392 RADIATION BIOLOGY 



ture, and that its mutation represents a change in that structure which, 

 Hke all chemical changes, is subject to the all-or-none rule of cjuantum 

 events. The same principle of sudden discrete change, preceded and 

 followed by stability, was then found to hold for the mutations induced 

 by radiation. This result, obtained with an agent known to cause 

 individual quantum changes of atoms and molecules, tended to confirm 

 the interpretation of mutation as involving definite chemical recombina- 

 tion. In a sense, however, this is almost stating the case backwards, 

 since it was chiefly these and related considerations, given below, which 

 had led the writer to the testing out of the possibility that ionizing radia- 

 tion produces mutations. 



The view that a gene mutation represents a definite chemical change in 

 the composition of the gene does not imply that, as in a reaction in a 

 test tube, a definite product could be produced to order by adding a 

 certain reagent or arranging conditions in a certain way. Mutation 

 does not occur on a molar scale but, on the contrary, each mutation 

 represents one submicroscopic transformation, instigated by a physico- 

 chemical situation, the minute localization of which, in the effective 

 neighborhood of this gene or that gene, must depend upon the many 

 chance factors of what Troland (1917) aptly called "the molecular 

 chaos." That this was the case was at that time already indicated by the 

 fact that spontaneous mutations appeared to happen largely at random, 

 without reference to the type of environment. Even under the most 

 constant conditions of living obtainable, mutations of the most diverse 

 kinds continued to arise in an apparently sporadic fashion, while, con- 

 versely, changing the environment had no discernible effect in causing 

 mutations of given types to be found. Moreover, when a mutation 

 arose, only one gene in the given chromosome, and indeed in the nucleus, 

 underwent change at a time. 



Especially telling was the evidence, the significance of which was 

 pointed out by the present writer (1920, 1922), showing that when muta- 

 tion occurs in a gene in a given chromosome of a diploid cell, the homolo- 

 gous gene of identical composition, which in Drosophila lies in a cor- 

 responding position in the homologous chromosome at a distance from 

 the first gene, usually, of only a fraction of a micron (because of the close 

 pairing of the chromosomes even in ordinary interphase stages in Dro- 

 sophila), fails to undergo any change at all. Thus the mutation could 

 not have been due to the presence, in molar amount in the cell, of some 

 special chemical, of such composition as to have a pronounced proclivity 

 for reacting with that particular type of gene rather than with genes of 

 other types. For in that case there would have been a tendency for both 

 identical genes to have been attacked at once. Instead, the decision that 

 this gene rather than another one underwent mutation on a given occa- 

 sion evidently depended upon the chance ultramicroscopic distribution of 



