THEORIES AND PROBLEMS OF CANCER 115 



cells vary from each other, so the cells produced by division 

 must vary from the cell that has produced them and from each 

 other. In these inoculation experiments we have therefore two 

 outstanding sets of variable potentialities : those of the individual 

 mice into which the tumour cells are introduced and those of 

 the cells themselves. Theoretically it should be possible to 

 select particular and obvious characters in either the hosts or 

 the tumour cells and with this idea in view I began some experi- 

 ments in selecting tumour cells which I am still continuing. 

 Though mice breed quickly, it would obviously be a more 

 lengthy, difficult and uncertain process to breed highly resistant 

 and highly susceptible races of mice. I used mice obtained 

 from the same source throughout the first series of experiments 

 and have repeated them with mice from an entirely different 

 source. The procedure was as follows : Twenty mice were 

 grafted at the same time with pieces of tumour of as nearly as 

 possible the same size. When two or three were large enough 

 to use for grafting, twenty more mice were grafted from the 

 largest. The process was carried on through several generations 

 as quickly as possible. On the other hand one of the most 

 slowly growing tumours was chosen at a later date from the 

 original batch of mice and was used to graft another twenty and 

 so on for several generations, selecting always a slowly growing 

 tumour. In this way I modified the rapidity of the growth and 

 produced three strains of tumour which developed at different 

 rates on the average. The differences between the rates of 

 growth were so very great as to be beyond explanation as the 

 result of chance. Selection also accounts for the fact that whilst, 

 when this tumour first came from Prof. Ehrlich's laboratory, 

 I succeeded in only about 30 per cent, of the graftings, the 

 percentage of successes increased in subsequent generations to 

 nearly 100. Working with another breed of mice, I have had 

 precisely the same experience. 1 



Other observers who have found that a tumour became more 

 visible after passing through a series of mice of the same breed 

 attribute this change to the acquirement of a power of resistance 



1 The first series of experiments was carried out in the Cancer Research 

 Laboratories in the University of Liverpool with mice bred in Essex. They 

 have been repeated with another breed of mice from Langside, Glasgow. The 

 figures of these experiments will be published shortly, being at present in the 

 hands of the Editors of the Journal of Pathology and Bacteriology. 



