estern side an elongated bank of sand and gravel; then, where a
barrier of rocks stretched transversely across the stream and where
extremely bad rapids occurred--three of them in succession, each worse
than the last--was another island--Providence Island--1,400 m. in length.

When we reached any rapid we had to be quick in judging which was the
best channel to follow, as the current was so strong that we had not
sufficient strength to pull back against it. I generally selected the
channel, my men by this time having gained sufficient confidence in my
judgment, since so far we had had no serious mishap. But I foresaw that
we should soon have an accident, as they were getting foolhardy, and in
their ignorance attributed the wonderful luck we had had entirely to
their own skill in navigation.

On that particular occasion we had hardly time to recover from shooting
the first rapid with the velocity of an arrow, and were wet all over with
the splash of the water, when we came to the second and third rapids,
where the channel was so narrow and rocks were scattered so near the
surface, that it was really a marvel to me how we got through without
capsizing. The men in their excitement were shrieking wildly as we dashed
through the foaming waters, and there were also yells of positive terror
from the man ahead, who with a long pole in hand tried to save the canoe
from dashing now upon one rock then upon another.

Below the rapids the three other islands were Dora Island, 200 m. long;
Edna Island, 500 m. long; and Lucia Island, 700 m. long.

The river was flowing in a westerly and south-westerly direction, the
banks showing a quantity of rubber trees all along. A tiny islet 50 m.
long had been eroded from the right bank, just above a strong
_corrideira_, easily identifiable by later travellers who may visit it,
since a huge rock stands there in the centre of the river.

On the left side of the river foliated rock 10 ft. high was exposed for
the length of 1 kil. Dense forest was to be seen on both sides of the
river all along the rapids.

Two more islands, each 100 m. in diameter--Romeo and Juliet
Islands--close to each other, were then seen on one side of the main
channel, which was 200 m. wide.

From this point the river actually flowed in a S.S.W. direction (230 deg.
b.m.), and for 2,500 m. we had to negotiate strong and troublesome rapids
with variations of shallow water, usually with a bottom of sharp rocks.
The water