Evaluation Of Yield And Sub-Yield Components Of Cowpea (Vigna Unguiculata (L) Walp) Accessions

Abstract

One hundred and eight cowpea accessions were evaluated for yield and sub-yield components at the Department of Plant and Environmental Biology. They consisted of 60 F2 segregating, 44 accessions as test materials and 4 controls. The controls were Asontem Wang Kae, Padituya and Kirkhouse from CSIR-SARI. Data collected comprised of morphological and phytochemical traits. Descriptive, multivariate and genetic component analyses were carried out to evaluate the extent of variation. Rank summation index and contrast analysis were performed to select the best performing accessions. For the morphological traits studied, semierect growth habit with frequencies of 64.78%, 43.75% and 51.45% which occurred in the test materials, controls and the entire population. The majority of the accessions in the test materials exhibited red seed coat colour (19.59%) whereas in the controls most of the accession showed cream seed coat (50%). Overall, the majority of the accessions exhibited cream seed coat colour with a frequency of 35.35%. The chi-square test of association between qualitative traits showed was 125 significant associations in the test materials whereas the controls only showed 3 significant associations. The overall population exhibited 162 significant associations. In the multiple correspondence analysis (MCA) terminal leaf shape, growth habit, pod curvature and raceme position mostly led to variation in the first dimension which accounted for inertia of 51.40%. Thirteen farmer and consumer-preferred traits were used in the MCA plot. Asontem, UG8 and UG30 had a close relationship with similar phenotypic classes in the first quadrant of the MCA plot. UG1 and UG81 were outliers and had a distant relationship with majority of the phenotypic classes of the selected traits. The test materials had a significantly higher mean than the controls for the following traits; number of pods per plant, number of locules, seed weight and seed yield but lower than the controls for the following traits; days to first flower, days to 50% germination, days to first mature pod and days to 50% mature pod. Furthermore, the test materials were significantly higher in mean concentration for the following amino acids; gallic acid and vanillic acid, glycine, l-histidine, l-aspartic acid, l-valine and l-methionine. A total of 191, 537 and 413 significant pairwise correlations were observed in the test materials, controls and entire population respectively for all quantitative traits. The multiple regression analysis revealed that the underlying determinant of yield was most influenced by yield and yieldrelated components in the test materials and controls with contributions of 100% and 36.67% respectively. Phytochemical traits made the highest contribution of 25.95% to influencing variation in yield for the entire population. The first six principal components in the morphological traits accounted for a total variability of 47.17%, 55.40 %, 42.99% in the test materials, controls and entire population respectively. The first three principal components in the phytochemical traits accounted for a total variability of 51.81%, 100 %, 53.14% of the total variation in the test materials, controls and entire population respectively. The biplot showed that the relationships among accessions and morphological, phytochemical and all traits explained 46.63% and 99.49 and 94.48 % of the total variance respectively. The canonical discriminant analysis grouped against yield showed that the first CV accounted for 17.19% and 40.82% for the respective morphological and phytochemical traits. The cluster analysis based on morphological and phytochemical traits clustered the 108 accessions into 7 and 6 major groups respectively. The genetic component studies showed that seed yield had a high genotypic (63.1%) and phenotypic (243.8%) variances but low heritability of 18.41% while phytochemical traits had high GCV (>20), PCVs (>20) and heritability (>80%). Based on the lowest RSI scores, the 10 best performing accessions with high yield potential among the population evaluated were UG84, UG24, UG44, UG69, UG47, UG14, UG70, UG66, UG36, UG102 . The marginal analysis revealed that no accession was significantly higher in yield than the overall mean. However, it confirmed cluster grouping of accessions based on trait similarities and dissimilarities.