It is often assumed that prey species consumed by generalist predators are largely, though not entirely, equivalent in terms of their value to the predators. In contrast to this expectation, laboratory feeding experiments uncovered distinctly varied developmental responses of a generalist predator, the wolf spider Schizocosa , to different experimental diets. Naive Schizocosa attacked and fed upon all the prey species offered; however, highly divergent patterns of survival, development, and growth of Schizocosa spiderlings reared on different single-prey diets revealed a wide spectrum of prey qualities. Spiderlings fed the collembolan Tomocerus bidentatus sustained the highest overall rates of survival, growth, and development. Fruit flies ( Drosophila melanogaster ) were intermediate-quality prey: spiders fed Drosophila initially exhibited rates of survival, growth, and development equal to those of spiders on a diet of T. bidentatus , but after about 3 months, rates declined markedly. Fungus gnats (Sciaridae; Bradysia sp.) and conspecfic spiderlings are low-quality prey for Schizocosa : a sole diet of either of these prey types resulted in positive but markedly submaximal rates of growth, retarded rates of development, and survival rates much lower than that supported by a diet of Drosophila . Worst were the collembolans Folsomia candida and Isotoma trispinata , and the aphid Aphis nerii : spiderlings fed solely one of these species did not grow and died without molting. A. nerii is classified as poor quality because survival was no better than that of starved controls. F. candida and I. trispinata were toxic: survival of Schizocosa hatchlings fed these collembolans was lower than that of starved controls. A mixed diet of T. bidentatus and fruit flies yielded positive synergistic effects with respect to growth, but development and rate of survival were similar to those of spiders on a sole diet of T. bidentatus . Including toxic prey did not produce a better diet, while inclusion of toxic prey with prey of higher quality created diets that were no better than the toxic prey alone. The results of these experiments suggest that prey species that are similar in morphology and behavior, and that are initially killed and consumed, may differ dramatically in their suitability as food for generalist arthropod predators.