The preceding studies on typhus fever, chiefly done with a Mexican strain obtained from Dr. Mooser, concern themselves largely with reinvestigations of some of the fundamental problems of this disease. Filtration experiments carried out with methods almost regularly successful with true filterable viruses, in regard to material, suspension fluid, reaction, nature of filters and pressure employed for filtration, indicate that the virus is not filterable in the ordinary sense in which this expression is employed. It is probably smaller than bacteria and the results of filtration experiments suggest that its magnitude is consistent with the tunica bodies observed by Mooser. Negative filtrates did not immunize, a result consistent with the previous work of Olitsky. The virus is present in blood plasma, hardly if at all in leucocytes, and becomes closely associated with the red blood cells, though we do not believe that it is contained in them. It becomes firmly associated with normal red blood cells when these are exposed to infectious plasma, a result similar to that obtained in Rocky Mountain spotted fever by Spencer and Parker. In tissue culture, tunica material with Mooser bodies remains alive and virulent for about 10 days, but so far we have not been able to determine that it can keep alive without the presence of living cells. These results do not carry this subject any further than it has been carried for European typhus in tissue cultures with the same method by Wolbach, Schlesinger and Pinkerton (12). Within glass capsules in the peritoneum of guinea pigs, the virus may remain alive for about the same length of time as in the tissue cultures. Rough comparative virulence estimations between blood plasma in which it would be hardly possible to find a limited number of Mooser bodies, even though they were present, showed the blood plasma to be less infectious than the tunica material, in which considerable numbers of Mooser bodies were visible. The testicular swelling characteristic of Mexican typhus and showing the above mentioned bodies—probably Rickettsia—may be absent in individual guinea pigs under ordinary conditions and in guinea pigs inoculated by other than the intraperitoneal route. On re-inoculation into the peritoneum after non-orchitic passages, the swelling reappears. Whenever it did not so reappear, we found that the strain had either degenerated in virulence or it had been contaminated by intercurrent infection. Though we can not prove it at the present time, we believe that the tunica lesion is an integral part of this disease in guinea pigs, and not an accidental accompaniment. Convalescent blood from Mexican typhus guinea pigs mixed in the test tube with virus affords protection if the blood is taken between the first to the tenth day after defervescence. After the third week, the blood no longer contains protective bodies although the guinea pigs may still be immune. In one case a serum was obtained which was both protective in such a test but at the same time seemed still to contain virus, a result which we cannot explain. No complement-fixing antibodies were found when virus serum was used as antigen and convalescent serum as antibody. The low concentration of the virus in the serum may account for this. In a limited number of observations guinea pigs which were negatively inoculated with virus-serum mixtures proved on re-inoculation to be immune. In one of these cases the protective serum mixture with the virus was taken 1 day, in the other 5 days after temperature had returned to normal and the re-inoculations were done 36 and 40 days after the primary injection. This recalls similar experiences of Nicolle and encourages further immunological study in this direction. In a number of experiments active immunization with formalinized tunica material containing large numbers of the Mooser bodies seems to have modified the course of subsequent inoculations in the direction of protection. A single accidental human infection seemed particularly associated with tunica material, although this cannot be positively asserted. All that part of our work which has bearing on the infectious agent is consistent with the assumption that the small, Giemsa-staining bodies observed by Mooser in the tunica of Mexican typhus guinea pigs represent the virus of the disease.