The aim of the randomised study was to compare mortality rates from colorectal cancer (CRC) in persons screened with faecal occult-blood tests every two years during a 10-year period with those of unscreened similar controls. Thirty thousand nine hundred and sixty-seven persons aged 45-75 years in 1985 were allocated to screening and another 30,966 to a control group. Only participants who completed the first round with Hemoccult-II were invited for further screening. Participants with positive tests were offered colonoscopy. The primary endpoint was death from CRC. Sixty-seven percent completed the first screening round, and of these more than 90% accepted repeated screenings. During the 10 year study, 481 persons in the screening group had a diagnosis of CRC, compared with 483 unscreened controls. CRC mortality was significantly lower in the screening group (205 deaths) than in controls (249 deaths) (mortality ratio 0.82 [95% conf. lim. 0.68-0.99], p = 0.03). Our findings indicate that biennial screening by faecal occult-blood tests can reduce CRC mortality.