Cyclic adenosine 3′, 5′ monophosphate (cyclic-AMP) has been estimated in mucosal biopsy samples removed from the descending colon and rectum at endoscopy to investigate the possibility of using this substance for monitoring pre-malignant changes in the large bowel. Four groups of patients have been studied: those with normal large bowel and rectal mucosa; those with non-malignant inflammatory bowel disease; those with an adenomatous polyp in the descending colon or sigmoid colon; and those with a rectal adenocarcinoma. No difference was found in the cyclic-AMP content of ‘normal’ rectal mucosa, ‘normal’ colonic mucosa, ‘diseased’ colonic mucosa, carcinomas, and uninvolved mucosa adjacent to the polyps. Less cyclic-AMP was found in the polyps than in adjacent uninvolved mucosa. Conversely, more cyclic-AMP was found in the carcinomas than in adjacent uninvolved mucosa. It is concluded that although cyclic-AMP may be a very useful parameter for delineating the extent of the disease in individual patients, it is not a suitable biochemical marker for the screening of neoplastic changes in the large bowel in the population as a whole.