ULTRAVIOLET RADIATION AND CANCER 533 



the carcinogenic spectrum. The erythemal mechanism involves at least 

 two light-absorbing substances, whereas the carcinogenesis process prob- 

 ably involves only one since reciprocity is obeyed with polychromatic 

 light. 



Tumor Types and Penetration of the Radiation. Those cutaneous 

 tumors of man in which sunlight may have an etiological role are car- 

 cinomas of either the squamous cell or basal cell type. All originate in 

 the epidermis. When experimental tumors were produced in rodents 

 with ultraviolet radiation, it was generally expected that they would be of 

 these same types. Indeed, the earlier reports indicated that this was 

 generally true. Only Roffo (1934) claimed that a considerable proportion 

 of the tumors induced by ultraviolet radiation were sarcomas, i.e., tumors 

 of tissues underlying the epidermis. There remained a certain skepticism 

 regarding this point when in 1940 experimental studies were initiated at 

 The National Cancer Institute, but it was then found that the incidence 

 of sarcomas induced in the ears of albino mice by ultraviolet radiation 

 was very high — over 90 per cent (Grady et al., 1941, 1943). 



This apparent disparity between human cutaneous cancers and the 

 tumors experimentally induced in the skin of mice needed to be resolved. 

 It was thought that the differences in distribution of tumor types in the 

 two cases might be explained by difference of penetration of the ultra- 

 violet radiation, since mouse epidermis is considerably thinner than 

 human epidermis. In order to settle this question, the transmission of 

 ultraviolet radiation was measured for mouse skin and for human epi- 

 dermis (Kirby-Smith et al, 1942). Figure 13-7, in the preceding chapter, 

 compares the spectral transmissions of samples of mouse and of human 

 epidermis. In the mouse, as in man, the transmission decreases markedly 

 after repeated exposures to ultraviolet radiation because of thickening 

 of the corneum, but mouse skin always remains more transparent to the 

 carcinogenic wave lengths than does human skin, even when the latter 

 has not been exposed to ultraviolet radiation. On the basis of these 

 measurements of physical penetration of the carcinogenic agent, it is 

 altogether plausible that, if ultraviolet radiation is the common carcino- 

 genic agent, tumors of human skin should occur almost entirely in the 

 epidermis, whereas in the mouse deeper tissues are involved. 



Differences in response of tissues to the action of the ultraviolet radi- 

 ation might also be expected, and there is evidence of such differences. 

 Table 14-2 shows the incidence of tumor types obtained under different 

 experimental conditions. The size of the dose had little influence on the 

 distribution of tumor types, and there is not much evidence of any relation 

 to the time from the first dose to the appearance of the tumor. ^ On the 

 other hand, the ratio of carcinomas to sarcomas differed markedly with 

 the frequency with which the doses were given. When the dose was 



■' The "development time" defined in the next section. 



