This work was begun as a continuation of the studies that I1 described in 1920 in a paper entitled "Bacteriologic Studies in Chronic Arthritis and Chorea." One phase of that study was to take cultures of the blood of arthritic patients. It was found at that time that 14 per cent of fifty-six cases showed positive blood cultures of streptococci. During the development of that work it was thought that massage and manipulation of the affected joint caused germs to be thrown into the blood stream and thereby increased the percentage of positive results, and this work was begun as a check on that opinion. A careful inspection of the technic in taking the cultures showed that when large amounts of blood were added to small amounts of medium, cultures were less frequently found positive than when the blood was greatly diluted by the medium.
These two facts being