Continuity and Change in Late Iron Age Roman Transition within the Environs of Quarry Wood Oppidum: excavations at Furfield Quarry, Boughton Monchelsea

ArticlesVolume 134

25 Jun

Written By KAS

Continuity and change in the Late Iron Age-Roman transition within the environs of Quarry Wood opppidum: excavations at Furfield Quarry, Boughton Monchelsea

isca howell

With Julian Bowsher, Jon Cotton, Rupert Featherby, Kieron Tyler and Anthony Mackinder

This article reports on the excavations at Furfield Quarry, Boughton Mon-chelsea, to the south-east of Maidstone (NGR 78320 51680). The site was bounded to the north and north-west by Brishing Lane, to the north-east by Brishing Road, a disused quarry (Furfield Quarry) to the south and to the east by an area of woods known as Hogstrough Shaw. The wood known as Long Shaw divides the site in two; West Field and East Field (Fig. 1).

Evaluation of the whole site by trial trenching took place in August 1996 (Philp 1996). This identified archaeological remains of Iron Age, Romano-British and post-medieval date in the East Field resulted in two specifications for archaeological investigation (KCC 2005a; KCC 2005b): a strip, map and record of the East Field, carried out from April-July 2005, and a watching brief in advance of, and during development of the West Field for new housing, undertaken from October 2006 to April 2007. Museum of London Archaeology (MOLA) undertook both phases of work.

The results of the excavations confirmed the results of evaluation that significant remains survived on the East Field (site code KT-BMS05), hereafter known as ‘the site’. This report focuses on the results of a programme of analysis that has identified significant evidence for the study of the development of west Kent in the late Iron Age/early Roman period.

The site lies on the north side of the Loose Stream, a tributary of the River Medway. Topographically, it is located at c.100m od, on the Green-sand Ridge; an extensive, prominent, often heavily wooded, sandstone escarpment and range of hills that forms a horseshoe-shape around the Weald of Surrey, Sussex and Kent. The underlying geology is Lower Greensand with Hythe Beds (Geological Survey of Great Britain Sheet 288). The rock from the Hythe Beds has been quarried for building material in the past, and consists of bands of ragstone and softer ‘hassock’ or sandstone. Although the upper Hythe Beds form a stiff clay, it is covered in a fertile loam, thus the site would appear to have been well located being close to running water and surrounded by woodland, and the resources they provide, such as fuel, building material and game.

1.35km to the west of the site is Quarry Wood Camp (Fig. 2), an earthwork comprising single east and west banks, with part of a ditch on the east side. Kelly (1972, 55-84) described the earthwork as part of a ‘Belgic oppidum’. He based this interpretation on the recovery of pottery typically found on ‘late Belgic domestic sites in East Kent’, and by comparison to several other ‘known’ oppida in the south of England. The pottery referred to had been found during excavations in 1963-7 on the earthwork, but Kelly also referred to other finds of ‘Belgic’ pottery in the area, including Furfield Quarry.

Whilst the site lies away from and on the other side of the Loose Stream to Quarry Wood Camp, it does lies within an area of surrounding earthworks (Fig. 2). Kelly also thought this to be part of the oppidum. The surviving traces of the earthwork appear to form the northern and southern boundaries of a territory centred on Quarry Wood Camp and the Loose Stream, although the lack of evidence for its eastern and western limits suggest territory may be too strong a definition. An assessment of the section of earthwork 250m to the south-east of the site, undertaken in 2006, concluded it was reasonable to provisionally associate the earthworks with the Late Iron Age activity related to the ‘Loose oppidum’ based around Quarry Wood Camp (Elsden 2006, 12).

The term ‘oppidum’ was derived from classical sources, as noted by Timothy Champion when describing oppida in Kent (Champion 2007, 121), as ‘the Latin word for towns used by (Julius) Caesar to refer to some of the major Iron Age sites he encountered’. Although, these descriptions may be more appropriate to Gaul where the term was coined, Quarry Wood Camp does have similar defences to those adopted in northern Gaul, known as the Fécamp style defences. These defences comprise a massive dump rampart fronted by a wide, shallow, flat-bottomed ditch (Cunliffe 2005, 150). It is perhaps unsurprising that a Late Iron Age site in Kent should share characteristics with contemporary sites in northern France when the two regions have so many similar cultural aspects.

Another characteristic that may define Quarry Wood Camp as an oppidum is its setting on sloping ground. This distinguishes it from the adjacent putative oppida at Oldbury to the west and Bigbury to the east, which are set on hilltop or hill-promontory locations, and are probably better defined as hillforts. A third factor could be the concentration of Iron Age coinage in the area (Cunliffe 1982, 46-7) representing the market economy associated with oppida (Cunliffe 2005, 603). However the discovery of coin moulds during excavations in Rochester (Detsicas 1983, 3; 14; 59 and fig. 2) suggest a shift of focus, possibly from Quarry Wood Camp, before the Claudian invasion, just as there appears to be a shift from Bigbury to Canterbury at that period. These occurrences, if correct, would have probably led to these settlements being adopted as Roman towns.

Evidence for Roman activity in the area lies to the south-west of the site where the projected line of the Roman road from Rochester crosses the valley of the Loose Stream and forks c.1.km to the south-east for routes through the Weald and presumably on to the coast (Fig. 2). South-west of the site, c.400m, and to the west of the Roman road, is the site of a bathhouse. The field in which the building was found first attracted attention around 1834 when a human skeleton, urns and a dish were found. Subsequent finds in the field included glass vessels and more Roman pottery. In August 1841 Charles Roach Smith was informed that farming work had uncovered a foundation. This was part of a Roman building, most of whose foundation was subsequently uncovered. It had a hypocaust and concrete walls faced on the outside with Kentish ragstone. The remains of sinks/basins found inside led to it being interpreted as a bathhouse, 63ft long by 29ft wide. Its northern end was 27ft from the south bank of the Loose Stream. Coins from the field date widely, from Claudius (41-54) to Valens (364-5). Subsequent work in the surrounding area found patches of burning associated with urns that were interpreted as funeral pyres (Roach Smith 1842, 414-20).

More evidence for Roman burials in the area was provided in 1842 when Clement Taylor Smythe recorded a walled cemetery 700m to the north-west of the site, on the east side of the Roman road. His observations were published in 1883 (Scott Robertson 1883, 68-88). The cemetery was within a walled, rectangular area of 77 x 85ft. Funerary monuments (described as tombs: a square tower and a circular tower) and seven cremation burials (four with grave goods) were recorded. Two of the burials were in square, stone cists, while four were in what were described as amphorae/urns. Dating evidence was not discussed beyond the definition as Roman, but further archaeological work at the cemetery site in 1998 found pottery dated overall to the early second to mid third centuries and cremation vessels of Kentish types (Mackinder 2000).

Finally the site lies in an important area for the quarrying of ragstone, a hard limestone, which was exported during the Roman period to London and widely used, including public buildings such as the city wall, and also locally as at the walled cemetery.

Overview of the excavations on the site

Approximately 500 features were excavated on the East Field site. The most obvious were the two enclosures: earlier enclosure (S1) and later enclosure (S3) (Fig. 3). These two features have defined two of the periods described here, with a third period defined as later. Thus, because of the limited stratigraphic data, the defined building and structures on the site have been assigned to each period on the basis of how they relate to each enclosure; whether by stratigraphy, relative dating or best fitting interpretation. Most of the 500 features could not be assigned to any period as they were undated; either by the absence of dating evidence or by not forming part of group defined by dating evidence in the other features. Many features, particularly pits and postholes, containing dateable pottery were also not assigned to periods as they were isolated and, it was decided, the evidence was too unsecure to be reliable.

Generally all discrete features were 100% excavated, with the larger pits being 50% excavated. Linear features such as ditches and gullies were a minimum of 10% excavated.

Dating was mostly established by pottery. In the case of the local in-digenous wares the fabric codes were provided by the Canterbury Arch-aeological Trust (CAT) and the form types defined by Thompson (1982).

Pre-late Iron Age activity

Prehistoric flints dating from the later Mesolithic to Neolithic were found on the site; all presumed residual in later features. The only significant prehistoric find was an arrowhead and two end scrapers that were delib-erately buried in a posthole with the arrow point uppermost (Fig. 4). Other possible prehistoric features might exist amongst a series of winding ditches, generally scattered across the whole site. They tended to be quite shallow and did not contain any dateable material. These are now thought to be the result of natural erosion processes.

Possible ritual deposition of flint tools by Jon Cotton

Three flint tools were found, apparently placed upright, in the fill of a posthole feature (Fig. 4). The feature measured 0.14m in diameter and 0.12m in depth and was located just over 10m south-west of Open Area 1. These consisted of a leaf arrowhead, a burin worked on the proximal end of a robust blade, and a narrow, parallel-sided blade:

1. Leaf arrowhead: L 27mm; W 17mm; T 2mm; Wt 1.72g; worked on the distal end of a flake of speckled translucent grey brown flint, part of whose ventral surface survives unretouched. The dorsal surface is covered by shallow invasive radial retouch. The arrowhead appears to have been deliberately modified by two shallow notches.

2. Burin: L 62mm; W 20mm; T 12mm; Wt 15.73g; worked on the proximal end of a robust blade of good quality lustrous opaque grey brown flint whose narrowing distal tip has been snapped off, perhaps deliberately. The proximal end of the blade has been prepared to receive the burin through the removal of the bulb of percussion and the original striking platform.

3. Blade: L 76mm; W 16mm; T 5mm; Wt 8.72g; of good quality lustrous speckled grey brown semi-translucent flint. It has been soft hammer struck from a core with a deliberately abraded platform edge and has parallel, feather sharp, lateral edges.

Given the circumstances – the fresh condition of the pieces, their shared upright position within the feature and ‘close proximity’ to one another, the apparent detachment of two of them (nos 2 and 3) from the same core, and the presence of two finished forms, one purposely if idio-syncratically modified – it is possible that these three pieces comprise a small, deliberately placed, cache. Furthermore the presence of the leaf arrowhead indicates that the deposit is likely to have been buried during (or after) the earlier Neolithic.

Caches of Neolithic stone tools more usually comprise axes of various forms, and examples have been recorded within the county, as at Pembury ten miles to the south-west (Tester 1951). The intention behind the deposition of such caches is of course open to question: do they mark a necessarily ‘ritual’ act, or was there some utilitarian purpose?

Notwithstanding the caveats expressed by Saville (2002, 91-3) regarding the over enthusiastic interpretation of often low resolution lithic evidence recovered from Neolithic sites such as causewayed enclosures, this group is interpreted as the depositional signature of a non-utilitarian act, albeit one more likely to have had a modest personal motivation than any wider communal significance. As such, it provides a rare glimpse of the actions of an individual preserved within the archaeological record.

The earlier enclosure (S1): ad 10-50

Structure 1 comprised a rectilinear arrangement of ditches that would have enclosed an area 210m by at least 95m (Fig. 3). The full dimensions are unknown as it continued beyond the north-eastern limit of the excavation. Approximately 30m from the south-eastern corner, a 20m length of ditch, appears to form a partition. Presumably this section of the enclosure joined the main enclosure, just beyond the south-western limit of excavation. At its north-eastern end it returned to the north-west, but was truncated at this point by the later enclosure (S3) and was not seen elsewhere. Alternatively the most eastern length of ditch may have been part of an extension.

In general, the ditch was up to 1m wide, and varied in depth from 100-800mm. The profile, where apparent, was ‘V’-shaped. It was filled with a compact mid greyish brown clayey silt with local stone inclusions, but there was no direct evidence for a bank or wall.

The earlier enclosure ditch (S1) produced a pottery assemblage of 166 sherds. The majority of the assemblage (99% by sherd count) comprised Late Iron Age fabrics. Dismissing the two sherds in Roman fabrics as intrusive, then the date of the ditch would be ad 10-25 based on a combination of grog-tempered fabrics (CB1, CB2.3 and CB24). Of the late Iron Age fabrics, two vessel forms were identified. These accounted for over half the assemblage by sherd count (60%). Both vessel forms are plain everted-rim necked jars (form B1-1); one is in a Belgic coarse grog-tempered (pale grog) ware (fabric B2.1); and the other is in a Belgic grog- and shell-tempered ware, fabric B2.3 (Fig. 5, <P1> and <P2>).

Other than pottery, a stone hone <4>, a spindle whorl <6>, an iron knife <7> and an iron reaping-hook <12> were recovered from the ditch. The spindle whorl was manufactured from a pottery sherd. The stone hone was roughly rectangular with a flat surface, and was worn thin across its centre as a result of use. The iron finds both came from the most eastern length of ditch. All these finds indicate some form of industry on the site.

Bloomery (S2)

Structure 2 was interpreted as a bloomery, a type of furnace once widely used for smelting iron from iron ore bearing rocks. It comprised the truncated remains of the ragstone footings of two parallel walls, with the possibility of a third footing, or stokehole, located to the north-west (Fig. 6). This structure measured c.7 x 6m (including the possible third footing). It was originally thought this structure was a corn drier, with the walls supporting a raised floor allowing hot air to pass beneath. However, analysis of the samples from Structure 2 did not support this initial interpretation. The masonry was almost entirely robbed and none of the above ground structure survived (Fig. 7).

The interpretation here is derived from small amounts of highly mag-netic spheroidal hammerscale that were recovered from sieved samples. Spheroidal hammerscale is associated with smelting sites (Bayley et al. 2001, 14). It is a by-product from the bloom, a porous mass of iron and slag used in the early process of making wrought iron. There may have been other bloomeries on the site, or in the vicinity, as spheroidal hammerscale was noted in other features across the site.

Structure 2 contained an assemblage of 92 sherds, presumably all dumped after the period of use. The assemblage comprised primarily 1st- century ad Iron Age and Roman fabrics, approximately 88% by sherd count. With the exception of eleven later Roman sherds, which can be considered intrusive because of the disturbed nature of the upper de-posits, the dating for the disuse of the furnace, based on these fabrics alone, would be ad 45-70, possibly 45-50.

The possible stokehole contained burnt daub and wattle. From its infill, there were 138 sherds from only three fabrics, all of Late Iron Age/Early Roman date. The assemblage is dominated by fabric B9.1, representing approximately 94% of this assemblage by sherd count. Only three vessel forms are represented; two sherds were from one small plain everted-rim jar, form C2-2 in fabric B9.1 (75 bc-ad 50) and 121 sherds came from two plain everted-rim necked jars, in form B1-1, also in fabric B9.1 (see Fig. 5, <P3>). Therefore dating for this pit would be 75 bc-ad 50.

Bloomeries were a common form of smelting furnace up to the end of the medieval period in England. The resources for iron production on the site would have been bountiful with woodland for the production of charcoal in the vicinity and iron ore bearing rocks from the Weald to the south , although iron from glauconite in the local geology and ‘bog iron’ cannot be ruled out.

Evidence for kiln activity on the site or the vicinity (OA1)

Evidence for a kiln came from a group of features (Fig. 8) approximately 17m to the north-west of bloomery (S2). These included a sub-rectangular pit [498] that measured 1.15 x 0.85m and 0.50m deep. It was vertically sided with a flat base. The lower fills of the pit were formed of lenses of debris including frequent large sherds of pottery.

This assemblage contained 737 sherds from only four fabrics, all of Late Iron Age/Early Roman date. The assemblage is dominated by an Iron Age/Belgic coarse sandy fabric with glauconite (fabric B9.1), representing 86.7% of the assemblage by sherd count. The other three fabrics represented in the pit are B1, B3 and B5 at 7.5%, 0.3% and 5.5% respectively. The assemblage contained only two identifiable vessel forms (see Fig. 8); C3, a plain jar <P4>, with no true external rim and unusual internal thickening, and B1-1, a plain everted-rim necked jar <P5>. The form B1-1 represents 91% of the form assemblage (556 sherds/62 estimated vessels). It was found exclusively in B9.1 thus dating to 75 bc-ad 50. The form C3 was in fabric CB1 and thus dated to ad 10-100. These sherds were much burnt and possibly represented waste from a kiln. Only a relatively small portion of the form B1-1 was burnt.

Burnt clay with wattle impressions was also found in this pit, which suggests there was a wattle and daub structure, perhaps a fence around a hearth/kiln. This could be represented by the curving gully, located c.8m to the north-east of the pit, which was dated ad 45-100 by fabric B8. This later fabric may represent a later disuse of the possible fence.

Other finds suggest other industrial activity. Two rectangular-shaped blocks of Kentish ragstone were also recovered from the pit. One has two smoothed sides and a smoothed top, and the other a smoothed edge. These may have been used as whetstones. Also a piece of furnace slag with fuel impressions was found. The sparkly ‘pyritic’ structure visible over part of the surface is thought to be because of a prolonged cooling period after the smelting process (David Dungworth pers. comm.) and suggests, as with the hammerscale, that the blooming process seen at Structure 2 may have occurred across the Late Iron Age enclosure.

Two other pits (Fig. 8) may possibly be associated with the rectangular pit. The pottery recovered from the fills is composed of four late Iron Age/Belgic fabrics and one Belgic/early Roman fabric, all common to the area. The most common fabric is a coarse sandy fabric with added glauconite (B9.1), which dates 75 bc-ad 50. The next most common fabric was B5.1, Belgic fine grog-tempered fabric with added fine sand, which dates to 75 bc-ad 120.

The identification of fabrics B8 and B9.1 within the features would give a date of ad 45-50; however, this is too fine a dating for the quantity and type of pottery but it does suggest that this structure went out of use towards the middle of the 1st century ad. The other two fabrics within this feature were B1.1, a Belgic fine/coarse grog-tempered fabric dating to 75 bc-ad 120, and B24, a Belgic grog-tempered fabric with sparse flint and coarse sand, dating to 75 bc-ad 25.

Individually the four features only suggest kiln activity on the site, but together they indicate a location for a kiln, possibly lost to ploughing. Other kiln sites may have existed on the site as suggested by the pottery found in the infill of the possible stokehole in Structure 2. It, too, produced predominately wares in fabric B9.1.

Cremations (OA2)

So far there has been no direct evidence of any settlement, unless the curving gully within Open Area 1 is interpreted as part of a round-house. Although not necessarily the remains of those living in the area, two cremations were found to the east of the site; outside of the earlier enclosure (S1): a single urned cremation pit [184], contained 39.5g of human bone and an urned cremation pit [189], contained 94.5g of human bone, with accessory vessel. The robust nature and overall proportions of the bone fragments suggest that the remains were probably adult, but the fragmentary nature of the remains and the small quantities involved meant there could be no further analysis.

The burial vessel found with [184] comprised 337 sherds, weighing 201gm, from a flagon in Hoo white-slipped ware which dates to ad 50-100. Flagons are a common burial vessel both as an urn and as an access-ory vessel. No rim survived, only body sherds, so it was impossible to de-termine a more accurate date. The fabric is common in Kent, coming from one of the three major early pottery production areas; Canterbury, Eccles and Hoo. The other burial [189], contained 76 sherds, weighing 578gm, from two vessels; a tall barrel jar with bead rims and cordons on upper body (form B5-3) in B1.1 dating to 50 bc-ad 50, and the base of another jar also in fabric B1.1, dating to 75 bc-ad 120. The barrel jar is a common vessel throughout the south-east and east of England during the late Iron Age and appears to have been influenced by European butt-beaker styles. Only a small portion of the rim survives. It is likely that this was the burial urn and the base was part of an accessory vessel. This burial could date anywhere between 50 bc-ad 50 but given the dating of the features on the site it is more likely that this burial dates to the first half of the 1st century ad rather than the second half of the 1st century bc. The dating of both burials points to the continuity of activity but because of the limitations of osteological data it is impossible to see how these burials relate to each other. Are they related by familial ties; are they unrelated individuals of different family groups within one larger settlement or are they totally unrelated individuals drawn to site as workers?

The evidence of small quantities of burnt human bone in three pit fills and the presence of burnt earth, in the area replaced by a Building 3 sometime after ad 120, may support the interpretation of pyre clearance debris.

Period discussion

The period associated with the use of Structure1 has been dated by pottery to the pre-invasion period, i.e. 10 bc-ad 25/50, although it is entirely possible that the use of some or all of the associated features could have continued for some time after the Claudian invasion.

One find from this period stands out from all the others, despite being found in the infill of a posthole of the later structure (S4). A silver coin <1>, issued by Amminus, is the second of his coins found in the area. Amminus was ruler of the Cantiaci of eastern Kent, a kingdom which presumably fell within his father’s sphere of influence. Suetonius tells us he was deposed and exiled by his father c.ad 40 (see below).

A Silver Coin of Amminus By Julian Bowsher

The only coin was found in the excavations was identified as a silver issue of Amminus (Mack 1975, 313; Van Arsdell 1989, 192-1). This was found in context [1029], part of stock-handling structure (S4), dated to the period immediately after the Claudian invasion. The coin dates to the second half of the ad 30s and may possibly have been circulating. The left half of obverse legend is mostly off flan, but it is otherwise in good condition.

<1>, [1029], see Fig. 9.

Amminus, c.ad 36-39, silver piece; Diam. 13x11, Wt. 1.03g, Ax. 8. Wear A.

Obv AMM [INVS], Ears of corn. Rev. DVN, winged horse to r.

Amminus is known only from a passage from Suetonius’ Life of Gaius (Caligula) as ‘Adminius, the son of Cunobeline’ who presented himself, and a small body of troops, to Gaius after being driven from his native country by his father. Gaius fancifully thought that the whole of Britain was being surrendered to him! (44.2).

It is only after examining a series of coins, silver and bronze, with similar characteristics, that the late Derek Allen asked ‘Did Adminius strike coins?’ (Allen 1976). These coins, including our example as Type 3 (Allen 1976, 96), found from the 1840s to the 1960s, had a wide provenance. They were found as far afield as Jersey, Dorset, Lancashire and Kent (including a bronze specimen from Boughton Monchelsea) and indeed, stylistically, of Kentish association. They were also similar to those of Cunobelin, the king of the Trinovantes whose realm certainly extended into Kent. Allen proposed therefore, with other historical sources, that these coins were associated with Amminus who had been given Kent to govern by his father. After further discoveries, again mostly from Kent, Daphne Nash retorted that ‘Adminius did strike coins’ (Nash 1982) and another article suggested that Amminus’ control extended into Sussex (Henig and Nash 1982).

With the advent of the Portable Antiquities Scheme database of finds, it can be seen that Kent has provided more Amminus coins than anywhere else. 71 coins of Amminus are recorded with the largest number from Kent and indeed, the largest number of this type, from that county too.http://finds.org.uk/ironagecoins/rulers/ruler/id/23 accessed 21/10/2011. Not all coins are illustrated or identified so that exact numbers are not fully ascertained at present.] Amminus minted in silver and bronze only as would befit his junior status, gold coins of Cunobelin circulating in Kent represented the supreme power (Nash 1982, 113). The Dun on the reverse of the coin might suggest a place of minting but this suffix meaning fort, is common to many ‘Celtic’ settlements. An identification with the defended enclosure at Quarry Bank as the origin of these coins remains wishful thinking until further research and excavation in the area is undertaken. At present a site much further east in Kent seems most likely (David Holman, pers. comm.).

Whether this site lay within an oppidum under the direct control of Amminus is conjecture but the enclosure appears to be associated with industry, chief among which is iron smelting. Settlement is suggested by the cremations although there no clear evidence of buildings.

The later enclosure (S3) ad 50-160

In an apparent replacement of the earlier enclosure (S1), the later enclosure (S3) appears remarkably similar (Fig. 3). Structure 3 appears to have enclosed a smaller area, but with much more substantial ditches. The ditches were 3.0 to 3.50m wide and up to 1.50m deep; the dimensions could suggest this was a defensive enclosure. The ditch formed an enclosure 123m in length and over 46m wide. Its full dimensions are unknown as it continued to the north-east beyond the limits of the excavation.

At some stage the enclosed area was increased. The original ditch on the south-eastern side was infilled and a new ditch a further 14.0m to the south-east was dug. This ditch was not continuous; a gap towards the north-east may have been an entranceway. The exact date at which the eastern extension was added to the enclosure is not known; the pottery recovered was dated ad 100-20. Some of the sections excavated across this ditch showed evidence of re-cutting, dated slightly later by pottery dated to ad 140-200.

The total assemblage of pottery from the infill of the ditch comprised 1,360 sherds, a marked increase in what was recovered from the earlier enclosure ditch. The most common fabric found within the ditch was fabric B1.1, a Belgic grog-tempered ware that dates to 75 bc-ad 120, representing 35% of the assemblage by sherd count. Interestingly, fabric R16, a product of the Upchurch kilns in north Kent, represented approximately 9% of the assemblage by sherd count, and was the most common of all the fabrics that could be ascribed to an industry indicating there is now Roman activity on the site. Also indicating Roman activity is the appearance of a number of 2nd-century ad fabrics such as central Gaulish samian (SAMCG) and black-burnished ware 2, both fabrics dating from ad 120.

The ditch infill also contained the largest array of forms. Of particular note within the Roman form assemblage are sherds of black-burnished-type round-rimmed bowl (Marsh and Tyers 1978, fig. 242, nos IVH1-7) in black-burnished ware 2 with acute lattice decoration (ad 120-250) (see Fig. 10, <P6>), sherds from two bead and flange mortaria in a pink buff fine/coarse sandy fabric (ad 140-200) (Fig. 12, <P7>), and sherds of a Dragendorff form 18/31 dish with rouletted decoration in SAMCG (ad 120-50). It is clear from these vessels that this larger ditch went out of use c.140-50/60.

Other finds include an iron stylus <2>, roman vessel glass <3> and <16>, iron knife <11> and stamped samian sherd <27>. Together with the pottery evidence it is clear that between the disuse of the earlier enclosure (S1) and the disuse of the later enclosure (S3) the material culture of the site has shifted from Iron Age/Belgic to Roman.

Stock-handling structure (S4)

Two parallel, straight rows of postholes were recorded over a length of 31.40m (Fig. 11). They were 5.50m apart with a series of postholes at the northern end that suggest a terminus. The southern end was bounded by the ditch of Structure 3. These proportions are unlikely to suggest the remains of a building. Therefore a stock-handling structure is suggested, despite no known parallels in Kent.

The excavation of Structure 4 yielded 107 sherds of pottery in a wide range of fabrics. However, nearly three-quarters of the assemblage (71% by shred count) are Late Iron Age fabrics and are considered residual, as is common in posthole structures. The Roman assemblage had a range of unsourced wares common to the whole Roman period. However several wares could be more closely dated. One sherd of Verulamium Region white ware, dating ad 50-160, was identified along with seven sherds of different fabrics from the Upchurch industries, which date to c.ad 70-275. However, one sherd of Hoo Island ware was also identified, this dates to ad 50-100. The range of forms present is of little help in trying to narrow the dating of this structure. Only four identifiable forms were recovered: two Late Iron Age forms and two that are common to both periods.

Ultimately, Structure 4 has been dated by its association with later enclosure (S3). Structure 4 aligns with both earlier enclosure (S1) and the later enclosure but no postholes were identified between the two sections indicating the later structure bounded its limits. Of note is the alignment of ragstone-foundation building (B2) whose western end is superimposed on part of the eastern alignment of Structure 4 and perpendicular to it (see below).

Aisled building (B1)

Building 1 is located over the corner of the partition ditch of the earlier Structure 1 (Fig. 12); hence it is placed in the period after the disuse of Structure 1. However, it is possible that it is contemporary with the stone buildings in the next period. Building 1 is an aisled building with six large postholes on each side, with smaller postholes forming the outer walls on three sides. Some of these were definitely angled, suggesting they were intended to support inner walls. This building was 17m long by12.60m wide with access from the north-west side. The nave is 7.20m wide and was possibly subdivided as it had four small postholes along its length.

Building 1 has a greater quantity of Roman fabrics than any other build-ing. Roman fabrics represent approximately 57% of the assemblage of 60 sherds. It is interesting to note that even though Roman fabrics formed the larger part of the assemblage, there were no identifiable Roman forms. The forms identified related to a plain everted-rim jar (form C2-3) in fabric B1.1, dating 75 bc-ad 50 (see Fig. 10, <P8>), and a storage jar with rolled rim and decoration on the shoulder (Marsh and Tyers 1978, fig. 237, IIM) in fabric B2.3 (possibly Patchgrove), dating to ad 60-100 (<P9>).

Parallels to this building can be found c. 5 miles to the north-east at Thurnham, where two aisled buildings were excavated close to a villa house (Millet 2007, 154-5). One was associated with agricultural use as it contained a possible corn-drying oven, but the other closer to the villa house appeared to have a domestic function.

Post and sill-beam building (S5)

Structure 5 was bounded by the south-eastern corner of Structure 3, within its possible extension (Fig. 12). It was constructed of postholes and sill-beams and measured 11.50m long x 8.40m wide. Only one sherd of a local Romanised miscellaneous coarse sandy ware (fabric R73) was recovered, giving a general date of ad 50-400. Therefore the structure is dated by its relationship to the ditch and a date of ad 100-160 is suggested. The structure may be all that remains of a palisade, but the lack of evidence makes any further interpretation difficult.

Period discussion

The period associated with the use of the later enclosure (S3) represents the site in the early Roman period. Gone is the evidence for ‘heavy’ industry and the emphasis appears to be on agriculture, possibly pastoral. Larger quantities of pottery may be a reflection of increased supply rather than increased population. Building 1 could be a domestic residence, while the remains of Structure 5 are too insubstantial for any valid interpretation.

The Roman road, known to the south-west of the site (Fig. 2), was probably founded during this period, and Structure 3 appears to be aligned with this road. However, Structure 3 was also aligned with the Late Iron Age Structure 1. This suggests a continuity similar to Hunts Hill Farm in Essex, where the Late Iron Age defensive enclosure went out of use, but the fields continued, their alignment preserved with the similar alignments of the ‘new’ Roman fields and trackways (Howell et al. 2010, 72-3).

Roman farmstead buildings ad 160-270+

Ragstone-foundation building (B2)

Building 2, located in the centre of the site, was rectangular and had well-set ragstone foundations for walls, and external buttresses on the north-west, south-east and south-west sides (Fig. 13). It measured 18.0m long x 7.50m wide. The buttresses suggest this building would have been substantial, possibly two-storey, although it is unclear whether the superstructure would have been masonry or timber.

Only 32 sherds of pottery were collected from Building 2. The reason for the small quantity was the result of the building’s previous excavation between 1970 and 1974 by the Lower Medway Archaeological Research Group. The results of the excavation were similar to the present excavation, with the additional evidence for a surrounding yard. The group also reported finding two ‘bronze’ coins, dated ad 270, which is similar to a copper-alloy coin <23> found during the present excavations, and dated 270-73. Curiously the majority of the Roman pottery on both excavations was 1st-century ad with a latest date of 175, although at present the dating of the pottery from the 1970s excavation may be misleading.

Other finds included square-shaped blocks of typical tessera size but with no definite signs of wear or of mortar attached. If they are tesserae then they presumably come from a plain red tessellated pavement.

Second aisled building (B3)

Building 3 had two parallel rows of eight postholes on each side and measured 17.0m long x 7.4m wide (Fig. 14). Rubble infill in the adjacent ditch (S3) may be the remains of a sill-beam for an outer wall. This would make the building around 12.50m wide.

50 sherds of pottery were recovered from Building 3, split approximately two-thirds Iron Age and one-third Roman. This included a sherd of central Gaulish samian and one of black-burnished ware 2, both of which date to ad 120-250. This date is further supported by the evidence that the eastern return ditch of Structure 3 had been infilled after ad 100-120.

Although dated later than Building 1, by the pottery, the evidence of a possible outer wall suggests they are similar buildings and again draws parallels to the aisled buildings at Thurnham (Millet 2007, 154-5). Although the evidence for Building 3 being aisled is not as strong as Building 1, the similarity of the ‘central’ postholes to the Thurnham aisled buildings and the date suggest they are comparable.

Flint-foundation building (B4)

Building 4 was located to the north-west of the other two buildings. It was probably a masonry building (Fig. 15) with flint foundations measuring c.13.20 x 14.20m and continuing towards the north-east beyond the limit of excavation. There is no obvious dating as only eight sherds were recovered of which only one was in a Roman fabric, R8.2, a ‘catch-all’ fabric representing an unsourced fine red sandy fabric found throughout the whole Roman period. However this report places it in this period as masonry buildings on lesser villa sites, like Keston (Philp et al. 1999, 190-4) and Beddington (Howell 2005, 28) do not generally occur until the late 2nd-century ad.

Period discussion

This period is marked by the occurrence of masonry buildings, albeit their foundations, in the second half of the 2nd-century ad. This date is broadly compatible with the establishment of the major Roman features in the area; the bathhouse to the south-west of the site and the walled cemetery to the north-west. Clearly these buildings are not just part of a Romanised landscape but the expression of an established Roman lifestyle. It maybe that the infilling of the later enclosure (S3) at the end of the preceding period meant that agriculture was not the main occupation of the area. It is disappointing that the current excavations appear to have contributed little to understanding the important ragstone quarrying industry in the immediate area.

Discussion of the pottery by Rupert Featherby

The fabrics identified are all quite common to the Late Iron Age/Belgic and the Roman periods in west Kent. The most common fabric within the whole assemblage is B9.1, a coarse sandy ware with glauconite, at 25.3% by sherd count. The dominance of this fabric is expected given the proximity of the site to the Medway Valley, where glauconite tempered wares are common. The dominance of grog-tempered over flint tempered wares would indicate ceramic traditions common in west Kent (Pollard 1988, 30-46). Roman wares in general represent approximately 22% by sherd count, with reduced wares being surprising low at 2.5%, suggesting that throughout the period of activity at this site local traditions dom-inated.

The limited range of early Roman fabrics conforms to what is known about assemblages from the 1st and 2nd centuries ad in the area, especially the complete lack of Alice Holt/Surrey, which dates to ad 50-160 (Pollard 1988). Late Roman fabrics are absent for the assemblage indicating that by ad 250 significant activity at this site had ceased. Imported wares represent only 1.6% of the assemblage by sherd count, which, despite the site’s closeness to the Medway, is unsurprising. Imported wares during the pre-invasion period tended to be represented by products of Gallo-Belgic derivation, i.e. imported from Gallis Belgica, or from Colchester. Post-invasion fine wares remained relatively scarce with products from central and south Gaul remaining of limited distribution (Pollard 1988, 36-9). Samian comprises the majority of imported wares at 1.1%, with central Gaulish samian, dating ad 120-250, being the most common. Only four sherds of amphora were identified and represented only 0.05% of the assemblage by sherd count. The low level of imported fabrics suggests a community little affected by the increasing trade with the continent both pre- and post-invasion. The assemblage is very much in line with what would be expected for an isolated community. However the site lies within a known pre-invasion site and close to a main arterial Roman road, therefore making the assemblage peculiar in this respect.

Conclusions

In 2007, Timothy Champion said of Quarry Wood that ‘there is nothing to suggest continuity into the first century ad’ (Champion 2007, 121). This article does not necessarily contradict this statement; however, the evidence from the present excavations does show that there was Iron Age to Roman continuity adjacent to the neighbouring Roman road. Margary suggested that this road, which passes on a north-west – south-east alignment c.200m to the west of the site, was associated with industrial sites in the Weald, and primarily laid out to link areas producing iron (Margary 1967 edn, 34, 44), i.e. post-ad 43. The evidence of the bloomery (S2) indicates the site in the Late Iron Age was not just handling the iron trade but was an active part. It is possible that the absence of this industry on the site post-invasion was the result of political change that controlled the production of iron. Whether this political change meant that the defended enclosure at Quarry Wood ceased to control the area, or even went out of use, does not change the fact that the Roman road, an early expression of Romanisation, must have followed an alignment already defined by the layout of the Late Iron Age landscape.

A similar Late Iron Age site, with a bloomery and six cremations, may have existed c.7km to the south-east of the present site (Aldridge 2005). This second site also appears associated with a Roman road (Aldridge 2006), suggesting the road network in this area was intended to link a number of established rural ‘native’ settlements.

The replacement of earlier enclosure (S1) with later enclosure (S3) post-invasion does not necessarily mean the post-Iron Age focus may have shifted east from Quarry Wood, in response to the new infrastructure. There may not have been a full-scale, Rochester-type Roman town in the vicinity of Quarry Wood, but, as demonstrated here, there certainly was immediate post-invasion use. The possible change to an agricultural landscape, suggested here, may reflect the broader development of the area but with no direct evidence of crop or livestock production, this may also reflect the isolation of the area as indicated by the analysis of the pottery.

The development of more substantial buildings on the site, probably post-ad 160, is also a probable reflection of the broader landscape; not just from the evidence of the nearby bathhouse and walled cemetery, but also occurring at other probable villa sites in Maidstone. Maidstone is thought to have been a dispersed and wealthy agricultural settlement throughout the Roman period (Clark and Murfin 1995; Pollard 1988) as indicated by substantial Roman period buildings at The Mount (Houliston 1999), Barton Lane (Clark and Murfin 1995, 12) and Bower Lane (Edwards 2007, 79). In the surrounding area other substantial Roman period buildings have been identified at Teston (VCH 1932, 125), East Barming (VCH 1932, 104), and Chart Sutton (Pollard 1988). In fact the Medway Valley, north of the Weald appears to be a favoured location in the Roman period, like the Darent Valley to the west.

The pottery suggests that all activity on the site had ceased by ad 250. However the coins from Building 2 suggest occupation continued after ad 270, and the neighbouring bathhouse produced a coin of Valens (ad 364-5). It is unclear why the limited coin evidence should suggest continued use of the site later than the date suggested by the pottery.

acknowledgements

MOLA would like to thank Taylor Wimpey South London Ltd who generously funded the archaeological work, in particular James Brett, David Bishop and Paul Cheeseman. Wendy Rogers and Adam Single the County Archaeological Officers from Kent County Council Heritage who monitored the site.

The field work was supervised by Kevin Appleton with the subsequent assessment by Tony Mackinder and analysis by Kieron Tyler. The text provided by the assessment and analysis was edited by Isca Howell to form this article. Specialist input came from Rupert Featherby (pottery), Julian Bowsher (coin), Nicola Powell (accessioned finds and bulk glass), Natasha Powers (cremated bone), Tony Grey and Jon Cotton (worked flint), Lyn Blackmore and Lynne Keys (slag), John Giorgi (plant remains) and Alan Pipe (animal bone). Louise Fowler is duly acknowledged for help with the identification of the silver coin.

The illustrations were prepared by Judit Peresztegi, including original drawings of the pottery and flint tools by Hannah Faux, and digital site graphics supplied by MOLA’s geomatics team, in particular Mark Burch.

The following field staff were involved; Carla Banks, Ryz Bartkowiak, Aaron Birchenough, Agnieszka Bystron, Peter Cardiff, Vince Cherubini, Elaine Eastbury, Bruce Ferguson, Val Griggs, David Harrison, Simon Stevens, Kate Stevens. Survey support was by Dave Mackie and Cordelia Hall. Stewart Hoad managed the field project, with post-excavation management by Nicola Powell.

Special mention is accorded to Steve Dockrill (senior lecturer in archaeology at the University of Bradford) who, as a schoolboy, helped on LMARG’s excavations at the site in the 1970s. As the original site records were lost, Mr Dockrill kindly supplied plans and notes that will be incorporated into the present site archive.

bibliography

Aldridge, N., 2005, ‘A Belgic cremation cemetery and iron bloomery furnace at Jubilee Corner, Ulcombe’, Archaeologia Cantiana, 125, 173-82.

Aldridge, N., 2006, ‘The Roman road from Sutton Valence to Ashford: evidence for an alternative route to that proposed by Margary’, Archaeologia Cantiana, 126, 175-83.

Allen, D.F., 1976, ‘Did Adminius strike coins?’, Britannia, 7, 96-100.

Arsdell, R.D. van, 1989 Celtic Coinage of Britain, London.

Bayley, J., Dungworth, D. and Paynter, S., 2001, Archaeometallurgy, Centre for Archaeology Guidelines 2001/01, London.

Champion, T., 2007, ‘Prehistoric Kent’, in Williams, J. (ed.), The Archaeology of Kent to AD 800, Rochester, NY.

Clark, P. and Murfin, L., 1995, The History of Maidstone, Stroud.

Cunliffe, B.W., 1982, ‘Social and economic development in Kent in the pre-Roman Iron Age’, in Archaeology in Kent to AD 1500 (ed. P.E. Leach), CBA Res. Rep., 48, 40-50, London.

Cunliffe, B.W., 2005, Iron Age communities in Britain, 4th edn, London.

Detsicas, A., 1983, The Cantiaci, Gloucester.

Edwards, C., 2007, ‘Excavations at Fremlin Walk, Maidstone’, Archaeologia Cantiana, 127, 73-106.

Elsden, N., 2006, ‘Furfield Quarry, Boughton Monchelsea, Maidstone, Kent, Earthworks assessment’, unpubl. MoL report.

Henig, M. and Nash, D., 1982, ‘Amminus and the kingdom of Verica’, Oxford J. Archaeol., 1, 243-246.

Houliston, M., 1999, ‘Excavations at the Mount Roman Villa, Maidstone’, Archaeologia Cantiana, 119, 71-172.

Howell, I. (ed.), 2005, Prehistoric landscape to Roman villa: excavations at Beddington 1981-7, MoLAS Monogr. Ser. 26, London.

Howell, I., Swift, D. and Watson, B., 2011, Archaeological landscapes of east London: six multi-period sites excavated in advance of gravel quarrying in London Borough of Havering, MoLAS Monogr. Ser. 54, London.

KCC, 2005a, Specification for an archaeological strip, map and record of ditto.

KCC, 2005b, Specification for an archaeological watching brief of the West Field, Furfield Quarry, Boughton Monchelsea, Maidstone, Kent.

Kelly, D.B., 1972, ‘Quarry Wood Camp, Loose: A Belgic Oppidum’, Archaeologia Cantiana, 86, 55-84.

Mack, R.P., 1975, The coinage of Iron Age Britain (3rd edn), London.

Mackinder, T., 2000, ‘Land adjacent to Pested Bars Road, Boughton Monchelsea, Maidstone, Kent’, a post-excavation assessment and updated project design, unpubl. MoL rep.

Marsh, G. and Tyers, P.A., 1978, ‘The Roman pottery from Southwark’, in Southwark excavations 1972-4 (eds J. Bird, A.H. Graham, H. Sheldon and P. Townend) (2 vols), Joint Publ. LAMAS/Surrey Archaeol. Soc. 1, 533-86, London.

Millet, M., 2007, ‘Roman Kent’, in Williams, J. (ed.), The Archaeology of Kent to AD 800, Rochester NY.

Nash, D., 1982, ‘Adminius did strike coins’, Oxford J. Archaeol., 1, 111-114.

Philp, B.J., Parfitt, K., Willson, J. and Williams, W., 1999, The Roman villa site at Keston, Kent: second report (excavations 1967 and 1979-90), Kent Monogr. Ser. Res. Rep. 8, Dover.

Philp, B, 1996, ‘Furfield Quarry, Maidstone, Kent, report on a Programme of Archaeological Evaluation’, KARU unpubl. [KCC rep no. 1996/81.]

Pollard, R. J., 1988, The Roman pottery of Kent, KAS Monogr. Series V, Maidstone (ed, A.P. Detsicas).

Roach Smith, C., 1842, ‘Account of various Roman remains discovered in a field called the Slade, in the Parish of Boughton Monchelsea, in Kent’, Archaeologia, 29, 414-20.

Saville, A., 2002, ‘Lithic Artefacts from Neolithic Causewayed Enclosures: Character and Meaning’, in Varndell, G. and Topping, P. (eds), Enclosures in Neolithic Europe: Essays on Causewayed and Non-Causewayed Sites, 91-105, Oxford.

Scott Robertson, W.A., 1883, ‘Traces of Roman occupation in and near Maidstone’, Archaeologia Cantiana, 15, 68-88.

Tester, P.J., 1951, ‘Three associated Neolithic axes from Pembury’, Archaeologia Cantiana, 64, 57-62.

Thompson, I., 1982, Grog-tempered ‘Belgic’ Pottery of South-eastern England, BAR Brit. Ser. 108(i)-(iii), Oxford.

VCH, 1932, Kent, iii, London.

Fig. 1 Site location.