Lessons from the sandbox: Is unexplained nitrogen real?

In their review of 24 studies of forest nitrogen (N) budgets, Binkley and others (2000) found that only one of them supported the conclusion that an N accumulation of more than 25 kg N ha(-1) y(-1) is possible without known symbiotic N-2-fixing plants. They contended that, given how well the N cycle is known, new N accumulation pathways are unlikely. They also concluded that the Hubbard Brook sandbox study (Bormann and others 1993) was insufficiently replicated and had low precision in vegetation and soil estimates. Here we reevaluate and extend the sandbox analysis and place the findings in a broader context. Using multiple methods of estimating vegetation N accumulation in pine sandboxes, we arrived at results that differed from the reported rates but still strongly supported large biomass N accumulation. The original study's conclusions about soil N changes were strengthened when new evidence showed that N accumulated in lower horizons and that the sandboxes were successfully homogenized at the beginning of the experiment. Unexplained ecosystem N accumulation ranged from about 40 to 150 kg ha(-1) y(-1), with 95% confidence intervals that did not include zero. No evidence was found that could balance the sandbox ecosystem N budgets without adding unexplained N. Unreplicated experiments, such as the sandboxes, can explore the possibility that N can accumulate in ways not explainable by mass balance analysis, but they cannot quantify the frequency and extent of the phenomenon. New studies should combine substantive microbiological, mass balance, and process research using multiple direct measures of N-2 fixation.