72 EXPERIMENTAL RESEARCHES 



all been used up, growth can only be further 

 stimulated by giving a fresh supply of this sub- 

 stance, e.g., by re-inoculation into the mouse. 



Just as it is impossible to explain this immunity 

 of rats by antibodies, one cannot thus explain 

 that form of immunity in mice which is evidenced 

 in the number of positive results following inocu- 

 lation with different tumour strains. Generally 

 speaking, the proportion of positive results for 

 one strain, with the exception of unavoidable 

 chances, varies only within narrow limits. 

 Thus, the strain of carcinoma which we culti- 

 vate in grey mice regularly grows in about 20 

 to 25 per cent. ; Jensen's tumour, according to 

 the publications, in about 40 to 60 per cent., 

 our other carcinomata and sarcomata in 90 to 

 100 per cent., and our chondroma unexception- 

 ally in 100 per cent. The constancy of these 

 differences can never be explained by the 

 assumption of antibodies, but only as the 

 expression of a certain vitality of the tumour 

 cells which is constant for each strain. 



The atreptic condition is most clearly shown 

 in the cases repeatedly mentioned by Michaelis, 

 Bashford and Haaland, where the transference 

 to other races of mice of such virulent tumours 

 was either quite impossible, or was only successful 



