486 Genetic Theory and Evolution 



contrary to all we know about the action of the gene, if we argue now 

 within the classic theory of the gene. Genes change only by mutation 

 into one or more types of alleles, which all have the same general 

 t\'pe of action. All alleles at the scute locus change the bristle pattern; 

 and all at the white locus, the eye color. Nobody has ever found a 

 scute allele affecting eye color. If we accept the explanation of 

 pseudoalleles as duplicated genes, the theory would require that they 

 become diversified. In all cases known, the assumed duplicate has 

 generally the same action, and is different only in detail, but different 

 only to the same extent as another allele. Actually, the pseudoalleles 

 were always described first as ordinary multiple alleles and, in ad- 

 dition, as producing their own alleles of similar action. What dis- 

 similarities there are, occur within the limits of multiple allelic action. 

 For example, one group of the white alleles (see I S C c ee ccc) 

 affects eye color with different expression in the female and in the 

 male; the other group affects eye color with identical expression in the 

 two sexes. The believers in new genes via duplication could say that 

 in pseudoallelism we are witnessing only the beginning of an origi- 

 nation of new genes, and in a million years one of them will become 

 a gene for the production of a new structure, not yet known today. 

 I cannot be satisfied with such an argument, which sounds like 

 Haeckel's type of phylogeny. 



Another argument in favor of the theory could be derived from 

 Lewis' views regarding the functioning of pseudoalleles (see III 4 A). 

 As we have seen, Lewis assumed (as Pontecorvo had done tentatively 

 with great caution ) that in a series of pseudoalleles the new duplicate 

 gene takes over the next synthetic step in a series of synthetic actions 

 of which the original gene produced one step. The idea behind this 

 assumption is, it seems, that an observed orderliness in the action of 

 multiple alleles should be laid down in a similar orderliness in chro- 

 mosomal structure. Previously Serebrovsky and Dubinin had used this 

 approach when they concluded that the step arrangement of scute 

 phenotypes (see I 3 C a) requires a stepwise arrangement of subunits 

 of a gene. Both conclusions are rather unconvincing, actually a revival 

 of the preformist error in chemical guise. But even accepting Lewis' 

 ideas for argument's sake, we still face the general difficulty just 

 discussed: How can the new gene get out of its connection with 

 consecutive steps of a synthetic reaction to catalyze a completely new 

 chain of reactions, as we must expect when a new gene for com- 

 pletely new evolutionary steps originates? Such an event seems to me 

 to be much more miraculous than a repatterning within supermole- 



