28 DIMORPHIC LEAVES IN EELATION TO HEEEDITY. 



sarv to rosort to the idea that variations in such cliaracters in cuUi- 

 vatod stocks must bo (hie to ])revious hyljriiUzation. Some writere 

 consider tliat uniform expression of characters, as in a carefully 

 selected line-bred variety, represent the normal condition of heredity, 

 and assume that this condition is found in wild species. The diver- 

 sities that appear through variation are ascribed to hybridization, to 

 the disturbing influence of the environment, or to mutative transfor- 

 mation into new species. Yet diversity is always found among the 

 members of wild species as soon as the observer gains sulhcient 

 familiarity. The uniformity found m "pure-bred" A^arieties is an 

 artificial ])roduct established and maintained by selection. The 

 inferior variations that api)ear in selected strains of Upland cotton 

 show the same range of diversity that is found among the members of 

 primitive, unselected stocks. Such variations may reasonably be 

 considered as reversions. 



The return of latent characters to expression shoidd not be looked 

 upon as rare or exceptional, but as a normal phenomenon of heredity. 

 Uniform expression of characters is rare and exceptional because the 

 tendency to reversion is so general and jiersistent. Transmission is 

 permanent, not variable like expression. Characters that have been 

 suppressed for thousands of generations, like the incisor teeth of 

 cattle, continue to be transmitted. Students of embryology recognize 

 permanence of transmission by the law of recapitulation. The 

 development of the- embryo of a higlier animal does not take a 

 straight course from the egg toward the adult form, but remains 

 closely parallel with the courses followed in lower groups. Many 

 primitive characters are brought into sUght or tem])orary expression, 

 though they may disa})])ear entirely ])efore even the embryonic devel- 

 opment is comjdete. 



In view of the continued transmission of primitive characters, the 

 tendency of the diversities of the wild types of cotton and other 

 plants to reappear in selected varieties is more easily understood. 

 In a divei-se, unselected type each individual inherits and transmits 

 from its manv ancestoi^s a large number of characters that are not 

 expressed in its own body, and this transmission of latent characters 

 continues even in tlie most carefully selected variety. Though 

 op])()sed by selection, the natural tendency to alternation in expression 

 also contumes and becomes effective in the occasional individiuUs that 

 show mutative variations. 



It is true that examples of mutative reversion are usually not fre- 

 quent enough to affect statistical investigations of other forms of 

 expression of characters, but they are of essential inij)ortance in many 

 questions of heredity and breeding. If the ])()ssil)ility of reversion 

 and suj)])ression of characters ])e left out of account, every definite 



