Eighty-three hospitalized patients suspected of having serious infections were studied for the presence of circulating endotoxin; the limulus lysate test was used. Although 39 patients in this group were infected with organisms capable of releasing endotoxin into the circulation, only two patients had a positive test. Since the test had to be repeated several times before endotoxin was detected in the two positive cases and since the diagnosis could have been established with blood cultures alone in one of these patients, our results cast serious doubt on the usefulness of the limulus test as a diagnostic aid.