1906.] Experimental Analysis of the Growth of Cancer. 199 



be better appreciated by considering also tbat they have been obtained as the 

 result of a triple process of selection ; firstly, the rapidly growing tumours 

 of a batch have been selected because of the greater powers of growth 

 exhibited; secondly, only the healthy parts were used for transplantation ; 

 and thirdly, a further sifting has been effected by the elimination of those 

 cells which degenerate after transplantation. Taken together with the 

 simultaneous reductions of the number of cells continuing growth at each 

 fresh implantation, the repeated implantation of minute cellular grafts 

 renders it practically impossible that any one tumour at the present stage 

 of propagation should still contain cells representing all the growing centres 

 of a tumour even two or three transplantations antecedent to it. 



The percentage of tumours developing after transplanting is, however, 

 only one means of measuring the proliferative power of a tumour 

 experimentally.* It is an arbitrary measure selected for its convenience of 

 application. It merely records that, of a number of fragments taken from 

 a tumour, a certain proportion grew and the remainder were absorbed after 

 implantation in fresh animals. It necessarily neglects variations always 

 obtaining in the weights of tumours in every batch. It is obvious that 

 a sporadic tumour may be obtained, or a time may come in the course of 

 the prolonged selection of tissue for implantation in the future continued 

 propagation of Jensen's tumour when all transplantations will yield tumours. 

 Should this ever be so, measuring the energy of growth by the percentage of 

 tumours developing would fail to reveal any fluctuations. The fluctuations 

 in percentage of success which had previously occurred would retain their 

 importance, and a different method of measurement might still reveal 

 fluctuations dependent on the same factors as great as those represented in 

 these experiments by percentages of success varying between 5 and 100 

 per cent, of the animals used. 



We have studied the growth of the tumour in parallel series of experiments 

 at different times. In order to compare the results we have estimated the 

 percentage of success attending the subinoculation of all tumours trans- 

 planted. The repeated subdivision of the transplanted tumours results in 

 the separate propagation of many strains, which become increasingly 



* The weights which the tumours attain in equal times present great fluctuations as 

 weU. In series with a high percentage of success many tumours attain a weight of 

 1 gramme in the course of 10 days, while series with low percentage of success seldom 

 show tumours of 0'5 gramme weight in the same time. This may be due to the greater 

 number of cells continuing growth in each animal in series of high percentage, and there- 

 fore does not necessarily indicate a more rapid rate of proliferation of the individual cells. 

 For this reason we have not been able to use the weight of tissue produced in a given 

 time as a means of comparing proliferative power at different times. 



