In 6 patients with bacterial endocarditis studies were made of the bacterial content of arterial and venous blood. Paired samples were collected, approximately simultaneously, from two different locations in the circulatory system, and colony counts were determined. As many as 48 specimens were taken for culture during a single period of study. Venous blood was drawn not only from different locations in the extremities, but also from the superior and inferior venae cavae, the right auricle, and the hepatic and renal veins. As would be expected, colony counts were highest in arterial blood. Blood from the antecubital veins gave colony counts only slightly lower than arterial blood. In the femoral veins, on the other hand, there were appreciably fewer organisms. This difference is attributed to the type of tissues drained by the two veins. Colony counts in blood from the superior and inferior venae cavae were also lower than arterial counts, the ratio being comparable to that found in femoral vein blood. In the renal veins colony counts were only slightly below the arterial level indicating that few organisms are removed from the blood during passage through the kidneys. The greatest reduction in bacterial content was found in hepatic vein blood. In 3 of the 6 subjects this reduction amounted to more than 95 per cent, and in all subjects the difference was very considerable. Mixed venous blood in the right auricle of the heart gave colony counts which were usually one-half to two-thirds as high as in corresponding samples of arterial blood. An interesting finding in these studies was a remarkable constancy of the bacterial content of arterial blood, during periods of 1 or 2 hours. Despite the fact that a considerable portion of the bacteria which leave the heart in arterial blood appear to be removed during a single circuit of the body, the number of bacteria in successive samples of arterial blood shows little change. This indicates that in bacterial endocarditis organisms are discharged into the blood from the endocardial vegetations at a comparatively even rate, rather than in a haphazard fashion as a result of the breaking off of infected particles.