Sunflower (Helianthus annuus L.) is one of the most important oil crops in South Africa and genetic improvement for grain yield and oil-content was initiated in the country in the early 1970s. Commercial production of sunflower in South Africa is done under natural rainfall conditions in areas where frequencies of drought are high hence the requirement for drought tolerant cultivars. An assessment of the genetic gains in seed and oil yield achieved since 1970, the effects of re-cycling inbred lines and strategies for developing drought tolerant sunflower cultivars has not been done for South African sunflower breeding programmes. Two data-sets were used for the genetic gain studies: side-by-side evaluation of historical and current sets of popular cultivars in the same environment under one set of trial management practices; and yield trends in commercial farmers’ fields based on annual production estimates. The estimated relative genetic gain for seed yield based on side-by-side trials was 1.5% year−1 and the relative gain in seed yield per year under commercial production was 1.9% year−1. The contribution of new cultivars to total seed yield progress in sunflower were 56.3% for the period 1970 to 1989; 23.9% from 1990 to 2009 and the mean over the four decades under consideration from 1970 to 2009 was 41.6%. Quantifying the usefulness of inbred lines in advanced cycle plant breeding was done using four base breeding populations based on: phenotypic or genetic variability; heterosis; and combining ability. Significant genetic variation was evident for seed yield and oil yield while genetic variability for oil content was low. Genetic advance (GA%), with a 10% selection intensity, was high for seed yield and oil yield for each of the four populations ranging from 36-42% and 38-43%, respectively. The GA% for oil content was low ranging from 1.3% to 5.1% indicating the need to introgress high oil content germplasm in the present breeding populations in advanced cycle pedigree breeding. Founder parent heterosis (FPH), mid-standard heterosis (MSH) and high standard heterosis (HSH) indicated that some new testcross hybrids from the advanced cycle pedigree breeding populations were performing better than their founder parents in hybrid combination as well as the standard commercial hybrid checks. From variance component analysis, general combining ability (GCA) was predominant over specific combining ability (SCA) for seed and oil yields indicating that superior hybrids can be identified based on positive and significant GCA effects of the female lines. For oil content, SCA was predominant over GCA indicating that it would be best to select for specific hybrids combinations with high oil content rather than selecting female lines with high GCA effects. Variable moisture conditions characterise the sunflower production environments in South Africa. Breeding for such environmental conditions requires a combination of strategies including use of secondary traits and developing appropriate test environments. Three secondary traits, head diameter, stem diameter and stay green canopy which are easy to measure in the field were evaluated for their appropriateness for selecting for drought tolerance under three moisture conditions: random stress environments (RSE), managed drought stress environments (MSE) and well watered nonstress environments (NSE). Type A genetic correlations indicated that stay green canopy (SG) had the potential to be used as secondary trait to indirectly improve oil yield under the three moisture conditions. The indirect selection efficiency (ISE) for SG using genetic correlations based on H2 were 0.79, 0.82 and 0.78 in the RSE, MSE and NSE, respectively, while that using genetic correlation based on h2 were 0.67, 0.98, and 0.93 in the RSE, MSE and NSE, respectively. In both cases selection in the MSE had the highest efficiency using genetic correlations based on either H2 or h2. Estimates of indirect selection based on type B genetic correlations indicated that indirect selection for oil yield (OY) in the MSE and NSE for the target RSE was as effective as direct selection of OY in the RSE based on additive genetic correlations of 0.96 obtained in both selection environments. Overall, the results from the exploratory drought tolerance study should inform the development of breeding strategies to improve drought tolerance and associated yield stability of sunflower cultivars grown in South Africa and associated environments. / Thesis (Ph.D.)-University of KwaZulu-Natal, Pietermaritzburg, 2013.

