August 20, 1915] 



SCIENCE 



247 



nishes actual " homozygotes " whicli will stand 

 every test of the theoretical " homozygote." 

 Tower's work with chrysomelid beetles gives 

 him pure-breeding species which behave as 

 homozygotes in one hybrid cross with other 

 pure breeding species, and as heterozygotes in 

 other such crosses; and this order controlled 

 environmental conditions never equalled in 

 experiments with mice. 



Moreover, the assumption of the exact simi- 

 larity of every first generation hybrid in a 

 given cross with respect to a given " unit char- 

 acter " leaves no place in nature for variation 

 in any " unit character." Variation therefore 

 would become wholly a matter of environment. 



The divergence from the accepted canon of 

 my results of color transmission in crossing 

 wild housemice with laboratory albinos, in- 

 volves a difference of data which in no way 

 ailects the question of the transmission of 

 cancer. 



Moreover, it is increasingly difficult to 

 know the established canon in the behavior of 

 characters in heredity. Exceptions to what 

 was the canon have become so numerous as to 

 be a part of the rule; and Riddle's work on 

 melanin formation makes it particularly dan- 

 gerous to be dogmatic on the transmission of 

 such pigmentation in heredity, particularly in 

 mammalian species where pigment is melanin. 



My attack upon the problem of the inherita- 

 bility of cancer was made almost with the sole 

 end of solving the practical questions as to its 

 inheritability and its nature, in order that we 

 might get light upon the methods for its pre- 

 vention and its cure, since these facts are so 

 desperately needed. 



In the face of the tremendous difficulties 

 which the study of these things involve, it is 

 essential that all minor considerations should 

 be laid aside. It is essential also that the pres- 

 entation of results should be simplified as 

 much as possible and be kept as free as they 

 can be from the disputes involved in the study 

 of general problems in heredity. These details 

 are not desired by the two classes most con- 

 cerned, viz., humanity, who suffers from can- 

 cer, and the medical profession who must deal 

 with it. 



My practical results in the matter of the 

 inheritability of cancer are these: 



1. I have established strains of mice which 

 neither in inbreeding nor in crosses with other 

 noncancer-bearing strains ever in any genera- 

 tion have produced cancer. 



2. I have made hybrid crosses between can- 

 cerous and non-cancerous individuals and 

 have extracted from such crosses lines of mice 

 which neither in inbreeding nor in hybridiza- 

 tion with other non-cancer-bearing strains of 

 mice have ever afterward shown cancer. 



3. I have produced a cancer strain in which 

 every member (of a reasonable cancer age) 

 still living after my cancer work began bred 

 true to cancer, and carried it into every strain 

 with which it was ever hybridized. 



4. I have extracted from crosses between 

 cancerous and non-cancerous mice, lives 

 which produce as high a per cent, of can- 

 cer-bearing individuals as it is reasonable 

 to expect in dealing with a characteristic 

 like cancer which may not appear until a 

 mouse is three years old or even more. Many 

 of these mice in cancer strains inevitably 

 die of other causes before the right provo- 

 cation has induced cancer in them. But 

 the hybridization test has shovioi them can- 

 cerous potentially, for they transmitted it to 

 their offspring. 



This test, of course, has not been made with 

 every individual. To subject every individual 

 in every strain to every test is obviously im- 

 possible. The best one can do is to make a 

 reasonable number of such tests, the object 

 being to give to the medical world as quickly 

 as possible the evident facts. 



The production of about a thousand spon- 

 taneous cancers in specified strains, and the 

 non-occurrence among this entire number of 

 any cancer in certain other specified strains, 

 no matter what test is applied to them, demon- 

 strates to every reasonable probability the in- 

 heritability of cancer, and when these results 

 are characteristically and systematically ob- 

 tained in such an immense stock as to furnish 

 over ten thousand autopsies and a living stock 

 of about eleven thousand mice, with a steady 

 production of between seventy-five and a hun- 



