24 SCIENCE PROGRESS 
unimportant. Division being merely the end phenomenon in 
the growth of the individual cell, what requires study is not 
the manner in which the cells divide, but the mechanism of 
cell assimilation and its relation to nutritional environment. 
These problems have been attacked experimentally by 
altering the conditions under which cancer is artificially pro- 
pagated, and, as I have pointed out above, the nutritional 
environment can be so altered that the cancer cell is unable 
to grow. The study of the growth of cancer further permits 
inferences as to the nature of the changes, in cells and in 
organisms, leading to the inception of the continuous assimilation 
of which cancer cells are capable. 
Murray and myself pointed out the importance of separately 
considering two factors in studying the growth of cancer— 
viz. the conditions of origin and the conditions of growth— 
and we pointed out that artificial propagation enabled us to 
study the conditions of growth experimentally. Ehrlich has 
since penetrated more deeply into the relation of these factors 
by differentiating the primary cell changes in a circumscribed 
area from the conditions permitting the cells in such an area 
to grow into a tumour, and by assuming the importance of 
constitutional conditions favourable to growth. We have 
pointed out, with reference to carcinoma mamme in the mouse, 
that the differences revealed and shown to be retained in 
artificial propagation indicate primary differences inherent in 
the cells of different tumours, and that they can be conceived 
as due either to a variety of causes or to one causative factor 
acting in varying degree. As our work proceeded, evidence 
accumulated pointing more and more definitely to the assump- 
tion that one primary change occurring in varying degree 
explained the different behaviour of apparently similar and 
nearly related tumours; but we are ignorant of how the 
primary change is elicited. If histological differences be 
ignored for the present, the only fundamental differences, from 
the standpoint of growth, are in the energy of assimilation of 
the cells, and in the chemiotactic influences they exert on the 
connective tissues of the host which convey the food supply. 
In accordance with this conception what are abstractly called 
variations or differences in malignancy between tumours of 
the same histological type are concretely expressed as differ- 
ences in rate of growth. The injury to the mouse in which 
