Comparison of selection indices using different weights for biometrical characters in bajra crop

  • A. D. Kalola, D. J. Parmar, G. N. Motk and P. R. Vaishnav
Keywords: selection index, genotype, genotypic correlation, phenotypic correlation


The data on different biometrical characters of large scale varietal trial having 21 genotypes of bajra were collected from regional research station, Anand for present study. The experiment was conduct in a Randomized Block Design with three replications. The characters for grain yield (kg/plot) (X1), 1000 seed weight (gm) (X2), ear head length (cm) (X3), ear head girth (cm) (X4) and effective tillers/plant (X5) were used for this analysis. The selection index technique was employed to study the crop improvement using different characters giving different weights to each character. Selection index for individual character, combination of two characters, three characters and so on can be calculated and the combination which provides the higher relative efficiency is selected. Selection index for single or combination of characters which provides the high per cent relative efficiency (PRE) is selected. The genotypic correlation and phenotypic correlation were worked out between  yield and yield contributing characters. Selection indices were worked out taking five biometrical characters and were constructed taking all possible combinations of the characters. Total 31 selection indices were constructed using equal weight (W1) for all characters. The genotypic correlation coefficients (W2), phenotypic correlation coefficients (W3), direct effects of biometrical characters based on genotypic (W4) and phenotypic correlations (W5) as weight for different characters. The expected genetic gain and PRE of different indices were calculated. The results indicated that in general PRE with equal weight was higher than rest of the weights in all combinations of indices. The rank correlation among different methods for selection score values were positive and highly significant indicating agreement with all methods except equal weight. Therefore index based genotypic correlation taken as weight is considered better than all other indices.

Author Biography

A. D. Kalola, D. J. Parmar, G. N. Motk and P. R. Vaishnav

Associate Professor

Department of Agril. Statistics

AAU, Anand

Research Article