Emerging from the hills at Street House, it descends to the valley of the Kale, which it crosses just below Towford School. A little further on it passes on the left a large rectilinear enclosure, probably of Roman origin. Gradually ascending, it crosses Pennymuir, and, sweeping round the top of the ridges to the east of Cunzierton, runs from Shibden Hill in a straight line over hill and dale to the Oxnam. Here on the high ground to the south of the river, overlooking the ford, lay the fort of Cappuck, obviously intended to guard the crossing.

To the north side of the Oxnam the ground rises again, and the road, wet and marshy, hardly used save now and then by some chance wayfarer, continues its straight course over another billowy ridge of land to Jedfoot, where it is lost in the grounds of Monteviot. It reappears a little to the north-west of Ancrum House, running in exactly the same direction as it had done from Shibden Hill to Jedfoot, and eventually disappearing in the road from Jedburgh to St. Boswells, which follows the same straight course for a mile or two further. Beyond this the Roman road probably made in a nearly straight line for Newstead, skirting on the left the village of Eildon. From every vantage point of rising ground that the road crosses in its course from Street House northwards, there stands out as a landmark in the distance that most familiar feature of the border country, the triple-peaked Eildon. To many a Roman soldier marching northwards it must have served as a beacon. So far as Scotland is concerned, the excavation done on the line of this road is as yet confined to that carried out by the late Marquess of Lothian[1] on the fort of Cappuck in 1886. Partial though it was, it proved conclusively that the site was Roman. And the presence of a Roman fort high in this upland country proved no less conclusively the Roman character of the road which runs beside it. The relics found in the course of the excavation are too few to enable us definitely to estimate the limits of its period of occupation. But that the road constituted the line of advance to Newstead in the Antonine period seems clear. Further, the records of Agricola's campaign sixty years earlier were doubtless familiar to Lollius Urbicus and his men, and the presumption is that, in making their advance, they followed in his footsteps. The question can only be definitely solved by more excavation of the forts in the hill-country, Cappuck, Pennymuir, Chew Green. In the meantime, however, such evidence as we do possess

1 Proceedings of the Berwickshire Naturalists' Club, 1893, pp. 382–9.

8