Held at:





Internet – Archaeology Data Services [University of York]


Earthwork Castles of Gwent and Ergyng AD 1050-1250: Walterstone Castle

Place name:



1050 - 1250


This summary concerning Walterstone Castle is an extract from a Doctoral thesis by Dr N Phillips, University of Sheffield (2005), entitled ‘Earthwork Castles of Gwent and Ergyng, AD 1050-1250’. This thesis is concerned with the earthwork and timber castles built in the southern March of Wales, addresses the presence of the castles and discusses their roles as weapons of conquest and structures of administrative control.


WALTERSTONE - Grid: SO 33932 24999




The castle stands at the eastern end of the village of Walterstone which can be reached from the A465 Abergavenny to Hereford road. The road to Longtown branches off the A465 to the north of Pandy, the first crossroads going right, and then first left at Alltyrynys. The site of the mound is to the north-east of the village close to The Carpenter’s Arms and 100m west of the church.


The motte, and possible bailey at Walterstone occupies a broad expanse of land on top of a ridge to the east of the Olchon Valley. Unusually, the motte has been built beneath a shallow rise. Using Ewyas Harold and Pont Hendre as examples with mottes placed beneath slopes and ridges, the bailey occupies the ground behind the motte, although it has to be said the motivation is not understood. Two of the closest castles to Walterstone are Ewyas Harold and Pont Hendre, both regarded by this study as early.

The first photograph, (Vol. 2. plate 260) shows the motte in the distance, covered with trees and just in front of it, before the mud, is a faint dark line. The line represents a southern ditch denoted ‘w’ on the Remfry plan (Vol. 2. figure 117). The motte itself is quite sizable with a fairly large flat top. A ditch surrounds the motte completely and is for the most part wet (Vol. 2. plates 261 and 262). The last photograph shows the spread of stone that litters the motte, although it is difficult to say in what capacity it was used; motte fill or building material (Vol. 2. plate 263). There is a noticeable raised rim at parts of the motte top which could be the remains of a shell keep or just an old palisade setting. Remfry, who also noted the rim, drew a possible similarity to the shell keep at Llancillo, which is suspected to be contemporary both in builder and period (1998. 14).

Topographic survey: (Survey 52)

The motte takes the form of a high sided conical motte 8.76m, with a top surface diameter of 574.189m² and containing an estimated 5539.545m³ of fill. The actual raised portion, however, can be reduced to 1391.525m³ as it is surrounded by a ditch 2.89m deep which has been cut into the natural surface. The highest point above natural is to the east at 8.06m where the ditch is only 0.9m deep.

Surprisingly, calculations on ditch volume and motte volume revealed that some 1103m³ of cut from the ditch is not accounted for on the motte which may either suggest that the motte was bigger, which is unlikely, or that outer earthworks such as a rampart or raised bailey are missing.

The ditch to the east of the motte is only 0.9m deep whereas elsewhere it is at least 1.18m rising to 2.89m.

Strategic position:

The position of the motte is fairly weak, lying quite close to a ridge that overlooks it. It is probable that there must have been earthworks surrounding the site otherwise there would have been no defence and whereas this has been seen to be reasonable at a watch-tower, the size of Walterstone motte would suggest that this was a more substantial castle. The farm land around the site is heavily ploughed which would have removed any low-lying earthworks. Another possibility for defence may be the waterlogged nature of the site, a strategy discussed for Mount Ballan, Llancillo and Newton Tump, (see above).

Documentary evidence

Modern reference: HWCM5590

The second map shows the layout of the site with its outer ramparts as it was 1900 (Vol. 2. figure 118). Since then extensive farming activities and modern building works have served to reduce the earthwork perimeter. An example of this is that the RCHME inventory records that the bailey to the east and south-east is mostly bounded by a scarp (1931. 247), but there is no evidence of this today. The Remfry map, (see above), shows a different, more detailed interpretation of the site. The field survey for this study failed to identify baileys ‘n’ and ‘b’ shown on the second map but ‘w’ is still identifiable and referred to above. In his account of the castle, Shoesmith mentions that stone defences were unlikely, the site having been abandoned by 1137 (1996. 220). Unfortunately, there is no documentary evidence for Walterstone and no source cited by Shoesmith for his given date. As to the lack of stonework the last photograph, (see above) shows this not to be the case.

No early documentary evidence exists for this castle but much has been made of the name Walterstone drawing the obvious association with Walter de Lacy (Marshall 1938. 149). If this association is correct then the motte would be dated either pre-1085 or pre-1241 when the two respective Walter de Lacys died. Remfry argues the case for the former suggesting the likely transformation from ‘Walter’s tun’ to Walterstone (Remfry 1998.14). If this is the case, then the castle would represent one of the earliest castles of the de Lacy holdings possibly as early as 1067.

Additional references:

VCH 1908. 246.

Hogg and King 1963. 97.

King 1983. 212.

Interpretation: Motte and possible bailey (Early)

Interpretation based on actual remains, survey and location would suggest that the motte and possible bailey at Walterstone are early constructions. The shape and size of the motte along with the surrounding ditch emphasize the function of the private defensive strong point


The material is copyright by the author, and is reproduced here from the Archaeology Data Service website of the University of York for research purposes under their terms of use


Top - Back

Ref: rs_wal_0096