101 



The above grouping is probably a natural one, but the proper 

 rank of the two groups and their relationship to each other are 

 problematical. The plants included in the first division fall 

 naturally into three or four evolutionary series, starting in each 

 case from the simplest species in the genus, very appropriately 

 B. simplex. But this species may also represent the origin of the 

 second group, and some writers have preferred to place it in this 

 section because of its short commonstalk, and occasionally sub- 

 ternately divided lamina. However, it is surely more closely re- 

 lated to species in the first group, notably B. tenebrosum, through 

 which it is connected with other forms of the same general type. 

 No real connecting species is known between simplex and the 

 species of the ternatum group. The latter form a unified and 

 natural section and have even been given generic rank, but on in- 

 sufficient data. H. L. Lyon * noted a considerable difference in the 

 development of the young sporophyte of B. obliquiim on the one 

 hand, and of B. simplex, B. neglectum, and B. virginianuin 

 on the other. This difference is interesting, but it cannot be con- 

 sidered of value, even to separate the groups, unless it is shown 

 to be constant for other and diverse species in both groups. 



Two other groupings of the species have been made. Milde 

 recognized two subgenera, Eiibotrychiiim and Osmundopteris. 

 The latter included only B. virginiamim, and was based on a very 

 artificial character, the fact that in this species, the bud-sheath is 

 partly open. Prantl divided the genus into groups nearly like 

 those used here, but differed in placing B. virginiamim with B. 

 ternatum and others of that type to form his subgenus Phyllo- 

 trychiiim, of which the principal distinctive character is the pos- 

 session of a hairy bud. But B. virginiamim seems to find its 

 closest affinities with B. lanceolatnm, and it is more reasonable to 

 suppose that it represents an extreme development of this type 

 than to relate it to the much more different B. ternatum group. 



In the key that follows, the difficulties in connection with the 

 two smallest species, B. simplex and B. pjimicola, are met by giv- 

 ing descriptions sufficiently complete to prevent the confusion 

 of mature specimens of other species with these, but it is prob- 



*Bot. Gaz. 40: 455-45S. 



