An association between adenomatous polyps of the large bowel and colorectal cancer has been reported, in the absence, however, of population-based estimates of risk. Subjects with histologically confirmed first diagnosis of large-bowel polyps notified to the population-based Cancer Registry of the Swiss Canton of Vaud (about 600,000 inhabitants) during the calendar period 1979-1990 were actively followed up to the end of 1990 for the subsequent occurrence of malignant neoplasms. Among 2,496 individuals with intestinal polyps, followed for a total of 10,310 person-years at risk (6,201 among males and 4,109 among females), 150 malignant neoplasms were registered versus 152 expected. Thus, the standardized incidence ratio (SIR) for all cancers combined was 0.99. A significant excess was observed for colorectal cancer, with 35 cases observed (19 males, 16 females) versus 17.0 expected (SIR = 2.1; 95% CI: 1.5–3.0). There was also an excess, although not significant, for small-bowel cancer (2 cases observed vs. 0.4 expected; SIR = 5.4). In none of the other cancer sites was SIR significantly or appreciably elevated: in subjects with colorectal polyps the SIR was 1.6 for stomach, 1.0 for lung, 0.9 for breast and 1.2 for prostate. The SIR of colorectal cancer was 3.1 in the first year since polyp registration, and declined thereafter to 1.8, in the absence, however, of any further trend with time since diagnosis. The cumulative risk of colorectal cancer in subjects with colorectal polyps was 2% at 5 years and 3% at 10 years. The quantitative estimates of this study are of interest for their population-based nature, and are potentially useful for defining and targeting screening colonoscopy programmes.