Comment on ‘Kane et al . 2008. Precipitation control over inorganic nitrogen import‐export budgets across watersheds: a synthesis of long‐term ecological research. Ecohydrology 1(2): 105–117’

A recent paper in Ecohydrology (Kane et al., vol. 1, pp. 105–117) made a first attempt to synthesize and compare a large body of watershed biogeochemical data from diverse Long-Term Ecological Research (LTER) sites, which is certainly an impressive exercise. These authors correctly point out that while there is large body of work exploring the ways that disturbance and elevated N influence the relationships between N deposition and watershed dissolved inorganic N (DIN) retention, there have been a few studies exploring climate and precipitation effects on N losses from watersheds using a cross-site approach. However, there appear to be some flaws in the logic and data analysis that we would like to point out, not as a critique of this exercise per se, but to raise further avenues for research and to raise new questions that perhaps could be answered by different analyses or more focused research. Our main concerns are over inferences made from the graphs in the paper. The problem, of course, with a cross-site, observational data set is that it is not experimental. In other words, no factors could be held constant while variables in question could be varied, so the sample may not be representative of the full range of variation in presumed causal factors, and may even coincide with other, underlying factors not included in the analysis. In this case, simple geographical patterns in N deposition unfortunately confound most of the interesting climatological inferences that were made. The issues of interpretation arise from the fact that much higher N deposition occurs in the eastern compared with the western United States or Puerto Rico. Also, N retention calculated as the difference between inputs and outputs