Northamptonshire Archaeology 2002, 30 ARCHAEOLOGY, 30,2002

Contents

Middle Iron Age occupation at Mawsley New Village, CransleyLodge, , Northamptonshire 1 GRAHAM HULL AND STEVE PRESTON

Excavation of Roman settlementat Sponne School, Towcester, 1997 21 ROB ATKINS AND ANDY CHAPMAN

Excavationsat Derngate, Northampton,1997-2000 31 JONATHAN HILLER, ALAN HARDY AND PAUL BLINKHORN

Excavationsat the Former Cantor and Silver Site, Brackley, Northamptonshire 63 JON MURRAY

Brickmakingin Holy Sepulchre Parish, Northampton 83 ROB ATKINS

Notes 101 Excavation of a "Triple-Ditch System" at The Larches, Stowe Nine Churches Martin Tingle 101 An Iron Age Site at Brafield Allotments: Pottery from Excavations in 1962 and 2001 Dennis Jackson and Martin Tingle 105 The Prebendal Manor Research Project, Nassington. Jane Baile 116 Northampton:the Double Streets and the Norman Town. T.C. Welsh 119 The Luffield Priory Grange at Monksbarn,Whittlebury, Northants. Richard Jones 126 Fotheringhay:a new Perspective from the 1640s. Glenn Foard and Tracey Britnell 140 Archaeology in Northamptonshire,2002 145 Errata Volume 29 155

Northamptonshire Archaeology 2002, 30 Middle Iron Age occupation at Mawsley New Village, Lodge, Kettering, Northamptonshire

by

GRAHAM HULLAND STEVE PRESTON

with contributions by Paul Blinithorn, Varian Denham,Steve Ford, Sheila Hamilton-Dyer, Lynne Keys, NicolaPowell and Mark Robinson

SUMMARY feeding the River Ise (to the east) and another stream (currently via Pitsford Reservoir) to the south west, A series of middle Iron Age circular structures both tributaries of the Nene. Geological maps (BGS comprising an unenclosed occupation site with at 1993) indicate that the geology is Boulder Clay and least twophases ofactivity, were radio-carbondated this was confirmed during the excavations. The of the site was estab- to aperiod between the 4th and the 1st centuriesBC. archaeological potential lished by a series of field surveys and evaluation There was some evidence of metal working on the site. A similar was revealed trenching carried out by John Samuels Archaeo- group of features by Consultants Slatcher within an to the logical (Nicholls 1999; 1999; geophysical survey enclosure, just 2000; Young 2000). As a result,an area of 6250sq m east the excavated site. A small of very quantity of was identified as archaeologically sensitive (Fig. 1) late Iron Age material was recoveredfrom features and a programmeof full excavation was instigated. unrelated to the occupation area, while a series of This followeda specification prepared by Mr Myk shallow linearfeatures, cutting across the site may Flitcroft, Planning Officer (Archaeology) for North- be ofIron Age date but are more likely the remainsof amptonshire Heritage, the archaeological adviser to medieval cultivation. the BoroughCouncil. The excavation wasfunded by David Wilson Homes. The archive is currently held by Thames Valley INTRODUCTION Archaeological Services Ltd, 47—49 De Beauvoir Road, Reading, RG1 5NR, pendingidentification of This report documents the results of an excavation a suitablerepository. The site code is MNVOO/64. carried Out by Thames Valley Archaeological Services Ltd at Housing Area 3B, Mawsley New Village, Cransley Lodge, Kettering, Northampton- ARCHAEOLOGICAL BACKGROUND shire. Outline planningpermission had been granted by Kettering Borough Council for the development The site lies at the head of a valley aligned SW—NE, of the area, subject to a condition requiring arch- draininginto the riverIse. Fieldwallcing had revealed aeological investigation. The whole development area a concentration of Neolithicfinds not far from New covered O3ha, south-west of the village of Cransley Lodge, and Bronze Age sites were known in the (centred on SP 807 760) (Fig. 1). The site lay on vicinity. The Northamptonshire Sitesand Monuments essentiallyflat land on top of a ridge at an average of Record contains evidencefor Iron Age and Roman c.140m above Ordnance Datum. The wider topog- finds from the parish, although not closely located. raphy slopes down sharplyboth to east and west in a Aerial photography mapped by the National Map- series of narrow valleys cut by numerous streams ping Programme of the Royal Commission on

Northamptonshire Archaeology 2002, 30 GRAHAM HULL AND STEVE PRESTON

Fig I Sitelocation.

2 Northamptonshire Archaeology 2002, 30 MIDDLE IRON AGE OCCUPATION AT MAWSLEY NEW VILLAGE, CRANSLEY LODGE, KETTERING, NORTHAMPTONSHIRE

Historical Monuments () suggested the THE EXCAVATION presence of what appeared to be hon Age and/or Roman enclosures and settlement both on the site The excavation took placeover the winterof 2000—1 and nearby. A deserted medieval village also lay under the supervision of GrahamHull and opened a close to the site. The site appears to have been single continuousarea of 6250sq m (Fig. 2). Topsoil farmland from late Roman times onwards. and overburden were removed by a 360° mechanical The site was identifiedby a desk-based assessment excavatorfitted with atoothless bucket to exposethe (Slatcher 1999) and geophysical survey (Nicholls uppermost surface of archaeological deposits. The 1999) as of high archaeological potential. In partic- archaeological deposits encountered included the ular, two clusters of circular and semi-circular feat- ring gullies, a series of straight, parallel ditches, a ures, one of them enclosed by ditches, appeared to few pits, and somestructural features(post holes and represent Iron Age or Roman settlement (Fig. Ic). post-pads). The vast majority of the features were The entire area of the site was then fleidwalked and ditches or gullies. All archaeological deposits were evaluation trenching was conducted, to ascertainthe cleanedand excavated by hand. A minimum of 5% survival and nature of the archaeological features, of linearfeatures was excavated in slots, in the cases Part of the site was also investigated in an open area of the ring gullies, risingto 25%, and in some cases excavation (Slatcher2000). Neither the trenching nor closer to 50% was sampled in an attempt to retrieve the area excavation showed many archaeological more dating evidence. All other features were features, and those very poorly preserved. The dom- half-sectioned as a minimum. Forty-eight bulk soil inant features were the remainsof medieval ridge and samples were taken from sealedcontexts for environ- furrow cultivation. Remarkably few finds were recov- mental evidence, but few of these yielded any ered, from fieldwalking, trenching and excavation material. combined. Even in areas where cropmarksand the geophysical survey suggested definite archaeolog- ical features, the trenching revealed no trace of sur- EXCAVATION RESULTS vivingarchaeology. Ij particular, nothing was found of the eastern of the two settlementcomplexes. Itwas The excavated features can be grouped together concluded that much of the alreadypoorly preserved conveniently as six structures (A to F) and a single archaeology might have been finally ploughed Out series of parallel ditches. Each structure consisted of relatively recently. one or more penannular gullies, usually associated Archaeological features were located in one with a number of post-holes and 'scoops' that could portion of the site, around Cransley Lodge, where have been hearths. Features not clearly associated preservation was considerably better, and these with either the structures or the ditcheswere rare. In features were dated to the Iron Age and Roman all cases, the ditches clearly post-dated the structures, periods. These included the western of the two and a modern drainage systemcut acrossboth. clusters of ring-gully structures located in the In spite of a lack of deep stratification, almost all geophysical survey. As result, this area was targeted the site's features had at least one direct stratigraphic for full excavation. Specific research objectives relationship with other features, and the broad phas- centredon the date and natureof the occupation and ingof the site is reasonably clear. What is less clear is abandonment, the function of the ring gullies, whether the features within each phase need all be whether the site was enclosed or open, and whether contemporary, or represent a constantly shifting there was any evidence for the integration of an pattern of similar land-use within each period. This occupied area within a formal arrangement of pad- considerably affects not only the length of each docks and fields. In addition it was hoped to provide phase, but all interpretation of the site's status and absolute dating for pottery with traditionally recog- function, as discussedbelow. nized Middle and Late Iron Age fabrics and forms, The modem drains had caused severebut localized especially in respectof their potential continued use damage to many features, and the whole site had into Roman times. been reduced by ploughing to the extent that few featureswere more than 0.3—0.4m deep. Recognition of features in the first place was often a problem, as most of the feature fills, invariably yellow-grey silty

3 Northamptonshire Archaeology 2002, 30 GRAHAM HULLAND STEVE PRESTON

75700

75650

80300

0

Fig2 Mawstey New village, 2000. All features.

