A study conducted on five rice varieties indicated that productivity can be enhanced from 17 to 29 % by increasing the number of high density (HD) grains/panicle. HD grains/panicle were partitioned into different components. The number of secondary branches/panicle and the number of HD grains on secondary branches were identified as the most important contributing factors for the number of HD grains/panicle. Selection of plants with panicles having these characters would increase the proportion of HD grains/panicle and productivity.

Genetics of yield components was investigated from combining ability in line X tester design under different environmental conditions in Pearlmillet. The variance due to females was significant for number of tillers, number of grains and single plant yield. The males were also significantly different in respect to panicle length and grain yield. Both additive and non-additive gene effects were important for the expression of yield characters. 732A and PT1650 were best general combiners for grain yield. Best cross combinations for various components were selected and their use in breeding programmes has been advocated..

Pronounced heterosis in grain yield with different degree and magnitude was expressed in various individual crosses under varied environments. Environment-specific as well as widely adaptable hybrids and parents were identified. The hybrid 732 A X PT 1650 was versatile for providing heterotic hybrid in individual as well as a cross the environments. Differential behaviour of heterosis under varying environments and breeding value are discussed.

Genetic variation and diversity studies among 21 genotypes of green gram collected from different sources were carried out for five economic characters. The seed yield and plant height showed moderate genotypic coefficient of variation. High heritability coupled with high genetic advance as percent of mean was noticed for plant height and seed yield. By application of D2 analysis, the twenty one genotypes were grouped into three clusters. The types chosen from the same eco-geographic region were found scattered in different clusters. The clustering together of types from the same eco-geographic region into one cluster was also observed. Maximum inter cluster distance was observed between cluster I and III indicating their divergence in origin.

The P1, P2, F1, F2, BCI and BC2 generations of five pearl millet crosses were studied for six metric traits. The additive dominance model was adequate for plant height, leaf breadth, earhead length and earhead breadth in one cross each. An epistatic digenic model was assumed for other crosses. Heterosis breeding is suggested for improvement of all traits. Duplicate epistasis playes a relatively greater role than complementary epistasis. Among the interactions, dominance X dominance played a major role. Therefore, reciprocal recurrent selection is suggested for development of a superior variety.