Correlation and regression analysis of trait and yield of soybean

Correlation and regression analyses were performed based on variations in 17 morphological traits and yield obtained from field experiments on three soybean varieties with four fertilization modes. Results show that eight morphological traits including plant height (x₁), branch number (x₅), uni-grain pod (x₉), di-grain pod (x₁₀), tri-grain pod (x₁₁), tetra-grain pod (x₁₂), 100-grain weight (x₁₅) and economic coefficient (x₁₆) have significant impact on yield with relative effects (from standard regression coefficient) in the following order: x₁₁＞x₁₂＞x₉＞x₁₀＞x₁₆＞x₁₅＞x₁＞x₅. Enhancing effective pod number is an important first step for improving soybean yield, followed by improving economic coefficient. In comparison, increasing 100-grain weight and plant height are less significant for improving yield. Selection of traits with high biomass, more trigrain pod, tetragrain pod and low podding node site should be emphasized in breeding and developing regulation measures to enhance soybean biomass. Descending podding node site should also be adopted in soybean cultivation.