HUNTING FOR THE PYROTHERIUM BEDS. l8l 



haps fifteen or twenty miles, crossing and recrossing the stream as it 

 zigzagged back and forth from one side to another of the narrow val- 

 ley. For a time I experienced little difficulty in following the bed of 

 the canon. The stream had considerable current, and flowed over an 

 almost continuous bed of coarse gravel consisting, for the most part, of 

 crystalline rocks evidently derived from the bed of shingle which covered 

 the adjacent plains. After a time, however, the current of the stream was 

 arrested. It gradually became more sluggish and, owing to the soft and 

 miry nature of the bottom, it was crossed with constantly increasing 

 difficulty and danger, and finally resulted in the formation of an impen- 

 etrable swamp, through which the waters moved slowly on in deep, quiet 

 stretches, with a current scarcely sufficient to indicate the direction of flow. 

 This swamp supported a luxuriant growth of rushes and extended on 

 either side to the very base of the cafion walls, compelling me to leave 

 the valley and take to the bluffs above, where, from the excessively rugged 

 nature of the country, the travelling was extremely difficult. For more 

 than a week I spent the time exploring that intricate labyrinth of chasms, 

 with which the surface of this country has been so deeply dissected. The 

 bottoms of the deepest of these chasms could scarcely have been less 

 than three thousand feet below the level of the higher of the surrounding 

 tablelands. Most of them were not only destitute of running streams, 

 but it was quite evident that these conditions had continued for an indefi- 

 nite period. In some of the larger canons still occupied by flowing 

 streams it was plainly evident that the farther from their sources one pro- 

 ceeded, the less important the streams became as erosive agents, while in 

 the very midst of this deeply dissected region and where the depth of the 

 cailons was greatest, the destructive work ceased entirely and they were 

 absolutely engaged in constructive work, rapidly silting up their own 

 valleys. From their very nature it was evident that these canons had 

 not been carved out within the recent geological cycle and by the small 

 and insignificant streams, when any at all exist, that now occupy their 

 beds, but that, for the most part, they were the remains of a previous 

 drainage system that existed prior to the time which witnessed the great 

 upheaval in the region now occupied by the Andes. This upheaval had 

 resulted in an elevation sufficient to cause the accumulation of great 

 bodies of snow and ice over that region, giving rise to the formation of 

 glaciers that pushed eastward over the adjacent plains and temporarily 



