Micropalaeontology reveals the source of building materials for a defensive earthwork (English Civil War?) at Wallingford Castle, Oxfordshire

Microfossils recovered from sediment used to construct a putative English Civil War defensive bastion at Wallingford Castle, south Oxfordshire, provide a biostratigraphical age of Cretaceous (earliest Cenomanian) basal M. mantelli Biozone. The rock used in the buttress – which may have housed a gun emplacement – can thus be tracked to the Glauconitic Marl Member, base of the West Melbury Marly Chalk Formation. A supply of this rock is available on the castle site or to the east of the River Thames near Crowmarsh Gifford. Microfossils provide a unique means to provenance construction materials used at the Wallingford site. While serendipity may have been the chief cause for use of the Glauconitic Marl, when compacted, it forms a strong, almost ‘road base’-like foundation that was clearly of use for constructing defensive works. Indeed, use of the Glauconitic Marl was widespread in the area for agricultural purposes and its properties may have been well-known locally.


INTRODUCTION
Microfossils have been recovered from a range of archaeological and historical remains including building materials (e.g. Wilkinson et al., 2008), art and ceramics (e.g. Perch-Nielsen, 1973;Quinn & Day, 2007). The microfossil technique has wide 'forensic' application as even very small samples of rock (tens of grams) can yield hundreds of fossil specimens. The technique can be used to provenance building materials from classical contexts through to the present (e.g. Wilkinson et al., 2008). Here we apply techniques of microfossil biostratigraphy to establish the age of materials used in the construction of an enigmatic (English Civil War?) defensive buttress at Wallingford Castle in south Oxfordshire, a feature excavated as part of a major archaeological programme at the town -the Wallingford Burh to Borough Research Project.
The town of Wallingford lies adjacent to the River Thames to the west of London. Although the town has its origins in Saxon times, the history of the area (see Airs et al., 1975) suggests that the ford, attested to in the town's name, has been in use since at least the Bronze Age. Finds of Roman coins are known from the general area, and an Anglo-Saxon cemetery dating from the fifth to sixth century AD lies to the southwest of the town. In the late ninth century, as part of a wider strategy to counter the Danish threat, ramparts were raised by the army of King Alfred the Great around the town. After the Norman invasion a castle was built in Wallingford's northeastern corner in the late eleventh century. Further banks and ditches were created around the castle, thereby obscuring the line of the nearby section of the earlier town rampart, perhaps during the Anarchy of 1139-1153 when Stephen and Matilda were at war, or when the castle was extended in the thirteenth century. The castle's last major role was during the First English Civil War, the last gasp of which came with the Royalist surrender on 27 July 1646, after a siege of 65 days by Parliamentary forces.
Castle Meadows now contain the extensive surviving earthworks of the once magnificent Castle, demolished after the Civil War. Figure 1 shows the location of the castle earthworks in relation to the wider town and recent archaeological excavations (the trench under discussion here is marked as '4' on the map). A resistivity and magnetometer survey in April 2008 covered the northern part of the castle ramparts and the adjacent low-lying alluvial-deposits bordering the west bank of the Thames (Fig. 2), all just outside the outer moat of the castle. Among other features, the survey revealed an area ('bastion') projecting northwards from the castle ramparts to the west of a larger such promontory popularly referred to as 'the Civil War bastion', although this latter feature is undated. It is the material from the newly identified bastion that we have analysed using micropalaeontological techniques to establish its provenance.

