A series of 78 patients with advanced mitral incompetence treated by mitral valve replacement with a Starr-Edwards prosthesis is reported. The overall early mortality was 24 (30.7 per cent) and a further 3 died later. The causes of these early and late deaths are fully analysed. Most occurred in the early part of the series, and there has been a dramatic decline in both mortality and morbidity in the later cases since adoption of continuous clamping of the ascending aorta during the period necessary for excising the valve and suturing the prosthesis in position. The average duration of aortic clamping has been 20 minutes. During the last year (1967) there have been only 5 deaths in 32 patients, representing a mortality of 15.6 per cent. It is emphasised that only patients who have severe incapacity due to mitral incompetence require operation, and neither age nor the degree of pulmonary vascular resistance is a bar to surgery. Advanced renal disease, generalised pulmonary emphysema, and recent myocardial infarction are the only three absolute contraindications. The 54 survivors from operation have been closely supervised for periods varying from several months to as long as 4 years, and as mentioned above 3 have died during this time. The remaining 51 have shown a striking improvement in exercise tolerance, with reduction in heart size. The various surgical procedures available for dealing with mitral incompetence are discussed. Plication of the valve would appear in our experience to be limited to a small number with non-calcific, mildly sclerotic, and only slightly distorted valves. Pig aortic heterografts or pulmonary autografts may offer scope in the future, but we contend that the Starr-Edwards ball-valve prosthesis, although admittedly not as yet ideal, is the best available at present. With the introduction of the latest modification of this valve (now marketed in Britain), thrombo-embolic phenomena should be reduced to a minimum.