DERIVATION OF THE FLORA AND FAUNA 259 



guay, the fifth, J. planifolius, shared by Chile and Austraha-Tasmania-Xevv Zea- 

 land, but it is also true that even if we look for the origin of the family in 

 Antarctica, a possibility pointed out by Weimarck [281), we must count with 

 secondary centres of evolution in the boreal zone where the overwhelming major- 

 ity of sections and species are found. Some 50 species are in the South hemi- 

 sphere with main centres in South Africa {20) and Australia (17). 



Libertia with 3 species in Chile and 2 in New Guinea-Australia-Xew Zealand 

 tells a story of an Antarctic past, and the group Sisyrinchineae has a stronghold 

 in the South, where the genera concentrate; Sisyrinchiuvi itself centres in South 

 America, where 4 small genera are endemic [Syj)iphyostenwn, Chaviaelum, Sole- 

 noinelus and Tapemia)\ South Africa is another stronghold [Aristea, Witseniay 

 Bobartia, Klattia and Cleanthe), and 3 {OrtJiosantJius, Diplarrhena and Palersojiid) 

 are found in the East Australian - Indomalayan region, but this does not entitle us 

 to deri\e Iridaceae from the far south. 



Of the 4 island species of Peperouiia only /-'. fernandeziana (Chile) is neo- 

 tropical, while the other species appear to be more nearly allied to palaeotropical 

 ones (Java, Australia, Oceania); P. bertcyoaiia occupies a unique geographical posi- 

 tion (p. 203). The genus is of tropical origin and centres in America, but if we 

 try to understand the history of the endemic species of Juan Fernandez, the possi- 

 bility of Antarctica as a migration route should be considered. 



Whereas Urtica masafjierana, one of the few indigenous annuals, is related 

 to a species from Ecuador, U . fernandeziana with the habit of a miniature tree 

 [22g. 862) appears to lack near relatives. The family is, I suppose, of tropical 

 origin, but the actual centre of Urtica is in the north temperate and subtropical 

 zones; there are 14 species in Eurasia, including the Mediterranean region, and 

 7 in North and Central America with a secondary centre in the Andes (Colomb.— 

 Fueg. 6); 3 species are tropical (Braz.— Urug., Ethiopia, Java). There are a few 

 species in the south hemisphere, 2 S. Afr., i Australia, 2 N. Zeal, and i Auckl. Is. 

 U. fernandeziana seems to represent an ancient type. 



Boehnieria excelsa was described by BuRGER (.//) as an elegant shrub which came 

 from S. America; it is a clumsy tree and not related to neotropical species. It 

 comes nearest to B. dealbata (Kermadec Is.) and points west, not east. 



Parietaria humifusa (Chile); see above p. 204. Belongs to a neotropical group 

 which has not been cleared up [241) but is also closely related to the Australian- 

 Polynesian debilis of FOR-STER. Weddell's debilis is a mixture of varieties scat- 

 tered over the globe and most likely consists of several good species. 



Phrygilantlius Berteroi. The family Loranthaceae is tropical, but /^/^r)'^//r7?//////x 

 is of Antarctic origin. Of the 7 sections, 4 (about 20 species) are American and 

 range from Lower California to S. Chile, i with 2 species belongs to Australia 

 and New Zealand, i (i species) to New Zealand, and i has one species in Aus- 

 tralia, another in New Guinea and a third in the Philippines — a typically austral- 

 bicentric genus. 



The genus Santaluni ranges from Australia and ^Melanesia to Micronesia and 

 Malaysia and east to Polynesia and Hawaii with a distant outlying station in 

 Juan Fernandez; see Tuvama's map {2j6). Of the 4 sections distinguished by 



