46 D. Prain — Flora of Narcondam and Barren Island. [No. 2, 



To the south-east of the oblong spit, and therefore on the east side 

 of the island, is a third, much wider bight, three-quarters of a mile 

 from cape to cape, but only receding a furlong and a half. The northern 

 half of this bay, bounded by the billy spit, is overlooked by steep hill- 

 sides ending in cliffs that, though not lofty, are particularly abrupt. The 

 southern half, limited by the main island-mass, has a beach of rounded 

 boulders ; behind this is a straggling sea-fence in which stands a solitary 

 coco-nut tree ; a narrow belt of true beach-forest lies beyond. It was 

 with little expectation of being able to land that we put into this bay ; we 

 were therefore agreeably surprised to find that — at least at the time of 

 our visit, the end of March — not only could a landing be made without 

 difficulty, but that the bay afforded a more comfortable anchorage than 

 Anchorage Bay itself. The boulder beach slopes rather gradually out- 

 wards, and is of a considerable width ; probably therefore the surf here 

 is very strong during the north-east monsoon. That the sea-fence is 

 here irregular and thin is no doubt due partly to the surf, and partly to 

 the fact that it has an insecure root-hold among the rounded stones 

 that are piled behind the beach into an embankment which protects 

 the forest beyond. This beach-forest occupies a strip of level land that 

 stretches backwards from 50 to 1 00 yards to the base of the main hill. 

 Three gorges debouch on this level area and have filled up the interstices 

 of the old beach with the soil on which the trees grow. At the mouth 

 of one of these ravines there is a gap in the beach-forest occupied by a 

 small depression that in March is covered with only a coating of fine 

 sun-cracked mud, but in the rains evidently forms a small lagoon ; 

 this appears to be the only spot in the island where water ever lodges. 



Though entirely volcanic in structure there is no indication at the 

 summit or elsewhere that the island has recently been active. There is 

 no crater at the top*, and his examination led the writer to think, not 

 that all traces of craterine shape have been obliterated by long erosion, 

 but that there never has been any crater on the peak. The local features, 

 coupled with the nature of the rocks that constitute the island,f 



* Mallet: Memoirs of the Geol. Surrey of India, xxi, 281. 



f Ball : Eecords of the Geol. Survey of India, vi, 90, only mentions a bed of 

 volcanic agglomerate, (of which several crop out round the coast), at Coco Bay, where- 

 in are embedded trachytic boulders. Mallet — Memoirs of the Geol. Survey of India 

 xxi, 281—283 — describes the Narcondam lavas as " compact, or very slightly vesicular 

 " lavas in which crystals of white translucent felspar, and black or dark-brown 

 " hornblende, are disseminated through a ground-mass which is (generally light) 

 " grey in unaltered specimens, but pale red in those that have undergone weathering 

 " and in which the iron has been peroxidised." Farther on, Mallet remarks : — " The 

 " lavas of Narcondam are essentially hornblende andesites, and are of a decidedly 

 " more acid character than those of Barren Island." This character] of acidity 



