Four hundred and fifty asymptomatic general practice patients and 330 hospital inpatients had their stools tested for occult blood with the Haemoccult and Fecatwin methods. In general practice, 9/64 (14%) of patients with a positive result had a colonic neoplasm (three carcinomas, one Dukes' Stage A, two Dukes' Stage C, six adenomas) and in hospital 12/142 patients (8%) were found to have colonic tumours, (nine carcinomas, two Dukes' Stage A, two Dukes' Stage B, five Dukes' Stage C and three adenomas). The overall detection rates for colonic neoplasia were 2% in general practice and 3.4% in hospital. In 2 years of follow-up, none of the general practice patients have presented with colonic symptoms. Two hospital patients with colonic carcinomas produced negative tests with both methods. Out of the total of 21 colonic neoplasms, nine were detected by Fecatwin alone, but this trend in favour of the more sensitive test did not reach the 5% level of statistical significance. In contrast, the number of false positive results were significantly greater with Fecatwin than Haemoccult. From our data it would appear that the Fecatwin method warrants assessment in a full controlled trial of its value as a population screening test for colonic cancer.