By means of 15N tracer technique the fate of N in ammonium sulfate, milk vetch, sesbania and azolla, and the availability of their residual N were studied in a microplot experiment. Results showed that a) at the end of the first crop of early rice, both plant recovery and loss of N from ammonium sulfate were the highest whereas those from azolla were the lowest with those from milk vetch and sesbania in between; the sequence was reversed in terms of recovery of N in soil; the net residual N from ammonium sulfate was very low, about 1 / 7-1 / 4 of that from green manures, indicating that chemical N fertilizer contributes little to the soil N reserve; b) plant recovery of the residual N was low and it did not always decrease with time; the total plant recovery (from the second to the fifth crops) of the residual N from various test fertilizers was only 8-11 % of the total N originally applied; c) the plant recovery of the residual N from ammonium sulfate was the highest, followed by those from milk vetch Results showed that a) at the end of the first crop of early rice , both plant recovery and loss of N from ammonium sulfate were the highest alternately those from azolla were the lowest with those from milk vetch and sesbania in between; the sequence was not found in N in soil; the net residual N from ammonium sulfate was very low, about 1 / 7-1 / 4 of that from green manures, indicating that chemical N fertilizer contributes little to the soil N reserve; b) plant recovery of the residual N was low and it did not always decrease with time ; the total plant recovery (from the second to the fifth crops) of the residual N from various test fertilizers was only 8-11% of the total N initially applied; c) the plant recovery of the residual N from ammonium sulfate was the highest , followed by thos e from milk vetch