Seventy-five high-elevation lakes in the Sierra Nevada mountains of California were sampled for microcrustacean species and major ion compositions. Cluster analysis was used to delineate five recurrent community types. Distributions of both individual species and of community types were related to chemical, morphometric, topographic and geologic variables, as well as fish presence or absence, by stepwise logistic multiple regression. Fish distribution was an important predictor of the distributions of all individual species for which significant regression models could be built; other common predictors included nitrate concentration, elevation, basin area and lake depth. One common community type (n = 22) consisted of Dapbnia rosea, Diaptomus signicauda, Bosmina longirostris and Holopedium gibberum. A second, more species-rich, community (n = 27) included many of these same species, in addition toCyclops vemalis, Diaphanasoma brachyurum, Polyphemus pediculis and/or Ceriodaphnia affinis. Two further communities (n = 9 and n = 6 respectively) contained Daphnia middendorffiana with either Diaptomus shoshone or Diaptomus eiseni. A fifth community type (n = 11) either lacked microcrustaceans or contained only Chydorus and/or Alona spp. The distributions of the first four of these community types could be well predicted on the basis of fish presence and elevation (community types I and II), or fish absence and lake depth (community types III and IV). The distribution of the fifth community type was independent of fish presence or absence; phosphate concentration was the only significant predictor of this community.