21. In the instances which follow, Hipparchus, treating of these subjects, either asserts things similar to those which we have already refuted, or takes for granted matters which are not so, or draws improper sequences. For instance, from the computation [of Eratosthenes] that the distance from Babylon to Thapsacus[547] is 4800 stadia, and thence northward to the mountains of Armenia[548] 2100 stadia more, it does not follow that, starting from the meridian of that city, the distance to the northern mountains is above 6000 stadia. Besides, Eratosthenes never says that the distance from Thapsacus to these mountains is 2100 stadia, but that a part thereof has never yet been measured; so that this argument [of Hipparchus], founded on a false hypothesis, amounts to nothing. Nor did Eratosthenes ever assert that Thapsacus lies more than 4500 stadia north of Babylon. 22. Again, Hipparchus, ever anxious to defend the [accuracy of the] ancient charts, instead of fairly stating the words of Eratosthenes concerning his third section of the habitable earth, wilfully makes him the author of an assertion easy of disproof. For Eratosthenes, following the opinion we before mentioned, that a line drawn from the Pillars of Hercules across the Mediterranean, and the length of the Taurus, would run due west and east,[549] divides, by means of this line, the habitable earth into two portions, which he calls the northern and southern divisions; each of these he again essays to subdivide into as many smaller partitions as practicable, which he denominates sections.[550] He makes India the first section of the southern part, and Ariana[551] the second; these two countries possessing a good outline, he has been able not only to give us an accurate statement of their length and breadth, but an almost geometrically exact description of their figure. He tells us that the form of India is rhomboidal, being washed on two of its sides by the southern and eastern oceans [respectively], which do not deeply indent its shores. The two remaining sides are contained by its mountains and the river [Indus], so that it presents a kind of rectilinear figure.[552] As to Ariana, he considered three of its sides well fitted to form a parallelogram; but of the western side he could give no regular definition, as it was inhabited by various nations; nevertheless he attempts an idea of it by a line drawn from the Caspian Gates[553] to the limits of Carmania, which border on the Persian Gulf. This side he calls western, and that next the Indus eastern, but he does not tell us they are parallel to each other; neither does he say this of the other sides, one bounded by the mountains, and the other by the sea; he simply calls them north and south.