by slow and easy marches, and then put his troops into winter quarters."

Antiquaries are divided in opinion as to the country into which Agricola marched after the battle, some thinking the Horestii were in Angus and Mearns; others, that they inhabited Fife. Richard places them in Fife, and to the westward, as far as Comrie in Perthshire; for he says, their towns were Alauna, Lindum, and Victoria, Keir, Ardoch, and Dealgin Ross. The Horestii mentioned by Tacitus were unknown to Ptolemy, who gives these towns to the Damnii. Richard's accounts bear every mark of truth; and, lib. i. c. 7. he acknowledges, that " he altered his authorities, differing sometimes from Ptolemy and others; but, as he hopes the alterations he made were for the better, he trusts he does not, on that account, merit reprehension."* As we are ignorant of the reasons which induced Richard to place the Horestii in Fife, we need not blame him; but, from whatever source he derived his information, there is reason to conclude that he was led into a mistake. Agricola carried his ravages to the estuary of the Tay, three years before he marched against Galgacus, so the inhabitants of Fife could not now, with any propriety, be called new nations; nor would a commander, so skilful, have advanced to Mons Grampius, leaving Fife behind him unsubdued: to reduce that peninsula would not have occupied the Roman general during the remainder of the summer; nor could it, at any time, on account of its circumscribed limits, have made any effectual resistance to such a force as

Lib. i. c. 7. Ex Ptolemæo et aliunde nonnullis ordinem quoque, sed quod spero in melius, mutatum hinc inde deprehendes.

the army of Agricola. There are few traces of the Romans found in Fife; but the permanent camp at Loch Ore, and two or three smaller posts, with Alauna, Lindum, Hierna, and Victoria, in which it seems probable that Agricola wintered with his army, would sufficiently overawe the inhabitants of Fife, and, consequently, to take hostages from them would be unnecessary. Pinkerton, in his Inquiry, Chalmers, in his Caledonia, and almost all who have treated of the campaigns of Agricola, say, that, after the battle of Mons Grampius, he retreated, or led back his forces into the territories of the Horestii. General Roy says the same thing; and it seems to have been this circumstance which led him to suppose, that the traces of the battle might be looked for in the Mearns. One would be tempted to think, that each succeeding antiquary took the matter upon trust from his predecessors, without consulting the text; for the expression of Tacitus does not imply any retrograde movement, — in fines Horestorum exercitum deducit,* «he marched his army into the territories of the Horestii." It must be acknowledged, that deducere does sometimes signify to convey, or to withdraw; and Agricola may be said to have conveyed his army, or to have withdrawn his troops into Angus, instead of following the Caledonians among the Grampians; but deducere commonly means to lead forth, and in this sense we find the word elsewhere used by Tacitus, deducere legiones in aciem, "to lead forth the legions to battle." It seems unreasonable to suppose that Agricola, after gaining a victory so decisive, com

* On looking into Murphy's elegant Translation of Tacitus, we are happy to find that here he uses this very phrase.

menced his retreat, without advancing a step, to reap the fruits of his toil. To us it appears, that, after defeating Galgacus at Blairgowrie, the Roman commander advanced into Strathmore, probably then, as it is now, the richest portion of Scotland; sent part of his troops on board his fleet, at Montrose Basin, Lunan Bay, or probably at Invergowrie; and, having taken hostages from the inhabitants of Angus, returned by easy marches, to winter at the places formerly mentioned, when he founded Victoria. Some of the camps discovered in Strathmore, appear to have been occupied by the same army which encamped at Ardoch and Grassy Walls, being of the same size, and constructed in the same manner: we agree, therefore, with those antiquaries who think the Horestii inhabited Angus and Mearns; and we are pleased to find this to be General Roy's opinion, to which we pay much respect.

Before attempting to trace the march of Agricola farther, we must notice the camps to which we formerly alluded, as being in the vicinity of Meiklour. Three miles to the eastward, Cupar Angus is situated near the middle of a Roman camp. Maitland, Hist. of Scot. p. 199, describes this camp as a square of 1200 feet, fortified with two strong ramparts, and large ditches, which are still to be seen on the eastern and southern sides. Upon the Polybian system, it would hold about 11,000, and upon the Heginian, about 29,000 men. From the double rampart, this does not seem to have been one of the temporary encampments of Agricola. We incline to think it a camp of Lollius Urbicus, and afterwards one of the castra stativa, in which a town was founded by the natives, on the departure of the Romans.

At Camp Muir, near Balgershoe, and a mile and a half to the southward of Cupar Angus, there is a small Roman camp, which is crossed by a road, leading from Lintrose to the post road at Burrelton. This camp is oblong, the mean length being 1900, and mean breadth 1220 feet, and would hold 10,000 men, or a legion with its auxiliaries, upon the Polybian system. About three quarters of the intrenchment can still be distinctly traced; and one thing is singular about this camp, that there is only one gate extant, though as much of the ramparts remain as would have been sufficient to have shewn two or three more, if such had originally existed. This gives reason to believe, that the Romans did not always confine themselves to the usual mode in fortifying camps. Perhaps the large opening on the south side, where the ground is of a morassy nature, has been left so from the beginning, and might serve instead of the two principal gates, which are commonly found, one on each side. Of the gate that remains, the traverse is straight, and there seems reason to conclude, that this camp was occupied by Agricola, with one of the divisions of his army, on returning from the country of the Horestii toward his winter quarters. The size of this camp, as well as that at Haerfaulds, to be afterwards noticed, being somewhat less than the small camp at Ardoch, General Roy deduces, from a calculation founded on their measurements, that the number of men sent by Agricola, on board his fleet, amounted to between three and four thousand.

The nameless station of Richard's 10th Iter, General Roy conjectures to have been the post at Barry Hill, a mile and a half north-east of Alyth, the distance from In medio being about 9 Roman miles. In his memoirs,

he candidly acknowledges, that he had not seen all the antiquities he mentions; but, in several instances, had been obliged to trust to information received from others; and, we may safely conclude, he had never visited the two posts on Barry Hill, or he would have seen at once that they were not Roman. Tradition says they are Pictish; and also, that, in the largest fort, Vanora, or Guinevar, Queen of the British Arthur, was for some time confined. Barry Hill, being one of those that belong to the secondary, or lower range of the Grampians, is of considerable elevation, being about 685 feet above the level of the sea, and commands an extensive view of the Sidla Hills, and of Strathmore. Upon its summit there has been a fort of an elliptical figure, 180 feet long by 74 broad; around this space a mound of earth has been raised, which is still 12 feet in thickness at the top, and 8 feet high: upon this the foundations of the walls, composed of boulders of rough granite, may still be seen, and of the same breadth as the top of the mound; the walls themselves have been of free stone without cement. Gordon's estimate of the extent of this post is extremely erroneous, and it may have been his account in the Iter Sept., which misled General Roy. Among the ruins of this curious antique fortress are to be found pieces of vitrified breccia, or plumpudding stone; but the bridge across the moat seems to be the only part of the work which has been intentionally subjected to the process of partial fusion. The south and east sides, where the hill gently slopes, are defended by a fosse 10 feet broad, and in depth from 12 to 16 feet below the foundations of the wall. A bridge has been thrown over the fosse 18 feet long, and at the middle 2 feet