entire day in the water, struggling to take
the canoe down the rapid and up once more. By eight o'clock at night we
were still working, endeavouring to save the canoe.

We had had no lunch, and now had no dinner. My men felt perfectly
miserable, and in their speech did not exactly bless the day they had
started with me on that expedition. We had worked hard, and had only
covered a distance of 7,500 m. in twelve hours. At sunset, while the
storm was raging, we beheld a most wonderful effect of light to the west,
very much like a gorgeous aurora borealis. The sky, of intense vermilion,
was streaked with beautiful radiations of the brightest lemon-yellow,
which showed out vividly against the heavy black clouds directly above
our heads. The river reflected the red tints, so that we appeared to be
working in a river of blood.

As we had nothing to eat, I thought I would spend my time in taking the
correct elevation of that place with the boiling-point thermometers. The
man X, the humorist of the party, remarked that if I were killed and went
to Heaven or some other place, the first thing I should do would be to
take the exact elevation with what he called "the little boiling stove"
(the hypsometrical apparatus).

We had a minimum temperature of 62 deg. F. during the night of August 10th.

Next morning I sent my men to reconnoitre, in order to see if they could
get some edible fruit. As they stayed away a long time I knew they had
found something. In fact, they came back quite in a good humour, as they
had found some _jacoba_ or _jacuba_ trees, with abundant fruit on them,
most delicious to eat.

In the meantime I had gone exploring the rapids endeavouring to find a
more suitable channel. Eventually, on the east side of the stream, I
found a place where we could take the canoe down. There too was a fall of
9 ft., down which we let the canoe with considerable difficulty; then it
had to pass over a number of smaller terraces and down winding channels,
where we sweated for some hours before we got through our work.
Innumerable channels separated by sand-mounds 20 to 30 ft. high had
formed along that rapid and also through the vertical wall of cutting
volcanic rock which formed a barrier across the stream. Below the fall
were two long sand-banks, one with some _burity_ palms upon it.

The river flowed 20 deg. west of north for some 4,000 m. We had gone but
2,000 m. of that distance when we came to another rocky barrier,
sp