28 SCIENCE PROGRESS 



settlement of the question whether there is or is not a family 

 liability to the disease. This obstacle is inoperative in the case 

 of the mouse, with a short life of two to three years. Statements 

 have been made of the greater frequency of cancer in the mice 

 housed in certain cages, but no account has been taken of the 

 total population of the cages for the period during which the 

 tumours were observed, nor of the sex and age-constitution of the 

 mouse population. In short, the data necessary to justify state- 

 ments of relative frequency have not been ascertained. To baldly 

 assert that epidemics occur, or that the disease must have been 

 conveyed from one mouse to another because so and so many 

 cases of cancer have occurred in one cage, as has been done in 

 the most positive manner by the New York State Laboratory, 

 Buffalo, U.S.A., is to ignore the rudiments of statistics. By 

 breeding and in-breeding mice of cancerous stock we hope to 

 finally settle the importance attaching to the apparent frequency 

 of cancer in some strains of mice as compared with others. 

 We have long had such observations in progress ; but, since 

 statistical studies on the incidence of cancer in animals require 

 to be conducted with all the precautions accurate statisticians 

 employ in dealing with the incidence of cancer in a human 

 population, we are not yet able to institute comparisons between 

 different communities of mice. These observations do not lend 

 support to assertions of infection, any more than do our studies 

 on the nature of experimental transference, and the means 

 whereby immunity or protection may be conferred. The 

 universality of the disease in man and animals, the biological 

 law of its age incidence, the unique character of the proliferation 

 and its continuation on transplantation, the peculiar nature of 

 the measures which protect mice, as well as the intimate 

 relations between normal and cancerous tissue, and the increased 

 digestive activity induced to compensate for the building up of 

 protoplasm in the tumours borne by otherwise normal animals, 

 all point to the probability that cancer arises de novo in the 

 individual attacked. This statement may seem to be mere 

 dogmatic reiteration of an oft-expressed hypothesis ; but experi- 

 ment has removed this conception of the relation of cancer to 

 the individual from the realms of hypothesis to the level of a 

 well-considered theory, harmonising many isolated facts, and 

 fruitful as the basis of further inquiry. 



Finally I may point out, that if the comparative and experi- 



