PIIYSIOAL ASPECT. 417 true ^nWcmens, a ro^^ion <io«tituto of floworinf^ plants or fihrubii, without birdii or buttorflios, exiH)^c<l only to tho blind fonx»« of bout un<l tbt» windH. Arnmlin^ to a suinmury estinmto, the horizontul und ruvinod plutouux occupy uUmt half of ita whole extent; a ninth is covered with sandu, while the rest i» shurcd by the rocky highlundM, steppes, crtKlcd depressions, ouscs, und strips of cultivated borderlanda. Zittel estinuitcs its nicun elevation alnive the sea at 1,1(K) feet. Notwithstanding the difTerenccs of relief, a character of unity is imported to the Sahara throughout its whole extent, chiefly by the scarcity or complete abaenco of water everywhere except in the hilly districts, where the higher summita pene> trate to tlie upjxT atmospheric currents. For the origin of the Sahara has to bo sought, not in the ground itself, but in the aerial regions above. Its creation must evidently be traccnl to the same causes that have given rise to analogous desert wastes in the Asiatic continent. It is, in fact, merely a western prolongation of the almost treeless tracts which traverse Mongolia, Kashgaria, Turkestan, Irunia, und Arabia, interruptwl at long intervals by watercourses fringed with trees, or by mountain ranges and vcrtlant upland valleys. To the prevailing dry winds is due this long desert zone with its ])arallcl strips of bordering stepjH^lands, obliquely crossing the eastern hcmisphert> for a space of some 7,000 or 8,000 miles. To these Asiatic and African wastes Humboldt has given the collective name of " track of the polar winds," as if the aerial currents which in the tropics become the trade winds regularly followtnl the line tractnl in white sands across the two continents. This view, however, is not ipiite cornvt. The general atmospheric movement from the North. Pole towanls the equatorial regions does not follow such an oblique direction as is here indicatetl. Although deflected towards the soutli-west by the rotation of the glob« it is far frtmi being always regularly directed east and west along the axis of the Sahara, i>arallel with the equator. The mettH)roli>gical obst>rvations made in the Saham itself and on it* borders show that the normal direction is from the Mediterranean southwards. Except in the eastern parts, and in the Tuareg territory, when however, they are very variable, the pn>vailing wiiuls are not those whieh come from the east or north-cast after losing nearly all their moisture on the long journey across the Asiatic continent. Nevertheless the direction of the prevailing currents in the Sahara sufiiciently explains its arid character. In Asia the i)olar winds, whose moisture lias In^en jmnipitattMl on the Altai, Tian-Shan, Turkestan, and Anatolian highlands, bring little or no rain to east China, Irania, and Arabia. For the same reason the dry winds from Kurope are unable, during the short jmssuge across the Mediterranean, to take up a sufficient quantity of vai^ur to supply the African continent. Their slight store is exhausted on the northern uplands, leaving little for the southern regions of the Sahara, where scarcely any rain fulls except iu the month of August, when the sun is at its zenith. The general aspect of the desert bears evidence of great changes, which can be due only to the action of water. Large wadics with their bunks and flats preserve the record of running streams in these now arid regions. I)e*^p gorges excavated iu the rocky plateaux also speak of the erosions caused by torrents ond streami,
