S82 D. Praiu — The Vegetation of the Coco Group. [No. 4, 



This list includes 94 species for which sea-iutroductiou is conceivable, 

 and for the presence of most of the species it contains this mode of intro- 

 duction is almost certainly responsible. The list might even be made more 

 extensive than it is, for if Steo-cuUa rubiginosa, which is a purely " litto- 

 ral " species here as it is elsewhere throughout the Andaman and Nicobar 

 groups — to which area the variety found in the Coco Islands is strictly 

 confined, — be sea-introduced, there is no reason why some of the other 

 species of Stermdia should not bo added. As a matter of fact the writer 

 has collected specimens of species of StercuUa in Narcondam and in Batti 

 Malv, the first island a locality where certainly, the second one where 

 probably, every species present has been somehow or other introduced. 

 But no StercuUa seeds were recognised in the " drifts " and therefore 

 the whole of the species have been left out except this purely ' littoral ' 

 one, while even it has been omitted from consideration in the analysis 

 of the table which follows. Again, Leea hirta might well be sea-intro- 

 duced if Leea samhucina is ; their fruits are very similar and Leea fruits are 

 common in the " drifts." All the fruits found, however, were precisely 

 the same and seemed to be undoubtedly those of Leea samhucina, which 

 is a very common species in the mud flats that skirt the mangrove- 

 swamps, where it occurs as a considerable shrub or small tree with 

 stilted roots that imitate the style and appearance of those of the man- 

 groves. Both species, however, may have been introduced by fruit- 

 eating birds ; only one therefore, owing to its habitat, is taken as an 

 example of this mode of introduction, the other being relegated to the list 

 of species that are bird-introduced. Another species to which the same 

 remarks apply is Ardisia humilis, which is a purely beach-forest species 

 and, as such, is equally common here, on JS'arcondam, in the Andamans, 

 and in the Nicobars; perhaps it is, on the whole, more likely to have been 

 introduced owing to birds having eaten its purple-berried fruit. Allophy- 

 lus Cohbe, which is almost certainly bird-introduced, may be quoted in 

 support of this, for though it also occurs in the intei-ior it is a common 

 tree in the Pandanus fence and in the beach-forest. Lracontomelum man- 

 giferum might be a sea-introduced species, for Mr. Hemsley i^ecords a 

 Dracontomelum ? fruit from the New Guinea *' drift ", with empty seed- 

 cells however {Ghallenger Reports ; Botany, vol. i, part 3, p. 290). And 

 if Dracontovielam be included so might Spondias and Cavarium, for 

 though birds and bats eat the pulpy fruits of these species they cannot 

 swallow the stone and, as in the case of Terminalia Gatappa, can hardly 

 do more than assist in dispersing them locally. Desmodium triquetrum 

 and Desmodium polycarptim are both very common on the rocky parts of 

 the coast just above the spray-line and their fruits therefore are extremely 

 common in the " drifts." But ic is not at all clear that thej must therefore 



