ST PAUL’S CATHEDRAL: EVIDENCE FOR A SAXON DITCH AND MEDIEVAL ACTIVITY FROM EXCAVATIONS AT PATERNOSTER SQUARE 2000—2001 Sadie Watson With contributions by Lyn Blackmore, Nigel Cameron, Jane Corcoran, Anne Davis, Geoff Egan, Jacqueline Pearce, Natasha Powers, Kevin Reilly, Mary Ruddy, Robert Scaife and John Shepherd SUMMARY from Paddington to Cheapside and provided drinking water between the 13th and 16th centuries. Excavations carried out during the redevelopment of Paternoster Square in 2000 and 2001 revealed multi- INTRODUCTION period archaeological remains. This article outlines discoveries dating from the Saxon and medieval During 2000 and 2001 Museum of London periods. A large man-made watercourse or ditch run- Archaeology (MOLA) (prior to 2009 known as ning southwards to the west of St Paul’s Cathedral the Museum of London Archaeology Service) was excavated and radiocarbon dates confirm that it undertook several excavations during the was in use during the 7th—8th centuries and perhaps redevelopment of Paternoster Square; a large was associated with the founding of St Paul’s in AD area of land directly to the north of St Paul’s 604, while the associated pollen showed evidence of Cathedral (Fig 1; Table 1). The works were cereal cultivation and arable or waste land. This ditch commissioned by Stanhope and Mitsubishi was backfilled during the early 12th century as part Estate (Sites B and D), Standard Life Invest- of the Bishop of London’s rebuilding scheme. Burials ments (Site A), and HSBC (Site C). The four of 8th- to 10th-century date were found to the north main excavations took place during 2000 and of the cathedral. The earliest Saxon ceramics recovered 2001, although they were all preceded by from the rubbish pits found scattered across site were extensive phases of evaluation, watching-brief residual sherds of Middle Saxon Ipswich ware. Finds and geotechnical monitoring (full details are from the various early and late medieval rubbish pits in Watson 2006, 6). All four excavation areas provided evidence of diet and daily life. One pit, which yielded extensive archaeological remains dat- was backfilled in c.1500, contained an assemblage ing from the Roman period through to the of high status glassware, probably associated with present day. The Roman and post-medieval the bishop’s palace. An isolated medieval masonry results have been published elsewhere (Watson footing may be part of the Vicars Choral complex. Part 2006; Watson & Pearce 2010). Pottery fabric of the City’s medieval water-supply system was also code expansions are available from http:// discovered; it consisted of a section of lead pipe over 7m www.museumoflondon.org.uk/Collections- long, representing a portion of the conduit which ran Research/LAARC/ (accessed 2012). 33 34 Sadie Watson Table 1. Details of excavations undertaken on the Paternoster Square Masterplan site Site prefix Site code Site name National Grid Reference A SLY00 Juxon House 531950 181200 B NGT00 Newgate Triangle 531985 181250 C PNS01 Paternoster Row 532065 181235 D PSU99 Stonemason’s Yard Ramp 531920 181210 PREVIOUS EVIDENCE FOR THE show a slope within the gravels from north- ‘WESTERN STREAM’ east to south-west with a large linear hollow in the London Clay running in the same Paternoster Square is located at the high- direction, confirming that natural drainage est point in the City of London. The mod- on the site is in this south-westerly direction ern topography of the area is fairly level (Fig 2). This depression in the London Clay at c.17.00m OD, with the ground falling dates from the period when the Thames away steeply to the River Fleet at the was carving out a new floodplain, across west and the Thames to the south. The which the Taplow Gravels were subsequently underlying geology is Eocene London Clay, deposited, around 128,000—280,000 years overlain by Pleistocene terrace gravels. ago. This would have remained as a shallow Geoarchaeological contour plots of the site depression in the surface, even after infilling Fig 1. Site locations (scale 1:4000) St Paul’s Cathedral: A Saxon Ditch and Medieval Activity from Paternoster Square 2000-2001 35 Fig 2. Contour plot of surface of terrace gravels (scale 1:2500) by the gravels and the later brickearth had backfilling, possibly during the 11th—12th masked its alignment, reflecting the point centuries (Bentley 1987, 331). It was thought at which groundwater emerged from the that this channel continued to the south- interface of the Hackney and Taplow Gravels, west, to join the Thames south of modern fanning out towards the Fleet to the west Queen Victoria Street (Tyler 2000, 23). (Watson 2006, 9—10). The channel was likely to have still been A major topographical influence on the visible during the 13th century, as it formed landuse of the area was a large stream channel the boundary between the Farringdon running in a southerly direction to the west Within and Castle Baynard Wards. Bentley