symposium. Iowa State University Press, Ames, Iowa. Tsonev, S., Todorovska, E., Avramova, V., Kolev, S., Abu-Mhadi, N., et al. (2009) Genomics assisted improvement of drought tolerance in maize: QTL approaches. Biotechnol. & Biotechnol. Eq, 23, 1410-1413. VandenBerg, J. (2012) Field-resistance of the African maize stem borer to Bt maize: what did we learn? Scientific Conference: Advancing the understanding of biosafety GMO risk assessment, independent biosafety research and holistic analysis. 28 - 29 September 2012 Hyderabad / India. Witcombe, J. R., Hollington, P. A., Howarth, C. J., Reader, S. & Steele, K. A. (2008) Breeding for abiotic stresses for sustainable agriculture. Phil. Trans. R. Soc. B 363, 703-716. doi:10.1098/rstb.2007.2179 Witcombe, J. R., Joshi, K. D., Gyawali, S., Musa, A. M., Johansen, C., et al. (2005) Participatory plant breeding is better described as highly client-oriented plant breeding. I. four indicators of client-orientation in plant breeding. Expl Agric. , 41, 299-319. Wolf, J. M., Levine, G., Naderman, G. C. & Gonzalez, E. E. (1974) Adverse soil-water conditions and corn production in Central Brazil. Page 47 in Abstracts of the 66th Annual Meeting of the American Society of Agronomy. Chicago, Illinois. Wolfe, D. W., Henderson, D. W., Hsiao, T. C. & Alvino, A. (1988) Interactive water and nitrogen effects on senescence of maize. 11. Photosynthetic decline and longevity of individual leaves. Agronomy Journal, 80, 865-870. Yan, W. & Tinker, N. A. (2006) Biplot analysis of multi-environment trial data: Principles and applications. Canadian Journal of Plant Science, 623-645. University of Ghana http://ugspace.ug.edu.gh 152 Zhang, Y. & Kang, M. S. (1997) DIALLEL-SAS: A SAS program for Griffing's diallel analyses. Agronomy Journal, 89, 176-182 Zhang, Y., Kang, M. S. & Lamkey, K. R. (2005) DIALLEL-SAS05: A comprehensive program for Griffing’s and Gardner-Eberhart analyses. Agronomy Journal, 97, 1097-1106 University of Ghana http://ugspace.ug.edu.gh 153 APPENDICES Appendix 3.1. Checklist for Focus group discussion 1. Maize production system 1.1 What is the importance of maize among the cereals grown in your village? 1.2 Do you grow maize under irrigation or rainfed conditions? 1.3 Is maize grown in mono or poly-cropping system? At what period of the year do you grow maize? 1.4 Do you grow maize as food or cash crop? 1.5 What type of usage is maize used for? 1.6 Estimate the proportion of area devoted to maize production compared to other crops in your village? 1.7 How does maize production impact your income? 2. Maize varieties and crop management 2.1 What varieties are you currently growing on your fields? What are their characteristics (endosperm colour, plant height, maturity, yield etc.)? 2.2 What is your most preferred variety? Why? 2.3 Where do you get seeds for planting? 2.4 Can you describe how you manage your maize field from land preparation to harvest (land ploughing, planting, plant density, weeding, organic and mineral fertilizer application etc.)? 3. Farmers preferred traits 3.1 Give maize cultivars characteristics you pay attention to when selecting the variety you decide to grow on your fields 3.2 Give in descending order the most important traits that you take into consideration in adopting new maize cultivars 4. Maize production constraints 4.1 What problems do you face in maize production? 4.2 Give in descending order the most important ones 4.3 What are the solutions you are currently using to deal with those problems? University of Ghana http://ugspace.ug.edu.gh 154 4.4. What are your expectations from the maize breeding program of INRAN? 5. New technologies 5.1 Are you aware of maize new cultivars released by INRAN or other research institutions or seed companies outside the country? 5.2 Do you have access to those improved cultivars? 5.3 Have you ever received training on maize production? If yes what the training was about? Who gave the training? Where and When? 5.4 Do you receive any technical or material support from the agricultural extension services? 5.5 Do you need to collaborate with INRAN researchers and possibly participate training, participatory varieties selection and on-farm trails? 