Resistance of the size structure of the fish community to ecological perturbations in a large river ecosystem BRENT A. MURRY* AND JOHN M. FARRELL † *Biology Department, Institute for Great Lakes Research, Central Michigan University, Mount Pleasant, MI, U.S.A. † Department of Environmental and Forest Biology, College of Environmental Science and Forestry, State University of New York, Syracuse, NY, U.S.A. SUMMARY 1. Frequency distributions of log-scaled abundance and body size for aquatic biota, including fish, yield size spectra that provide gross estimates of food-web capacity (number of individuals supported by the food web indexed by the midpoint height or y-intercept) and ecological efficiency (rate of decreasing organismal abundance with increasing body size as indicated by the size-spectra slope). 2. Size spectra are known to respond in predictable fashion to exploitation by commercial (marine) fisheries, but there has been relatively little research of size spectra in freshwater systems and less evaluation of how community size-structure responds to the wider array of ecological perturbations prevalent in freshwater systems. 3. We used a 29-year experimental gillnet data set to test the responses of food-web capacity and eco- logical efficiency (as indexed by size spectra) of a large river fish community to variations in preda- tory demand (population expansion of double-crested cormorant, Phalacrocorax auritus) and perturbations affecting primary production (reductions in total phosphorus and warming summer water temperature). 4. Fish community capacity (size-spectra elevation) and ecological efficiency (size-spectra slope) showed low interannual variation (coefficients of variation 5.6 and 15.4%, respectively). Though highly stable, food-web capacity decreased over time concurrent with declining total phosphorus concentrations and increasing cormorant abundance. Interannual variation in ecological efficiency was not related to the measured perturbations. 5. Relative abundance of the core species that comprised 95% of the community remained relatively consistent over time. The fish community size-structure was also stable and resistant to ecological perturbation. However, we found measurable changes in the composition of secondary species, and these were associated with an increase in cormorant abundance and decrease in total phosphorus. The lack of change in core community species representation and in overall size structure suggest community-level compensatory mechanisms were operating. Keywords: body size, capacity, ecological efficiency, ecological stressors, size spectra Introduction Aquatic food webs are highly size-structured owing to individual metabolic constraints associated with body size (Peters, 1983; Brown et al., 2004), mortality induced by gape-limited predation (Arim et al., 2010), competi- tion among individuals for limited resources and energy loss through successive trophic transfers (Elton, 1927; Lindeman, 1942). Size spectra [log abundance (or nor- malised biomass) regressed against log body size] pro- vide a simple approach for assessing community body size distributions and a generalised depiction of food- web structure and stability, as well as food-web capacity (sensu Brown, 1981) and ecological efficiency (White et al., 2007; Petchy & Belgrano, 2010). Size-spectra theory (Sheldon, Prakash & Sutcliffe, 1972; Kerr & Dickie, 2001) Correspondence: Brent A. Murry, Caribbean Landscape Conservation Cooperative, International Institute for Tropical Forestry, San Juan, PR 00926, U.S.A. E-mails: brent_murry@fws.gov, murry1ba@cmich.edu © 2013 John Wiley & Sons Ltd 155 Freshwater Biology (2014) 59, 155–167 doi:10.1111/fwb.12255