of the modern churchyard, known as the concluded by suggesting that the channel ‘western stream’. Initially observed by Peter was a ‘vast man-made feature’ (1987, 334). Marsden during the 1961—2 development of Recent work on other sites in the vicinity the site, profiles through the channel were (Askew 2007; Lyon 2007; Pitt 2006) and drawn and Roman pottery recovered from on Paternoster Square (Watson 2006, 10, its basal fills (Bentley 1987, 330; Merrifield figs 9 & 12) has concluded that this feature 1965, 192). Marsden concluded it was a was not naturally formed; although it may natural stream flowing in a deep valley, have partly utilised the natural depression utilised by the Romans to provide water for formed by the linear hollow in the London a pottery industry in the area (kilns were Clay (Fig 2), the watercourse cannot have found on Paternoster Square). Later work been part of this particular channel, as it ran in the area by Brian Pye revealed that the perpendicular to the depression (Watson channel had been recut several times and 2006, 12). In conclusion, this stretch of the the upper fills were the result of deliberate ‘western stream’ is now interpreted as a 36 Sadie Watson large man-made feature, which could more band overlying an orange, iron-concreted aptly be described as a ditch than a stream horizon at the top of the gravel. Chalk channel. The northern portion of this clasts were found at the base of the fine- feature might have served as a water supply grained channel fills at the western end of for the cathedral area, while the southern the transect (AH 1). These were not well- end of it certainly had both a drainage and a rounded, as would be expected if they had rubbish disposal function. been transported by water or had lain for any Other stream channels which are probably time on the bed of a channel. Nor were they of natural origin were recorded on various weathered, as would be expected if the base sites to the north running in a westerly of the feature had been ‘dry’ and a soil or direction towards the V-shaped Fleet valley; plant growth developed across it. They are these channels were recut during the Roman likely to have been discarded or have fallen period (Pitt 2006, 48; Askew 2007, 259). The into the channel feature from the adjacent date for the initial cutting of the watercourse bank or have been deposited during its on Site A had remained unclear, although construction. Either way they are likely to it was previously thought that this channel date the base of the sequence to the historic was cut to divert the course of streams to the period, probably late Roman onwards. This north and was backfilled during the Roman was confirmed by the pollen assemblages period, possibly after the construction of from the overlying fills, which are clearly of the city wall in c.AD 200 (Shepherd 1988, historic age and are typical of those found 19). in urban environments where admixtures of pollen occur through extremely complex SAXON ACTIVITY (c.AD 600—900) taphonomy (Scaife 2002). There is a general absence of arboreal pollen from the channel Environmental reconstruction of a large fills and the presence of a single grain of watercourse to the west of St Paul’s walnut (Juglans regia), which is generally Jane Corcoran, with plant remains by Anne Davis, regarded as a Roman introduction into pollen analysis by Rob Scaife and diatom analysis Western Europe as a whole (Godwin 1975), by Nigel Cameron suggests a Roman or post-Roman date. The augerhole sequence suggests that The upper portion of a large truncated north— the lowest fills of the channel feature were south-aligned watercourse was excavated in deposited by natural processes. A sequence, plan on Site A (Fig 3). However, due to depth c.1m thick, of inter-bedded humic muds and restrictions, the lower portion of this large sands was found at the base of the feature in feature could only be investigated in two the western part of the transect; this is likely geoarchaeological auger transects (Figs 2— to represent low-energy water flow and the 3). The southernmost transect has been used build-up of a channel bar. A radiocarbon date to reconstruct a profile across the feature, of CAL AD 650—780 (Beta-192219) obtained which was formerly known as the ‘western from organic sediment within these deposits stream’ (see above) (Fig 4). indicates that they accumulated during either Geoarchaeological investigation has re- the 7th or 8th century (see Appendix 2 for vealed that this watercourse was about 20m details). The diatoms present in the channel wide and in profile had a flattened ‘U’-shape bar deposits were characteristic of shallow – the eastern edge presenting a concave water habitats with high flow rates, although slope, with an overall gradient of 45 degrees, they indicated that some ‘backwater’ areas but shallower at the top and steeper towards would have existed for algae and aquatic the base (Corcoran 2002).