Birds may have the ability to view their environments at a wide range of spatial scales; accordingly, they may make habitat-selection decisions at multiple spatial scales. I investigated the implications of hierarchy theory and a landscape perspective on nestsite selection in cavity-nesting birds in the Uinta Mountains in northeastern Utah. I used · three different approaches to address the concept of a multi-scaled nest-site selection Ill process. First, I conducted an exploratory study in which I investigated nest-site selection at three spatial scales for Red-naped Sapsucker (Sphyrapicus nuchalis), Northern Flicker (Colaptes auratus), Tree Swallow (Tachycineta bicolor), and Mountain Chickadee (Parus gambeli). By conducting a hierarchically structured analysis, I was able to investigate the habitat relationships that might result from a hierarchically organized nest site selection process . I found that the four species were associated with patterns of vegetation at three spatial scales and that these associations combined in such a way as to imply a process of nest-site selection that may be more complex than that posited by the niche-gestalt concept. Second, I conducted an experiment in which I investigated nest-site selection at two spatial scales. I compared the use of four types of aspen stands in a two-by-two factorial design according to within-stand structure and landscape context. Stands were classified as either dense or sparse and as having predominantly meadow or forested edges. To address nest-site selection by secondary cavity nesters , who may be limited by cavity availability, I augmented the natural cavities with nest boxes. I found that birds predominantly nested in sparse stands and in stands with meadow edges. Although only five nest boxes were used for nesting, all five of these boxes were in sparse stands with meadow edges. The third way in which I investigated the process of nest-site selection was to build and test predictive models using associations between birds and landscape patterns. By using landscape patterns to predict habitat, I was able to build models that were easily applied ; predictions could be made without any additional data collection in the field. The models were very accurate for both Red-naped Sapsuckers and Tree Swallows (86- 98% and 53-93% nests correctly predicted, respectively) but were less accurate for Mountain Chickadees and Northern Flickers (33-42% and 19-37%, respectively) .