13 259 



in the handbooks said to have wide ranges, e. g. "tropical America", "from the 

 West-Indies to Chile", "from Mexico to Brazil", etc. I find, however, that the area 

 of most species is much more restricted, and perhaps all grow within limited 

 regions only, beyond which the species does not go. The wide distribution of the 

 species of the handbooks, is, of course, connected with the fact of the accepted 

 species being collective species, each of which in my opinion consists of two or 

 several good species, their right of being species depending not only on their 

 specific characters but gaining besides considerably in force by the demonstration 

 that these specific characters are in close relation to the geographical range. 



The experiences I have gained by studying these species are in considerable 

 contrast to the view, that the ferns in their geographical distribution dilTer from 

 the flowering plants, the single fern-species being supposed not to be so closely 

 confined to limited floras. This view, which may be traced to Hooker-Baker's 

 manner of putting together related species in great collective species without attri- 

 buting even the most different geographical area any importance, pteridologists, such 

 as Christ and Underwood, have latterly proved to be absolutely inconsistent. Espe- 

 cially Christ has recently' strongly maintained that the ferns as a whole have the 

 same geographical distribution as the phanerogams, and my researches of the li- 

 mited group here dealt with have quite proved to me, that this view is the right one. 



The two regions in tropical America, West-Indies-Andes and South Brazil, in 

 the vegetation of which the ferns form an important feature, have a long series of 

 species peculiar to each, perhaps without one species common to both. If I in 

 the following pages have referred specimens from the West-Indies and Andes as 

 well from Brazil to some species {D. cheilanthoides, D. diplazioides, D. oligocarpa, 

 D. opposita, I), pachyrachis), this has been done with reservation, as will be 

 learned from the remarks under these species. In these cases, namely, the Bra- 

 zilian plants do not exactly agree with the andine ones, and even in some cases 

 it might possibly be right to separate them out as belonging to distinct species. 

 The connection between the floras of South Brazil and Andes-Antilles has appa- 

 rently been interrupted a long time ago, but it is not longer than that the original 

 ancestors have just reached to their separation into, so to speak, geographical sub- 

 species, a separation, which in the aforesaid rare cases has proceeded so slightly 

 that a specific separation is scarcely possible. In other respects the evolution in 

 the two regions has taken a different direction. In the West-Indies and in the 

 Andes from Mexico to Bolivia, the affinity of the floras is very intimate, as a 

 great number of species are common to both regions, which is probably owing 

 mainly to their proportionally late geological separation; in these regions the 

 group, which I call the group of D. Sprengelii, has been subjected to an intensive 

 development. The number of species belonging to this group is certainly conside- 

 rably larger than the 18 species enumerated below. In South' Brazil this group is 

 only very sparsely represented, only D. cheilanthoides may properly be referred 



') Denkschr. matli.-nat. Klasse d. Kais. Akademie Wien 7i); 1906. 



