- This study examined the differences, at various scales, between northern spotted owl (Strix occidentalis caurina) nest sites and non-nest sites, i.e., sites where no nests were found during surveys, in the Coast Ranges of Oregon. I compared land cover around 41 owl nest sites (ONS) and 41 non-nest sites (NNS) at 5 different landscape scales (0.6-, 1.2-, 1.8-, 24-, and 3.0-km-radius circles). I also compared characteristics of 29 old-growth (overstory trees >86-cm diameter-at-breast-height [dbh]) patches where owl nest were found (ONP) with characteristics of 29 old-growth patches where no nest were found (NNP). All sites were in the Eugene District, Bureau of Land Management. Land cover was classified into 8 categories and typed from aerial photographs. I quantified land-cover indices using a geographic information system and used logistic regression for binary responses to make statistical inferences. At the 0.6-km-radius circle, percentages of old-growth and old-remnant (young conifer stands with some large-diameter trees) forests were the best predictors of difference between ONS and NNS, being greater at ONS. The even distribution of forest cover types, as calculated using Simpson's Evenness Index, was also a good predictor of this difference, being more even at ONS. At the 1.2-km-radius circle, GlSfrag, i.e., the mean distance among old-growth patches, was the best predictor of difference between ONS and NNS, being greater at NNS. At 1.8-, 2.4-, and 3.0-km-radius circles, the percentage of pole-young ("even-aged" conifer stands; trees 13-53-cm dbh) forests was the best predictor of difference between ONS and NNS, being greater at NNS. However, differences in all of these landscape indices between ONS and NNS, including GlSfrag and percentage of pole-young forests, were greatest at the smallest circle size, and differences decreased as circle size increased. At the old-growth patch scale, the ratio of core area (interior patch area >100 m from the outer patch edge) to patch size and complexity of patch shape were the best predictors of difference between ONP and NNP, being greater at ONP. Furthermore, spotted owls tend to nest in old-growth patches, particularly those [greater than or equal to] 40-50 ha in size and with a ratio of core area to patch size [greater than or equal to] 0.20-0.30. This study supports assertions that spotted owl nest sites are associated with old-growth forests, Nest sites were also associated with old-remnant forests. Spotted owls avoid pole-young forests for nest sites.