300 -THE UPPEE AMAZONS. 



the liills of Saiitarem, are erratic bowlders, masses of 

 liornblende. Upon the Rio Negro, just below Rio Brauco, 

 we discovered ledges of compact hornblende, to which we 

 have already alluded.* Will not these throw some light 

 upon the origin of the bowlders of Erere ? In collating 

 all the evidence of ancient glaciers, found in the moun- 

 tains about the valley, we must not fail to mention a fact 

 obtained by Agassiz from the careful and reliable observer 

 Dr. Felice. He states that in the southeastern part of the 

 valley, running from Serra Grande to the Rio Arucaty- 

 Assii, is a wall-like accumulation, composed of the usual 

 debris of moraines, which has an extent of one hundred 

 and eighty miles. Passing from the valley of the Ama- 

 zons, but still remaining within the lines of the tropics, 

 we find other unmistakable traces of glaciers. In the vi- 

 cinity of Rio de Janeiro, Agassiz found the surface of 

 the rocks upon which the drift rests characterized by 

 tliose peculiar iindulations recognized as one of the dis- 

 tinctive marks of moving ice-masses, and which are termed 

 roches moutonnees. Erratic bowlders fill the drift, differ- 

 ing so essentially from the rocks in place as to leave no 

 doubt as to the fact of their transportation from a dis- 

 tance. Among the equatorial Andes we find evidence of 

 the snow-line having formerly been lower than at present. 

 We have already alluded to the vast accumulation of sand 

 and gravel, termed the arenal, resting upon the flanks of 

 Chimborazo, several hundred feet belovv' the line of per- 

 petual snow nj)on that mountain. Its whole character 

 indicates a glacial formation. Dr. Loomis found traces 

 of glaciers, such as ground and striated rocks, and mo- 

 raines, upon the Peruvian Andes. We certainly must not 

 ignore such facts as these, in our search fpr an interpreta- 

 tion of intertropical geology. These facts appear to indi- 

 cate an intertropical winter — at least, such a reduction of 



*See page 1S9. 



