The incidence of and mortality from cerebrovascular diseases (CVD) have been studied in a cohort of 12,210 workers first employed at one of the main plants of the Mayak nuclear facility during 1948–1958 and followed up to 31 December 2000. Information on external γ-ray doses is available for virtually all of these workers (99.9%); the mean total γ-ray dose (± SD) was 0.91 ± 0.95 Gy (99th percentile 3.9 Gy) for men and 0.65 ± 0.75 Gy (99th percentile 2.99 Gy) for women. In contrast, plutonium body burden was measured only for 30.0% of workers; among those monitored, the mean cumulative liver dose from plutonium α-particle exposure (± SD) was 0.40 ± 1.15 Gy (99th percentile 5.88 Gy) for men and 0.81 ± 4.60 Gy (99th percentile 15.95 Gy) for women. A total of 4418 cases of CVD, including 665 cases of stroke, and 753 deaths from CVD, including 404 deaths from stroke, were identified in the study cohort. Having adjusted for non-radiation factors, there were statistically significant increasing trends in CVD incidence but not mortality with both total external γ-ray dose and internal liver dose. Much of the evidence for increased incidence in relation to external dose arose for workers with cumulative doses above 1 Gy. Although the dose response is consistent with linearity, the statistical power to detect non-linearity at external doses below 1 Gy was low. CVD incidence was statistically significantly higher among workers with a plutonium liver dose above 0.1 Gy. There was a statistically significant increasing trend in incidence with increasing internal dose, even after adjusting for external dose, although the trend estimates differed between workers at different plants. The risk estimates for external radiation are generally compatible with those from other large occupational studies, although the incidence data point to higher risk estimates compared to those from the Japanese A-bomb survivors.