238 THE RIVER OF DOUBT [chap, viii 



lower levels. Only the higher stretches were dry. On 

 the sheer banks where we landed we had to push the 

 canoes for yards or rods through the branches of the 

 submerged trees, hacking and hewing. There were 

 occasional bays and ox-bows from which the current 

 had shifted. In these the coarse marsh grass grew talL 



This evening we made camp on a flat of dry ground, 

 densely wooded, of course, directly on the edge of the 

 river and five feet above it. It was fine to see the speed 

 and sinewy ease with which the choppers cleared an open 

 space for the tents. Next morning, when we bathed 

 before svmrise, we dived into deep water right from the 

 shore, and from the moored canoes. This second day 

 we made sixteen and a half kilometres along the course 

 of the river, and nine kilometres in a straight line almost 

 due north. 



The following day, March 1, there was much rain — 

 sometimes showers, sometimes vertical sheets of water. 

 Our course was somewhat west of north, and we made 

 twenty and a half kilometres. We passed signs of Indian 

 habitation. There were abandoned palm-leaf shelters on 

 both banks. On the left bank we came to two or three 

 old Indian fields, grown up with coarse fern and studded 

 with the burned skeletons of trees. At the mouth of a 

 brook which entered from the right some sticks stood in 

 the water, marking the site of an old fish-trap. At one 

 point we found the tough vine hand-rail of an Indian 

 bridge running right across the river, a couple of feet 

 above it. Evidently the bridge had been buUt at low 

 water. Three stout poles had been driven into the 

 stream-bed in a line at right angles to the current. The 

 bridge had consisted of poles fastened to these supports, 

 leading between them and from the support at each 

 end to the banks. The rope of tough vines had been 



