It is clear from the _Life of Griffith ap Cynan_[848] that the Earl of Chester had conquered and incastellated Gwynedd before the accession of William Rufus. This valuable document unfortunately gives no dates, but it mentions in particular the castle at Aberlleinog,[849] one at Carnarvon, one at Bangor, and one in Merioneth. The motte at ABERLLEINOG, near Beaumaris, still exists, and the half-moon bailey is traceable, but the curious little round towers and revetting wall in masonry on the motte were probably built to carry guns at the time of the Civil War, when this castle was besieged by the Royalists. At CARNARVON the magnificent castle of Edward I. has displaced all former erections, yet some evidence for a motte-and-bailey plan may be found in the fact that the northern portion of the castle has evidently been once separated by a ditch from the southern, and is also much higher.[850] On the hills above BANGOR, Pennant thought he had discovered the remains of Earl Hugh's castle, but having carefully examined these walls, we are convinced that they never formed part of a castle at all, as they are much too thin; nor are there any vestiges of earthworks.[851] We are disposed to think that instead of at Bangor, the castle of Earl Hugh was at ABER, often spoken of as ABERMENAI in the _Chronicles_, and evidently the most important port on the Straits. At Aber there still remains a motte which must have belonged to an important castle, as it was afterwards one of the seats of Llywelyn ap Jorwerth, Prince of Gwynedd. The castle in Merioneth cannot be certainly identified. In one of the invasions of William Rufus, which both the _Anglo-Saxon Chronicle_ and the _Brut_ describe as so unsuccessful, we hear that he encamped at MUR CASTELL, a place undoubtedly the same as what is now called TOMEN-Y-MUR, a motte standing just inside a Roman camp, on the Roman road leading from Shropshire into Merioneth and Carnarvon. This motte is surrounded by a ditch; there are traces of the usual earthen rampart round the top, now mutilated by landslips.[852] We may, with great probability, assume that this motte was thrown up by William Rufus, and that the Roman camp served as a bailey for his invading host. Whether it was garrisoned for the Normans we cannot say, but it evidently formed an important post on a route often followed by their invading armies, as Henry I. is said to have encamped there twice.[853] It is one of the few mottes which stand in a wild and mountainous situation, and its purpose no doubt was purely military.[854] * * * * *