ON INVESTIGATIONS IN THE INDIAN OCEAN. 201 
The structure of the rock varies considerably, but as a whole is remark- 
able for the small amount of metamorphosis which has taken place. 
Even on the surface large fields of coral in natural position were fre- 
quently noticed, while on the face of the cliffs it was always obvious 
that the corals were stil] essentially in the same position as when under 
the sea. Rock consisting of a conglomerate of reef débris was not 
common, but was always more prevalent near the lagoon shores, point- 
ing to the probability that the central portions of the atoll were dead 
before elevation. Metamorphosed rock was found in various parts of 
the western half of the atoll, and always contained a considerable pro- 
portion of phosphatic matter, derived presumably from guano. Of the 
latter substance, considerable deposits exist near the lagoon shore on 
the west, and in large natural pits in the rock scattered over the atoll. © 
The lagoon is very shallow, being only a few inches deep at low 
spring tides; the bottom is covered with a very fine white mud, which, 
on account of the small depth, is churned up by strong winds, making 
the water quite white and piling the shores with foam. The amount of 
material carried to sea by each tide must be very large. This fact, taken 
in conjunction with the obvious erosion of the cliffs which form the 
shores of the lagoon, points to a rapid increase in size of the latter. 
Islands in the lagoon only occur near the pass, and it seems certain that 
much of the atoll depression must have existed at the time of elevation. 
These features of the lagoon can only be explained by supposing that 
the pass is of very recent date, an hypothesis borne out by the character 
of the latter itself. It is narrow, with no marked channel, being almost 
dry at low tide. The bottom is rocky, and there is no live coral either 
in it or in the lagoon. The reef outside is narrow, and but slightly 
channelled by the escape of water from the pass. It is not composed 
of modern reef substance, but apparently of elevated coral rock. Live 
coral is absent, though a certain amount of Lithothamnia exists. A 
cursory examination of the reef in other places pointed to its structure 
being the same, though a greater amount of sand and Lithothamnia- 
covered débris were present, often bound together by beds of Cymodocea. 
Variations in level across the atoll were difficult to estimate, but the 
highest rock level (12 feet above high-tide level) seemed near the sea, 
with a varying gradient to the lagoon. The seashore is sandy on the 
west, with cliffs appearing here and there. The east coast I could not 
visit owing to lack of time. 
The land vegetation consists of dense scrub, composed of large 
bushes and small trees near the lagoon, though to seaward it is formed 
rather by small shrubs and matted herbaceous plants. A small collec- 
tion of plants, as also of animals, was made, but the short duration of 
the visit coupled with the season prevented any thorough collecting being 
attempted. The settlement is situated on the west coast; it is small, 
and the inhabitants are employed in catching turtle and cultivating a 
small coconut plantation, which is being tried, while a certain amount of 
maize and tobacco is grown on the share system. 
COSMOLEDO. 
This group was reached on September 1. It is an atoll also, but 
differs from Astove in having only a small proportion of its rim capped 