4 Northamptonshire Archaeology 2002, 30 MIDDLE IRON AGE OccUPATION AT MAWSLEYNEW VILLAGE, CRANSLEY LODGE, KETrERING.NORTHAMPTONSHIRE

clay, were virtually identical, not onlyto one another, In each case, theouter gully (typically O.20—O.30m deep) was but also to the natural. This had been noted markedly shallower than the inner one (typically 0.45—0.60m problem deep), which may suggest that the inner ring was structural, during the earlierevaluation exercise(Young 2000) forming a slot for upright posts, with an internal diameter of and was onlyexacerbated by the wintry conditions of around 9—1Im in each case, while the outer was an eaves-drip the excavation. gully. Suchan interpretation is problematic, since each structure seems to have been replacedby one with a single deeper trench. However, the later structures may not have had drip gullies becausesome means was found to allow roofwater to drain off PHASESUMMARY viathe 'spur' segments ofthe new single slot (see Phase 2). Although similar in plan, Structure F was fundamentally different from the others. It consisted of a single pcnannular PHASEI MIDDLE IRONAGE i (4th—3rdcentury BC) (Figure 3) ditch, up to 3m wide and I m deep, with evidence of several episodes of rccutting. In this case, the stratigraphic evidence No features on the site can be demonstrated to be earlier than the consistentlysuggested that the outermost cuts were the earlier six penannular structures. There is no stratigraphic basis for (Fig. 5). The irregularityof the shape in plan also suggests that deciding whether these ring-gullies were all contemporary or not all of the cuts followed the whole of the circuit of the successive, nor does theceramic chronology help to distinguish circumference, although actual terminals were not recorded them, so the first elements of each structurehave been phased other than at the extreme points. Sucha massive cut is difficult togetherfor convenience, with the rebuilding episodes likewise to reconcile with the structural function suggestedfor the other all phasedtogether. There are some clear differences between the ring-gullies. Apart from any other consideration,it created a ceramicassemblages from each structure,which maybe chrono- very much smaller internal space (c. 6m by 4.5m), despite a logical, but could be accountedfor in a number ofother ways. similar external diameter of c. 13m. In other contexts, such The radiocarbon dating stronglysuggests two separatephases of massivecuts could beargued to have been structural (perhaps a activity;unfortunately the relationshipof structures to phases is monumental tower, for example) but this seems much more less clear. The initial phase of Structure A and the recut of likely actually to have been a ditch, and it seems unlikely Structure C, for example,produce very comparableradiocarbon simply to have been intended to drain this small area. Similar dates (see below) white the initialcut of Structure B falls much ditches are usually seen as stock pens such as the 'eastern later. Itcan only beadmitted that the dating isimprecise; the site enclosure' at Great Houghton (Chapman 2001, 13-14) but is discussed as if all the structures were initially contemporary Structure F seems over-elaboratefor such a purpose. It could and later remodelled, but this clearly oversimplifies and it is perhaps be interpreted as the quarry ditch for the creation of a possible that no more than one structure was ever in use at any round barrow, not normally an Iron Age monument type. The one time. recutting would of course, then, not be related to the original All the structures were morphologically very similar(Fig. 5). purposeas a quarry, but it may have been regardedas necessary With the exception of the markedly smaller Structure B, they to the proper functioningof the monument to keep its ditch consistedof a double gully, almostcircular, withclear terminals open. It is notable that, in spite of the considerably greater markingan entrancefacing south-east. The case ofStructure A is volume of the cut, deposits within Structure F contained only less clear, as its fill was virtually indistinguishable from the very small quantitiesof finds (for example46 sherds ofpottery natural, but it seems likely it also originally had a double-ring compared to over 250 in each of structures A, B and C). construction. In the cases where it can be demonstrated, it Furthermore,finds from Structure F were mostly from close to appearsthat an original planof two separateconcentric rings was the tops ofthe fills, suggesting they had been introduced when replacedby one single, deeper gully, which may then have been the Structure was already out of use. This may also mean that recut at least once in each case. Thelater gullies, or their recuts, thestructure prc-dates the earliest datable finds. The ditch (and also appearto have extendedbeyond the simplepenannular form thearea within it?) may even havebeen kept clean deliberately. to include extension'spurs'. Thesesingle-ring replacements, and With this proviso, however, it isalso worth notingthat in terms all their rccuts, have been placed into Phase 2, but it is unclear if ofthe types offinds, and even the proportions ofpottery fabrics this accuratelyrepresents the developmentof the site. It may be represented,Structure F appearsvirtually identical to Structure that the chronological sequence runs across the site from struct- A albeit from a much smaller assemblage. The only aspect of ure to structure,with internal modifications to each one follow- the finds that distinguishesStructure F is a concentration of ing a similarpattem time and again, ratherthan auniform pattern smithing earth bottoms and undiagnostic slags. It is therefore of structurally similar replacement in near-contemporary possible that thearea withinthe ditch was a smithy, thoughwhy structures. There is no convincing difference between the pottery this shouldrequire such a massive ditch isunclear. assemblages recovered from the different gullies within each structure. All appear to have been allowedto silt up naturally, and often a small amount of charcoal or ash was all that PHASE 2: MIDDLEIRONAGE ii(2nd—i at centuryBC) (Figure 3) distinguished their fillsfrom the natural. Each ring structure seems to have been a building (except In each of the roundhouse structures, A—D, the original perhaps F, see below). In some of the gully segments, post- double-ring (foundation slot and drip gully) construction was settingsand/or stakcholcs were dimlydiscemed in thebase ofthe replacedby a single, widerand deeper trench. Thiswas typically cuts, while both structures B and C had associatedpost-holes and also recut at least once. In all cases, the fill materials for all of post pads concentrated in the area of the entrance. Very similar these cuts were essentially identical, and distinguishing them double-ringbuilding plans are known from other Iron Age sites proved next to impossible; so the interpretations and aequences in the region (e.g., Grendon Quarry; Jackson 1995). In all cases have had to be based almostentirely on the profilesof the cuts. the extcmal diameter was l3—15m,except StructureB (7m). These can be determined clearly in the cases ofStructures B and

5 Northamptonshire Archaeology 2002, 30 GRAHAM HULL ANDSTEVE PRESTON

75700

75650

0 50m

Fig3 Mawsley New Village, 2000. hon Age features phases Ito3.

6 Northamptonshire Archaeology 2002, 30 MIDDLE IRON AGE OCCUPATION AT MAWSLEY NEW VILLAGE, CRANSLEY LODGE, KE'Il'ERING, NORTHAMPTONSHIRE

75700

75650

0 SOm

Fig 4 Mawsley New Village, 2000. Phase 4, Medieval? and later features.

7 Northamptonshire Archaeology 2002, 30 GRAHAM HULL AND STEVE PRESTON

B W

S 34

tI

19

W 0• NNW /k NW

119

StructureD

133

Structure E E

109

NE SW1 SW NE 254

B 199 7I492201

E

— — — W B o mi 246 — — — — 270 o 5m 21g

Fig 5 Mawsley New Village, 2000. Details ofIron Age structures Ato F.

