as to the Roman walls of Chester or the probably RomanoBritish lines of Tamworth. Just so, to take an instance from France, the mediaval motte is at Châteaumeillant in the Département du Cher, a subsequent addition to a Roman or Gallo-Roman castrum. Now, it is Mr. Clark's contention that many a burh mentioned in the Chronicle can still be identified, and stands a silent witness to the struggle of eleven centuries ago. The strongest case, in our opinion, though it is not appealed to, we think, in these volumes, is that of the mound at Kennardington, in Kent, which may safely be identified with that which the Danes are recorded to have raised at Apuldore' in 893. But some of these instances are puzzling enough. There is, no doubt, at Towcester a typical mound-down by the river, as usual-but was it, as Mr. Clark assumes, the work of Edward? That king was keenly alive to the value of Roman sites, and at Towcester, as at Colchester about the same time, he clearly repaired the Roman walls around the town. Was the mound, as at Chester, an addition to these defences? We cannot tell. Again, the fortified mound was destroyed as easily as it was formed. We read, in some cases, of this being done: we have ourselves seen it done in the case of an interesting mound on the right bank of the Leahalfway between Bow Bridge and the Old Ford'-within the last few years. Thus it is difficult to be sure about the burh now standing at Tempsford, an advanced post of the Danes, which was stormed by the English in 921 (i. 22, 78), not, as Mr. Clark has inadvertently stated since, erected in that year by Edward.* How closely all this warfare was connected with rivers is seen also in the system of twin mounds, one on each bank of the stream, employed, it would seem, at Buckingham, Hertford, and York. We cannot but think that the existence of such mottes in Flanders is similarly due to the presence of the Northmen on the sluggish rivers of the Low Countries, and that everything points to these strongholds, in England, dating only from their invasions. Mr. Clark has demolished the view that assigns them to the Britons or the Romans; but he has not, we venture to think, gone far enough. The only instance he vouches for their use, before the coming of the Northmen, is that of Taunton, where the earthworks are assigned to Ine the lawgiver, nearly two centuries earlier than any other fortress mentioned in the Saxon Chronicle.' But when we seek, at Taunton, for the mound as the leading and typical feature,' Mr. Clark, though he tells us that there is reason to

† Ibid. xlvi. 211.

suppose

suppose that there was a mound' (i. 22), and that 'the mound has been removed,' is compelled to admit, in his monograph on this interesting castle, that there seems to be no 'record or tradition of a mound in the technical sense,' and that the absence of a mound is rather peculiar.'

6

There is one point which, it seems to us, Mr. Clark does not explain. Why is the castle mound in many cases not, as we might expect, within the city walls, but just outside them? One of the most notable of these is at Rochester, where 'the Boley Hill' stands by the river south of the castle enclosure and of the old town walls. The occasional confusion of which we have spoken as inseparable from Mr. Clark's system of publication, is seen in the statement (i. 19) that the 'work' wrought by the besieging Danes in 885 is no doubt the great mound that still remains outside the castle and the Roman area' (i. 19); for, in his able monograph on Rochester Castle, the writer argues against the supposition that this mound was a work of the Danes (ii. 406, 420). The case is a test one, for Florence of Worcester distinctly speaks of an arx, a castellum firmum raised by the Danes in their siege of 885. On the other hand, the Boley Hill was included in the city defences by projecting earthworks, just as, we may add, was the 'Danejohn -a moated mound was still spoken of as a 'Dongeon' in the seventeenth century-at Canterbury, 'older even than the bank and ditch of the city, which were laid out at an angle to include it.' Chester is a striking case in point. The mound there stood by the river outside the Roman lines, which must, Mr. Freeman suggested, have been extended by the Lady of the Mercians to take it in. At Leicester, again, in Mr. Clark's words, the mound stood between the Soar and the Roman Rate (i. 83), though he elsewhere (i. 18) classes it among those within Roman enclosures.' The mound of Oxford Castle, close, as usual, to the river, seems to have similarly stood beyond the line of wall. All we can say of this disposition is that it points either to the work of assailants coming from the river, or of defenders anticipating attacks from that quarter.

The other important point established by Mr. Clark is the essential distinction between the two types of keeps in our English castles. These he terms respectively the 'rectangular' and the 'shell' keep. We cannot say that these terms strike us as satisfactory. We have, on the one hand, solid towers (i. 139), square, or at least rectangular, in form; on the other, a hollow space enclosed by a 'ring wall' (i. 140), circular or polygonal. The names given them would scarcely

make

make this distinction clear. It is pointed out, with justice, of these two types, that 'the shell-keep was the most numerous of the two; but the tower type, being of a more solid and more durable character, has lasted longest, and is at this time so much the most common that it has been designated by writers of authority as the type, instead of as but one of two types, of a Norman keep.' We would add that, while the 'shell' was a mere development of the Anglo-Saxon stronghold, the timber stockade of the latter being replaced by a wall of stone, the ' rectangular' keep, on the contrary, was a wholly new form. Yet those who love to detect evolution in all things may trace some resemblance between the earthen mound, on which the Englishman wrought his stronghold, and the lower half of a Norman 'tower,' above which was the true dwelling. In the chapter devoted to the shell' keep we are shown how its shape was necessary for an artificial mound, which would not have supported the great weight of a solid tower. Within its wall, some ten feet thick and at least twenty feet high, was an open court, with timber dwellings round it built up against the wall. 'Durham Keep is said to have been originally open, but to have been closed to accommodate the bishops, who were forced, by the disturbed state of the country, to reside within it.' Windsor and Belvoir are familiar examples, but have been too completely reconstructed to serve any purpose of instruction. Arundel remains practically unspoilt, and Tamworth, of much interest in many ways, illustrates the erection of domestic buildings within the 'shell' of masonry. But better still is Berkeley, which, as Mr. Clark observes, has been inhabited from its foundation to the present day,' and' is as little altered as is consistent with modern usages and modes of life.' The space within the shell, 45 yards in diameter, is there still partly occupied by the domestic apartments, and the shell stands, in Mr. Freeman's words, on the mound on which the great Earl of the West-Saxons had once dwelled.' It is a remarkable feature of the shell-keep on its mound that it usually stood on the actual enceinte, or line of outer defence of the fortress, which was thus broken by it. The curtain wall consequently was carried up the mound on both sides till it abutted on the actual keep, which consequently stood half within and half without the fortress. The original approach to these structures was by a timber bridge raised on posts, as is seen in the famous Tapestry of Bayeux, and is well described in a passage belonging to the early years of the twelfth century, which is appositely quoted by Mr. Clark from the 'Acta Sanctorum.'

