298 THE UPPER AMAZONS. 



jects in the scenery of the Lower Amazons, and mark the 

 thiclaiess of the formation. Resting nnconformably upon 

 the sandstone thus denuded, lies a deposit of clay, of a 

 bro\vnish-rcd color, showing slight traces of stratification. 

 Agassiz, finding in these vast deposits no fossils of any 

 kind, with the exception of a few loaves, disposes of tlio 

 question of their origin, by supposing them to be glacial 

 formations, deposits in a vast fresli-water lake, into which 

 the valley Avas converted upon the breaking up of the 

 glacier, which, flowing from the Andes, crowded between 

 the highlands of Guiana and Brazil — the mouth of the 

 valley being closed by an immense moraine barrier. By 

 these peculiar circumstances, under which Agassiz sup- 

 poses these formations to have been deposited, he accounts 

 for the absence of fossils in them ; and to the giving way 

 of the barrier, and the escape of the waters of the lake, 

 he refers the origin of those "hills of denudation" to 

 which we have alluded, located in the lower portion of the 

 valley. 



The finding of marine or brackish-water shells at 

 Pebas militates against this theory, unless it can be shown, 

 as Agassiz suggests, that they are accidental, or belong- 

 to a post-glacial formation, resulting from a submergence 

 of the continent, such as is known to have taken place in 

 North America, after the close of the ice-period. But, 

 although these shells prove the marine or estuary origin 

 of the formation with which they are identified, yet they 

 prove nothing more. We should be cavitious, and not 

 proceed in advance of our facts. These, as we now pos- 

 sess them, simply allow us to limit glacial action hero. 

 They may permit us, perhaps, to infer that portions of the 

 formations, or even that the main deposits of the Ama- 

 zonian Valley, are not of fresh-water or glacial origin. 

 But we are committing an error, if, from these data, with 

 a hasty examination of the equatorial flora, with reference 



