ULTRAVIOLET RADIATION AND CANCER 535 



j^laiid tumor, and one osteochondrosarcoma. The small number of car- 

 cinomas fits with the thinness of the malpighian layer in the mouse ear, 

 which never becomes as thick as it does in human epidermis and hence 

 never absorbs as large a proportion of the incident radiation. When the 

 attempt is made to compare the incidence of occurrence of tumors or of 

 the type of tumors induced by ultraviolet radiation in different species of 

 animals, this question of penetration must be taken into account. An 

 interesting experiment in this regard was made by Hueper (1941), who 

 elected to study the effects of ultraviolet radiation on congenitally hairless 

 rats with the expectation that he would obtain more tumors than with 

 haired mice. To his surprise, only one of the hairless rats developed a 

 tumor and this was a carcinoma. On examination he found that the 

 corneum of the hairless rats was very much thicker than the corneum of 

 the haired animals, which could account for his findings. The generally 

 greater incidence of carcinomas in rats than in mice (Putschar and Holtz, 

 1930; Roffo, 1934; Beard et al., 1936) may be explainable in similar terms. 



The morphological character of these tumors leave little doubt as to 

 their malignanc3^ Metastasis is relatively rare, but it does occur. This 

 is also characteristic of most cutaneous tumors of man, the malignant 

 melanomas being an outstanding exception. Transplantation has been 

 successfully carried out in some instances. For a more complete dis- 

 cussion of tumor types and other aspects of their pathology, the original 

 papers (Grady et al, 1941, 1943) should be consulted. 



Quantitative Aspects. In discussing the quantitative aspects of the 

 induction of tumors by ultraviolet radiation reference is made principally 

 to a series of studies carried on at The National Cancer Institute, since 

 they provide the most extensive data collected thus far and since, because 

 of the method used, they are quantitatively intercomparable. Precau- 

 tions were taken in these experiments to assure that the data would be 

 reproducible. In the first place a genetically homogeneous strain of mice, 

 strain A, was used. This is an albino mouse which tends to develop 

 tumors of the lung but which does not spontaneously develop cutaneous 

 tumors. The females of this strain also develop cancer of the breast, but 

 the males do not; therefore, in order to reduce the number of variables, 

 only the latter sex was used. The diet and other treatment of the animals 

 was made as uniform as feasible so as to reduce biological variability to 

 a minimum. 



The mice were subjected to measured doses of ultraviolet radiation from 

 a mercury arc given at regular intervals. After a considerable number of 

 doses — in only a very few cases was the dosage period less than 1 00 days — 

 tumors appeared on the ears of the mice, the time of appearance depending 

 on the size and the freciuency of the dose. The time from the application 

 of the first dose of ultraviolet radiation to the appearance of such a tumor 

 in a given mouse is here called the "development time," and in the fol- 



