542 RAIM ATIOX mOLOGY 



that tho induction poriod would have to be a function of both the numl)er 

 of tumor t-ells and the proHferation rate of these cells since these factors 

 are jrioinfj; to determine the suhsc(iuent {j;ro\vth period. That is, /, varies as 

 f{\\(r\ . . . ). We see. therefore, that before the hypothesis assumed in 

 settinfj; up the first of these efjuations can be accepted, some mechanism 

 must be found by which the induction period always adjusts itself so as to 

 be proportional to the growtii period that is going to follow it. Such a 

 mechanism could hardly be consistent with the idea that the tumor cells 

 proliferate autonomously at their own inherent rate, once they are formed, 

 without reference to their past history, since the rate would have to be 

 determined by the same factors that determine the length of the "induc- 

 tion" period. Tt does not appear that any one of the (lualitative hypoth- 

 eses assuming separate induction and growth periods provides such a 

 mechanism. 



Let us now examine another attractive and popular idea of carcino- 

 genesis — that cancer cells are somatic mutants — to see how this fits with 

 these data. The somatic mutation theory of cancer was formulated in 

 the early part of this century and has been variously applied (see Strong, 

 1949) . It met with little favor for a long time but of recent years has been 

 rather widely accepted. This theory postulates that tumor cells originate 

 from normal tissue cells by mutation, thus assuming characteristics which 

 distinguish the tumor cells from normal cells. The idea might have par- 

 ticular interest with regard to the induction of tumors by ultraviolet radi- 

 ation since this agent causes mutations in a multiplicity of living organ- 

 isms. The theory implies the ostensibly irreversible change of a normal 

 tissue cell into a cancer cell. The cancer cell has new characteristics 

 which it transmits to all its daughter cells. One of the outstanding char- 

 acteristics of all tumor cells is that, in some stages at least, they proliferate 

 more rapidly than their fellows, and, since this aspect may be treated 

 (luantitatively, our attention may be focused upon it to the exclusion of 

 others. The same kind of quantitative relations might be expected for 

 any change inherited in the same manner. Let us examine these relations 

 in terms of our data. One of the facts clearly shown by the experiments 

 is that repeated doses of ultraviolet radiation are required to cause a 

 tumor to appear within the lifetime of the mouse. In terms of the muta- 

 tion hypothesis, this might be interpreted to mean that a given cell must 

 be repeatedly acted upon by the radiation before it mutates; but this 

 seems improlmble since typical mutations, whether induced by ultraviolet 

 or some other agent, occur suddenly. A more definite objection is that, 

 as has been shown, the idea of an induction period followed by a period of 

 unrestricted growth is not compatible with the data. 



On first consideration it seems that these objections might be avoided. 

 We may imagine that with each dose of ultraviolet radiation applied to 

 the skin of the mouse, a corresponding number of normal cells mutate to 