6. Simulation of maize planting and fertilizer application 6.1. Using your planting tools show how you sow maize seeds and how you space plants within and among rows 6.2. For each of the mineral fertilizers you are using in maize production indicate approximately how and what quantity of the fertilizer you apply in your fields. University of Ghana http://ugspace.ug.edu.gh 155 Appendix 4.1.Grain yield and other agronomic traits of 105 single cross hybrids derived from 15 × 15 diallel cross among 15 early maturing yellow- grained maize inbred lines and 5 checks evaluated under drought stress and well-watered environments Hybrids Yield Kg ha -1 DYA DYS ASI PLTH (cm) PASP EPP EASP SGC2 Index WWC WSC Across WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WSC WSC TZEI 160 x TZEI 129 5599 1661 3911 57 70 57 75 1 6 180 145 2.8 3.3 1.2 0.9 2.3 3.9 7.0 -0.55 TZEI 160 x ENT 15 4734 2020 3571 58 68 58 72 1 4 164 145 3.1 2.8 1.1 1.0 2.7 3.5 6.8 2.96 TZEI 160 x ENT 17 4732 2393 3730 59 70 60 74 1 4 165 144 3.3 3.0 1.2 1.1 2.5 3.6 6.7 4.18 TZEI 160 x TZEI 182 3808 1942 3008 57 66 57 72 1 7 155 141 3.4 3.5 1.2 1.0 3.1 3.6 6.3 1.46 TZEI 160 x TZEI 167 3801 1128 2655 60 72 60 78 0 6 155 125 3.3 4.2 1.1 1.0 2.9 4.2 7.3 -3.91 TZEI 160 x TZEI 14 3865 1212 2728 59 71 60 79 1 9 156 130 3.3 3.5 1.1 0.8 2.9 3.9 6.5 -3.12 TZEI 160 x TZEI 124 5483 1819 3913 55 66 56 74 1 8 178 166 2.4 3.2 1.1 0.9 1.9 3.5 8.0 -0.92 TZEI 160 x TZEI 16 4884 1546 3454 58 71 59 76 1 6 173 140 3.0 3.3 1.4 1.1 2.5 3.8 6.3 0.76 TZEI 160 x ENT 4 5070 1923 3717 57 69 57 75 0 6 172 143 3.1 3.5 1.1 1.0 2.2 3.2 5.7 2.69 TZEI 160 x TZEI 161 2533 1043 1894 59 70 58 77 -1 6 131 111 4.0 4.2 1.1 0.9 4.1 4.4 7.3 -4.68 TZEI 160 x TZEI 23 2596 1213 2003 58 67 58 73 -1 5 132 121 3.8 4.3 1.1 0.8 3.9 4.3 7.0 -4.15 TZEI 160 x ENT 13 5190 2222 3918 57 69 57 73 -1 5 172 149 2.9 2.8 1.2 1.2 2.3 3.3 7.7 3.69 TZEI 160 x TZEI 135 5350 2210 4004 56 68 57 71 0 4 175 151 2.6 3.5 1.2 1.0 2.5 3.3 7.0 3.07 TZEI 160 x TZEI 157 5708 2570 4363 56 68 57 72 1 5 182 157 2.8 3.0 1.2 1.1 2.5 3.0 6.7 5.48 TZEI 129 x ENT 15 4770 1591 3408 59 72 61 78 2 6 193 152 3.3 3.5 1.1 1.1 2.2 3.7 6.7 0.24 TZEI 129 x ENT 17 5141 2399 3966 59 71 60 75 1 4 191 156 3.0 3.0 1.2 1.0 2.2 3.1 6.2 5.26 TZEI 129 x TZEI 182 4915 2216 3758 57 71 58 77 1 6 176 144 3.3 3.2 1.1 1.2 2.4 3.1 6.0 5.01 TZEI 129 x TZEI 167 5528 1873 3962 59 72 60 78 2 6 186 160 2.4 3.3 1.1 1.1 1.9 3.6 6.2 2.21 TZEI 129 x TZEI 14 4943 1558 3492 60 71 61 78 2 7 192 150 2.8 3.7 1.1 0.9 2.0 3.6 6.0 -0.16 TZEI 129 x TZEI 124 4958 1048 3282 58 69 59 79 2 10 188 156 2.5 3.3 1.1 1.1 1.9 3.7 7.7 -3.36 TZEI 129 x TZEI 16 4826 1042 3204 59 73 62 79 3 6 192 147 2.9 4.0 1.2 1.1 2.4 4.4 7.0 -3.39 TZEI 129 x ENT 4 5607 1635 3905 58 70 59 78 1 7 189 153 2.5 3.7 1.1 1.0 1.9 3.6 7.2 -0.93 TZEI 129 x TZEI 161 5056 1550 3554 57 67 58 74 2 7 180 157 2.8 4.0 1.1 0.8 2.3 3.5 7.2 -2.1 TZEI 129 x TZEI 23 4803 2158 3669 57 68 57 75 1 7 172 141 3.3 3.5 1.2 1.1 2.8 3.5 6.5 2.32 Overall mean 4571 1684 3334 58 70 60 76 1 7 174 145 3.0 3.5 1.1 0.9 2.6 3.7 6.5 0.00 Lsd 865 775 605 1.5 2.4 1.7 3.9 1.3 3.2 12 15 0.5 0.9 0.1 0.3 0.4 0.7 1.1 5.4 Genotypes ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** * ** ** ** Environments ** ** ** ** ** ** ** ** ** ** ** ** * ** ** ** ns ** ** G x E ** * ** ** ** ** ns ** ** * ns ** ns ** ns ** * * ns WWC: well-watered conditions; WSC: water stress conditions; Across: across environments YIELD:Grain yield; DYA: days to anthesis; DYS: days to silk; ASI: anthesis silking interval; PLTH: plant height; PASP: plant aspect; EPP: number of ears per plant; EASP: ear aspect; SGC2: stay-green 2; Index: selection index under water stress conditions *Significant at the 0.05 probability level; **Significant at the 0.01 probability level ; ns: not significant University of Ghana http://ugspace.ug.edu.gh 156 Appendix 4.1. Continued Hybrids Yield Kg ha -1 DYA DYS ASI PLTH (cm) PASP EPP EASP SGC2 Index WWC WSC Across WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WSC WSC TZEI 129 x ENT 13 5255 2212 3951 57 70 59 75 2 5 201 160 2.6 3.3 1.1 0.9 1.8 2.9 6.8 3.36 TZEI 129 x TZEI 135 4417 1197 3037 58 69 60 78 2 9 196 155 3.1 4.2 1.1 0.9 2.8 3.9 6.0 -3.33 TZEI 129 x TZEI 157 4890 1661 3506 56 68 58 73 3 5 183 149 3.0 3.5 1.0 1.1 2.6 3.6 6.2 1.78 ENT 15 x ENT 17 4043 984 2732 61 74 61 82 1 8 183 139 3.1 4.3 1.1 0.7 2.6 4.1 7.5 -6.53 ENT 15 x TZEI 182 4272 1680 3161 57 68 59 75 2 7 164 151 2.9 3.3 1.1 1.0 2.6 3.8 6.2 0.49 ENT 15 x TZEI 167 4354 1069 2946 61 73 64 82 3 9 175 137 3.0 4.3 1.1 0.9 2.7 4.2 7.3 -5.68 ENT 15 x TZEI 14 4020 849 2661 60 74 63 83 3 9 177 137 3.4 4.3 1.1 0.7 2.6 4.3 7.2 -7.24 ENT 15 x TZEI 124 5290 736 3339 58 71 61 83 3 12 189 156 2.5 4.2 1.2 0.5 2.1 4.2 7.7 -9.55 ENT 15 x TZEI 16 4164 1776 3140 60 72 63 79 3 7 177 159 3.4 3.3 1.1 1.0 2.8 3.7 6.3 0.5 ENT 15 x ENT 4 2021 594 1410 64 77 66 84 2 7 167 131 4.4 4.3 1.0 0.8 3.4 4.2 7.2 -6.8 ENT 15 x TZEI 161 4825 1544 3419 59 70 60 78 1 8 184 150 3.1 3.5 1.1 0.8 2.2 3.7 6.5 -1.59 ENT 15 x TZEI 23 4916 2048 3687 58 67 58 73 1 7 168 137 2.8 3.2 1.2 1.1 2.9 3.5 6.7 2.54 ENT 15 x ENT 13 4846 1536 3427 61 74 62 82 1 7 195 142 2.9 3.8 1.0 0.9 2.1 3.7 6.3 -1.06 ENT 15 x TZEI 135 4777 1591 3412 57 70 60 77 3 7 184 144 3.0 3.7 1.1 0.9 2.5 3.6 7.3 -1.3 ENT 15 x TZEI 157 5534 1678 3882 57 68 59 75 2 7 181 147 2.3 3.5 1.1 0.9 2.2 3.5 7.2 -0.94 ENT 17 x TZEI 182 4427 1899 3344 58 69 59 74 0 6 162 140 3.1 3.2 1.2 1.0 2.6 3.4 5.2 3.51 ENT 17 x TZEI 167 5141 1579 3614 60 75 60 78 0 3 176 134 2.5 3.2 1.3 1.0 2.4 3.9 6.5 1.27 ENT 17 x TZEI 14 3947 1290 2809 61 74 62 81 2 7 174 140 