For extant enzyme activity and net N-mineralization data, we used a two-way analysis of variance to test for the main effects of watershed (long-term N-enriched watershed versus reference watershed) and forest type, and the interaction of
these effects on independent variables.