Estimating species richness in large biomes is a central challenge in ecology and conservation biology. However, accurate census data is often available only from small discrete plots distributed within the biome. Using tree species richness data collected from 48 plots (0.25 ha each) widely distributed through 60 000 km2 in the rainforests of the Western Ghats of southern India, we test the application of a proposed method for estimating species richness at large scales from measured species commonalities between pairs of censused plots. We show that the method allows extrapolation of species richness from a scale of 0.25 ha plots to that of the entire biome, or 105 km2.