Breeding ecology of Long-billed Curlews (Numenius americanus) in northcentral Oregon was studied during spring and summer, 1978-1979. Five habitats were used by territorial curlews, and habitat selection by adults with and without broods was studied. Both adults and adults with broods used habitats significantly different from availability. A preference for habitats of low vertical profile and low vertical density (plant parts/volume/ height) was observed; habitats with tall, dense shrubs or weedy annual vegetation were generally avoided. Nest density varied from 0-9 nests/40 ha. Highest nest density occurred in cheatgrass habitats. Nest success as computed by the Mayfield method was 0.69 in 1978 (n = 40) and 0.65 in 1979 (n = 61). Predators destroyed 10 (4) and 16% (10) of the nests in 1978 and 1979, respectively. Eight of 14 (57%) nests over both years were destroyed by mammalian predators; at least three (21%) were taken by crows (Corvus brachyrhynchos) or magpies (Pica pica). Nest predation rate was significantly lower in annual grass habitats. Differences between nest sites and randomly selected habitats were related to vertical stratification of vegetation. Vegetative cover was important in only the bunchgrass habitat. Overall, nest sites were less complex above 5 cm than surrounding areas. Significant differences in nest-site structure between habitats were observed for five of six paired comparisons, indicating some plasticity in nest-site selection by breeding curlews.