Page:Africa by Élisée Reclus, Volume 2.djvu/509

 THE EEOS— lOUIDI. 419 ubout by the storm, filling the atmosphere with dciiHe clouds that darken the sun. In the presence of these sundNtonns, almost as dang(>rouH as those of snow, the trav(ller rwulls the lej^'nds of caravans and whole armies swallowed up by the udvuncing billows of siind. But after the storm has |)assed the general aspect of the landscajx? is found to have undergone little change. The hillocks are still in their places, and seem to have been scarcely modified in their main outlines. But a single day counts for little in the history of the earth, and even during the contemporary period examples arc not wanting to show that if most of the sand- hills remain or reform in the same place, others occasionally get shifted. The guides often point to hillocks occupying the site of some former depression along the caravan route. That such dinplacements must take place is evident, unless we suppose a perfect equilibrium of the atmospheric currents. But such an equi- librium does not exist, because the winds blowing from the Mediterranean are known to predominate in the Sahara. In many districts the exposed spaces reveal a rocky ground, evidently of different geological origin from that of the surface sands. Thus the chalk plateau south of the Mzab territory is covered here and there with dunes brought from the great western reservoir of sands. East of El-Golea M. Rolland recognised two such shifting ridges about 30 miles long, with a mean breadth of over 2 miles. But such formations cannot be developed in all places, the direction of the sands being necessarily influenced by the relief of the plateaux, the valleys and depres- sions, the aerial currents. Shifting from dune to dune under the action of the wind, the fine particles of dust are at times swept into heaps, like the drift snow in sheltered spots. But elsewhere simdhills are met which have been permanently fixed or bound together by the roots of trailing plant.>», and near the oas««8 it might be possible to arrest the progress of the dunes by planting drin and other species which flourish in such a soil. The Ekgs — Iguidi. The principal sandy regions are the great Libyan desert, between the Eg}'ptian oases and the Tibesti highlands, the two Ergs, or " veins " of the Berber Sahara, the Edeyens of the Tuareg territory, the Iguidi, west of the AVeean travellers pniceeding to Ghadames. Here the slow progress of the dunes lies in the direction from north-west to south- east, as shown by the present position of the great ridges relatively to the original centres of disintegration. In the western Erg the movement is eastwards to the chalk plateaux ; in the Wed High and in Wargla, northwards to the Oijses ; in Iguidi, nuiinly from tbe north-west to the south-east, under the influence of the oceanic roonsoona. According to M. Duveyrier, the normal direction for the whole of the Siihan