RADIATION AND THE STUDY OF MUTATION IN ANIMALS 1211 



order were obtained as to be obvious even without special techniques. 

 But the result was not attained before the construction of machinery for 

 its measurement. 



Following Muller's work, which was, interestingly enough, done 

 at the same time as the independent work of Stadler on plants (164, 165), 

 a host of confirmations appeared. Similar results have been obtained in 

 all forms on which adequate experiments have been performed, and there 

 is now every reason to believe that in principle the effect is a general one 

 for all organisms. 



THE QUANTITATIVE STUDY OF MUTATION 



As has already been pointed out, an essential factor in the final success 

 of Muller's X-ray experiments was the development of a technique for the 

 study of mutation. The technique is in principle applicable to any 

 organism ; the details will, of course, vary with the form studied. 



Ideally, for the quantitative study of mutations en masse, it would be 

 desirable to detect all types of variants. This would include the lethal 

 effects, the ordinary visible mutants, and the very slight types. Of these, 

 the last present the greatest difficulty for measurement with any 

 accuracy; there are at present no data available which permit even an 

 estimate of their frequency, w^hich may be higher than that of any other 

 type. For the "visible" mutations, it is equally clear that criteria may 

 differ from experiment to experiment, and notoriously from observer to 

 observer. The frequency of occurrence of visible mutation might be 

 determined with accuracy in an organism in which all the possible types 

 of variation w'ere known, and they could be distinguished from develop- 

 mental accidents due to nongenetic influences. Even in Drosophila 

 melanog aster, how'ever, genetically the best studied of all animals, it is not 

 possible to do this yet. 



It is, however, possible to obtain accurate data if a specific mutant 

 character, sharply distinguishable from the norm, is selected, and the 

 frequency of its appearance measured. This is particularly easy for 

 diploid organisms in the case of sex-linked mutations; the offspring belong- 

 ing to the heterozygous sex display any new mutant which has occurred 

 in the germ cells of the parent of the homozygous sex. In the case of 

 organisms haploid in one sex, a new- mutant in any chromosome is 

 similarly detectable at its occurrence. 



Lethal mutations, however, provide the most feasible approach to an 

 objective criterion. Accordingly they were extensively used by Muller 

 in his studies on mutation in Drosophila melanogaster. Essentially, the 

 methods which he developed are based on the usual procedure for locating 

 a lethal mutation (Morgan, 90). A character linked to a lethal appears 

 in numbers proportional to the crossing over between the two genes. If, 

 then, it is desired to test a group of ehromo.somes for lethals, crosses to a 



