Our knowledge concerning the permeability of the meninges has been rather limited on account of the difficulties involved in getting reliable data. When large doses of drugs can be given safely, it is easy to determine whether any part of them has passed into the spinal fluid. But in using potent drugs like arsenic or mercury, we find the proportion of the chemical to the body weight so small that even if any of it does pass into the cerebrospinal fluid, it is almost impossible to recover it. Since the advent of salvarsan, we have been enabled to inject relatively large doses of arsenic intravenously, and it is therefore possible to determine with greater ease and precision the permeability of the meninges to this drug. After a careful study of the literature on this subject, I find that the authors have been unable, except in a few instances, to recover