390 D. Prain — The Vegetation of the Coco Grotip. [No. 4, 



Ifc will be seen that the majority of the species in this table are 

 actually present in one or other of the two volcanic islands of the An- 

 daman Sea, Narcondam and Barren Island, and even in these cases where 

 they are not pr-esent allied species are. There is a Bomhax in Narcon- 

 dam and though its specific identity or otherwise with the Andamans 

 one cannot be here discussed, it is evident that any Bomhax may be 

 wind-introduced. And whatever agency explains the presence of Bom' 

 hax will, pari passu, explain that of Eriodendron. 



Not a single orchid was found on Narcondam thougb on Barren 

 Island two were found — a species of Dendrobium on trees on the outer 

 cone, and Pholidota imhricata, which occurs at the top of the inner 

 cone within the crater-cup where the ground is kept moist by the con- 

 densation of escaping steam. Then the Hoyas are both present in great 

 abundance on the exposed I'ocks and tall trees of both islands. The 

 most doubtful species undoubtly are Aristolochia tagala, Gloriosa superha, 

 and, especially, the two species of Dioscorea. Yet these must all be 

 immigrant. The writer has collected Aristolochia tagala on Batti 

 Malv, a small outlying uninhabited fragment of the Nicobar Group, 

 on Barren Island, and on Narcondam. And even if it be claimed that 

 on Batti Malv the species may be a remanent one on the other two 

 islands it, like every other species, must be immigrant. The case of 

 Gloriosa superba has already been discussed when dealing with the 

 species introduced by the sea. The Dioscoreas are still more difficult 

 to explain, but it hardly seems as if they could be bird-introduced, and 

 it is almost as difficult to think that they have been introduced by the 

 sea. They are never littoral, being even in these islands strictly con- 

 fined to the higher dry ridges. Yet they are certainly not necessarily 

 remanent, for the writer has collected not these only but a third 

 species, Dioscorea bulbifera, or at all events a bulbiferous one, which is 

 present along with these two in great quantity in Narcondam and 

 especially in Barren Island. In both these islands the species must 

 all be immigrant and from the physiographical history of Barren 

 Island should there be, biologically speaking, extremely recently so. 

 Though no Strophanthus occurs in Barren Island, an Aganosma is 

 common there. The distribution of the Cryptogams of this class calls 

 for little remark, the peculiarities displayed in this respect by the 

 Fungi being probably altogether owing to this class being imperfectly 

 known in most floras. The pi'esence, for instance, of two species here 

 that are recorded only from North America probably implies that they 

 are both in reality cosmopolitan or nearly so. 



To the 68 species enumerated above should be added six imperfectly 

 represented Cryptogams, giving a total of 64 species ; the following table 

 contains an analysis of their distribution. 



