Creel survey and cove rotenone data from 17 reservoirs in Piedmont North and South Carolina were subjected to correlation and regression analysis to examine environmental factors influencing fish production and taxonomic composition and to develop models for the prediction of total harvest and biomass. Summer surface chlorophyll concentration was the best univariate predictor of total harvest (r2 = 0.93) and one of the best predictors of total biomass (r2 = 0.79). The morphoedaphic index was a good predictor of both biomass (r2 = 0.79) and harvest (r1 = 0.83), but explained no more variation than conductivity alone. Mean depth was not a strong predictor of biomass (r2 = 0.55) or harvest (r2 = 0.61), and total phosphorus concentration was not significantly correlated with biomass or harvest. Multivariate models based on phosphorus loading and reservoir morphometry/hydrology explained up to 91% of variation in biomass and 92% of variation in harvest. Harvest composition varied along gradients of morphometry, hydrology, and productivity. Percents of harvest consisting of largemouth bass {Micropterus salmoides) and temperate basses (Percichthyidae) increased with reservoir retention time, while percent of harvest consisting of crappie (Pomoxis spp.) declined. The combined harvest of striped bass (Morone saxatilis) and hybrid bass (M. saxatilis x M. chrysops) declined as chlorophyll increased, possibly reflecting a loss of cool, oxygenated habitat with increasing eutrophication. Biomass composition did not vary significantly along environmental gradients, with the exception that percent of biomass consisting of common carp (Cyprinus carpio) declined with increasing reservoir fertility. No evidence was observed of shifts in biomass composition to undesirable species as productivity increased. The absolute abundance of most major taxa increased with increasing reservoir fertility, with the exception of percichthyids and cyprinids. Habitat preferences were evident in the negative correlations of crappie harvest and catfish biomass with mean depth and retention time.
