MULTIPLE REGRESSION EQUATION AND SELECTION FOR GRAIN AND FODDER YIELD IN SORGHUM VULGARE

Abstract

With a view to formulate an index to aid breeding for high yield in forage sorghum, multiple regression analysis was carried out to estimate the contributions made by eight common independent component characters. The studies in the four environments, showed· the characters: green fodder yield/plant, digestible dry matter/plant and crude protein yield/plant contributed substantially to the determination of dry matter yield/plant. Dry matter yield/plant, leaf/stem ratio and panicle weight/plant were found to be most important characters in the determination of green fodder yield/plant. Dry matter yield/plant exerted highest degree of influence to the determination of digestible dry matter/plant. Seed yield/plant influenced significantly to the determination of panicle weight/plant and vice versa was also true. The multiple correlation coefficients R in each case showed a high degree of goodness of fit as indicated by as high as 72% to 99% of the variability for various dependent characters in each of the fitted regression equations. The genetic upgrading of forage productivity in combination with grain in sorghum requires the judicious combining of the forage yield and forage quality components with seed yield compQnents as indicated by multiple regression equations.