24. It is in this general kind of description of the third section that Eratosthenes supposes 10,000 stadia from the Caspian Gates to the Euphrates. This he again divides according to former admeasurements which he found preserved. Starting from the point where the Euphrates passes near to Thapsacus, he computes from thence to the place where Alexander crossed the Tigris 2400 stadia. The route thence through Gaugamela,[556] the Lycus,[557] Arbela,[558] and Ecbatana,[559] whither Darius fled from Gaugamela to the Caspian Gates, makes up the 10,000 stadia, which is only 300 stadia too much. Such is the measure of the northern side given by Eratosthenes, which he could not have supposed to be parallel to the mountains, nor yet to the line drawn from the Pillars of Hercules through Athens and Rhodes. For Thapsacus is far removed from the mountains, and the route from Thapsacus to the Caspian Gates only falls in with the mountains at that point.[560] Such is the boundary on the northern side. 25. Thus, says Eratosthenes, we have given you a description of the northern side; as for the southern, we cannot take its measure along the sea, on account of the Persian Gulf, which intercepts [its continuity], but from Babylon through Susa and Persepolis to the confines of Persia and Carmania there are 9200 stadia. This he calls the southern side, but he does not say it is parallel to the northern. The difference of length between the northern and southern sides is caused, he tells us, by the Euphrates, which after running south some distance shifts its course almost due east.