English French




Solanum aethiopicum is one of the most important Solanum species, with four morphological groups. Two of the groups, Gilo and Shum, are mainly cultivated because of their nutritional value and income generating potential for farmers in developing countries. Of focus for this study was the Shum, a leafy morphological group whose productivity and quality is directly affected by drought. Inadequate productivity of farmers’ selections is mainly a result of drought stress, inadequate information regarding genetic diversity intended for selecting stress resistant varieties, and scarce information for breeders on breeding value of germplasm for water deficit stress tolerance improvement. Global limitations on water resource availability call for need to develop productive varieties that are drought tolerant. This called for need to improve drought tolerance in S. aethiopicum Shum (SAS). The specific objectives were to: (i) determine the genetic distinctiveness between SAS and its progenitor, S. anquivi (SAN); (ii) evaluate genetic diversity within SAS germplasm; (iii) identify parental material for development of drought tolerant S. aethiopicum Shum varieties; and (iv) determine the combining ability of selected Shum group germplasm for drought tolerance.
In determining genetic distinctiveness between SAS and SAN, 25 accessions five of which were the wild progenitors, was evaluated in a screen house and morphological data collected. The UPGMA generated clusters exposed presence of moderate structure associated with a cophenetic correlation coefficient of 0.73. Five distinct clusters were identified; the progenitor accessions for S. aethiopicum Shum were grouped in their own cluster. The Richness, Shannon-Weaver and Simpson indices were also different among qualitative variable categories. A ‘prcomp’ function based principal component analysis (PCA) in R on quantitative variables indicated that days to germination and emergence (DG), cotyledonous leaf blade length (CLBL), cotyledonous leaf width (CLBW), fresh shoot biomass (SBF), plant height (PH), petiole length (PL), days to first flower appearance (FLW), leaves per plant (LPP), plant canopy width (PW), and plant branching (PB) were the major drivers of variability in germplasm studied. Further, results from canonical discriminant analysis to discern between the S. aethiopicum and its progenitor accession groups showed that DG provide the best separation; with the former emerging earlier than the latter. The study revealed that morphological markers are useful in distinguishing between the S. aethiopicum Shum and its progenitor accessions.
Similarly, clustering was used to identify structure within in 20 accessions of SAS. Sixty-one morphological parameters were measured from a screen house experiment set up in a completely randomized design with 12 plants per accession for two seasons. Multivariate analysis of variance followed by canonical discriminant analysis was used to model each of 61 traits, as predicted by clusters and experiment to sieve out non-significant variables. Clustering results showed four separate clusters supported with a cophenetic relation coefficient of 0.87 at p < 0.001 confidence level. Petal length (PEL), seed color (or sepal length), fruit calyx length, seeds per fruit, LPP (or leaf fresh weight, LYF), fruit fresh yield, seedling vigour, fruits per plant, harvest index and plant growth habit were canonical variates that best predicted observed structure. The findings show prospect for variety discrimination in SAS at each of seedling, vegetative and reproductive stages. Further, SAS germplasm were evaluated to discover accessions (G) which excelled across water deficit regimes (WLs) where a split-plot arrangement in screen house was used. Four WLs and 20 genotypes of SAS as whole plot and sub-plot factors, respectively, were implemented in two replications for two seasons. In both first and second seasons, a highly significant effect (p < 0.05) of at least two WLs on performance among at least two genotypes for majority of the traits was established. Similarly, WL x G interactions were very highly significant (p < 0.001) for leaf relative water content (LRWC), LPP and PH, and non-significant for leaf blade width (LBW) and leaf blade length (LBL). Preference as breeding traits for stability superiority across WLs was suggested as LRWC > PH > LPP. Basing on the LRWC, superior and most stable genotypes were identified as E6 followed by E12, E15, E16 and E14GP. In contrast, least stable genotypes were observed as E7H, E17GP, E3H and E9. The H2 for each of recommended traits for water deficit stress tolerance breeding was > 0.9 and expected genetic advance as % of grand mean ranged from 16.68 (for LRWC) to 70.38 % (for PH) per generation; indicating a prospective response to selection. Water deficit stress reduced performance of SAS germplasm but observed differences enabled for selection of stable and superior genotypes. After identification of contrasting accessions for drought stress response, another study was carried out to determine the mode of gene action and combining ability for drought resistance among accessions. A 9 × 4 North Carolina II mating design with some missing values was applied which generated 24 successful crosses. Evaluation was done at five moisture regimes premised on crop growth stage and applied moisture as a percentage of field capacity of potting substrate. Restricted maximum likelihood (REML), rather than ANOVA, was opted for when analyzing the data. Effects of specific combining ability (SCA) were significant across and within moisture regimes for all study traits unlike for general combining ability (GCA) where significant effects were obtained with chlorophyll content (CHL) only. In the narrow sense (h2), the most highly heritable traits were identified as LPP, CHL, LYF and leaf dry yield (LYD) while leaf area (LA), leaf mass area (LMA) and LRWC were least heritable. Broad sense heritability (H2) was however, > 0.80 for all measured traits, indicating that dominance gene action (VD) exceed additive gene effects (VA) for moisture deficit stress tolerance in SAS. It was revealed that LA, LRWC, CHL and LMA were suitable for selection of best combiners (SCA) under well-watered (WW) and drought-stress (DS), DS, both DS and drought recovery (DR), and DR, respectively.
The parents with best GCA effects for leaf wilting score, LPP, LA, LRWC, CHL and LMA were identified as E3H followed by E6, E13 followed by E20, E11 followed by E2, E6 followed by E10, E10 followed by E13, E11 followed by E1, and E11 followed by E4, respectively. The crosses with best SCA were identified as E10xE20 (for LA under well-watered), E3HXE15 (for LYF across watering environments, LRWC under DS and CHL under drought recovery), and E11xE4 (for LMA under drought recovery). The studies reported in this thesis unearthed vital information as basis for establishment of a breeding programme for the crop.