Self-report measures of depressive symptoms, such as the Center for Epidemiological Studies-Depression Scale (CES-D), correlate with current and lifetime smoking status. In one previous study of adult female twins, genetic factors accounted for the covariation of liability to a diagnosis of major depressive disorder and liability to lifetime smoking (Kendler, Neale, MacLean, Heath, Eaves, & Kessler, 1993b, Archives of General Psychiatry, 50, 36-43); however, it remained unclear whether genetic effects also account for the covariation between subclinical depressive symptomology and smoking behavior. In this study, we use twin structural equation modeling to explore whether genetic and/or environmental influences contribute to the covariation between depressive symptoms, as measured by the CES-D, and current and lifetime smoking status among 120 monozygotic and 114 dizygotic Caucasian male twin pairs (aged 59-69). In this sample, depressive symptoms showed small but significant correlations with current and lifetime smoking status. Univariate twin analyses indicated that additive genetic and non-shared environmental factors contributed significantly to liability to current and lifetime smoking. However, the majority of variance in CES-D scores was attributable to non-shared (individual) environment. In bivariate analyses, non-shared environmental factors accounted for the majority of covariation between liability to depressive symptoms (CES-D scores > or = 8; above the 75th percentile) and liability to current and lifetime smoking status. Taken together with the previous literature, these results suggest that the etiology of covariation among depressive symptoms and smoking behavior may vary by measurement and severity of depressive symptomology.