MENDELIAN HEREDITY 



But the clear case referred to above is the following: 

 Quite recently there appeared in a culture bottle that had 

 been producing for more than four months (probably for 

 twelve generations) only wingless flies, an individual with 

 one " wingless" wing and one normal wing on the other 

 side. Here the evidence is conclusive that reversion had 

 occurred. The wingless stock in which the asymmetrical 

 form arose had purple eyes and the same eye color was 

 present in the new type. As the eye color was relatively 

 new at the time the chance that contamination had oc- 

 curred was rendered very unlikely. Had contamination 

 by a red-eyed fly occurred, making the new type a hetero- 

 zygote, the eye color would have been the dominant red. 

 When the asymmetrical fly (cO was bred to wingless 

 females only wingless flies appeared, for three or more 

 generations. The reversion, therefore, was somatic and 

 did not involve the germ-plasm, yet this fact does not 

 invalidate the question here raised. 



In the light of this evidence, as well as the evidence 

 from ever-sporting varieties (that may also be consid- 

 ered,! think, as mutating and reverting as regular proc- 

 esses), I believe it unwise to commit ourselves any longer 

 to a view that a recessive character is necessarily the 

 result of a loss from the germ-cell. We need only assume 

 that some readjustment occurs, and as the result a new 

 factor is produced. A simile may make this clearer, if 

 not taken too literally. If we suppose that a factor is a 

 labile aggregate, and that a rearrangement in it occurs, 

 then the new aggregate in connection with the other parts 

 of the cell produces a character that differs from the old 

 one. Here there need be no loss, but only a change in 

 configuration with a corresponding change in the end 

 product in which the changed part plays a role, along 

 with the other parts of the cell. A factor, in this sense, 

 may exist in two or more forms according to the state 

 of equilibrium; one of its states is dominant-producing, 

 and the other is recessive-producing. Such a view may 

 make it easier for us to appreciate that a mutation need 



