456 SUBANTARCTIC ISLANDS 0]F NEW ZEALAND. [Systematic Botany. 



larger than any of the above, being about ninety-five miles long by twenty miles in 

 greatest breadth. It forms a high mountain-chain, rising very steeply out of the sea, 

 and many of the peaks attain a height of over 6,000 ft. A large portion of the in- 

 terior is occupied by extensive snowfields, and glaciers discharge into all the fiords. 

 As there is little low-lying land or level soil, there are few suitable stations for a 

 phanerogamic flora, even if the cold, wet, and exceptionally boisterous climate did 

 not necessarily limit plant-growth. The island has been twice carefully explored — 

 first by the German Expedition of 1882-83, and then again by Nordenskiold's 

 Swedish Antarctic Expedition of 1902-3 — but only fourteen phanerogams and 

 four pteridophytes have been observed, not one of them being endemic. 



The Sandwich Group, situated to the south-west of South Georgia, and consist- 

 ing of about sixteen small islands, has never been botanically explored. Cook, who 

 discovered the group, mentions that he saw two hills free of snow, apparently covered 

 with green turf, so that probably there is a scanty phanerogamic vegetation, doubt- 

 less composed of some of the grasses found in South Georgia. 



The small group of the Bouvet Islands, situated in the same latitude as South 

 Georgia, and about midway between it and Marion Island, is also very imperfectly 

 known. For many years its existence was looked upon as doubtful, but in 1898 

 the German expeditionary ship " Valdivia " rediscovered it, fixing its position as 

 54° 26' S. lat. and 3° 24' W. long. Unfortunately, it was impossible to effect a 

 landing, but no sign of vegetation could be discerned from the ship. 



Far to the westward of Cape Horn — in fact, almost a third of the distance between 

 it and the meridian of New Zealand — is Dougherty or Keates Island, one of the most 

 isolated spots in the Southern Ocean. Captain Keates gives its position as 59° 20' S. 

 lat. and 120° 20' W. long., which would make it nearly 1,700 miles from Cape Horn 

 and about 2,500 from Campbell Island. It has never been landed upon, and nothing 

 whatever is known of its physical features and natural productions. Its position 

 with regard to New Zealand and South America is exceptionally interesting, and 

 it is much to be wished that some of the expeditions now regularly visiting the Ant- 

 arctic Continent would give it a careful exploration. A group called the Nimrod 

 Islands is sometimes marked on our charts about 1,000 miles south-east from 

 Antipodes Island, but it is very doubtful if it has any real existence.* 



The general character of the vegetation of the islands described above has been 

 vividly portrayed by Sir J. D. Hooker in his memoir on the vegetation of Kerguelen 

 Island (Phil. Trans. Roy. Soc, vol. 168, p. 10), where he says, " The three small 

 archipelagoes of Kerguelen Island (including Heard Islands), Marion and Prince 

 Edward Islands, and the Crozets are individually and collectively the most barren 

 tracts on the globe, whether in their own latitude or in any higher one, except such 

 as lie within the Antarctic Circle itself ; for no land, even within the north polar area, 

 presents so impoverished a vegetation." All students of botanical geography know 

 that this statement is by no means overdrawn. To mention a single proof out of 

 many that could be quoted, the Island of Spitzbergen, which is situated 30° nearer 

 the Pole (in N. lat. 77° 80') possesses 120 vascular plants, whereas Kerguelen has 



* Since this was written, both Dougherty Island and the Nimrod Group have been unsuccessfully 

 searched for by Captain Davies, of the Antarctic exploring vessel " Nimrod." 



