Evaluation of Three Cycles of Recurrent Selection for Fruit Yield within a Population of Virginia-Type Peanut
- E. J. Monteverde-Penso and
- J. C. Wynne *
Recurrent selection with minimal crossing and single-seed descent of S0 plants has been proposed as a breeding procedure for selfpollinated crops where hybridization is difficult. Three cycles of recurrent selection for fruit yield using a low selection intensity and little crossing were conducted in a genetically broad-based population of Virginia-type peanut (Arachis hypogaea L. ssp. hypogaea var. hypogaea). This study was conducted to (i) evaluate progress from selection for yield and (ii) estimate the realized heritability for yield as a function of the additive genetic variance in the population. Progress from selection was estimated using bulked seed of all selected lines within each parental group (BC experiment) and the mean of 20 randomly chosen lines from each parental group (EM experiment) to represent each of the cycles. The response to selection for fruit yield was linear across cycles with positive regression coefficients that were significantly different from zero for both experiments. Estimates of the progress of selection were 135 and 117 kg ha−1 per cycle for the BC and EM experiments, respectively. The proportion of selected individuals was 51, 43 and 50% for Cycles 0, 1, and 2, respectively. Selection based on fruit yield did not affect pod weight, pod length, seed weight, and seed count. Predicted progress was greater than observed progress in both experiments. The difference is an indication of possible nonadditive genetic variance as well as genotype ✕ environment interaction variance that inflated the estimate of heritability. Realized heritability, an estimate of narrow sense heritability in the base population computed by two different procedures, was about 30%. The recurrent selection procedure is simple, cost-effective, and increased the mean yield of the population. The small differences in the results of the evaluation observed between the two experiments suggest that there is no advantage in using individual lines per plot to evaluate the response from selection in peanut.
Copyright © .