638 Miniature Physical Geography . [November, 
snags get stranded, and these protect the rest from the force 
of the current. In the next stage the sand-banks are higher, 
having probably been built up during a flood. Then a few 
cotton-wood shrubs, the growth of which is very rapid, are 
seen here and there. These in two or three years cover the 
island, and attain a height of 14 or 15 feet. Beneath them 
springs up a considerable undergrowth of cane. The island 
is nowcomplete,and ready to be erelong washed away by some 
change in the direction of the current.” I find, too, another 
note written on the Para branch of the Amazons : — “ At 
Para the river is, I am informed, 9 miles wide, and is studded 
with many islands. Not unfrequently some of these islands 
are washed away, and fresh ones formed. For instance, a 
few years ago, Parroquet Island stood some few miles below 
the town : this has disappeared, and on the other side of the 
stream a new island, — the Ihla Nova, —has arisen and is 
covered with a luxuriant vegetation.” 
VI. Changes of Depth in River Channels . — Sand-dragging . 
As illustrating the varying depth of the channel in large 
rivers, the following observation, made on the sands near 
Ramsgate, is perhaps not uninteresting ; — The stream had 
to pass through some narrows between a miniature chalk 
cliff and some other obstruction. The force of the current 
here dug a hole in mid-stream. This hole gradually length- 
ened down stream. Then a little ridge began to separate it 
into two parts, an upper and a lower. The lower depression 
advanced, and was followed by the ridge. By this means 
the upper depression was lengthened, until it was also sepa- 
rated into an upper and lower part by a ridge. In this way 
Several depressions in the bed of the stream were formed. 
These advanced down stream, and the lowest of them was 
some considerable distance from the narrows, but below this 
point the depressions were so far filled in as to be indis- 
tinguishable. This is a good instance of the way in which 
fine and coarse sand is dragged along the bottom. Of 
course in the case of a large and deep river this action 
would not be so noticeable; but I imagine that the estimate 
of Messrs. Humphreys and Abbot, that the Mississippi 
pushes along the bottom sand and gravel equal to about 
one-tenth of the mud held in suspension by the river, is by 
no means excessive. 
