| Abstract: | Twenty cowpea (Vigna unguiculata L. Walp.) genotypes, comprising Kenyan cultivars and Kenyan and introduced breedings lines, were grown in 15 replicated field trials carried out at three locations in semi-arid eastern Kenya over four seasons. Grain and straw yields and the carbon isotope discrimination (Δ) of this material were determined, and days to flowering and maturity observed.None of the seasons in which the trials were conducted was appreciably drier than the long term average for the sites, and some trials received excessive rain. Averaging the data for the 15 trials, one genotype yielded much less grain (77 g m−2) than the 19 others, whose yield ranged from 125 to 177 g m−2 (average 148 g m−2). Trial mean yield trials had a high straw Δ (r = +0.567) and had received more rain between flowering and maturity (r = +0.428) than lower yielding ones. Among genotypes, averaging over trials, the correlation coefficient (r) between grain yield and grain Δ was +0.394, and between grain yield and straw Δ, +0.460. Early genotypes had the highest grain Δ and straw Δ. The correlation among genotypes between date of flowering and grain Δ was −0.632 and with straw Δ, −0.502. When comparisons were made among trials, there was no clear relationship between the strength of the correlations among genotypes between grain yield and straw Δ or grain Δ and the degree of stress experienced by a trial (as indicated by the rainfall it had received). It appeared that this was because the variances of grain Δ and straw Δ were greater in the most droughted trials, whereas the variance of yield was greatest in the least droughted trials.It is concluded that the genetic correlation between grain yield and either grain Δ or straw Δ is unlikely to be strong enough to make either Δ a useful surrogate or adjunct in selecting for high yield in the segregating generations of a breeding programme for semi-arid eastern Kenya, but that Δ may be of value in the selection of parent genotypes. |
