REPOKT ON THE BOTANY OF THE ISLANDS OF THE SOUTHERN OCEAN. 145 



represented eastward of the Tristan da Cunha group, while six extend only in an easterly- 

 direction, some of them only as far as Amsterdam and St Paul Islands, others as far as 

 New Zealand and Australia. Two, namely, Apium austrcde and Nertera depressa, occur 

 where there is land all round the south temperate zone, and the latter beyond ; and the 

 doubtful Convolvulus is a widely spread species. Of the twenty-six ferns and lycopods 

 only six are peculiar to the islands ; fourteen are African, and sixteen are American ; two 

 are all round the south temperate zone, without touching South Africa ; and three are 

 restricted to this group, Amsterdam Island, the Mascarenes, and Madagascar. Blechnum 

 austrcde is so near the South American Blechnum hastatum in some of its forms as to 

 be indistinguishable, so we can hardly regard this as an African form. 



From this partial analysis of the constituents of the flora of Tristan da Cunha, it will 

 be seen that it consists of three or four distinct elements, no one of which sufficiently 

 preponderates to justify the assumption that it is essentially the original flora which has 

 subsequently been enriched by the other elements. It has been suggested that the flora 

 is mainly Fuegian, but the facts before us do not specially support such a view. They 

 indicate rather the former existence of a floral region distinct from the American and 

 Australian, and, if not African, at least more closely allied to that than to any other. 

 Remnants of this flora still exist in St Helena, Tristan da Cunha, Amsterdam Island, 

 and the Mascarene Islands, and in Eastern and Southern Africa. The distribution of 

 Phylica nitida and Spartina arundinacea, the prevailing plants in Tristan da Cunha, 

 favour this hypothesis ; yet there are perhaps equally weighty facts against it. The 

 strongest American affinity, at least to the present flora, is offered by Empetrum, which, 

 though abundant in these islands, is not known to occur eastward of them. Nertera 

 cannot be regarded as an American type, with an endemic Tristan species, and the other 

 Tristan species ranging all round the south temperate zone ; and the remaining species 

 of the genus more numerous in the eastern hemisphere, extending from New Zealand to 

 the Philippine Islands. The presence of Chevreidia stolonifera and Lagenophora eommer- 

 sonii, both comparatively rare, will hardly establish an American affinity. A majority 

 of the species of the flora of Tristan da Cunha are characteristic of the present general 

 flora of the south temperate zone rather than any particular part of it. But when we 

 have compiled the lists of plants for the islands eastward to Amsterdam, we shall be 

 in a better position for discussing this question. Among introduced plants, more or less 

 established, we have seen specimens of Gerastium triviale, Oxalis comiculata, Maloa 

 sylvestris, Gnaphalium luteo-album, Senecio vulgaris, Anagallis arvensis, Runiex aceto- 

 sella, Poa annua, Poa pratensis, Festuca myurus, and Festuca unioloides. Mr Moseley 

 mentions, too, that the Germans whom they found on Inaccessible Island had introduced 

 two species of clover, which were rapidly spreading. 



(bot. CHALL. EXP. PART II. — 1884.) .. li 19* 



