Using two tests for direct density dependence and standard techniques of time series analysis, we identified density dependence in defoliation time series of the spruce budworm across its outbreak range in eastern North America over the years 1945-1988. We carried out analyses for the entire region and for grid cells of defoliation maps at five spatial scales created by aggregating the smallest grid cells. The rate of detection of direct density dependence, as assessed by two previously published methods, decreased with increasing spatial scale. Using both methods, density dependence was detected more frequently at the periphery of the outbreak range, where defoliation rate was lower. This result suggested that density-dependent regulation may be stronger in those areas. The first order autoregressive process was the basic model for defoliation dynamics overall and the most common model across spatial scales. Second-order processes were encountered much less frequently, and those commonly identified as resulting from delayed density dependence generally occurred across spatial scales at a rate expected by chance alone. Our results were similar to those of other published studies, which have found the detection of density dependence to decrease at larger spatial scales. The results also reinforced the importance of considering spatial scale when diagnosing population processes using time series of abundance for single species.