VOYAGE FEOM NEW YORK TO EIO DE JANEIRO. 17 



If this be correct, it is very inconsistent with what we know 

 of the geological structure of other continents, where the 

 stratified rocks are in much larger proportions." 



Upon this followed some account of the different kinds of 

 valley formation and of terraces. " Do the old terraces 

 above the rivers of South America correspond to the river 

 terraces on any of our rivers, — those of the Connecticut, 

 for instance, — showing that their waters had formerly a 

 much greater depth and covered a much wider bottom ? 

 There must of course have been a cause for this great 

 accumulation of water in ancient periods. I account for it 

 in the northern half of the hemisphere by the melting of 

 vast masses of ice in the glacial period, causing immense 

 freshets. There is no trustworthy account of the river 

 terraces in Brazil. Bates, however, describes fiat-topped 

 hills between Santarem and Para in the narrow part of the 

 valley, near Almeyrim, rising 800 feet above the present 

 level of the Amazons. If this part of the valley were 

 flooded in old times, banks might have been formed of which 

 these hills are a remnant. But because such a theory 

 might account for the facts it does not follow that the 

 theory is true. Our work must be to study the facts, to 

 see, among other things, of what these hills are built, 

 whether of rock or of loose materials. No one has told 

 us anything as yet of their geological constitution." * 



To-day we have seen numbers of flying-fish from ti:e 

 deck, and were astonished at the grace and beauty of their 

 motion, which we had supposed to be rather a leap than 

 actual flight. And flight indeed it is not, their pectoral 



* Mr. Agassiz afterward visited these hills himself, and an account of their 

 structure and probable origin wiU be found in the chapter on the physical 

 history of the Amazons. 



2 



