THE STUDY OF CANCER 27 



are not capable of easy solution in the mouse. One can use 

 the organism as an indicator of the nutritive needs and meta- 

 bolic activity of cells, because in studying propagated cancer we 

 have such a large number of tumour cells of the same kind in 

 one animal. Conversely, as I have shown above, we can use the 

 tumour cells as indicators of changes in the animals. So far 

 as these results go they encourage further experimental investi- 

 gation into the growth of cancer as a problem of cell nutrition, 

 in which the organism as a whole is implicated, although the 

 demand proceeds from a circumscribed area, limited in the first 

 place to a little bit of tissue no bigger than a pin's head at the 

 time of inoculation. It is obvious that this line of inquiry may 

 have ultimately direct bearings on certain aspects of the disease 

 in man ; but I must enforce caution in drawing conclusions, 

 and above all warn the layman that I have said nothing in 

 favour of popular views on diet or disordered digestion as 

 causes of cancer. 



This is an opportune stage at which to terminate an outline 

 of the firstfruits of the experimental investigation of cancer, to 

 the development of which the Imperial Cancer Research has 

 very materially contributed. In 1905, 137,128 deaths of males 

 occurred in England and Wales above the age of 35 ; of these 

 11,908 were due to cancer. In the same year there were 138,477 

 deaths of females above 35 years of age, of whom 16,875 died 

 of cancer. Therefore of persons dying above the age of 35, 

 one out of every eleven men died of cancer, and one out of every 

 eight women. Is this great frequency as a cause of death to 

 be explained as due to communication of the disease from 

 one person to another ? Our experimental investigations have 

 revealed no analogy with any known form of infection. The 

 chance of ultimately dying of cancer is one in eight for women 

 and one in twelve for men above the age of 35. A simple 

 calculation shows the probabilities of the occurrence of one or 

 several cases of cancer in families of whom few or many survive 

 beyond the age of 35. Murray and myself have pointed out 

 that when the full facts are known, the incidence of cancer in 

 some animals will probably approximate to that in man, 

 although the forms of cancer to which different species are 

 most liable may differ. Since the greater frequency of cancer 

 coincides with the later years of life, the span of life of human 

 beings has long been one of the obstacles preventing a final 



