Relatively little is known about the epidemiology of carcinoid tumours in contrast to the extensive information available on their biochemical effects and natural history. Accordingly, we have used cancer registrations in England from 1979 to 1987, and in Scotland from 1980 to 1989, to estimate the incidence of carcinoid tumours in Britain. Age-standardised incidence rates for England, based on 3,382 registrations, were 0.71 (0.68-0.75) for men and 0.87 (0.83-0.91) for women, per 100,000 per year. The equivalent rates for Scotland, based on 639 registrations, were 1.17 (0.91-1.44) for men and 1.36 (1.09-1.63) for women. There was a consistent female excess of carcinoid tumours in the reproductive years, which was reversed after the age of 50. The female excess was most striking for gastrointestinal carcinoid tumours in women aged 15-19 years (F:M ratio = 2.14). The sex differences are probably due in part to incidental diagnosis of carcinoid tumours during abdominal procedures, which are more common in women than men at ages 15-49 years. However, there is some evidence to suggest a true sex difference in incidence, particularly the fact that the sex ratio for thoracic tumours varies with age in a similar way to that for gastrointestinal tumours. Hormonal factors may, therefore, be important in the aetiology of carcinoid tumours.