266 C. SKOTTSBERG 



with this species and with solanifolia gave evidence of fundamental affinities between 

 the island endemic and species of the mainland to the north (I.e.). 



Mimidus is essentially a western N. American genus with few species else- 

 where; the somewhat polymorphous M. glabratus ranges from N. America to 

 Bolivia, Chile and Argentina. 



The different opinions on the systematic \)o?,\Uon o^ Ei/phrasm formos/ssw/a 

 have already been referred to above. Within the area of the geographically isolated 

 Chilean-Magellanian Trifidae occurs the semicalcarate E. perpusillaVWxX. (S. Chile). 

 Both would indicate a road from the Australian-New Zealand area across the Ant- 

 arctic to America, just as the tropical mountains of Malaysia served as a road 

 between Asia and Australia, as Du RiETZ thinks (77. 536): "The Euphrasia popu- 

 lation of Juan Fernandez may therefor very well have formed the northern end 

 of a population so far south that the lack of close relationships between E.for- 

 mosissima and the species of Middle Chile is fully explained, and the semical-. 

 carate anthers of the more southern E. pejpusilla may be the last South American 

 remnant of this old connection." Another question, not yet answered, is this: are 

 the very well-marked Trifidae, reaching from subantarctic America to the Andes 

 of S. Chile, likewise descendants from a remote southern population or did they 

 originate in Chile. ^ 



Planiago feriiandezia is another arborescent member of a herbaceous world- 

 wide genus and often cited as an example of a mysterious connection between 

 Hawaii and Juan Fernandez, but if we look at the total distribution of sect. Palaeo- 

 psy Ilium (see above p. 209) its character of an austral group is revealed, even if 

 it extends north to N. America and S. Europe. In the south it is circumpolar, 

 and the route from New Zealand (Auckland Is.) via Rapa to Hawaii can be traced 

 — a radiation from Antarctica seems not unthinkable. 



Plaiitago trmicata, represented in Juan Fernandez by an endemic (?) variety, 

 belongs to the large section Novorbis and needs no further comments. 



The neotropical Hedyotis ihesiifolia is a recent addition to the flora and may 

 have been accidentally introduced. 



Nertera is an austral-circumpolar genus allied to Coprosiua and best devel- 

 oped in New Zealand where 5 species occur, 4 of them endemic; the fifth is 

 A^. granadensis (depressa), claimed to be very widely spread. One species is en- 

 demic in Tristan da Cunha. Recent observations tend to show that granadensis 

 includes taxonomically distinct forms; the Malaysian plant is not identical with 

 the Andean, and other forms will perhaps become distinguished after a critical 

 revision. Be this as it may, the genus is Antarcto-tertiary and, if we link Tristan 

 da Cunha with Africa, tricentric. 



Coprosma is a parallel to Ha/orr/iagis but differs in being present in Hawaii; 

 another difference is that the species of Juan Fernandez do not have their closest 

 relatives in New Zealand or Australia^ — Hookeri forms its own section, pyrifolia 

 is of Polynesian affinity. As the genus is absent from America as well as from 

 Africa it should lie near at hand to refer it to a West Pacific element, a position 

 favoured by the relationships oi pyrifolia, and to the believers in transoceanic 

 migration combined with evolution of local endemics wherever Coprosnia happened 



