ACROSS UNKNOWN SOUTH AMERICA 
with three islands and two gravel beaches in its centre. 
The two principal islands, Paolo and Francesca, were 
each 100 metres long and 50 metres wide. 
We now made the acquaintance of the capivara 
(Hydrochoerus capibara ), a rodent which we found com¬ 
mon farther down in those waters. It was a stupid 
animal. When fired at several times by my men it re¬ 
mained perfectly still, gazing at its enemies. It was 
only when a bullet hit the ground too near that it would 
move away, surprised rather than concerned. 
After going down a corrideira (small rapid) we en¬ 
countered thousands of white and lemon-yellow butter¬ 
flies. On islets of red earth swarms of them were basking 
in the sun, which was getting hotter and hotter as we 
got farther north. 
Again we were soon after faced by a high natural wall 
of brilliant yellow and red colouring. In its western 
part it showed a white stratum three feet thick upon a 
layer of yellow lava of an equal thickness. A stratum of 
lighter yellow was nearest the surface of the water, while 
above was a thick layer of grey earth. On the right 
side, at this point, a tributary streamlet flowed into the 
Arinos. The basin formed by the crescent-shaped wall 
was perfectly circular. When the river emerged from 
it, it folded back from 40° bearings magnetic to 290° 
bearings magnetic. 
Owing to the steepness of the banks we experienced 
difficulty in finding a suitable camping place for the night. 
Eventually at sunset we had to clear with our big knives 
a patch in the dirty forest on the edge of the stream. I 
never liked to camp out of sight of the canoe in case 
anything happened during the night — an attack, a flood, 
a forest fire, or anybody trying to steal or get away with 
the canoe; the danger from my own men being quite as 
great as from any enemy I could have found. I well knew 
that if we lost that canoe we were done for entirely. 
42 
