Irrigation and rain-out shelters were used to simulate precipitation patterns of wet and dry years in the northern Chihuahuan Desert. Irrigation provided approximately double the long-term average monthly precipitation. Rain was excluded during the wet season, July-October, to simulate a dry year. N net mineralization in laboratory incubations was undectable at calculated water potentials less than -1 MPa. With increasing moisture, mineralization gradually rose to the highest observed rates near field capacity. There was no mineralization maximum at moisture contents below field capacity. Irrigation significantly increased the water potential and rainfall exclusion reduced water potentials to less than -8 MPa. The general absence of important irrigation effects may have resulted from the high natural precipitation during the experiment or because irrigation inputs were insufficient to increase microbial activity during very dry periods. Precipitation exclusion reduced ion capture during the warm-wet season. After allowing precipitation inputs to resume, NH₄-N capture was increased in the cool-dry seasons of both 1987-1988 and 1988-1989. NH₄-N capture more than doubled that predicted from the overall covariance of moisture input and ion capture, suggesting increased availability of N. An unusually hot, dry period in May and June 1989 was followed by a three-to fourfold increase in the warm-wet season N0₃+N0₂-N capture compared to 1988. These data suggest that short droughts of about 3 months in length (both simulated and natural) increased N availability relative to moisture availability.