Maize is the principal crop for food security and livestock feed in South Africa. It is also an industrial crop and the produce is exported to many countries in the world. Therefore there is high seed demand which prompts competition for breeding productive hybrids. However direct introduction of tropical hybrids into the warm temperate South African environments has not been successful. Competitive advantages can be obtained by implementing a “shuttle breeding” programme whereby part of the breeding is done in Zimbabwe and South Africa to minimise research and production costs. Introgression of temperate germplasm in tropical elite inbred lines can also be pursued to obtain adapted hybrids. The aim of this study was therefore to assess the effectiveness of introgression of temperate germplasm into tropical elite maize inbred lines as a strategy to enhance adaptability of new hybrids to South Africa, and also to determine both breeding and economic value of a “shuttle breeding” programme. To this end, the introgressed inbred lines and their hybrid progenies were evaluated in South Africa to determine the effect of the selection environment on their performance and genetic variation. Both genetic and economic gains were evaluated with a view to make recommendations to the small and medium scale enterprises with interests in the market. Introgression of temperate germplasm into tropical germplasm elite lines did not disrupt the heterotic groupings because most of the introgressed lines (86%) fitted into known existing heterotic groups. Only 14% of the introgressed lines did not show any inclination to towards the known heterotic clusters of their founder tropical parents. These lines were considered to be new recombinant inbred lines that showed little resemblance with their founder parents. Selection environment did not influence heterotic clustering of the introgressed lines, and genetic diversity was identified among introgressed lines developed in the same environment. Genetic variation was observed for the major economic traits and heritability of 21% to 91%. The introgression was effective for improving grain yield potential and ear prolificacy. Spearman’s rank correlation analysis on grain yield and ear prolificacy data showed significant positive correlation between selection environments such as Ukulinga in South Africa and Kadoma Research Centre in Zimbabwe. Therefore Kadoma Research Centre will be recommended for use in breeding new maize germplasm lines for South Africa. Correlation among traits showed that ear prolificacy and plant height had significant (P<0.05) direct effects on grain yield thus direct selection of these traits will be emphasised in breeding new hybrids. Introgression of temperate germplasm into tropical elite maize inbred lines was effective for improving their adaptation to warm temperate environments. Positive genetic gains of 5-58% were realised for grain yield potential and 26-46% for ear prolificacy. Whereas 1% to 37% gains were realised for secondary traits such as plant and ear height, anthesis and silking days there was barely any improvement for root and stalk lodging, and grain moisture content at harvest. However, introgressed lines displayed impressive performance per se and inter se indicating potential for commercial production. The new inbred line 71-DMLF7_88 combined early physiological maturity, high ear prolificacy and grain yield potential qualifying it as a perfect parent for the warm temperate environments. At least six hybrids were stable and adaptable while four were considered to be ideal genotypes relative to standard commercial hybrids such as PAN6Q445B which is a market leader. The exceptional hybrids, 12C20264, 12C22766, 13XH349 and 11C11774 will be advanced in South Africa. The study also indicated significant economic gains when a shuttle programme is implemented to breed new hybrids following the introgression strategy. The “Shuttle breeding” programme attained a positive net present value (NPV) of $1, 834, 166. 00. This indicated an increase in shareholder value through an opportunity cost of 17% and 3% relative to conventional breeding programmes which are based in South Africa and Zimbabwe, respectively. Positive NPV and genetic gain achieved using the “shuttle breeding” programme makes it a viable option for small and medium scale seed companies with intention to breed and commercialise competitive products in South African. In general, the study revealed that introgression of temperate germplasm into tropical elite inbred lines using a “shuttle breeding” programme was effective for enhancing adaptability of tropical germplasm to the South African warm temperate environments. / Thesis (Ph.D.)-University of KwaZulu-Natal, Pietermaritzburg, 2013.

Maize (Zea mays L.) is one of the most important food crops in sub-Saharan Africa (SSA); however its production is constrained by many factors. Grain yield is compromised by poor genetic performance and poor agronomic management. This calls for need to develop hybrids and exploiting heterosis of single crosses which are adapted to challenging environments. Currently, there is no popcorn hybrids developed in South Africa which is adapted to local conditions. As such, there is need to develop hybrids that cater for smallscale farmers in marginal environments. The objectives of the study were to determine i) standard heterosis, levels of variation and heritability for phenotypic traits in dent and flint maize hybrids; ii) the association between genetic distances and phenotypic traits in dent and flint maize hybrids; iii) mid-parent heterosis in popcorn hybrids, iv) the effect of secondary traits on grain yield in dent, flint and popcorn hybrids; v) genetic diversity and the relationship between traits in widely grown selected hybrids in Southern Africa; and vii) to compare effectiveness of phenotypic analysis models for determining genetic distances between hybrids. Popcorn, dent and flint hybrids were evaluated at two sites. The data was analysed using SAS, Genstat and Power marker statistical packages. The results revealed that the relationship between genetic distance and heterosis is dependent on the environment. Hybrids in top 10 at both sites were different indicating that there was a significant genotype x environment interaction. 13 new heterotic patterns that performed better than the controls can be utilized in heterosis breeding; however there is need to test them in different environments to check on their stability. Grain texture cannot be used to discriminate hybrids for yield because all patterns of dent x dent, dent x flint and flint x flint were present in the top 10 hybrids. Lines DXL124 and DXL158 dominated parentage of the top 10 hybrid rank for yield qualifying them as potential testers for specific combining ability in future studies. Heterosis in popcorn hybrids that performed better than the mid-parent can be utilized in heterosis breeding to exploit vigour, though there is need to test the hybrids in a number of different environments. The main direct factors contributing to yield were ear prolificacy, ear aspect, number of plants and shelling percentages qualifying them to be selected to boost grain yield. Phenotypic data and 91 SNP markers were used to estimate the genetic distance between the hybrids. The results indicated that hybrids that were in the same cluster belong to the same brand and were related in origin and pedigree. Both molecular and phenotypic data were effective in discriminating the hybrids into different clusters according to genetic background. SNP markers revealed nine clusters of hybrids, 12-trait model revealed eight clusters and five-trait model revealed six clusters at 85% genetic distance. The study indicates strategies that can be adopted to boost grain yield in dent, flint and popcorn hybrids. / Thesis (M.Sc.)-University of KwaZulu-Natal, Pietermaritzburg, 2013.

