The purpose of this study was to determine if the occult gastrointestinal blood loss produced by aspirin administered in the form of conventional tablets may be reduced or prevented by using aspirin in solutions. Fifty‐seven healthy subiects received 2.6 or 3.9 Gm. of aspirin daily for 8 day periods. Blood loss in the feces was determined by the Cr51 method. The average daily blood loss was 0.5 ml. during drug‐free control periods and 3 ml. above control values when aspirin tablets were taken. After calcium‐aspirin solution with low buffer capacity and a calcium‐sodium‐aspirin solution with medium buffer capacity, average daily blood loss was 48 and 32 per cent of that produced by aspirin tablets. Highly buffered sodium‐aspirin solutions, with or without acetaminophen and caffeine, caused no bleeding. Thus, aspirininduced occult gastrointestinal blood loss may be reduced significantly or prevented by administering the drug in a solution of sufficient buffer capacity to reduce appreciably the acidity of gastric fluids. Since aspirin in solution not only causes little or no bleeding but also is more rapidly absorbed and yields higher drug concentrations in the plasma than do tablets, it is concluded that aspirin‐induced gastrointestinal bleeding is a local rather than a systemic effect under the experimental conditions.