but also to the requirements of civil life, and of agriculture and commerce. The former were mostly carried along the backbone of the country, where there was one, and well deserve the name of ridgeway, where, as in hostile borders like Cumberland and Westmoreland, they follow the almost inaccessible crests of the mountain chain; the latter, styled via diverticulæ or branch ways, with their subdivisions of vicinales, agraria, deviæ, etc., although laid down on the same principle, were not so rigidly bound by it, deviations being often rendered necessary for the sake of convenient access to particular places; rough and ready modes of construction were also allowed here. One characteristic belongs to all these classes of roads, as compared with our modern ones-they are carried straightforward, uphill and down dale, to the point aimed at.

Let us now take a brief survey of some of the chief roads of old Cornwall, and see how far their direction squares with these rules. We may begin doubtingly with the road from the Land's End district, as it comes straight through Penzance, there meeting the way from Newlyn and the steep hill west of it to join the littoral road to Marazion, thence proceeding by St. Hilary, Bosence, Townsend, Bluestone, and Blackwater to the Four Burrows, and thence to Mitchell; over the bleak Goss Moors to Bodmin, and onward through a still higher and rougher tract to Launceston. This This may be called the main ridgeway of the county, from which other great roads diverge, having again their own byeways. Thus, not far from Marazion, is given off a road leading straight to Helston and right up its long and steep street, and thence over high ground, down a like hill at Penryn, an important seaport in very early times-ages before Falmouth existed. Another secondary road passes from Hayle north of Camborne, right down hill and up again through the old town of Redruth, to fall in with the trunk line not far from Scorrier, where it gives off the road which went down into Truro by Chapel-hill, and sent branches to the Four Burrows

road, up Kenwyn-hill to Zelah, up Mitchell-hill to the old town so called, and possibly in the opposite direction by Carnon to Penryn-the southern line proceeding up St. Clement's-hill to Grampound, probably Voluba, which the tide once approached, where the breasting of the long hill is equally marked; staight to St. Austell, and again up hill out of it; thence over high ground to the head of the Fowey estuary at Lostwithiel, the Uxela of Ptolemy, and right up the eastern hill to the elevated station at Liskeard, to terminate, as far as Cornwall is concerned, in a roof-like descent at Saltash, and by the branch through Callington, at the head of the tidal water of the Tamar, on the road to Tavistock. It would be tedious to enter into detailed illustration of the same go-ahead plan of road-making from the branch lines, as in the instances of Tregony, Fowey, and Looe, or from the main north line given off near the Blue Anchor to St. Columb, Wadebridge, Camelford, and Stratton, which place was in direct communication with the Roman roads through Devon, and the large camp at Clovelly Dykes.

Enough has been said, I think, to show that the opinion is at least tenable that the old roads of Cornwall were essentially Roman works. The deviations, which make their original straightness and steepness less obvious now, have, with scarce an exception, been effected within my own recollection; but a wholesome expansion of lung, super antiquas vias, may serve pleasantly to impress my argument. It seems to me, indeed, although it may sound paradoxical, that the Roman roads in Cornwall have been overlooked by over looking for them. Hunting about for bits of dykes and stonework, which the Roman engineer would no more than one of to-day dream of making on firm, welldrained ground; the antiquary failed to notice the plain fact that the common highway bore the impress of the

1 These old roads are often, and oftener were, agreeably marked by a large space of unenclosed ground on each side of the roadway, allowing a gallop on the turf.

Roman system. A good many scraps of the peculiar work referred to have, however, been found, and others may yet be discovered; but they must always be insignificant by the side of the great system of internal communication which, if my reasoning is sound, marked the complete, and, in many respects, beneficent occupation of Cornwall by the Romans, the greatest of administrative powers.

C. BARHAM, M.D. Cantab., Vice-President Royal Institution of Cornwall, Sec. for Cornwall, Camb. Arch. Assoc.

PEMBROKESHIRE HOUSES.

IN no district of Wales exist more numerous, and in many respects more interesting, examples of domestic architecture than in the southern parts of Pembrokeshire. For this peculiarity more reasons than one may be suggested. The character and abundance of the stone, the excellence of the lime, and the fact that the builders were exposed to sudden and numerous attacks from their neighbours, who would naturally give as much annoyance as they could to the strangers who had taken possession of the fairest portions of their land, would have considerable influence in determining the character of their houses. Security against sudden attack would be their first object; and although no such ordinary dwellings could withstand a formal blockade,

1 Dr. Borlase satisfied himself that he found various portions of Roman road between Lostwithiel and Liskeard, and between the former town and Fowey, by Castle Dour,-a fort most judiciously placed, and occupied as lately as the Great Rebellion; and I doubt not that he is generally correct as to the lines of way at any rate. Whether his opinion is equally well founded in assigning to the Romans the great work, half road, half fortification, which extends from Looe to Lerrin, a distance of seven miles, called the Giant's Hedge, may be open to question. The minute examination and discriminating criticism of Mr. McLauchlan (Report, R. I. C., 1846) incline him to believe that the ground was occupied, and the works chiefly constructed by some powerful Celtic tribe long before the Romans visited the island.

like a more regular fortress, yet it was not difficult to make them secure enough so as to hold out until aid could be obtained from the nearest castle. How thickly these are scattered over this part of Pembrokeshire is well known, so that it may be assumed that relief was always to be had from no great distance. Additional security was furnished also by the numerous semifortified churches with their lofty towers, which could provide a temporary refuge: hence the constant use of stone vaults for the lower portions of buildings, which would prevent the occupants of the upper part from being burnt out. In some instances a movable ladder was the only means of communication, as, for example, in the curious church of Manorbier. In this instance the proximity of the Castle might be thought to have made this precaution less necessary than in other and less favoured districts.

These motives would induce settlers in a disturbed district to provide as far as they could for their safety : hence we find in this locality so many houses provided with stone-vaulted roofs and ceilings. Some, however, of these may, perhaps, be classed among the smaller castles not unlike the Pele castles of the north, or those common in some parts of Ireland. Such, for instance, is Bonville Court, near Tenby, described in the Archæologia Cambrensis for 1868, as also the more important house of Eastington or Estington (Iestyn's town), south-west of Pembroke, assigned by Mr. Parker to the time of Edward II. An earlier house probably existed on this spot, as the estate came by marriage with a coheiress of Meirchion ap Rhys into the Perrot family long before the time of Edward. In both these houses these stone vaults are found in all the rooms. Bonville Court is now degraded as a depository of rubbish. Eastington is occupied by a farmer.

One of the most remarkable examples of what Leland calls a castlet, is the isolated square building at Angle, traditionally known as the Rectory. It is, either as a rectory or a fortified house, unique as regards the whole