We compared variation in butterfly communities across 3 years at six different habitats in a temper-ate ecosystem near Boulder, Colorado, USA. These habitats were classified by the local Open Space consortium as Grasslands, Tallgrass, Foothills Grasslands, Foothills Riparian, Plains Riparian, and Montane Woodland. Rainfall and temperature varied considerably during these years. We surveyed butterflies using the Pollard-Yates method of invertebrate sampling and compared abundance, spe-cies richness, and diversity across habitats and years. Communities were most influenced by habitat, with all three quantitative measures varying significantly across habitats but only two measures show-ing variation across years. Among habitats, butterfly abundance was higher in Plains Riparian sites than in Montane Woodland or Grassland sites, though diversity was lowest in Plains Riparian areas. Butterfly species richness was higher in Foothills Riparian sites than it was in all but one other habitat (Tallgrass). Among years, butterfly abundance and species richness were lower during the year of least rainfall and highest temperatures, suggesting a substantial impact of the hot, dry conditions. Across habitats and years, butterfly abundance was consistently high at Plains Riparian and Foothills Riparian sites, and richness and diversity were consistently high in Foothills Riparian areas. These two habitats may be highly suitable for butterflies in this ecosystem, regardless of weather conditions. Generally low abundance and species richness in Montane Woodlands sites, particularly in 2002, sug-gested low suitability of the habitat to butterflies in this ecosystem, and this may be especially impor-tant during drought-like conditions. Finally, to examine the effect that the presence of the very abundant non-native species Pieris rapae L. (Lepidoptera: Pieridae) has on these communities, we re-analyzed the data in the absence of this species. Excluding P. rapae dramatically reduced variation of both butterfly abundance and diversity across habitats, highlighting the importance of considering community membership in analyses like ours.