Within recent years investigations of the blood-cerebrospinal fluid barrier have occupied a prominent place in psychiatric research. Various methods have been developed for studying the functional state of the barrier. One of the procedures used most frequently for this purpose is the test devised by Walter1 for determining the ratio of distribution of bromide between the blood and the cerebrospinal fluid. Some of the advantages and disadvantages of this method have been discussed in previous communications.2 Apparently, few workers have attempted to test its accuracy by carrying out parallel experiments with other procedures for estimating the bromide content of the two fluids. Nevertheless, the consistent deviations from the normal that have been discovered in certain psychoses by means of Walter's method cannot be dismissed as purely accidental. Even though their significance may still be obscure, such deviations are of sufficient interest to render further studies in this field