flower and seed-coat color of beans and peas 

 as found by Tschermak and Lock, and is no 

 doubt the correct explanation of the purple 

 mottling in my hybrid beans. Indeed, so 

 many instances are now on record in which a 

 cross results in reversion, that generalizations 

 can be made with some degree of security. 



These reversions indicate that the original 

 character was compound, being determined by 

 the simultaneous action of two or more, pos- 

 sibly many, dominant units, and that the later 

 specific or varietal derivatives were produced 

 by the disappearance of one or more of these 

 original units as a dominant characteristic. 

 Thus in the example assumed above in which 

 the original character was determined by the 

 dominant units ABCDEFGH, the later de- 

 rivatives may be ABCDEFGh, ABGDEFgH, 

 ABCDEfgH, etc., through all the possible per- 

 mutations. May we not perhaps get in this 

 way a comprehensive view of at least the later 

 stages of evolution as a process of analysis 

 due to the disappearance of one unit after 

 another ? 



All the visible variations of the present 

 plant and animal world were once involved in 

 some generalized form or forms, and the 

 process of differentiation pictures itself to us 

 as a true process of evolution brought about 

 by the change of individual character-de- 

 termining units from a dominant to a reces- 

 sive state. This conception results in an 

 interesting paradox, namely, the production 

 of a new character by the loss of an old unit. 



When I first became interested in the Men- 

 delian discipline one of the most difficult 

 things for me to understand was the fact that, 

 somehow, every dominant character in a plant 

 or animal finds its recessive counterpart in 

 all of its near relatives not possessing the 

 character in question. For a time credulity 

 balked and I was compelled to look upon char- 

 acter-units as figures of speech. The origin 



