Conclusion. 511 



tioned Draha verna, Viola tricolor, HcliantJicmwn vul- 

 garc, etc. as the best examples of these. But in cases 

 where the constancy of the new species has not been ex- 

 perimentally tested, or where their study has been made 

 difficult by the effects of natural crossing, we get the 

 so-called nebulous groups of the systematists — i. e., 

 groups in which the various authorities do not agree with 

 each others' diagnoses. Salix, Ruhus, Rosa, Hieraciimi 

 are the most familiar examples of this case. 



But the most important point for us is the almost 

 complete agreement between the new Lainarckiana-grou\) 

 and the old biennis-gronp. The forms of the latter, re- 

 garded by some authors as species and by others as 

 varieties, give the impression of being the remains of a 

 previous mutation period. They obviously belong to- 

 gether, differ from one another in the same kind of way 

 as the newer species do; they are constant, mutually 

 fertile, and exhibit transgressive variability in many of 

 their characters in such a way that at first sight they 

 do not seem to be sharply separated from one another. 

 Nevertheless they come absolutely true from seed. 



The supposed mutation period of Oenothera biennis 

 must obviously have taken place in their American home : 

 the products of this period, the Linnean species of to-day, 

 have spread, thence, over a large part of the earth. 



And if we give the rein to our imagination we can 

 conceive each genus and each larger group as being, also, 

 the result of a mutation period. 



