Outbreak populations of the mountain pine beetle, Dendroctonus ponderosae Hopkins, in central Oregon, were analyzed with an autologistic regression model to evaluate colonization patterns in thinned and unthinned plots of lodgepole pine, Pinus contorta Dougl. Results confirmed previous observations that beetle attacks in unmanaged stands were related to tree diameter and spacing among trees. Beetles showed a decided preference for colonizing trees with large diameters. Trees with small diameters were rarely attacked unless they were close to other trees under attack. Thinned plots were initially unattractive to beetles, but when attacks occurred, colonization was not very different from the pattern in unthinned plots. Conditional probabilities of attack for trees near other attacked trees were actually greater in thinned plots, indicating that the wide spacing between trees in these plots did not seem to interfere with the switching of attacks between trees. Vigor was not a significant covariate for estimating probability of attack in the unthinned plots. It was significant for one of the thinned plots, however, and marginally significant for another. Only one tree was colonized in a third thinned plot that had the widest spacing and the greatest vigor. A surprise was that some of the oldest trees on thinned plots appeared resistant to attack; their resistance was unrelated to diameter, vigor, or position relative to other attacked trees. For. Sci. 39(3):528-545.