Conclusion. 507 



whole ''habit" of the plant, in so far as it is not directly 

 modified by the varietal character, remains unaltered ; 

 color varieties can only be recognized by their colcjr; 

 thornless ones only by the absence of thorns ; and so 

 forth. True species are on the other hand distinguishable 

 from one another in almost every organ and at almost any 

 age. Secondly, the character of a variety usually does 

 not exhibit transgressions. White flowered varieties, al- 

 though they often have a bluish or reddish blush in their 

 petals, are almost always paler than the palest variants of 

 their parent species. The very reverse is the case with 

 specific characters of which the extreme variants may not 

 only meet but often overlap. 



New species differ from varieties in yet another way. 

 The latter usually occur in several genera and families 

 under exactly similar forms, which are as a rule, given 

 the same or synonymous names. But I have sought in 

 vain for forms analogous to my new species with the 

 single exception of the dwarfs. Perhaps also, 0. lacri- 

 folia which is distinguished by the absence of crumples 

 in its leaves, and O. hrevistylis with its partial loss of the 

 inferior ovary should be regarded as true varieties ; they 

 are just the two forms which have never arisen in my 

 cultures. But apart from these, the new species are with- 

 out parallel either in the genus Oenothera or elsewhere 

 in the vegetable kingdom. 



If we regard varieties as having arisen by the loss (or 

 latency) of an old character, it seems reasonable to re- 

 gard species as having arisen by the origin of a new one. 



The question as to the constancy of the now forms is 

 one of the greatest importance. And it may be stated 

 straight away that the character is not one which is at- 

 tained by repeated selection. The view that at least a 



