The radiographic study of coronary arteries in man has been attempted only in recent years. The earliest workers in this sphere employed thoracic aortography, the opaque medium being introduced by needle puncture (Radner, 1945) (1960), the aim being to fill the aortic sinuses so that a high proportion of the medium enters the coronary arteries. In our animal experiments we found that cardiac arrest induced by acetylcholine (Sloman and Jefferson, 1960) and timed diastolic injections of contrast medium (Michell and Jefferson, 1962) gave encouraging results, but we considered that cardiac arrest was potentially too dangerous to use in man. Although Thal et al. (1958) have applied timed diastolic injections successfully in man, our experience convinced us that it was unreliable in filling diseased coronary arteries. Failure to fill a severely diseased right coronary artery is illustrated by Case 4 ( Fig. 1 and 2). This patient, a man aged 44, had had severe angina of effort for five years. Aortography with 75 ml. opaque medium showed occlusion of the anterior descending artery but failed to outline the right coronary artery (Fig. 1). Selective coronary arteriography was subsequently performed, and the right coronary artery was successfully cannulated and clearly outlined by an injection of only 2 ml. (Fig. 2).Thus, in 1960, faced with the problem of how to improve aortography without increasing the hazards of the procedure, we decided to investigate selective coronary arteriography, introduced by Sones in 1960 and practised by him with success and safety. Apart from Sones, who has had an extensive experience with selective coronary arteriography (Sones and Shirley, 1962), Lang and Sabiston (1961) and Tapia, Bolton, and Mazel (1961) have also reported favourably.