604 THE AMERICAN NATURALIST [Vol. XL V 



races of 0. biennis for crossing with 0. f/ranrfiffora, hy- 

 brids more closely resembling 0. Lamarckiana, at least 

 in foliage, could doubtless be obtained. It would appear 

 that, as far as the characters are concerned, the "Euro- 

 pean biennis" is as likely to have originated by a cross 

 between 0. biennis and 0. Lamarckiana, as 0. Lamarcki- 

 ana is to have originated from 0. (jrtuidiflora X 0. bi 



From the evolutionary standpoint, however, the im- 

 portant question is not whether a given "species" has 

 arisen through crossing, because this is the condition 

 under which the evolution of open-pollinated species 

 must have taken place. Whether or not we assume that 

 mutation is the result of previous crossing, it is neces- 

 sary to determine whether the new types which appear 

 are progressive and will form races which will become in 

 turn the progenitors of future types. 



Even if it be assumed that 0. Lamarckiana originated 

 from a cross between 0. grandiflora and 0. biennis, such 

 crosses must have occurred in nature in North America 

 centuries before the advent of the white man. For there 

 is ample evidence that both these species originally occu- 

 pied the Virginia-Carolina region. 



The natural and necessary tendency of systematists 

 and collectors is to abstract a few from many existent 

 types, as the foundation for their species. The result is 

 that the actual limits between Linmean species appear 

 well-defined until the discovery of intermediate races 

 bridging such gaps makes it appear that, in many cases 

 at least, the lines drawn between Linmean "species" are 

 purely arbitrary. This is shown by cultures of many 

 races belonging to 0. biennis L. and 0. muricata L. from 

 various parts of North America. 



One piece of negative evidence which does not favor 

 the hypothesis that 0. Lamarckiana originated from 0. 

 grandiflora X 0. biennis, is the fact that none of the mu- 

 tants from 0. Lamarckiana have hitherto shown any 

 tendency to revert to either of the putative parents, but 



