Abstract. To determine the generality of avian diversity patterns, we investigated patterns of elevational zonation shown by birds and mammals along the eastern slope of the Andes Mountains in southeastern Peru. The strong environmental gradient sampled, entirely within Peru's Manu National Park and Biosphere Reserve, supports highly diverse faunas. Elevational distributions of 901 bird species, 129 bat species, and twenty-eight species of native mice exhibit contrasting patterns in species richness, species composition, and species turnover. Birds and bats showed smooth declines of species richness with elevation, whereas the richness of mouse assemblages was unrelated to elevation. For all three groups, the greatest differences were between lowland and highland faunas, although cutoff points for this contrast varied among groups (≈ 500 m for birds, 750 m for bats, and 1000 m for mice). Differences in composition also separated bird and bat faunas on either side of c. 1400 m (the boundary between montance forest and cloud forest); for mice, this faunal transition may take place nearer to 2000 m. Bird and bat faunas lacked the more discrete zonations suggested for mouse assemblages, as indicated by elevational range profiles and nested subset analyses. Distinct highland assemblages are apparent in two-dimensional histograms of range limits of birds and mice, but not for bats. Highland bat species occupy broader elevational ranges than lowland bat species, but for both birds and mice, species at intermediate elevations had the broadest amplitudes. Finally, clumping of range maxima and minima along the gradient identified zones of pronounced species turnover in each group, but these were generally not strongly associated with the locations of ecotones.
Differences in zonation of these groups appear to reflect their different biological attributes and phylogenetic histories. Such differences obviously complicate discussions of ‘general’ diversity patterns, and limit the usefulness of birds to forecast or predict diversity patterns in other more poorly known groups—other groups may show elevated diversity and endemism in areas where avian diversity patterns appear unremarkable. The pronounced contrasts between bats and mice, and the generally intermediate character of avian patterns, suggest that future analyses might profitably partition birds into finer, more homogeneous groups of historically and/or ecologically similar species. Group differences in zonation may ultimately prove explicable with information on both species-abundance patterns and resource distributions.