the majority of the burhs of Aethelfleda and Edward were wrought at certain places and in certain strategical positions.

Many of them were evidently designed for the purpose of commanding or barring the passage of rivers. Thus the two burhs at Buckingham, the two at Hertford, and the second at Nottingham, all point to river warfare. They forcibly remind us of the two earlier works of Alfred on the river Lea, of the fortified bridge-heads of Charles the Bald, and of the two later (?) mound fortresses of the Conqueror at York.

In some instances, they were designed to overawe or contain already existing Danish settlements, as at Bedford and Stamford.

In others, they defended outlying provinces, as at Bridgnorth and Bakewell. All this points to fortress building for the defence of districts or towns, rather than to the construction of urban enceintes.

It is true that Florence usually translates the burh of the Chronicle by the Latin urbs, but he also mentions six forts as having been constructed by Aethelfleda and Edward.1

That the Roman walls of certain cities were also renovated is likewise abundantly evident: as witness the cases of Chester, Towcester, Manchester, and probably Maldon (Florence-urbem reædificavit).

Bearing these points in mind, it will be seen that some of the fortresses erected in the great burh-building age might well have been of the citadel or moated mound type.

Evidence to this effect is forthcoming in the cases of Eddisbury, Bridgnorth, Bakewell, and Witham.

The case of Eddisbury has already been alluded to. Panpudding Hill, at Bridgnorth, was identified by both Eyton and Clark as the fortress which Aethelfleda erected, 913. This identification was based upon a deed dated at Brugg, 1299, which describes an acre of land in the fields of Oldbury, as bounded on one side by the land of John of Oldbury, and on the other by the road 1 Sceargete, arx. Bridgnorth, arx. Stafford, arx. Munitiones. Stamford, arx.

1906

Buckingham,

19

which leads to Oldbury under the old castle. This exactly locates the position of the Panpudding Hill, which is artificially scarped and terraced. The summit is further defended by a parapet, and it is separated from an adjoining but ill-defined platform by a deep fosse. Further evidence not only confirms this view, but also points to the conclusion that this Hill was the site of the castle of Robert of Belesme.

Florence of Worcester1 relates how Robert, Earl of Shrewsbury, began to repair and surround with a broad and lofty wall the fortress which Aethelfleda built at a place called Brycge, on the west bank of the river Severn; and that he hastened the completion of the walls and towers of the castles of Brycge and Caroclove, having the works carried on by day and by night.

Henry I sat down before Bridgnorth, and began to construct machines and erect a strong fort, "castellum firmare coepit."

In close proximity to Panpudding Hill, but separated from it by a ravine-like dip in the ground, is another earthwork of similar type. It is suggested that this is the castellum of Henry I.

That portion of the well-built Tower of Bridgnorth which stands on the rocky prominence in the immediate neighbourhood can scarcely be the remains of the hastilybuilt "towers and walls" of Belesme.2

The first notice of the Tower of Bridgnorth is in 1169, when £26 6s. 4d. were spent on the works.3

1 Flor Wigorn., vol. ii, p. 49 :--" Arcem quam in occidentali Sabrina fluminis plaga, in loco qui Brycge dicitur lingua Saxonica, Aegelfleda Merciorum domina quondam construxerat, fratre suo Eadwardo Seniore regnante, Scrobbesbyriensis Comes Rotbertus de Beleasmo, Rogeri Comitis filius, contra Regem Heinricum, ut exitus rei probavit, muro lato et alto, summoque restaurare coepit." P. 50:-" Muros quoque ac turres castellorum videlicet Brycge et Caroclove, die noctuque laborando et operando, perficere modis omnibus festinavit."

2 There is good reason to believe that Belesme had only lived one year in England prior to his submission to Henry I, 1102; see Eyton's "Antiquities of Shropshire," p. 245.

3 It is doubtful whether there was a tower at Bridgnorth even in the year 1154. See the passage in Duchesne, p. 991, where the Chronicler expressly mentions the "Tower of Gloucester," and the Castles of Brug, Wigemore, and Deobens,

Florence's identification of the site of Aethelfleda's arx with that of the castle of Belesme was probably the result of local topographical knowledge, for he lived not more than thirty miles distant.

