We collected increment cores from Pseudotsugamenziesii trees growing 9.5 km (closest site) and 15.5 km (intermediate site) from a copper smelter producing pollutant gases (primarily SO2) and also sampled a control site 27.0 km from the smelter. Annual rings were cross-dated and measured, and the growth trend due to age and geometry was removed. All sites had similar high-frequency variance, indicating comparable climatic response. Autocorrelation was highest at the closest site and lowest at the control site, an intimation of long-term disturbance to polluted trees. Approximate climatic variation was removed from the two polluted growth series using control site growth in a subtraction technique, which produced corrected tree-ring indices. Corrected indices were regression modeled using SO2 emissions from 1940–1984. Although both models were significant, SO2 accounted for 54% of closest site nonclimatic growth variance but only 6% at the intermediate site. At highest known emission levels (during World War II), growth at the closest and intermediate sites was reduced 49% and 14%, respectively. SO2 production declined after World War II, and growth increased, perhaps from decreased stand density or from low-concentration SO2 fertilization.