3.1 4.0 1.2 0.8 2.9 3.8 6.2 -2.4 ENT 17 x TZEI 124 4859 721 3085 60 74 60 84 1 10 190 142 2.8 4.5 1.2 0.7 2.2 4.6 7.8 -9.15 ENT 17 x TZEI 16 3948 1237 2787 61 76 63 80 1 4 178 140 3.1 3.3 1.2 1.0 3.1 4.1 6.0 -0.11 ENT 17 x ENT 4 4582 1529 3274 61 74 62 78 1 5 181 140 3.3 3.3 1.1 0.8 2.4 3.9 6.7 -0.64 ENT 17 x TZEI 161 4964 2043 3712 58 69 58 74 0 6 172 152 3.1 3.3 1.2 0.9 2.6 3.5 7.0 1.52 ENT 17 x TZEI 23 4911 1797 3576 57 68 57 77 1 9 163 136 3.0 3.5 1.2 1.0 2.7 3.6 6.3 -0.12 ENT 17 x ENT 13 4647 1886 3463 60 71 61 77 1 6 182 142 3.0 2.7 1.2 1.0 2.4 3.5 5.8 3.22 Overall mean 4571 1684 3334 58 70 60 76 1 7 174 145 3.0 3.5 1.1 0.9 2.6 3.7 6.5 0.00 Lsd 865 775 605 1.5 2.4 1.7 3.9 1.3 3.2 12 15 0.5 0.9 0.1 0.3 0.4 0.7 1.1 5.4 Genotypes ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** * ** ** ** Environments ** ** ** ** ** ** ** ** ** ** ** ** * ** ** ** ns ** ** G x E ** * ** ** ** ** ns ** ** * ns ** ns ** ns ** * * ns WWC: well-watered conditions; WSC: water stress conditions; Across: across environments YIELD:Grain yield; DYA: days to anthesis; DYS: days to silk; ASI: anthesis silking interval; PLTH: plant height; PASP: plant aspect; EPP: number of ears per plant; EASP: ear aspect; SGC2: stay-green 2; Index: selection index under water stress conditions *Significant at the 0.05 probability level; **Significant at the 0.01 probability level ; ns: not significant University of Ghana http://ugspace.ug.edu.gh 157 Appendix 4.1. Continued Hybrids Yield Kg ha -1 DYA DYS ASI PLTH (cm) PASP EPP EASP SGC2 Index WWC WSC Across WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WSC WSC ENT 17 x TZEI 135 4336 1268 3021 58 71 60 79 2 8 188 141 3.4 3.8 1.1 0.9 2.6 3.9 6.2 -2.4 ENT 17 x TZEI 157 4787 1539 3395 59 71 60 77 0 6 184 147 2.8 3.7 1.2 1.0 2.4 3.8 6.5 -0.47 TZEI 182 x TZEI 167 4179 2063 3272 58 71 59 76 1 5 164 148 3.0 3.0 1.1 1.0 2.7 3.5 5.2 4.25 TZEI 182 x TZEI 14 4181 1604 3077 59 71 60 77 1 6 165 130 3.1 3.3 1.1 1.0 2.8 3.7 4.7 2.06 TZEI 182 x TZEI 124 4597 1759 3381 58 66 60 74 2 8 185 152 2.9 3.3 1.1 0.9 2.5 3.5 5.8 0.65 TZEI 182 x TZEI 16 4176 1937 3216 59 69 60 74 1 6 174 147 3.1 3.2 1.3 1.0 3.0 3.6 5.3 3.05 TZEI 182 x ENT 4 4322 1879 3275 57 69 59 76 1 7 175 146 3.0 3.3 1.1 1.2 2.6 3.4 6.2 2.51 TZEI 182 x TZEI 161 4045 2401 3340 57 67 58 71 2 4 166 145 3.4 3.2 1.1 1.1 2.8 3.4 4.7 6.4 TZEI 182 x TZEI 23 3850 1927 3026 54 65 55 73 1 7 148 131 3.1 3.5 1.1 0.9 3.1 3.9 5.8 0.66 TZEI 182 x ENT 13 5416 2753 4275 57 67 58 72 1 5 189 158 2.6 2.7 1.1 1.0 2.2 2.9 6.0 7.07 TZEI 182 x TZEI 135 3604 2388 3083 56 67 57 73 1 6 168 155 3.4 3.3 1.1 1.0 2.9 3.4 6.2 3.63 TZEI 182 x TZEI 157 2397 1298 1926 58 68 59 74 1 7 150 141 4.5 4.5 1.0 0.9 3.6 4.0 5.5 -2.29 TZEI 167 x TZEI 14 2248 1183 1792 64 76 66 80 2 4 151 131 3.8 4.2 1.0 1.1 3.5 4.3 5.7 -1.11 TZEI 167 x TZEI 124 5149 1200 3456 60 71 61 79 1 9 193 153 2.1 3.8 1.1 0.8 2.1 4.0 7.7 -4.73 TZEI 167 x TZEI 16 3465 1110 2455 61 76 63 81 2 5 166 133 3.4 4.0 1.1 0.9 3.2 4.4 6.3 -2.92 TZEI 167 x ENT 4 3996 952 2691 60 74 62 82 3 8 174 138 3.0 4.3 1.0 0.8 2.5 4.4 7.7 -6.85 TZEI 167 x TZEI 161 4100 1493 2982 58 70 58 76 0 7 158 137 3.3 3.5 1.1 0.8 2.9 4.0 6.2 -1.2 TZEI 167 x TZEI 23 3879 2031 3087 56 68 57 71 2 3 148 140 3.4 2.8 1.1 0.9 2.9 3.5 6.2 3.79 TZEI 167 x ENT 13 5711 2297 4248 60 72 61 77 2 5 181 141 2.4 2.8 1.1 1.0 2.0 3.2 6.2 4.59 TZEI 167 x TZEI 135 4216 2076 3299 59 73 61 76 1 3 169 141 3.0 3.0 1.1 1.0 2.7 3.8 6.3 3.19 TZEI 167 x TZEI 157 4715 1773 3454 59 70 60 76 1 6 178 149 2.6 3.3 1.0 0.9 2.4 3.7 6.5 0.64 TZEI 14 x TZEI 124 4614 1628 3334 59 68 61 78 2 10 177 146 2.6 3.3 1.1 0.7 2.4 3.7 6.2 -2.03 TZEI 14 x TZEI 16 1995 819 1491 65 81 67 85 2 4 150 119 4.1 4.2 1.1 0.8 3.8 4.5 6.2 -4.23 TZEI 14 x ENT 4 3457 1108 2450 61 74 63 79 3 6 164 144 3.4 4.2 1.2 0.9 3.0 4.3 6.7 -3.87 Overall mean 4571 1684 3334 58 70 60 76 1 7 174 145 3.0 3.5 1.1 0.9 2.6 3.7 6.5 0.00 Lsd 865 775 605 1.5 2.4 1.7 3.9 1.3 3.2 12 15 0.5 0.9 0.1 0.3 0.4 0.7 1.1 5.4 Genotypes ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** * ** ** ** Environments ** ** ** ** ** ** ** ** ** ** ** ** * ** ** ** ns ** ** G x E ** * ** ** ** ** ns ** ** * ns ** ns ** ns ** * * ns WWC: well-watered conditions; WSC: water stress conditions; Across: across environments YIELD:Grain yield; DYA: days to anthesis; DYS: days to silk; ASI: anthesis silking interval; PLTH: plant height; PASP: plant aspect; EPP: number of ears per plant; EASP: ear aspect; GC2: stay-green 2; Index: selection index under water stress conditions *Significant at the 0.05 probability level; **Significant at the 0.01 probability level ; ns: not significant University of Ghana http://ugspace.ug.edu.gh 158 Appendix 4.1. Continued Hybrids Yield Kg ha -1 DYA DYS ASI PLTH (cm) PASP EPP EASP SGC2 Index WWC WSC Across WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WSC WSC TZEI 14 x TZEI 161 4286 2130 3362 57 68 59 76 2 8 156 143 3.0 3.2 1.0 