improbable in the theory that Alfred reared the outer bank of the fortress, recent excavations have shown that the place was occupied by the Romans, and therefore make it certain that its origin was very much earlier than Alfred's time. Moreover, the convergence of several Roman roads at this spot suggests the probability of a Roman station,' while the form of the enclosure renders an earlier origin likely. Domesday Book does not speak of Salisbury as a burgus, and when the burgus of Old Sarum is mentioned in later documents it appears to refer to a district lying at the foot of the Castle Hill, and formerly enclosed with a wall. Nor is it one of the boroughs of the Burghal Hidage. an important place in Saxon times that there was a mint there; and there is evidence of the existence of at least four Saxon churches, as well as a hospital for lepers.3

But that Sarum was is clear from the fact

For more exact knowledge as to the history of this ancient fortress we must wait till the excavations now going on are finished, but in the meanwhile it seems probable that the theory adopted by General Pitt-Rivers is correct. He regarded Old Sarum as a British earthwork, with an inner castle and outer barbicans added by the Normans. After building this castle in the midst of it the Normans appear to have considered the outer and

1 General Pitt-Rivers in his Address to the Salisbury meeting of the Archæological Institute in 1887, says that traces of these roads may still be seen. He adds that Old Sarum does not resemble the generality of ancient British fortifications, in that the rampart is of the same height all round, instead of being lower where the ground is steeper; this led him to think that the original fortress had been modernised in later times. Sir Richard Colt Hoare noticed that the ramparts of Sarum were twice as high as those of the fine prehistoric camps with which he was acquainted. Ancient Wiltshire, p. 226.

78.

2 Benson and Hatcher's Old and New Sarum, p. 604.

3

Cf. Benson and Hatcher, 63, with Beauties of England and Wales, xv.,

larger fortification too valuable to be given up to the public, but retained it under the government of the castellan, and treated it as part of the castle.

There is no mention of the castle of Salisbury in Domesday Book, but the bishop is named as the owner of the manor.1 The episcopal see of Sherborne was transferred to Sarum in 1076 by Bishop Hermann, in accordance with the policy adopted by William I. that episcopal sees should be removed from villages to towns: a measure which in itself is a testimony to the importance of Salisbury at that time. The first mention of the castle is in the charter of Bishop Osmund, 1091. The bishop was allowed to lay the foundations of his new cathedral within the ancient fortress. As might be expected, friction soon arose between the castellans and the ecclesiastics; the castellans claimed the custody of the gates, and sometimes barred the canons, whose houses seem to have been outside the fortress, from access to the church. These quarrels were ended eventually by the removal of the cathedral to the new town of Salisbury at the foot of the hill.

The position of the motte of Old Sarum is exceptional, as it stands in the centre of the outer fortress. This must be owing to the position of the ancient vallum, encircling the summit of one of those round, gradually sloping hills so common in the chalk ranges, which made it necessary to place the motte in the centre, because it was the highest part of the ground. The

1 D. B., i., 66. "Idem episcopus tenet Sarisberie." Part of the land which had been held under the bishop was now held by Edward the Sheriff, the ancestor of the earls of Salisbury. This in itself is a proof that the castle was new. See Freeman, N. C., iv., 797.

" This policy had been dictated by an œcumenical council.

3 He gives to the canons of the church two hides in the manor, “et ante portam castelli Seriberiensis terram ex utraque parte viæ in ortorum domorumque canonicorum necessitate." M. A., vi., 1294.

present excavations have shown that it is in part artificial. But though the citadel was thus exceptionally placed, the principle that communication with the outside must be maintained was carried out; the motte had its own bailey, reaching to the outer vallum. The remains of three cross banks still exist, two of which must have enclosed the magnum ballium which is spoken of in the Pipe Rolls of Henry II. Probably this bailey occupied the south-eastern third of the circle, which included the main gateway and the road to the citadel. In the ditch on the north side of this enclosure, an arched passage, apparently of Norman construction, was found in 1795; it was doubtless a postern or sallyport.' The main entrance is defended by a separate mount with its own ditch, which is conjectured to be of later date than the vallum itself. The area of the top of the motte is about 1 acres, a larger size than usual, but not larger than that of several other important castles.' In Leland's time there was "much notable ruinous building" still remaining of this fortress, and the excavations have already revealed the lower portions of some splendid walls and gateways, and the basement of a late Norman keep which presents some unusual features. The earthworks, however, bear witness to a former wooden stockade both to the citadel and the outer enclosure. The top of the motte is still surrounded by high earthen banks.

As that great building bishop, Roger of Salisbury 1 Gentleman's Magazine, 1795.

* The area of the outer camp is 29 acres.

3 It is unlikely that this is the turris mentioned in the solitary Pipe Roll of Henry I. "In unum ostium faciendum ad cellarium turris Sarum, 20s." This entry is of great interest, as entrances from the outside to the basement of keeps were exceptional in the 12th century; but the basement entrance of Colchester keep has every appearance of having been added by Henry I.