188 COSMOS. 



geographical information of Ptolemy ; but his work surpasses 

 all other geographical labors of antiquity by the diversity of 

 the subjects and the grandeur of the composition. Strabo, as 

 he takes pleasure in informing us, had seen with his own eyes 

 a considerable portion of the Roman empire, " from Armenia 

 to the Tyrrhenian coasts, and from the Euxine to the borders 

 of Ethiopia." After he had completed the historical work of 

 Polybius by the addition of forty-three books, he had the cour- 

 age, in his eighty- third year,* to begin his work on geography. 

 He remarks, " that in his time the empire of the Romans 

 and Parthians had extended the sphere of the known world 

 more even than Alexander's campaigns, from which Eratos- 

 thenes derived so much aid." The Indian trade was no lon- 

 ger in the hands of the Arabs alone ; and Strabo, when in 

 Egypt, remarked with astonishment the increased number of 

 vessels passing directly from Myos Hormos to India. t In 

 imagination he penetrated beyond India as far as the eastern 

 shores of Asia. At this point, in the parallel of the Pillars of 

 Hercules and the island of Rhodes, where, according to his 

 idea, a connected mountain chain, a prolongation of the Tau- 

 rus, traversed the Old Continent in its greatest width, he con- 

 jectured the existence of another continent between the west 

 of Europe and Asia. " It is very possible," he writes,^ " that 



* On the reasons on which we base our assertion of the exceedingly 

 late commencement of Strabo's work, see t Groskurd's German transla- 

 tion, th. i., 1831, s. xvii. 



t Strabo, lib. i., p. 14 ; lib. ii., p. 118; lib. xvi., p. 781 ; lib. xvii., p. 

 789 and 815. 



J Compare the two passages of Strabo, lib. i., p. 65, and lib. ii., p. 118 

 (Humboldt, Examen Critique de I' Hist, de la Geographic, t. i., p. 152- 

 154). In the important new edition of Strabo, published by Gustav 

 Kramer, 1844, th. i., p. 100, " the parallel of Athens is read for the par- 

 allel of Thina?, as if Thinae had first been named in the Pseudo-Arrian, 

 in the Periplus Maris JKubri." Dodwell places the Periplus under 

 Marcus Aurelius and Lucius Verus, while, according to Letronne, it 

 was written under Septimius Severus and Caracalla. Although five 

 passages in Strabo, according to all our manuscripts, have Thince, yet 

 lib. ii., p. 79, 86, 87, and, above all, 82, in which Eratosthenes himself 

 is named, prove decidedly that the reading should be the " parallel of 

 Athens and Rhodes." These two places were confounded, as old geog- 

 raphers made the peninsula of Attica extend too far toward the south. 

 It would also appear surprising, supposing the usual reading Qiv&v 

 KVK%.O to be the more correct, that a particular parallel, the Diaphragm 

 of Dicwarchus, should be called after a place so little known as that of 

 the Sines (Tsin). However, Cosmas Indicopleustes also connects hia 

 Tzinitza (Thina;) with the chain of mountains which divides Persia and 

 the Romanic districts no loss than the whole habitable world into two 

 parts, subjoining the remarkable observation that this division is accord- 