Maize (Zea mays L.) is the third most important cereal crops in the world after wheat and rice. In Ethiopia, maize remains the second largest food security crop after tef [Eragrostis tef (Zucc.) Trotter.]. The mid-altitude, sub-humid agro-ecology (1000 to 1800 m above sea level) is the most important maize producing environment in Ethiopia. However, productivity of maize is low, due to several biotic and abiotic constraints. Among the biotic constraints, Turcicum leaf blight disease of maize caused by Exserohilum turcicum Pass Leonard & Suggs shows high incidence of 95-100% and inflicts significant grain losses in the country. Therefore, high yielding, Turcicum leaf blight resistant and farmers-preferred maize varieties and their production technologies should be developed and made available to growers to enhance maize production and to achieve food security. The objectives of this study were to: (1) assess farmer’s preferences, and production constraints for maize in the mid-altitude, sub-humid agro-ecology of western Ethiopia, (2) determine the genetic variability among elite maize inbred lines and select promising parents for resistance to E. turcicum, (3) determine diversity among the elite germplasm lines using SSR markers, (4) determine combining ability and heterosis among elite maize inbred lines and their hybrids, and (5) investigate genotype x environment interaction and yield stability of experimental maize hybrids developed for the midaltitude sub-humid agro-ecology of Ethiopia. A participatory rural appraisal (PRA) research was conducted involving 240 maize farmers in three representative maize growing zones of western Ethiopia; West Shoa, East Wollega and West Wollega, each represented by two districts and two subdistricts. Maize was ranked number one both as food and cash crop by 82.9% of respondents. Turcicum leaf blight was ranked as number one devastating leaf disease by 46% of respondents. Breeding for improved disease resistance and grain yield, enhancing the availability of crop input and stabilizing market price during harvest time were recommended as the most important strategies to increase maize production by small-scale farmers in western Ethiopia. Fifty inbred lines were evaluated for reaction to Turcicum leaf blight during the main cropping seasons of 2011 and 2012. Inbred lines were clustered into resistant (CML202, 144-7b, 136-a, 139-5j, 30H83-7-1, ILOO’E-1-9, SZYNA-99-F2, and 142-1-e), and susceptible (CML197, CML464, A7033 , Kuleni C1-101-1-1, CML443, SC22-430 (63), (DRB-F2-60-1-2) – B-1-B-B-B, Pool9A-4-4-1-1-1). Inbred lines (CML312, CML445, Gibe-1-158-1-1-1-1, CML395, and 124-b (113)) had intermediate response to the disease. Overall, inbred lines such as CML202, 30H83-7-1, ILOO’E-1-9-1, CML312, CML395 CML445 and 142-1-e were selected with better agronomic performance and resistance to leaf blight for breeding. Twenty selected elite parental inbred lines were genotyped with 20 polymorphic SSR markers. The genotypes used were clustered into five groups consistent with the known pedigrees. The greatest genetic distance was identified between the clusters of lines CML-202 and Gibe-1-91. Eighteen selected inbred lines were crossed using the factorial mating scheme and 81 hybrids developed to determine combining ability effects and heterosis. Inbred lines with high GCA effect (CML 202, CML395, 124-b (113), ILOO’E-1-9 and CML 197) were selected as best combiners for hybrid development. Additionally five high yielding novel single cross hybrids with grain yield of > 8 t ha-1 and high SCA effects were identified such as CML395 X CML442, DE-78-Z-126-3-2-2-1-1 X CML442, ILOO’E-1-9-1-1-1-1-1 X CML312, X1264DW-1-2-2-2-2 X CML464 and SC22 X Gibe-1-91-1-1-1-1. These experimental hybrids are recommended for direct production or as hybrid testers for hybrid development. Genotype x environment interaction (GEI) effects of 81 newly developed and three check maize hybrids were evaluated across 10 locations in the mid-altitude sub-humid agro-ecologies of Ethiopia. The AMMI-3 and GGE biplot models were used to determine stability. Hybrids such as G68, G39, G37, G77, G34 and G2 were identified as the most stable and high yielding at favorable environments such as Bako, Jima, Arsi Negelle and Pawe in Ethiopia. The genotype and genotype by environment interaction (GGE) biplot clustered the 10 environments into three unique mega-environments. Environment I included Bako, Jima, Asossa, Ambo, Finote Selam, Haramaya and Pawe while environment II represented by Arsi-Negelle and environment III Areka and Hawassa. In general, the study identified valuable maize inbred lines with high combining ability for breeding and novel single cross hybrids for large-scale production or as testers for hybrid development at the mid-altitude, sub-humid agro-ecologies of Ethiopia or similar environments in sub-Saharan Africa. / Thesis (Ph.D.)-University of KwaZulu-Natal, Pietermaritzburg, 2013.

