[959] _English Towns and Districts_, p. 152. [960] _Mediæval Military Architecture_, ii. 514. [961] There is a strange use of "castellum," apparently in this sense, in William of Malmesbury's version (ii. 119) of Godwine's speech on the Dover riot (1051). The phrase is "magnates _illius castelli_," which Mr. Freeman unhesitatingly renders "the magistrates of that _town_" (_Norm. Conq._, 2nd ed., ii. 135), a rendering which should be compared with his remarks on "castles" on the next page but one, and in Appendix S. Mr. Clark is of opinion that "whether 'castellum' can [here] be taken for more than the fortified town is uncertain" (_M. M. A._, ii. 8). [962] Skeat's _Etymological Dictionary_; Oliphant's _Old and Middle English_, p. 37. It is not, therefore, strictly accurate to say of the expression "ænne castel," in the chronicle for 1048, that it was "no English name," as Mr. Freeman asserts (_Norm. Conq._, 2nd ed., ii. 137), or to imply that it then first appeared in the language. [963] _Norman Conquest_ (2nd ed.), ii. 189. [964] Ed. Howlett, p. 106. Robert also mentions (p. 126) the "towers" of Evreux, Alençon, and Coutances as among those constructed by Henry I. [965] "About the Tower," as the chronicle expresses it. [966] "Henricus Rex circa turrem Rothomagi ... murum altum et latum cum propugnaculis ædificat, et ædificia ad mansionem regiam congrua infra eundem murum parat" (_Robert of Torigny_, ed. Howlett, p. 106). [967] I can make nothing of Mr. Clark's chronology. In his description of the Tower he first tells us that "all save the keep [_i.e._ the White Tower] is later, and most of it considerably later than the eleventh century" (_M. M. A._, ii. 205), and then that "the Tower of the close of the reign of Rufus" (i.e. _before the end of_ "the eleventh century") ... was probably composed of the White Tower with a palace ward upon its south-east side, and a wall, probably that we now see, and certainly along its general course, including what is now known as the inner ward" (_ibid._, ii. 253). Again, as to the Wakefield Tower, which "deserves very close attention, its lower story being next to the keep in antiquity" (_ibid._, ii. 220), Mr. Clark tells us that Gundulf (who died in 1108) was the founder "perhaps of the Wakefield Tower" (_ibid._, ii. 252); nay, that "Devereux Tower ... may be as old as Wakefield, and therefore in substance _the work of Rufus_" (_ibid._, ii. 253); and yet we learn of this same basement, that "the basement of Wakefield Tower is probably late Norman, perhaps of the reign of Stephen or Henry II., although this is no doubt early for masonry so finely jointed" (_ibid._, ii. 224). In other words, a structure which was "the work of Rufus," _i.e._ of 1087-1100, can only be attributed, at the very earliest, to the days of "Stephen or Henry II.," _i.e._ to 1135-1189.