12 SCIENCE PROGRESS 
cancer frequently stops growing, and this is the most encouraging 
result obtained by experiment. Nevertheless, there is as yet 
little reason to suppose that limits corresponding to those set 
to the growth of organisms and their organs, are necessarily 
set to the growth of cancer in any of the vertebrates, except 
when artificially set by the death of the animal in which it 
originates or is growing. The primary transplantations of 
our sporadic tumours have succeeded for 66 per cent., and 
ultimately all may be found to be transplantable. Hence, since 
a mouse tumour has now been growing twice as long as a 
mouse lives, the more limited evidence of a similar phenomenon 
in other vertebrates, e.g. rats and dogs, is probably due mainly 
to insufficient experimentation, and to failure to supply suit- 
able soil. On the one hand, we have the limited powers of 
proliferation of normal tissues which never produce tumours 
when transplanted, and the unlimited proliferation of cancerous 
tissue; on the other hand, we have the dependence of the 
inception of cancerous change on senescence of the tissues 
as determined by the laws limiting the duration of life 
specifically. We have stated that in this association hes the 
crux of the problem of cancer, and R. Hertwig in referring 
to physiological senescence, and others, have also noted its 
biological importance. This association involves a problem 
which cannot be attacked directly at present, because we are 
still unable to determine the inception of cancer experimentally. 
We can, however, approach it indirectly by the prolonged study 
of the growth of cancer when once it has started, and endeavour 
to determine if it be purely vegetative, dependent entirely on 
the supply of adequate and suitable food, or be perhaps also 
maintained and renewed by some intra-cellular reorganisation 
occurring periodically, at a time when growth tends to cease 
naturally. In short, we must find out if the study of growth 
gives us any clue to its nature and origin. We can also study 
the conditions favouring and retarding its continuation or modi- 
fying it. Thus, the experimental study of cancer is at present 
essentially the experimental study of the growth of cancer ; 
but this is no subsidiary attribute—it is in reality that property 
of cancer by virtue of which it destroys human life. It is 
customary to consider the growth of tumours as expansive or 
“benign,” and infiltrative or “malignant,” a classification corre- 
sponding also to their clinical behaviour. As a matter of fact 
