The association of a toxin or toxic factor with a Virus has never been demonstrated although it may have been suspected. In fact, under most circumstances it is difficult to devise experiments which would demonstrate clearly the occurrence of a toxin in association with these agents since both would be of small size and usually of labile nature. Thus, ff all possible sources of confusion were controlled, a positive result would be of great value but a negative one would be without significance.For many reasons it has been suspected in this laboratory that the agent of lymphogranuloma venereum, and the other agents which have been shown to be closely allied to it (1-5) should be separated from the true viruses and established in a group by themselves, comparable to the R i c k e t t~ for example (6). It had also been suspected that the agent of lymphogranuloma venereum at least, with which we have had most experience, produces a toxin. In the acute form of generalized infection with this latter agent in man, chills, severe headache, and marked prostration occur and the whole picture suggests a toxemia (7,8). Furthermore, in the laboratory, the convulsive deaths occuring in mice within 30 hours after intracerebral infection (9) are difficult to account for by the lesions present in the meninges, and the cause of death in the chick embryo after infection of the yolk sac has always seemed due possibly to toxemia since the agent does not invade the embryo to any marked extent (10), and careful studies have shown that no significant disturbance of nutrition results from the infection of the yolk ceils (11).The work of Gildemeister and Haagen (12) on the toxin associated with Rickettsia mooseri grown in the yolk sac of the chick embryo stimulated anew our interest in this problem. These authors found that suspensions of yolk sacs heavily infected with R. mooseri, when suspended in Ringer's solution, were lethal for mice. Following intraperitoneal inoculation of 0.5 to 1.0 ml. of such suspensions the mice died in 4 to 48 hours, often with convulsions. Smaller amounts were less active and illtracerebral or subcutaneous inoculation was less effective than intraperitoneal. The toxic substance was very labile and could not be demonstrated after the suspensions had been treated in any manner which killed the Rickettsiae. Serum from individuals who had 463