The purpose of this investigation was to determine whether there are recognizable psychological characteristics of predictive value in identifying future physician suicides. A retrospective-prospective design was used. The 33 subjects were a subset of the 1198 medical students who were both participants in the long term Study of the Precursors of Hypertension and Coronary Disease and graduates from the medical school. The group consisted of nine suicides, two matched controls for each suicide, and six distractors. Using a wide range of materials collected for other purposes, a psychiatrist, blind to the number of the suicides and controls, as well as to the criteria for selection of controls, was able to identify the nine suicides correctly.On a rank ordering of the 33 subjects for suicide potential, the suicides occupied the first nine places. There was also a highly significant correlation between the reviewer's rating for suicide potential and the dichotomy, suicide-nonsuicide. The reviewer's ratings on eight categories from the Lorr Outpatient Mood Scale and on five selected dimensions adapted from the Katz Adjustment Scales significantly separated the suicides from the controls. Psychopathology scores were higher for the suicides than for the controls, and the Minnesota-Hartford Personality Assay revealed that the suicides were rated significantly higher than the controls on many personality factors, particularly self-destructive tendency, depression, and guilty self-concept. These findings are briefly discussed and possible explanations for the success in identifying the nine suicides are suggested. The results are provocative and strongly imply that the suicide group differs in many identifiable ways from the control group of medical students.