The value of a 3-day Hemoccult II test in the diagnosis of colorectal disease in patients receiving anticoagulant therapy was studied. The influence of rehydration and nonrehydration of the slide was investigated, as well as the reproducibility of positive tests. Eight hundred forty-nine patients on anticoagulant therapy performed three Hemoccult II tests each. One hundred twenty-eight (15%) had one or more positive slides, when these were rehydrated. Three patients with carcinoma and 13 patients with adenomas were found among these 128 patients. One hundred twenty-one of these patients performed six duplicate Hemoccult II tests with four samples from each of three consecutive stools. Three of the six slides were rehydrated and three were not rehydrated before development. Forty-one and 23 patients, respectively, showed positive test results. Two of three reinvestigated patients with large bowel carcinoma had positive slides in both of these series. Seven and 2, respectively, of 13 patients with adenomas had positive tests in the second series when the slides were developed with and without rehydration. Thus, the predictive value was low with nonrehydrated slides, and the precision was low when the slides were rehydrated. It is concluded that the sensitivity of a 3-day Hemoccult II test without rehydration is too low to permit its use as a screening method for colorectal neoplasms in patients receiving anticoagulant therapy.