Sweetpotato (Ipomoea batatas (L.) Lam.) is an important food crop in East Africa including Kenya. The crop incurs high yield losses in production due to biotic (insect pests and diseases) and abiotic (drought, and heat) constraints. Among abiotic constraints, drought is the most important. Prolonged periods of drought in arid and semi-arid areas of Kenya have led to reduced quantity and quality of sweetpotato storage roots and in severe cases caused total crop failure. The objectives of this study were to: 1) determine sweetpotato production system constraints and farmers’ coping strategies; 2) evaluate sweetpotato clones for yield performance and drought tolerance; 3) analyse genotype x environment interaction and stability for storage root yield of selected clones, 4) determine mechanisms of drought tolerance in sweetpotato, and 5) determine combining ability and heterosis for yield and drought tolerance traits under managed drought stress conditions. To determine the production constraints and farmers’ coping strategies, a survey was conducted in central, eastern and western Kenya. Out of 345 farmers interviewed, 60% were women, and 40% men. Farm sizes ranged from 0.4-0.8 ha, with 90% of sweetpotato cultivated on 0.2 ha or less. The main sweetpotato varieties were Vitaa, Kabonde and Bungoma and the majority of farmer’s used their own conserved planting material which was conserved by leaving them in the field after harvest. About 35% of the farmers identified weevils as the major pest, and sweetpotato virus disease (SPVD) as the major disease, while 28% of the farmers identified drought as a major constraint. The farmers used clean seed, high yielding varieties, high planting density, and manure application as the main strategies to cope with sweetpotato production constraints. Eighty four sweetpotato clones were evaluated under managed drought stress environments at KARI-Kiboko and KARI-Thika. Drought reduced the fresh weight of storage roots (FSR) (72.5%), fresh biomass weight (FB) (74.0%), marketable fresh storage root (MFSR) (80.7%), number of storage roots (NSR) (24.5%), days to permanent wilting point (DPWP) (0.3%), but seemed to increase percent root dry matter (% RDM) (-4.7%), harvest index (HI) (-2.6%), and chlorophyll content (CC) (-2.7%). Across the environments, genotypes 194555.7 (1.06), 421066 (1.05), Chingovu (0.94), 420014 (0.91), Excel (0.9), 199062.1 (0.87) and Unawazambane06-01 (0.81) gave higher FSR yields (kg plant-1) than the local checks. Genotypes Nyarmalo and Polista were among the lowest yielding in the irrigated and non-irrigated conditions. Clones W119, 441725, and Xiadla-xa-kau were the highest yielding under the drought stress conditions. The performance of 24 improved clones was evaluated in replicated trials at KARI-Thika and KARI-Kiboko using 24 sweetpotato clones grown under managed drought stress conditions for two seasons. AMMI, GGE biplots and regression analyses were conducted to determine stability of the clones. Mean FSR was significantly different (P < 0.001) in the two research sites with the environment contributing to 92.7% of the total variation, genotype 1.8%, and interactions 0.4%. AMMI and GGE biplots, and regression indicated the most stable clones to be 441725, Unawazambane06-01 and 189150.1, while Xiadla-xa-kau was the least stable. To gather more information on drought tolerance mechanisms expressed in the sweetpotato genotypes, an experiment was conducted in the greenhouse using clones expressing different levels of drought tolerance in the field. The results showed more under-developed roots (pencil roots) in the drought stressed regimes than in the irrigated regimes and in the drought susceptible genotypes. Drought tolerant genotypes produced more FSR and NSR. The number of vine branches (NVB), vine tip pubescence (VTP) and mature leaf pubescence (MLP) increased with reduction of drought stress however, drought tolerant clones had reduced NVB. Drought stressed clones had shorter basal vine length (BVL), reduced CC, and reduced leaf growth (LG). Drought stress reduced growth of vines in terms of internodes length, internodes diameter, vine length, petioles length, and leaf CC. Overall, for the first time, the study demonstrated that drought stress in the first three months after planting, leads to the proliferation of non-edible pencil roots, which do not become edible storage roots even when drought stress is removed. Therefore, water is critical during this period for improved sweetpotato storage root yield, probably as a drought tolerant mechanism. Combining ability for yield and drought tolerance of 15 F1 sweetpotato families generated through a half diallel mating of six parents was evaluated at KARI-Kiboko in 2012. Significant (P≤0.05) general combining ability (GCA) and specific combining ability (SCA) effects were recorded for root yield in both drought stress and no stress conditions, indicating that both additive and dominance gene effects were important in the inheritance of resistance to drought stress. Progenies from families G2, G5, G7, G8, G10, G12 and G15 had good SCA for fresh storage root yield, total biomass, number of days to permanent wilting point (DPWP), harvest index, and drought stress index (DSI). Progeny 8 from family G4, 5 and 8 from G15, had the highest mid and best parent heterotic effect (117-270%) for fresh storage root yield in both drought stress and no stress conditions. Progeny 6, and 7 from family G10, had the highest mid and best parent heterotic effect (165-234%) for fresh total biomass yield under no drought stress conditions. In conclusion, the progenies from families G2, G5, G7, G8, G10, G12 and G15 that had high yield and biomass specific combining ability under drought and no drought stress indicated that they could be having drought tolerant genes, and therefore could be incorporated into advanced drought screening trials with the aim of releasing the best performing drought tolerant varieties. Secondly, the findings in this study lay a foundation for sweetpotato breeding programmes on drought tolerance. Thirdly, for the first time, this study uniquely combines yield performance, combining ability estimates, days to permanent wilting point and heterosis under contrasting moisture regimes to unmask the gene action of drought tolerance in sweetpotato, a milestone in science. / Thesis (Ph.D.)-University of KwaZulu-Natal, Pietermaritzburg, 2013.