0.9 2.6 3.4 6.0 2.32 TZEI 14 x TZEI 23 3712 2251 3086 59 69 60 76 1 7 155 140 3.0 2.7 1.2 1.0 2.8 3.2 6.0 4.06 TZEI 14 x ENT 13 4858 2304 3763 60 71 62 78 2 7 180 144 2.6 3.3 1.1 1.0 2.4 3.5 6.3 2.51 TZEI 14 x TZEI 135 3949 1526 2911 59 71 62 81 3 10 171 142 3.4 3.5 1.1 1.0 3.0 3.8 5.8 -0.79 TZEI 14 x TZEI 157 3126 1218 2308 60 72 62 79 2 6 159 122 3.1 3.8 1.1 0.9 2.9 4.2 6.0 -2.31 TZEI 124 x TZEI 16 5121 1164 3425 59 70 61 80 2 10 185 146 2.4 3.7 1.1 0.8 2.1 4.0 6.8 -4.13 TZEI 124 x ENT 4 5097 649 3190 59 68 60 82 2 14 182 154 2.5 4.2 1.1 0.7 2.1 4.3 7.3 -9.46 TZEI 124 x TZEI 161 5205 1811 3750 57 68 57 74 0 6 180 158 2.6 3.2 1.1 0.9 1.9 3.4 8.0 -0.22 TZEI 124 x TZEI 23 5806 1897 4131 55 64 56 70 1 6 170 151 2.5 3.2 1.1 1.0 2.2 3.6 7.5 0.73 TZEI 124 x ENT 13 6584 1993 4617 57 68 58 77 1 9 194 164 1.1 2.7 1.1 1.0 1.4 3.3 6.2 2.48 TZEI 124 x TZEI 135 4634 949 3055 58 71 60 81 2 10 187 154 2.8 4.2 1.1 0.7 2.3 4.2 7.2 -7.1 TZEI 124 x TZEI 157 4550 1877 3405 59 71 60 77 2 6 190 153 2.1 3.0 1.0 0.9 2.3 3.7 5.5 2.2 TZEI 16 x ENT 4 4426 1254 3066 60 71 62 80 2 9 185 149 2.9 3.7 1.3 0.8 2.5 4.0 6.8 -4.02 TZEI 16 x TZEI 161 5652 2125 4140 57 69 59 76 1 7 175 142 2.9 2.7 1.2 1.0 2.5 3.3 6.7 2.83 TZEI 16 x TZEI 23 5032 1567 3547 58 69 59 77 1 8 160 135 3.0 3.5 1.3 1.0 2.6 3.9 6.5 -0.81 TZEI 16 x ENT 13 5097 2155 3836 60 71 62 77 2 6 197 161 2.5 2.8 1.1 1.1 2.3 3.2 6.0 4.71 TZEI 16 x TZEI 135 4501 1146 3063 60 72 62 80 2 8 180 143 3.0 4.0 1.3 1.0 2.6 4.1 6.0 -2.83 TZEI 16 x TZEI 157 5123 1914 3748 58 69 60 74 2 5 183 155 2.6 3.0 1.2 0.9 2.8 3.6 6.2 2.29 ENT 4 x TZEI 161 5288 2256 3989 58 69 59 75 1 6 178 145 3.1 3.7 1.2 1.0 2.1 3.1 6.0 3.36 ENT 4 x TZEI 23 5344 1840 3842 57 68 57 75 0 7 167 140 3.3 3.5 1.1 1.0 2.3 3.7 6.5 0.54 ENT 4 x ENT 13 5067 1444 3514 60 71 61 78 1 7 192 146 2.8 4.0 1.0 0.9 2.1 4.0 6.5 -2.18 ENT 4 x TZEI 135 4102 1376 2934 59 71 60 80 1 9 175 146 3.3 3.8 1.2 1.0 2.7 3.8 6.5 -2.11 ENT 4 x TZEI 157 4372 1690 3222 59 70 61 78 2 8 173 145 2.9 3.7 1.1 1.0 2.5 3.8 6.8 -0.98 TZEI 161 x TZEI 23 3343 1228 2436 57 68 57 74 0 6 142 124 3.8 4.5 1.2 0.7 3.7 4.2 7.0 -4.75 Overall mean 4571 1684 3334 58 70 60 76 1 7 174 145 3.0 3.5 1.1 0.9 2.6 3.7 6.5 0.00 Lsd 865 775 605 1.5 2.4 1.7 3.9 1.3 3.2 12 15 0.5 0.9 0.1 0.3 0.4 0.7 1.1 5.4 Genotypes ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** * ** ** ** Environments ** ** ** ** ** ** ** ** ** ** ** ** * ** ** ** ns ** ** G x E ** * ** ** ** ** ns ** ** * ns ** ns ** ns ** * * ns WWC: well-watered conditions; WSC: water stress conditions; Across: across environments YIELD:Grain yield; DYA: days to anthesis; DYS: days to silk; ASI: anthesis silking interval; PLTH: plant height; PASP: plant aspect; EPP: number of ears per plant; EASP: ear aspect; GC2: stay-green 2; Index: selection index under water stress conditions *Significant at the 0.05 probability level; **Significant at the 0.01 probability level ; ns: not significant University of Ghana http://ugspace.ug.edu.gh 159 Appendix 4.1. Continued Hybrids Yield Kg ha -1 DYA DYS ASI PLTH (cm) PASP EPP EASP SGC2 Index WWC WSC Across WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WSC WSC TZEI 161 x ENT 13 4854 2383 3795 57 69 57 73 0 4 182 150 2.8 2.8 1.2 1.0 2.5 3.1 6.5 5.01 TZEI 161 x TZEI 135 4982 2088 3742 57 69 57 74 1 5 175 153 3.0 3.7 1.1 0.9 2.6 3.7 6.3 1.57 TZEI 161 x TZEI 157 5625 2207 4160 55 68 56 71 1 4 176 156 2.6 3.7 1.2 1.2 2.3 3.6 7.2 2.97 TZEI 23 x ENT 13 4589 2211 3570 56 68 57 73 1 5 163 141 3.0 3.0 1.1 1.1 2.7 3.3 6.7 3.98 TZEI 23 x TZEI 135 4389 1741 3254 55 66 57 72 1 6 164 136 2.9 3.7 1.1 1.0 2.9 3.7 6.8 0 TZEI 23 x TZEI 157 4633 1467 3276 55 65 56 71 1 6 160 144 3.1 3.5 1.1 0.9 2.8 3.7 6.7 -0.64 ENT 13 x TZEI 135 5426 1590 3782 58 71 60 77 2 6 187 147 2.6 3.5 1.1 1.0 2.3 4.0 6.7 -0.6 ENT 13 x TZEI 157 5454 2560 4213 58 69 60 75 1 6 195 161 2.1 2.7 1.2 1.0 2.0 2.9 5.5 6.61 TZEI 135 x TZEI 157 5303 2004 3889 56 68 57 76 1 7 187 147 3.1 3.0 1.1 0.9 2.4 3.6 5.2 2.93 P3 Kollo Check 4690 1439 3297 54 64 55 70 1 6 172 156 3.0 3.8 1.3 0.9 2.9 4.1 6.8 -2.42 TZEI 9 x TZEI 16 Ck 4981 1835 3633 58 69 59 76 1 7 164 144 2.5 3.5 1.1 1.0 2.4 3.6 7.0 -0.19 TZEQI 82 x TZEQI 93 Ck 3792 2309 3156 63 74 65 78 2 4 175 147 2.9 2.3 1.0 1.0 3.3 3.5 5.2 6.28 TZEI 16 x TZEI 8 Ck 4423 2008 3388 58 69 59 76 1 7 156 135 2.8 3.2 1.2 1.0 2.8 3.4 7.3 1 TZEI 23 x TZEI 13 Ck 5063 1992 3747 58 72 59 76 1 4 168 134 2.9 3.3 1.3 1.1 2.7 3.5 6.5 2.96 Overall mean 4571 1684 3334 58 70 60 76 1 7 174 145 3.0 3.5 1.1 0.9 2.6 3.7 6.5 0.00 Lsd 865 775 605 1.5 2.4 1.7 3.9 1.3 3.2 12 15 0.5 0.9 0.1 0.3 0.4 0.7 1.1 5.4 Genotypes ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** * ** ** ** Environments ** ** ** ** ** ** ** ** ** ** ** ** * ** ** ** ns ** ** G x E ** * ** ** ** ** ns ** ** * ns ** ns ** ns ** * * ns WWC: well-watered conditions; WSC: water stress conditions; Across: across environments YIELD:Grain yield; DYA: days to