8 Northamptonshire Archaeology 2002, 30 MIDDLE IRON AGE OCCUPATIONATMAWSLEYNEW'VILLAGE, CRANSLEY LODGE. KETFERING, NORTHAMPTONSHIRE

C, less obviously in the case of Structure A, but not determined the furrows requiring unusually specific definition by deep forthe less ID. At leastone ofthe of for drainage, ownership or hedge-based crops, and the well-preserved phases cutting gullys 2 of this single gully also involved extending it beyond the presenceof medieval ridges may be the reason the Phase I and penannularring shape, creating a spur running a short distance featuressurvived at all. south-eastfrom the southern terminus. This must be assumed to There is clearly no suggestion of continuity between Phases 2 have been fordrainage purposes, perhaps allowingrun off from a and 4, and the Phase 4 features displayedno apparent awareness redesignedroof construction, removing the need for a separate ofthe previous use ofthe site. Findsfrom these ditches were rare, drip-gully. This form recalls the classic shape of the banjo but not apparently much different from the assenthlage of the enclosure', long considered a type-site' for the period. It is earlier Phases, with pottery of fabrics I and 4 the dominant possible that this is another (conscious?)echo in domesticarch- constituent, as in the ring-gullystructures, so that it is possible itecture of an existing larger scale ground-plan template. that this phase also falls within the Iron Age. However, most or earlier comparable to that resonatingbetween ?barrowditch (F) and the all of the finds could have been redeposited from the roundhouses. features, where these werecut by Phase 4 features. As noted above, it is not possibleto determineif the sequence of structures involved contemporary buildingseach modifiedin PHASE5: MODERN the same way approximately simultaneously, or a more complex, longer pattern offewer structures being moved around thesame the site in at a similar site. The latter is suggested by the two distinct radiocarbon Modern drains crossed long parallel lines, the of within the separation (5—bm) to the Phase 4 ditches but on an alignment dates. However, given uniformity development had disturbedmost of structures, it seems to assume that the site consisted approximately 20 degreesremoved. These simplest led to a of of a central focus (F), flanked by two pairs of the prehistoric features and may have degree long-lived redistribution offinds them. contemporaryroundhouses (A andC followed by B and ID) and amongst one smaller ancillary structure (E). The roundhouses were originallyconstructed in one form and later replacedin a second form. ThE FINDS

PHASE 3: LATE IRONAGE? THEPOTFERY Asingle posthole (6), and a scatterof pottery from the surfaceof by Paul Blinkhorn the natural next to it (58), produced a small assemblage of 'Belgic' pottery. The precise significance of this is unclear; it The potteryassemblage comprised 1165 sherds with could mean that the latter part ofthe occupation (possibly even a total weight of 7037g. The estimated vessel Phase 2, given the uncertainty over Phase 4) extended into the rim- after 5DBC when such came into use; it could equivalent (EVE), by Summation of surviving period pottery All the material was belong to a separate unrelateduse ofthe location. There was only sherd circumference was 1.75. one other post hole (4) likely to be associatedwith (6), so little of middle-late Iron Age date with the exception of 68 can be deduced from this evidence. sherds (253g.EVE = 0.20) of late hon Age 'Belgic' material. The range of potterytypes present atthe site PHASE 4:MEDIEVAL?(Figure 4) suggests occupation began around 300BC, although the end date is somewhat more problematic. The site A series of straight parallel ditches cut across the site could have fallen from use at time between east to west, truncating each ofthe Iron any approximately clearly 200BC and AD5O. the of fabrics Age structures. Each ditch was 0.6 to I .30m wide (with the Otherwise, range emphasis on the lower part of the range) and survived to only and forms are typical of the pottery found at con- 0.3 to O.4m deep. These were set on average c. 6.4m apart. temporarySites lfl the region. Again. it is possible these were not all necessarily contemp- orary, as they seem unusually close together for drainage, so that there may have been two phases of somewhat broader FABRICS divisions. Ifthey were defining land boundaries,the plots were of an unusual and unlikely shape. Alternatively,they could be The range ofIron Age fabrics at the site was as follows: bedding trenches for regularly laid-out rows of plants such as vines (cf, Brown and Meadows 2000; Meadows 1996) or hops Fabric I: Coarse shell. Moderate todense temper of angular coarse latter are held to be a medieval intro- shell up to 10mm. with rare quartzite, grog, flint, (although the generally fragments = duction; Mabey 1996). The environmental evidence (below) organic material orironstone. 628 sherds,3421g. EVE 0.74. cannot however,support this suggestion, althougha further hint Fabric 2: Fine shell. Sparse to moderate angular shell fragments comes from the of stakeholes and considerable up to 5mm, although most are usually below 2mm. Other recognition =0. irregularity, as ofroot action, along the bases of some of these materialoccurs as Fl. 60 sherds,409g. EVE trenches. The possibility that these represented medieval ridge Fabric 3: Grogged. AsF2, with sparse to moderatered grog up to and furrow systems, as suggested by the geophysicalsurvey to 2mm. 1 sherd, lOg, EVE= 0. the south of the excavated area (but on a noticeablydifferent Fabric 4: Sand and fine shell. As P2. but with moderate Ic), was considered during excavation but sub-rounded quartz up to 0.5mm, givingsherds a sandy texture. alignment; Fig. = rejected. In retrospect,however, this remains a possibilitywith 298 sherds, 1874g. EVE 0.50.

