Population Breeding Methods in Pearl Millet Improvement (Pennisetum americanum)

Abstract

The population improvement programme at ICRISAT is described, the 2 major methods of selection being half-sib and S1. The former is used on less important composites as it is less resource consuming; the latter employs 2-stage progeny testing. It incorporates selection for resistance to Sclerospora graminicola and Tolyposporium penicillariae. An experiment is described in which 4 methods of selection were compared in a single composite, with up to 6 cycles of selection. Response to selection was poor with all methods, and there were no significant differences between the methods. Two cycles of selection using reciprocal recurrent selection were carried out in 2 pairs of populations. Although the individual performances of the populations improved, there was no improvement in interpopulation performance. Comparisons of composite bulks and varieties derived from them are described. The varieties were formed using a higher selection intensity than that used to produce the composite bulk. Varieties from later cycles of the composites were higher yielding than those from earlier cycles. When single progenies or a number of progenies were used to form varieties, the multi-progeny varieties outyielded the single progeny varieties as they had less inbreeding depression. The optimum strategy for producing varieties is dicussed