NATURE OF THE GENETIC EFFECTS 391 



with a much higher frequency in the descendants of irradiated than of 

 nonirradiated molds, but here genetic analyses were wanting, and the 

 results might have been attributed to such phenomena as somatic segre- 

 gation or nondisjunction. Gager and Blakeslee (1927), on going over a 

 body of data on descendants of radium-treated Datura, obtained before 

 1922, and comparing it with control data obtained since that time, were 

 able to show that many so-called "chromosome mutants" had indeed 

 been produced by the radiation, but most of these were types having an 

 entire but normal extra chromosome, brought about by nondisjunction, 

 an already known effect of radiation (see Sect. 3-2). There was however 

 one case of a structurally changed chromosome, and two cases of reces- 

 sive visible mutations. 



The reason that other investigations, carried out in the later 1920's, 

 succeeded in obtaining more conclusive results than all these lay in 

 the great developments which both genetic technique and genetic 

 theory, based on studies of nonirradiated material, had by that time 

 undergone. These made discriminations between mutagenesis, on the 

 one hand, and both environmentally induced "modifications" and 

 genetic effects of inbreeding, on the other hand, more precise, and also 

 made the analyses into different classes of heritable changes more 

 informative. 



In order to view the more definitive work on the production of changes 

 in genes by means of radiation in its proper perspective, the results of the 

 prior work on gene mutation and its converse, gene stability, in the 

 absence of artificially appfied radiation, should be briefly reviewed here. 

 Considerable evidence had accumulated — for instance, in the work of 

 Johannsen (1909) on beans and of Muller and Altenburg (1919) on 

 Drosophila — that genes are ordinarily very stable. Although they are 

 capable of undergoing "spontaneous" permanent changes these changes 

 are, for any given gene, rare, sudden, and discrete, causing the gene to 

 pass from one stable state to another. The stability of each gene is such 

 as to preclude its undergoing frequent small fluctuations of an inheritable 

 nature, as assumed on the view of "continuous variation." If these fre- 

 quent small changes occurred they would tend to accumulate, so as to 

 result in changes of perceptible size, and these perceptible changes of 

 any gene, when classified according to size and number, would then be 

 grouped in a "probability distribution," and would make possible the 

 progressive success of continuous selection of a given gene in a given 

 direction. This is contrary to the results of exact observations on gene 

 changes (Muller, 1918, 1920; Altenburg and Muller, 1920). 



The finding that an individual gene undergoes a definite transformation 

 when it mutates spontaneously, and that in the intervals between muta- 

 tions it maintains a fixed composition, was difficult to interpret except by 

 the view that each gene, like a molecule, has a distinctive chemical struc- 



