Aberrant crypt foci are microscopic lesions found in the colons of rodents treated with carcinogens, and in patients with premalignant colorectal conditions. They consist of single or multiple abnormal crypts and show cellular changes ranging from dysplasia to microscopic adenomacarcinoma. It is thought that these lesions represent the initial stages of the adenomacarcinoma path that results in the development of colorectal neoplasia. We have examined the effect of sulindac and indomethacin on the formation of aberrant crypt foci in rats treated with dimethylhydrazine (DMH). Aberrant crypt foci were induced in male Sprague-Dawley rats with two oral doses of dimethyl hydrazine at 25 mg/kg per dose. Rats were randomized to receive sulindac at 3 mg/kg (n = 20) or 10 mg/kg (n = 18) b.d., indomethacin at 1 mg/kg per day (n = 18) or 2 mg/kg per day (n = 19) or control (n = 37). Drug treatment was started on the day following the first dose of carcinogen and continued for 3 weeks. Colons were fixed flat overnight in 10% formalin and stained with 0.2%. Methylene Blue solution before being studied. There was a significant reduction in the number of aberrant crypt foci in rats treated with 10 mg/kg b.d., sulindac (P = 0.001) and indomethacin at 2 mg/kg per day (P = 0.0002). Sulindac, at 3 mg/kg b.d., and indomethacin, at 1 mg/kg per day, did not have a statistically significant effect (P = 0.089 and P = 0.052, respectively). None of the drug treatments affected the relative frequency of single crypt vs multiple crypt foci. Previous studies have shown that sulindac and indomethacin will significantly inhibit the growth and development of tumours in DMH treated rats. The current data suggest that one of the pathways of action of NSAID is to inhibit the formation of early preneoplastic lesions.