34 POISONOUS PKOTEINS 



form the basis of the frequently reported and 

 well attested observations of some of the great 

 clinicians of the past that the case mortality in 

 certain infections, most notably in typhus fever, 

 is much higher in the better nourished than in 

 the less robust. 



As I have indicated the cellular proteins when 

 introduced parenterally into animals are not 

 wholly harmless even when they do not kill. 

 When the cellular substance of the bacillus tu- 

 berculosis is injected into the abdominal cav- 

 ity of a guinea-pig it has no recognizable ef- 

 fect so far as the behavior or external condition 

 of the animal shows. The dead bacilli are taken 

 up in the folds of omentum and develop local 

 tubercles. When the cellular substance of the 

 colon bacillus is injected, a peritonitis results. 

 In short, the lesions which follow infections re- 

 sult also from the injection of the dead cellular 

 substance. I conclude from this that the lesions 

 of the infections are not due to the activity of 

 the living bacilli, but result from reaction be- 

 tween the bacterial proteins and the body cells. 



SPLIT PKODUCTS. In 1903, Wheeler and 1 

 found that the bacterial cellular proteins could 

 be split into poisonous and nonpoisonous parts 

 and later we showed that all true proteins can 

 be broken up in the same way. This work has 



