According to Bergmann's rule, individuals of a given species tend to be larger in colder (northern) climates. Traditional explanation points to the relatively lower surface-to-volume ratio in larger animals and, consequently, relatively lower costs of thermoregulation. We examined intraspecific covariation of body size with geographical location and climate in five species of Sorex shrews, animals that are among the smallest extant mammals. The condylobasal length of skull (CBL), compiled from literature data and measured on museum specimens, was used as an indicator of the overall body size of shrews. Surprisingly, in three out of five shrew species, the CBL was negatively correlated with latitude, and the same trend, although not statistically significant, was found in the fourth species. In general, shrews were smaller in colder areas, as evidenced by the positive correlations between the CBL and temperature. In two species, these positive correlations appeared when the effect of longitude was held constant in the partial correlation analysis. Characteristically, the strongest negative correlation with latitude and positive with temperatures was found in S. minutus, the smallest species under study. Shrews were in general larger in environments with high actual evapotranspiration. Body mass reviewed in S. araneus paralleled the pattern found in the CBL variation in this species, i.e. it decreased northward, both in summer- and winter-caught animals. In addition, contrary to the widely accepted − but not rigorously tested − belief, body mass recession from summer to winter (the Dehnel Effect) did not correlate with latitude. We concluded that shrews followed the converse to Bergmann's rule, and hypothesize that part of their body size variation along the west-east axis may be explained by character displacement. We also hypothesize that scarcity of food, especially in winter, is a major factor selecting for small body size in shrews in northern areas, as smaller individuals should require less food. © 2003 The Linnean Society of London, Biological Journal of the Linnean Society, 2003, 78, 365–381.