A study was conducted to determine the responses of 10 selected soybean (Glycine max L.) genotypes to potassium silicate (KSi) and Trichoderma harzianum (Eco-T®) applications. Preliminary studies involving two independent experiments were conducted under controlled conditions at the University of KwaZulu-Natal during 2010. Potassium silicate at three concentrations (0, 200 and 250ppm) were applied twice weekly over a period of four months to the genotypes laid out in a randomized complete block design. Subsequently, a field experiment was conducted at Ukulinga Research Farm of the University of KwaZulu-Natal, Pietermaritzburg during 2010/2011 to investigate the responses of the genotypes to KSi at 0 and 200ppm, with and without(Eco-T®) seed treatment. This experiment was set out in a randomized complete block design with three replications. Data collected included number of days to 50% flowering, number of days to 50% maturity, plant height, number of pods per plant, number of seeds per pod, 100 seed weight, root mass, shoot mass, seed yield and harvest index. The total number of root nodules formed and the number of active nodules were determined at end of the field experiment. In most cases a decrease was noted in total nodule formation as well as a decrease in the number of active nodules that formed. In the controlled environments there was a significant interaction between genotype and KSi concentrations for all measured traits. In most cases KSi applied at 200ppm was more successful in enhancing growth, improving seed yield and resulted in high harvest indices. The genotypes that produced the highest seed yield and harvest index in these environments were Williams and Barc-2 at 200ppm KSi. Results from correlation analysis revealed that harvest index and seed yields were generally positively associated with plant height, number of pods per plant and 100 seed weight, which in turn were the traits that contributed to most of the variation to seed yield and harvest index as revealed in the principle component analysis (PCA). The field experiment revealed a significant interaction between genotype x KSi x Eco-T®. Potassium silicate applied at 200ppm with Eco-T® usually promoted growth, seed yield and high harvest indices for all the genotypes. The PCA showed seed yield and harvest index were the traits that contributed to most of the variation. Genotypes Williams, LS6161R, Magoye and Barc-2 were the best seed yielders with the highest harvest indices that responded strongly to the combined use of KSi and Eco-T® under field conditions. Genetic comparison of the ten soybean genotypes with eight microsatellite markers revealed the close genetic relationship between Williams, LS6161 R and Magoye. A link between Barc-2 and Williams was noted by the common parent Clark. Therefore, for these genotypes, the application of KSi at 200 ppm with Eco-T® under field conditions effectively increased seed yield, ranging from 0.45 to 65.26% for some genotypes when compared to the control. An increase was also noted for other agronomic traits and harvest index. / Thesis (M.Sc.)-University of KwaZulu-Natal, Pietermaritzburg, 2012.

Maize production in Africa is constrained by abiotic and biotic stresses. Breeders need to have information on the nature of combining ability of parents, their traits and performance in hybrid combination. This requires careful determination of genetic variability of parents, and studying associations between grain yield and adaptive traits to breed superior cultivars which are better able to withstand such stresses. Therefore, this study was aimed at selecting parental testers with best combining ability in hybrid combination with recombinant inbred lines (RILs); and studying the correlation between grain yield and its components in eastern and western South Africa. It was also aimed at determining genetic variation and associations among adaptive traits in hybrids involving RILs. The final objectives of the study were to determine cultivar superiority of testcrosses involving RILs, and to select the best cultivars within and across four different environments. The 42 RILs were crossed to 9 Zimbabwean tropical testers resulting in 1009 hybrids with sufficient seed for planting in trials. From these a sample of 87 hybrids with adequate seed were selected and planted at four sites for combining ability analysis. The hybrids were evaluated at four sites in two regions; western region (Potchefstroom research station) and eastern region (Cedara, Ukulinga and Dundee research stations), during 2011/12 season. The experiments were laid out as augmented alpha lattice design. Trials were managed in accordance with production culture for each region. All quantitative data was subjected to GenStat and SAS statistical softwares. The results from combining ability study indicated that the line general combining ability (GCA) effects played a non-significant role (p > 0.05) in determining grain yield, grain moisture and anthesis date, while they were significant (p ≤ 0.05) for the other traits such as ear prolificacy. The tester main effects were significant for all the traits except ear prolificacy and plant height. Results also revealed that all the traits were controlled by both additive and non-additive genes, where additive gene action had the most contribution to the traits. The non-additive gene action played a minor role suggesting the total GCA effects attributed to both lines and testers predominantly higher over the specific combining ability (SCA) for all traits. In general the additive effects were preponderant over the non-additive gene effects. One cross (L114 x T12) had a significant and positive SCA effect for grain yield. The correlation between grain yield and secondary traits (number of ears per plant, grain moisture content, ear height, plant height, ear position and anthesis date) suggested that indirect selection can be employed to enhance grain yield by breeding for these particular adaptive traits. Path analysis showed that plant height had the highest direct and indirect effect on grain yield indicating its importance among other secondary traits for grain yield enhancement. Phenotypic coefficient of variation (PCV) was higher than genotypic coefficient of variation (GCV) for all the studied traits across all the four environments. All