I note with some interest the fact that your list of medieval fortified sites in Kent contains no mention of Faversham, even as a "possible", despite documentary references to a "castell" in the Ospringe area, and a very strong local tradition concerning "King Stephen's Castle", evidence of which may very probably have been obliterated by the building of the railway. The strategic importance of Faversham as a port heavily involved with Flemish trade (and the probable seat of William of Ypres' power - certainly a major source of his income - from 1141 to 1148), makes it most unlikely that there would NOT be some kind of military installation in addition to the royal manor-house (which was demolished to make way for the abbey). The fact that Davington Priory, situated on the only hill overlooking Faversham and its harbour, was founded in 1153, the year that Stephen began the agreed demolition of many of the castles that had been hastily assembled during his reign, gives an alternative (and in my opinion much stronger) possibility for the site of this putative castle. (Jack Longpers corr.)
Much has already been said in the former parts of these volumes, of the different opinions of learned men where the Roman station, called in the second iter of Antonine Durolevum, ought to be placed. Most of the copies of Antonine make the distance from the last station Durobrovis, which is allowed by all to be Rochester, to the station of Durolevum, to be xiii or xvi miles, though the Peutongerian tables make it only vii. If the number xvi is right, no place bids so fair for it as Judde-hill, in this parish, which then would have every probable circumstance in favor of it. The Romans undoubtedly had some strong military post on this hill, on the summit of which there are the remains of a very deep and broad ditch, the south and east sides are still entire, as is a small part of the north side at the eastern corners of it, the remaining part of the north side was filled up not many years since. The west side has nothing left of it; close within the southern part of it is a high mount of earth thrown up to a considerable height above the ground round it, the scite of Judde house, and the gardens are contained within it. The form of it seems to have been a square, with the corners rounded, and to have contained between three and four acres of ground within its area, the common people call it king Stephen's castle, but it is certainly of a much older date. At a small distance from it, on the opposite, or north side of the high road, there are several breast works cast up across the field facing the west. At the bottom of the hill, in the next field to this, are the ruins of Stone chapel, in which numbers of Roman bricks are interspersed among the flints, and in the midst of the south wall of it, there is a separate piece of a Roman building, about a rod in length, and near three feet high, composed of two rows of Roman tiles, of about fourteen inches square each, and on them are laid small stones hewed, but of no regular size or shape, for about a foot high, and then tiles again, and so on alternately. When the new road from the summit of Judde hill westward was dug down, quantities of fragments of Roman culinary ware, and a coin of Vespasian were found intermixed with many parcels of oyster shells and in the gardens of Judde house, at different times, coins of Adrian M. Aurelius, Arcadius, and others, have been discovered. And at about a mile distance north-eastward, on Davington hill, almost adjoining to the town of Faversham, within these few years, a Roman burial place has been discovered, and many Roman coins, urns, and other relics of antiquity dug up there, as there have been at different times at Faversham, and places adjoining to it, especially along the London high road. Besides this, the vicinity of this place to the stream at Ospringe, a is strong argument in its favor, and still more its nearness to Faversham, for Bede notes in several places, that the villæ regiæ of the Saxons were mostly placed on or near where in former ages the Roman stations had been before. (Hasted)
This has been taken as of Roman interest only and, therefore, has missed medieval scholars although the description of "a high mount of earth" is suggestive of a motte in an existing Roman camp - a not uncommon castle form. However even as a Roman camp, despite much Roman finds, this report of Hasted has been dismissed but Hasted was not a fanciful observer and his report is detailed. The destruction of the site in the late C19 does seem fairly intense but by then steam traction engines allowed deep ploughing.
There is some credence to Judd Hill as a possible early castle site.
1. The main purpose of royal castles was to administer the kingdom. The practical consideration of putting a castle on a main road to ease this is not uncommon, although generally the county town was the location of the royal castle. There is some suggestion that early in the Norman conquest there was less centralisation and more smaller centres of administration.
2. One of the best ways to administer the kingdom was to use the previous administrative systems and also take the symbolic authority of the previous administration. Thus there is a tendency for the Normans to use sites associated with Saxon administration. If this was the site of a Saxon Royal palace (and a highly unusually stone chapel of Saxon date is highly suggestive.) then placing a Norman castle on the site would not be unusual. There is a less strong association with Roman sites and the Imperial kudos but placing a motte within existing Roman defences is certainly not unknown.
3. Close by is a place called Queen Court. There are several examples of coupled royal palaces.
4. The tendency of high status sites to gain kudos and remain in use as high status. The Judd Hill site seems to be occupied by a deer park.