Notes: 1. We analysed the effect of several variables on the morphology of the rotifer Keratella tropica with the aim of accounting for its morphological variability in nature. We used field data from (i) an extensive survey of Argentine lakes, (ii) a 1.5 yr, monthly sampling of Salto Grande Reservoir, and (iii) a 3-month, outdoor experiment in fish culture tanks. In addition, we performed a series of laboratory experiments aimed at isolating the effects of different variables.2. The increase in crustacean biomass and the amelioration of their nutritional status (measured as the reciprocal of the starvation time) caused spine enlargement. In contrast, the presence of decomposers caused reduction of the spines. These results support the idea that a major proportion of the phenotypic expression of K. tropica depends on the concentration of an inducing factor, which in turn depends on the balance between release and decomposition rates of the chemical.3. Food concentration was identified as a secondary factor affecting morphology. In agreement with records for other rotifers, food concentration and spine length were inversely related. Both factors, crustacean chemicals and food availability, seemed to act independently, at different developmental stages. Our experiments suggest, however, that the most exuberant individuals can only be induced through chemicals released by predators and competitors.4. The presence of backswimmers (Notonectidae) resulted in individuals with very short or non-existing posterior spines. Filtrates of backswimmer-conditioned medium also produced a significant, although less striking, spine inhibition. Thus, different chemical factors seem to have opposite effects on K. tropica morphology. Those released by crustaceans induce spine elongation, while those released by backswimmers inhibit spine development.5. Genetic differences were also identified as a source of morphological variability. Under identical laboratory conditions, different clones had spines of different length. Our experiments also suggest an inverse relationship between spine length and growth rate. The latter result, however, is based upon only three clones of K. tropica and is thus preliminary.