1898.] Funafuti, Rotuma and Fiji. 459 



across, with caves here and there and numbers of small tunnels. 

 Where the rock is weathered, it is grey, rough, sharp and much 

 pitted, while its fracture is white, not uncommonly streaked 

 with red; the appearance presents no trace of organic structure. 

 When broken into, holes are often found, a few inches to some feet 

 in diameter, having rough and rugged sides ; other parts show 

 pockets of a red hard amorphous sandy rock, apparently very 

 largely consisting of ferric oxide. Low down, especially where 

 directly acted upon by the tide and spray, the rock is much more 

 crystalline, its rough projections being frequently entirely formed 

 of transparent, yellow, prismatic needles, or shorter and thicker 

 crystals. 



Where . exposed to the action of the waves, the rock is worn 

 away to an extraordinary extent, so that many of the rocks in 

 the lagoon are mushroom-shaped. The height at the edge above 

 low water is from 7 — 15 feet, depending largely on the waves and 

 hence on the exposure, the stretch of water that the wind may 

 have for raising a sea. On the seaward face of the island it is 

 generally low, as much wave action is impossible, owing to the 

 reef breaking the waves ; within the lagoon near the passage it is 

 greatest. The cliffs by the sand flat to the south-west have the 

 same undermining, proving that it was formed by the sea piling up 

 the sand. In the lagoon there is usually a flat platform at low 

 tide level, running from the base of the cliff to the outer edge of 

 the overhanging part, where it drops directly into 2 or 3 fathoms 

 of water ; on its surface are a few calcareous algae, a little weed 

 and some worm tubes. The heads of these mushroom-shaped 

 rocks in the lagoon are often very evenly convex ; occasionally one 

 can be seen by the wearing away of its stem to have fallen over 

 to one side. Not uncommonly large masses have split off from 

 the sides, and south of the entrance is one which has broken into 

 two across the middle. The portion of the platform, which should 

 have been covered by these split-off masses, is not, unless the fall 

 is very recent, visible and must have been washed or dissolved 

 away. Every stage from a well-formed islet to a small patch 

 2 — 3 fathoms deep can be seen ; no shoals appear to exist, which 

 cannot be ascribed to the solution of these islets. Within the 

 lagoon no corals and no reef-forming nullipores live, the bottom 

 consisting of a fine, soft, white, sandy mud, the deposit of which 

 on any coral would quickly prove fatal ; solution is going on con- 

 tinually and there is no balancing growth. The final result should 

 be a perfect atoll-reef just awash, the land being dissolved away 

 while the living reef round it continues to grow. (Plate.) 



The whole island is surrounded by a fringing reef, off Quoin 

 Hill very narrow, but to the south-west half-a-mile broad with 

 a well-marked channel, 2 fathoms deep; to the north-east it 



