Biopsies obtained at colonoscopy from 121 patients with ulcerative colitis (UC) for greater than 7 years were reviewed. Dysplasia or neoplasia was found in 27 patients (22%) after a mean of 16 years; 22 (18%) had dysplasia (all low grade), 2 had polyps, and 3 had carcinoma without prior dysplasia. Seven had dysplasia (or neoplasia) on the next examination, and another 4 after multiple negative examinations. Dysplasia preceded carcinoma in 4 (18%) of the 22 patients, and carcinoma occurred in 7 (6%) of 121 patients. The average time from the first encounter of dysplasia to the finding of carcinoma was 6.3 years, and 3 of 4 patients with negative second colonoscopies then had unremarkable examinations for 2-5 years. Dysplasia without cancer was found later in another 3. Dysplasia was found in 12 of 13 colectomy specimens, including the 3 from patients with cancer in whom no dysplasia had been found at surveillance colonoscopy. Active disease did not eliminate the capability to detect dysplasia or negate the value of surveillance for cancer when colonoscopy was conducted for routine clinical indications. Most dysplasia was detected in the rectum and sigmoid, supporting the value of interim sigmoidoscopies and biopsies during routine management of UC. Compliance for surveillance was diminished when the patient was asymptomatic, thereby increasing the risk of cancer. Low-grade dysplasia, like high-grade dysplasia, is predictive of future carcinoma and warrants careful follow-up. A "negative" second examination is no basis for a sense of security or relaxation of vigilance.