anthesis; DYS: days to silk; ASI: anthesis silking interval; PLTH: plant height; PASP: plant aspect; EPP: number of ears per plant; EASP: ear aspect; GC2: stay-green 2; Index: selection index under water stress conditions *Significant at the 0.05 probability level; **Significant at the 0.01 probability level ; ns: not significant University of Ghana http://ugspace.ug.edu.gh 160 Appendix 4.2 Comparison of the efficiency of the SCA_PY and the HSGCA methods in classifying inbreds into heterotic groups based on their models Fan et al., (2009) reported the above equations: SCA = Cross mean (Xij) – Line mean (X.) – Tester mean (Xi.) + Overall mean (X..) GCAL = Line mean (X.j) – Overall mean (X..) HSGCA = Cross mean X – Tester mean (Xi.) = GCAL + SCA. What could we obtain when we develop these equations ? Eq1: SCA = Cross mean (Xij) – Line mean (X.) – Tester mean (Xi.) + Overall mean (X..) Eq2: GCAL = Line mean (X.j) – Overall mean (X..) Line mean (X.j) = GCAL +Overall mean (X..) Eq3: GCAT = Tester mean (X.i) – Overall mean (X..) Tester mean (X.j) = GCAT +Overall mean (X..) Eq4: HSGCA = Cross mean X – Tester mean (Xi.) Eq1, Eq2 and Eq3 SCA = Cross mean (Xij) – Line mean (X.) – Tester mean (Xi.) + Overall mean (X..) SCA = Cross mean (Xij) – (GCAL +Overall mean) – (GCAT +Overall mean.) + Overall mean SCA = Cross mean (Xij) – GCAL – Overall mean – GCAT – Overall mean + Overall mean Eq5: SCA = Cross mean (Xij) – GCAL– GCAT – Overall mean Eq3 and Eq4 HSGCA = Cross mean (Xij) – Tester mean (Xi.) HSGCA = Cross mean (Xij) – (GCAT +Overall mean) Eq6: HSGCA = Cross mean (Xij) – GCAT – Overall mean Let us see the relation between the SCA and the HSGCA methods: SCA = Cross mean (Xij) – GCAL– GCAT – Overall mean Eq5 and Eq6: HSGCA = Cross mean (Xij) – GCAT – Overall mean SCA identifies specific combinations where the hybrids performed better or worse than would be predicted based on GCA effects of both line and tester. HSGCA method identifies combinations where the hybrids performed higher or lower than would be expected based on the GCA effect of the tester involved in the testcross. The SCA method is likely to be more efficient than the HSGCA method. University of Ghana http://ugspace.ug.edu.gh 161 Appendix 5.1.Grain yield and other agronomic traits of 103 MSCs of varying maturity groups and 5 checks evaluated under drought stress and well- watered environments Entry Extra-early YIELD DYA ASI PLTH EHT PASP EPP SGREEN HPH Early Intermediate WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC (TZEEI 76 x TZEEI 82) x TZEI 182 4293 2654 66 67 1.5 5.0 169 151 85 101 3.0 2.0 1.0 6.0 36 397 (TZEEI 76 x TZEEI 82) x TZEEI 96 4249 859 66 70 2.0 7.5 178 136 100 89 2.5 4.0 1.1 7.0 34 61 (TZEEI 95 x TZEEI 79) x TZEEI 63 4068 1760 63 67 3.0 4.5 162 144 73 86 2.5 3.0 1.0 7.0 73 114 (TZEEI 76 x TZEEI 82) x TZEI 157 4034 1785 65 70 2.5 5.5 158 148 80 96 3.0 3.0 1.0 4.5 27 234 (TZEEI 78 x TZEEI 79) x TZEEI 67 4016 2345 66 69 1.5 4.0 161 153 90 99 3.0 2.5 1.0 7.5 12 14 (TZEEI 78 x TZEEI 79) x TZEI 23 3999 1297 69 73 0.5 3.0 163 121 98 80 3.0 3.0 1.0 6.0 12 -37 (TZEEI 95 x TZEEI 79) x TZEI 23 3983 1640 63 67 2.5 5.0 134 136 68 84 3.0 3.0 1.0 6.5 69 100 (TZEEI 95 x TZEEI 79) x TZEEI 67 3971 2096 64 68 1.5 5.0 158 133 85 84 2.5 2.5 1.0 5.0 69 155 (TZEEI 76 x TZEEI 82) x TZEI 23 3844 1912 64 67 1.5 4.5 138 129 70 80 3.0 3.0 1.0 6.5 21 258 (TZEEI 78 x TZEEI 79) x TZEI 157 3739 2307 68 69 1.5 4.0 145 166 83 109 3.0 2.5 1.0 6.0 5 12 (TZEEI 76 x TZEEI 82) x TZEEI 67 3718 1624 65 70 1.5 2.5 168 127 98 86 3.0 3.0 1.0 6.5 17 204 (TZEEI 76 x TZEEI 82) x TZEEI 81 3679 851 67 71 1.0 3.0 152 154 73 112 3.0 3.5 1.0 6.0 16 59 (TZEEI 95 x TZEEI 79) x TZEEI 64 3671 1122 65 70 0.5 5.0 156 130 83 81 3.0 3.0 1.1 7.0 56 37 (TZEEI 95 x TZEEI 79) x TZEI 157 3658 1543 67 71 0.5 4.0 161 134 100 89 3.0 3.0 0.9 5.5 55 88 (TZEEI 78 x TZEEI 79) x TZEEI 64 3621 1595 65 68 0.5 5.0 160 142 95 98 3.0 3.0 1.0 6.5 1 -22 Overall mean 3788 1429 67 71 1.9 5.7 158 142 83 90 2.9 3.2 1 6.8 32 78 Lsd 1538 918 2.2 4.2 2.3 5.5 20 23 20 17 1.3 1.2 0.2 1.7 Rep ns ns ns ** ** ns * ns ns ns ns ns * ** Block(Rep) ** ** * ** ns ns ** ** ** ** * ** ns ** Entry ** ** ** ** ** ** ** ** ** ** ** * * ** WWC: well-watered conditions; WSC: water stress conditions; HPH (%): heterosis for grain yield over the best RSC parent YIELD: Grain yield; DYA: days to anthesis; ASI: anthesis silking interval; PLTH (cm): plant height; EHT (cm): ear height; PASP (1-5): plant aspect; EPP: number of ears per plant; SGREEN: stay green characteristics at 90 and 100 days after planting on a scale of 1 to 9 ; *Significant at the 0.05 probability level; **Significant at the 0.01 probability level ; ns: not significant University of Ghana http://ugspace.ug.edu.gh 162 Appendix 5.1. Continued Entry Extra-early YIELD DYA ASI PLTH EHT PASP EPP SGREEN HPH Early Intermediate WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC (TZEEI 