ARCHAEOLOGICAL CONTEXT OF THE BASTION
A trench was excavated across the platform of the bastion during the summer of 2008 (http://www2.le.ac.uk/projects/wall-ingford_dig_2008). Figure 2 shows the location of the trench in relation to the results obtained from resistivity survey. The partially excavated platform was surfaced with light, variably olive, grey, weakly cemented sandy silt, overlying a sandy gravel preparation layer. Surrounding the platform was a large artificial depression, now mostly filled in, and thought to be a defensive ditch (though only the near side of the ditch has been revealed through excavation). The top of the platform created a level floor with some sandy preparation below. The material collected for microfossil examination was from this platform surfacing, which was itself buried under a shallow layer of topsoil.
Apart from tree-root holes and rabbit burrows, no features were found cutting the surface of the platform. Box sections were cut through the platform along the northern and eastern sides of the trench, revealing it to be about 0.5 m in depth and clearly of artificial construction. Beneath both the platform and the base of the ditch was a uniform layer of dark brown sandy loam containing medieval pottery (Fig. 3). This probably represents 'made ground' which may have been formed by up-cast from the digging of the outer moat of the castle during the medieval period.
Finds overlying the layer sampled for microfossils were mixed, of seventeenth-to nineteenth-century date with a number of fragments of clay pipes from the earlier to middle part of this range. Finds of pottery from the layer underlying the platform were dated as medieval (twelfth-fourteenth century). If the platform is taken to have served a military function, which seems plausible, then that is assumed to relate to the Civil War siege already mentioned. Interpretation of the levelling layer is speculative; but a level defensive platform, perhaps a gunnery position, would make sense. Strategically this 'bastion' fits topographically with that to the east (Fig. 1), and with features of similar appearance detected in the 2009 geophysical survey to the west, towards the town's north gate, a focus of assault during the Civil War.
Sampling of the bastion construction material was for routine analysis to investigate the presence of any inclusions such as carbonized cereals, wood charcoal, molluscs, bone fragments, etc. A two-litre pilot sub-sample was washed-over and floated (Kenward et al., 1980) and the residue sifted and allowed to dry before examination under stereoscopic zoom magnification and fossils picked out. Much of the matrix failed to disaggregate, and prolonged soaking in water and then hydrogen peroxide was used.
Small quantities of bone, molluscs, a trace of wood charcoal, a few worm ova and insect parts, molluscs and seeds were recovered. Normally from this type of site uncharred decomposable organic remains such as seeds and insects would be dismissed as intrusive. Given the cementation, it seems quite possible that at least some of these finds were trapped at the time of deposition of the sediment. The seeds and small fruits were fragile, but identifications were achieved using published illustrations (e.g. Cappers et al., 2006) and the botanical reference The presence of seeds and fruits may indicate summer or autumn as the time of year when the sediment was laid down. Although hazel ripens in late August, it can be stored well beyond then for later consumption. The other species are ruderals and weeds so their occurrence should be natural.

BIOSTRATIGRAPHICAL AGE OF SEDIMENT FROM THE BASTION
Material from the platform was examined further for its microfossil-content in an attempt to provenance its source. The sediment sample from the bastion contained a rich microfossil assemblage of ostracods and foraminifera (Pl. 1; Appendix A). The hyaline and agglutinated benthic foraminifera indicate the sediment sample to be of Cretaceous, Cenomanian age (Fig. 4). Gavelinella baltica, G. intermedia, G. cenomanica and Hagenowina advena first evolved at the base of the Cenomanian and ranged throughout that stage. Hagenowina anglica became extinct in the T. costatus macrofaunal Subzone (basal M. rhotomagense macrofaunal Biozone) in the mid-Cenomanian. The stratigraphically restricted index species Flourensina intermedia is confined to foraminiferal Biozone BGS1 (equivalent to the N. carcitanense macrofaunal Subzone, basal M. mantelli Biozone). It was accompanied by a single fragment tentatively assigned to Bulbophragmium aequale cf. aequale, which supported this age determination.
A number of the ostracod species present (e.g. Schuleridea jonesiana, Cytherella ovata, Bairdoppilata pseudoseptentrionalis, Homocythere harrisiana, Neocythere vanveenae) are biostratigraphically long-ranging taxa of mid-Albian to mid-Cenomanian age. However, others have a more restricted temporal distribution and support the age assignment established by the foraminifera. Cythereis hirsuta and C. thoernensis are known to occur in the Albian in France, but in the UK are found only in the uppermost Albian (perinflatum macrofaunal Biozone) and were more common in the lower Cenomanian. Neocythere steghausi first appeared in the Albian, but its extinction in the basal part of the N. carcitanense Subzone of the M. mantelli macrofaunal Biozone was an important bioevent and useful in dating the sediment from Wallingford Castle.

