252 THROUGH THE BRAZILIAN WILDERNESS 



extraordinary, almost impossible, that so broad a river 

 could in so short a space of time contract its dimensions 

 to the width of the strangled channel through which it 

 now poured its entire volume. 



This had for long been a station where the Nhambi- 

 quaras at intervals built their ephemeral villages and tilled 

 the soil with the rude and destructive cultivation of sav- 

 ages. There were several abandoned old fields, where the 

 dense growth of rank fern hid the tangle of burnt and fallen 

 logs. Nor had the Nhambiquaras been long absent. In 

 one trail we found what gypsies would have called a "pa- 

 teran," a couple of branches arranged crosswise, eight 

 leaves to a branch; it had some special significance, be- 

 longing to that class of signals, each with some peculiar 

 and often complicated meaning, which are commonly used 

 by many wild peoples. The Indians had thrown a simple 

 bridge, consisting of four long poles, without a hand-rail, 

 across one of the narrowest parts of the rock gorge through 

 which the river foamed in its rapid descent. This sub- 

 tribe of Indians was called the Navai'te; we named the 

 rapids after them, Navai'te Rapids. By observation Lyra 

 found them to be (in close approximation to) latitude ii° 

 44' south and longitude 60° 18' west from Greenwich. 



We spent March 3 and 4 and the morning of the 5 th 

 in portaging around the rapids. The first night we camped 

 in the forest beside the spot where we had halted. Next 

 morning we moved the baggage to the foot of the rapids, 

 where we intended to launch the canoes, and pitched our 

 tents on the open sandstone flat. It rained heavily. The 

 little bees were in such swarms as to be a nuisance. Many 

 small stinging bees were with them, which stung badly. 



