Genetic variability in a crop population is important for successful plant breeding. Sixteen cucumber (Cucumis sativus L.) genotypes were evaluated in the early and late planting seasons to estimate the magnitude of their genetic variability and heritability. Genotypes were also classified into groups based on the performance and determination of the highest discriminating trait that accounted for greater variability using cluster analysis and principal component analysis (PCA), respectively. The measured agronomic traits were vine length, number of branches, number of leaves, and leaf area 8 wk after planting, days to flower initiation, days to 50% flowering, number of staminate flowers per plant, number of pistillate flowers per plant, fruit length, fruit girth, fruit weight per plant, number of fruits per plant, mean fruit weight, and total fruit yield ha -1 . A high coefficient of variation was recorded for most traits in both seasons and high variability was found among genotypes. High broad-sense heritability was associated with all the traits in both seasons, except for mean fruit weight in the early planting season and fruit girth, mean fruit weight, and total fruit yield in late planting season. Cluster analysis and its comparison of means showed that 'Beit Alpha', 'Ashely', 'Straight 8', and 'Sumter' from cluster F in the early planting season and 'Beit Alpha' and 'Ashely' from cluster E in the late planting season expressed the best agronomic traits and yield potentials. Hence, selection for any trait would favor genotypes in these clusters. Principal component analysis involved vine length as the most discriminating trait that accounted for greater variability in cucumber in both the early and late planting seasons, and it should be considered in cucumber improvement programs.