CHAPTER XXI 



ARE UNIT-CHARACTERS (GENES) CONSTANT OR 



VARLABLE? 



In some of the preceding chapters we have considered facts 

 which show to what a large extent the varieties of animals and 

 plants formed under domestication owe their origin to dis- 

 continuous variations or sports, which, by reason of their 

 Mendelian behavior in heredity, may be combined in various 

 ways through the agency of hybridization. It is a question of 

 much interest, both theoretical and practical, whether these 

 sports or unit-character variations, are entirely stable or 

 whether they themselves are subject to variation. For if a 

 unit-character is not variable, we can only vary the combi- 

 nations into which it enters, the character itself being un- 

 affected. But if a unit-character is variable, it is important 

 to know whether its variation is continuous or discontinuous. 

 For if it varies by distinct steps only, that is discontinuously, 

 it would be a waste of time to try by selection to establish 

 any other conditions than those which arise spontaneously, 

 by " mutation " as De Vries would say. 



The mutation idea has greatly weakened the faith of biolo- 

 gists in selection. Darwin had great confidence in the power 

 of selection gradually to modify the characteristics of races. 

 Practical breeders of animals and plants have always worked 

 by this means, and Darwin based his views concerning the 

 eflScacy of selection largely on the results of their experience. 

 But breeders do not confine their attention to the propaga- 

 tion of variations which they have seen arise spontaneously. 

 They often form ideals of uncreated varieties and then work 

 zealously for the production of these. Some of these ideals 

 may be unattainable, but too many of them have been real- 

 ized to make us think that all work of this sort is fruitless. 

 Today animal breeders hold among their unrealized ideals, 

 a tri-color variety of mouse; a blue variety of fowl which will 



182 



