T. E. Lawrence’s Undergraduate Dissertation – Page 66

The Norman keep was thus rather an ineffective fortress; it could be mined with the greatest ease, as King John proved at Rochester, and it provided little accommodation. Apparently its builders felt its defects, for they never allowed a keep to stand alone. Sometimes it is within a Roman wall, as at Porchester (...-"") and Pevensey, more often there was a ring wall drawn around it as at Ludlow: and this wall was of course provided with curtain towers. Other keeps have only earthworks older or contemporary and once no doubt palisaded, for their outer defence. Usually they stand on the strongest point of the castle, but sometimes, as at Richmond, they are so placed as to defend the weakest part of the outer wall. In either case these outworks are seldom elaborate enough to stand the determined attack of a considerable force. The square keep is an ideal defence against a border raid, and in the North of England it survived in this purpose in the peel-towers to the 16th century: but against an enemy with leisure, or sufficient resources to drive a mine, the reduction of such a castle was only a question of time. The form probably only remained in favour for a very brief period, but this included the years in which the first Crusaders went out to Palestine. One may determine therefore with reasonable certainty that Normans and Provencals alike were accustomed to build castles of a square keep and a curtain wall with towers around it: only in the Norman castles greater stress was laid on the keep, and in the Provencal on the curtain wall.