VARIABILITY AND SELECTION STUDIES IN 2 KENYAN BEAN LANDRACE POPULATIONS

Two Kenyan landraces of field beans (Phaseolus vulgaris L.), PAS/001 a nd PAS/002, were subjected to variation analysis at the University of Nairobi in 1986 and 1987. The populations were variable for the number of filled pods, seeds per pod, weight of 20 seeds, pod length and num ber of pod-bearing nodes. PAS/002 was also significantly (p = 0.5) var iable for yield and days to flowering. Two cycles of both direct and i ndirect mass selection were carried out on the two populations in 1986 and 1987 at 10% selection pressure for grain yield improvement. In po pulations PAS/001, the number of filled pods per plant and pod length were used as indirect selection criteria. Selection for number of fill ed pods per plant was comparable to direct selection for yield in impr oving the yield of this population. Pod length was an ineffective sele ction criterion of yield improvement. In population PAS/002, the numbe r of filled pods per plant and the number of pod-bearing nodes per pla nt were used as indirect selection approaches to yield improvement. Bo th traits were more rigorous selection criteria than direct selection for yield.