zontally about six or eight inches under the surface, and send up suckers at every three or four inches of their course. Thus each species by the multiplication of roots and stems becomes an effective agent in the retention and increase of soil, and frequently both combine in their work of encroachment on the water and the formation of land; as is well seen along the north-east shore of Cunningham Lake. In the swampy clay and sand of the Florida coast, the mangroves form dense jungles from five to twenty miles broad, and running up the creeks and inlets."

Captain Nelson also observes that other and much smaller plants contribute largely to the formation of new land in the Bahamas. "The marshy lands that are gradually taking the place of the creeks and brackish lakes, abound with and may be said at some points to consist largely of a highly calciferous moss-like Conferva, which in concert with mangrove roots, grasses, and other plants consolidates and completes the chalky soil."

In Brazil also the mangrove is an efficient agent in aiding the silting up of the large lagoons which border the coast. When by deposition of sand the bottom is brought up to the level of low tide the seeds of the mangrove take root, and the shoal is soon covered with the trees. Among their roots fine silt is deposited, and the sandbank is overlaid with a layer of soft sand, which may increase in thickness till its surface is only covered at high water. Reeds, rushes, arums, and coarse grasses then assist in converting the swamp into land.1

1 Hartt's "Geology of Brazil,” p. 222.

CHAPTER XII.

FLUVIATILE DEPOSITS.

Deposits formed by Rivers and Glaciers.

Deposition of Detritus. It must be borne in mind. that the three processes, erosion, transportation, and deposition are closely connected; each may proceed in close neighbourhood to the other, and the hollow which has been formed by erosion one day, may be filled up with sediment on the next. It is seldom that the exact conditions of equilibrium exist, when the velocity of the stream is sufficient only for the transport of sediment without exercising

B

Fig. 53.

erosion or permitting deposition. For suppose that at a given time or place, the velocity is just enough to keep the load of detritus in suspension, it is clear that a slight rainshower causing an increase in the volume of water, would at once convert the load into an agent of erosion, while conversely a decrease in the volume would allow some of the load to subside. Speaking generally, therefore, we may say that when the load is not employed in erosion, some of it is being deposited.

In the upper part of a river-course where the slopes are

steep, and the current swift, only the larger stones are allowed to rest; lower down the valley, where the stream loses some of its velocity, smaller stones and coarse sand are deposited, though much of the latter is still carried onward. Eventually this also subsides, and when a great river reaches the broad plains which are little above the level of the sea, it carries nothing but the finest silt and mud, part of which is deposited, and part is poured into the sea.

A winding river is constantly changing the position of its curves, for the stream is continually eating away the bank on one side of a curve and allowing deposition to take place on the opposite shore. The cause of this is found in the fact that the whole of the water in the stream does not move with the same velocity, and that the line of quickest motion does not keep exactly midway between the banks, but follows the course indicated in fig. 53, the arrows showing the impact of the current at the points a A. Since, therefore, the velocity of the stream is least off the points в B, and since decreased velocity means decreased power to transport detritus, it follows that deposition will take place opposite B B, and the angles will gradually be filled up with gravel, sand, or silt, as the case may be. This is shown by the black parts, c c, in fig. 53.

Gravel Beds and Alluvial Levels. When a stream has reached a tract where the lessened slope allows gravel and sand to be deposited, small areas of these deposits are usually found to occur at intervals, first on one side of the river, and then on the other. They are slightly above the flood level of the stream, because, since they were formed, it has deepened its channel and consequently flows on a lower plane than it previously did.

Still lower down, the watercourse itself is generally bordered by strips of level ground, which form a narrow tract or plain, with the stream winding from side to side within it, and over which the waters spread in time of flood. The soil of this flood plain or alluvial level consists partly of earth conveyed by rain down the sides of the valley, and partly of the silt and mud deposited by the overflow of the stream; when the stream is in flood, and overflows this flat and shallow ground, the turbid current

G

Fig. 54.

View of a River-Valley with a Gravel Terrace and more recent Alluvium.

T

is checked, and being unable to carry all its load of sediment, some of it is precipitated to the bottom and remains there on the recession of the waters; this muddy soil is called alluvium.

In descending a valley it will always be found that the width of the alluvium increases in proportion as the volume of the stream increases, and within this constantly widening plain, the river pursues a tortuous and meandering course. At the same time the tracts of gravel increase in size and in height above the river-level, and often form extensive banks and terraces on either side of the alluvium.

Fig. 54 shows a river emerging from hilly ground and flowing towards the spectator: alluvial levels are seen on each side of the valley: on the right (at r) is a tributary stream winding through the deltoid flat it has helped to

Fig. 55. Section through terraces of River Gravel.

form, and on the left (at a) is a bank consisting of gravel deposited by the river when it flowed at a higher level, and before it had cut out its present channel.

In many valleys patches of gravel may be found at various levels above the river, and sometimes at considerable distances from the present stream. These are portions of still older gravel beds deposited at a time when the river ran at that level, and before it had cut its channel down to its present depth. These higher banks of gravel have necessarily suffered much from erosion and detrition, and are generally reduced to the state of disconnected patches; but in some cases they are better preserved and remain as more or less continuous terraces, so that the valley side presents a succession of such gravel terraces, and a section across it has the appearance exhibited in fig. 55.

It must not be supposed from this diagram that the