78 x TZEEI 79) x TZEEI 89 3560 740 63 68 4.0 4.5 150 121 70 75 3.0 3.5 1.0 7.0 0 -64 (TZEEI 78 x TZEEI 79) x TZEEI 81 3509 1390 69 75 1.5 4.0 161 156 78 105 3.0 3.5 1.0 6.0 -2 -32 (TZEEI 76 x TZEEI 82) x TZEEI 64 3495 664 66 70 -0.5 8.0 139 139 85 94 3.5 3.5 1.1 6.0 10 24 (TZEEI 78 x TZEEI 79) x TZEI 182 3458 1296 69 71 2.0 5.0 167 139 88 95 3.5 3.0 0.9 6.5 -3 -37 (TZEEI 95 x TZEEI 79) x TZEEI 89 3371 1145 62 68 2.5 4.0 128 124 63 73 3.0 3.5 1.0 6.0 43 39 (TZEEI 95 x TZEEI 79) x TZEI 182 3286 1628 67 70 1.5 5.0 145 134 70 86 3.5 3.0 1.0 5.5 40 98 (TZEEI 95 x TZEEI 79) x TZEEI 81 3147 1131 65 70 3.0 3.0 145 134 65 86 4.0 3.0 1.0 6.0 34 38 (TZEEI 76 x TZEEI 82) x TZEEI 89 3058 419 65 69 2.0 6.5 133 122 65 80 4.5 4.5 1.0 6.0 -3 -21 (TZEEI 78 x TZEEI 79) x TZEEI 63 3041 749 65 70 2.5 6.0 153 138 75 90 4.0 3.5 1.0 7.0 -15 -64 (TZEEI 76 x TZEEI 82) x TZEEI 58 2918 903 66 69 2.0 4.5 154 144 75 89 3.5 3.0 1.1 6.5 -8 69 (TZEEI 95 x TZEEI 79) x TZEEI 58 2677 1453 62 70 4.0 3.0 150 138 65 87 4.0 3.0 0.9 6.5 14 77 (TZEEI 95 x TZEEI 79) x TZEEI 87 2623 1034 65 69 2.0 5.0 153 138 73 90 4.0 3.5 1.0 6.5 11 26 (TZEEI 78 x TZEEI 79) x TZEEI 58 2444 616 64 71 3.5 6.0 141 131 75 82 4.0 4.5 0.9 7.0 -32 -70 (TZEEI 76 x TZEEI 82) x TZEEI 87 2150 1785 66 68 3.0 5.0 149 147 83 98 4.5 3.0 1.0 6.5 -32 234 (TZEEI 78 x TZEEI 79) x TZEEI 87 2094 396 68 71 3.0 7.0 163 125 88 78 4.5 4.0 1.2 7.5 -41 -81 (TZEEI 76 x TZEEI 82) x TZEEI 63 1984 1002 66 69 1.0 5.5 139 136 80 88 4.5 3.5 1.5 5.5 -37 88 Overall mean 3788 1429 67 71 1.9 5.7 158 142 83 90 2.9 3.2 1 6.8 32 78 Lsd 1538 918 2.2 4.2 2.3 5.5 20 23 20 17 1.3 1.2 0.2 1.7 Rep ns ns ns ** ** ns * ns ns ns ns ns * ** Block(Rep) ** ** * ** ns ns ** ** ** ** * ** ns ** Entry ** ** ** ** ** ** ** ** ** ** ** * * ** WWC: well-watered conditions; WSC: water stress conditions; HPH (%): heterosis for grain yield over the best RSC parent YIELD: Grain yield; DYA: days to anthesis; ASI: anthesis silking interval; PLTH (cm): plant height; EHT (cm): ear height; PASP (1-5): plant aspect; EPP: number of ears per plant; SGREEN: stay green characteristics at 90 and 100 days after planting on a scale of 1 to 9 ; *Significant at the 0.05 probability level; **Significant at the 0.01 probability level ; ns: not significant University of Ghana http://ugspace.ug.edu.gh 163 Appendix 5.1. Continued Entry Extra-early YIELD DYA ASI PLTH EHT PASP EPP SGREEN HPH Early Intermediate WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC (TZEEI 95 x TZEEI 79) x TZEEI 96 1835 744 68 72 2.0 2.5 144 126 80 87 4.5 3.5 0.7 7.0 -22 -9 (TZEEI 78 x TZEEI 79) x TZEEI 96 1765 377 71 73 2.5 9.5 138 138 85 92 5.0 4.5 0.9 8.0 -51 -82 TZEEI 78 x TZEEI 79 Check 3571 1053 66 69 1.5 2.5 167 159 85 95 3.0 2.0 1.3 7.0 95 95 TZEEI 76 x TZEEI 82 Check 3166 534 65 69 1.0 4.0 132 134 75 88 3.5 4.0 1.1 5.5 95 95 TZEEI 95 x TZEEI 79 Check 2353 821 69 71 4.5 6.5 153 136 80 91 3.5 4.0 0.8 8.0 (TZEEI 76 x TZEEI 82) x TZEI 124 5399 1834 66 68 2.0 6.0 173 164 90 98 2.0 2.5 1.0 6.5 71 243 (TZEEI 78 x TZEEI 79) x ENT 13 5328 2376 70 71 0.5 3.0 201 182 130 123 1.5 3.0 1.0 6.0 49 16 (TZEEI 95 x TZEEI 79) x CMZ ENT 5 5293 1016 66 69 1.0 7.0 177 145 90 96 2.0 3.5 1.0 7.5 125 24 (TZEEI 76 x TZEEI 82) x TZd 164 5272 1440 68 72 -0.5 2.0 171 139 93 92 2.0 3.0 1.2 7.0 67 170 (TZEEI 76 x TZEEI 82) x TZEI 129 5202 2044 67 70 1.5 6.0 182 153 110 96 2.0 2.5 1.0 6.0 64 283 (TZEEI 76 x TZEEI 82) x ENT 13 5084 2486 68 69 2.0 4.5 163 151 83 104 1.5 2.5 1.1 9.0 61 365 (TZEEI 76 x TZEEI 82) x CMK ENT 20 4928 533 68 72 0.5 11.0 186 114 98 72 2.0 4.5 1.0 7.0 56 0 (TZEEI 78 x TZEEI 79) x TZEI 124 4855 984 69 69 2.5 9.5 173 149 80 93 2.0 3.5 1.0 8.5 36 -52 (TZEEI 95 x TZEEI 79) x CMK ENT 20 4616 499 68 75 2.0 11.0 160 133 80 79 2.5 4.0 1.0 7.5 96 -39 (TZEEI 76 x TZEEI 82) x CMZ ENT 5 4499 681 68 73 1.5 11.0 165 120 93 77 2.0 3.5 1.0 5.5 42 27 (TZEEI 95 x TZEEI 79) x TZEI 158 4474 1616 66 72 1.5 3.0 164 134 88 83 3.0 3.0 1.0 6.5 90 97 Overall mean 3788 1429 67 71 1.9 5.7 158 142 83 90 2.9 3.2 1 6.8 32 78 Lsd 1538 918 2.2 4.2 2.3 5.5 20 23 20 17 1.3 1.2 0.2 1.7 Rep ns ns ns ** ** ns * ns ns ns ns ns * ** Block(Rep) ** ** * ** ns ns ** ** ** ** * ** ns ** Entry ** ** ** ** ** ** ** ** ** ** ** * * ** WWC: well-watered conditions; WSC: water stress conditions; HPH (%): heterosis for grain yield over the best RSC parent YIELD: Grain yield; DYA: days to anthesis; ASI: anthesis silking interval; PLTH (cm): plant height; EHT (cm): ear height; PASP (1-5): plant aspect; EPP: number of ears per plant; SGREEN: stay green characteristics at 90 and 100 days after planting on a scale of 1 to 9 ; *Significant at the 0.05 probability level; **Significant at the 0.01 probability level ; ns: not significant University of Ghana http://ugspace.ug.edu.gh 164 Appendix 5.1. Continued Entry Extra-early YIELD DYA ASI PLTH EHT PASP EPP SGREEN HPH Early Intermediate WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC (TZEEI 76 x TZEEI 82) x TZEI 158 4467 1268 64 69 3.0 6.0 158 145 65 90 2.5 3.0 1.0 6.5 41 137 (TZEEI 76 x TZEEI 82) x CMZ ENT 2 4421 1336 68 72 -1.0 2.5 177 139 98 91 2.5 3.5 1.0 7.0 40 150 (TZEEI 95 x TZEEI 79) x CMZ ENT 2 4232 987 68 69 0.5 8.0 155 139 73 91 2.5 3.5 0.9 7.0 80 20 (TZEEI 76 x TZEEI 82) x TZd 181 4214 1092 69 70 1.5 6.0 152 139 65 91 3.0 3.5 1.1 7.0 33 104 (TZEEI 76 x TZEEI 82) x TZEI 13 4159 1968 67 71 1.5 7.5 147 139 78 98 3.0 3.0 1.0 6.5 31 269 (TZEEI 76 x TZEEI 82) x TZEI 16 4004 1058 68 72 4.0 4.0 168 134 105 88 2.5 3.5 1.2 7.5 26 98 (TZEEI 78 x TZEEI 79) x CMZ ENT 5 3966 823 68 70 2.0 13.5 163 155 88 103 3.0 3.5 1.0 6.0 11 -60 (TZEEI 76 x TZEEI 82) x CamInb 3900 767 69 78 2.5 8.0 160 145 93 95 2.5 4.0 1.0 6.5 23 44 (TZEEI 95 x TZEEI 79) x ENT 13 3863 1336 67 70 2.0 10.5 153 129 75 86 3.0 3.0 1.0 7.0 64 63 (TZEEI 78 x TZEEI 79) x CMZ ENT 2 3823 1538 68 72 1.0 4.0 171 158 90 106 2.5 3.0 1.0 6.5 7 -25 (TZEEI 78 x TZEEI 79) x TZd 181 3761 1431 67 72 4.0 4.0 163 138 80 76 3.0 3.5 1.0 7.5 5 -30 (TZEEI 78 x TZEEI 79) x TZEI 129 3735 1255 70 72 2.5 10.0 173 141 98 88 3.0 3.5 1.1 8.0 5 -39 (TZEEI 78 x TZEEI 79) x TZd 164 3655 1416 68 72 2.5 4.0 163 151 83 101 3.0 3.0 1.0 8.0 2 -31 (TZEEI 78 x TZEEI 79) x TZEI 16 3574 989 71 74 2.0 9.5 158 151 78 103 3.0 3.5 1.0 7.0 0 -52 (TZEEI 78 x TZEEI 79) x CMK ENT 20 3455 836 70 77 1.5 8.0 157 124 70 81 3.0 3.5 1.0 7.5 -3 -59 (TZEEI 95 x TZEEI 79) x TZEI 124 3438 1441 62 69 5.0 10.0 143 145 60 81 3.5 3.0 1.0 8.0 46 76 Overall mean 3788 1429 67 71 1.9 5.7 158 142 83 90 2.9 3.2 1 6.8 32 78 Lsd 1538 918 2.2 4.2 2.3 5.5 20 23 20 17 1.3 1.2 0.2 1.7 Rep ns ns ns ** ** ns * ns ns ns ns ns * ** Block(Rep) ** ** * ** ns ns ** ** ** ** * ** ns ** Entry ** ** ** ** ** ** ** ** ** ** ** * * ** WWC: well-watered conditions; WSC: water stress conditions; HPH (%): heterosis for grain yield over the best RSC parent; YIELD: Grain yield; DYA: days to anthesis; ASI: anthesis silking interval; PLTH (cm): plant height; EHT (cm): ear height; PASP (1-5): plant aspect; EPP: number of ears per plant; SGREEN: stay green characteristics at 90 and 100 days after planting on a scale of 1 to 9 ; *Significant at the 0.05 probability level; **Significant at the 0.01 probability level ; ns: not significant University of Ghana http://ugspace.ug.edu.gh 165 Appendix 5.1. Continued Entry Extra-early YIELD DYA ASI PLTH EHT PASP EPP SGREEN HPH Early Intermediate WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC (TZEEI 95 x TZEEI 79) x TZEI 16 3422 726 68 73 3.0 6.0 161 123 98 80 3.0 4.0 1.0 7.5 45 -12 (TZEEI 78 x TZEEI 79) x TZEI 158 3418 993 68 71 2.5 7.0 168 137 93 86 3.0 3.5 1.1 7.0 -4 -52 (TZEEI 95 x TZEEI 79) x TZd 181 3404 756 66 70 2.5 5.5 152 116 65 67 3.5 3.5 0.9 8.5 45 -8 (TZEEI 95 x TZEEI 79) x CamInb 3290 1275 71 85 1.5 1.5 164 131 88 76 4.0 3.0 1.0 6.5 40 55 (TZEEI 95 x TZEEI 79) x TZEI 129 3179 1647 67 71 2.5 6.0 165 139 100 88 3.5 3.0 0.9 7.0 35 101 (TZEEI 78 x TZEEI 79) x CamInb 3146 578 72 80 3.5 6.0 169 138 113 87 3.5 4.0 0.9 7.0 -12 -72 (TZEEI 95 x TZEEI 79) x TZd 164 2989 1814 65 71 2.5 3.0 138 126 68 74 3.0 3.0 0.9 8.0 27 121 ENT 13 x CamInbgp.1-17 Check 4081 1770 70 70 1.5 11.0 178 153 108 92 3.0 2.5 1.0 6.0 98 98 TZEI 124 x ENT 13 check 4524 1498 75 81 2.5 4.5 177 154 100 103 2.5 3.0 1.0 5.5 106 106 (TZEEI 76 x TZEEI 82) x CMZ ENT 12 5182 1694 64 70 2.0 3.0 143 129 70 77 2.5 3.0 1.0 8.0 64 217 (TZEEI 76 x TZEEI 82) x CMK ENT 8 4841 2124 65 69 0.5 4.5 150 135 83 84 2.5 2.5 1.0 7.5 53 298 (TZEEI 76 x TZEEI 82) x CMZ ENT 6 4779 2121 68 70 2.0 6.5 167 150 95 93 2.0 3.0 1.0 7.0 51 297 (TZEEI 78 x TZEEI 79) x CMK ENT 8 4662 2754 67 70 0.0 2.0 159 176 88 104 3.0 2.5 1.0 7.0 31 34 (TZEEI 76 x TZEEI 82) x CMZ ENT 3 4654 1055 66 69 2.0 7.5 147 137 78 84 2.0 3.5 1.0 7.0 47 98 (TZEEI 78 x TZEEI 79) x CMZ ENT 3 4580 1586 71 70 0.0 4.0 159 173 80 93 2.5 3.5 1.0 8.0 28 -23 (TZEEI 76 x TZEEI 82) x 88069 4576 2627 71 74 1.0 4.0 163 146 88 94 2.0 2.0 1.0 7.5 45 392 Overall mean 3788 1429 67 71 1.9 5.7 158 142 83 90 2.9 3.2 1 6.8 32 78 Lsd 1538 918 2.2 4.2 2.3 5.5 20 23 20 17 1.3 1.2 0.2 1.7 Rep ns ns ns ** ** ns * ns ns ns ns ns * ** Block(Rep) ** ** * ** ns ns ** ** ** ** * ** ns ** Entry ** ** ** ** ** ** ** ** ** ** ** * * ** WWC: well-watered conditions; WSC: water stress conditions; HPH (%): heterosis for grain yield over the best RSC parent; YIELD: Grain yield; DYA: days to anthesis; ASI: anthesis silking interval; PLTH (cm): plant height; EHT (cm): ear height; PASP (1-5): plant aspect; EPP: number of ears per plant; SGREEN: stay green characteristics at 90 and 100 days after planting on a scale of 1 to 9 ; *Significant at the 0.05 probability level; **Significant at the 0.01 probability level ; ns: not significant University of Ghana http://ugspace.ug.edu.gh 166 Appendix 5.1. Continued Entry Extra-early YIELD DYA ASI PLTH EHT PASP EPP SGREEN HPH Early Intermediate WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC (TZEEI 95 x TZEEI 79) x CMZ ENT 3 4568 601 65 70 2.0 6.5 153 122 70 73 2.5 4.5 1.0 6.5 94 -27 (TZEEI 95 x TZEEI 79) x CMK ENT 23 4464 1370 69 75 0.5 2.5 166 142 80 86 2.5 3.5 1.0 7.0 90 67 (TZEEI 76 x TZEEI 82) x CMK ENT 24 4360 930 69 77 1.0 3.0 167 134 90 80 2.0 3.5 1.0 4.5 38 74 (TZEEI 76 x TZEEI 82) x CMZ ENT 8 4350 2959 68 70 0.5 3.0 158 161 78 100 3.0 2.0 0.9 6.5 37 454 (TZEEI 78 x TZEEI 79) x CMZ ENT 8 4326 2451 67 73 3.0 7.0 148 173 65 103 2.5 2.5 1.0 6.5 21 19 (TZEEI 78 x TZEEI 79) x CMZ ENT 12 4325 1316 66 70 2.0 10.0 173 144 95 87 2.5 3.0 1.1 7.0 21 -36 (TZEEI 76 x TZEEI 82) x CMK ENT 22 4151 1967 68 71 1.0 3.5 158 156 88 94 2.5 3.0 1.1 6.0 31 268 (TZEEI 95 x TZEEI 79) x 9450 4098 1288 71 79 2.0 5.0 153 131 88 87 2.5 3.5 1.0 7.0 74 57 (TZEEI 76 x TZEEI 82) x CMK ENT 23 4092 3383 71 71 -0.5 2.0 163 169 83 109 3.0 2.5 1.1 6.5 29 533 (TZEEI 76 x TZEEI 82) x 9450 4090 1579 70 75 1.0 8.0 153 149 83 104 3.0 3.0 1.0 6.5 29 196 (TZEEI 95 x TZEEI 79) x 9848 4090 2084 65 70 2.0 4.0 152 134 83 86 3.0 3.0 1.0 7.5 74 154 (TZEEI 95 x TZEEI 79) x TZd 177 4064 1493 67 71 4.5 6.5 158 150 73 86 3.0 3.0 0.9 7.0 73 82 (TZEEI 95 x TZEEI 79) x 88069 3971 1448 68 71 1.5 5.0 164 156 93 101 3.0 3.0 1.0 6.5 69 76 (TZEEI 95 x TZEEI 79) x TZd 176 3947 1398 67 72 2.5 3.0 155 135 78 80 3.0 3.5 1.1 8.5 68 70 (TZEEI 78 x TZEEI 79) x CMK ENT 24 3938 726 72 76 1.5 9.0 162 149 88 96 2.5 4.5 1.0 5.0 10 -65 (TZEEI 95 x TZEEI 79) x CMZ ENT 6 3913 2204 66 70 3.5 5.5 168 173 93 99 2.5 2.0 1.0 6.5 66 169 Overall mean 3788 1429 67 71 1.9 5.7 158 142 83 90 2.9 3.2 1 6.8 32 78 Lsd 1538 918 2.2 4.2 2.3 5.5 20 23 20 17 1.3 1.2 0.2 1.7 Rep ns ns ns ** ** ns * ns ns ns ns ns * ** Block(Rep) ** ** * ** ns ns ** ** ** ** * ** ns ** Entry ** ** ** ** ** ** ** ** ** ** ** * * ** WWC: well-watered conditions; WSC: water stress conditions; HPH (%): heterosis for grain yield over the best RSC parent; YIELD: Grain yield; DYA: days to anthesis; ASI: anthesis silking interval; PLTH (cm): plant height; EHT (cm): ear height; PASP (1-5): plant aspect; EPP: number of ears per plant; SGREEN: stay green characteristics at 90 and 100 days after planting on a scale of 1 to 9 ; *Significant at the 0.05 probability level; **Significant at the 0.01 probability level; ns: not significant University of Ghana http://ugspace.ug.edu.gh 167 Appendix 5.1. Continued Entry Extra-early YIELD DYA ASI PLTH EHT PASP EPP SGREEN HPH Early Intermediate WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC WWC WSC (TZEEI 95 x TZEEI 79) x CMK ENT 22 3714 2152 68 73 1.0 3.5 157 149 63 83 3.0 3.0 1.0 7.5 58 162 (TZEEI 76 x TZEEI 82) x TZd 177 3698 1664 67 71 3.0 5.5 152 148 70 88 3.0 3.5 0.9 8.0 17 212 (TZEEI 95 x TZEEI 79) x CMZ ENT 12 3639 993 64 67 2.5 8.0 147 146 58 83 3.5 4.0 1.0 7.5 55 21 (TZEEI 95 x TZEEI 79) x CMZ ENT 8 3624 1357 71 68 1.0 10.0 153 140 75 78 2.5 3.5 1.1 7.5 54 65 (TZEEI 78 x TZEEI 79) x CMZ ENT 6 3522 1297 69 72 4.0 15.0 171 158 90 99 2.5 3.0 0.8 6.5 -1 -37 (TZEEI 78 x TZEEI 79) x 9450 3477 1297 74 84 1.5 4.0 178 148 108 103 3.0 3.0 1.0 6.0 -3 -37 (TZEEI 76 x TZEEI 82) x TZd 176 3434 1604 67 71 3.5 3.5 157 141 80 84 3.0 3.5 1.1 8.5 8 200 (TZEEI 78 x TZEEI 79) x 88069 3313 1606 71 75 0.5 1.5 158 177 80 118 3.0 3.0 1.0 6.0 -7 -22 (TZEEI 78 x TZEEI 79) x CMK ENT 23 3109 2778 70 72 1.5 8.5 154 168 83 103 3.5 2.5 1.0 7.0 -13 35 (TZEEI 95 x TZEEI 79) x CMK ENT 8 3108 2937 68 72 0.5 1.0 135 141 78 86 3.5 2.5 1.0 8.0 32 258 (TZEEI 78 x TZEEI 79) x TZd 177 2918 779 69 76 2.0 6.5 149 128 75 79 3.5 3.5 1.1 7.5 -18 -62 (TZEEI 78 x TZEEI 79) x TZd 176 2255 1527 69 75 3.5 3.0 153 138 75 86 4.5 3.0 1.1 8.0 -37 -26 (TZEEI 78 x TZEEI 79) x CMK ENT 22 2161 827 72 74 3.5 14.5 160 153 75 88 4.0 3.5 1.1 6.0 -39 -60 Overall mean 3788 1429 67 71 1.9 5.7 158 142 83 90 2.9 3.2 1 6.8 32 78 Lsd 1538 918 2.2 4.2 2.3 5.5 20 23 20 17 1.3 1.2 0.2 1.7 Rep ns ns ns ** ** ns * ns ns ns ns ns * ** Block(Rep) ** ** * ** ns ns ** ** ** ** * ** ns ** Entry ** ** ** ** ** ** ** ** ** ** ** * * ** WWC: well-watered conditions; WSC: water stress conditions; HPH (%): heterosis for grain yield over the best RSC parent; YIELD: Grain yield; DYA: days to anthesis; ASI: anthesis silking interval; PLTH (cm): plant height; EHT (cm): ear height; PASP (1-5): plant aspect; EPP: number of ears per plant; SGREEN: stay green characteristics at 90 and 100 days after planting on a scale of 1 to 9 ; *Significant at the 0.05 probability level; **Significant at the 0.01 probability level ; ns: not significant University of Ghana http://ugspace.ug.edu.gh 168 Appendix 6.1. Pictures of the single cross hybrid TZEI 124 x ENT 13 Single cross hybrid (TZEI 124 x ENT 13) coined NAGODE (Thank You) by farmers in Niger. Results of the evaluation of the hybrid at (A) Konni in 2012 ; (B) Maradi in 2012 ; (C&D) FESA Seed Co at Chadakwari in 2013 A B C D University of Ghana http://ugspace.ug.edu.gh