This study presents data on cancer incidence among 37,795 blood donors in an attempt to test the hypothesis that blood donation might be associated with cancer development. At a median follow-up time of nine years (range 5-13 years) a total of 1152 cancer cases have been diagnosed. The expected number of cancer cases derived from an age-matched population was 1459 giving a relative risk ratio (RR) of 0.79 (p less than 0.001). Calculations were made with and without latency periods between first blood donation and the diagnosis of cancer (0, 5, 10, 15 years). Overall, significantly decreased cancer incidence was observed (p less than 0.001) though the number of observed cases of haematological malignancies was not significantly different from that expected. For polycythaemia vera, however, the O/E ratio was 1.81 possibly indicating an association with blood donation. A more likely explanation is that this reflects increased diagnosis of polycythaemia vera in the blood donor population.