The authors examined state dependency on depression, trait stability, and state-trait characteristics of perfectionism in a short-term longitudinal study of university students. Relative stability of perfectionism was assessed with test-retest correlations across 3 time points, and results showed higher rank order and relative stability for perfectionism scores compared with depression scores. Regression and path analyses to disentangle directions of effects revealed that initial maladaptive perfectionism scores remained robust predictors of later perfectionism scores, even after the authors controlled for prior and concurrent depression and other dimensions of perfectionism. Perfectionism proved to be quite stable and was a significant predictor of later depression. Perfectionism was also not meaningfully altered by state changes in depression. The overall findings indicate that perfectionism appears to have substantial relative stability, and perfectionistic discrepancy in particular is a clear vulnerability factor for depression.