Genetic potential of cowpea breeding populations for grain yield and agronomic traits improvement in Uganda.

Cowpea yields are low due to the use of varieties that are susceptible to multiple biotic stresses such a scab and virus, and insect pests especially thrips infestations. In order to develop cultivars with a wide range of farmer preferred characteristics and resistant to multiple biotic stresses, large populations with wide genetic diversity are required, but there are limited effective selection indices to cater for the multiple traits that are required while human and financial resources are limiting for the selection processes. This study was conducted to evaluate the efficacy of two selection criteria i.e. the usefulness criterion and base selection index to identify the best segregating populations for multiple traits (high yields, resistant to virus, scab, thrips and with farmer preferred traits). Twenty-five cowpea parental lines that were preferred by farmers and with traits namely resistance to scab and virus disease, thrips and high yield potential were used to generate diverse populations. A total of 135 F2 populations and 25 parents were evaluated in the field at Makerere University Agricultural Research Institute (MUARIK) during the 2017A season to identify the best performing populations for multiple traits. The design was an alpha lattice design with 5 blocks of 32 plots which was replicated twice. Data were collected on scab, virus, thrips, yield and yield related traits. There were significant differences among populations (P<0.05) for all the traits evaluated except for thrips. Among the best populations identified were 2392 x Ebelat*NE 51, 2392 x NE 5, 3306 x Ebelat*NE 5, Ayiyi x 2392, Danila x NE 48, Danila x NE 5, KVU 27-1 x WC 27, MU 15 x Ebelat*NE 51, MU 20B x WC 27, MU 9 x NE 55, NE 21 x NE 55, NE 5 x 2392, WC 48A x 2392 among others. Usefulness criterion computed for yield identified NE 36 x 2392 and WC 63 x 2392 as the best and worst populations, respectively. Usefulness criterion computed identified NE 5 x Sanzi and WC 48A x 2392 as the top ranked populations for yield and for yield and its components respectively. The cross WC 48A x 2392 was identified as the top ranked population using base selection index values that included resistance to virus, thrips, scab and the yield and its component when only yield and its components were fitted in the model. There were no differences observed between the two methods used for selection and sixteen populations out of the best 30 selected populations appeared in common in both methods. In breeding for multiple traits in cowpea, both methods can be used. A total of best 40 cowpea populations (960 F2:3 lines) were selected and were randomly divided into three sets of 320 lines and each being planted in the locations: MUARIK, Serere - National Semi Arid Resources Research Center (NaSARRI) and in Arua. The experiment was set in an alpha lattice design of 10 blocks by 40 plots and was replicated twice. The parents, populations and lines exhibited high significant (P<0.01) variations for all the traits evaluated in all locations except for the trait thrips in the locations Serere and Kabanyolo and for most traits in Arua. Significant variations were not observed among the populations, parents and lines for some traits in Arua due to the existence of large amount of error that resulted from the non-uniformity of the field which the area is well known for being hilly. Significant variation (P<0.001) existed among the populations by location interaction an indication of the existence of variation due to environmental influence. At individual locations, 6 genotypes were resistant to virus disease infection in Arua, 7 in Kabanyolo and 1 in Serere. Most cowpea genotypes were resistant to moderately resistant to scab disease infection. Furthermore, there were 85 genotypes that were identified with resistance to thrips damage in Kabanyolo, 21 in Arua and 118 in Serere. Populations performed better than their parents for some traits especially number of pods and peduncles and days to flowering which suggested the presence of transgressive segregants. Similarly, lines KVU 27-1x WC 27/8 and NE 55 x NE 5/6 which gave a greater number of pods and peduncles as compared to their parents and WC 27 x VCR 1432/15 which flowered earlier than their parents, exhibited transgressive segregants ability. Across locations the study revealed that NE 48 and WC 63 x NE 48 were the best yielders. Genetic coefficient variability was lower than the phenotypic coefficient of variability which suggested that the variability observed was due to the influence of environment. Low heritability was recorded for crosses in reaction to thrips, scab, seeds per pod and number of pods and peduncles per plant. Moderate heritability values were recorded for crosses in reaction to virus, days to flowering and grain yield while high heritability was obtained for 100 seed weight. Negative and significant associations existed between yield and scab disease, between yield and thrips and between yield and virus disease. Highly significant and positive associations existed between yield, number of seeds per pod and weight of 100 seeds. This indicated that there is need to improve genotypes that exhibit resistance to diseases to improve on yield. Accordingly, in the cowpea improvement programme, the results obtained in this study show that there is potential of promising cowpea lines with desirable attributes preferred by farmers and these can be advanced in several diverse environments in addition to the current locations for adaptability and stability.