1911.] Ancestry and the Incidence of Cancer in Mice. 



47 



recent cancerous heredity, in greater proportion than mice which did not 

 develop cancer. In fact, the difference between the two groups is a minimum 

 difference, and it should be possible, by continued selective mating, to 

 breed two strains of mice with a still greater difference in their liability to 

 cancer. 



In conclusion, it is well to consider the importance of these results in 

 the light of the comparative pathology of cancer. Investigations of the 

 most diverse kind on man and animals show that the actual initiation of 

 cancer is, in many forms of the disease, a terminal phase of a long- 

 continued process of localised chronic irritation. Even in mice in 

 whose ancestry cancer is absent, cancer may arise in consequence of such 

 irritation, particularly when they attain extreme old age, in large numbers, 

 and a diminished predisposition to the disease would merely effect a 

 diminution in the number of individuals attacked as compared with a 

 corresponding number of individuals with inherited liability similarly 

 irritated. The phenomena of the experimental production of sarcoma by 

 transformation of the stroma of transplanted tumours of a particular strain 

 in nearly all animals, however, indicate that the diminished liability is 

 never likely to become absolute in practice. Conversely, it should be 

 possible by shielding individuals, even of highly susceptible stock, from 

 chronic irritation of specific tissues to diminish considerably the incidence of 

 cancer of these tissues amongst them. 



Other investigations have shown that a constitutional condition favouring 

 the grovjth of cancer and accounting for its incidence is not present in mice 

 which suffer spontaneously from the disease. The determining factors are 

 those which initiate cancerous proliferation, and it is highly probable that the 

 predisposing condition which is transmitted is some peculiarity of the cells 

 of the tissues in which cancer develops, of such a kind that, under the wear 

 and tear of life, the regenerative and proliferative changes which accompany 

 the inception of the disease are more prone to occur, or take place with 

 greater intensity. The present observations harmonise with the conclusion 

 drawn from other lines of work, that cancer always arises dc novo in the 

 organism attacked by a transformation of the ordinary tissue elements, 

 and lend no support to the view that groups of cells outside the anatomical 

 and physiological nexus of the organism from an early period in the 

 ontogeny form the physical basis of the development of malignant new 

 growths. 



The figures have also been submitted to mathematical analysis, involving 

 determination of the standard errors of the differences between the cancerous 



