IN A PAPER published in the Archives in 1964, Conrad et al1 present the results from a double-blind study of long-term oral anticoagulant therapy for the prevention of recurrences of myocardial infarction. In the introduction to their paper we find the following words: "Unfortunately, past medical trials conducted to test the effectiveness of long-term anticoagulant therapy in preventing recurrences of myocardial infarction have been designed so poorly that they do not qualify as controlled medical trials." We should, ourselves, be hesitant to apply such rigorous critical standards, but their adoption by the authors arouses admiration and gives rise to a keen sense of anticipation as to the quality of their own study. It was, therefore, with some dismay that we found ourselves unable to accept many of the interpretations which the authors put on their own results. Some readers may have accepted the authors' conclusions without a careful scrutiny