ARE RECESSIVES LOSSES OF GENES? 85 



of that where the wild type (dominant) gene present in 

 the duplicated piece was opposed to two vermilion genes. 

 The two situations are, however, not identical in all re- 

 spects, for the triploid differs from duplication by the 

 occurrence of nearly an entire X-chromosome instead of 

 only a short piece of this chromosome. The excess of 

 genes in the extra X-chromosome may account for the 

 difference in the two cases and this holds equally whether 

 the recessive genes be interpreted as real absences or 

 as mutated genes. 



The Bearing of Reverse Mutation (Atavism) on the 

 Interpretation of the Mutation Process. 



If recessive genes arise by losses, then there is little 

 expectation that a pure recessive stock would ever pro- 

 duce again the original gene, since this would mean ap- 

 parently the production of something highly specific from 

 nothing. On the other hand, if mutation is due to a change 

 in the constitution of the gene, it seems less difficult to 

 imagine that the mutated gene might sometimes return 

 to the original condition. It may be that we know too little 

 about the gene to give much weight to such arguments; 

 nevertheless, the occurrence of return mutants would ap- 

 pear more plausibly explained on the latter view. Unfor- 

 tunately the evidence bearing on the question is not en- 

 tirelv satisfactory. There are, it is true, a number of 

 instances in Drosophila where a mutant recessive stock 

 has given rise to an individual with the original or wild 

 type character; but an occurrence of this sort, unless 

 controlled, cannot be accepted as sufficient evidence, since 

 the chance of contamination of the stock by a wild type 

 individual is not to be ignored. If, however, a mutant 

 stock is marked by several mutant characters, one only 

 of which reverts, the occurrence furnishes the desired 

 evidence, provided no other combinations of these mu- 



