The results of a 1987 television-advertised colorectal screening program using fecal occult blood tests (FOBT) are compared with the initial 1986 program (results in parentheses). In the 1987 program, 73,508 fecal occult blood test (FOBT) kits were distributed free of charge, of which 63% were returned for analysis (57,000, 53%). Twenty-five percent of persons from the initial screening participated again in the 1987 program: 1,303 or 2.8% of persons had a positive screen (1,165, 3.9%). The predictive value of a positive screen was 23% for an adenomatous polyp and 8% for colorectal cancer (22%, 8%). Seventy-nine percent of the cancers detected were Dukes A or B or carcinoma in situ (78%). In order to promote a more thorough diagnostic work-up in positive screenes, a suggested diagnostic algorithm for the work-up of a positive FOBT was sent to participating physicians. Despite this, 35% of positive screenees had a diagnostic work-up limited to a repeat FOBT, and/or sigmoidoscopy only (32%). In conclusion, television-advertised mass screening programs consistently enroll large numbers of participants. The rate of compliance (percent of kits returned) and the limited diagnostic evaluation of persons with a positive screen appear to be the major factors limiting the success of our screening program.