Medical authorities have commented that the compassion of the physician-investigator for human subjects, together with his professional expertise and other individual qualifications, is more important than, and preferable to social controls, for the protection of subjects' rights and welfare. Data largely from interviews with 337 research physicians provide a basis for refinement and operationalization of the concept, compassion, and permit estimation of the frequency, first, of general concern for human subjects among physician-investigators and, then, of their specific concern for the actual subjects of their own investigations. Some research-related social conditions which tend to engender or activate this concern are specified. Findings of earlier studies, which can lead to an unwarranted conclusion that virtually all research physicians are deeply concerned and involved with their subjects, are thus qualified. The data indicate that physician-investigator compassion for subjects is far from universal. In conclusion, the need for social controls, such as peer review, is reaffirmed. Recently the scandal caused by the revelation of the forty-year-old Tuskegee Study, in which treatment was withheld from a group of black male syphilitics, has caused a flurry of activity in the Government and in Congress to propose regulations and bills to strengthen existing practices designed to safeguard the rights and welfare of the human subjects of biomedical research (H.E.W., 1973b; Senate Subcommittee on Health, 1973). Moreover, even before this present flurry, much attention had been given this matter during the past decade: in the writings and discussions of biomedical researchers themselves, lawyers, moral philosophers, and social scientists (e.g., Barber, 1967; Beecher, 1966; Freund, 1970; Ladimer and Newman, 1963; Pappworth, 1967); in the mass media; and in governmental regulations, notably those of the National Institutes of Health and the Food and Drug Administration (Curran, 1970). Since 1966 the National Institutes of Health have mandated institutional peer review for all the biomedical clinical research they have supported. This includes a large part of all the biomedical clinical research done in this country and in nearly every biomedical research organization. In 1971 the Food and Drug Administration, which up to that point had required only written voluntary consent from subjects, followed suit and added the requirement of peer review for all clinical