36 Physical Geography of the Amazons Valley. (January, 
which, by the sinking of the land, was changed to an estuary 
while the sea was being filled with sediment. These questions 
must be settled by a more careful study of the Upper Amazons 
and its tributaries and especially of the Madeira and Negro. But 
whatever the changes may have been it is certain that the Ama- 
zonian river system is much older than the period of subsidence 
of which I have spoken. The slow pulsations of the earth have 
sent many throbs to this equatorial region ; centuries of subsi- 
dence have been followed by centuries of upheaval, and these 
again by depression; river ‘has become sea and sea has passed 
into river and estuary again and again since first the rains and 
springs united to form an infant Amazons. We have ‘some 
glimpses of this older history, but as yet geological exploration 
on the Amazons is too new to give us any very clear sequence of 
events. 
It appears certain that the immense low plain of the Upper 
Amazons was occupied by a Tertiary sea, older and much larger 
than the one which has been described above. Tertiary marine 
shells have been found at several points on the Marañon and Sol- 
imoens, and the Tabatinga clays which contain these shells ex- 
tend far up the Japurá and Purts. The upheaval which placed 
these clays beyond reach of the river waters may have taken 
place long previous to the estuary depression, and many changes 
may have intervened. It appears probable that this Tertiary sea 
opened into the Atlantic through what is now;the valley of the 
Orinoco, and that the Cassiquiare, which at present unites the 
two great river-systems, may correspond to one side of the strait 
or channel. It has been supposed that the sea had two outlets, 
one by the Orinoco valley and the other by that of the Lower 
Amazons, and that the Guiana highland formed a great island be- 
tween them. Of this I think there is no sufficient proof. On 
the Lower Amazons the bluffs between the Tapajés and Tocan- 
tins and the table-topped hills of the northern side are formed of 
clays and ferruginous sandstones; but it is yet to be shown that 
these are continuous with the Tertiary formations of Tabatinga. 
Similar clays and sandstones occur all through Central and East- 
ern Brazil and in the Argentine Republic, but they belong to 
many different ages; no one who has studied geology in Brazil 
will be likely, on mere lithological grounds, to unite formations a 
thousand miles apart. The clays of the Lower Amazons, then, 