That Aethelfleda built a fort at Bridgnorth, and did not establish a town, is further proved by the fact that no mention of Brycge is made in Domesday.

The "motte" at Bakewell stands at one side of a more or less oblong enclosure; it is partly artificial, and has a cup-like depression on its summit. It is separated from the town of Bakewell by the river Wye. The church, with its pre-Norman cross and other Saxon remains, is nearly half a mile distant. It may therefore to this day be described, in the words of the Chronicle, an. 924, as being in the neighbourhood of Bakewell.

A reference to the Chronicle makes it clear that Edward did not entrench the town of Bakewell, but that he constructed and manned a fort in its vicinity. So, too, Florence of Worcester states that he placed some stout soldiers in it: "Inde cum exercitu ab Beadecanweallan profectus, non longe ab ea urbem construxit, et in illa milites viribus robustos posuit" (an. 921). In translating the burh of the Chronicle by the Latin urbs, Florence no doubt gave the word burh its contemporary meaning. The entry, however, distinctly points to the construction of a fort, in contradistinction to a town. No record of a Norman castle is forthcoming. The burh which Edward the Elder workte & getimbrede aet Witham (Florence: ædificaretur et ædificata firmaretur) was evidently no ordinary town enceinte. Although much defaced, there are good grounds for regarding it as a citadel camp. Mr. I. C. Gould' states that the original fort seems to have consisted of a large enclosure of about 400 by 350 yards, with an inner ward or "keep" of 200 by 175 yards. The inner rampart rises from a base about 10 ft. above the surrounding enclosure.

Strutt illustrates this central citadel or "keep,"showing a low flat-topped mound, with parapet, fosse, and rampart. The central citadel must have occupied an area of nearly 1 "Victoria County History of Essex," p. 288. 2 "Manners, Customs," etc., p. 24. 1775.

five acres. Although of large dimensions, it is evident that the earthwork at Witham approximated more to the moated mound, or citadel type of fortification, than to that of a simple town enclosure.

As the result of this enquiry, the following conclusions may, I think, reasonably be formulated :

(a) The moated mound type of fortress did not owe its origin to one nationality; witness its widespread

distribution.

(b) It was common to many nations and periods; witness the throwing-up of mounds by Indian chieftains in the sixteenth century.

(c) Whilst, so far as the British Isles are concerned, the case for Norman origin and occupation may be regarded as definitely proved, there are good grounds for concluding that some examples are of much earlier date.

(d) The archæological evidence, as revealed by excavation, points to a very early date in the case of some "mottes." Therefore. whilst historical reference to castles of this type is of great value, it does not justify the conclusion that they were all constructed in contemporary and historic times.

I would therefore deprecate the somewhat hasty generalisations of modern antiquaries as tending prematurely to close an interesting inquiry.

Paradoxical as the statement may seem, I venture to predict that the fuller our knowledge of this subject becomes, the more complete will be the proof of Norman parentage, and the clearer will be the evidence of an earlier origin.

BY R. OLIVER HESLOP, Esq., M.A., F.S.A., HONORARY CORRESPONDENT,

(Read November 21st, 1906.)

"PILGRIMAGE," in which some threescore wayfarers take part, may not be productive of fresh light upon the Roman Wall, and the conflicting views advanced respecting the lines of Wall and Vallum from Tyne to Solway; but the expedition that took place under the joint organization of the Cumberland and Westmoreland Society and the Newcastle Society of Antiquaries has served at least to bring prominently forward to a large and interested company the number and complexity of the problems in these lines yet unsolved.

Beginning on Saturday afternoon, June 23rd, the section from Wallsend-on-Tyne was covered. On Monday, the 25th, the party, reinforced to the number of over sixty members, set out westward from Newcastle, and continued the "pilgrimage" day by day, until they arrived at Bowness on Solway on the Saturday following, completing the entire length of the Great Wall of 72 miles in seven days.

At Wallsend (Segedunum) Mr. W. S. Corder was conductor, and the proceedings began on the spot, marked by a bronze tablet in the shipyard of Messrs. Swan, Hunter, and Wigham Richardson & Co., Limited, where the tail-end of the wall had been continued from the south-east corner of the stationary camp to low-water mark in the river Tyne. From Newcastle (Pons Aelii) the works as far westward as Benwell (Condercum) are