the traits displayed high heritability at Potchefstroom except anthesis date which was highly heritable at Ukulinga. Cedara was the second best site for heritability of all the traits except for the number of ears per plant. The genetic advance for grain yield was the highest at Cedara followed by Potchefstroom, Dundee and Ukulinga. The hybrids exhibited different patterns of variation and distribution for all the traits. This indicated that selection strategies to exploit GCA should be emphasised. Association studies among grain yield and secondary traits such as ear length, number of ears per plant, plant height, anthesis date, silking date and ear leaf area revealed that there were significant phenotypic correlations between grain yield and secondary traits, and among the secondary traits. Ear length had the highest direct effect on grain yield at Ukulinga; number of ears per plant had the highest direct effect on grain yield at Cedara and Potchefstroom; whereas plant height had the highest direct effect on grain yield at Dundee. Grain yield was least affected by indirect factors at all the sites except Ukulinga, where anthesis date had the highest indirect effect on grain yield through silking date followed by plant height through leaf area. The study reveals that there is significant variation among the hybrids for mean performance, indicating that there is opportunity for selection. Overall the findings suggest that direct selection would be appropriate to enhance grain yield. Path analysis revealed that plant height had the highest direct and indirect effects on grain yield, indicating that plant height can be further exploited as the main trait in future breeding programmes for grain yield increment. Hybrid 10MAK10-1/N3 was the best hybrid at Ukulinga in terms of grain yield, relative yield and economic traits. Whereas hybrid T17/L83 was the best hybrid at Cedara in terms of grain yield and relative yield; however, T11/L102 was selected as the most elite hybrid with respect to grain yield, relative yield and economic traits. Hybrid T3/L48 was identified as thebest hybrid at Dundee with respect to grain yield, relative yield and prolificacy. At Potchefstroom the standard check PAN6611 was identified as the best hybrid in terms of grain yield and relative yield followed by developmental hybrid T1/L28; however, developmental hybrid T1/L28 was the best in terms of earliness, prolificacy and ear aspect. Stability coefficients and cultivar superiority index across the sites revealed that four developmental hybrids were identified as best hybrids and they performed better than the standard check. These hybrids will be recommended for further testing in advanced trials. With respect to cultivar superiority, the desired hybrids are required to combine high grain yield with economic and adaptive traits such as high ear prolificacy, low grain moisture, and low ear aspect score (desired) for them to adapt to production environments in South Africa. There was significant variation among the top 25 yielding hybrids. At least 5 hybrids combined high grain yield with the desired complimentary adaptive traits such as quick moisture dry down, prolificacy and ear aspect. The results showed that there is variation in the performance of high yielding genotypes within all the sites, and that agronomically superior cultivars can be identified. The study shows that there is significant variation among the RILs since they interacted differently with the 9 tropical testers. Even among the top 25 selections of RILs in each environment there was still variation for combinations of the desired traits. Significant associations among grain yield and other economic and adaptive traits were observed with implications for breeding strategy. Above all the significant variation gives large score for future breeding of new unique lines. / Thesis (M.Sc.)-University of KwaZulu-Natal, Pietermaritzburg, 2012.

There are few maize varieties that are drought tolerant in semi-arid eastern Kenya and farmer perceptions of drought tolerant maize cultivars have not been studied in this region. Farmers in this region use maize landraces that have not been studied for their potential future hybridization. The main objectives of this study were therefore to: (i) study farmer perceptions of drought and preference for maize varieties, (ii) improve drought tolerance in maize populations in the semi-arid eastern Kenya using S1 family recurrent selection, and (iii) classify maize landraces according to their heterotic patterns. A participatory rural appraisal (PRA) was conducted in Machakos and Makueni districts in semi-arid eastern Kenya. A total of 175 farmers were involved in focus group discussions. An open ended questionnaire and a checklist were used to guide the farmers during the discussion sessions. Scoring and ranking techniques were used to assess farmers’ preferences of maize varieties and constraints to maize production. The farmers grew maize as their major crop followed by beans. Nearly 60% of the farmers grew local maize landraces, whose seed they recycled from season to season; 40% grew improved varieties, but mainly composites rather than hybrids. The key farmers’ criteria for choosing a maize variety in order of importance were drought tolerance, early maturity, high yield, and disease resistance. The major constraints to maize production were drought, lack of technical know-how, pests, poor soils, and inadequate seed supply. Maize traits preferred by farmers in a drought tolerant variety included high yield, recovery after a dry spell and the stay green characteristic. Two maize landrace populations MKS and KTU from semi-arid eastern Kenya and three CIMMYT populations V032, ZM423, and ZM523 were subjected to two cycles of S1 progeny recurrent selection for drought tolerance in yield and traits indicative of drought tolerance were measured during flowering and grain filling from February 2005 to September 2007. Evaluation to determine selection gains was done in one trial replicated five times. It was laid out as a 4x4 lattice design and drought was imposed at reproductive stage by withholding irrigation one week before flowering and resumed during grain filling. The trial was repeated under well-watered conditions which served as a control experiment. After two cycles of selection