the types, and as the descriptions are misleading, he evidently changed his 

 opinion when he published the Icones, He referred to the specimens issued 

 by Drummond as H. amoenuin as a synonym. In Dr. Torrey's set, No. 196 

 appears to be a small form of S. recurvans mixed with Hypmmi pratense. 

 In the Musci Boreali-Americani, SuUivant and Lesquereux also issued two 

 numbers in each set, one as H. recurvans and the other as an unnamed 

 variety. In the first set I cannot see much difference between them, but in 

 the second set, No. 447 H. recurvans var. is evidently 5. delicatulum. 



Austin has suggested a resemblance in the coarser forms of 5. recur- 

 vans to H. nemorosum. I have only seen one specimen that at all approxi- 

 mated this species in size, and it suggested S. Carolinianum. They were 

 collected by Dr. Small in the Canyon of Tallulah Falls, growing on quart- 

 zite rocks ; they are bright glossy yellow or red-brown plants, with branches 

 7cm. long, and the pedicels 25mm. long; the leaves are strongly recurved. 

 It occurs also in lax, loose mats in the same locality where the normal form 

 was also collected. Austin also named some specimens in his herbarium 

 h. recurvans var. compactum ; they were collected at Otter Pond, near 

 Closter, N. J., forming densely caespitose tufts, fully three centimeters 

 high and matted together by brown tomentum. They evidently grew in 

 rich moist soil, as I have noticed that this species when growing on the 

 ground has a tendency to become pulvinate, while on roots of trees or on 

 rocks it spreads out in thin mats. He also named some specimens in his 

 herbarium H. recurvans var. minus ; these are referable to S. delicattiluvi^ 

 but owing to the strongly recurved leaves, and the misleading characters of 

 H. laxepatulum he did not recognize them for this species. In fact, 

 Austin failed to discriminate between them. His idea about H. recurvans 

 seems to have been that it was as variable a species as H. cupressiforme. 



Kindberg has described Raphidostegium Whitei as having the leaves 

 faintly or not recurved, the pedicel short and the tufts green. This agrees 

 with the form we have been calling H. laxepatuluiti, but I have not seen 

 the type. 



In geographical distribution, S. recurvans has the greater range, 

 having been collected in British America from New Foundland to Manitoba, 

 along the Alleghanies from Maine to Georgia, northwestward to Michigan, 

 Minnesota and Wisconsin. It appears to be unknown in the Rocky Mountains 

 and the West Coast, where it is replaced by S. Roellii. In Part VII. of the 

 Catalogue of Canadian Plants, Raphidostegium Roellii is credited to Cape 

 Breton Id. N. S. ; this determination needs investigation. In the immediate 

 vicinity of New York City 5. recurvans is not common, though it has been 

 collected on the Palisades by Austin and Oilman ; on Long Island by Dr.Orout, 

 and in Westchester County, near Bedford, by me. It is common in the 

 Catskills and the Mountains of Pennsylvania. S. delicatulum ranges from 

 Hudson's Bay to Manitoba, along the Alleghanies from Vermont to North 

 Carolina, is abundant in the Catskill and Adirondack Mountains as well as 

 the mountains of southern Virginia. It is apparently more restricted in its 

 range, occurring at higher elevations and in more moist shady localities. 



New York Botanical Gardens. 



