In the preceding experiments observations have been reported upon the nature of herpes virus which confirm the suspicion that the virus is intracellularly located in the infected nervous system. In regard to the immunological conditions existing in this disease, our experiments have reaffirmed that herpes virus can be neutralized with the serum of actively immunized animals and have offered an explanation for the irregularity of the results of others, as well as our own. It has been found that brain extracts possess some virus-neutralizing power, but considerably less than the serum of the corresponding animals. Attempts at passive immunization with neutralizing serum were uniformly negative, even when the serum was introduced into the cisterna magna 12 to 24 hours before infection with the virus. It has been shown that active immunity can be attained only when some degree of reaction to the living virus has occurred. Rabbits which survived neutralized serum-virus mixtures did not acquire immunity nor did those treated with virus phenolized to the extent of actual destruction. This point suggests a reinvestigation of the Semple method of rabies immunization. In so far as our studies touched upon the herpes-encephalitis problem we have uniformly failed in attempts to transfer herpes virus directly from man to rabbits. These results are in contradiction to those of most of the earlier workers, but in keeping with the recent reports of Flexner and Amoss. Attempts to overcome the difficulty of transfer by the recently published technique of Perdrau were unsuccessful. Furthermore, animals repeatedly treated with human encephalitis material, either fresh or glycerolated, as practised in the Perdrau method, failed to acquire the slightest degree of immunity to subsequent herpes inoculation. By the inoculation of very small doses or by infection of partially immunized rabbits, as described above, we have succeeded in modifying the characteristic herpetic syndrome in rabbits in a manner which simulates many of the clinical features of human encephalitis. Our own experience forces the conclusion that no valid proof exists upon which can be based an assertion concerning the identity of the virus of herpes with that of encephalitis lethargica. Either the two viruses are entirely unrelated, or else prolonged sojourn in the central nervous system of man attenuates the virus for rabbits to an extent analogous to that in which rabies virus is attenuated for man by passage through rabbits. The isolated successes of Levaditi and of Doerr and their assistants might thus be regarded as fortunate exceptions in which material incompletely attenuated had been at their disposal. We suggest this point of view as an alternative working hypothesis largely because the results we are reporting, as well as those of Flexner and Amoss, are in flat contradiction to the reported successes of earlier workers and the more recent experiments of Perdrau. The experiments of the latter, as described, cannot be explained by the occasional existence of spontaneous encephalitis in his rabbits, nor by the assumption that a herpes virus fortuitously coexisted with that of lethargic encephalitis in his material, inasmuch as this material alone at first injection or in the unglycerolated state failed to infect. It is also possible to conceive that human beings may, by repeated skin infections, attain a not inconsiderable partial immunity to herpes virus, which would explain the nature of the clinical course (as in our partial immunity rabbits) as well as the innocuousness of direct injections of herpetic virus into man, as reported by Bastai and Busacca, and the finding of herpes virus in human beings not suffering from lethargic encephalitis. These suggestions are discussed in order to give this important problem the broadest possible consideration. For the time being, however, such reasoning cannot be taken as more than a logical possibility impressed upon us by our partial immunization experiments. All other experimental evidence obtained by direct inoculations with the limited material at our disposal tends to render identity of the two varieties of viruses unlikely.