9 Northamptonshire Archaeology 2002, 30 GRAHAMHULL AND STEVE PRESTON

Fabric 5: Pounded shell. Sparse to moderate fine shell up to 77—85). This has been noted at a numberof sites in the valleys 1mm. Vessels are often self-slipped, so that inclusions are only ofthe Nene andits tributaries, but the material is entirely absent visible section. 104 in sherds, 998g.EVE = 0.31. from this site. This can be interpreted in several ways. Perhaps Fabric 6: As fabric 2, with sparse to moderate sub-rounded the most obvious is that the site fell from use before this ironstone up to 3mm. 6 sherds, 72g, EVE = 0. material came into use, which is thought to be after 175BC 'Belgic': Wheel-thrown,grog-tempered ware, 5OBC—AD5O. 68 (Jackson and Dix 1987, 77), and given the rsnge of pottery sherds, 253g,EVE = 0.20. types at this site, this is entirety feasible. However, the size of the assemblage must also be considered.The largest assemb- Such arange is typical ofthe Iron Age pottery ofthe region, and lage of curvilinear decoratedpottery in the county came from csn be paralleledat many sites, soch as Weekley (Jackson and nearby, at Weckley on the northern edge of Kettering. That Dix 1987) and Twywell (Jackson 1975). The full quantification assemblage comprised 366 sherds of the curvilinear material, of pottery by number and weight of sherds hy fabric type per all noted in Ceramic Phase 1, which produceda total of 2520 context is held in archive. sherds of pottery with a total weight of 63,160g.amounting to nearly 15% of the assemblage by sherdcount. As can be seen, the with a meansherd of is of QUALITATIVEANALYSIS Weckley assemblage, weight 25g. a far higher quality (in terms of preservation) than the group from this site. If the at is The majority of the context assemblages from this site are so occupation Mawsley contemporary with CPI at then it would not be unreasonable to small and fragmentedas tomake impossible any formof Weekley, analysis expect around ISO sherds of curvilinear material from the site. other than providing a probablehon Age date for the deposit. It is still possible that occupation at Mawsley was This is borne out by the mean sherd weight of the group, 6.0'Ig, contemp- orary with CPI at Weekley. Certainly, the small amount of which is very low. In addition, no cross-fits were noted, sug- diagnostic pottery from this site, scored gesting that the majority of the assemblage is much a comprising mainly very vessels with rims, is, curvilinear material product ofsecondary deposition. The poor quality of this group fingernail-impressed can aside, very similar tothat from CPI at Weekly (cf.,Jackson and be seen when this figure is compared with those from other Dix Iron Age sites in the region. For the mean sherd 1987,figs 29—31). example, weight However, to be somewhat in of the pottery from the Iron Age site at Wilby Way near Weekley appears exceptional termsof the amount ofcurvilinear pottery present at the site. The Wellingborough (Blinkhorn and Jackson in press), was lO.9g, next largest group of such material in the county is 21 sherds and that from CPI at nearbyWcekley was around 2Sg (Jackson fromHardingstone (Jackson and Dix 1987). Curvilinear pottery and Dix 1987). Consequently,it is very difficult to give precise was also found at (Jackson 1975), 5km dating to much of the assemblage, and equally problematic to Twywell approximately east of but in this case two both from derive any sort of functional, morphological or typologicat Kettering, only sherds, the information. same vessel, were noted.The exact size of the pottery assemb- lage is notgiven in the report, but it is obvious fromthe accomp- anying illustrations that the assemblage was far larger and better GENERAL CHRONOLOGYAND THE ASSEMBLAGEIN 1TS preserved than the Mawsley group. Thus, if Mawsley was REGIONAL CONTEXT occupied during the 2nd—Istcentury BC and the use of curvi- linear pottery atthe site wasof a similar level to that at Twywell One ofthe few things that can be said with any certainty about rather than Weekley, it is highly unlikely that it would be this rather poor assemblage is that there are few grounds to represented in such a small assemblage as this. It was said that suspect that there was any early Iron Age activity at this the lack ofcurvilinear pottery at Twywell is due to the fact that settlement. The pottery of the early period is well-attestedin the the site fell from use around the time at which the material was region, and is characterized by the so-called 'Angular Horizon' introduced (Jackson 1975, 73). However, the large variation in i.e., jars and bowls with pronouncedshoulders, sometimes with the occurrence of the material throughout the county would fingertip decoration around the carmnation, such as the group suggest that this neednot have beenthe case, although there was from near Gretton (Jackson and Knight 1985, figs 6—9). Two a completelack ofBelgic material from that site. Consequently, fragments of the same shouldered bowl were noted from it cannotbe said with certainly that occupationat this site ceased StructureB (see below). They wereaccompanied byscored ware, before the early—mid 2nd century BC when curvilinear pottery suggesting that rather than being of early hon Age date, the was introducedto theregion. vessel datesto the 4th—3rd centuriesBC, a period which sawthe Finally, there is the small group of 'Belgic' pottery to end ofthe Angular Horizon and a widespread increase in theuse consider. None of the material was associated with any ofthe ofscored ware in theregion. major structures at the site, although a large proportion of the Other than this, however, any statement with regards to group came from an isolated post-hole (6). This is the highly chronology must treated with the utmost caution. The few fragmentedremains ofa single large sherdwhich seems likely assemblages which producedany diagnosticpottery all appear to to have been used as post-packing, and suggests limited be ofthe same broad characteri.e., mainly plainbodysherds in a occupation of the site during the first half of the first century variety offabrics, usually with a few scoredware sherds, and the AD. It is impossible to say if this represents continuityor occasional flat-topped rim with fingernail impressions. This re-occupation, although the evidencesuggests that the former suggests that activity at the site was generally limited to the may be the more likely. None of the features that produced period4IhI3rd—I stcentury BC, althoughthere are other factorsto 'Belgic' pottery had any of the more traditional Iron Age consider. pottery stratified with it, but this is a common situation One ofthe characteristicsof later Iron Age potteryassemb- throughout the county. Belgic pottery appears to be similarly lages from this region of Northamptonshirc is the presence of rare on sites whichdo not produce evidenceof continuity into 'La Tene' curvilinear decoratedpottery (Jackson and Dix 1987, the Roman period, indicating a major re-organization of the

10 Northamptonshire Archaeology 2002, 30 MIDDLEIRON AGE OCCUPATION AT MAWSLEY NEW VILLAGE, CRANSLEY LODGE, KETTERING, NORTHAMPTONSHIRE landscapeduring the first centuryAD (Foster 1999, 132—3). Itis possible therefore that occupation at Mawsley lasted from around 300BC until the middleof the 1st century AD, but the picture is far from clear, and this dating must be regardedas at best tentative.

POTTERY FROM STRUCTURES

Structure A: ?4th —3rd centuryBC+ Structure A produced 259 sherds of pottery (2078g), with the EVE = 0.42. The occurrenceis shown in Table I. The majority ofthe pottery occurred in the mainring gully (10). Themajority of the contexts produced small, fragmented assemblages,with the only exception to this being deposit 53 from the gully terminal (3). The assemblage from this deposit mainly com- prised the highlyfragmented remainsof a singlescored warejar in fabric I (Fig. 6.1), along with a fragmentof a loop handle in fabric 4 (Fig. 7.2), and a single foot-ring base-sherdin fabric 5 (Fig. 7.3). The scored ware vessel, with its developed, fingernail-impressed rim is very typical of the period 300— 200BC in Northamptonshire,and such types are known from the 1st century BC (Jackson 1975, 64; Jackson 1995, 14). The onlyother diagnostic pottery from this structural group was two small fragments of simple upright everted rims from contexts 76 and 77. These are broadly contemporary with the material from context 53. Context 75 (Group 10) produced a largish fragment of the base and lower body of a scored ware vessel also a further sherd of 0OW 100mm (Fig. 7.4), and the context produced scored ware, as did context 71. Itseems likely thereforethat this group belongs to the middleto late Iron Age, possibly the 3rd century BC, although it is entirely possible it could be later. Fig6 Pottery. The rest ofthe assemblagecomprises fragmentedplain body- sherds.

Illustrations Fig. 6.1: Context 53. Fl. Fragments of the rim and base of a Fig. 7.4: Context 75. F4. Base and lower body of scored ware scoredware jar. Black fabric with reddish-brown innersurface. vessel.Black fabric with brown patches on the outer surface. Fig. 7.2: Context 53. F4. Fragment of loop handle. Dark grey fabric with variegated orange and light brown surfaces. Structure B: Middle/Late Iron Age? Fig. 7.3: Context53. F5. Footringbase. Uniform black fabric, Structure B produced 263 sherds with a total weight of I 199g lightlyburnished outer surface. (EVE = 0.21). A total of 6 rimsherdswere present,all of which

Table I: Pottery occunenceby numberand weight (in g)of sherds per context by fabiictype, Structure A Fl F2 F3 F4 F5 F6 Cut Dep Group No Wt No WI No Wt No Wi No Wt No Wt Date 3 53 10 85 1336 16 201 2 28 MIA? 11 60 10 14 58 6 38 2 10 IA 12 61 10 27 35 1 3IIA 18 68 10 2 4 IA?? 27 75 10 4 81 MIA? 28 76 10 11 28 13 93 IA 30 77 10 1 4 IA 41 86 10 8 39 IA 23 71 22 I 16 MIA? 25 73 22 17 10 IA?? 33 80 38 48 61 I 5 IA?? Total 193 1557 0 0 0 0 61 452 4 38 1 31

11 Northamptonshire Archaeology 2002, 30 ORAl-lAM HULLAND STEVE PRESTON

Table2: rcunence Pottery by number and weighl (in g) ofsherds per context by fabric type, Structure B

Fl F2 B F4 F5 F6 FCntx Group No WtNo Wt No Wt No Wt No Wt No Wt Date 117 164 126 55 216 12 170 8 46 MIA?? 121 168 126 II 94 IA 124 171 126 14 41 IA 125 172 126 2 97 LA. 132 179 126 3 II 2 22 IA 142 190 126 5 15 IA?? 143 191 126 1 4 4 8 LA 118 165 135 61 106 IA 134 181 135 824 1 3 IA?? 136 182 335 6 43 2 4 IA?? 141 189 135 21 85 IA?? 119 166 200 1 6 NBA 119 170 200 31 102 MIA? 133 180 200 10 40 IA 133 196 200 5 8 IA Total 189 65525 163 0 0 35 253 14 76 0 0

were 6% or less complete, and the assemblage was largely with no 2 fragmented diagnostic pottery other than 15 sherds of 3 scored ware. The largest rimsherd (Fig. 7.10) had fingernail impressions on the rim, and what may he the beginnings of scoring. This generallysuggests a middleto late Iron Age date, but furtherrefinement is impossible. 1/lustration Fig. 7.10. Context 172. Fl. Fingernail-impressed rimsherd. Black fabricwith reddish-brown surfaces.

