The feedback effect of plant residues on soil organic matter is of current interest in many agricultural systems due to recent large increases in crop yields.Computer-based numerical methods which fit differential equations to data allow the plant yield-soil organic matter relations in crop sequential experiments to be examined more fully.The equationdN/dt = -K1(t)*N+K2+K3(t)*Y(t) whereNis soil organic nitrogen,Y(t) is plant yield at timet, K2is a constant noncrop addition term, andK1(t) andK3(t) are decomposition and addition crop coefficients which vary with the stage of the sequential experiment, has been fitted to soil and plant yield data from the Morrow and Sanborn plots. Coefficients of the above model and three restricted variants of it were estimated by least squares, using soil nitrogen and crop yield data over 50 to 63 years. Statistical comparisons of the models were also made.From the equations the relationships between equilibrium soil levels and crop yields were calculated and the feedback effect of increasing crop yield on soil nitrogen level was estimated.At Morrow, increasing corn yields in a continuous corn system had negligible feedback effects on soil N level. However, some feedback effects were shown by oats and clover.At Sanborn, oats had the greatest feedback effect of the four crops examined. In this experiment timothy had the least feedback effect.Manure in both experiments substantially increased soil equilibrium levels.The value of quantitative estimates of feedback effects and the need for additional soil data in understanding changes in agricultural ecosystems are discussed.