Advances in understanding the factors that limit a species’ range, particularly in the context of climate change, have come disproportionately through investigations at range edges or margins. The margins of a species’ range might often correspond with anomalous microclimates that confer habitat suitability where the species would otherwise fail to persist. We addressed this hypothesis using data from an interior, climatic range margin of the American pika (Ochotona princeps), an indicator of relatively cool, mesic climates in rocky habitats of western North America. Pikas in Lava Beds National Monument, northeastern California, USA, occur at elevations much lower than predicted by latitude and longitude. We hypothesized that pika occurrence within Lava Beds would be associated primarily with features such as “ice caves” in which sub-surface ice persists outside the winter months. We used data loggers to monitor sub-surface temperatures at cave entrances and at non-cave sites, confirming that temperatures were cooler and more stable at cave entrances. We surveyed habitat characteristics and evidence of pika occupancy across a random sample of cave and non-cave sites over a 2-yr period. Pika detection probability was high (~0.97), and the combined occupancy of cave and non-cave sites varied across the 2 yr from 27% to 69%. Contrary to our hypothesis, occupancy was not higher at cave sites. Vegetation metrics were the best predictors of site use by pikas, followed by an edge effect and elevation. The importance of vegetation as a predictor of pika distribution at this interior range margin is congruent with recent studies from other portions of the species’ range. However, we caution that vegetation composition depends on microclimate, which might be the proximal driver of pika distribution. The microclimates available in non-cave crevices accessible to small animals have not been characterized adequately for lava landscapes. We advocate innovation in the acquisition and use of microclimatic data for understanding the distributions of many taxa. Appropriately scaled microclimatic data are increasingly available but rarely used in studies of range dynamics.