We examined habitat relationships and nest site characteristics for 6 species of cavity-nesting birds--American kestrel (Falco sparverius), northern flicker (Colaptes auratus), red-headed woodpecker (Melanerpes erythrocephalus), black-capped chickadee (Parus atricapillus), house wren (Troglodytes aedon), and European starling (Sturnus vulgaris)--in a mature plains cottonwood (Populus sargentii) bottomland along the South Platte River in northeastern Colorado in 1985 and 1986. We examined characteristics of cavities, nest trees, and the habitat surrounding nest trees. Density of large trees (>69 cm dbh), total length of dead limbs >10 cm diameter (TDLL), and cavity density were the most important habitat variables; dead limb length (DLL), dbh, and species were the most important tree variables; and cavity height, cavity entrance diameter, and substrate condition at the cavity (live vs. dead) were the most important cavity variables in segregating cavity nesters along habitat, tree, and cavity dimensions, respectively. Random sites differed most from cavity-nesting bird sites on the basis of dbh, DLL, limb tree density (trees with >1 m dead limbs >10 cm diameter), and cavity density. Habitats of red-headed woodpeckers and American kestrels were the most unique, differing most from random sites. Based on current trends in cottonwood demography, densities of cavity-nesting birds will probably decline gradually along the South Platte River, paralleling a decline in DLL, limb tree density, snag density, and the concurrent lack of cottonwood regeneration. J. WILDL. MANAGE. 54(1):112-124 Cavity-nesting birds are a major component of many avian communities. In cottonwood bottomlands they represent 32-43% of the breeding avifauna (Sedgwick and Knopf 1986). In studies of habitat associations of cavity-nesting birds, overstory basal area, understory vegetation, tree dbh and height, and species composition all have been related to densities of cavity nesters (e.g., Anderson and Shugart 1974, McClelland et al. 1979, Raphael and White 1984). The importance of snags (dead trees) to cavity-nesting birds has been especially well documented (Haapanen 1965, Thomas et al. 1979, Scott et al. 1980, Raphael and White 1984). Both the dead and living portions of live trees provide nest substrates for cavity-nesting birds in some forest associations (Carey 1983, Sedgwick and Knopf 1986). Sedgwick and Knopf (1986) concluded that snag densities were not good predictors of breeding habitats for cavitynesting birds in cottonwood bottomlands. Alternatively, limb tree (trees with 21 m dead limbs >10 cm diameter) density and dead limb length (DLL) in live trees, both of which were correlated with cavity density, were better measures of suitable nesting habitat for cavity nest-