RADIATION AND THE STUDY OF MUTATION IN ANIMALS 1225 



of these cases await the more accurate data on locus of breakage afforded 

 by cytological study of the chromosomes of the salivary glands. 



Despite these objections, it seems probable that the mutational effects 

 actually are due to the change of neighboring genes resulting from trans- 

 location. The alternative hypotheses reduce to two: (a) presence of 

 deficiency at the locus of break, owing to destruction of genie material 

 at the time of break; or (6) mutation in a broad sense. The former 

 hypothesis, as has been seen, is specifically eliminated in three of the 

 previous cases. As a general proposition for lethal effects it is rendered 

 unlikely by experiments of Schultz (Morgan, Bridges, and Schultz, 94) 

 to test the association of deficiencies with translocations. Forty Minutes, 

 which belong to a dominant phenotype probably due in most cases to 

 deficiency, were found in an X-ray experiment. They were then tested 

 for translocation; not one was found, although under similar circum- 

 stances a few might have been expected as a chance occurrence. 

 Certainly there is no marked association between the production of 

 deficiencies and of translocations. 



The present data do not offer the possibility of completely excluding 

 the mutation alternative. A position effect must, properly speaking, be 

 due to an upset of the normal relation between a gene and its neighbors; 

 when new neighbors are substituted for the old, the gene behaves differ- 

 ently and an apparent mutation results. However, it is possible that the 

 mere separation is sufficient to bring about a change in the gene, owing 

 to the disruption of intergenic bonds (Muller and Altenburg, 117; 

 Sivertzev-Dobzhansky and Dobzhansky, 159). This, however, is not 

 properly to be classed as a position effect; proof of specific mutual rela- 

 tions between neighboring genes is necessary first. 



Certain data on the behavior of the dominant brown allelomorphs 

 point in this direction (Schultz and Dobzhansky, 149). It may be sur- 

 mised, however, that the study of the reversibility of such mutational 

 effects may provide a real distinction; since the reverse rearrangement 

 will also involve a break, there should be, unless a true position effect 

 exists, no corresponding specific reversal of the mutational effect. 



THE "group effect" THE ASSOCIATION OF MUTATIONS WITH 



EACH OTHER 



From the preceding section, it would appear that apparent mutational 

 effects are associated to a high degree with chromosome breakages. 

 Whatever the interpretation, the fact raises another question: To what 

 degree are mutations associated with each other? Muller (113) has 

 addressed himself to this problem in the following way, which resembles 

 the methods used in the study of mutations at the breakage point of 

 translocations: He selected mutations at the locus of scute, and then 