PROVENANCE OF THE BASTION MATERIAL
Cenomanian age geological deposits in the neighbourhood of Wallingford comprise the West Melbury Chalk Formation, including the Glauconitic Marl Member at its base (Fig. 4). The Castle site is located on the Glauconitic Marl and materials used to construct the platform may have been quarried on site ormore likely perhaps -to the east of the river where the Glauconitic Marl and West Melbury Chalk form an extensive outcrop in the Crowmarsh Gifford area. However, the Glauconitic Marl is obscured beneath more recent superficial deposits (Summertown-Radley Sand and Gravel Member, Northmoor Sand and Gravel Member and the river alluvium) in the Wallingford area, so that quarrying would have been required. The only place shown on the geological map for this area where superficial deposits do not cover the Glauconitic Marl is to the north of Wallingford Castle (approximately National Grid Reference SU 608 905).
Although the precise function of the bastion remains uncertain, its stratigraphical context supports attribution to activity during the English Civil War in the mid-seventeenth century when the castle was besieged, captured and subsequently destroyed by Parliamentary forces. During this conflict the defenders of the castle would have needed a reliable and nearby source of sediment to construct durable defensive platforms, some of which may have been used as gun emplacements. The Royalist occupiers of Wallingford Castle were fortunate that outcrops of the Glauconitic Marl were available nearby and material could have been ferried to the construction site fairly easily. When compacted, the Glauconitic Marl forms a durable, almost 'road base'-like medium, making it eminently suitable for defensive works. However, the work of the Royalist builders was ultimately to no avail, as the castle fell to Parliamentary forces. Indeed, whatever military advantage was intended through construction of the bastion, there is no clear archaeological evidence that it actually played a part in any military action.
Subsequently, in the 2009 excavation season, more marly deposits used in construction were identified (http://wallingforddig.pbworks.com/): first extensively capping an embankment set back from the original site in Castle Meadows; and secondly on the School Playing Field site on the west side of the northsouth-running Castle Street, just north of the town ramparts, but here to a much greater thickness of at least 1.5 m, apparently used for levelling uneven ground. The extent of medieval and post-medieval usage of the Glauconitic Marl Member may therefore have been much more substantial than the original discovery suggests, and may be illuminated by further investigation.

CONCLUSIONS
Microfossils identify the geological provenance of building materials used in the construction of a putative Civil War bastion at Wallingford Castle, south Oxfordshire. A ready supply of this material, the Glauconitic Marl Member of the West Melbury Marly Chalk Formation, was used widely in the area for agriculture and its properties may have been wellknown locally. The microfossil technique used here has great potential for establishing the provenance of a range of building materials from classical to early modern contexts.

ACKNOWLEGEMENTS
Excavations were conducted in Wallingford, south Oxfordshire, in the summer of 2008 as part of the Wallingford Burh to Borough Research Project run from the Universities of Exeter, Leicester and Oxford supported by an Arts and Humanities Research Council grant. This followed and accompanied continuing survey in the town and immediate environs. The authors thank the Wallingford Burh to Borough Research Project for access to samples and help with excavation records. I.P.W. publishes with permission of the Director of the British Geological Survey (NERC).

Manuscript received 8 October 2009 Manuscript accepted 23 January 2010 APPENDIX A
The material used to construct the Wallingford Castle bastion contained fragments of echinoid spines, bivalve molluscs (including occasional Inoceramus) and very rare fish. Biostratigraphically significant taxa are mentioned in the text, but a full list of the calcareous microfossils recorded is given below.