Ireland, where they are called Crannoges, after the name given to them by the chroniclers, who preserve examples of their comparatively recent use. A great stimulus has lately been given to the investigation of this kind of archæological phenomenon, by its discovery in other parts of the world, and especially in Switzerland, where it has been discovered on so large a scale, and so affluent in relics, as to afford ample means for examining the domestic habits of the people who abode in such dwellings.

Whether in imitation of the Romans, or from some conception of their own, possibly earlier than the Roman invasion, the inhabitants of Scotland possessed a wall strengthened by a system of forts. It is fortunate that it was seen by the antiquary Gordon, and caught a strong hold of his attention. He has accordingly followed its track, and described a great deal that agricultural improvement has obliterated. He finds its northern commencement about a mile from Galashiels, on the river Gala, a tributary to the Tweed on its northern side; and there is a conjecture that it may have been carried from the other side of the stream across to the east coast. The most southerly trace of it is at

1 One Thomas Phettiplace, in his answer to an inquiry from the Government as to what castles or forts O'Neil hath, and of what strength they be, states (May 15, 1567), " For castles, I think it be not unknown unto your honours he trusteth no point thereunto for his safety, as appeareth by the raising of the strongest castles of all his countreys; and that fortification which he only dependeth upon is in certain fresh-water loghes in his country, which from the sea there come neither ship nor boat to approach them. It is thought that there, in the said fortified islands, lyeth all his plate, which is much, and money, prisoners, and gages: which islands hath in wars to fore been attempted, and now of late again by the Lord-Deputy there, Sir Harry Sydney, which, for want of means for safe-conducts upon the water, it hath not prevailed."—Cited in Lubbock, Prehistoric Times, 120.

Peel Fell, in Cumberland; its profile is a ditch between two walls. It has three local names: "The Catrail," "The Deil's Dyke," and "The Picts' Work Ditch." It passes through the most classical portions of the border land, by Yarrow, Deloraine Burn, Melrose, and Liddesdale, then near the Leepsteel and Hermitage Castle. Gordon found its most distinct vestiges to be "24 and 26 feet broad and very deep, the ramparts on every side 6 or 7 feet in perpendicular height, and each of them 10 or 12 feet thick." From the phenomenon that the moss has at one place thickened to a level with the top, so that the sides of the wall are exposed by digging, it is supposed that the work is of extreme antiquity. There are several hill-forts on the line of this rampart, so disposed as to leave little doubt that they are elements of the system of fortification connected with the walls and ditch.

A set of edifices of a character totally distinct from any of the works just described, have been usually associated with them as a portion of the defences of ancient Scotland. In the north-western parts of Invernessshire, in Ross-shire, Sutherland, and Orkney, there may still be seen, sometimes in deep valleys, at others on picturesque points of rock, round towers extremely symmetrical in their structure. They are circular, on a broad base, which narrows upwards with a graceful curve, like that on which Smeaton designed his lighthouse system, as being the form which nature proclaimed in the profile of the trunk of a tree, to be that in which strength and beauty united reach their highest development. They are They are called Burghs," Danish Towers," and "Pictish Towers." These build

66

1 Iter. Sept., 102.

2

Wilson, ii. 79.

ings, at a distance, might pass for remnants of Gothic castles of the days of the English Edwards, but a close examination shows them to be far older.

They are the arch.

They

work of masons unacquainted with the have no roofs, nor visible arrangements for retaining a roof; and the minor openings, which in the Norman style and the others derived from it would be closed by the arch, are simply flagged. From their perfect roundness the stranger would expect to find them built of finely-hewn stone in courses, but when he examines them he will find no mark of a tool on anything connected with their structure. Each is made of the surrounding shingle without mortar; and if any artificial means had been taken to procure the material, it must have been little more than the breaking-up of great stones by tossing them from precipices. The secret of keeping the pure rounded outline with materials so rude is, that the stones are all thin and flat, and chosen so as to give no opportunity for deviations exceeding half or a quarter of an inch; hence, though symmetrically round as seen from a little distance, the surface is rough. The inside is as singular as the out. There are, in fact, in all of them two concentric walls-the outer circular, and widening to the base; the inner circular, and perpendicular like a well. The two thus approach each other as they rise in height. All round the inner wall are tiers of square openings into the space between the walls; and thus the interior resembles in some measure a columbarium, or Roman tomb-chamber for the reception of urns, only that the orifices in the burgh are fewer and farther apart. The space between the walls to which these openings lead is paved or flagged, so as to make several storeys of chambers or galleries, as they

are called, all of course decreasing in breadth the higher up they are. They are divided from each other into lengths by slabs, and in the larger lengths there are generally remnants of rude stairs, leading, as it were, through the successive storeys. Such is generally the type of these strange edifices, though there are slight variations from the standard. At Achir na Kyle in Sutherlandshire, and Dunalishaig on the Dornoch Firth, the chambers between the walls incline to the round or oval form instead of the rectangular.' The chief variation in these buildings is, however, in their dimensions, and on this something will have to be said.

And now comes the question-For what purpose were these strange structures raised? The answer has ever been immediate and uniform, as if it did not require a second thought; in fact, it has been an answer or decision without a question, and is to the effect that they are fortresses. The appearance decides that at once; but if they are viewed with a little consideration, the appearance is seen to decide too much. They resemble, as has been said, the towers of a castle of the Edwards; but such towers were made to flank curtains of wall, and would have been very incapable of defence standing alone. The fact is, that it was not till heavy artillery came into use that a small round building could defend itself; hence the Martello towers. analogy is as if some very ancient wooden pipe were pronounced to be an instrument of war because its shape resembles that of an Armstrong gun.

The

In truth, from anything we can learn about defensive works earlier than the time of the Normans, we are not to expect to find them in the shape of castles built 1 See the ground-plan in Cordiner's Antiquities, 118.

house-fashion. The defences were such as we have seen-great heaps of stones surrounding, on the crests of hills, open spaces of country, in which large bodies of people could assemble. Some of the burghs would not hold twenty persons. The largest of them might perhaps, from their size and strength, be made in some measure defensible. We know that the largest of them all was sought as a place of refuge, and was defended. Erland, the son of Harold, having carried off a beautiful Norwegian widow, took refuge with her in the burgh of Mousa in Shetland, where her son besieged him for some time. Places not built as castles have often been so used in emergencies: the Normans, for instance, fortified the massive tombs which the Romans used to build over their dead. One of them is now the Castle of St Angelo; and the tomb of Cecilia Metella had doubtless often been a tower of strength. Mousa is by far the largest of the burghs, and with a little assistance could easily be made into a sort of castle.

The difficulty in supposing them to be strongholds is to find how they could assail any enemy occupied in pulling down their uncemented walls. There are no orifices of any kind outwards by which missiles or other weapons could have been used. The shape of the buildings is the worst possible for assailing an enemy. A square block is rather feeble; it requires something to flank it-that is, to be at right angles with its faces; and in the square towers afterwards built in Scotland. there were very ingenious economical devices for effecting this-as, for instance, when it had to be done on a very small scale, by overlapping works at the top, or by bastions or turrets at the angles. The square tower may have both these assailants; the round can