There is considerable controversy in the literature regarding the extent to which chronic pain and depression are associated and the possible causal relationship of such an association. The present study examines these issues with a sample of 243 patients diagnosed with rheumatoid arthritis (RA) who were mailed questionnaires for six waves of data collection. The results indicated that RA patients experience higher levels of depressive symptomatology than community samples. Using a two-latent-variable, cross-lagged design, covariance structural modeling was conducted on self-report measures of pain and depression over 6-month intervals. Results most strongly supported a causal model in which pain predicts depression during the last 12 months of the study.