We propose to assemble some of the most detailed and complete food web data yet collected, and use these data to evaluate, extend, and integrate food web theory. Data will be collected from several of the most thoroughly studied ecosystems in the world including Lake Tahoe, Castle Lake, and Clear Lake in California, and Mirror Lake in New Hampshire. In constructing food webs for these lakes we will emphasize complete species lists (several webs will include over 500 species), even resolution of producers, consumers and decomposers, and estimates of the dominance of species and strengths of trophic interactions. With these data we will evaluate the effects of scale (extent and resolution of the data) on food web structure, extend food web theory to more precisely account for real food webs (i.e., food webs with at least an order of magnitude more species than those in current food web collections) and integrate food web theory by delineating the complementarily between binary descriptions and functionally prioritized descriptions of food webs. We will examine major food web patterns including connectance, links per species, chain length, and proportions of top, intermediate and basal species. The data will also be used specifically to test several pertinent hypotheses about food web structure including the intermediate trophic state hypothesis, the constant connectance hypothesis, and previously described effects of scale and aggregation on food web structure. I Larger goals of this proposal include improving procedures for constructing food webs and integrating food web theory by assembling data using both structural (binary) and functional approaches for the same systems. These studies will help stimulate the clearly needed interaction and integration of theory and empiricism for development of this vital subdiscipline of ecology.
