DERIVATION OF THE FLORA AND FAUNA 267 



to laml this is the only course to take, even if the sudden appearance (jf an iso- 

 lated type like Hookeri become?, somewhat embarrassing. With an extension north 

 and east of an Antarctic borderland Co/>rosina could pass as of Antarcto-tertiary 

 origin. 



C. pyrifolia offers a good example of an incorrect taxonomic position leading 

 to false conclusions. It was described as Psychotria and referred to a neotropical 

 group: it came, as BCrger writes (./z. 20), "von den Urwaldern Perus und Co- 

 lombias". 



Of the more than 500 species described under Gal'nim about 400 are distri- 

 buted over the boreal zone and some 50 or 60 are known from South America, 

 extending from the tropical Andes to Fuegia, the Falkland Is. and South Georgia; 

 about 30 are African, the remaining species being divided between India, Malaysia, 

 Australia, New Zealand, etc. Galium is a boreal genus with a strong represen- 

 tation along the Andes, a not uncommon case. G. inasaftiero7itan is allied to 

 species from the mountains of Central Chile, probably also to species foimd farther 

 north. Nothing indicates that Antarctica ever had a share in the history of this 

 genus. 



Whether Campanulaceae-Lobelioideae evolved in the tropics or in the far South 

 will not be discussed here. The remarkable concentration of arboreous genera, 

 two of them large, in Hawaii, has given rise to much speculation, and the occur- 

 rence of other endemic genera in Polynesia as well as the Australian affinity of 

 BrigJiauiia of Hawaii have led some authors to look for the origin of the sub- 

 family (or family) in the Antarctic. In our special case we can leave this question 

 aside because Lobelia alata is a widely distributed seaside species, tricentric in 

 the South Hemisphere. 



The distribution of \Vahle7ibergia, a large and widespread genus, is inter- 

 esting. It is essentially southern, of the about 230 species described 150 are South 

 African and 20 tropical African, 6 are reported from Madagascar and the Masca- 

 rene islands. North it extends to the ^Mediterranean, the Orient and Asia, together 

 some 25 species. Scattered species are known from New Guinea (i), Australia (7), 

 New Zealand (7), Lord Howe I. (2) and St. Helena (3). Species are few in America 

 (North Amer. i. South Amer. 8); a single species, ]V. gracilis, ranges over the 

 south hemisphere. 



If we look at the related genera, some 13 in number, the dominant position 

 of South Africa becomes still more conspicuous: 7 are exclusively or preponder- 

 antly South African, i is from tropical Africa, E. Indies and Brazil, 4 Asiatic 

 and I from S.E. Europe, all according to the old synopsis in Natiirl. Pflanzenfam- 

 Our island species are, as we have seen, not matched in Africa, nor in America 

 or other parts of the world except on St. Helena, see above p. 210. All known 

 facts suggest that the genus had a long and complicated history in the far south 

 and that S. Africa is a secondary centre where evolution has been progressive 

 and prolific in species formation. It is important to remember that our island forms 

 are not a group of closely related taxa but that W. Masafuerae and particularly 

 Berteroi stand apart from the rest. The distribution of the genus can hardly be 

 ■explained without admitting Antarctica into the picture. 



