The effects of selection for prolificacy in two composite populations of Zea mays L.

The past excellent work done on maize in Kenya necessi­ tated continued effort to be put into the development of this important food crop.The comprehensive breeding system, used to improve the main maize breeding populations in Kenya, has proved to be acpowerful tool. The main objective of this study was to find out the effect of breeding for prolificacy in KCE and KCB, the two broad-based maize populations in the Kitale applied maize breeding programme. It was expected, as has been ypothesized by many maize breeders, that the subsequently changed populations would not only be more prolific but, as a result, also bring about stability in grain yield in different growing environments. Three cycles of full-sib selection for prolificacy were completed for both KCE and KCB. The response to selection in terms of prolificacy was rapid in both populations. Evaluation of progress made under selection indicated that the KCB improved from about 9.2% in the KCB Co to 32% prolificacy in the KCB(F)C3. The KCE Co recorded about 20.7% against 30.5% for the KCE(F)C3. These evaluations were done in twenty-four environments "lith large differences in potential. The observed increases in the frequency of this trait was accompanied by similar increases in grain yield potential from the Co to C3. The yield increase was a result of improved % crop index in the selected populations. Similarly, the response of % crop index increased from Co to G3° Physiological changes in the populations developed by this method of selection will need further research. The number of grains produced per plant, as expected, increased with increasing fre­quency of prolificacy but, these grains were smaller than those from the predominantly one ear original populations. The effect of this type of selection on frequency of barrenness was small. In both KGE and KGB, contrary to the common hJ~othesis that pro­lificacy should markedly reduce barrenness, the reduction in the proportion of barren plants was insignificant. This obeervation led to the suggestion that barrenness was independently inherited and that effective selection against it would have to be done on a different score. It "vas important to consider the relative merit of this. type of indirect selection and direct selection in effecting progress in grain yield. Estimates for percent gain were compared and these indicated that selection for prolificacy in these two composites was just as effective in bringing about improvement in yield as direct full-sib selection for grain yield. It must be pointed out, however, that the estimates of percent gain were per year rather than per cycle. On this basis, full-sib selection for prolificacy had the advantage that a cycle was completed in a year. The full-sib selection for grain yield required two years for a complete cycle. Environmental stability parameters were computed for the different genotypes evaluated in the different groups of environ­ ments. The stability parameters considered for each genotype were:­ the mean yield of each genotype in all environmentsthe coefficient of linear regression of the mean genotype yields on the environme­ntal means or indicesand the deviations from the linear regression. These stability parameters showed that the selected genotypes maintained their high yields above the original popula­tions. The corresponding linear regression coefficients on the environmental indices were higher for the selected. the third stability parameter, the deviation from linear regression, was lower for the selected genotypes than for the original populations. This implied better environmental stability for the former. The KCB(F)C3 and KCE(F)C3 developed in this study have higher potential yield in favourable environments than the original populations from which they were derived. Other workers have shown that maize Hith a different type of stability could be developed in lower yield potential environments. It is suggested that further breeding in the two new popula­tions should be by reciprocal recurrent selection (RRS) and reciprocal testing (RT). The latter method should take advantage of the higher frequency of prolificacy in the two new populations. Both methods of selection should be able to exploit any type of gene action that may be available in the two populations.