StructureC: ?4th —3rd centuryBC Structure C produced 316 sherds with a total weight of 2025g (EVE = 0.56). Once again, the assemblage is highlyfragmented, although there are a small numberof sherds thatindicate that some of the vessels had quite well developed shoulders (Fig. 7.6). In particular, two fragments of a shouldered bowl (Fig. 7.5) seem relatively earlywhen compared to the rest of the assemblage. Such vessets were noted in quantity atthe early hon Age site at Gretton, Northants (Jackson and Knight 1985, 1—8).However, scored figs 10 ware is also present (28 sherds,225g), so the group appearsmost cc; likely to date to around the 4th—3rd centuries BC, a time when carinated vessels were falling from use, and scored ware was becoming morecommon.

1/lustration Fig. 7.5: Context 50. F4. Two non-joining sherds from a shouldered bowl. Dark grey fabric with dark grey outer sur- - 0 --100mm faces. Fig. 7.6: Context 50, F4. Two non-joining rimsherds from a ?shouldered jar. Black fabric with dark reddish-brown sur- faces. Fig 7 Pottery.

12 Northamptonshire Archaeology 2002, 30 MIDDLEIRON AGE OCCUPATION AT MAWSLCY NEW VILLAGE, CRANSLEY LODGE, KETTERING, NORTHAMFTONSHIRE

Table3: Potteiy oun-enceby number and weight (ing) of sherds per context by thbrictype, Structure C

F! F2 P3 P4 F5 F6 F Cntxt Gmup No Wt No Wt No Wt No Wt No Wt No Wt Date 208 255 0 2 12 IA?? 0 55 5 10 12 4 8 41A 1 50 5 48 274 9 67 42 357 28 229 NIIA? 7 59 5 10 42 10 13 97 NRA?? 7 151 5 I 21 1 31 IA 9 57 5 22 137 1 52 30 166 NRA? 106 152 5 14 105 1 21 4 58 NRA? 137 183 5 18 173 1 35 LA?? 211 259 5 15 33 LA?? 212 260 5 12 47 5 35 8MIA?? 213 26! 5 8 43 1 10 MIA? 236 283 5 8 18 NRA?? 122 162 240 2 8 3 12 NRA? Total 160 895 11 140 10 III 802 31 276 2 12

StructureD: Middle/LoteIron Age? = 0.56). The assemblage is small and highly fragmented, with Structure ID produced 35 sherdswith a total weightof 247g (EVE little diagnostic pottery. Context94 did producedthe fingernail- = 0.56). The assemblage is small and highly fragmented, with impressed rim of a large scored warejar (Fig. 7.7) ofa typeakin only a single rimsherd (3% complete) from the whole group. to that from structure A. Four scored sherds(45g) were also Four sherds of scored ware (25g) were also noted, as was a noted. This would therefore similarly suggest a date of the partially complete base from a vessel of uncertain type. It is 4th—3rd century BC forthe group, although this must be treated impossible to date this structure other than broadly within the with caution due to thesmall assemblage size. middle-late Iron Age. Illustration Structure E: ?4th —3rdcentury BC Fig. 7.7. Context94, Fl. Rim from large scored ware jar. Black StructureE produced 35 sherds with a total weightof 247g (EVE fabric with reddish-brown outer surface.

Table4: Pottery cxcurTenceby numberand weight (in g) ofsherds per context by fabric type, Structure D

Fl P2 P3 F4 F5 F6 F CntxtGmup No WtNo Wt No Wt No Wt No Wt No Wt Date 107 153 2 16 MIA? 46 91 47 9 147 IA 108 154 110 4 5 TA 109 155 110 8 22 3 12 1 23 MIA?? 109 156 110 6 8 2 14 IA?? Total 14 30 0 ' 0 0 0 11 47 10 170 0 0

Table 5: Pottery xcurrenceby numberand weight (in g)of shenis percontext by fabric type, Structure E

Fl P2 P3 F4 F5 F6 F Cntxt Group No Wt No Wt No Wt No Wt No Wt No Wt Date 100 94 102 1 61 2 14 16 44 MIA? 103 96 102 4 68 1 5MIA? 104 97 102 2 3 LA 105 98 102 5 20 1 26 2 29 139 186 102 I 19 IA?? Total 6 81 2 14 0 0 22 167 4 32 1 5

13 Northamptonshire Archaeology 2002, 30 GRAHAM HULL AND STEVE PRESTON

Table6: Pottery occurrence by number and weight (in g) ofsherds per context by fabrictype, Structure F

Fl F2 F3 P4 F5 F6 Belgic F Cntxt No Wt No Wt No Wt No Wt No Wt No Wt No Wt Date 149 198 27 IA 201 199 8 11 2 4 4 24 IA?? 203 253 3 4 6 16 IA?? 209 257 3 4 4 5 IA 210 258 3 9 2 23 IA 218 268 2 27 MIA? 222 271 I 2 IA 222 272 2 10 IA 228 278 2 24 MIA?? 242 286 8 12 6 27 IA?? Total 29 77 0 0 0 0 15 75 II 53 2 24 0 0

Structure F ute a maximum of 933 individual bones, and F witha of prob- Structure produced 57 sherds total weight 229g (EVE fewerthan this; 345 were 005). The assemblage is small and highty fragmented, with ably many tiny fragments little diagnosticpottery. A singledeveloped rimsherd was noted retrieved from sieved soil samples. Fewer than one during the cleaning of the junction of 198/199, but otherwise third of the hand-collected bones were identifiedto there was nothing to suggest a date other than three sherds of taxon. Although much of the unidentified bone is scored ware. the structurecan be a date of the Thus, only given fragmented ungulate limb bones, the pieces could middle-late Iron Age, although the developed rim (Fig. 7.8) suggests later rather than earlier. However, since the not bejoined or securelyidentified. Of the 306 ident- stratification ofthe sherd is somewhat uncertain, this date must ified remains over a third are of looseteeth. This is a be treated withCaution. high proportion and indicative of poor survival as teeth are more resistantthan bone.It is expectedthat Illustration in in Fig. 7.8: Contextl98/199.F4. Devetoped rimsherd from closed there is a bias the material favour of large vessel.Black fabric with reddish-brown surfaces. species and the larger anatomical elements. Surface features such as butchery, gnawing and pathology POTTERYFROM OTHER FEATURES had been largely obliterated and epiphyseal ageing data limitedand unreliable. As was the case with the ring-gullies, the pottery from the other Despite the obvious limitations of the material features, such as pits and linear ditches, was extremely there are several observations that can be made. fragmented with little diagnostic material present. The few and can diagnostic sherds present were either scored ware, or extremely Horse, cattle, sheep, pig, dog be positively small fragments of rims, which other than providing a general identified and there is also one fragment of bird date of the middle to late Iron Age, offered little in terms of bone, which could not be identified. The distribution refiningthe chronology. of the taxa by context, together with other inform- Illustration ation, is held in archive. Fig. 7.9: Context 178, F2. Pierced bodysherd from lower part of Cattle bones dominatethe assemblage with sheep small to medium sized scored ware jar. Dark grey fabric with in secondposition and horse third. In most lion Age reddish-brown inner surface. assemblages sheep is more frequent than cattle, the dominance here could be a taphonomic artefact as ANIMAL BONES cattle bones are large and more likely to survive. by Sheila Hamilton-Dyer This might also apply to the numbers of horse and pig bones. Although the condition of the bone is not Animal bone was recoveredfrom 70 contexts. The good, it was noted that more meat bones than head condition of the bone is generally poor, with a and foot bones of cattle are present,particularly from significant degreeof surface erosion. The integrity of gully terminus 1 (part of Structure C). Several of the bone is weak and most bones had fresh breaks these are of elements such as the femur which where it had not.been possible to recover them in a usually preserveless well than the metapodia (foot) complete state. Following examination it became and this is therefore a positive observation rather clearthat the 1366 hand collected specimens constit- than bias due to poor preservation.

