A POSSIBLE OUTER BAILEY DITCH TO MARLBOROUGH CASTLE 



103 



describing the corner of a rectangular plan. 

 Containing a single deposit of coarse flint rubble 

 within a sticky chalky matrix, this feature is 

 interpreted as a 'robbed-out' wall foundation. 



The remaining features consisted of single 

 postholes and smaller linear gullies. They formed 

 no readily discernible pattern and, though planned 

 and photographed, were not further investigated. 



Artefactual data 



Mediaeval pottery was recovered from four deposits, 

 1001-1004, within the ditch and pit groups; 

 together with a bone object (SF3), a knife handle, 

 an iron buckle (SF2) and a piece of iron strapping 

 (SF3) from the overburden layers. The latter, 

 unstratified objects, are not further described but 

 have been curated appropriately. 



Pottery 



A total of 85 sherds (2094g) from four separate 

 contexts was examined. Apart from a single residual 

 Romano-British sherd, and one post-medieval sherd 

 (both from context 1000), all the pottery is medieval. 

 Sherds from context 1000 are abraded, but the 

 remainder are, on the whole, markedly unabraded, 

 with crisp fractures and unworn surfaces. 



The pottery has been quantified by fabric type 

 within each context. All fabric types identified are 

 known local and regional wares, and form an 



assemblage composed of locally produced 

 coarsewares supplemented by finewares from 

 different, more distant sources. The medieval 

 assemblage has a potential date range of 12th to 

 early 14th century, although the majority would fit 

 with a more restricted date range of mid to late 

 1 3th century. 



Three coarseware types were identified, two of 

 which are likely to derive from a single source, 

 almost certainly local. These are the 'Kennet Valley' 

 wares, flint-tempered and chalk-tempered, first 

 defined at Newbury in west Berkshire (Vince 1997, 

 fabric groups A and B respectively) and 

 subsequently renamed following the discovery of a 

 production site outside Newbury (Mepham 2000). 

 These two wares form part of a 'ware tradition' 

 found along the Kennet Valley from Newbury to 

 Devizes, and were probably produced at a number 

 of different centres within this wide distribution 

 area; one potential source has been identified on 

 the basis of place-name evidence near Marlborough, 

 in Savernake Forest (Vince 1997, 65). Only two 

 rims are present here - one jar rim and one bowl 

 rim. The tripod foot from context 1001 is not from 

 a tripod pitcher (such vessels are not known in 

 Kennet Valley wares) but probably from a cauldron 

 (e.g. Mepham 2000, fig. 17, 52). Kennet Valley 

 wares were long-lived (12th to 14th century), but 

 the vessel forms seen here would fit with the date 

 range of mid to late 1 3th century which is suggested 

 by the finewares. 



Fig. 3. Northwest facing section of Ditch 1037 showing Ditches 1022 and 1023 plus overburden 



