in the open air; at night they scrape holes for themselves in the snow; they howl like wolves, in accompaniment with a dog that sits in the middle of the circle and sets them off. In Mexico the dogs were subjected to an operation to make them fatter and better eating. On the borders of the province of Durango, and farther to the north on the slave lake, the natives, formerly at least, conveyed their tents of buffalo skins on the backs of large dogs when changing their place of residence with the change of season. All these traits resemble the customs of the inhabitants of eastern Asia. (Humboldt, Essai polit. T. ii. p. 448; Relation hist. T. ii. p. 625.) [16] p. 8.--“_Like the greater part of the Desert of Sahara, the Llanos are in the torrid zone._” Significant denominations,--particularly such as refer to the form in relief of the earth’s surface, and which have arisen at a period when there was only very uncertain information respecting the countries in question and their hypsometric relations,--have led to various and long-continued geographical errors. The ancient denomination of the “Greater and Lesser Atlas” (Ptol. Geogr. lib. iii. cap. 1) has exercised the prejudicial influence here alluded to. No doubt the snow-covered western summits of the Atlas in the territory of Morocco may be regarded as the Great Atlas of Ptolemy; but where is the limit of the Little Atlas? Is the division into two Atlas chains, which the conservative tendencies of geographers have preserved for 1700 years, to be still maintained in the territory of Algiers, and even between Tunis and Tlemse? Are we to seek between the coast and the interior for parallel chains constituting a greater and a lesser Atlas? All travellers familiar with geognostical views, who have visited Algeria since it has been taken possession of by the French, contest the meaning conveyed by the generally received nomenclature. Among the parallel chains, that of Jurjura is generally supposed to be the highest of those which have been measured; but the well-informed Fournel, (long Ingenieur en chef des Mines de l’Algérie), affirms that the mountains of Aurès, near Batnah, which were still found covered with snow at the end of March, are higher. Fournel denies the existence of a Little and a Great Atlas, as I do that of a Little and a Great Altai (Asie Centrale, T. i. p. 247-252). There is only one Atlas, formerly called Dyris by the Mauritanians, and this name is to be applied to the “foldings,” (“rides”) or succession of crests which form the division between the waters flowing to the Mediterranean, and those which flow towards the Sahara lowland. The strike or direction of the Eastern Mauritanian portion of the Atlas is from east to west; that of the elevated Atlas of Morocco from north-east to south-west. The latter rises into summits which, according to Renou, (Exploration Scientifique de l’Algérie de 1840 à 1842, publiée par ordre du Gouvernement, Sciences Hist. et Geogr. T. viii. 1846, p.