Relative efficiency of different selection indices for seed yield in pigeonpea (Cajanus cajan L.)

Different selection indices were constructed in 96 germplasm accessions of pigeonpea (Cajanus cajan L.) on the basis of multiple regression analysis and discriminant functions to determine the relative importance of various component characters. Genetic advance was used to compare the efficiency of different selection indices. The multiple regression equation involving all the characters, viz. days to flowering, duration of flowering, days to maturity, plant height, primary branches, secondary branches, pods per plant, seeds per pod, 100-seed weight, protein content and harvest index indicated that 74% of variation in seed yield were explained by these characters. All the partial regression coefficients except for flowering duration, protein content, seeds per pod and primary branches were significant and the degree of determination was not affected when multiple regression equation was fitted with characters having significant partial regression coefficients. Different discriminant functions fitted for the characters showing significant partial regression on seed yield showed that secondary branches, followed byseed yield, pods per plant, harvest index, days to maturity, plant height, and 100-seed weight were the important characters. The best character combination was of secondary branches and seed yield, which was 137% more efficient than straight selection for seed yield. The efficiency increased with every additional character. In the present investigation the index involving four characters, viz. secondary branches, seed yield, pods per plant and harvest index, was effective and efficient.