Genetic variability, correlation studies and path coefficient analysis in bittere gourd (Momordica charantia L.)

Abstract

Studies were undertaken with 25 diverse bitter gourd types in the Department of Horticulture (Olericulture), College of Horticulture, Kerala Agricultural University, Vellanikkara, Trichur, to estimate the genetic variability, correlation coefficients and path coefficients, and also to formulate a reliable selection index for the crop, during 1977-78. The results have shown that the differences between types were highly significant for all the twenty one characters studied. The estimates of variance components and coefficients of variation have indicated that the major portion of total variability in all characters, except number of seeds per fruit, was due to genetic causes. Heritability in the broad sense was found quite high for all characters except number of seeds per fruit. Expected genetic advance has shown that, by selecting five per cent superior plants from the available population, yield could be improved by 4.08 kg per plant. Characters such as yield per plant, vitamin C content, number of fruits per plant, number of female flowers per plant, iron content and phosphorus content which exhibited parallelism in the high estimates of heritability and genetic gain may be suggested to be due to the action of additive gene and can be straightly improved through selection. Type B.G.23 was found to be exceptionally high yielding. Yield per plant was found to be highly correlated with length of main vine, weight of fruit, length of fruit, number of fruits per plant, number of female flowers per plant and number of primary branches per plant. The correlation coefficients among these yield components were also highly significant. Path coefficient analysis has shown that the weight of fruit, number of fruits per plant and length of main vine had high direct positive effects on yield. Number of primary branches per plant, number of female flowers per plant and length of fruit exhibited low and negative direct effects on yield. Multiple regression analysis resulted in fitting a function, indicating that, about 74 per cent of the fruit yield in bitter gourd could be predicted phenotypically, if the observed value for length of main vine is multiplied by 0.8990, number of fruits per plant by 0.0492 and weight of fruit by 0.0335 and from their sum of products 7.2168 is subtracted.