In this review, we address the extent to which stream ecology has contributed to biodiversity-ecosystem function (BEF) theory and empirical testing. BEF research first targeted the implications of the ongoing loss of biodiversity for ecosystems and humans. Terrestrial ecology has played a leading role in this field, whereas the contribution of riverine science to the debate has been more limited. Nevertheless, a considerable merit of stream ecology has been to consider a wide range of ecological groups (riparian litter producers, aquatic micro-fungi, macroinvertebrates, and fishes). Through a meta-analysis of these unique data, we show that the relative importance of species number versus assemblage composition increases as we go towards higher trophic levels. Whether this pattern is general or specific to stream ecosystems needs to be evaluated through cross-ecosystem comparisons looking more closely at mechanistic processes. It is evident from stream studies that trophic and non-trophic (e.g. facilitation) interactions govern the functional consequences of biodiversity. These studies also indicate that richness-function relationships are altered by a multitude of factors, such as evenness, non-taxonomic diversity (genetic/phenotypic diversity), species extinction order, the environmental context, as well as experimental setups. This review highlights the relevance of stream ecology to BEF research. Copyright © 2009 John Wiley & Sons, Ltd.