Another island, 300 m. long—Martia Island—with a picturesque spur of rock at its south-easterly end, was next reached as we were going swiftly down a corrideira in the channel to the right which we were following.
After the corrideira, as I was busy writing a description of the landscape, I was thrown off my seat. My men also had a similar experience, the canoe nearly turning turtle and becoming filled with water. Alcides had steered us right into the centre of a whirlpool.
These unexpected baths were not much to my taste—not so much for the discomfort they caused my person, as for the trouble they gave me in protecting my notebooks and instruments. Also, in these accidents we lost a considerable amount of our supply of salt, which melted away in the water, and the supply of flour and rice suffered from these unnecessary immersions.
A channel 30 m. wide separated Martia Island from a second island—Camilla Island—100 m. long, which must once certainly have formed part of it, but which had been separated by the eroding waters of the stream. Both islands were wooded, and were extremely pretty. Great heaps of rock, 20 m. in diameter and even more, occupied the centre of the stream after we had passed the last island.
We had only gone 12 kil. 300 m. that day, so difficult had been the navigation.
During the night in less than one hour we caught two large jahu, one huge pacu (Prochilodus argentius), the latter shaped like a sole, but of a much greater size, and with brilliant red patches on its body—a most delicately-tasting fish to eat—and a number of large trahira (Machrodon trahira), also called by the Brazilians rubaffo because of the noise they make in the water. Altogether over 200 lb. of fish were got out of the water in less than sixty minutes.
We found many jenipapeiros (or genipapeiro) trees, from the stewed bark of which we made excellent tea. Its fruit was good to eat, and we used it for making sweets.
During the night of July 19th the minimum temperature was 67° F.
We started off gaily enough in the morning, passing first a great boulder, 10 m. in diameter, sticking right out of the water; then an island 200 m. long contained in a basin 500 m. wide. We left the island—Ruby Island—which was 80 m. long, on our left, and went down a channel with strong eddies and whirlpools. Looking back at the eastern channel, we were glad we had not followed it, as it was extremely rocky.
