Path analysis of the correlation between fruit number and plant traits of cucumber populations

The relationships between fruit yield and yield components in several cucumber (Cucumis sativus L.) populations were investigated as well as how those relationships changed with selection for improved fruit yield. In addition, the correlations between fruit yield and yield components were partitioned into partial regression coefficients (path coefficients and indirect effects). Eight genetically distinct pickling and slicing cucumber populations, differing in fruit yield and quality, were previously subjected to modified half-sib family recurrent selection. Eight families from three selection cycles (early, intermediate, late) of each population were evaluated for yield components and fruit number per plant in four replications in each of two testing methods, seasons, and years. Since no statistical test for comparing the magnitudes of two correlations was available, a correlation (r) of 0.7 to 1.0 or -0.7 to -1.0 (r2 greater than or equal to 0.49) was considered strong, while a correlation of -0.69 to 0.69 was considered weak. The number of branches per plant had a direct positive effect on, and was correlated (r = 0.7) with the number of total fruit per plant over all populations, cycles, seasons, years, plant densities, and replications. The number of nodes per branch, the percentage of pistillate nodes, and the percentage of fruit set were less correlated (r < [0.7]) with total fruit number per plant (fruit yield) than the number of branches per plant. Weak correlations between yield components and fruit yield often resulted from weak correlations among yield components. The correlations among fruit number traits were generally strong and positive (r greater than or equal to 0.7). Recurrent selection for improved fruit number per plant maintained weak path coefficients and correlations between yield components and total fruit number per plant. Selection also maintained weak correlations among yield components. However, the correlations and path coefficients of branch number per plant on the total fruit number became more positive (r = 0.67, 0.75, and 0.82 for early, intermediate, and late cycles, respectively) with selection. Future breeding should focus on selecting for the number of branches per plant to improve total fruit number per plant.