1891.] D, Praia — The Vegetation of the Coco Group. 385 



of further evidence, however, it is treated as only doubtfully sea- 

 introduced. Another doabtful species is Dodoncea viscosa, a cosmopolitan 

 species. Still another, equally doubtful, is Gloriosa superba which is ex- 

 ceedingly common in the coast zone on both the Coco Islands, and which 

 the writer has collected, in the coast zone also and only there, in South 

 Andaman, in Rutland Island, in Batti Malv, in Car ISTicobar, in Narcon- 

 dam, and in Barren Island, and which Dr. Alcock has collected, near the 

 sea, in the Laccadives. On the whole therefore we might feel justified 

 in considering it a sea-introduced species. But it is very abundant also 

 throughout the whole of India ; it extends from the Nilghiris and 

 Central India to Rajputana, the Panjab, and the Gangetic plain, as well 

 as to the Himalaya from Kamaon to Bhutan, and is common in Bengal, 

 Assam and Burma. It cannot very easily be bird-introduced and one 

 must therefore incline to the opinion that the agency responsible here 

 is that of winds, a view which is favoured by the nature of its seeds. 

 But even then it is not easy to suppose that winds could carry these as far 

 as some of the islands mentioned and still that its distribution should be 

 limited to South-Eastern Asia. Oroxylum indicum might possibly be 

 sea-introduced, but on the whole has more probably been brought by 

 wind. It need not be indigenous for it occurs in abundance in Narcondam. 

 Though its fruits occur in the " drifts " they are always split open and 

 it is unlikely that the seeds could remain attached to the fruit-segments 

 during their transit from any of the neighbouring coasts. 



Few of the cryptogams can be considered " littoral " and the state- 

 ments that have been made of the possibility of Fungi, etc., being brought 

 to ocean-islands attached to logs of wood or trunks of trees are not as a 

 rule made by those who have seen and carefully examined ocean-drifts. 

 Even Polyporus sanguinale, which apparently has a prediliction for dead 

 or dying trunks of Cocos nucifera, being commoner there than in any 

 other situation, was not found growing on any of the trunks that lie on 

 the beaches exposed to the sun after having been soaked in salt water. 

 The logs that are cast up on the beach and the roots that protrude from 

 the sand at those points where denudation is going on, are scrubbed 

 bare by the coral-sand and bleached white by the sun ; they harbour no 

 Fungi and seem preserved from decay by the treatment to which 

 they have been subjected. There is, however, a striking exception in a 

 " dry-rot " which attacks Mimusops littoralis trunks and some other 

 timbers. In the case of the Bullet-wood it was seen both on Great and 

 Little Coco ; the same appearance was presented by the remains of 

 a wooden vessel in Little Coco. The appearance and consistence of 

 this "dry-rot" so closely resemble the results of charring that it was 

 difficult to realize tliat the wood in question had not been subjected to 



