THE GENETICS OF SPONTANEOUS TUMOR FORMATION 249 



ancestry of the material was fragmentary and insufficient. In all, the long 

 continued process of inbreeding so necessary to establish the genetic homogeneity 

 oj the strains before they were used was lacking. 



This handicap weighed heavily against the later work of Slye (86-92) and 

 of Lathrop and Loeb (48-50). It has provided the most important point of 

 difference between the work completed before 1925 and that of the fifteen 

 years that have followed. 



It is not at all surprising that the earlier work lacked this preliminary 

 process of genetic purification. There was more than one reason why this 

 was the case. 



In the first place extra-peritoneal tumor nodules in mice are striking and 

 superficially obvious. To the early workers in experimental genetics they 

 gave a false sense of simplicity and a feeling that a tumor was something as 

 definite and predictable as coat color or any of the routine Mendelian 

 characters. 



Then too, the great interest of all in the cancer problem was a constant 

 challenge to begin work on it without delay. The period of years necessary 

 in order to carry out the required preliminary inbreeding was not at all 

 appealing to the geneticists who had found abundant "surface gold" in the 

 shape of genetic differences capable of immediate analysis. Once an investi- 

 gator had embarked upon a program of rapid genetic methods, it was not 

 likely that he would scrap that work, retrace his steps and make a fresh start 

 with more uniform material. 



Yet it is probable that more progress would have been made had this been 

 done. The experiments of Tyzzer, of J. A. Murray and of Haaland reached 

 the limit of their potentiality for detailed analysis by showing that female 

 mice with breast tumors had more female ancestors with similar tumors than 

 did tumor-free animals. The findings of Loeb and Lathrop indicated that, 

 in addition, there were "strain differences" in the age at which such tumors 

 developed, and that the actual incidence of the tumors might have a quanti- 

 tative basis on multiple factors. Not even the later painstaking statistical 

 analysis of their data by Bernstein (9) could add basic accuracy or further 

 knowledge of the nature of the genetic process. 



The most extensive series of experiments between 1900 and 1930 were 

 those of Slye who raised and observed thousands of animals. These repre- 

 sented many pedigree lines of descent but had to rely by their very nature on 

 the ex post facto combination and summation of a large number of scattered 

 small-progeny matings in order to establish trends, groups or genetic 

 principles. 



