Thirty diverse genotypes of cucumber (Cucumis sativus L.) collected from different indigenous sources were characterized with respect to economically important traits by using principal component and regression analyses during kharif 2009-10. The effect and contribution of each character on fruit yield per plot was measured. Principal component analysis characterized the genotypes into four principal components based on their total variation (83.72%). The first principal component accounted for more than 48% of the total variation and was the combination of number of marketable fruits per plant, fruit length, harvest duration, total soluble solids, seed germination, seed vigour index-I and II and yield per plot. The second, third and fourth principle components contributed only 15.27%, 13.50% and 6.72% of total variations, respectively. To quantify the importance of each variable in predicting average fruit weight and yield per plot, multiple linear regression models were developed. Model-I indicated that average fruit weight can be predicted satisfactorily on the basis of number of marketable fruits per plant, fruit length and breadth, while, Model-II indicated that yield per plot can be best predicted by with the help of number of marketable fruits per plant, fruit length, average fruit weight, harvest duration, seed germination, seed vigour index-II and severity of powdery mildew and anthracnose. Therefore, on the basis of information on genetic diversity through principal component and regression analyses, suitable selection strategy can be formulated for getting higher yield in cucumber.
Read full abstract