REACTIVITY OF MALIGNANT NEOPLASMS 25 1 



twenty-four hours later, and 3 mice t^vo to t^venty days later. The 

 subcutaneous injection of 1500 reacting tuiits killed all the mice 

 tested. In another instance, ^vhen 3000 reacting iniits were admin- 

 istered in two doses of 1500 reacting units twenty-four hours 

 apart, the first injection produced no effect and 7 out of 13 mice 

 died t\venty-foin' hotns folloAving the second injection. 



No comparison of the lethal effects of subcutaneous injections 

 upon normal and tiniior-bearing mice is attempted since the 

 number of normal mice employed ^\as qtiite small. 



B. tyfjliusiis "agar ^vashings" filtrates also sho\ved a decidedly 

 higiier toxicity for tumor-bearing mice, the greatest mortality 

 being elicited within twenty-four hours. Two otit of 30 tiunor- 

 bearing mice and 8 out of 20 normal mice sur\'ived the intrave- 

 nous and intraperitoneal injections of doses ranging from 125 to 

 250 reacting luiits. 



B. enteritidis culture filtrates in the doses employed were 

 equally toxic to normal and tinnor-bearing mice. Doses of 125 

 to 250 reacting units injected intravenously and intraperitoneally 

 killed all the mice tested witiiin twenty-fotn^ hours. 



B. coU, Streptococcus Jieuwlyticus, B. tuberculosis (human 

 strain) and StapJiylococcus aureus cidtiue filtrates injected intra- 

 venously and intraperitoneally killed a high proportion of mice. 

 Of these filtrates, StapJjylococcus aureus produced no effect upon 

 the ttmior. These latter experiments, although carried otit ^vith 

 an inadequately small ntnnber of mice, suggested that the tumor- 

 destroying potency of bacterial filtrates ^vas independent of their 

 lethal effect. 



Thus, the foregoing experiments on a large series of mice con- 

 firmed the previous findings that the filtrates are decidedly more 

 toxic to ttmior-bearing than to nc^rmal mice. They suggested, ho^v- 

 ever, that the hemorrhagic and necrotic reactions in timior tissue 

 coidd not be held directly responsible for the high death rate 

 since certain preparations lacking in tiunor-destroying principles, 

 or those possessing a lo^v concentration thereof, also displayed a 

 high toxicity for tumor-bearing mice. Attempts \vere then made 

 to obtain preparations of high ttimor-destroying and low lethal 

 potencies. 



In the comse of netUralization experiments, the observation 

 was fi^equently made that in certain proportions, mixtines of B. 

 typJwsus filtrates ^vith homologous antisera ^verc comj^letelv de- 



