RESIDENCE AT TARAPOTO 85 
place to be reached, has been sought to be obtained 
by following a ridge separating two streams. The 
summit attained, another similar one has been 
picked out and reached in a similar manner, often 
no doubt with much trouble, and after considerable 
entanglement in the valleys. Thus the roads here, 
like the first-made roads in all parts of the world, 
go straight over the tops of hills, instead of winding 
around their base. The dense forest makes the 
finding of a way among hills infinitely more difficult 
when no compass is used, and though it would seem 
more feasible to have sought out a passage along 
the watercourses, a very little practice shows the 
impossibility of this. Besides that the vegetation 
is much ranker near water, the course of the streams 
— not merely their bed, but the whole of the narrow 
valley in which they run — is so obstructed by large 
masses of rock and stones as to be all but impass- 
able, and completely so when the valley narrows to 
a gorge with perpendicular sides which merely 
admits the passage of the stream in an alternation 
of cascades and deep still pools. To avoid a 
pongo — as these gorges are called — one must 
climb a mountain-side and then go down again, and 
perhaps steep cliffs render descent impossible for a 
long distance. Hence it may be seen how, by 
seeking out the sharp ridge of a mountain, when not 
too steep, we really avoid invincible obstructions, 
although we have to ascend and descend great 
heights. It is true that a little previous surveying 
and a little good engineering would smooth down 
most of the difficulties that offer themselves, and I 
have no doubt that good winding mule-roads, at a 
slight inclination, might be made in any part of the 
