u6 SCIENCE PROGRESS 



on the part of the tumour cells. A more probable explanation 

 seems to be that only those cells in the original graft that were 

 most resistant to the new environment survived to divide and 

 produce more cells. Of succeeding generations of cells, whether 

 in the same mouse or after having been transferred to another, 

 all which varied towards less resistance degenerated, whilst 

 those that varied towards greater resistance survived to transmit 

 the favourable variation to other cells, which varied in their 

 turn. This process of selection would go on until a race of cells 

 almost entirely resistant to the environment was produced. 

 When the tumour cells are introduced to a new environment 

 in the shape of a new race of mice, the process would be gone 

 through again, unless of course the environment were so 

 unfavourable to begin with that none of the cells was sufficiently 

 resistant to survive. This interpretation seems to account for 

 the fact that Ehrlich and Apolant l were able to produce a very 

 rapidly growing tumour by a very quick succession of inocula- 

 tions — they were obviously obliged to use only the most rapidly 

 growing cells. 2 It accounts for the fact noted by Jensen, that a 

 well-established tumour gives a higher percentage of successful 

 grafts than a young one. 3 It explains why various parts of the 

 same tumour may give different results when grafted 4 and that 

 though tumour cells will not survive for long in an unsuitable 

 host, some of them survive and multiply when transferred back 

 to a suitable one. 5 Unconscious selection also accounts for 

 the so-called rhythms of growth in Bashford's and Calkins' 

 experiments. 



Bashford 6 has suggested that another kind of selection 

 accounts for the production of strains of rapidly growing 

 tumours. He says : " In the light of the wide experience 

 gained, it can be asserted that the technique which consists 

 in the employment of large doses of tumour emulsion and rapid 

 passage was responsible for the selection of certain primary 

 tumours which survived the procedure and not for the in- 

 duction of a marked change in their rate ol growth." His 



1 Op. cit. 1905. 



2 It of course applies equally to the method of producing a virulent strain of 

 bacteria referred to by these authors. (See previous reference.) 



3 Jensen, op. cit. 1903. 



4 Bashford, Proc. Roy. Soc, B. vol. lxxviii. 1906. 



5 Ehrlich, op. cit. 1905 ; Ehrlich and Apolant, op. cit. 1905. 



6 Fourth Scientific Report, Imperial Cancer Research Fund, 191 1. 



