Division of Genetics, Indian Agricultural Reserach Institute, New Delhi-12
**Present address: U.P. Institute of Agricultural Sciences, Kanpur-2.
*Part of a thesis submitted by Ziauddin Ahmad in partial fulfilment of the requirement of the Ph.D. degree of the Indian Agricultural Research Institute, New Delhi.
A partial diallel analysis (s= 11) in the F1, F2 and F3 generations of crosses involving 22 diverse dwarf derivatives of pearl millet was evaluated for the quality component, crude protein, with reference to the nature of gene action and combining ability.
The range of variation for crude protein was found to be much higher in the hybrids (11.10 to 20.60%) than in the parents (12.51 to 18.43%), although the parents had much higher protein content than many of the hybrids in each generation. The differences between overall means of the parents and hybrids were negligible in all’ the three generations. Mean sums of squares between parents vs. hybrids were not significant in the F1, F2 and F3 generations which indicate that hybrids did not show heterosis.
The genetic analysis for crude protein in each generation has revealed substantial variability among the parents and among the crosses. Predominantly non-additive gene action was observed for this character in the F1, F2, and F3. However, substantial additive component was also available for this trait in the successive generations. The relative ranking of parents on the basis of mean performance and gca effects were same in all the three generations of partial diallel. Five parents were found to have desirable gca effects with high means for crude protein. These parents were found to be of diverse origin involving mostly D2, D4, and IP. 81 in their parentage. Productivity had no barrier for crude protein indicating that desirable recombinants could be obtained with respect to protein content and plant type.
The predominance of non-additive gene action for crude protein indicated that the diversity of alleles rather than heterozygosity per se is important in bringing about adaptive changes. It is, thus, suggested that cyclic inbreeding and selective combination of lines to form gene pools appears to be more effective.
Since five parents are superior in each partial diallel set for crude protein, a composite of these lines or an intermating population involving all possible crosses between them which will be subjected to biparental mating is recommended. Selection for crude protein content alongwith other quality components in such a composite population will improve the quality and quantity of protein substantially.
