ARE UNIT-CHARACTERS (GENES) CONSTANT? 185 



tions) the whitest individuals extracted from the dark hooded 

 race were no darker than the darkest individuals extracted 

 from the white hooded race. In other words repeated cross- 

 ing with the non-hooded (wild) race had caused the changes 

 in the hooded character, which had been secured by selection, 

 altogether to disappear. This result shows conclusively that 

 the changes in question had not occurred in the gene for the 

 hooded pattern, but in the residual heredity. Other cases of 

 apparent gradual change in unit-characters under the action 

 of selection may be explained in a similar way. Accordingly 

 we are led to conclude that unit-characters or genes are re- 



Fic. 125. A series of grades for classifying the plus and minus variations of the white 

 spotting pattern of hooded rats. 



markably constant and that when they seem to change as a 

 result of hybridization or of selection unattended by hybridi- 

 zation, the changes are rather in the total complex of factors 

 concerned in heredity than in single genes. 



Nevertheless changes do sometimes occur in single genes. 

 Such, we assume, are the several unit-character variations 

 described in previous chapters, which form the basis of the 

 varieties of domestic animals and cultivated plants. These 

 occur singly and sporadically as changes each in a particular 

 locus or part of a system of genes. By hybridization these 

 isolated changes are later combined in any desired fashion. 

 Change in a genetic locus, that is the appearance of a new 

 gene, is in the terminology of Morgan called a mutation but 

 this use of the term differs fundamentally from that of De 

 Vries. There is no known means by which a mutation, in this 



