﻿Selaginella rupestris and its Allies. 



By Lucien Marcus Underwood. 



The variations of the species of Selaginella with many-ranked 

 leaves have long been a puzzle to botanists if we may judge from 

 the numerous herbarium and MS. names that have been given to 

 various members of the group. Two clearly marked species from 

 North America have been separated from the tangle already ; be- 

 sides these there remain the boreal S. selaginrides on which Beauvois 

 founded the genus, and which needs no consideration here, and the 

 widely varying forms that for the past forty years have found an 

 unsatisfactory resting place under 5. rupestris. For many years it 

 has been evident that we have in this group several species as well 

 marked as those which have been segregated already. In 1889 

 one of these species was found growing at Pasadena, California, and 

 when in 1 89 1 another was seen growing erect in sand fields in cen- 

 tral Florida, the polymorphous nature of the group was even more 

 forcibly impressed on us, and with it the impossibility of main- 

 taining such divergent forms under a single species. Recognizing 

 the fact that without abundance of material at hand and especially 

 without consulting the lamest collections not only of America 



fc> " ,l - ""!-> 



• uld 



but also of Europe, the description of new speci< v 

 be a hazardous matter, we have hesitated until the great mas 

 of material, especially at Kew, could be satisfactorily examined. 

 meanwhile continuing to refer the various forms provisionally as 

 variations of S. rupestris. We are now convinced that it is more 

 logical as well as more scientific to designate these clearly marked 

 forms by specific names. Even now, however, we are obliged 

 to leave several well marked forms undescribed for lack of suffi- 

 cient material and shall call attention to them briefly, in the hop 



that our field collectors in the west, in the southw. t and in Mexi - 

 may give them better attention. The larger number of the spec. 

 proposed below are comparatively lo- -I in their d tributi n : most 

 of them are confined to the r« [ion w< t of the Mi ]'<• 



(125) 



