isles, such
as are in the sea; but I do not know that any thing has been said of high
islands, or such as I have been speaking of. In this island, not only the
loose rocks which cover the surface, but the cliffs which bound the shores,
are of coral stone, which the continual beating of the sea has formed into
a variety of curious caverns, some of them very large: The roof or rock
over them being supported by pillars, which the foaming waves have formed
into a multitude of shapes, and made more curious than the caverns
themselves. In one we saw light was admitted through a hole at the top; in
another place, we observed that the whole roof of one of these caverns had
sunk in, and formed a kind of valley above, which lay considerably below
the circumjacent rocks.

I can say but little of the inhabitants, who, I believe, are not numerous.
They seemed to be stout well-made men, were naked except round the waists,
and some of them had their faces, breasts, and thighs painted black. The
canoes were precisely like those of Amsterdam; with the addition of a
little rising like a gunwale on each side of the open part; and had some
carving about them, which shewed that these people are full as ingenious.
Both these islanders and their canoes agree very well with the description
M. de Bougainville has given of those he saw off the Isle of Navigators,
which lies nearly under the same meridian.

After leaving Savage Island, we continued to steer W.S.W. with a fine
easterly trade-wind, till the 24th in the evening, when, judging ourselves
not far from Rotterdam, we brought-to, and spent the night plying under the
top-sails. At daybreak next morning, we bore away west; and soon after, saw
a string of islands extending from S.S.W. by the west to N.N.W. The wind
being at N.E., we hauled to N.W., with a view of discovering more
distinctly the isles in that quarter; but, presently after, we discovered a
reef of rocks a-head, extending on each bow farther than we could see. As
we could not weather them, it became necessary to tack and bear up to the
south, to look for a passage that way. At noon the southernmost island bore
S.W., distant four miles. North of this isle were three others, all
connected by breakers, which we were not sure did not join to those we had
seen in the morning, as some were observed in the intermediate space. Some
islands were also seen to the west of those four; but Rotterdam was not yet
in sight. Latitude 20 deg. 23