14 Northamptonshire Archaeology 2002, 30 MIDDLEIRON AGEOCCUPATION AT MAWSLEY NEW VILLAGE, CRANSLEY LODGE. KETI'ERING. NORTHAMFTONStIIRE

Butcherymarks made by knives were observedon fell into the undiagnostic category. A summary is a few bones, including a horse tibia. One bone given in the table below. showed evidenceof canid and several of gnawing, Several small or what be ore were found in the smaller of bone had been charred. fragments may pieces several contexts (53, 170, 191, 282); the fragmentfrom context The three dog remains are of jaws and a tooth of 191 is very magnetic. The identification of the stones is not 'average' size. A variety of dog sizes have been certain but some appear 10 be roasted, a process sometimes recorded for Iron material but most fall in the carried out before smelling 10 drive off carbon dioxide,or waler Age and on the The is middleof the sulphur (depending ore). porosily increased, range. allowing betler penetration of the reducing gases during the smelting process. A small amountof unstratified tap slag was recovered (at site METALWORK coordinates I 15E 66N) but it IS 001 known how it relates 10 lhe by Nicola Powell rest of the assemblage or the dale of its deposit. Such a small amount, by itselfis not enough to enable us to say that smelling Three of iron were recovered from the was takingplace on the site. pieces Nine hearth bottoms were recovered from several excavation in addition to one from wet smithing piece sieving. contexts, from Structures A and (mainly) F. This slag is the most A very corroded piece similar to a nail stem was distinctive wasteproduct ofsmithing activity. Itwas the result of recoveredfrom the slot through gully terminal 109 high temperature reactions between the iron, iron-scale and silica in D and a broken of from either a clay furnace lining or the silica flux used by the (156) Structure piece strapping of the iron and inhibit further oxidation with what the remains of nail or rivet hole smithto clean the surface may be a of the iron during hot working. The slag dripped down into the was found in posthole 213 (261), part of Structure C. hearth base formingsmithing slag which,if not cleared out, grew A triangular-shaped piece of ferrous material was into the characteristic piano-convex-shaped smithing hearth found in slot 12 10 bottom in front of and below the tuyère where the air from the (61) through gully (Structure A). bellows enters hottest of the When removed associated with iron A (the part hearth). It may be working. lump of from the hearth they were usuallytaken outsideand depositedin iron-rich material from sample 42 taken from slot the nearest pit or ditch. The proximity of cut features or dumps 145 (193) through Phase 4 gully 206 could be from a with amountsof smithing hearth bottomsto a buildingis often a hearth-bottom. good indicationthe structuremay have been a smithy. Hearth lining can vary from highly vitrified hearth lining nearest the tuyère region to burnt clay. By itself it is not diagnosticof smithingactivity but associationof vitrified lining SLAG with other diagnostic material provides support for the process. by Lynne Keys Cinder is a very porous, highlyvitrified material formed at the interfacebetween the alkali fuel ashes and siliceousmaterial of a A small assemblage of material (6.58kg) thought to hearth lining. On many excavations it represents the lighter relate to was recovered portionof vitrified hearth lining. ironworking during Structures A and F produced the smithing hearth bottoms, excavations. The assemblagewas visually examined while Aand B and posthole234 produced the possibleiron ore. It and categorized on the basis of morphology alone. would seem from this that there is very little overlap between the Each category of slag within each context was two types ofactivity. The smithingevidence probably relatesto a and the hearth bottoms were limited period of activity; and the possible ore could be weighed smithing accounted for by a small amount of smelting or fragments of individually weighed and measured. Most of the local ore being carried onto the site during other activities, Mawsley assemblagewas highly fragmented and so especiallyas ironstone occurs locally.

Table7: Summary ofIron Slag by weight(g)

structure other description A B C I) E F atures total firedfbumt clay 85 115 26 11 22 6 29 294 cinder and undiagnttic 0 129 8 0 0 2 0 139 roastedore? 8 14 0 0 0 0 0 22 smithing hearth bottom 736 0 0 0 0 2620 0 3356 tapslag o o 0 0 0 0 270 270 undiagnostic 170 666 16 0 222 1178 18 2270 hearthlining 0 52 12 0 34 12 1 Ill total 999 976 62 II 278 3818 318 6462

IS Northamptonshire Archaeology 2002, 30 GRAHAM HULLAND STEVE PRESTON

FIRED CLAY probably obtained from a local river. A possible Varian Denham by ritual significance has been suggested for white in earlier and medieval A total of 109 pieces (336g) was recoveredfrom the quartz pebbles, prehistoric site. It is notable that several were in contexts in Ireland and Scotland, and possiblyfrom pieces pottery but has not been demon- fabrics, and some had been heated to Anglo-Saxon England, yet temperatures strated for the Iron 1988 and that caused vitrification. The fired clay did not have Age (Baker pers. the structural traces that are associated with burnt comm.). All the pieces came from ring gully fills, daub, nor was unburned daub recovered. Each of with a particular concentration (14 pieces) in the Structures A—F some fired but there terminals (117 and 118) of gullies in Structure B, produced clay, which also contained of burnt flint. The was no marked clustering. quantities terminals of roundhouse gullies and of ditches in general have long been recognised as preferentially BURNTFLINT selected for the reception of 'deliberate' or 'struct- byNicola Powell ured' deposits.Terminal 117, in particular, seems to have seen the deposition of an unusual concentration Forty-seven pieces of burnt flint with a total weight of finds of most but there is little to of were all but one from the wet types, suggest 66g recovered, this was ritual. A singlesmall irregularlump of iron- sieving programme. All were examined for evidence stone derived from the local North- of None was found and the material was presumably working. amptonshire ironstone deposits also came from not retained. Structure B.

STRUCK FLINT PLANT SteveFord MACROFOSSILS by by Mark Robinson Nine struck flints were recoveredfrom nine separate of 5-10 were taken contexts. One of the is a made on a Forty-nine samples, mostly litres, pieces scraper from the full range of contexts. In many cases, these large natural flake. One piece is possibly a segment small from the blade. The remainder flakes. All the samples represent multiple samples of a large are of same fill. are in condition and to be made gully pieces good appear The sampleswere floated onto a 0.3mm mesh and from flint available locally. The collection is not dried. A of 40 flots were then scanned distinctive and some or even all of sub-sample chronologically undera binocularmicroscope for charcoal and other the pieces could be contemporary with Iron Age carbonized remains. The charred seeds observed of the site. flint occupation However, assemblages (chaff is absent) were identifiedand counted. Char- of Middle or Late Iron date are unambiguously Age coal from the flots was broken transversely and atbest recordedin the literature 1981; rarely (Saville examined at up to x50 magnification. While this is Humphrey and Young 1999). Durable prehistoric an means for the identification of flint across the often at appropriate is present widely landscape, the other charcoal identifications are low densities and the results Quercus, (Ford 1987) represents tentative. Some Quercus charcoal was also of a combination of of casual the manur- picked factors loss, out of Context 282 from Feature 234 during the ing of arable fields, a short-term mobile settlement excavation. pattern and ad hoc use. In this case the Mawsley Carbonizedremains are five of struck flints could be residual. very sparse, only easily the samples containing seeds and no sample con- taining more than three seeds. While the individual STONE samples were mostly small even 40 litre samples would probably not have given any assemblages with David Williams of the University.of Southampton as many as.ten items. With.the exception of one un- has identified twenty-seven pieces of stone recov- identified weed seed, the seeds are all cereal grains. ered from the excavations. Most are burnt quartzite There are several examples of Hordeum sp. (barley) pebbles. Other than burning, none of the pieces can including hulled Hordeum vulgare (six-row hulled be shown to have been worked or used. Most of the barley) and a single grain of Triticum sp. (wheat). quartzite pebbles appear water-turned and they were Unfortunately, it was not possible to determine

