1851.] 
SERRA DE COCOI. 
233 
peaks, like those generally found in this district. The 
next day we reached Marabitanas, the frontier fort of 
Brazil : there is now only the remnant of a mud en- 
trenchment, and a small detachment of soldiers. As the 
Commandante was not there, we did not stay, except to 
purchase a few plantains. 
On the 1st of Eebruary we reached the Serra of Cocoi, 
which marks the boundary between Brazil and Vene- 
zuela. This is a granite rock, very precipitous and form- 
ing nearly a square frustum of a prism, about a thou- 
sand feet high. It rises at once out of the forest plain, 
and is itself, on the summit and the less precipitous 
portions, covered with thick wood. Here the piiims, or 
little biting flies, swarm, and made us very uncomfort- 
able for the rest of the day. We had now beautiful wea- 
ther, and in the evening slept on a fine granite beach 
very comfortably. The next night we staid at a rock on 
which we found some curious figures engraven below 
high-water mark. Here having a clear horizon up the 
river to the north, I saw my old friend the pole-star, 
though I was only in 1° 20' north latitude. We had 
now every day fine rocky beaches, along which I often 
walked, while young Luiz would shoot fish for us with 
his bow and arrow. He was very skilful, and always 
had his bow by his side, and as we approached a rock 
or shallow would fit his arrow and send it into, some 
glittering acarra or bright -coloured tiicunare. 
At length, on the afternoon of the 4th of February, we 
arrived at Sao Carlos, the principal Venezuelan village 
on the Rio Negro. This was the furthest point reached by 
Humboldt from an opposite direction, and I was there- 
