1885.] Physical Geography of the Amazons Valley. 35 
movement went on, the branch estuaries were blocked up at their 
mouths by the islands which formed in front of them. Where 
the branch received a muddy tributary it also was filled up; but 
the clear-water tributaries, like the Tapajós, Xingu and Tombetas, 
brought down no sediment, and their estuaries, closed at the 
mouths, assumed the form of lakes. 
In this way the whole of the Amazonian flood-plain has been 
built up. Passing now a step farther back it is easy to see that 
the flood-plains of the Solimoens and Madeira were formed in 
the same way. But the vast extent of this alluvial land on the 
Upper Amazons seems to indicate a widening of the great bay at 
its upper end; a kind of inland sea connected with the ocean to- 
wards the east by a comparatively narrow strait. This sea, at 
the period of which I am speaking, had no connection with the 
Orinoco valley, for the Amazonian flood-plain is now separated 
from that river by rocky terre-firme, indicated by the falls of the 
Orinoco and Negro. Several large rivers, flowing down from the 
Andes, emptied into the sea near its western end, and eventually 
transformed it into a river by filling its bed with sediment. These 
Andean torrents still exist as the Huallaga, Ucayale, Napo, Tigre 
and extreme Upper Amazons. 
Such a branched estuary bay as I have described could only 
have been formed by the subsidence of land over a great area, 
and the encroachment of the sea on the valleys of some former 
Amazons and its tributaries. “This subsidence must have taken 
place subsequently to the deposition of the clays and sandstones 
which form much of the éesve-firme along the Lower Amazons. , 
For the very tributaries on which the above arguments have been 
based flow through valleys which they have cut in the clays and 
sandstones themselves. It appears probable, also, that the period 
of subsidence was anterior to the formation of the Tabatinga 
clays on the Upper Amazons. 
- For the sake of clearness I have described the silting-up of the 
valley as occurring during a period of repose subsequent to the 
subsidence. But it is quite possible that the subsidence and fill- 
ing up were, in part, contemporaneous. The Upper Amazonian 
sea may have had its outlet through a Lower Amazonian river 
1It may be, however, that this widened flood-plain is due to the extreme shallow- 
ness of the valley of the upper river; the Amazons, dammed back somewhat by the 
accumulations below, would tend to spread out on either side and build up its own 
