When endogenous pyrogen was given to a normal recipient, a biphasic fever was produced which simulated that seen with bacterial endotoxins. With the use of a technique of serial passive transfer, it has been shown that the "double-humped" response results from two separate actions of the injected pyrogen. The first of these appears to be a direct stimulation of the thermoregulatory centers. The second involves the release of further endogenous pyrogen in the normal recipient to cause, in turn, the second fever peak. Since the injection of endogenous pyrogen did not produce a significant change in the number of circulating leukocytes, it is inferred that this substance is different from either bacterial or tissue polysaccharides.

In rabbits rendered tolerant by a previous injection of virus the second fever peak failed to appear and the response to endogenous pyrogen was monophasic. Evidence indicates that the absence of a second fever peak in the tolerant recipient was not due to rise in temperature on the preceding day of virus injection or to the development of either serum inhibitors or tolerance to virus itself.

It is postulated that prior mobilization of endogenous pyrogen by virus may have modified the ability of the tolerant recipient to liberate further amounts of this substance in response to an injection of endogenous pyrogen.