528 THE AMERICAN NATURALIST [Vol. LII 



origin from a single definite area or locus in the "X" 

 chromosome, accepting linkage as a criterion. They have 

 not arisen in a continuous series but as sudden changes 

 from one extreme to another at comparatively long inter- 

 vals. The character may remain modified in one direc- 

 tion and then suddenly revert to an original condition. 

 Thus white changed to eosin and later back to white as 

 noted by Morgan (1916). Furthermore, the changes are 

 not extremely infrequent. A similar transformation has 

 been noted by Emerson (1917) in maize where self color 

 apparently changed to variegation and later back to self 

 color. A variation which may be of the same type has 

 been described by Shull (191*1) for Lychnis. Quite re- 

 cently Zeleny (1917) in studies on Drosophila melano- 

 gaster Meig. (=anipelophtfa Low) 2 has noted a reversed 

 mutation where full-eyed flies result from the return of 

 the bar gene to the original full-eyed condition. In each 

 of the cases mentioned the germinal purity of the stock 

 was believed to be without question. 



Such results are not to be attributed to a continuous 

 series of mutations, to progressive changes, or to genetic 

 losses. They clearly suggest that the gene, if it is the 

 individual gene which is involved, is made up of smaller 

 combinational units which through their permutations 

 give^rise to the characters in question. Presumptive evi- 

 dence is certainly furnished against the idea that any- 

 thing new has developed in the organism to form the par- 

 ticular characters. Furthermore, one may well believe 

 that any particular mutation under observation suffici- 

 ently long, will exhibit recurrent changes. 



The second argument (b), to the effect that the gene is 

 comparatively stable and that "mutations" are only 

 transitory combinational changes, is based on the main- 

 tenance of apparently identical genes through long 

 periods of time. Thus Metz (1917) found that the three 

 mutations which had, up to that time, occurred in Droso- 

 phila virilis Sturt. appeared exact duplications of the 



