During the past four years, studies of tuberculosis in patients and animals with special reference to diagnostic cutaneous reactions and allied phenomena of sensitization1 have led me to investigate more critically a phase of experimental tuberculosis that has not received much attention in recent years. There is compelling clinical and pathologic evidence of the important part played by toxemia in the disease. Therefore, the likelihood of a filtrable toxin acting apart from the bacterial cell substance is, in reality, not a new idea. A hint in this direction had already been given by Marmorek2 and by Denys,3 both of whom, however, did not offer proof of the presence of a toxin. During the same period and somewhat earlier, Spengler4 had attempted immunization of guinea-pigs with a filtrate prepared from tubercle bacillus cultures.
All these studies, concerned primarily with attempts at possible therapy, lacked the requirements to