DERIVATION OF THE FLORA AND FAUNA 27I 



of boreal origin and has spread south. Pernettya is a bicentric genus, but it 

 is much better developed in America and may have originated with Gaultheria 

 in the Antarctic. The same should apply to Nicotiana according to Merrill 

 (jo<5. 310) who thinks that its actual distribution was attained in the Tertiary pe- 

 riod by way of Antarctica; however, N. cordifolia cannot be removed from II. i. b 

 and find a better place with I. i.a. On the other hand, a truly Antarctic genus 

 may have produced numerous species in one sector and few in another, or the 

 few may be a remnant of a larger population. In group I. 3 Peperomm berteroana 

 calls for attention. As I have shown (^^5) it is so close to P. tristanejish that 

 their common origin cannot be doubted and that little prevents us from regarding 

 them as forms of one species, very distinct but clearly related to the other two 

 species endemic in Juan Fernandez, and they point west. This is the reason why 

 all three were referred to I. Wahlenbergia (I. 4) is another mysterious case, but as 

 it is neither boreal nor palaeotropical or neotropical but South African (and on St. 

 Helena), only a far southern ancestry remains to explain the disjunction. 



Of the endemic genera belonging to group II. i only Juaiiia is taxonomically 

 isolated, Rhodostachys of Chile is brought to Ochagavia by some authors, and 

 Selkirkia comes very near Hackelia. With regard to the endemic Cichoriaceous 

 genera authors' opinions differ. Bentiiam looked upon the arborescent Pacific 

 Compositae as relics of an old Polynesian flora but did not refer directly to Ant- 

 arctica as source; Guppy {^121] quoting Bentham believed that, with the Hawaiian 

 Lobelioideae, they belonged to an ancient flora of the Pacific which had origi- 

 nated in America and gained dominance during what he termed "Age of Lobe- 

 liaceae" and "Age of Compositae", respectively. Antarctica seems not to have 

 meant anything at all to him, but as I have discussed his ideas in some detail 

 in an earlier paper [2/f.8), I shall not enter upon this subject here. 



Calystegia tuguriorum is doubtful case, but few people will be inclined to 

 think that it originated independently in Chile and New Zealand. 



In a paper of 1928 [^218) Setchell discussed what he called the two prin- 

 cipal elements in the Pacific flora, the Indomalayan and the "Subantarctic". With 

 this he did not understand what, from a geographical viewpoint, I call subant- 

 arctic, nor the species of old Antarctic genera found in the bogs of Hawaii, the 

 Pacific flora with which he was particularly concerned, but the group he later 

 {2ig) called "the Old pacific and antarctic element", for it included also the ar- 

 boreous Lobeliaceae and Compositae, which he believed had a common origin 

 in high southern latitudes. They had migrated north along different lines, the 

 Compositae taking an easterly course along the route Juan Fernandez-Hawaii, the 

 Lobeliaceae, which are absent from Juan Fernandez but have left traces in Poly- 

 nesia, a more westerly; a third line is called the Dammara (Agathis)-Podocarpus 

 line, an Australasian line running over Fiji. What interests us here is that the 

 isolated genera of Compositae were regarded as Antarctic. As we have seen, I 

 did not venture to include the Cichoriaceous genera, but I cannot assure that 

 Setchell was wrong. 



