Results from the quaiac slide or Hemoccult (HO) test for fecal occult blood were compared with quantitative determinations of gastrointestinal loss after intravenous administration of 51Cr-labeled red cells. Subjects were 80 consecutive patients, without dietary restriction, who were referred because of clinical suspicion of gastrointestinal blood loss or complex anemia. A total of 555 stool specimens analyzed for 51Cr loss were graded negative, trace, or positive by the HO method. Of 338 specimens containing 0-2 ml/day by isotope assay, 7.4% were positive to the HO qualitative test. Loss of at least 10 ml/day in 51Cr equivalent was necessary to assure that the majority of HO reactions would be positive. Of specimens containing more than 30 ml/day, 93% were positive. The ratio of 51Cr-labeled red cell equivalents to stool volume and the percentage of positive HO reactions increased together. When this ratio exceeded 10%, two thirds of the HO responses were positive.