No. 641] INHERITANCE OF CANCER IN MICE 523 



ous and he believes that a causal connection between the 

 developmental rate and tendency to cancer exists. 



9. There may, however, possibly be an exception to this 

 independence of the hereditary transmission of the tend- 

 ency to cancer. As we have stated, we arranged our 

 various strains of mice in three groups, in strains with a 

 high, with a medium and low tumor incidence, and we 

 found that in these three groups the cancer age varies 

 yari passu with the decrease in cancer rate. If we now 

 determine in these same three groups on the same per- 

 centage basis the age of death from all other causes 

 taken together except cancer, we find the differences be- 

 tween the three groups considerably less than if we com- 

 •pare the percentages of the cancer age. There is. how- 

 ever, a distinct difference. In the grou]) ol' tlir Iiiali can- 

 cer rate strains the age of death from all (Oilier causrs but 

 cancer is decidedly earlier than in the liUMliuiii and low 

 rate cancer strains. The difference between the medium 

 and low cancer rate strains is very slight, very much less 

 . than that between the high and medium rate tumor 

 strains, but this slight difference is again in the same 

 (lii tctioii. This relationship between cancer rate and age 

 of (h ath from other diseases may be explained in two 

 wa\s: {(I) W'c may assume that the mice dying from 

 caneer are the strongest, most resistant individuals of 

 the family or strain and those which are left back are 

 therefore relatively less resistant to disease; the higher 

 the cancer rate in a strain, the less resistant are the mice 

 not dying from cancer, and the earlier. tluMvfore. tlieii- ags> 



death from other causes. Or (h) tiie majoi-ity of the 

 strains in which the cancer rate was hluh hapix-iied to 

 be less resistant strains and therefcu'e the average age 

 of death from other disease> is rarlin'. The average 

 difference between the modiuni an.l h»\v rate tumor 

 strains, as far as general power of r. sistance is con- 

 cerned, happened to be v(>ry -nialL Aithouixh it is per- 

 haps impossible to decide deliiiiteiy ])etween these al- 

 ternatives, we believe the second one to In. much more 



