much more rugged and abrupt, their precipitous flanks bare of débris, and their serrated crests and peaks recalling, according to Walther, the bold and abrupt dolomite mountains of South Tyrol. The horizontally arranged strata of the desert assume very

FIG. 79.

WIND EROSION: TABLE MOUNTAINS, ETC., OF THE SAHARA (Mission de Chadamés).

different forms, and have been carved into tabular, conical, and pyramidal hills, with a general resemblance to the buttes, mesas, and pyramids of the Colorado region. (Fig. 79.) When the strata are

FIG. 80. WIND EROSION: HARDER BEDS AMONGST INCLINED CRETACEOUS STRATA. LIBYAN DESERT. (J. Walther.)

inclined the outcrops of the harder beds project, and we have in like manner a reproduction of the escarpments and dip-slopes which the same geological structure gives rise to in well watered lands. (Fig. 80.)

The projecting ledges and escarpments, however, are always honeycombed and dressed in a different way, betokening everywhere the characteristic action of the wind.

The final result of wind erosion is the reduction of inequalities and the production of a plain-like surface.

FIG. 81. WIND EROSION: STAGES IN THE EROSION AND REDUCTION OF A TABLE-MOUNTAIN. (J. Walther.)

In the Eastern Sahara wide areas of rocky land have been thus levelled. (Fig. 81.) Such areas are usually more or less abundantly besprinkled and paved with angular stones, usually dark brown or black, and so highly polished that they glance and glitter in the sun.

It is obvious that such stones are derivative; they are the relics of massive beds of sandstone, through which they were formerly distributed, and which have since been gradually disintegrated and removed. In some places, indeed, massive inclusions of the kind (manganese concretions), of all shapes and sizes, project from the surface of the sandstone in which they are still partly embedded. On the lee side of such concretions the sandstone has been sheltered from the attack of the wind, while it has been planed away in

FIG. 82.

front.

MANGANESE CONCRETIONS WEATHERED OUT OF SANDSTONE;
ARABAH MOUNTAINS, SINAI PENINSULA. (J. Walther.)

No stone withstands the action of the wind so well as the hard flints, jaspers, and silicious concretions, which are so commonly met with in the sedimentary strata of the Libyan desert. When the latter have become disintegrated and gradually removed by the wind, the hard nodules and concretions remain, and thus the broad plains are covered over with sheets of "gravel" and shingle. The Sserir, according to Walther, are nothing more than rocky lands levelled by wind-erosion; the more yielding materials have

been swept away, while the hard inclusions left behind. are now concentrated at the surface.

Another result of deflation may be referred to. Now and again in wind-swept plains and plateaux the rocks, according to their nature, are variously affected. Some are disintegrated and rotted more readily than others. These, therefore, tend to be more rapidly reduced below the general level, and shallow basins are thus formed which are sometimes occupied by water for shorter or longer intervals. Such is probably the origin of the Caldeirãos of Bahia, where the general configuration of the surface has some resemblance to that of an ice-worn regiongently rolling ground, namely, showing innumerable shallow depressions winding amongst multitudinous bare-backed, dome-shaped rocks.

The disintegrated material removed from a rocky desert is eventually spread out and piled up in sheets and heaps of sand, which travel onwards in the direction of the prevalent wind. In the Eastern Sahara bare rocky plateaux prevail, and sand-wastes are usually of inconsiderable extent. In the Western Sahara, on the other hand, the whole area is more or less smothered in sand. There vast stretches of dunes move with the trade-winds. Advancing to the south-west, they reach the banks of the Niger and the Senegal, and are here and there forcing these rivers southward. Again, passing to the west, they touch the Atlantic coast between Cape Bojador and Cape Blanco, and stream out to sea so as to form

a belt of sand-banks extending several miles from the shore. For long ages, therefore, a great current of sand has been constantly flowing out of the desert.

The dunes of a desert appear to move more readily than those of maritime regions. Possibly this may be due to the better rolled character of the constituent grains, to the drier condition of the sand, to the want of any binding materials, and the absence of a fixed nucleus, such as is so commonly acquired for the formation of coastal dunes. In the central portions of a desert they are generally arranged in series of long parallel undulations, that extend in a direction at right angles to that of the prevalent wind. Elsewhere they may be more irregular in their grouping and arrangement, individual sand hills not infrequently assuming a crescentic or sickle-like shape. They vary much in height, not often exceeding 250 feet, although occasionally reaching 500 or even 600 feet.

It need hardly be said that dunes are not restricted to desert regions. Wherever incoherent deposits are dried and exposed to the air, they are liable to drift with the wind. Hence blowing sands are well developed upon certain sea-coasts and lake shores, and in the broad, flat valleys of many large rivers. If the surface over which sand is blown be level and free from obstructions, the sand does not necessarily accumulate in heaps and banks, but is often spread out in successive horizontal layers, forming a sandplain. But wherever obstructions intervene, such as