OF THE ROMAN EMPIRE 459 dence at Cairo, a French consul has ventured to assign about four millions of Mahometans, Christians, and Jews, for the ample, though not incredible, scope of the population of Egypt,^^ IV. The conquest of Africa, from the Nile to the Atlantic amco^ ocean, 15S was first attempted by the arms of the caliph Othman.i-' vas^on by The pious design was approved by the companions of Mahomet aj). 647 and the chiefs of the tribes ; and twenty thousand Arabs marched from Medina, with the gifts and the blessing of the commander of the faithful. They were joined in the camp of Memphis by twenty thousand of their countrymen ; and the conduct of the war was entrusted to Abdallah,!'^*^ the son of Said, and the foster-brother of the caliph, who had lately supplanted the con- 1*7 See Maillet (Description del'Egypte, p. 28), who seems to argue with candour and judgment. I am much better satisfied with the observations than with the reading of the French consul. He was ignorant of Greek and Latin Hterature, and his fancy is too much dehghted with the fictions of the Arabs. Their best know- ledge is collected by Abulfeda (Descript. JEgypt. Arab, et Lat. a Joh. David Michaelis, Gottingas, in 4to, 1776), and in two recent voyages into Egypt we are amused by Savary and instructed by Volney. I wish the latter could travel over the globe. 158 ty conquest of Africa is drawn from two French interpreters of Arabic literature, Cardonne (Hist, de I'Afrique et de I'Espagne sous la Domination des Arabes, torn. i. p. 8-55), and Otter (M«5m. de I'Acad^mie des Inscriptions, torn. xxi. p. 111-125, and 136). They derive their principal information from Novairi, who composed, A.i). 1331, an Encyclopaedia in more than twenty volumes. The five general parts successively treat of, i. Physics, 2. Man, 3. Animals, 4. Plants, and, 5. History ; and the African affairs are discussed in the vith chapter of the vth section of this last part (Reiske, Prodidagmata ad Hagji Chalifoe Tabulas, p. 232- 234). Among the older historians who are quoted by Novairi, we may distinguish the original narrative of a soldier who led the van of the Moslems. [The work of Novairi (see Baron de Slane's translation. Journal Asiatique, 1841, and App. to tome i. of his transl. of Ibn Khaldun, p. 313 sqg.) is marked by many romantic and legendary details. It is safer to adiiere to the briefer notices of the older ninth- century writers, especially Biladhuri (see references in Journal Asiat. , 1844) and Ibn Abd al Hakam (see extract in Journal Asiat., ifi., and App. to Slane's Ibn Khaldun, p. 301-12), and use with caution both Novairi and Ibn Khaldun (whose History of the Berbers and Musulman dynasties of North Africa has been trans- lated by the Baron de Slane, 1852-6, 4 vols.). Ibn Khaldun (14th century) used Novairi ; and Novairi used Biladhuri, and Ibn al Athir, among other sources. Ibn Kutaiha has also some important notices (see Gayangos, History of the Mohammedan dynasties in Spain, 1840, vol. i. App. E), and Al Bakri (see Slane, in Journal Asiat., 1858). The French conquest of Algiers and occupation of Tunis have led to some valuable studies on this period : Fournel, Les Berbers : Etudes sur la conquete de I'Afrique par les Arabes, 1881 : Mcrcier, Hist, de I'Afrique septen- trionale, 1888-91 ; Diehl, Bk. v. in L'Afrique Byzantine, 1896. Besides these, we have Weil, Amari (Storia dei Musulmani di Sicilia, first chapters of vol. i. ), Roth's Oqba ibn Nafi, 1859, Tauxier's Le patrice Gregorius (Rev. Africaine in 1885).] isa [Amr however had already rendered Barca tributary and reduced Tripoli and Sabrata in A.D. 642-3 or 643-4 (according to Ibn Abd al Hakam, ap. Slane's Ibn Khaldun, p. 302-3. See Weil, i. p. 124). Omar decided against a further advance westward.] i"0See the history of Abdallah in Abulfeda (Vit. Mohammed, p. 109) and Gagnier (Vie de Mahomet, torn. iii. p. 45-48).
