In Chronic Disease of the heart, prognosis and therapy depend not only upon the anatomical and etiological diagnosis of the damage but also on the proper assessment of the repair, the work demanded of the damaged organ, and the reserve left to deal with the load thrown upon the defective mechanism. Even the most careful calculations, however, may be overthrown by unforeseen accidents, and the outlook may suddenly be completely changed. Among such accidents is the occurrence of thrombi within the chambers of the heart, with consequent interference with the peripheral circulation and embolism. The frequent occurrence of thrombi within the ventricles and auricles, particularly in mitral stenosis, is proved by numerous autopsy reports. Among 178 rheumatic hearts examined by Graef et al., 24 showed mural thrombi, which in 14 cases were lodged in the left auricle, in 5 in the right, and in 5 in both auricles. Among 60 patients with auricular fibrillation and rheumatic heart disease, Garvin found 26 cases of mural thrombosis at autopsy (43.3 per cent). In a series of 771 patients with all types of heart disease, including coronary artery disease, hypertensive heart disease, and syphilitic heart lesions, the incidence was 34.4 per cent. In sharp contrast to their frequency in autopsy series is the small number of intra-auricular thrombi diagnosed during life. Few have been recognized clinically, and still fewer roentgenographically. These thrombi are not an agonal or post-mortem phenomenon; they are not the terminal outcome of auricular fibrillation; they are, according to Levine, an important cause of an appreciable number of deaths (20 per cent), and they might well be amenable to modem anticoagulant treatment. That there is sufficient time for the application of such treatment if the diagnosis is made early is evident from a report by Schwartz and Biloon in which signs of embolism occurred in 1919 and death ensued seven years later (1926) from proved auricular thrombosis. Similarly one of our own patients who now shows roentgenologic evidence of auricular thrombosis and electrocardiographic evidence of auricular fibrillation had the first signs of cerebral embolism, leading to temporary hemiplegia, in 1931 and signs of pulmonary infarction in 1946. Auricular thrombi are of two kinds, the non-occluding and the occluding. The non-occluding thrombi are those which do not impinge on the auriculo-ventricular orifice and therefore do not impede the flow of blood from auricle to ventricle. By far the great majority of auricular thrombi are of this type. The occluding thrombi are those which, because of their size and position, impinge on the valve orifice and hinder the flow of blood through it. They may be pedunculated or completely free as loose bodies in the auricular cavity. The spherical free clots are the so-called ball thrombi.