Abstract

                                Three way hybrids numbering sixty involving six parents were tested in randomized block design with three replications. The mean data on number of mature pods per plant were analyzed as per a triallel analysis model. The character showed predominantly epistatic genetic variance. The magnitude of additive x dominance types of epistasis was maximum followed by dominance x dominance and additive gene effects respectively. The dominance and additive x additive type of gene effects were negative. The combining ability analysis revealed that ALR 2 was a good grand parent as well as third parent and GG 2 good third parent in three-way crosses. The cross combination (JL 24% Co 2) x GG 2 gave the highest three line specific effect. Whereas, the other combinations of the same three parents (JL. 24 x GG2) x Co 2 and (GG 2 x Co 2) x JL 24 exhibited negative estimates. As the epistatic gene action predominated, attempting multiple crosses or intermating of early generation segregants besides postponing the selection to later generation may yield fruitful results for the improvement of pod number in groundnut.

Key words : Groundnut, Triallel. Order effects. Epistasis