REPRODUCTION 93 



at, if they did not absolutely arrive there. 

 (Plate III.) 



Does this condition indicate that the unit 

 character of the original hooded stock has 

 gradually been pushed along in one derived 

 stock in one direction and in the other in the 

 opposite, or is this whole change a purely 

 superficial one and does the unit character 

 still remain in both stocks unmodified? An 

 answer to this question was found by cross- 

 ing: females from each extreme stock with a 



o 



single wild dark male and by breeding grand- 

 children separately from each of the two sets 

 of descendants. Under such circumstances we 

 should expect to find that three fourths of 

 the grandchildren of both sets would be dark 



o 



and that one fourth in each set would show 

 the condition of the hooded character. Such 

 proportions were practically realized, and the 

 hooded individuals in the two sets were found 

 to be, not like the original hooded stock, but 

 like their modified grandparents. The hooded 

 individuals from the light grandmother were 

 mostly light; those from the dark grand- 

 mother, mostly dark. These results show that 

 a unit character is not necessarily fixed, as 

 De Vries originally maintained, but can be 



