• Samples of blood were obtained from the femoral vein and from the heart in the course of 75 autopsies. In 24 cases no alcohol was found in either femoral or cardiac blood. This suggested that microbial activity and postmortem autolysis do not produce alcohol in the cadaver. In 51 cases alcohol was found in the blood, and in 35 of these the cardiac blood had a higher alcohol content than the femoral blood. The difference in alcohol content between femoral and cardiac blood was too large to be explained as an error of random sampling. The courts have accepted a blood alcohol level of 0.150% as the criterion of drunkenness or as presumptive evidence that the subject cannot drive an automobile safely. While the difference between the averages for cardiac and for femoral samples was 0.009 percentage points, in many individual cases the difference between cardiac and femoral samples ran higher. The greatest difference observed was 0.090 percentage points. Such differences could cause a person to be pronounced drunk on the basis of his cardiac blood and not drunk on the basis of his femoral blood. Alcohol has been shown to diffuse through the stomach wall after death if it is present there as a beverage taken before death. It can diffuse into neighboring structures including the heart. Since this can lead to serious error in medicolegal situations, it is recommended that postmortem samples of blood to be analyzed for alcohol be taken from the femoral veins. This is done at autopsy immediately after opening the abdomen. Through the pelvic cul-de-sac a pipet is introduced under the inguinal ligament into the femoral vein. By raising the leg of the cadaver slightly and reinforcing the gravitational effect with gentle massage, an adequate sample of blood for analysis can be obtained.