CROSSING. 101 



bination of genes which produced the new colour, we could 

 easily be led to accept a novel character which showed a mono- 

 factorial segregation in crosses with "wild" stock, as proof of 

 the existence of a hitherto undescribed gene. 



So long as on one hand the "mutations" in Drosophila are 

 accepted with so little reserve, whereas on the other hand the 

 cases, in which new characters, new colours, new shapes have 

 been shown to result from novel combinations of known genes, 

 have been so very little worked out, and only in the crudest 

 possible way, we are justified, we think, in attributing a great 

 deal of the startling departures of this material from other 

 material, and of the constant need for more and for more com- 

 plicated subsidiary hypothesis to this peculiar terminology, 

 to this system of naming genes after novel characters. 



For this system evidently promotes in the younger authors 

 a tendency to look upon the relations between genes and char- 

 acters as direct and reciprocal, and to kndre-estimate the ne- 

 cessity of rigid proofs in the work of factorial analysis. 



It seems probable that a certain amount of the genetic var- 

 iability in Drosophila studied in recent years is due to irregu- 

 larities in the usual constitution of the chromosomes, and there- 

 fore to real mutations. Shull has very clearly shown how 

 "crossing over" between chomosomes, including "longitudinal 

 crossing over" could explain simultaneous duplication of genes 

 and loss-mutation, without recourse to real spontaneous loss 

 or the less easily conceived spontaneous acquisition of genes. 

 Disturbances in the usual behaviour of chromosomes have 

 been made very probable in Drosophila. On the other hand 

 it is clear, that the evidence for the existence of so many dis- 

 tinct genes in Drosophila, distinguishing animals with differ- 

 ent characters, is wholly inadequate, and that the possibility 

 is not excluded, that a great part of the unexpected behaviour 

 of characters in crosses, which is now met by a novel hypothe- 

 sis, depends upon the circumstance, that new characters are 

 not expected to originate by novel combinations of partially 

 known genes. 



