THE TRIBUTARY SUMIDORO 
The northeastern passage was shallow, with a stony 
bottom. We followed the northern channel along the 
vertical wall. On leaving the island we came to a stretch 
2,500 metres long of beautiful water flowing due north, 
with ideally fascinating banks embellished by dense vege¬ 
tation— neat, clean, and healthy—of the richest green. 
After crossing a bay 100 metres wide, with volcanic 
rocks showing through on both banks and in the river bed, 
the stream was squeezed through a rocky neck 25 metres 
wide, and spread again immediately afterwards to its 
normal width of 50 metres. We were beginning to find 
big rocks more frequently, many in the river channel, a 
bad sign for us, for I feared we might soon encounter 
rapids. 
Wonderful oleo pardo trees ( Myrocarpus frondosus 
Fr. All.), with their octopus-like branches hanging down 
to the water, were fairly common in that region. There 
were two kinds of oleo trees in Brazil, the brown or oleo 
pardo and the red or oleo vermelho, the latter technically 
known as Myrospermum erytroxylon Fr. All. 
We subsequently entered a basin 150 metres wide, 
which contained a circular island 100 metres in diameter, 
Horus Island. 
Eight hundred metres farther we came to another 
large circular bay with a large globular mass of lava on 
its left side. The current was very swift over a nasty 
rocky bottom. The canoe was suddenly flung by the 
current between an accumulation of rocks and an island, 
and, as we found it impossible to turn, floated down at 
an uncomfortable speed through a narrow channel, dodg¬ 
ing as best we could the many ugly rocks just below the 
surface of the water. At the end of this channel we 
encountered violent eddies forming wide circles of most 
treacherous water, although on the surface it looked placid 
enough. 
The tributary Sumidoro, thirty metres wide at its 
45 