under drought stress conditions, KTU population had a realized gain in yield of 0.2 t ha-1, MKS population 1.2 t ha-1 and ZM423 0.4 t ha-1, whereas in V032 and ZM523, grain yield reduced by 1.1 t ha-1 and 0.6 t ha-1, respectively. Under well watered conditions, the realized gains in grain yield were positive in all the populations except V032, where there was a reduction of 0.1 t ha-1. Selection increased the genetic variability and heritability estimates for yield in S1 lines of MKS and ZM423 populations, but decreased in KTU, V032 and ZM523 populations. The research to identify heterotic patterns was undertaken using ten maize landraces from the semi-arid eastern Kenya, six maize landraces from coastal Kenya, and three maize populations from CIMMYT. These populations were planted at Kiboko Research Farm during the short rains of October-December 2005 and crossed to two population testers, Embu 11 and Embu 12. The evaluation of the test crosses was done during the long rains of March-June 2006. Percentage heterosis for yield ranged from -17.7% to 397.4%, -79.4 to 22.2% for anthesis-silking interval, -23.9% to 29.2% for ear height, -0.1 to 1.1 for ear diameter, -7.1 to 21.2% for ear length and -5.9% to 30.3% for plant height. iii General combining ability (GCA) effects were significant (p=0.05) for all the traits, while specific combining ability (SCA) effects were not significant (p>0.05), implying that variation among these crosses was mainly due to additive rather than nonadditive gene effects. Since SCA was not significant (p>0.05) for yield, maize populations were classified based on percentage heterosis for yield alone. The maize populations therefore, were grouped into three different heterotic groups P, Q and R. Twelve landrace populations and two CIMMYT populations showed heterosis with Embu 11 and no heterosis with Embu 12 were put in one group P. Two landrace populations that showed no heterosis with either tester were put in group Q. Two landrace populations and one CIMMYT population showed heterosis with both testers were put in group R. None of the populations showed heterosis only with Embu 12 and no heterosis with Embu 11. The main constraint to maize production was drought and the farmers preferred their landraces whose seed they recycled season to season. After two cycles of recurrent selection, the landrace populations showed improved progress in yield. Thus, further selection will be beneficial in the populations where genetic variability increased. Therefore, these populations can further be improved per se and released as varieties and/or incorporated into the existing maize germplasm to broaden their genetic base, given that their heterotic patterns have been identified. Considering that farmers recycle seed, breeding should be towards the development of open-pollinated varieties which are drought tolerant. / Thesis (Ph.D.)-University of KwaZulu-Natal, Pietermaritzburg, 2007.

Maize (Zea mays L.) streak virus disease (MSD) is the most important virus disease in Africa but farmers are unaware of its status. A project was initiated to assess the current status of MSD and to breed for its resistance. Four populations comprised of two BC1F1 and two F2 progenies developed by backcrossing and selfing the F1 progenies of two crosses between a donor line (CMl 202) and two susceptible lines (CMl 321 and CMl 384) were developed. Conventional and molecular marker assisted selection (MAS) methods were used to screen for resistance to MSD in each of the four populations. To facilitate unbiased comparison, separate screening nurseries were established for MAS and conventional screening. The objectives of the study were five-fold; 1) to assess the status of MSD in Uganda and understand farmers' preferences and varietal selection criteria for maize using a participatory rural appraisal (PRA), 2) to screen for MSD resistance in early generations of segregating maize populations using conventional method, 3) to screen for resistance to MSD using SSR marker assisted selection , 4) to compare the effectiveness of marker assisted selection and conventional methods for selection for resistance to MSD, and 5) to compare costs associated with MAS and conventional selection methods. Results of PRA showed that unreliable rainfall and insect pests were the dominant constraints to maize productivity in Uganda. Diseases were ranked fifth among the production constraints . Maize streak virus disease was considered the most important disease constraint. Farmers showed common preference for high yielding and early maturing cultivars. However, farmers had other special preferences which were diverse and included large, white and high test density kernels for marketing, and sweet taste, particularly for home consumption. Farmers' research priorities included tolerance to drought, resistance to insect pests and diseases, sweetness, prolificacy, resistance to lodging, and drooping leaves because theyt cover the soil fast and prevent weed growth. Conventional screening for resistance to MSD showed that backcross and selfing populations segregated in 1:1 and 3:1 Mendelian ratios confirming the presence of one major gene with simple inheritance . Severity and incidence of disease were positively correlated suggesting a non-reference by the insects. In the selfing populations, the presence of complete esistance against MSD was suggested because frequency distribution patterns were highly skewed in favour of resistance. There was a decrease in disease severities with selection from BC1F1 to BC2F1 and from F2 to F3 generations indicating that high response to selection was achieved. On the other hand, one marker, umc1917, consistently polymorphic and eo-dominant was selected and used in MAS protocol. Results showed that the observed outcomes fitted the expected ratio of 1:2:1 for a F2 population and 1:1 for a BC1F1 population (X2 not significant). Evaluation of F3 and BC2F1 progeny selected using markers showed low disease severity suggesting that marker assisted selection was effective. However, the study showed that the presence of the O'Tl, was not consistent with symptom expression in the field. Evaluation of lines in three-way crosses identified ten potential lines that were high yielding, highly resistant to MSD and stable across three locations. Both MAS and conventional selection were equally effective in identifying high yielding lines although resistance was higher under MAS. Costs of MAS and conventional method varied depending on the units for comparison. The total costs of conventional method were higher than that of MAS in both first and second selection cycles. Comparing costs per row for conventional and costs per plant or data point for MAS showed that conventional selection was 2.4 times more expensive than costs per sample for MAS. However, costs per plant for MAS were 6.6 times higher than for conventional selection. / Thesis (Ph.D.)-University of KwaZulu-Natal, Pietermaritzburg, 2006

