Multiple-site and-year analyses of stream insect emergence: a test of ecological theory.

The use of existing data sets to test applicability of existing ecological theory is an uncommon but potentially cost-effective approach for exploitation of previously accumulated knowledge. Studies on the emergence of insects from small streams have been a major research topic in aquatic ecology, particularly in Austria and Germany; the availability of emergence data from these two countries, covering over 1 million identified specimens, from 18 sites, and for 32 collection years is an example of such exploitable information. Concurrent estimates of annual emergence biomass and annual benthic secondary production for 18 aquatic insect populations showed a statistically significant relationship, contradicting the premise that emergence data do not provide any quantitative measure for a given stream area. Therefore, the emergence data were examined to test various predictions from ecological theory. Observed richness of emerging species of three orders of lotic insects - the Ephemeroptera, Plecoptera, and Trichoptera (EPT) - over 15 years at one site did not agree with predictions based on either flow predictability or change in flow and the "habitat templet concept". Trends in observed richness of emerging EPT species over 1 year at 18 sites agreed weakly with predictions using either pH values or the annual temperature amplitude and the "intermediate disturbance hypothesis", or using either annual temperature amplitude or total biomass of EPT emergence and the "disturbance-productivity-diversity model". A prediction of the "river continuum concept" that abundance of the shredder functional-feeding group should decrease and that of grazers should increase along a dense- to open-canopy gradient was not consistently supported by the emergence biomass data. For shredders and grazers of all insects (12 sites) and EPT (18 sites), this trend was apparent (but not significant) only if sites with intermediate canopy density were omitted. We identified three critical elements in our study that generally interfere with such tests of these theoretical constructs: (i) species richness is a poor measure of resource limitation and/or utilization theories; (ii) restrictions of the taxonomic operational window (in our example usually to EPT) causes problems in extrapolation to a larger system (in our example to all insects); and (iii) historical constraints may affect the local result of tests of resource limitation and/or utilization theories simply because species that potentially interact are lacking in the region under examination. Problems notwithstanding, the use of existing data sets to test applicability of currently held ecological theories is a cost-effective and amenable approach for use in a variety of research topics in stream and general ecology. In this context, future tests should focus on: (i) measures that are more robust than just species richness, e.g., on measures commonly used to assign species to strategies such as r, K, or A; (ii) a variety of taxonomic groups; and (iii) gradients in historical constraints on current regional species composition.