NATURE OF THE GENETIC EFFECTS 423 



mutagenesis in mice by P. Hertwig (1939) or that by Russell, cited on 

 pp. 411 and 432. 



As for the study of carcinogenicity in nonvertebrates, it is still beset 

 with too many doubts to provide critical data in this connection. In 

 some of the work on Drosophila, for example, the criterion used for decid- 

 ing that an agent is carcinogenic has been its action in causing an increase 

 in the frequency of recognizable tumors of some special type in a mutant 

 strain already having, even without treatment, a high incidence of 

 tumors of this particular type (Burdette, 1951). It is uncertain, in the 

 first place, whether these tumors can be considered malignant, in the 

 sense in which this term is used when applied to vertebrates. Secondly, 

 the observed differences in the frequency of detected tumors may have 

 been caused by influences affecting the amount of their growth and 

 melanization rather than that of their origination. Finally, even if the 

 agent did affect the frequency of origination of the tumors, it might be 

 able to exert this influence only through some interaction, perhaps con- 

 fined to that particular type of tissue in which the given tumor arises, 

 between it and the products of the very special genetic agent possessed 

 by the mutant strain used. Hence the results of these tests, no matter 

 whether showing a parallelism or a lack of parallelism between muta- 

 genicity and this putative "carcinogenicity" in Drosophila, must be 

 regarded as far from definitive in their bearing on the somatic mutation 

 interpretation of cancer. 



There is, fortunately, a type of test available of the relation between 

 mutagenicity and carcinogenicity which is much less subject to difficulties 

 and objections of the various kinds mentioned. This consists in the 

 determination of whether agents which have given evidence of producing 

 point mutations in organisms in general act also as carcinogens, in those 

 forms in which the de novo origin of indubitable malignancies can be 

 definitely recognized. It was the discovery of the mutagenic effect of 

 ionizing radiation, an agent already known to be carcinogenic, which had 

 provided the first factual evidence in support of the somatic mutation 

 hypothesis of malignancies — a relation first pointed out by the present 

 writer (1927) — and the results of other workers, showing that radiation is 

 mutagenic in organisms in general, served to make this evidence much 

 more definite. The evidence then received a further important extension 

 in the results showing that ultraviolet radiation also is a general mutagen, 

 inasmuch as this agent likewise was known to be carcinogenic in verte- 

 brates. The lesser yet positive effect of high temperature, in promoting 

 the origination of both mutations and cancers, pointed in the same direc- 

 tion. But it is obviously a requirement of the somatic mutation view of 

 malignancies (even though it be admitted, as it must be, that only some 

 malignancies are of such origin) that all agents which produce point 



