The operations of this year have much perplexed historians. The obvious inference from the passage is that Argyllshire was the region he visited, and the author has entered thus minutely into the consideration of what Agricola had to accomplish, and his evident policy, to show that this was the natural step he would take. It has generally been supposed that he turned back upon his steps, and that Galloway was the country ‘opposite to Ireland’ that he visited; but, as we have seen, its inhabitants could not have been said to be ‘ad id tempus ignoti,’ and the language of the early geographers rather characterises Kintyre and the Hebrides as what impressed them most as overhanging Ireland. Chalmers, in order to avoid the plain inference from the passage, is driven to suppose that the Tavaus of the third campaign was the Solway, and that Agricola had advanced no farther, but this is quite inadmissible. The only alternative, that he crossed the river Clyde from north to south and entered Ayrshire, is equally inconsistent with Tacitus’s brief but precise language. Early writers speak of the Clyde as fordable as far down as Dumbarton, and his natural course would be to return by the same route as he came. Tacitus clearly states that he crossed ‘navi in proxima,’ which shows that it was the estuary, and not the river. The Roman fleet was then probably in the Firth of Forth, and the expression seems to imply that he took the first native vessel he could get. There is on an elevated moor in Cowall, between the Holy Loch and Dunoon, the remains of a small square fort which has all the appearance of a Roman exploratory station. It commands an extensive view, in one range, of the entire Firth to its mouth, the river Clyde for many miles of its straight course, and Loch Long penetrating in another direction into what was known to the Romans as the Caledonian Forest, and, if it is a Roman work, adds strength to the natural reading of the passage, and the expression, ‘copiis instruxit,’ is singularly applicable. Footnote 50: