Bronze Age enclosures and Saxo-Norman bakehouses: Excavations at Coldharbour Road, Gravesend

Bronze Age enclosures and

SAXO-Norman bakehouses: Excavations at Coldharbour Road, Gravesend

jake wilson

With contributions by

Luke Barber, Emily Johnson, Elke Raemen and Mariangela Vitolo

This paper presents the results of an archaeological excavation carried out by Archaeology South-East at Coldharbour Road, Gravesend, between September 2018 and January 2019.

The earliest phase of identified archaeological activity dated to the Late Bronze Age/Early Iron Age when the site was divided by two square livestock enclosures and an accompanying boundary ditch. There was evidence for cattle and sheep management; however, indications of direct human occupation such as buildings was limited.

After a considerable hiatus, an area of specialised activity, likely related to baking, was founded during the Saxo-Norman period within the footprint of the earlier Late Bronze Age/Early Iron Age enclosure. By the beginning of the 13th century, the bakehouses appear to have fallen into disuse and the now abandoned sunken feature buildings became the focus of pit digging, ostensibly for the disposal of occupation waste, but also possibly representing termination rites. This abandonment may suggest that activity related to any contemporary occupation in the vicinity of the excavation area moved away, possibly into nearby medieval settlements.

Archaeology South-East (ASE; UCL Institute of Archaeology) was commissioned by RPS Group to undertake archaeological mitigation on land at Coldharbour Road, Gravesend (centred on NGR 564170 172030; Fig. 1). The British Geological Survey showed the underlying geology as Thanet Sand overlying Upper Chalk. A thin, narrow area of head deposits was recorded stretching along Wrotham Road (BGS 2021). Much of the east and centre of the site was situated within a fold in the former chalk downland, forming a north-south dry valley the base of which was filled by head deposits. Site level fell from 34m aod at the southern end to around 32m aod at the northern. To the south of the site ground level rose, as it did to the west of the site (up to 47m aod) and to the east (up to 52m aod).

Historical and Archaeological Background

The Neolithic and Bronze Age are well represented within a 1km radius of the study area. A ‘family burial’, land boundaries and evidence for craft activity were all recorded at Northumberland Bottom (KHER Refs: MKE80425/ MKE42965; MKE42965) and Oxford Archaeology recorded beaker burials, contemporary earthwork enclosures and four-post structures during the A2 widening scheme (Allen et al. 2012). In 2009/10 ASE found evidence of Late Bronze Age/Early Iron Age activity in the area during the A2 activity park excavations, recording a metaled hollow way, pit clusters, fence-lines and post-built structures (KHER Ref: MKE90974). Strong evidence for a sustained Middle/Late Iron Age presence has also been found within the 1km radius of the site by both ASE at the A2 activity park and Oxford Archaeology during the A2 widening scheme (Allen et al. 2012; KHER Ref: MKE90974).

The Roman landscape within the study site is dominated by Watling Street, the Roman Road from London to Canterbury, now represented by the A2. Work completed by ASE at the A2 activity park uncovered a number of enclosures including an early Roman inhumation cemetery at the activity park which contained multiple burials and cremations (KHER Refs: MKE90978; MKE90979; MKE90981). High status burials were also recorded by Oxford Archaeology on the A2 widening scheme (KHER Ref: MKE4027).

The medieval period is sparsely represented within the 1km study area with the majority of evidence coming from the A2 widening scheme undertaken by Oxford Archaeology (Allen et al. 2012). The excavations produced evidence of sunken-featured buildings (SFBs), subrectangular pits, kilns and deneholes (KHER Ref: MKE40133). Further evidence of medieval activity was excavated in 1998 as part of the channel rail link project (HS1) by Museum of London Archaeology (MoLAS 2001). A large 12th-century enclosure was recorded along with pit clusters and a sunken-floored building (KHER Ref: MKE80456). The site is also within the vicinity of medieval parishes Shorne and Cobham, and close to the medieval town of Gravesend.

The Gravesend Tithe Map of 1841 and the Northfleet Tithe Map of 1838 recorded the site as a mixture of arable and hop fields. The presence of a ‘1174a Chalkpit Field’ is noted in the Gravesend Tithe at the extreme west of the site. A substantial private property Claphall is shown on the south of the site on the Northfleet Tithe.

The Ordnance Survey map of 1863-65 shows the study site as unchanged from the previous tithe maps continuing as agricultural land. Wrotham Road is clearly indicated as again is the substantial property Claphall. A small former chalk quarry is indicated on the western boundary of the site. The following Ordnance Survey map of 1895 showed the site largely as it was prior to the excavation including the demolition of Claphall and increased urbanisation of the surrounding areas.

results

The Late Bronze Age/Early Iron Age remains (Period 1)

Two enclosures, EN1 and EN2, probably relating to the management of cattle were identified (Fig. 2). EN1 was of particular interest, with evidence of internal divisions and a four-post structure (ST1; see below). Square-shaped Enclosure EN2 conjoined with EN1, matching similar enclosures found in the A2 widening scheme (Allen et al. 2012).

The function of both EN1 and EN2 appears to have been purely agricultural, focused on the holding of livestock, likely during movement back and forth between the River Thames and the North Downs. No direct occupational evidence relating to the Late Bronze Age or Early Iron Age was recovered on site, and it is likely that any settlement would have been located away from these animal enclosures, possibly at the top of the valley to the south. Artefactual remains dating from Period 1 were limited, without the quantities you would expect from an area favoured for human occupation. Indeed, the pottery assemblage was of small size and lacking many diagnostic features or large stratified groups. A similar picture was indicated by the environmental evidence with a distinct lack of burning, depositional waste, or any industry associated with domestic areas noted across the site. Both EN1 and EN2 incorporated sectioned off areas in their south-western, corners. These were interpreted as livestock pens (Fig. 2).

The four-post structure (ST1)

Four postholes in close proximity to each other were identified within EN1 (Fig. 3). Such four-post structures are common features of both the Late Bronze Age and Early Iron Age and have been found nearby at Site E at the A2 widening scheme (Allen et al. 2012). These features are generally interpreted as storage structures, possibly for grain or hay and would have been raised above ground to protect produce from rats or other pests.

Often interpreted as high, standing structures, the four-post construction within EN1 might have been used as an elevated fodder rick, related directly to the tending of livestock held within the enclosure. A number of similar, shallow four-post structures of varying sizes identified during the A2 widening scheme (Site E) bear a strong resemblance to the example uncovered by ASE at Coldharbour Road. These structures were also not associated with any nearby settlement activity (Allen et al. 2012, 137).

The boundary ditch (D1) and structured deposits (Emily Johnson)

Perhaps the most significant feature of Period 1 was the large, north-south boundary ditch identified in the west of the excavation area. This large ditch may have represented an early land division, with a prominent location at the top of a dry valley side, cut into the white chalk geology. The feature contained many fills, suggesting a slow decline and, in parts, may have been seasonally cleaned out and maintained (Fig. 4).

The alignment and size are of particular interest as it potentially continued northwards towards the River Thames. It is possible that this may have been part of a watershed boundary or field system where cattle were transported to water as part of a Late Bronze Age/Early Iron Age exploitation of the riverine landscape. Though the valley is currently dry and no indications of a former watercourse were found during the excavations, it must be remembered that chalk streams and winterbournes were more prevalent in the ancient Downland than is the case today.

One of the key areas defined by recent Late Bronze Age fieldwork in southern England was the Thames valley and its estuary approaches where land divisions have been found on the shores of the navigation route through the Wantsum Channel, and at the tributary river mouths at Mucking (Essex) and Gravesend (Yates 2007, 110). There are strong parallels between the ASE excavations at Coldharbour Road and those by Oxford Archaeology in the vicinity (Mudd 1994), both identifying large boundary ditches and accompanying square enclosures. This suggests that the east facing valley side where Coldharbour Road is located was part of a larger Late Bronze Age systematic approach to division of the landscape.

Land divisions such as this may have formed part of a ritual landscape as well as a functioning pastoral one and this is seen with the deposition of a cattle skull, cut within the base of the boundary ditch. Cattle skulls have also been found with human burials in the Late Bronze Age suggesting there is some significance to their deposition (Towers et al. 2010). Such offerings may have been special deposits that emphasise important points in the land. They provide clues to the complexity of a cultural landscape and the importance that herding bestowed upon the cattle owners. Such deposits help indicate that the Late Bronze Age landscape was not solely an impersonal expression of demographic and economic forces (Fokkens 1999, 41).

A second, possible structured deposit was encountered within EN1: a dog skeleton of ambiguous date. Given that the elements belonged to a single dog, and the bones show no evidence of butchery or burning, they probably relate to a dog carcass that was not subject to carcass processing or consumption after death. The partial nature of the skeleton and the lack of articulation could suggest casual disposal of animal carcasses in convenient open location, yet a deliberate placement of this animal followed by taphonomic disturbance cannot be ruled out. The use of farm or ‘working’ animals in a pastoral setting is relatively overlooked in the Late Bronze Age and evidence of animal burial in agricultural enclosures is rare.

The Saxo-Norman remains (Period 2)

Re-use of prehistoric enclosures?

The edge of a Saxo-Norman settlement was identified on site and almost entirely contained within the footprint of Late Bronze Age enclosure EN1 (Fig. 5). The Saxo-Norman remains included two smaller enclosures (EN3 and EN4), three 11th-/early 13th-century SFBs (identified as ST2, ST3 and ST4), at least one of which may have been subject to a ‘ritualistic’ termination procedure, and a set of amorphous pits that were likely related to the deposition of waste produced by use of the SFBs.

The Period 2 pottery assemblage again proved to be small and of types well known from elsewhere in Kent. Generally, the assemblage lacked large groups or a significant quantity of feature sherds. It is probable that the whole assemblage can be placed between c.1125 and 1225.

That the SFBs and associated enclosures were within the footprint of the prehistoric enclosure may simply be coincidental, but it is also possible that the Late Bronze Age ditched enclosure EN1 was still visible and present in the landscape as an earthwork during the medieval period. It was unlikely that the prehistoric ditches were open, due to the lack of archaeological finds from the Saxo-Norman period within the fills and this lack of evidence implies that only the banks of the ditches were visible, the features themselves having been filled. Similar medieval re-use of prehistoric earthworks has been found elsewhere in the south-east, such as at Penlands Farm, Haywards Heath, on the edge of the Sussex High Weald (Margetts et al. in prep).

The buildings’ structure and function ST2/ST3/ST4

The use of the SFBs as bakehouses or areas of similar production (kitchen/crop drying/brewing) reinforces the north Kentish tradition of SFB use during the Saxo-Norman period. This class of buildings has now been studied in detail and it was concluded that such kitchens/bakehouses comprise a completely different building tradition to the Anglo-Saxon Grubenhäuser (Rady and Holman 2019). The three encountered on site and three more found along the route of the A2 near Gravesend represent the most westerly finds of these building in all of Kent for this period (Schuster and Stevens 2009).

SFBs ST2 and ST3 displayed the same morphology and structural design while the third (ST4) was so heavily truncated by pitting after its termination that any interior structure could no longer be ascertained. ST2 and ST3 were both east-west aligned and contained a single ‘keyhole’ style oven base and several postholes all located within the rectangular cut of the building itself (Figs 6 and 7). A single large posthole was also identified in the centre of both buildings. C14 dating on cereal grain (wheat) recovered from ST2 posthole [80] (fill deposit [81]) shows a calibrated date range of ad 1025-1205 at 95% confidence (940 ± 30 bp, Beta-583563) with similar results on the barley grain also sampled from context [81] (cal. ad 995-1160, 980 ± 30 bp, Beta-583562, 95% probability). ST3 also gave a consistent date range from wheat grain recovered from fill [343] at cal. ad 990-1155 (1000 ± 30 bp, Beta-583567, 95% probability) and barley grain from the same deposit (cal. ad 990-1150, 1010 ± 30 bp, Beta-583566, 95% probability). These dates can place the cereal grain within ST2 and ST3 into the Saxo-Norman period and suggest sustained crop processing within the area.

Evidence of heat exposure or burning was present in all three SFBs on site and is an indication of in-situ burning from an oven base including possible rake out from the ovens, spread out over the floor surface. The spread of burnt material is consistent with other SFBs found in the local area from nearby excavations on the A2 widening scheme (Allen et al. 2012) and also suggest that these building had no raised floor level or platform, instead consisting of a simple beaten earth floor (a trait observed in ST2 and ST3).

Fragments of German lava quern were encountered in all four quadrants of ST2 and is a stone often found in the area and fairly common for the period (Barber below). The discovery of the stone quern within the building might lend credence to the idea that these structures were related to the processing of cereal crops.

During the early medieval period SFBs were subject to ‘termination’ rituals (Hamerow 2006) that would involve the deliberate destruction of the building and the complete removal of any interior structure by frequent and overlapping pitting. ST3 contained a small amount of pitting in its south-eastern corner where a mix of pottery and waste materials including hammerscale was present. The hammerscale possibly indicates small scale smithing on or near to the site. ST4 by contrast was subject to intensive later pitting, kept entirely within the original structural cut. The termination ritual removed everything internal except for a single area of heat exposed ground (similar to that found in ST2 and ST3; Fig. 8). A single, inverted cattle skull (RF <6>) was recovered from the base of a central pit within ST4 and was likely part of a ritual deposit (Fig. 9). Cattle skulls are sometimes deposited in SFBs in Anglo-Saxon England, including sites such as Pennyland, West Stow and Yarnton (ibid; Fig. 8).

Evidence for settlement in the immediate area of Coldharbour Road is based entirely around the SFBs and their reported association with the outskirts of settlements. The practice of placing these kitchens or bakehouses on the edges of enclosures or settlements was not uncommon (Allen et al. 2012, 576) and may stem from the need to remove large and hot ovens away from the main dwelling areas into a space of relative isolation and safety.

The two Saxo-Norman enclosures EN3 and EN4 appear to directly bound the three SFBs and further reinforce the idea that there was a distinct separation between bakehouses and the parent settlement. Oxford’s A2 widening excavations show that there was other Saxo-Norman activity nearby and as direct occupational evidence was not encountered on site it is reasonable to suggest that any settlement would have been located in the vicinity perhaps just westwards of the SFBs and at the very top of the valley. Other areas of medieval activity have been identified further west of the site, such as the medieval settlement at Wingfield Bank (Hardy and Bell 2001) and in the wider landscape the site lies close to several parish boundaries including those of Shorne and Cobham.

Postholes and pits within the enclosures, especially EN4, are all likely to be associated with the bakehouses. Large, amorphous pits held waste charcoal from the ovens and two sets of postholes that were semi-circular in plan (G25/G26) between ST2 and ST3 might have provided a wind break or a sheltered area of operation immediately outside of the buildings. These posthole structures may have been erected to provide seasonal shelter to the workers within the buildings. Conditions during the excavation were exceedingly windy in the winter, with strong gusts blowing northwards, it is possible the bakehouses also experienced such winds and erected windbreaks to prevent damage to their ovens.

specialist reports

The Registered Finds by Elke Raemen and Luke Barber

A small assemblage of 14 Registered Finds was recovered during the excavations. The earliest were recovered from Period 1 features, however, the majority date to Period 2.

Included are two pieces of struck flint, details of which have been deposited for archive. Unworked cattle skull RF <6> is discussed with other animal bone (Johnson below). The remainder has been discussed below by period.

Period 1

A ceramic weight (RF <9>, cat. no. 1) was recovered from ditch [329] (fill [330]). It is incomplete but has a rounded base and was probably biconvex, with a possible scar to one surface possibly from a means of suspension. Most clay weights of the period are perforated and would have functioned as loom weights. A wider and taller Early Iron Age example was recovered at Danebury (Poole 1984, 403, fig 7.48, no 7.69). There is limited but definite evidence of weights and weighing apparatus from the Late Bronze Age onwards, although the former tend to be in copper-alloy or stone. Clay weights, however, are difficult to classify, as many would have been made on an ad hoc basis and could have had a range of uses. It is unclear whether this weight had any means of suspension, but if not, it could still have been used as a counterbalance or to weigh something down.

An iron spearhead (RF <1>, cat. no. 2) was recovered from ditch [79] (fill [76]). At 67mm long, it measures only slightly above Manning’s smallest spearheads (Hod Hill Group 1), which range between 45-65mm (Manning 1985, 163). Although they are occasionally classed as arrowheads (e.g. Allason-Jones 1988, 193), the current consensus is that they are lightweight spearheads. Manning’s typology is almost entirely based on examples from Hod Hill where they date to the mid 1st century, however, they have been found in later contexts too (e.g. Scott 1980, 339-41).

1. RF <9> Ceramic weight (Fig. 10)

[330] (ditch [329]), EN1; Period 1

Incomplete. Biconvex weight. Most of one surface missing. Irregular central hollow may represent scar (e.g. suspension), or could be accidental. Orange fabric with common crushed calcinated flint to 2mm. Diam. 71mm, H38mm+.

2. RF <1> Iron spearhead (Fig. 11)

[76] (ditch [79]), EN1; Period 1

Incomplete. L67mm (blade L38mm, max width 16mm, internal socket diam. 8mm). Leaf-shaped with rounded shoulders; slightly asymmetrical blade. Triangular section with one flat side. Socket chipped.

Period 2

A total of nine Registered Finds were recovered from Period 2 features, including three iron objects. RF <8> (pit [243], (fill [244]) comprises a horseshoe fragment without caulkin (a blunt projection on a horseshoe that is often forged, welded or brazed onto the shoe). X-ray shows one surviving rectangular nail; however, too little survives to establish its type. An undiagnostic iron tool fragment (RF <7>) was found in ST2 ([80] (fill [81]). Finally, RF <3> consists of a small iron eyed spike with clenched tip (L56mm; compare Goodall 2011, fig. 9.11, H228, 189 for a larger example). It was recovered from oven [341] (fill [339]).

Querns by Luke Barber

A total of 26 fragments (RF <5>, <10>, <11>, -<13>) weighing 2,680g were recovered from five different contexts. Most pieces are amorphous, but gully [24] (EN4) produced a 37mm thick stone (RF <10>), oven [371] a 33mm thick stone (RF <5>) and ST2 [80] 27 and 30mm thick stones (RF <12a> and <12b>) – although none are discernible as upper or lower stones, and no diameters can be calculated. It is quite notable that German lava completely dominates the early medieval querns although the sample is admittedly small. The dominance may be in part due to a direct access to incoming goods travelling up the Thames to London.

Animal Bone by Emily Johnson

An assemblage of 739 animal bones weighing approximately 1,188g in total was analysed from the excavation. Material derived from both hand-collected and bulk-sampled contexts. The preservation of the assemblage was generally poor (Table 1).

Results

Both wild and domestic animals were present in the animal bone assemblage from Coldharbour Road. 142 bones were identifiable to taxa, and 220 to taxa size (Table 2). The assemblage is discussed below by period and landuse.

Period 1: Late Bronze Age/Early Iron Age 1150-600 bc

A total of 326 specimens were recovered from contexts dated to Period 1. Taphonomic modifications related to poor preservation were evident on bones from almost all contexts. Root etching affected 180 specimens, erosive action destroyed bone surfaces of 63 specimens, and 68 specimens showed evidence of excavation/post-excavation fracture. This contributed to the fragmentation and the relative paucity of bones identifiable to taxa.

The boundary ditch D1 yielded poorly preserved zooarchaeological material from three contexts (NISP=143). Cattle and ovicaprids were the only identifiable taxa represented. The most significant deposit was that of a cattle skull context [38], the fill of pit [37], which was cut into the base of the boundary ditch. Both frontal bones were present and over a hundred highly fragmented large mammal cranial specimens. This could be indicative of a ‘special’ deposit – cattle skulls associated with ritual deposition are known earlier in the Bronze Age, for example at Irthlingborough barrow where they were buried with human remains (Towers et al. 2010). However, given the difference in context and the fragmentation of the cattle skull from Gravesend, understanding its depositional nature is extremely difficult.

In addition to these specimens six animal bone fragments were recovered from dump context [67] <8>, which although undated was associated with the recut of this feature. Specimens included an ovicaprid mandibular molar and indeterminate fragments.

Ten fragments of animal bone were recovered from the large internal east-west gully (G13), from two contexts (NISP=10). Context [145] contained four refitting fragments of cattle radius diaphysis, and context [226] contained four ovicaprid teeth.

The largest assemblage of faunal material in this period derived from contexts related to the large sub-square enclosure EN1 (NISP=173). Taxa present included cattle, ovicaprids, and dog, which was well represented in ditch fill [214] of intervention [213]. Domestic dog bones in this context comprised tooth fragments and some postcranial material, including an ulna diaphysis, a fully fused radius and a fused distal tibia. These specimens show no evidence that dogs were eaten as butchery and burning were not observed. However, the apparent lack of articulation could suggest the casual disposal of these animals in convenient open locations when they died.

Period 2: Early Medieval 1050-1200

The animal bone assemblage from this period was slightly larger than that from the Late Bronze Age. As with the Bronze Age period assemblage, taphonomic surface modifications were common. Canid gnawing was identified on 6 specimens. Erosion (n=90), root etching (n=42) and recent breaks (n=55) also frequently affected specimens in this period assemblage.

Contexts related to a group of pits within this EN3 produced a very small assemblage (n=15) of cattle and ovicaprid postcranial material, and indeterminate bone specimens.

A total of 53 faunal specimens were recovered from contexts associated with an area of pitting within EN4 (G26; n=3) and the central enclosure partition (G19). Identifiable taxa included ovicaprids, cattle, and one tibia identified as mouse/vole. Just one poorly preserved medium mammal long bone fragment was recovered from sampling of the fill [81] of context [80] ST2.

An assemblage of 48 animal bone fragments was recovered from two contexts associated with ST3 – tree-throw [455] fill [456] and the fill of the SFB itself. The only identifiable specimens were a cattle distal femur and a pig tusk (indicating a male animal); the remainder were partially identifiable or wholly indeterminate.

Contexts associated with ST4 yielded approximately 291 animal bone specimens. Cattle were represented largely by fragments of cranium and dentition, and one cattle mandible from context [339] of oven [341] (Fig. 8) was aged as Adult 40 months – 6.5 years, based on dental attrition (Grant 1982; Halstead 1985; Jones and Sadler 2012). Some postcranial elements were present. Pigs were represented by an unfused first phalanx, a mandibular molar and a fragment of atlas. Ovicaprid bones included fragments of femur and indeterminate teeth.

Bird bones present included domestic fowl and goose. Wild animals were also present, including rabbit phalanges, one of which was carbonised/ approaching calcined, a humerus of a rat or water vole, which was carbonised, and three mouse (sp.) humeri. At least some of these wild taxa may represent non-anthropogenic inclusions, whether through archaeological pit falls or bioturbation.

Context [415], the fill of a pit which intercut other features of this SFB, contained a cattle cranium (RF <6>), found upside-down at the base of the pit (Figs 9 and Fig. 12). The cranium was fairly well-preserved, in several large and many small fragments. Those identifiable included the left and right frontal, parietal, temporal (including the petrous), lachrymal, zygomatic, palatine, maxillary, sphenoid and occipital bones, and the pars basilaris. Approximately 100 other fragments of cranium were not identified to specific cranial element, but likely all derived from the same cranium. Maxillary teeth present included the left molars, all of which were in wear, indicating an animal older than 30 months (Silver 1969; Grant 1982; Halstead 1985). The grand palatine foramen was bifurcated on the left palatine, but single on the right, likely a non-metric trait.

Certain parts of the cranium were notably absent, however, including the horn cores, nasal bones, premaxilla, and the occipital condyles. No clear evidence of these elements was identified in the associated cranial fragments. For the horn cores, it is possible that a hornless breed or polled animal is represented, yet there was no evidence of the part of the frontal bone where the horns would attach. Therefore, despite the completeness indicated by the good preservation, especially evident in the presence and state of the complex and fragile nasal turbinates, the cranium was not complete when deposited.

Cut marks were identified on the cattle cranium, on the dorsal aspect of the right frontal bone and the lateral aspect of the left zygomatic. These cuts could be associated with skinning the animal.

When interpreting this cattle cranium it is important to consider both the cranium and the deposit environment, building a deposit biography to attempt to understand the processes contributing to its placement (Morris and Jervis 2011, 74). The cranium itself was from an animal that had likely reached full size and showed evidence of ‘normal’ carcass processing in the form of butchery marks that suggest skinning. It was incomplete when placed, which may suggest certain elements were removed for further processing (particularly the horns). The processed cranium, likely skinned, and without mandibles so possibly also defleshed, must have been selected for deposition in this pit – regardless of whether that selection meant anything more than refuse deposition.

Turning to the deposition context, perhaps the most ‘special’ aspect of the deposition aside from selection of the cranium itself, is the upside-down position, resting on the very base of this feature dug into the disused SFB. When discussing Anglo-Saxon (4th-7th centuries) ‘special deposits’ in British settlements, Hamerow (2012, 133) suggests that these are typically found on or just above the base of their features.

However, it has been argued that skull deposits should be thought of as no more special than any other redeposited midden material and indeed that these so-called special cranial deposits are often found with fragmented, redeposited material (Morris and Jervis 2011, 72). Indeed, some other faunal specimens were also recovered from this context that may indicate redeposited waste. This included indeterminate bone fragments, one of which had been burnt at high temperature so not associated with the cranium, and a bird distal humerus likely belonging to a goose that also had evidence of butchery on the distal articulation. Other artefacts and ecofacts were recovered from this context, including small assemblages of pottery (16g), fire-cracked flint (105g) and very small quantities of charcoal, fired-clay and shell.

It can also be said that early medieval special deposits are highly varied (Hamerow 2012; Morris and Jervis 2011; Tipper 2004), and unambiguous comparable examples are rare. Given the post-disuse pit digging, the seeming placement of the deposit, and the conceivable residuality of other artefacts in this fill, this context and the cattle skull within it could still represent a deposit “closer to the ‘ritual’ end of the spectrum” (Morris and Jervis 2011, 72) and a useful addition to our understanding of early medieval depositional events.

Discussion

The animal bone assemblage from Coldharbour Road has limited archaeological significance due to poor preservation and high fragmentation. A large proportion of the assemblage likely derives from food waste and thus can suggest that people in the Late Bronze Age were eating cattle and ovicaprids, and those in the early medieval period were eating the main three domesticates, some domestic birds, and possibly some wild taxa (such as rabbits, although these could also be the result of bioturbation). The presence of microfauna including possible mice, rats and voles may indicate pests attracted to the bakehouses that died naturally or were caught and disposed of.

Though too much weight should not be placed on this poorly preserved assemblage, the emphasis on cattle and sheep has been noted to be a feature of the mixed pastoralism of the North Kent region and Thames Estuary during this period (Margetts 2021, 229-30). The area has also been highlighted for the number of manors that practised cattle dairying (ibid; Campbell 2000, fig. 4.03). The possibly ‘special’ deposits of cattle skulls may have ritual significance as discussed above. Particularly the upturned cattle skull placed in ST4 adds to the corpus of evidence and our understanding of this practice, of SFBs and the symbolic importance of animals in the early medieval period.

Plant Remains by Mariangela Vitolo

An extensive programme of environmental sampling was carried out at the site, targeting a range of features from Periods 1 and 2. A post-excavation assessment carried out on 41 bulk soil samples found that macroplant remains in Late Bronze Age/Early Iron Age features were scarce. 14 samples from several Period 2 contexts associated with the SFBs were deemed worthy of full analysis.

Results

ST3 was excavated and sampled in quadrants. The range, density and distribution of plant macrofossil taxa were balanced across almost all areas of the feature, except for the south-eastern quadrant, where a lower density of plant remains was noted. Considering that suitable 40L samples were taken from each of the quadrants, this cannot be a sample size issue. Free-threshing wheat (Triticum aestivum/durum) was dominant in the feature, immediately followed by hulled barley (Hordeum vulgare) and two tentatively identified grains of rye (cf Secale cereale). A single caryopsis of oat (Avena sp.) could indicate either a wild or a cultivated species. Few wild taxa were noted, including bramble/raspberry (Rubus sp.) seeds and hazel (Corylus avellana) nutshell fragments. These remains could have become accidentally charred, deriving from the surrounding vegetation or from material used for fuel. Arable weeds were also infrequent and included fat-hen (Chenopodium album), corncockle (Agrostemma githago), stinking mayweed (Anthemis cotula), brome (Bromus sp.), dock (Rumex sp.) and small scabious (Scabiosa columbaria).

A relatively small sample, measuring 5L, was taken from fill [424] of ST4. This feature produced a moderate amount of crop remains, mostly free-threshing wheat (Triticum turgidum/aestivum type) and some hulled barley (Hordeum vulgare). Fill [353], the capping of the feature, produced similar density of plant remains per litre as fill [424]. The range of taxa was roughly similar in both fills, but the top fill also produced two grains of oat (Avena sp.) and an indeterminate cultivated legume (Vicia/Lathyrus/Pisum sp.), in addition to two weed seeds such as stinking chamomile (Anthemis cotula) and knotgrass (Polygonum aviculare).

Pit [409] which truncated ST3 after its disuse, was sampled throughout its section and produced a far higher density of plant remains than the fills of the structural cut. The lower density of plant remains in the actual SFB could be due to the later pitting which erased most of the internal structure. Fill [407] was sampled on its north-east and south-west end. Fill [415] was the bottom fill which contained a deposited cow skull, whilst [404] was the upper fill. The density of remains and range of taxa were broadly similar in all fills of pit [409]. Free-threshing type wheat was again dominant, followed by oat (Avena sp.). Hulled barley and rye (Secale cereale) were present in smaller amounts. A sprouted barley grain and a detached coleoptile indicate germination. Chaff remains included indeterminate grass stem fragments, almost certainly from cereals, bread wheat, barley and rye rachis fragments. A number of indeterminate rachis fragments were likely to be either of barley or of rye but did not present enough diagnostic characters. A single diagnostic floret base indicates the presence of cultivated oat (Avena sativa) in this feature, although some wild oat could well have become accidentally mixed in with the crop. Weed remains were infrequent but included a wide range of taxa. Wild grasses of various size were present such as brome (Bromus sp.), ryegrass/fescue (Lolium/Festuca sp.) and meadow grasses/cat’s tails (Poa/Phleum sp.). Other weed taxa included corncockle (Agrostemma githago), vetch/tare (Vicia/Lathyrus sp.), eyebright/bartsia (Euphrasia/Odontites sp.), fat-hen, stinking mayweed, dock (Rumex sp.), sheep’s sorrel (Rumex acetosella), small scabious, and thistle (Centaurea sp.)

A large amount of charred remains was recovered from oven [341], which truncated ST4 on its south-eastern corner. The feature was again dominated by naked wheat, followed by similar amounts of hulled barley and oat, whilst rye grains were infrequent. The presence of a twisted barley caryopsis indicates the presence of six-row barley (Hordeum vulgare ssp vulgare), although the co-presence of the two-rowed subspecies cannot be ruled out. Weed seeds were as infrequent as in other contexts and included several caryopses of brome, including one of rye-brome (Bromus secalinus type), stinking chamomile, dock, sheep’ sorrel and knotgrass.

Pit [444] was a large tree-throw later converted into a pit, possibly to contain waste from the SFBs. Density of plant remains was low in this feature, which produced mostly naked wheat grains, with a caryopsis of six-row barley and one of rye. Wild grasses, including some caryopses of brome, were the only weeds present.

Two sampled fills of pit [336] produced a high density of charred crop remains. The assemblage was dominated by naked wheat grains, with smaller amounts of hulled barley, oat and rye. Weed seeds were infrequent and included a similar range of taxa encountered in other contemporary deposits on site.

[Details of the methodology used for this report and tabular data on the findings are published on the KAS website.]

Discussion

The most common cereal arising from the SFBs and associated features at Coldharbour Road is free-threshing wheat. This crop was dominant in all the analysed contexts. Infrequent but perfectly identifiable rachis remains indicate that at least some of it was of the hexaploid bread wheat. Hulled barley, including the six-row type, was the second most common crop from these features, followed by oat and then rye. Free-threshing wheat was also the dominant crop in SFBs excavated by Oxford Archaeology along the A2 (Smith 2012) as well as at a medieval oven at West Malling (Stevens 2009, 43). Interestingly, other Saxon and medieval features excavated further east on the island of Thanet yielded assemblages dominated by barley (Neveu in prep; Carruthers 2019; Stevens 2009, 143). This was likely due to local environmental factors, the lighter, sandy soils of Thanet being more suited to the cultivation of barley whereas the clay soils at Gravesend would have been better for growing wheat. It has also been suggested that the dominance of the one or other crop could be related to cultural reasons and to the status of cereal types (Stevens 2009, 143; Carruthers 2019).

The mixed nature of the assemblages indicates the possibility that different cereal species were grown together in order to buffer potential crop spoilage and/or to produce cheaper flours. Such mixtures could have been in the form of a maslin (barley and wheat) or a dredge (barley and oat). The evidence for such practices is notoriously difficult to identify from archaeobotanical assemblages as the mixing might have happened later, such as during storage or food preparation. Maslins and dredges existed in medieval times and it is therefore a possibility that such practices were in use at the site.

Ovens in the SFBs could have been used to make bread or to dry grains either before grinding or to prevent spoilage. It is, however, unlikely that they were used specifically for the purpose of malting. Despite the fact that beer was a commonly drunk in medieval England, often replacing water, and the technology for brewing was well-established, there is no indication of large-scale malt production at Coldharbour Road. In fact, the single sprouted grain of barley and the few detached sprouts or coleoptiles are more indicative of accidental germination. Similar evidence is available from the SFBs on the nearby A2 (Smith 2012), as well as at Thanet Earth (Carruthers 2019). At the latter, tentative evidence for malting is limited to a single feature (8 germinated grains out of 27). It seems likely that regardless of the location, these features were rarely if ever used to dry germinated grains for ale making.

The scarce presence of chaff is typical of assemblages of free-threshing cereals and in this case clearly the crops represent a relatively clean product from a late crop processing stage. This is indicated by the low presence of chaff and weed seeds. This alas hinders any in-depth discussion of weed ecology and only limited information can be inferred. The presence of stinking mayweed is common at post-Roman sites and indicates the cultivation of heavy clay soils where this weed thrives. The wild vetch/tare group of weeds normally live on poor, nutrient depleted soils, whilst other species such as fat-hen indicate nitrogen rich soils. Free-threshing wheats are high yielding but also very resource demanding. They would have needed nutrient rich soils to produce the large yields needed for an ever-growing population and it is therefore likely that more care was directed towards the fields where this crop was growing, whilst depleted soils were devoted solely to more forgiving crops, such as barley or rye. Soil enrichment could have been achieved in several ways, for example with the use of manure or through crop rotation. There is poor evidence for the latter practice being in use at the site, as cultivated pulses were scarce. Crop rotation was however likely at sites on the A2 (Smith 2012) as well as Thanet Earth (Carruthers 2019).

The local environment was probably dominated by open grassland and pasture, with several shrubs including hazel and raspberry or bramble. Sedges, as well as some species of buttercups and mints are indicative of wet conditions and could have grown by a nearby watercourse, for example on the bank of the Thames.

Scientific Dating by Lucy Allott

Eight samples were submitted to Beta Analytic for radiocarbon analysis. These samples primarily relate to cereal grain rich deposits from sunken feature buildings ST2 and ST3 as well as pit [422], which cuts ST3 and contained an inverted cattle skull from which a sample was submitted. A fragmented cattle skull from [38] the fill of pit [37] was also dated to help establish the antiquity of this deposit.

Conventional radiocarbon ages (Stuiver and Polach 1977) are presented in Table 3 and are quoted in accordance with the international standard known as the Trondheim convention (Stuiver and Kra 1986). Calibration of the radiocarbon ages has been undertaken using the maximum intercept method (Stuiver and Reimer 1986), the program OxCal v4.4 (Bronk Ramsey 2009), and the IntCal20 data set for terrestrial samples from the northern hemisphere (Reimer et al. 2020). Date ranges given in the table and text are those for 95% confidence and are quoted in the form recommended by Mook (1986), with the end points rounded outwards to 10 years.

Where two or more samples were submitted for dating from individual features, the radiocarbon dates have been subjected to chi-square test for consistency (Ward and Wilson 1978). In most instances, the later/younger of the measurements provides the best date for the context.

Measurements on Hordeum sp. (Beta-583562) and Triticum sp. (Beta-583563) from the fill [81] of posthole [80] in ST2, are statistically consistent at the 5% level (T’=0.9; v=1; T′(5%)=3.8; Ward and Wilson 1978) and could be of the same actual age. The younger of these (Beta-583563) calibrated to ad 1025-1205 provides the best estimate for the date of this deposit.

Measurements on cereal grains of Triticum sp. (Beta-583564) and Avena sp. (Beta-583565) as well as cattle skull bone fragment (Beta-583569) from [415] the fill of pit [422] are statistically consistent at the 5% level (T’=02.3; v=2; T′(5%)=6.0; Ward and Wilson 1978). This suggests the samples could be of the same age, relating to the same deposition event. The youngest calibrated date, ad 1035-1220, obtained on the cattle skull bone (Beta-583569) provides the best estimate for the context.

Measurements on Hordeum sp. (Beta-583566) and Triticum sp. (Beta-583567) from [343] the fill of SFB [342], ST3 are statistically consistent at the 5% level (T’=0.1; ν=1; T′(5%)=3.8; Ward and Wilson 1978) and the cereal grains could be of the same actual age. The calibrated date ranges are almost identical with that obtained on the wheat (Beta-583567) providing a slightly later end date and therefore the best estimate of cal ad 990-1155.

A single skull bone fragment (Beta-583568) from [38], the fill of pit [37] provided a calibrated date of 1380-1120 cal bc.

discussion

Topography and place: themes of re-use.

To understand why areas of prior occupation have been reused by peoples from other [later] periods in part relies on an understanding of the topography of the local area. To the north of the Coldharbour Road site lay the riverside settlement of early Gravesend (OE Grafes ende), the River Thames and east passage to Essex with routes across the landscape that had long been established by the time of the Saxo-Norman re-use of the prehistoric enclosures (EN1 and EN2).

The site of Coldharbour Road resides in a large north-south valley and sits roughly two miles south from the River Thames and would have made an excellent transport route to and from the river for livestock in the Late Bronze Age. Later development of the area in the Roman period meant that routeways became more defined and it is likely that the Downs Road dry valley, with its metalled Roman road, was used as the natural route from the estuary to the chalk uplands (Allen et al. 2012, 569).

The valley itself is of particular interest and the large, chalk banked, boundary ditch (D1) that lined the top of the site would have stood as a stark feature in the landscape, perhaps marking a boundary or line of communication. Though this large ditch fell into disuse between the Late Bronze Age and the Saxo-Norman period it is clear that linear connections between the chalk and the Thames were already well established.

It is not uncommon for people in later periods to adopt and re-use earlier earthworks if they are available or still prominent in the landscape. Enclosures or boundaries are re-used, and often this re-use had an entirely different function to how they were originally intended. Sites such as Perry Oaks Sludge Works (Heathrow Terminal 5), exhibit an evolving Bronze Age landscape that was subject to re-use or redefinition (Barrett et al. 2001, 222). The developing field system respected the boundaries of an earlier D-shaped enclosure and was remodelled and re-used in the Middle to Late Bronze Age. Excavations at Chilmington Green near Ashford have also uncovered re-use of prior earthworks and boundaries from the Late Bronze Age to the Roman occupation of Britain (ASE 2019a) and works by Oxford Archaeology during the A2 widening scheme (Site C) show that many medieval features, including ditches, appear to respect earlier prehistoric earthworks (Allen et al. 2012, 499), suggesting that these must have still been visible in the landscape.

The possible Saxo-Norman re-use of the Bronze Age enclosures at Coldharbour Road was designed to house three SFBs and as discussed above these would have been placed away from the main settlement to avoid the clear threat of fires from the large ovens housed within. The small earthworks that were excavated in addition to the SFBs sat within the square enclosures of the Late Bronze Age/Early Iron Age (EN1 and EN2) and were possibly an augmentation to the already established boundaries. Though, it is extremely likely that only the prehistoric banks and not the ditches themselves were visible in the Saxo-Norman period, leaving only a raised impression of an ancient enclosure.

The prehistoric remains in the wider landscape

The prehistoric remains excavated on site were likely to have been related to the maintenance and herding of livestock and pastoral farming. The lack of occupational evidence, from material remains to waste disposal suggest that the area was not used for human habitation. Occupation sites lay to the south, where settlement evidence has been uncovered by Oxford Archaeology (Allen et al. 2012). At least two enclosures (EN1 and EN2) (with possible subdivisions) and a large north-south aligned ditch (D1) suggest that this valley at Coldharbour Road was part of a larger agricultural landscape.

Yates (2007, 167) states that farming intensification is linked to the appearance of field systems and that fixing of boundaries for long term use characterises a fully agricultural society. Exploitation of the River Thames would have been paramount for any intensification of agriculture and long boundary ditches, coaxial field systems and droveways represent a new way in which the landscape was utilized in the Bronze Age. These large, often imposing field boundaries indicate a thriving Bronze Age population in south-east England and were created as a response to the pressures of intensified farming and livestock management (Pryor 1998, 82).

Excavations elsewhere on Coldharbour Road showed a corner of a square enclosure near a large north-south droveway (Mudd 1994). The features bear a strong similarity to the prehistoric boundary ditch and adjacent enclosures uncovered by ASE. Yates (2007, 26-27) suggests these droveways with associated roadside holding pounds are characteristic of the English Channel – North Sea exchange region and it is likely that the boundary ditch D1 and enclosures EN1 and EN2 are of similar function. This corresponds well with Champion’s synthesis of the Late Bronze Age where he concluded that the focus of activity was either riverine, estuarine or coastal (Champion 1980, 229). The nature of the prehistoric site, it seems, was intrinsically linked to the exploitation of the River Thames and was likely to have been a communal area, rather than belonging to a specific settlement of Late Bronze Age peoples.

The size of these ditches mentioned above were clearly designed for a large-scale operation of herding through the valley towards the Thames. The site of Hays, Dagenham, supports the interpretation that such routeways were designed to handle the passage of large herds (Meddens 1996, 326). The ‘roadside’ holding pens found in both Coldharbour Road excavations also reinforce the notion that this was a working landscape, divided into clear partitions for the transport of cattle and likely traded goods coming from the Thames.

These ‘roadside’ enclosures would have been used as rest-stops, likely to have been communally owned and maintained. They may have also been used to encourage husbandry between different herds of cattle – an idea proposed by Francis Pryor at Fengate (Pryor 1998, 129).

The intensification of the landscape and the effort required to construct and maintain large field systems, droveways and communal holding pens illustrates the importance of herding in the Late Bronze Age. The clearly defined spaces which were used for the transportation of livestock and other traffic from the Thames indicate that special consideration was given to herders or pastoral farmers (Fleming 1971; Peters 2000).

The prehistoric remains uncovered at Coldharbour Road correspond well with prior interpretations of the Late Bronze Age landscape of the region. The large north-south boundary ditch and at least two animal enclosures indicate the presence of pastoral farming and exploitation of a riverine landscape. Looking at more established examples of this across the south-east (and further afield) we can make a good case that they relate to communal ‘roadside’ livestock pens. The boundary ditch (D1) shares a similarity to the droveway excavated by Mudd (1994) and other examples of long boundaries used to direct livestock between grazing areas and towards water. This all indicates the presence of a well-used and well-defined landscape here in the Late Bronze Age based around the herding of livestock connected to the River Thames.

The early medieval remains in the wider landscape

The Saxo-Norman remains excavated on site were almost exclusively associated with the three SFBs (likely to have been bakehouses). Settlement evidence on site was sparse and indicates that while baking and cooking was likely to have taken place, the area of occupation would likely have been further away from these buildings. Earlier Anglo-Saxon settlements along the Thames are often widely spread with SFBs dispersed amongst fenced enclosures, posthole houses and open areas (Booth et al. 2007, 81-99). Oxford Archaeology also found buildings of similar morphology during their excavations on the A2 widening scheme (Allen et al. 2012) reinforcing the concept of a wide dispersal of these buildings alongside established settlements.

The bakehouses excavated at Coldharbour Road appear to date from around the 11th-late 12th centuries and were part of a unique Kentish tradition of the Saxo-Norman period (Bennett et al. 2008; Rady and Holman 2019). The date of these structures was derived from small quantities of pottery and radiocarbon dating, but the overall phasing pairs well with medieval Phase 2 on the A2 widening scheme (Allen et al. 2012).

It is clear from Coldharbour Road, the A2 widening scheme and HS1 that during the 11th-13th centuries, the area at the top of an east facing valley side underwent significant development with an intensification of activity. This can be seen by the construction of numerous SFBs and as the population grew, so too did the need for bread, beer and more permanent structures to process crops. Environmental sampling of the bakehouses at Coldharbour Road indicate that wheat, oats, barley and rye were all being processed, showing that agriculture was varied and diverse in this area of Kent and was likely an effect of the intensification of settlement in the area at this period. Alongside crops there was also evidence for livestock farming and despite the poorly preserved bone assemblage the remains suggest that Saxo-Norman people were consuming the three main domesticates along with domestic birds and possibly rabbits. The presence of a cattle mandible in an oven perhaps indicates cooking.

Artefacts from within the SFBs at Coldharbour Road were sparse, and consisted of pottery sherds, struck flint and fragments of German lava quern, which were encountered in all four quadrants of ST2. The inclusion of this stone is relevant in identifying the site’s relationship with the wider community as in this area German lava stone is fairly common for the Saxo-Norman period. The import of this particular stone, suggests continued trade connections with the continent; such trade routes may have been established as early as the Roman period and later utilised by the local medieval populations, such as the metalled Roman Road at Downs Road (Allen et al. 2012, 569).

How the Saxo-Norman people utilised prehistoric enclosures in the South-East is a neglected subject. The Coldharbour Road excavations indicate that some residual remains of the Late Bronze Age enclosures are very likely to have been present when the SFBs were established. What was left for the medieval people to interact with was likely to have been the banks of the former enclosures (EN1 and EN2) and the bank of the large north-south boundary ditch (D1).

The site of Coldharbour Road lies near the ancient parishes of Swanscombe (now Swanscombe and Greenhithe), Southfleet, Ifield (now part of Northfleet), Cobham and Shorne. All there were listed, besides Cobham and Shorne, in Domesday as manors or estates in 1086. Local routes became frequently used after the Norman Conquest and into the 13th century including a passage from the river Thames to London, named the Long Ferry (Hiscock 1968, 229).

The activity at Coldharbour Road reflects the increase and intensification of medieval use of the area. Multiple nearby parishes and manors and a re-use of older routeways and sea routes (Tatton-Brown 2001) would have caused an influx of people either trading, settling or passing through in pilgrimage to Canterbury (Watling Street near Coldharbour Road would have been used as the primary route for pilgrimage).

Historic OS maps from 1860 show that a field boundary, possibly hedgerows, leading into Gravesend’s Darnley Road towards the Thames followed a strikingly similar alignment to the Late Bronze Age land division (D1) as do local parish boundaries. Such long-lived linearity in the landscape may have originally been defined in the Bronze Age and remained functional and in-use over a long period of time. Indeed, pioneering use of scientific techniques in the dating of routeways is beginning to show their antiquity. At Lyminge, for example, a holloway that led to the famous Anglo-Saxon royal centre was certainly established by the Roman period and in its earliest phase (a lynchet against which the routeway formed) it originated during the Late Bronze Age or Iron Age (Bell 2020, 237).

Structured deposition, closing deposits, boundary deposits

At least two structured deposits were potentially uncovered at Coldharbour Road spanning the two separate periods (Late Bronze Age/Early Iron Age and Saxo-Norman). Unusually, both deposits were the same and consisting of cattle skulls placed in two differing pits. Special deposits appear to emphasise important parts of the landscape and can be used to provide clues about the complex intermingling of life and religion in the Late Bronze Age and Saxo-Norman period.

The prehistoric deposition involved the placement of a cattle skull into a pit, cut directly into the base of the large north-south boundary ditch (D1) showing that the boundary ditch was partially open at the time of deposition (Fig. 13).

No further bones were recovered, and it is clear that this was a deliberate act of ritual. The choice of cattle skull is not unusual considering the use of the area as a ‘roadside’ enclosure and the land division’s probable use to manage livestock movement towards the Thames. There appears to have been a religious connection to the animals that were moved across these substantial trackways which shows how the landscape was not solely an impersonal expression of demographic and economic forces (Fokkens 1999, 41).

There is increasing evidence for the act of a special disposition into field systems or showing that these enclosures or farmsteads are likely recipients for ritual acts and particular deposits can include quern stones, bronze objects and token human cremations (Brück 2001, 151). The act of deposition into the field boundary could indicate a reverence for the livestock that the community depended upon and reflect gratitude on behalf of the special status that herders were likely afforded (Fleming 1971; Peters 2000). Cattle skulls in the Bronze Age might also have had a particular significance ascribed to them as they have been found alongside burials such as in Irthlingborough barrow (Towers et al. 2010), though the reason for such use is still unknown.

There was also another possible ritualised deposit of dog remains excavated within the southern enclosure ditch of EN1. Analysis has revealed that elements may belong to a single dog and might provide insight into how domesticated animals were treated and give an indication that dogs may have been used to herd cattle.

Moving forward into the Saxo-Norman period, C14 dating on a placed inverted cattle skull within a large bell-shaped pit, potentially marking a ritual termination deposit, provides a date range of cal. ad 1035-1220 (900 ± 30 bp, Beta-583569, 95% probability). The bell-shaped pit the skull was recovered from was cut entirely within the footprint of ST4 and possibly formed part of a ritual deposition that included the discarding of waste material and other potential offerings (some which may have been burnt). Cereals recovered from the same context of the skull [415] have a similar C14 date with wheat at cal. ad 1020-1165 (960 ± 30 bp, Beta-583564, 95% probability) and oat with a date range of cal. ad 1025-1170 (950 ± 30 bp, Beta-583565, 95 % probability).

Termination rituals are not uncommon practice in the early medieval period and it has been observed that when SFBs would fell out of use, they would sometimes be deliberately destroyed and occasionally offerings deposited at the base of the building as seen by Hamerow (2006) on sites such as Pennyland, West Stow and Yarnton. Two out of the three SFBs excavated at Coldharbour Road, exhibited some kind of termination pitting (ST3 and ST4). While ST3 was truncated by a single pit, by far the most extensive pitting was encountered in ST4 where nearly the entire structure had been truncated away (Fig. 8). One of the largest pits in ST4 was within the centre; a large bell-shaped pit destroyed most of the internal remains and at the bottom, directly in the middle of the pit, an inverted cattle skull was placed (Fig. 12).

The deposition of cattle skulls, it seems, continues a tradition of the Saxo-Norman peoples that was originally practised in prehistory, though it is unclear if their two acts had the same connotations. As well as linking the importance that was placed on cattle between the two periods, this deposition may also tell us about ritual practice in what should be a Christianised Kent. It is possible that this find demonstrates that Saxo-Norman communities may have practiced a mix of Christian and pagan ritual, especially when related to day-to-day practices such as animal husbandry or the termination or disuse of buildings.

Both depositions can tell us a little of the worldview the people had who placed them and that fixed places in the landscape, whether they be large boundary ditches or buildings, held a significance to the people who used them that was greater than their practical purpose. Certainly, there must have been an esoteric element to how respect was paid to things that formed an integral part of their life. Cattle skulls were the chosen offering for the deposits, and it is possible that the reverence placed on these animals and what they signified in terms of wealth, health and community echoed through the ages.

Future research directions

The prehistoric nature of the site at Coldharbour Road corresponds well with the analysis of riverine exploitation discussed by Yates (2007), especially considering its proximity to the Thames. Further research might be focused on identifying Late Bronze Age settlements and how they interacted with droveways and boundary ditches and the herds that used them. Did local centres of power control these vast trackways or were they a part of the landscape that extended beyond the authority of tribal leaders? Ritual function within trackways or farm enclosures has been increasingly identified in the Late Bronze Age, but can any further meaning be attributed to it other than reverence or respect of what was being used on a daily basis? Did cattle play a more important role in Late Bronze Age society than simply livestock for food; were they displays of wealth such as in other cultures? Did accumulation of the animal merit higher social status as discussed by Fleming (1971) and Peters (2000) or were whole communities responsible for the herds they depended on evidenced by the use of communal ‘roadside’ enclosures?

Medieval activity is well documented in the area of the A2 and the HS1 excavations. The use of SFBs in this period is becoming better understood (Rady and Holman 2019) and they are often attributed multiple functions such as kitchens-cum-brewhouses, bakeries, crop drying facilities or even living spaces. Further environmental analysis of surrounding SFBs and the landscape could go some way in identifying the specific uses or determine that they were multi-functional in nature, the diet of the people who used them and if function changed seasonally.

The potential ritual aspect behind the termination process of these buildings is something that offers a wide range of areas to explore. SFBs identified just to the south of the site at the A2 widening scheme exhibited no termination ritual, neither did any identified at the 2017 Cliffsend excavations in Thanet (ASE 2019b; Dawkes in prep.) nor the 11th/12th-century examples at Thanet Earth (Rady and Holman 2019). Further research could be directed into understanding ritual ‘pagan’ practice in Saxo-Norman Christian Kent and how that correlates with termination practices of SFBs. It would appear that if cattle skull RF <6> was indeed linked to a termination rite, this represents a continuation of practices associated with Anglo-Saxon Grubenhäuser being applied to what is currently thought of as a separate Saxo-Norman building tradition.

acknowledgements

Archaeology South-East would like to thank RPS Group for commissioning the project and Wendy Rogers (Heritage Conservation Group Kent County Council) for her guidance and monitoring. The excavations were directed by the author and managed by Paul Mason (fieldwork) and Jim Stevenson and Andy Margetts (post-excavation). This article was edited prior to submission by Andy Margetts.

bibliography

Allason-Jones, L., 1988, ‘The small finds’, in: Bishop, M.C. and Dore, J.N., Corbridge: Excavations of the Roman fort and town, 1947-80, EH Arch. Report, no. 8, 159-218.

Allen, T.M, Hardy, A., Hayden, C. and Powell, K., 2012, A road through the past: Archaeological discoveries on the A2 Pepperhill to Cobham road-scheme in Kent.

ASE, 2019a, ‘Post-Excavation Assessment and Updated Project Design Report for Arch-aeological Excavation at parcel Q and R, Land at Chilmington Green, Kent’, Archaeology South-East unpubl. report.

ASE, 2019b, ‘Post-Excavation Assessment and Updated Project Design Report for Archaeological Excavation at Land North of Cliffsend Road, Cliffsend Ramsgate, Kent’, Archaeology South-East unpubl. report.

Barrett, J.C., Lewis, J.S.C. and Welsh, K., 2001, ‘Perry Oaks – a history of inhabitation, part 2’, London Archaeologist, vol. 9 (8), 221-227.

Bell, M., 2020, Making one’s way in the world. The footprints and trackways of prehistoric people, Oxford: Oxbow Books.

Bennett, P., Clark, P., Hicks, A., Rady, J. and Riddler, I., 2008, At the Great Crossroads: Prehistoric, Roman and Medieval discoveries on the Isle of Thanet 1994-5, CAT Occas. Paper 4.

BGS, 2021, British Geological Survey online viewer. https://mapapps.bgs.ac.uk/geology ofbritain/home.html [accessed 15/12/21]

Booth, P., Doff, A., Robinson, M. and Smith, A., 2007, The Thames through Time: The archaeology of the Gravel Terraces of the Upper and Middle Thames. The early historical period: AD 1-1000, Oxford Archaeol. Thames Valley Landscapes, Mono. 27.

Boessneck, J., 1969, ‘Osteological differences between sheep (Ovis aries Linné) and goats (Capra hircus Linné)’, in D. Brothwell and E. Higgs, Science in Archaeology: A survey of Progress and Research. London: Thames and Hudson, 331-58.

Bronk Ramsey, C., 2009, ‘Bayesian analysis of radiocarbon dates’, Radiocarbon, vol. 51, 337-360.

Brück, J., 2001, ‘Body metaphors and technologies of transformation in the English Middle and Late Bronze Age’, in J. Brück (ed.), Bronze Age Landscapes. Tradition and Transformation, Oxford: Oxbow Books, 149-160.

Campbell, B.M.S., 2000, English Seigniorial Agriculture: 1250-1450, CUP.

Carruthers, W., 2019, Charred, Waterlogged and Mineralised Plant Remains, in J. Rady and J. Holman, Beneath the Seamark 6000 years of an Island’s History. Archaeological Investigations at ‘Thanet Earth’, Kent 2007-2012. Canterbury Archaeological Trust Ltd. Technical report 2, 1087-1135.

Cohen, A. and Serjeantson, D., 1996, A Manual for the Identification of Bird Bones from Archaeological Sites, London: Archetype.

Champion, T., 1980, ‘Settlement and environment in Later Bronze Age Kent’, in J. Barrett, and R. Bradley (eds.), Settlement and Society in the British Later Bronze Age, BAR, vol. 83, 223-246.

Dawkes, G., in prep., Archaeological excavations at Cliffsend, Isle of Thanet and its environs, SpoilHeap monograph.

Fleming, A., 1971, ‘Territorial patterns in Bronze Age Wessex’, PPS, vol. 37 (1), 138-166.

Fokkens, H., 1999, ‘Cattle and martiality: changing relations between man and landscape in the Late Neolithic and the Bronze Age’, in Fabech, C. and Ringtved, J. (eds), Settlement and Landscape. Proceedings of a conference in Århus, Denmark, May 4-7 1998, Højbjerg: Jutland Archaeol. Soc., 35-43.

Goodall, I.H., 2011, Ironwork in medieval Britain: an archaeological study, The Society for Medieval Archaeology monograph, 31.

Grant, A., 1982, ‘The use of tooth wear as a guide to the age of domestic animals’, in R. Wilson, C. Grigson and S. Payne (eds.), Ageing and sexing animal bones from archaeological sites, Oxford: BAR British Series 109, 91-108.

Halstead, P., 1985, ‘A study of mandibular teeth from Romano-British contexts’, in I. Longworth (ed.), The Fenland Project, Number 1: The Lower Welland Valley, volume 1. East Anglian Archaeology 27, 219-221.

Halstead, P. and Collins, P., 2002, ‘Sorting the sheep from the goats: morphological distinctions between the mandibles and mandibular teeth of adult Ovis and Capra’, Journal of Archaeological Science, vol. 29, 545-553.

Hamerow, H., 2006, ‘Special deposits’ in Anglo-Saxon settlements’, Medieval Archaeol., vol. 50 (1), 1-30.

Hamerow, H., 2012, Rural settlements and society in Anglo-Saxon England, oup.

Hardy, A. and Bell, C., 2001, The excavation of a medieval rural settlement at the Pepper Hill electricity substation, Northfleet, Kent. Oxford: Oxford Archaeol. Occas. Paper No. 10.

Hillson, S., 1992, Mammal bones and teeth: an introductory guide to methods of identification, London: The Institute of Archaeology, University College London.

Hiscock, R.H., 1968, ‘The road between Dartford, Gravesend and Strood’, Archaeologia Cantiana, 83, 229-47.

Jones, G.G. and Sadler, P., 2012, ‘Age at death in cattle: methods, older cattle and known-age reference material’, Environmental Archaeology, 17 (1), 11-28.

Manning W.M., 1985, Catalogue of the Romano-British iron tools, fittings and weapons in the British Museum, London: British Museum Press.

Margetts, A., 2021, The wandering herd: The medieval cattle economy of the south-east England c.450–1450, Oxford: Windgather Press.

Margetts, A. Doherty, A. Douglas, C. Nicholls, H. and Stevens, S., in prep., Out of the Weald the secret Weald, Portslade: SpoilHeap monograph.

Meddens, F.M., 1996, ‘Sites from the Thames estuary wetlands, England, and their Bronze Age use’, Antiquity, vol. 70, 325-334.

Mook, W.G., 1986, ‘Business meeting: Recommendations/Resolutions adopted by the Twelfth International Radiocarbon Conference’, Radiocarbon, vol. 28, 799.

Morris, J. and Jervis, B., 2011, ‘What’s so special? A reinterpretation of Anglo-Saxon ‘Special Deposits’, Medieval Archaeology, vol. 55(1), 66-81.

Mudd, A., 1994, ‘The Excavation of a Later Bronze Age site at Coldharbour Road, Gravesend’, Archaeologia Cantiana, 94, 363-410.

MoLAS, 2001, ‘Area 330 (Zone 3) Northumberland bottom (ARC WNB 98) Archaeological Post Excavation Report’, unpubl. report.

Neveu, E., in prep., ‘Charred Plant Remains’, in G. Dawkes, Archaeological excavations on South Thanet and its Environs, SpoilHeap monograph.

Peters, F., 2000, ‘Two traditions of Bronze Age burial in the Stonehenge landscape’, Oxford Journal of Anthropology, vol. 19 (4), 343-358.

Poole C., 1984, ‘Objects of baked clay’, in: Cunliffe B., Danebury: an Iron Age hillfort in Hampshire. Vol 2. The excavations 1969–1978: the finds, CBA Research Report 52, 398-407.

Pryor, F., 1998, Farmers in Prehistoric Britain, Stroud: Tempus.

Rady, J. and Holman, J., 2019, Beneath the seamark: 6,000 years of island history. Archaeological investigations at ‘Thanet Earth’, Kent 2007–2012, Canterbury Archaeological Trust Technical Report 2. Volume 1: Chronological Narrative,

Reimer, P.J. et al., 2020, The IntCal20 Northern Hemisphere Radiocarbon Age Calibration Curve (0–55 cal kBP), Radiocarbon, vol. 62, 725-757.

Schmid, E., 1972, Atlas of Animal Bones for pre-historians, archaeologists and quaternary geologists, Amsterdam: Elsevier Publishing Company.

Schuster, J. and Stevens C.J., 2009, Kentish Sites and Sites of Kent. A Kentish Type of Medieval Sunken-Featured Bakery/Kitchen, pp. 249-250.

Serjeantson, D., 1996, ‘The Animal Bones’, in S. Needham, and T. Spense (eds.), Runnymede Bridge Research Excavations, Volume 2: Refuse and Disposal at Area 16 East, Runnymede, London: British Museum.

Scott, I.R., 1980, ‘Spearheads of the British limes’, in Hanson W.S. and Keppie, L.J.F. (eds), Roman Frontier Studies 1979, BAR 571, 333-43.

Silver, I.A., 1969, ‘The ageing of domestic animals’, in D. Brothwell and E, Higgs (eds.), Science in Archaeology: A survey of Progress and Research, London: Thames and Hudson, 283-302

Smith, W., 2012, ‘Charred Plant Remains’, in T. Allen, M. Donnelly, A. Hardy, C. Hayden, and K. Powell, A Road Through The Past: Archaeological Discoveries on the A2 Pepperhill to Cobham road-scheme in Kent, Oxford Archaeology Monograph 16, 562-566.

Stevens, C.J., 2009, ‘Environmental evidence’, in P. Andrews, K. Egging Dinwiddy, C. Ellis, A. Hutchenson, C. Phillpotts, A.B. Powell and J. Schuster, Kentish Sites and Sites of Kent. A miscellany of four archaeological excavations, Wessex Archaeology Report No. 24, 91-92, 125-127, 133-134.

Stuiver, M. and Kra, R.S., 1986, ‘Editorial comment’, Radiocarbon, vol. 28, ii.

Stuiver, M. and Polach, H.A., 1977, ‘Reporting of 14C data’, Radiocarbon, vol. 19, 355-363.

Stuiver, M. and Reimer, P.J., 1986, ‘A computer program for radiocarbon age calibration’, Radiocarbon, vol. 28, 1022-1030.

Tatton-Brown, T., 2001, ‘The evolution of ‘Watling Street’ in Kent’, Archaeologia Cantiana, 121, 121-33.

Tipper, J., 2004, The Grubenhäus in Anglo-Saxon England. Analysis and Interpretation of the Evidence from a Most Distinctive Building Type, Yedingham, The Landscape Research Centre.

Towers, J., Montgomery, J., Evans, J., Jay, M. and Person, M.P., 2010, ‘An investigation of the origins of cattle and aurochs deposited in the Early Bronze Age barrows at Gayhurst and Irthlingborough’, J. Archaeol Sci, vol. 37 (3), 508-15.

Ward, G.K. and Wilson, S.R., 1978, ‘Procedures for comparing and combining radiocarbon age determinations: a critique’, Archaeometry, vol. 20, 19-32.

Yates, D., 2007, Land Power and Prestige: Bronze Age Field Systems in Southern England, Oxford: Oxbow Books.

Zeder, M.A. and Lapham, H.A., 2010, ‘Assessing the reliability of criteria used to identify postcranial bones in sheep, Ovis, and goats, Capra’, Journal of Archaeological Science, vol. 37 (11), 2887-2905.

Fig. 1 Site location.

Fig. 2 Phased plan of Period 1.

Fig. 3 Four-post ST1 looking north.

Fig. 4 Section and photograph of D1 looking north.

Fig. 5 Phased plan of Period 2.

Fig. 6 Plan of ST2.

Fig. 7 Plan of ST3.

Fig. 8 Photo of ST4 looking south (top) and section of ST4 showing contexts mentioned in the text (bottom).

Fig. 9 Inverted cattle skull RF <6> at base of pit [422].

Fig. 10 Ceramic weight RF <9>.

Fig. 11 Iron spearhead RF <1>.

Table 1. Zooarchaeological assemblage by period

showing total fragment count (N), the number of hand-collected (HC) and bulk-sampled (ENV) specimens, the number of identifiable specimens (NISP) and the proportion of bones displaying varying preservation levels.

Period		N	HC	ENV	NISP Poor	Preservation %
						Moderate	Good
0	Undated	6	4	2	1	100	0	0
1	Late Bronze Age/Early Iron Age 1150-600 bc	327	327	0	211	79.5	20.5	0
2	Early Medieval ad 1050-1200	406	298	108	150	46.6	45.6	5.9
Total		739	629	110	362	61.6	34.1	3.2

Table 2. Taxa abundance in the overall and phased assemblages by NISP

Taxa	NISP	Period
		1	2	0
Cattle	68	19	49	0
Ovicaprid	29	17	11	1
Pig	4	0	4	0
Dog	20	20	0	0
Chicken	8	0	8	0
Goose	2	0	2	0
C.f. goose	3	0	3	0
Rabbit	3	0	3	0
Mouse sp.	3	0	3	0
Rat/ water vole sp.	1	0	1	0
Mouse/ vole sp.	1	0	1	0
Large mammal	196	149	47	0
Medium mammal	21	6	15	0
Microfauna	2	0	2	0
Bird	1	0	1	0
Indeterminate	377	116	256	5

Fig. 12 The near-complete cattle cranium from pit [422] fill [415] in situ.

Table 3. Summary of radiocarbon dates from Coldharbour Road, Gravesend

Lab Code Beta-	ASE Sample Ref.	Feature	Cxt	Bulk Sample	Material Type	Dated component	δ13CIRMS (‰)	δ15NIRMS (‰)	C14 age (BP)	Calibrated range (2σ)
583562	DS_799	Posthole [80], ST2	81	12	Charred Plant macrofossil	Hordeum sp. caryopsis	-23.3‰		980±30	cal ad 995–1160
583563	DS_800	Posthole [80], ST2	81	13	Charred Plant macrofossil	Triticum sp. caryopsis	-26.1‰		940±30	cal ad 1025–1205
583564	DS_801	Pit [422]	415	28	Charred Plant macrofossil	Triticum sp. caryopsis	-21.2‰		960±30	cal ad 1020–1165
583565	DS_802	Pit [422]	415	28	Charred Plant macrofossil	Avena sp. caryopsis	-24.0‰		950±30	cal ad 1025–1170
583569	DS_806	Pit [422]	415	-	Animal bone	Cattle skull	-21.2‰	7.8‰	900±30	cal ad 1035–1220
583566	DS_803	SFB [342] ST3 (SW quad)	343	33	Charred Plant macrofossil	Hordeum sp. caryopsis	-21.9‰		1010±30	cal ad 990–1150
583567	DS_804	SFB [342] ST3 (SW quad)	343	33	Charred Plant macrofossil	Triticum sp. caryopsis	-22.2‰		1000±30	cal ad 990–1155
583568	DS_805	Pit [37]	38	-	Animal bone	Cattle skull	-20.4‰	7.4‰	2990±30	1380–1120 cal bc

Fig. 13 Placed deposit within D1.