16 Northamptonshire Archaeology 2002, 30 MIDDLEIRON AGEOCCUPATION AT MAWSLEY NEW VILLAGE, CRANSLEY LODGE. KETFERING, NORTHAMPTONSHIRE whether the wheat was T. dicoccum (emmerwheat) using CALIB rev 4.3 (Data set 2, 1998 decadal or T. spelta (spelt wheat), the two wheatspecies most atmospheric data; Stuiver ci al. 1998). The result likely to have been grown in the region during the from Structure F (KIA 15744) is regarded as almost hon Age. certainly derived from contaminated material and These results show that although the settlement has been ignoredin phasingthe site, but is presented used cereals, there is no evidence for large-scale below for completeness (details of the reasons cereal processing. Indeed, it is possible that the causing the Kiel laboratory to doubt this result are settlement was not involved in cereal cultivation at held in archive). all but only imported grain for consumption. This low concentration of charred grain on the settlement is of significance to the understanding of the Iron THE NATURE OF THE SITh Age economy of the region. Identifiablecharcoal is only present in six of the The site provides a remarkable sequence of flots and the concentrations are mostly low. The structures and landscape use although the detailed charcoal includesboth Quercus sp. (oak) and thorny chronology remains uncertain. Excavatedopen hon scrub / hedgerow species such as cf. Prunus sp. (sloe Age settlements are still rare in the region, although etc). The charcoal was probably derived from their presence is increasingly recognized through domestic hearths. non-invasive surveymethods, and the publicationof several important excavations is imminent. The present site provides comparative data to Set RADIOCARBON DATING alongside, for example,Ecton (Atkins et al. 2001). The excavatedremains contain no indication that the Three samples of bone material from sealedcontexts site sat within an enclosure, although the earlier within Structures A, B, C and F were submitted for geophysical survey (Nicholls 1999) suggested a AMS dating of collagen at the Leibniz Labor für possibility that this might be so, and the adjacent Alterbestimmung und Isotopenforschung (Kiel cluster of very similar features was enclosed (Fig. University). Details of the methodology and results ic). A combination of enclosed and unenclosed are held in archive and only a summary presented settlement featuresin close proximity is perhapsnot here. Calibrated or calendar ages were calculated so unusual as might be imagined (cf., Ecton; Atkins

Table:Radio-carbon dates

Context Reference Result, years Calibrated age ranges calendar years (BC number BP except asstated) 68.2%robabi1ity 95.4%pthabiity Structure C, KIAI5741 2177± 20 350-316(Probability 52%) 356—287 (Probability53%) Gully5, 306-305(Probability 1%) 258-245 (Probability 3%) (106,152) 229-220(Probability 13%) 233—216 (Probability 12%) 208-197(Probability 16%) 214_169(Probability3l%) 191-175(Probability 17%) Stnicture B, K1A15742 2093±21 158-155(Probability3%) 169-50 (Probability 100%) Gully 135, 155-130 (Probability 31%) (118,165) 119-89 (Probabiity4l%) 77-58(Probability 24%) Structure A, K1A15743 2191±19 352-294(Prc*bility7l%) 357—276 (Probability 59%) Gully 10, 230-218 (Probability 16%) 259-241 (Probability8%) (33,53) 210-200 (Probability 13%) 234-197 (Probability 28%) 190—175 (Probability 5%) Structure F KlAl5744 1588± 35 AD428—434(Probability 10.9%) AD402-544(Probability 93.5%) Ditch300 AD447—463 (Probability 17.4%) AD548—559(Probability 1.9%) (209,257) AD481—531 (Probability 39.9%)

17 Northamptonshire Archaeology 2002, 30 GRAHAM HULL AND STEVE PRESTON

et at. 2001). A note of Caution should be sounded, shift to the adjacent site and back again. Within an since evaluation trenching indicated that many of the increasingly regulated landscape, the reuse of the featuresdetected by the geophysical survey no longer same plot need not seem implausibly coincidental; survived; it is possible that an enclosure did once equally, one of the structures (F, perhaps) could have exist around this site also. However,the geophysical remained, or left sufficient trace, to allow returning surveyresults may indicateno more than a trackway settlersto relocate in the same spot. along the east side of the roundhouse cluster. The concentration of smithing debris in Structures Certainly there is little indication of the substantial A and F might suggest a specialist purpose for these ditches enclosing Sites such as Wakerley (Jackson buildings, but the quantities are not great, and no and Ambrose 1978). other features survived to suggest intensive or When the evidence from the earlier survey and prolonged metalworking on the site. evaluation is combined with that from the excav- Unfortunately, there is little scopefor a character- ation, the interestingaspect is the dual nature of the izing the surrounding landscape from environmental settlement. There is little doubt that the second data, which was virtually non-existent. It is tempting, clusterof roundhouses was essentiallysimilar to that but probably dangerous, to interpret the lack of plant excavated. Two possible interpretations arise. The remains as in itself significant. The suggestion that two clusters could be chronologically distinct and crops were notprocessed on the sitecan, however, be represent a single social unit over an extended given some support from the lack of storage facilit- period, or there could be two contemporary social ies, and the absence of, for example, querns. Mawsley units living side by side, each perhaps consisting of shows a distinctlack of pits by comparison with one extended family or clan. The implication of the c. 180 at Twywell. Wakerley also has virtually none two discrete phases indicated by the radiocarbon and Jacksonand Ambrose note (1978, 171) that 'pits dates would tend to support the first suggestion. are practically non-existent on Belgic sites' in the The precise nature of the structures represented by region. As more sites are excavated, this distinction the successive ring gullies remains unfortunately may come to be seen as functional rather than chron- obscure. The simplest explanation would be to see ological. In Wessex, a disparity can clearly be seen •them as successive rebuildings of roundhouses, between the massive storage facility at some sites similarto the pattern seen at Wakerley, where it was compared to others and this distinction is not always also unclear how many of the roundhouses would drawn along the simplelines of hillforts versus open have been occupiedat any given time. The complex- sites. Even if it were argued that shallow pits had ity of the recutting is similar to a pattern observed at already been ploughedOut at Mawsley, pits of such enclosure G5 at Ecton (Atkins et al. 2001), and the limited depth would scarcely have been suitable for 'eastern enclosure' at Great Houghton (Chapman storage in any case. Nor were any above ground 2001),although the examples at Mawsley (excepting storage facilities (typified by 4-post structures) themore massiveStructure F) are much slightercuts, identified in the excavation. The lack of storage by and produce a considerably smaller enclosed space itself need not suggest that this was a sporadically than the c. 20m internal diameter of Ecton enclosure occupied, seasonal, specialist site, as there seems G5. Constant replacement of building frames in the little reason to suppose that Structures A to D were same spot is less easily explained than continual anything other than dwellings. However, a sporadic re-cutting of enclosureditches, but this does appear occupation cannot be ruled out either, and might to be the best explanation of the Mawsley structures. account for the complexity of recuts and phases of Based on the radiocarbon determinations, there each Structure. seems almost no chance that all of Structures A, B Cattle dominate the very limited animal bone and C were in contemporary use. It remainsunclear assemblage. This may be due to preservation or how the other structures relate to this sequence, or recovery bias, as most Iron Age assemblages tend to indeedhow the re-definitions of each structure fit the be dominated by sheep/goat, although cattle also picture. Allowing an approximately 50 year life for formed the vast majority of the identified bone from each building, the remaining structures could fill the Sywell (Atkins et a!. 2001). If taken at face value, gap between the later end of the date range for this might also suggest an unusual consumption Structures A and C and the early end of the range for pattern on this site, but again the evidence is B. It would also be possible to suggest a settlement inconclusive.

