barrier of rocks
at the end of this island. Then no sooner were we thanking our stars that
we had negotiated that portion of our journey safely than we were among a
lot of globular boulders, some 30 ft. high.

For 800 m. we had a placid time, the water of the stream being so
beautifully green, so transparent, that we could see the bottom quite
clearly. Our happiness did not last long. We had more rapids and a great
rocky bank spreading from south-east to north-west right across the
stream, and forming in one portion an island.

We went down another strong rapid between great and dangerously situated
rocks and a large island. Then came another wonderful group of high domed
rocks, one of the great domes displaying a sharp northern spur like the
ram of a battleship. Next to it were three cylindrical rocks, just like
towers, one of which leant over the dome.

Yet another rapid was shot through with no misadventure, and when we came
to the end of a large island 4,500 m. long and 80 m. wide--Priscilla
Island--preceded by a smaller islet of sand and gravel, we arrived at a
direct stretch of 4,000 m. of river, flowing to the west. Another rocky
islet with an accumulation of sand and a lot of scattered rocks by its
side, then a high island, were passed on our right, and farther on we
found another great group of globular rocks at the point where Daphne
Island, 350 m. in length, began.

I hardly had time to map out the numberless rocks and islands we met
before we came upon others. There again we saw three more islands in
succession--Mars Island, 500 m. long and 100 m. wide; Jupiter Island, 250
m. long; and a third and smaller one, separated from the second by a
channel strewn with huge boulders.

To the N.N.W., at 340 deg. b.m., we saw a hill 300 ft. high, some distance
from the stream. Innumerable rocks again occurred in the centre of the
channel, and then we came to an extensive triangular island--Barretos
Island--the base of which was 300 m. Its left side was 2,000 m. long, its
eastern or right side about 1,500 m. A hill range some 300 ft. high was
looming before us to the north-east. The second island--Antonio Prado
Island--had a total length of 2,000 m. with an average width of 200 m.

On this magnificent island we halted at five o'clock in the afternoon,
and I took altitude observations with the hypsometrical apparatus: 1,062
ft. above the sea level.

We were again lucky in fishing that evening. We caught six _trah