Biodiversity embraces multiple facets of the variability of nature, although most research has dealt separately with population-, species- and assemblage-level measures. This review concentrates on aquatic insect biodiversity and the assemblage-level measures, such as species richness, assemblage compositional variation, taxonomic distinctness and functional diversity. Most studies on aquatic insects have considered biodiversity patterns based on surveys of local assemblages along geographical and environmental gradients, while there is a virtual lack of studies that have considered regional grain sizes (i.e. the size of the observational unit). Latitudinal gradients at both regional and local grain are ambiguous in aquatic insects, as different studies have found either higher or lower local diversity in the tropics than in the temperate zone. Other geographical patterns in aquatic insect diversity may also be relatively weak, as suggested by subtle among-ecoregion differences in both local species richness and assemblage composition. An ecological explanation for the absence of strong geographical gradients is that local environmental features may not necessarily vary with geography, and these factors may override the influences of historical and climatic influences on local diversity. Evidence from large-scale studies suggests, however, that not only various local habitat and ecosystem variables, but also those measured at the watershed, regional and geographical scales are needed to account for variation in the species richness and assemblage composition of aquatic insects. Within regional species pools, species richness and assemblage composition in both lotic and lentic ecosystems vary most strongly along gradients in habitat size and acidity. Knowing how α-, β- and γdiversity of aquatic insects vary along geographical and environmental gradients has important implications for the conservation of biodiversity in freshwater ecosystems. Such knowledge is not yet well-developed, and two aspects should be considered in future research. First, further survey research on lotic and lentic ecosystems is necessary for improving our understanding of general biodiversity patterns. Second, given that freshwater ecosystems are facing a severe biodiversity crisis, the implementation of representative networks of freshwater protected areas would certainly benefit from increased understanding of patterns in aquatic insect biodiversity.