18 Northamptonshire Archaeology 2002, 30 MIDDLE IRON AGE OCCUPATION AT MAWSLEY NEW VILLAGE, CRANSLEY LODGE, KETFERING, NORTI-JAMPTONSHIRE

The site lies on a distinctive ridge on the watershed BIBLIOGRAPHY between the River1se to the east and the streamsnow Atkins, R. Parry, S. Holmes, M. and Meadows, 1., 2001. feeding Pitsford reservoirto the west. It is on heavy 'Excavations of hon Age Settlements at Sywell Aerodrome Boulder Clay, not normally thought of as a type of (1996) and Ecton (1992—3) Northamptonshire', Northampton- terrain much favouredfor Iron settlement, and shire Archaeol 29 (for 2000-I), 43—71. Age Baker, F.M.C,, 1988. 'White quartz pebblesin ritual contexts in often tacitly assumed to have been wooded. The Scotland'unpub BA dissertation, Univ Durham nearby site at Great Houghton, in an otherwise very Blinkhorn, P.W. and Jackson, Di., in press, 'Iron Age Pottery comparable location, lay on limestone just at the from Wilby Way, Wel)ingborough, Northants'Northampton- of the Boulder It is shire Archaeology, 31. margin Clay (Chapman 2001, 3). B.G.S., 1993, British Geological Survey Map, 1:50 000, Sheet therefore interesting to note the evidence here not 185, Drift Edition, Keyworth only for settlement, but apparently for re-use of the Brown, A.G. and Meadows, I., 2000. 'Roman vineyards in site over the course of several generations, rather Britain: finds from the Nene Valley and new research', than late Onto the Antiquity 74,491—2. a 'expansion' heavy clay. Chapman, A., 2001. 'Excavationof an Iron Age Settlement and a Whateverthe detail of the sequence withinthe site, Middle Saxon Cemetery at Great Houghton, Northampton- the date of final abandonment is important, as a shire, 1996', NorthamptonshireArchaeol 29 (for 2000—1), contributionto the broader study of population and 1-41. settlement since the use of the Cunliffe,B. W., 2000. The Danebury Environs Programme: the expansion, agricult- prehistory ofa Wessex Landscape;vol 1. introduction, EngI urally marginal Boulder Clay zone has long been Heritage and Oxford Univ Comm Archaeol Monogr 49, regarded as a sensitive indicator of population Oxford. D. 'Raunds pressures. Although there was a small amount of Denham, V. and Jackson A., forthcoming, Survey from the it was not associated Iron Age Pottery', in Parry, S, The Raunds Area Survey. 'Belgic' pottery site, Ford, S., 1987. The East Berkshire Archaeological Survey, with the structures, and the weight of evidence Berkshire County Council Dept of Highways and Planning, suggests the Phase 2 settlement was abandoned by c. Occas Pap I, Reading. 5OBC. The ditches Phase 4 no Foster, P. J., 1999. 'Lute Iron Age/early Roman Northampton- of display knowledge of ceramic to or relation the Phase I and 2 and it shire: A study in the use analysis investigate of, to, structures, social, economic and landscape changes', Northamptonshire is entirely possiblethat all the finds from their fills Archaeol 28, 129—35. have been disturbedfrom the earlierdeposits. If these Haselgrove,C. Armit, I. Champion, T. Creighton,J. Gwilt, A. later hon land rather Hill, J.D. Hunter, F. and Woodward, A., 2001, Under- representeda phaseof Age use, an action. than medieval and then it does standing the British Iron Age: agenda for ridge furrow, suggest Cambridge. the bringing of this marginal land into cultivation. Humphreyi. and Young R., 1999. 'Flint use in Later Bronze The 1st century AD providesa likelycontext for such Age and Iron Age England— still a fiction?', Lithics,20, 57— whether as a of the 61. a re-organization, consequence at North- Claudian or somewhat earlier. Such re- Jackson, D., 1975. 'An Iron Age site Twywell, conquest amptonshire',NorthamptonshireArchaeol 10, 3 1—93. organization is increasingly recognized in the immed- Jackson, D. and Ambrose, T.M., 1978. 'Excavations at Wakerley, iate pre-conquest period in parts of the country,such Northants, 1972—75',Britannia 9, 115—245. as Kent or Gloucestershire. In this respect, however, Jackson, D. and Knight D., 1985. 'An early Iron Age and Beaker a move from a settlement to an use of the site near Gretton, Northamptanshire', Northamptonshire agricultural Archaeology 20, 67—86. landscape seems to run counter to the general trend, iackson, D. andDix, B., 1987. 'Late Iron Age andRoman Settle- which is generally for an expansion in both the ment at Weekicy, Northants',Northamptonshire Archaeology numbers and size of settlements et al. 21 (for 1986-87), 41—94. (Haselgrove 1995. at Grendon North- One lie in the Jackson, D., 'Archaeology Quarry, 2001, 29). possible explanation may amptonshire, Part 2: other prehistoric, Iron Age and later sites enhanced abilityof farmersat this time to exploitthe excavated in 1974-5 and further observations between heavy clays,using iron-tipped plough shares, so that 1976-80',Northamptonshire Archaeology, 26, 3-32. land previouslyunsuited to arable agriculture could Meadows,1., 1996. 'Woolaston',Current Archaeol, 150, 212—5. be under the more Another Mabey,R., 1996. Flora Britannica, London. brought plough easily. Nicholls, J., 1999. 'Mawsley New Village Northamptonshire', factorcould be that the nucleation of population into GSB Prospection Geophysical Survey Rep 99/81 larger central settlements would have meant the Saville, A., 1981. 'Iron Age fliutworking - fact or fiction?', abandonment of smaller sites in the of the Lithics,2, 6-9. vicinity 1999. 'Adesk-based Assessment on new centre as has been have occurred Slatcher, D., Archaeological suggested may the line of the requisitioned sewer at Mawsley New Village, around the developedhill-forts of Wessex (Cunliffe Cransley, Northamptonshire',John Samuels Archaeological 2000). Consultants Rep 615/99/001, Newark.

19 Northamptonshire Archaeology 2002, 30 GRAHAM HULLAND STEVE PRESTON

Slatcher, B., 2000, 'A Fieldwalking Survey, Trial Trenching ACKNOWLEDGEMENTS Evaluation and Area Excavation at Mawsley New Village, Kettering, Northamptonshire',John Samuels Archaeological The fieldworkwas carried out with the assistance of ConsultantsRep 531/00/006,Newark. Sian Karen Simon Clare Stuiver, M. Reimer, P. J. Bard, E. Beck, J. W. Burr, G. S. Anthony, Bolchover, Cass, Hughen, K. A. Kromer, B. McCormac, G. van der Plicht, 1. Challis, Sarah Coles, Emma-Jane Evans, Stephen and Spurk. 1998. 'INTCAL98: Radiocarbon Age Calibration Hammond, Lisa-Maree Hardy, Paul Lambert, Tim 24,000—0cal BP', Radiocarbon40/3, 1041—83. Lankshear and Kate The was drawn 2000. 'Results of an Taylor. pottery Young, J.. Archaeological Evaluation J0 and the final illustrations were Excavation of the northern part of Mawsley New Village, by Richards, Cransley, Northamptonshire', John Samuels Archaeological produced with the assistance of Leigh Torrance and ConsultantsRep 531/00/011,Newark. Steve Ford.

Northamptonshire Archaeology 2002, 30 20