SELECTION FOR YIELD AT TWO FERTILIZER LEVELS IN SMALL-SEEDED COMMON BEAN

Abstract

Selection for seed yield (visual in the F2, visual and plot yield from the F3 to F5, and yield tests in F6) was carried out in two crosses of common bean (Phaseolus vulgaris L.) in high and low soil fertility environments at CIAT-Quilichao, Colombia. Parents were small-seeded and tolerant to low soil phosphorus, possessed indeterminate bush growth habit, and belonged to lowland tropical gene pools of Middle America. The mean yield of selected lines from low (LFS) and high (HFS) soil fertility environments was significantly higher than the mean of the parents in both crosses when tested in high fertility (HF) but not in low fertility (LF). The highest yielding LFS and HFS lines from A 286 × (G 5059 × A 80) and the LFS line from A 286 × ICA Pijao outyielded the best check cultivar, Carioca, in HF. No line yielded significantly more than A 286, the best parent used in both crosses. The mean effect of fertilizer levels on selection for seed yield was nonsignificant. Lines selected under two environments showed similar but average response and high stability of performance under variable environments. Low soil fertility accelerated maturity and reduced 100-seed weight and seed yield.Key words: Phaseolus vulgaris, bean (small-seeded common), selection for yield, soil fertility, Middle American gene pools, regression coefficient