114 62 



rich in species occurring in tropical America; each new collection received con- 

 tains new forms. As the key in my "Revision" includes 82 species only I have 

 worked out an entirely new key including all the species known to me. After the 

 key follows a systematical enumeration of all species including descriptions of 

 several new species or of species not seen, as I worked out my "Revision". As 

 will be seen the order of the species is not exactly the same as in "Revision, and 

 further that I have not arranged the species in sections and subsections. I think 

 that my present arrangement is a natural one in that sense that species intimately 

 related to each other are placed side by side. It is possible to separate out smaller 

 groups of closely related species, but such groups are connected with other groups 

 by several intermediate species that a thorough grouping of all the species should 

 be very defective. As rather distinct small groups I shall here mention 



1°) the group of D. rndis, which includes the species 151—160 and probably 

 others. They are large-grown species with generally dark-green, coriaceous or papy- 

 raceous, more or less hairy but always eglandulose lamina, which downwards is 

 suddenly narrowed, the lower 3 — 5 pinnæ being fully abortive and like small warts 

 along the stipe; the basal segments of the lower pinnæ are similarly reduced; aërophore 

 present; sori exindusiate. In most species the rhizome seems to be creeping. This 

 group appears to be a most distinct one, its species being widely different from 

 those allied to D. oligocarpa and D. opposita. Still the difference between these and 

 such species as D. Sprengelii and Ü. Christensenii is only small and these species 

 again are closely allied to D. panamensis, D. pachgrachis, D. tablaziensis and others, 

 which no doubt are intimately related to D. opposita and D. oligocarpa. On the 

 other hand D. rudis and its relatives are connected with the large bipinnate species 

 D. pteroidea by a species as D. euchlora, and it is perhaps unnatural to place 

 species as D. Thomsonii, D. macradenia in another subgenus, Glaphgropteris, as they 

 are as to essential characters very near D. rudis. 



2°) the group of D. cheilanthoides, including the species nr. 165 — 170. Large 

 species of thick texture and numerous veins; glanduliform pinnæ as in the former 

 group and aërophore as a rule present. Rhizome erect; hairs long, soft, pluri- 

 cellular; underside often glandulose and viscid; basal segments of most species not 

 much reduced, the lower one often prolongated; sori often indusiate. Species of a 

 characteristic texture and colour, which I can not explain. This group is through 

 D. limbata and Ü. consangiiinea connected with D. opposita and D. sancta. 



3°) The group of Z). rivularioides, including the species nr. 132—137, characte- 

 rized by a long-creeping rhizome and occasionally furcate veins. All the species 

 are from South-Brazil and adjacent countries. 



4°) The group of D. sancta, including the species nr. 56 — 62, small species of 

 thin texture and often with unequal-sided pinnæ. 



