﻿BOTANICAL GAZETTE 



most conspicuous of these differences would appear to be probably 

 quite independent of each other, and we cannot imagine them all 

 having originated at one stroke. Thus, the leaves of one are 



(1) glaucous and amplexicaul, of the other beautifully ciliate; 



(2) the perianth segments of one are widely spreading and recurved, 

 of the other campanulate; (3) the anthers of the one are entire on 

 long filaments, of the other forked and on short filaments; (4) the 

 stigma of the one is nearly entire, of the other 3-cleft. These four 

 main differences are probably not correlated with each other, and 

 may have originated through several independent changes in the 

 common ancestor. This hypothetical ancestor we may suppose 

 threw off a series of new forms differing from each other in various 

 unit characters, just as mutations are known to occur in Oenothera, 

 Drosophila, and other genera today. The forms exhibiting these 

 unit differences intercrossed, and, certain of the resulting combina- 

 tions proving more stable than others, two of the more extreme 

 combinations finally survived, while the others gradually dis- 

 appeared. This is of course only one of the possible hypotheses to 

 account for the occurrence of two such species. 



A hybrid between these two species has been described by 

 Fernald 9 from the Gaspe Peninsula under the name of 5. oreopolus. 

 This form has leaves less ciliate than in S. roseus, and flowers like 

 those of S. amplexifolius but deep claret-purple in color. There is 

 thus some evidence that the various character differences do behave 

 independently of each other, and it is also a significant fact that 

 the hybrids are sterile. In this connection I should like to point 

 out the possibility that the elimination of intermediate unit steps 

 between such species as these may be due not only to the instability 

 of certain combinations (since they would split in their offspring), 

 but to the sterility of certain combinations; or, in other words, 

 their inability to produce any offspring. There is evidence, which 

 I need not detail here, to show that sterility in crosses is a condition 

 which may originate relatively suddenly in connection with a 

 series of mutations. In other words, sterility has probably not 

 arisen gradually as the species became farther differentiated, but 

 certain forms are doomed to be sterile with certain other forms 



» Fernald, M. L., Rhodora 8:70. 1906; 9:106. 1907. 



