still bearing the name of Castle Hill, there stood in former times beyond all doubt a royal castle." (N. S. A.) Mr Neilson says, "It certainly bears out its reputation as an artificial mound."

ROSEMARKIE, Cromarty. - Was made a royal burgh by Alexander II., so the castle must have been originally royal. "Immediately above the town is a mound of nearly circular form, and level on the top, which seems to be artificial, and has always been called the Court Hill." (N. S. A.)

Even if we had no other evidence that motte-castles were of Norman construction, this list would be very significant. But taken in connection with the evidence for the Norman origin of the English, Welsh, and Irish mottes, it supplies ample proof that in Scotland, as elsewhere, the Norman and feudal settlement had its material guarantees in the castles which were planted all over the land, and that these castles were the simple structures of earth and wood, whose earthen remains have been the cause of so much mystification.

CHAPTER XI

MOTTE-CASTLES IN IRELAND

In the year 1169, when the first Norman invaders landed in Ireland, the private castle had been in existence in England for more than a hundred years, and had it been suited to the social organisation of the Irish people, there had been plenty of time for its introduction into Ireland. Nor are we in a position to deny that some chieftain with a leaning towards foreign fashions may have built for himself a castle in the Anglo-Norman style; all we can say is that there is not the slightest evidence of such a thing. We have two contemporary accounts of the Norman settlement in Ireland, the one given by Giraldus in his Expugnatio Hibernica, and the Anglo-Norman poem, edited by Mr Goddard H. Orpen, under the title of the "Song of Dermot and the Earl." 2 Now Giraldus expressly tells us that the Irish did not

1

use castles, but preferred to take refuge in their forests and bogs. The statement is a remarkable one, since Ireland abounds with defensive works of a very ancient character; are we to suppose that these were only used in the prehistoric period? But if castles of the Norman kind had been in general use in Ireland in the 12th century, we should certainly hear of their having been a serious hindrance to the invaders. The history of the invasion, however, completely confirms the statement of Giraldus; we never once hear of the Irish defending themselves in a castle. When they do stand a siege, it is in a walled town, and a town which has been walled, not by themselves, but by the Danes, to whom Giraldus expressly attributes these walls. Moreover, the repeated insistence of Giraldus on the necessity of systematic incastellation of the whole country is proof enough that no such incastellation existed.

It is true that in some of the earliest Irish literature we hear of the dun, lis, or rath (the words are interchangeable), which encircled the chieftain's house.

THE RATHS OF IRELAND

325

Many descriptions of royal abodes in Irish poems are evidently purely fanciful, but underneath the poetical adornments we can discern the features of the great wooden hall which appears to have been the residence of the tribal chieftain, whether Keltic, Norse, or Saxon, throughout the whole north of Europe in early times.1 The thousands of earthen rings, generally called raths, which are still scattered over Ireland, are believed to be the enclosures of these kings' or chieftains' homesteads. Were they intended for serious military defence? We are not in a position to answer this question categorically, but the plans of a number of them which we have examined do not suggest anything but a very slight fortification, sufficient to keep off wolves. At all events we never hear of these raths or duns standing a siege; the conquering raider comes, sees, and burns. We are therefore justified in concluding that they did not at all correspond to what we mean by a private castle. And most certainly the motte-castle, with its very small citadel, and its limited accommodation for the flocks and herds of a tribe, was utterly unsuited to the requirements of the tribal system.

A good deal of light is thrown on the way in which Irish chieftains regarded private castles at the time of the invasion by the well-known story of one who refused a castle offered him by the invaders, saying that he preferred a castle of bones to a castle of stones. Whether legendary or not, it represents the natural feeling of a man who had been accustomed to sleep trustfully in the midst of men of his own blood, tied to him by the bonds of the clan. The clan system in

Ireland undoubtedly led to great misery through the absence of a central authority to check the raids of one clan upon another; but though we occasionally hear of a chieftain being murdered "by his own," we have no reason to think that clan loyalty was not sufficient, as a rule, for the internal safety of the community. So that a popular chieftain might well refuse a fortification which had every mark of a hateful and suspicious invader.1

2

Unfortunately there is or has been until quite recently a strong prejudice in the minds of Irish antiquaries that works of the motte-and-bailey kind belong to the prehistoric age of Ireland. Irish scholars indeed admit that the word mota is not found in any Irish MS. which dates from before the Norman invasion of Ireland. We must therefore bear in mind that when they tell us that such and such an ancient book mentions the "mote" at Naas or elsewhere, what they mean is that it mentions a dun, or rath, or longport, which they imagine to be the same as a motte. But this is begging the whole question. There is not the slightest proof that any of these words meant a motte. Dun is often taken to mean a hill (perhaps from its resemblance to Anglo-Saxon dun), but Keltic scholars are now agreed that it is cognate with the German zaun and AngloSaxon tun, meaning a fenced enclosure. It may be applied to a fort on a hill, but it may equally well be

3