59< » 



THE AMERICAN NATURALIST [Vol. XLV 



ers, which belong in the 0. Lamarckiana series, occur in 

 the western region. One of these specimens is scarcely 

 if at all distinguishable from 0. rubrinervis. Contrary 

 to the opinion sometimes expressed, I have found her- 

 barium specimens of great value in determining the ex- 

 act characters of many of these races. There would be 

 no excuse for one familiar with the characters of the 

 plants from cultures, failing to discriminate easily be- 

 tween, for example, 0. grandiflora and 0. Lamarckiana, 

 from herbarium specimens showing merely a flowering 

 shoot. And much less conspicuous differences than 

 these, for example in the width or shape of leaves, can 

 be determined with equal accuracy. For example, no 

 one who is familiar with the plants in cultures would 

 confuse typical herbarium specimens of 0. Lamarckiana 

 and 0. rubrinervis. 



Hypotheses Concerning the Origin of 0. Lamarckiana. 

 — There has long been a disposition to look upon 0. 

 Lamarckiana as a "hybrid" 3 and to suppose that, on 

 this account, the phenomena of the sudden appearance 

 of aberrant types (as described by de Vries), displayed 

 by this form, are necessarily deprived of evolutionary 

 significance. This latter view is one which I do not 

 share. As already pointed out, I consider it necessary 

 to regard open-pollinated forms as hybrids in the sense 

 that their immediate ancestry has been participated in 

 by many races differing in various characters. The germ 

 plasm of such forms is like an unpurified chemical or 

 mixture of chemicals and can only be "purified" by con- 

 tinued self-pollination. This "impure" condition is the 

 one under which the evolution of open-pollinated plants 

 or interbreeding animals goes on, and any assumption 



