CHAPTER VII 



MENDELISM 



WE have already had occasion to point out how im- 

 portant it is, when engaged upon questions of heredity, 

 not to treat whole animals or plants as units, but to 

 deal with their separate characters one at a time. In 

 the course of the present chapter the reason for pro- 

 ceeding in this way will appear more clearly, and we 

 shall find that the adoption of this method is fully 

 justified by the results which it enables us to obtain, 

 and which could not have been arrived at in any other 

 way. We shall also find reasons for believing that 

 this method is the correct one from a theoretical point 

 of view. 



Naturally, considerable care is necessary in deter- 

 mining what are and what are not separable characters. 

 At the outset it is not always possible to make this 

 discrimination with certainty, but during the course 

 of the experiments which follow it is almost always 

 possible to arrive at a clear definition of each character, 

 and in many cases the distinction of characters is quite 

 obvious from the beginning. 



Up to the present time the experimental study of 

 heredity by the methods of definite breeding has yielded 

 163 ii 2 