In Southern Africa and Mozambique, tropical lowland accounts for 22% and 65%, respectively, of area under maize production, but grain yield is compromised by downy mildew disease (DM, which is caused by Peronosclerospora sorghi (Weston and Uppal) Shaw), and lack of appropriate varieties, especially hybrids. Among other factors, productivity can be enhanced by deploying DM resistant hybrids, which are higher yielding than open pollinated varieties. Development of a viable hybrid-breeding programme requires knowledge of genetic effects governing yield and DM resistance in inbreds, and effective germplasm management requires heterotic groups and heterotic patterns to be established. In addition, knowledge of farmer-preferred traits is required. Currently, such information is not available to the hybrid-breeding programme in Mozambique. The objectives of this study were, therefore; i) to identify farmers’ preferred variety traits and major production constraints, ii) to determine combining ability effects of inbred maize S4 lines for grain yield and DM resistance, iii) to determine heterotic groups and heterotic patterns among the elite inbred maize lines, and iv) to investigate gene effects governing resistance to DM in breeding source inbred maize lines from the breeding programme in Mozambique. During 2007/08, 142 households were involved in a survey conducted in three districts representing two maize agro-ecological zones in Mozambique. Formal surveys and informal farmer-participatory methods were employed and data subjected to analysis in the SPSS computer programme. Results indicated that there was a low utilization of improved varieties, especially hybrids, with grain yield estimated at 0.7 t ha-1. Farmers were aware of the major production constraints and could discriminate constraints according to their importance for their respective communities. For the lowland environment, farmers identified downy mildew, drought, and cutworm and stem borer damage as the main constraints. In contrast, for the high altitude environments, they ranked ear rot, seed and fertilizer availability, turcicum leaf blight, grey leaf spot diseases and low soil fertility among the major constraints limiting productivity. The most important variety selection criteria were grain yield, short growth cycle, white and flint grain with stress tolerance to drought, low soil fertility, diseases, and grain weevils. These afore mentioned traits, would be priority for the breeding programmes for the lowland and mid altitude environments in Mozambique. To determine combining ability for downy mildew resistance, heterotic groups and heterotic patterns, two testers (open-pollinated varieties) ZM523 (Z) and Suwan-1 (S), were crossed with 18 lines to generate 36 top crosses for evaluation. Crosses were evaluated at two sites under DM. Preponderance of GCA effects indicated that additive gene effects were more important than non-additive gene effects in governing both grain yield and downy mildew resistance in the new maize lines. Based on specific combining ability (SCA) data, lines for yield were classified into two heterotic groups, S and Z; whereas based on heterosis data, lines were fitted into three heterotic groups (S, Z and SZ). Further heterotic patterns and gene action for yield were determined by subjecting nine inbred lines and the two testers, S and Z, to an 11 x 11 diallel-mating scheme. The diallel crosses, three hybrid checks and the two testers were evaluated in six environments in Mozambique. Results revealed that non-additive gene effects were predominant for yield components. In addition, high levels of heterosis for yield was observed and three heterotic groups identified (Z, S and S/Z), and five exceptional heterotic patterns among the inbred elite maize lines were observed. Topcrosses with yield levels comparable to single cross hybrids were also identified, and these would be advanced in the testing programme with potential for deployment as alternative cheaper and sustainable technology to conventional hybrids for the poor farming communities in Mozambique. To determine gene effects for downy mildew resistance in potential breeding lines, two maize populations were derived from crosses between downy mildew susceptible line LP67, and resistant lines DRAC and Suwan-L1. To generate F2 and backcross progenies (BCP1 and BCP2), F1 progenies were self-pollinated and simultaneously crossed to both inbred parents (P1 and P2). All the six generations (P1, P2, F1, F2, BCP1, and BCP2) of the populations were evaluated at two sites under downy mildew infection. A generation mean analysis was performed in SAS. It was revealed that downy mildew resistance was influenced by genes with additive and dominance effects, plus different types of epistatic effects such as additive x additive, and dominance x dominance. Overall results indicated that genes with predominantly non-additive effects controlled resistance in DRAC, whereas resistance in Suwan-L1 was largely influenced by additive gene effects. These findings have serious implications on the effective use of these downy mildew resistance sources in breeding programmes that aim to generate varieties with downy mildew resistance. Overall, results suggested that inbreeding and selection within heterotic groups, followed by hybridization between inbreds within and across heterotic groups would be effective to generate new hybrids. The breeding programme will consider development of conventional hybrids, such as single crosses and three way crosses, and top crosses. Implications of the findings of the study and recommendations are discussed. / Thesis (Ph.D.)-University of KwaZulu-Natal, Pietermaritzburg, 2010.