The nature of the relationship between productivity and species richness has remained controversial for at least two decades. Recently authors have favored the suggestion that the form of this relationship is highly variable and scale dependent. However, this conclusion is not universally accepted. Here we present the results of a meta-analysis of animal productivity-species-richness relationships (PSRR) in terrestrial and freshwater ecosystems. Initially, 374 separate cases from 273 published studies were identified as potential tests of the animal PSRR. After critically assessing each study, 115 cases were accepted as robust tests of the relationship, and of these 95 had data available for formal meta-analysis. Contrary to expectation, we found no support for the form of the relationship being scale dependent; positive relationships predominated at all scales (geographical extents and grains). Furthermore, positive relationships were the most common form of the animal PSRR in both terrestrial and freshwater ecosystems and among vertebrates, invertebrates, homeotherms and poikilotherms. Therefore, our results also contrast with previous reviews that suggest no particular form of the PSRR is predominant. We demonstrate that the method used for classifying the form of PSRRs is critical to the result and that previous reviews may have been too liberal toward classifying the form of relationships as unimodal. The tendency for positive relationships between productivity and species richness across diverse animal taxa has important implications for understanding the mechanisms behind the latitudinal gradient in species richness.