Abstract As a legume green manure, Chinese milk vetch (MV; Astragalus sinicus L.) has been widely used to reduce nitrogen (N) fertilizer use. However, its effects on soil aggregate stability, nutrient retention, and contribution to soil fertility remain unclear. A 10‐year field experiment evaluated different MV substitution rates to identify the optimal balance between soil fertility and rice yield. The treatments included no fertilizer (CK), 100% chemical fertilizer (N 100 ), and recommended mineral N supply with different proportions of green manure (N 80 G 20 , N 60 G 40 , N 40 G 60 , and N 20 G 80 ). Compared with the N 100 treatment, the N 80 G 20 and N 60 G 40 treatments increased the mean weight diameter and geometric mean diameter of the soil by 4.2%–5.3% and 7.7%–12.1%. The N 60 G 40 treatment significantly increased the carbon pool management index by 45.0%. The N 80 G 20 and N 60 G 40 treatments increased total nitrogen (TN) and total phosphorus (TP) in macro‐aggregates (>0.25 mm) and improved the contribution rates of TN and TP in >2 mm aggregates. Compared with the N 100 treatment, the N 60 G 40 treatment improved TN, TP and total potassium by 6.0%, 9.3%, and 5.6%, respectively. The N 60 G 40 treatment improved the soil quality index most by 34.1%, which was positively correlated with grain yield. Substituting 20%–60% of mineral N with MV can sustain the grain yield, while higher substitution rates reduced it, especially in early rice. Consequently, incorporating MV to substitute 20%–40% N fertilizer enhanced soil structure by increasing the proportion of macro‐aggregates, thereby improving nutrient retention and soil fertility, which offers a sustainable practice for double‐rice systems.
Yuan et al. (2026) studied this question.