Although sponges contain a diverse array of secondary metabolites, often in high concentrations, little is known about the natural function(s) of these metabolites. We tested the effects on feeding of several sponge secondary metabolites by incorporating them into high-quality artificial diets and offering these to three groups of potential consumers in the laboratory and field. Compounds tested from sponges in the genus Dysidea, a brominated diphenyl ether and a mixture of avarol and isoavarol, usually deterred feeding. Luffariellolide and heteronemin, both from sponges in the genus Hyrtios, usually had no effect or stimulated feeding, while chondrillin, from a sponge in the genus Plakortis, had variable effects. Pufferfish were deterred from feeding by both Dysidea spp. secondary metabolites, and crabs by the brominated diphenyl ether, but natural assemblages of reef fish were deterred by none of the sponge metabolites in high-quality diets. However, when offered metabolites in artificial foods of lower quality, reef fish in the field were deterred from feeding by both Dysidea spp. metabolites. These results indicate that sponge secondary metabolites differ considerably in their effectiveness as deterrents to feeding, and that different consumers differ in their sensitivity to sponge secondary metabolites. Our results also corroborate previous findings that the effect of a given secondary metabolite can vary depending upon the quality of the artificial diet or, in nature, the nutritional quality of the organism containing it.