An ongoing roof experiment, where N and acid inputs were reduced to the recommended critical load levels, has been conducted since 199 1 in an N-saturated spruce stand in Solling, Germany. Our study was aimed at (1) quantifying the changes in gross rates of microbial N cycling under ambient and reduced N conditions, and (2) relating the soil N dynamics to the changes in N leaching and N status of trees. Two roofs were used, one to achieve "ambient" and the other reduced ("clean rain") inputs, with a roofless plot as a control for possible roof effects. In 2001, the ambient roof and ambient no-roof plots showed an apparent decrease in gross N mineralization rates and significantly lower microbial NH4+ immobilization rates and turnover rates of NH4+ and microbial N pools. The microbial NO3- immobilization rates and NO3- pool turnover rates were lower than the microbial NH4+ immobilization rates and NH4+ pool turnover rates, showing that less NO,cycled through microorganisms than NH4+. There was also low abiotic NO3- immobilization. High NO3- input from throughfall and low microbial turnover rates of the NO3- Pool, combined with low abiotic NO3- retention, may have contributed to the high NO3- leaching losses in these ambient plots. The clean rain plot showed a slight increase in gross N mineralization rates and significantly higher microbial NH4+ immobilization rates and turnover rates of NH4+ and microbial N pools. Neither nitrification nor soil NO3- was detectable. There was an increase in abiotic NO3- immobilization. Foliar N concentration had decreased but was still adequate. An efficient cycling of NH4+ through microorganisms, combined with the high abiotic NO3- immobilization, indicated efficient mineral N retention in the clean rain plot. These results indicated that long-term reduction of throughfall N and acid inputs had induced high but tightly coupled microbial NH4+ cycling and an increase in abiotic NO3- retention, which contributed to the reversal of N saturation.