[944] See Round’s Ancient Charters, p. 9, Pipe Roll Series, vol. x.
[945] Brut, 1113.
[946] The first mention of the castle of Llanstephan is in the Brut, 1147, if, as has been assumed above, the mention in 1136 refers to Stephen’s castle at Lampeter, as the Annales Cambriæ say.
[947] The motte of Conisburgh in Yorkshire is a very similar case known to the writer; it measures 280 × 150 feet. Such very large mottes could rarely be artificial, but were formed by entrenching and scarping a natural hill.
[948] Brut, 1256. See Arch. Camb., 1907, p. 214, for Col. Morgan’s remarks on this castle.
[949] The name Gueith tineuur is found in the Book of Llandaff, p. 78 (Life of St Dubricius), but it seems doubtful whether this should be taken to prove the existence of some “work” at Dinevor in the 6th century. See Wade-Evans, Welsh Mediæval Law, p. 337-8.
[950] Brut, 1145. “Cadell ap Griffith took the castle of Dinweiler, which had been erected by Earl Gilbert.”
[951] Gwentian Chronicle.
[952] The statement of Donovan (Excursions Through South Wales), that the castle stands on an artificial mount is quite incorrect.
[953] The Rolls edition of the Brut gives the corrupt reading Aber Cavwy for the castle of “Robert the Crook-handed,” but a variant MS. gives Aber Korram, and it is clear from the Gwentian Chronicle and Powell (p. 145) that Abercorran is meant.