D 2

6

Although,

Although, writes our author, the rectangular keep must be confessed to be inferior in grandeur and in completeness of outline' to the rival type, yet, in describing it, he is led to style it, of all military structures, the simplest in form, the grandest in outline and dimensions, the sternest in passive strength, the most durable in design and workmanship, and, in most cases, by some years the earliest in date.' To those who have made a study of these magnificent structures there will seem no exaggeration in so high a eulogy as this. For perfect adaptation of means to end, for the marvellous combining of actual strength with the effect produced by what Tacitus would have termed an 'arx æternæ dominationis,' it would be difficult to surpass the Norman turres, those 'towers that frown' over town or valley, those grim survivors, in places still, of an age when a conquered people trembled, in impotence, at their sight.

In its typical form, this structure was raised upon a solid platform of masonry of which the sides, sloping outwards, formed a battering plinth for the walls, and not only gave them a sure foundation, but raised them out of reach of the miner's pick or the strokes of the assailant's ram. The walls diminished in thickness at each stage, usually by an internal set-off. Their most distinctive feature is the flat pilaster, broad but of slight projection, though in rare cases developing into tower buttresses, as at the peculiar keep of Colchester. The groundfloor, as we may term it, of the keep, was little more than a cellar for the stores that would enable the garrison to stand a siege. Its walls, at the Tower of London, are from 12 to 15 feet thick; at Colchester, thicker still. A few narrow loopholes, splaying on the inside, could do little more than make darkness visible, but they were, skilfully, so constructed as to enable an archer to shoot from within, while virtually excluding the enemy's missiles and, above all, the dreaded firebrand. Over this was the first floor resting on mighty timber balks. Here the walls became a foot less thick, the loops a trifle wider and less sparingly bestowed. On this floor, it is thought, were the quarters of the garrison. Above this usually came the state or principal floor of the keep, tenanted by its lord or commander. But, at the Tower, it was the fourth and uppermost stage that fulfilled this function. In these two upper floors windows became possible; and mural galleries, in many cases, threaded the walls. In some keeps, especially in those constructed for a private owner, the walls are honeycombed by chambers evidently meant for sleeping in. In the noble keep of Hedingham, which Mr. Clark does not describe, this domestic arrangement is specially well seen. Where the third stage was the principal

one,

one, the floor above it is supposed to have formed a kind of fighting deck. In some cases, mainly in the small keeps, there may have been originally, in all, only three stages. Essentially military structures, these keeps, especially the early ones, offer little in the way of architectural detail: a garderobe here, a fire-place there remains to speak of human habitation; but gloomy, cold, and comfortless they must always have been, in spite of their stately grandeur. Passive strength was the object the builders kept before them: they sought to enable a handful of men to keep at bay, for an indefinite period, a considerable force. Against such a keep the artillery of the time was, practically, powerless: it was only by mining, as at the great sieges of Rochester in 1215 and Bedford in 1224, that these formidable strongholds could be assailed. The mediæval mine consisted of a gallery driven under an angle of the building, which was propped up by timber: this was fired, by the miners, when complete; and the weight of the superincumbent masonry, when deprived of support, caused the walls to crack and the angle to topple over. But if the foundations were strong and the masonry above them compact, the keep might yet hold together above the chasm.

The protection of the entrance to such a stronghold was a matter of anxious care. So far as is at present known, the practice in the earliest keeps was to place it at the first floor, as at Colchester, or, if below, at least above a high plinth, as at Malling, so that it could only be reached by a wooden ladder or staircase, which, in case of investment, could be removed. Within, the entrance was so arranged that a very few men could, generally, defend it with ease; and it was usually placed as far as possible from the main staircase and from the precious well. There was a further development known as the 'forebuilding,' of which the details, as Mr. Clark reminds us, 'have been but little studied.' This was a covered staircase outside the keep, passing through towers and doorways, and leading to a vestibule at the entrance of the keep, often with a prison beneath it. Although belonging to the 'rectangular' type, it has been applied to Berkeley, which is a 'shell,' but peculiar, and of late construction. A forebuilding may, in our opinion, prove valuable for determining, when doubtful, the probable date of a keep. It has been supposed,' Mr. Clark admits, to mark a late keep, but there is a forebuilding at Arques, usually regarded as a very early one.' If, however, as we hold certain, M. Deville, Mr. Freeman, and Mr. Clark are all alike mistaken about this famous keep; if it was built, not by the Conqueror's uncle, in 1039-1043, but by Henry I., about 1123, the above

objection