THE RIVER TROMBETAS 103 
I rose several times in the nights for star-obser- 
vations, but so cloudy was the sky that I got only 
a single meridian altitude of a Eridani, which gave 
for latitude 0° 47^ S. 
The 29th of December was cloudy and showery, 
and it seemed probable that the Christmas summer 
had been put off until another year. I found that the 
reason why our men had erected no rancho was 
that they hoped thus to prevent my making a pro- 
longed stay. They began now openly to express 
their discontent — the sound of the waterfall, they 
said, was muito triste," and, with the excessive 
cold, prevented their sleeping — and I saw plainly 
that if I did not move at once they would take 
Erench leave of me. On the 30th, therefore, at 
7 A.M., we started on our return voyage. The river 
had risen much, and we shot rapidly down. To- 
wards night we drew up at the second turtle-bank, 
of which we found only a very small space left 
uncovered by the water. 
[The return journey from this unfortunate and 
unproductive expedition occupied eight days, the 
Journal of which is chiefly occupied with hearsay 
geographical details as to the river Trombetas, 
now superseded by later information. A day was 
passed at the farm of Senhor Bentes, near the 
mouth of the Aripecurii, to dry their soaked 
clothes, mats, and sails before continuing the 
journey, allowing Spruce . to observe and collect 
a few more plants ; and the notes on these, as 
well as an interesting account of the different 
trees called " cedars " in various parts of the 
