To lay down, with any certainty, the line which the Greeks followed from the Euphrates to Trebizond, appears altogether impossible. I cannot admit the hypothesis of Mr. Ainsworth, who conducts the army across the Araxes to its northern bank, carries them up northward to the latitude of Teflis in Georgia, then brings them back again across the Harpa Chai (a northern affluent of the Araxes, which he identifies with the Harpasus mentioned by Xenophon) and the Araxes itself, to Gymnias, which he places near the site of Erzerum. Professor Koch (p. 104-108), who dissents with good reason from Mr. Ainsworth, proposes (though with hesitation and uncertainty) a line of his own which appears to me open greatly to the same objection as that of Mr. Ainsworth. It carries the Greeks too much to the northward of Erzerum, more out of their line of march from the place where they crossed the Eastern Euphrates, than can be justified by any probability. The Greeks knew well that, in order to get home they must take a westerly direction (see Anab. iii. 5, 15). Their great and constant purpose would be to make way to the westward, as soon as they had crossed the Euphrates; and the road from that river, passing near the site of Erzerum to Trebizond, would thus coincide, in the main, with their spontaneous tendency. They had no motive to go northward of Erzerum, nor ought we to suppose it without some proof. I trace out, therefore, a line of march much less circuitous; not meaning it to be understood as the real road which the army can be proved to have taken, but simply because it seems a possible line, and because it serves as a sort of approximation to complete the reader’s idea of the entire ground travelled over by the Ten Thousand.