1891,] D. Pvain— The Vegetation of the Coco Group. 393 



there all occur in India and Indo-China also, while two that occur in 

 India and Indo-China but do not occur in Malaya must have been intro- 

 duced from the north. This being the case the probability is that the 

 others have mainly been introduced fi'om the same direction, a cir- 

 cumstance quite in accordance with expectation, since it is from the north 

 that the stream of migration of marsh- and water-birds annually flows. 

 Durino" our visits to the islands snipe were found in the meadow near 

 the lake on Grreat Coco, while teal and other water-birds frequented the 

 lake itself and abounded in the lagoon on Little Coco. 



Table XX, Analysis of distribution of Marsh and Aquatic species. 



Present in both Hemispheres : — 7 



Cosmopolitan in the tropics: 6[ 



Nearly cosmopolitan (ahsent from Polynesia) 1( 



Confined to Eastern Hemisphere 



Africa, Asia, Australia 3 



Africa, Asia 2 



Confined to South-eastern Asia , 4 



ToTA.1, species probably introduced by water-birds , 16 



The second kind of species that may be introduced by becoming 

 attached externally to birds is somewhat more difficult to deal with. 

 TJrena lobata, which is here clearly not a weed, may have been introduced 

 in this way : its fruits sticking, burr-like, to the feathers of some bird ; 

 Biiettneria andamanensis, might also have been thus introduced, though 

 this is not so probable as in the other case. Three of the Desmodia — • 

 Desniodium triquetrum, D. laxijlorum and D. poly carp on — may very well 

 owe their introduction to this mode of dispersal. Boerhaavia repens, as has 

 already been said, is probably sea-introduced, though there is no reason 

 why it may not partly owe its dispersal to bird-agency. Its habitat 

 on these islands is always the rocky headlands or isolated rocks along 

 the coast on which sea-birds sit to devour the Grapsus crabs they capture 

 on the wave- washed ledges below, and nothing is more likely than that 

 the fruits may become at times attached to their feet and be carried at 

 least from point to point along the coast. The Pisonias may both very well 

 have been introduced in this fashion, though it is less likely as regards 

 P. aculeata than as regards P. excelsa. From what has been already 

 said of this tree in discussing it among the " littoral " species, it will 

 be evident that its fruits ai-e of such a nature as to admit of their being 

 carried for great distances attached to a bird's feet or body, if only the 

 bird should happen to come in contact with them, and the objection that 

 scraping- birds, which might do so, are not often migratory, while frugivor- 

 ous birds, which are migratory, would not come in contact with the fruits 

 because they are not likely to alight on a Fisonia, is not a valid one. 



