28o C. SKOTTSBERG 



as limited by CoPELAND 1938 [66] does not mclude/a/k/andman {Mecodiuvi accord- 

 ing to COPELAND p. 94), though it is closely akin to peltatum, and H. rugosiini 

 "is to H. tunbridgense what H.falklandicum is to H. peltatuni' [62. 13). According 

 to Christ (jp. 146), peltatiun and tu7ibridgense are essentially Antarctic-andine but 

 found north as far as Mexico, also reported from S. Africa and New Zealand, ex- 

 tending in insular climates to western Europe. The conclusion would be that also 

 falklcmdiciim and rugosinn should be brought to the Antarctic group. 



Hymenoplivllion-Merhigiiirn secimdion, plicatuni and tortuosttni are subantarctic- 

 American in their present distribution; Meringiuvi is austral and extends north to 

 tropical Asia, and its origin most likely Antarctic. 



Thyrsopteris. Mesozoic fossils supposed to come near this genus have been re- 

 ported from the boreal zone; besides, from the Tertiary in Chile (see below p. 399). 

 Hill {134-. 1477) calls it "a remnant of a once widespread group which migrated 

 southwards, possibly in Cretaceous times, from the Northern Hemisphere". With 

 regard to its present unique station Copeland {67. 175) writes: 



As Juan Fernandez can hardly have evolved so peculiar a fern, its most reasonable 

 origin, as a Juan Fernandez fern, is Antarctica, whether directly or through southern 

 Chile. European fossils have been referred to this genus, but Juan Fernandez was not 

 colonized from Europe. 



Lophosoria is exclusively neotropical at present, but if we are to believe, as 

 Copeland thinks, that the Cyatheaceae derive from Antarctica, it is unlikely that 

 Lopliosoria would be an exception. 



The Antarctic origin of Dicksonia seems very clear to judge from its actual 

 distribution pattern. To what was said above p. 222 I shall add that along the 

 track New Zealand-Australia-New Caledonia-Fiji (-Samoa)-Malaysia 14 species are 

 found, along the Andean path 9, and that the species of Juan Fernandez are not 

 linked with the neotropical but with southwest Pacific species. 



The present distribution of the genus Cystopteris suggests a boreal origin. 



In its wide as well as in its more restricted sense Dryopteris is world-wide, 

 and Copeland's remark {67. 181) that he has little doubt that the group as a 

 whole is of southern origin (see also 6g. 122) is perhaps little more than a guess. 

 I think we can take it for granted that hundreds of species have evolved in the 

 neotropical region, where the single Juan Fernandez species D. hnjeqi/al?fo/w has 

 its relatives. 



Polysiiclium. Eastern Asia is the centre of greatest concentration ; the section 

 occurring there is, in Copeland's opinion, the least primitive, and we have a number 

 of southern bi- or tricentric forms, among them P. vesliimn coll. The endemic P. 

 berterianum is very close to the circumpolar adiantiforme . To quote CoPELAND : "The 

 case for an Antarctic origin of PolysticJnan is so clear that the evidence has long 

 been familiar even to those hesitant to draw the obvious conclusion" {67. 181); 

 and, "The distribution . . . testifies clearly to its Antarctic origin" [6g. 109). 



ArtJiropteris is the only fern genus found in Juan Fernandez which is absent 

 from continental America, but otherwise typically austral. Its Antarctic origin is 



