During the past few years there has been observed a dying of fishes in the rivers of the Brazilian State of São Paulo. It is evident, from information obtained by the author and by other investigators, that the phenomenon does not appear to be new among us; it has also been verified in other Brazilian rivers, although not repeated with such regularity as at the present time. In the present paper it is demonstrated that we have to do with a contageous disease, caused by a filterable virus. The epizooty manifests itself by the appearance of some dead or sick fish; the number increasing very much during the suceeding days. It seems that at the outset a majority of the fishes is attacked, in the locality where it appears, for in 2 or 3 days, thousands are affected, and in a short time, 8 to 15 days, there is a rapid decline in the number of ill and dead animals. During this interval the disease spread to affluents of the first river, either above or below the first point of attack, and to a considerable distance from it. Clinically, the disease is characterised by the leassened motor capacity of the fishes, which move very slowly, allowing themselves to be carried by the current, or trying to remain the coves and at the margins of the river; and by the tendency of rising to the surface in an oblique or verticle, position, different from the healthy fish, which rise in a horizontal position. In the initial stage of the disease, the fish actively resist capture, but later on they can be easily taken with the hands. The visible lesions are restricted to spots, of variable size, situated at each side of the dorsum. These spots are frequent but not always present. A more constant character is the congestion of the fins, especially the pectoral fins. Interiorly, there is noted an increase of the mucus in the troath, paleness of the liver with congestion of the gall bladder, the gall being yellowish in colour. These internal lesions also are not constant. The disease is transmitted: directly, by the cohabitation of fish undoubtedly healthy with diseased fish, or by the contamination of the water containing healthy fish by the virus; indirectly, buy placing healthy fish in contaminated water, or in this water filtered through a Chamberland F candle, verified bacteriologically to be free from cracks. The virus do not cause the disease or is destroyed by temperatures above 15° C., or is active on fishes at the temperature of about 12° C. It retains its virulence perfectly at 0° C. This particularity explains the coincidence of the disease always appearing in the winter and in the smaller rivers, in which sudden variations of temperature appeared. However, cold water alone did not produce the disease in healthy test animals, placed in aquaria by the side of the infected fish, in all of the experiments. Because of this characteristic, the disease has been designated as « cryoichthiozoose ». Further histologic investigations shows in sick fishes inflammatory process in the mouth, whose cells show acidophil inclusions. The name « Contagious estomatitis of fishes » was proposed giving a anatomical substrat to the disease, instead the first name. This lesions do permit a further and distant diagnosis of the disease. It was verified that all other known agents causing dying of fishes, such as bacteria, protozoa, physical agents, including low temperature, chemical agents, all causing well known lesions in fishes, could be excluded. Aside from this, they are easily verified by simple laboratory techniques or by the lesions produced in the fishes attacked. In the « discussion » of the subject, it was evident that analogies exist between this disease and certain fish mortalities observed in other countries, especially that studied by Huxley and observed during several years in the rivers of Southern Scotland. It is also possible that some of the fish diseases described as caused by bacteria or animal parasites, but without sure proof of the pathogenicity of these agents, have the same etiology as the disease of the São Paulo rivers, here described.