Full-sib and reciprocal recurrent selection in relation to pearl millet improvement.

During the years 1973 to 1976 two populations of Pearl millet with wide genetic base, namely, Delhi composite (DC) and Vijay composite (VC) were used to compare the response to selection by the full-sib family method from biparental material and reciprocal recurrent selection (RRS). The results indicated that it was possible to advance grain yield with one cycle of RRS by about 23 percent in the case of population DC and 21 percent in population VC, while for the full-sib selection method, the improvement in grain yield was not so rapid. The studies on the nature of gene action indicated that both additive and dominance gene actions were important for grain yield, ear length and ear girth. The coefficient of variation as a result of RRS was reduced in population DC, while it was comparable to base population in the other population. The correlation studies indicated that the magnitude of favourable correlation of different characters with grain yield were higher in case of RRS compared to the full-sib system. The presence of negative correlation of plant height with grain yield in both the improved populations indicated the possibility of breakage of unfavourable gene combinations through RRS and full-sibs developed from biparental mating.