tion, owing to the pastoral and mountainous character of the district. In describing them, I shall for the sake of convenience follow the line of the Roman roads which traverse it, and advert to each fortification as it would successively present itself to a person following these routes. At the same time, I beg it to be understood that I by no means intend to assert that all these fortifications are Roman camps. On the contrary, many of them are of British origin, while others may possibly be creations of a later period, erected for the concealment of cattle during the stormy times of border feud and English invasion. Few of them can, however, in my opinion, be referred to this latter class, as they are almost invariably situated on commanding positions, which although admirably fitted for posts of observation and the military occupation of the country, are by no means adapted for concealment, while for that purpose much more appropriate sites might have been found in the dells and glens among the hills.

The geographical direction of the rivers Annan and Clyde, and the comparatively low elevation of the heights which separate them in at least one of the passes near their source, seems to have pointed out the course of their vallies to the engineers of all ages as a line of communication between the north-west of England and the northern part of Scotland; for we find that it has been adopted, not only by the surveyors of the Caledonian railway in our own day, but by the Romans of old. This iter can be distinctly traced along its whole course; and I may mention, without entering into the details of its course beyond the district to which this paper strictly refers, that it proceeds from Carlisle to Dryffe church on the Annan, where it divides into two branches. The most easterly of these ascends that river till it reaches the camp of Tatiusholm, the Tassiesholm' of Roy (p. 104 and plate 8), which, being the last station in Dumfrieshire, we may consider our

1 It may not be uninteresting to mention that this is the site of the anecdote of the well known antiquary sir John Clerk, which is stated by Lockhart, in his Life of Scott, chap. v, to have been the foundation of the incident in the Antiquary, "Prætorium here, prætorium there, I mind the bigging o't." The camp is undoubtedly Roman, and the amusing blunder arose from the endeavour to trace in its imperfect remains the complete details of the Polybian system of castrametation.

starting point. Leaving this, it continues along the Annan till that river is joined by its tributary the Evan. Crossing the latter, it ascends to the high ground on its right bank, along which, and bending to the left up the course of that stream, it proceeds till it enters the county of Lanark. This it does before reaching the summit, as several farms, though lying on the Dumfrieshire side of the height and belonging to that county in ecclesiastical matters, form quoad civilia part and portion of Lanarkshire; an anomaly which may be accounted for by their having at one time belonged to the powerful family of Douglas. From this point the iter will be found laid down on the plan, and distinguished by a continuous double line. (See plate 1.) On reaching the summit, it passes the almost insignificant elevation which here divides the Evan from a tributary of the Clyde, called Clydes Burn, which it crosses near the farmhouse of Little Clyde.

As some confusion appears to have arisen between these names, I may state that the former is the proper name of the stream, and that the latter is exclusively confined to the particular farm which occupies the upper part of its

course.

General Roy (p. 104), in tracing the course of this iter, states that "it falls in with the sources of the Clyde at a place named Little Clyde, where there has been another square redoubt." Chalmers (vol. i, p. 120), mentions "a Roman post, Little Clyde, on the track of the Roman road”, “ which corresponds so exactly with the Damnian town on Little Clyde." If we are to read these passages as asserting the existence of a Roman camp on the farm of Little Clyde, I am afraid that both the learned antiquaries have been misinformed, as I have not only been totally unable to trace the smallest vestige of an encampment in the vicinity of the small portion of the Roman road which passes through this farm, but I am informed by the tenant, whose family has been in possession of the land for several generations, that he never knew or heard of such a thing. If, on the other hand, we suppose that the words Little Clyde were used by mistake for Clydes Burn, these learned authors may refer to the camp I shall immediately describe, which may be said, though not with strict accuracy, to be situated on that stream.

I may also mention a fact with regard to Clydes Burn which is curious in an etymological point of view, viz., that until this stream joins the Clyde, that river is not called Clyde, but Daer, and only takes the former name below the point of junction, although the burn is so small that it can be crossed dry-footed in most weathers, and the river has already attained a breadth of at least thirty feet, and is of considerable depth.

Leaving Little Clyde, the Roman road descends the right bank of Clydes Burn till it approaches the foot of Bodsberry hill, when it begins gradually to ascend from the stream. This hill forms the last of the range of hills which here abut on the valley of the Clyde. It is comparatively isolated, being cut off from the others by a precipitous ravine. Most persons have supposed that the Roman road proceeded round the south side of this hill betwixt it and the Clyde, but this I am able to state positively is a mistake, and that it either passed through the ravine above-mentioned or crossed the top of the hill; and, I may add, that it may have taken both of the latter courses. The top of the hill forms a flat plateau of considerable extent, which is occupied by the first camp I have to describe. (No. 1,' and plate 2, fig. 1.) The whole of the plateau is occupied by the fortification, which consists of a single rampart. On the north-east side facing the ravine, and on the south and south-east above the Clyde the hill is so precipitous as to be inaccessible in a military point of view. At one part on the east the access, though still difficult, is of an easier nature, and here there is a gate through which the road must have entered if it crossed the hill. At the gate on the south-west the access is worse than at that last described, so much so, indeed, that one can hardly conceive what could be the use of a gate at that place; on the north and north-west, however, the hill slopes very gently, and we find that in consequence this quarter of the camp is defended by a second rampart and ditch. There are here gates through both intrenchments, and from them a Roman road can clearly and unequivocally be traced descending the hill. In the interior of the camp a draw well has been sunk,

1 The numbers referred to throughout this paper will be found marked on plate 1, which presents a plan or map denoting the several localities of the camps.