1891.] D. Prain — The Vegetation of the Coco Group. 293 



is very characteristic, more so than the trees, consisting of Ccesalpima 

 Nuga, Capparis sepiaria, and, very lai'gely, of Pisonia aculeata. In 

 the more muddy soil which occurs on the outskirts of the mangrove- 

 swamps other shrubs and creepers occur ; such as Leea samhucina with 

 stilted roots like the mangroves, Gynometra ramiflora, Hibiscus tiliaceus, 

 Flagellaria indica, Mncuna gigantea, remarkably common, Sarcostigma 

 edule, Plecospernum andamanicum, Antitaxis calocarpa, Salacia prinoides, 

 which extends also into the swamp proper, Acrostichum scandem, etc. 

 Not infrequent in such situations, when there is no high forest over- 

 head is Vitex Negundo which is particularly common on Little Coco. 

 In this muddy tract the tall trees remain much the same as in the 

 drier area just behind the beach. Further inward the vegetation is that 

 characteristic of a true mangrove swamp, Bruguiera, Geriops, Bhizophora^ 

 Aegiceras, Avicennia. The Avicennia, strangely, does not appear to be 

 common in many of the creeks, though there is one creek, on the east 

 side and near the south end of Great Coco, in which it is the prevailing 

 tree ; except indeed for a few Bruguiera gymnorhiza trees along the 

 open channel of the creek, the whole swamp consists of Avicennia 

 officinalis with thousands of its curious roots protruding through the 

 mud and water as described already in a former paper (/. A. S. B. 

 vol. lix, p. 272) ; considering the situation and loose structure of these 

 roots, which are of the consistence of solah--pith, there seems every pos- 

 sibility that they are concei'ned in the process of transpiration ; the large 

 area covered by the roots of each tree must also afford great stability to 

 a species which affects, as this one does, the situation of the mangroves 

 without having their stilted roots. In this particular swamp each tree 

 was loaded with the climbing form of Salacia prinoides and, as the latter 

 happened to be in flower at the time of the visit, the foetid nature of 

 the atmosphere expei-ienced may be imagined. 



Between the headlands, in most cases, a choked-up creek is to be 

 found ; generally this extends but a short way into the jungle, though 

 sometimes it winds about on the level ground for a considerable distance 

 as a mangrove-swamp. In two places the creeks on Great Coco are 

 apparently open at all times to the tide ; the chief creek is that which 

 debouches at the north end of the island. There are no open creeks in 

 the other two islands, though at the south-west corner of Little Coco 

 what has been a creek of considerable extent is now converted into 

 a large lagoon by a broad bank of sand having been blown and beaten 

 up by the south-west monsoon into a firm embankment across its former 

 outlet. 



On the ridges the trees are much the same, as to species, as on the 

 lower ground, except that the Miliusa which is common below is scarce 



