582 Species According to the Theory of Mutation. 



the breadth of the leaves and the length of the lluwer 

 stalks. But these characters do not segregate in the 

 offspring of the hybrids. They are presumably to be 

 regarded as results of progressive specific dift'erentiation.^ 

 L. vespcrtina is, perhaps, a white-flowered variety of a 

 red-flowered species which has disappeared. At any 

 rate I do not think we shall go far wrong if we conclude 

 that L. vespcrtina and L. diurna differ from one another 

 partly by typical specific, and partly by varietal char- 

 acters. 



Gartner has repeatedly expressed exactly the same 

 view and has illustrated it by the same instance.- He 

 says that any doubt as to the specific difference between 

 closely related species, as for instance between Lychnis 

 diurna and L. vespcrtina, can be most easily removed by 

 crossing; for if such species give exactly similar hybrids 

 with some other, i. e., with a third species, the difference 

 between the two is of a varietal nature only. But if this 

 does not occur we have proof that the essential nature 

 of the species crossed, although they appear closely related 

 with regard to their external features, is specifically dis- 

 tinct. For instance the two species of Lychnis just men- 

 tioned give wholly different hybrids with Cnciibalus vis- 

 cosus. On the other hand Gartner lavs stress on the 

 fact that these species behave as varieties in regard to 

 the color of their flowers when they are mutually crossed. 

 Moreo\'er Lychnis vespcrtina behaves as a variety with 

 regard to the bending over of the teeth of the capsule, 

 that is to say as a retrogressive variety of a species with 

 the character of L. diurna. 



^ For a historical and critical treatment of the point, see a paper 

 by R. Allen Rolfe, Hybridisation Viewed from the Standpoint of 

 Systematic Botany, Journ. Roy. Hort. Soc, April 1900, p. 197. 



' Gartner, he. cif., pp. 581-582. 



