Recurrent Selection for High Seed Weight in Cicer Milkvetch1

Abstract

Effectiveness of recurrent selection for high seed weight was determined in diallel cross and polycross populations of cieer milkvetch, Astragalus cicer L. To help prevent loss of vigor, selection was practiced for high vigor as well as high seed weight. In selecting parents of the second cycle from each population, selection pressure in one subpopulation was intense for mature‐plant vigor but was relaxed somewhat for seed weight. In the other subpopulation, selection pressures were reversed for the two characters.In the first cycle of selection, mean seed weight of the diallel and polycross populations was 3.99 and 3.97 g/1,000 seeds, respectively. This represented an increase in seed weight of about 13% over the original population (3.51 g/1,000 seeds).Two cycles of selection increased seed weight substautially. Mean seed weight of the four subpopulations ranged from 4.30 to 4.73 g/l,000 seeds, which represented increases of 22.5 to 34.8% over the original population. The percentage of increase in seed weight in cycle 2 over cycle 1 was greater for subpopulations in which selection pressure was higher for seed weight and lower for plant vigor. Vigor of some progenies in all subpopulations was high. A comparison of within‐plot variances for the subpopulations and clones indicated that genetic variance was still present after two cycles of selection. However, variance in the diallel snbpopulations was being depleted more rapidly than in the polycross subpopulations. This was attributed to the lower number of parents in the source population for the diallel group.