1. Certain aspects of the Jarisch—Herxheimer reaction (JHR) after treatment of louse-borne relapsing fever have been studied in nine patients. Peripheral vasoconstriction immediately preceded the chill phase and profound vasodilation accompanied the flush phase.
2. From four patients 50–150 ml of blood was taken at the onset of the chill and reinjected intravenously into the same patient the next day. One patient experienced a reaction identical with JHR 40–60 min after reinjection, suggesting that the mediator of JHR was present in that blood.
3. From five patients 40 ml of blood were taken serially throughout JHR and 2-ml samples of plasma, free of cells and spirochaetes, were injected into pairs of normal rabbits. All samples taken while the patients had fever were pyrogenic. After incubation with normal rabbit plasma these samples failed to produce fever in endotoxin-refractory rabbits.
4. These results suggest that in the plasma of patients with relapsing fever there is a powerful endotoxin. We were unable to show that the concentration of endotoxin increased during JHR, nor could we demonstrate the presence of leucocyte pyrogen. This failure could be a problem of quantity of plasma used or of an endotoxin-refractory state.
5. The bearing of these observations on the mechanism of JHR and the uselessness of cortisol in modifying the reaction are discussed.