Agricultural Revolution’: a Case Study from Stafford, England, C

Home , Extensive farming, Medieval archaeology

European Journal of Archaeology 23 (4) 2020, 585–609 This is an Open Access article, distributed under the terms of the Creative Commons Attribution- NonCommercial-NoDerivatives licence (http://creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is unaltered and is properly cited. The written permission of Cambridge University Press must be obtained for commercial re-use or in order to create a derivative work.

An Integrated Bioarchaeological Approach to the Medieval ‘Agricultural Revolution’: A Case Study from Stafford, England, c. AD 800–1200

1 1 1 HELENA HAMEROW ,AMY BOGAARD ,MICHAEL CHARLES , 1 2 1 EMILY FORSTER ,MATILDA HOLMES ,MARK MCKERRACHER , 1 1 1 SAMANTHA NEIL ,CHRISTOPHER BRONK RAMSEY ,ELIZABETH STROUD 2 AND RICHARD THOMAS 1School of Archaeology, University of Oxford, UK 2School of Archaeology and Ancient History, University of Leicester, UK

In much of Europe, the advent of low-input cereal farming regimes between c. AD 800 and 1200 enabled landowners—lords—to amass wealth by greatly expanding the amount of land under cultivation and exploiting the labour of others. Scientific analysis of plant remains and animal bones from archaeological contexts is generating the first direct evidence for the development of such low-input regimes. This article outlines the methods used by the FeedSax project to resolve key questions regarding the ‘cerealization’ of the medieval countryside and presents preliminary results using the town of Stafford as a worked example. These indicate an increase in the scale of cultivation in the Mid-Saxon period, while the Late Saxon period saw a shift to a low-input cultivation regime and probably an expansion onto heavier soils. Crop rotation appears to have been practised from at least the mid-tenth century.

Keywords: medieval farming, Anglo-Saxon England, crop stable isotopes, functional weed ecology, Stafford, open fields

INTRODUCTION three-field crop rotation, which enabled a larger proportion of arable land to be culti- The period between c. AD 800 and 1200 vated; widespread adoption of the mould- saw dramatic changes in farming practices board plough, allowing farmers to cultivate across large parts of Europe. These sus- heavier, more fertile soils; and low-input tained an increase in cereal production so cultivation regimes, in which fertility was greatthatithasbeendescribedasan‘agri- maintained by regular, short fallow periods cultural revolution’ which fuelled population during which sheep grazed on stubble and growth and underpinned the expansion of weeds, rather than by intensive manuring. townsandmarketsaswellastheriseof This enabled farmers to extend the area of lordship (White, 1940; Duby, 1954;Dyer land under cultivation by decreasing input, et al., 2018). Three key innovations made i.e. labour and manure, per unit area, a this increase in yields possible: two- and process referred to here as ‘extensification’.

© European Association of Archaeologists 2020 doi:10.1017/eaa.2020.6 Manuscript received 18 July 2019, accepted 31 January 2020, revised 4 December 2019 Downloaded from https://www.cambridge.org/core. IP address: 170.106.35.76, on 30 Sep 2021 at 05:08:41, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/eaa.2020.6 586 European Journal of Archaeology 23 (4) 2020

The result was much larger overall yields, productivity was boosted by manuring, and despite a decline in yield per land unit. In whether crop rotation was practised. many regions (including around one-third Analysis of the associated weed flora is pro- of England and much of Europe), these viding further indications of soil fertility and innovations culminated in the introduction disturbance, reflecting the extent to which of ‘open fields’, in which the holdings of fields were manured and tilled, as well as individual farmers took the form of strips likely sowing times. The lower limb bones dispersed across several large, unenclosed of cattle are being examined for the kinds of areas of arable land, intermingled with pathologies that can be caused by pulling a those of their neighbours. heavy plough, while other zooarchaeological To operate this more productive system, data are providing evidence of changing pat- farmers shared expensive resources such as terns of animal husbandry. Pollen data are teams of oxen and mouldboard ploughs, being analysed to gauge the impact of cereal and the cultivation of the open fields had to farming on the medieval landscape more be agreed and managed communally. The widely, for example, on the ratio of arable to need to coordinate cultivation activities is pasture, and on the overall scale of agricul- widely thought to lie behind the formation of turallanduse.Patternsemergingfromthese the nucleated villages that still characterize bioarchaeological data are being compared many parts of the countryside today. In this with evidence from excavated settlements— way, innovations in farming transformed not buildings, enclosures, droveways, etc.—to only large parts of England’s landscape, but explore the inter-relationship between also its social geography. When, where, and arable production, stock management, and how this unprecedented form of agriculture settlement forms (see Hamerow, 2012). A emerged, however, and whether it had suite of radiocarbon dates on charred a significant impact before the Norman cereals, bones, and pollen cores is being Conquest, remain contentious issues, largely used to locate the spread of low-input because scholars have been forced to rely on farming practices in time as well as space. a limited range of indirect evidence, written A national database of charred plant and archaeological (Banham & Faith, 2014; remains from this period is being compiled Hall, 2014; Dyer et al., 2018). The aim of as part of the project, while plant and the ‘Feeding Anglo-Saxon England’ project faunal remains from a series of case studies (hereafter FeedSax) is to break this impasse are being examined in more detail. These by generating the first direct evidence for the case studies span several ecological zones, conditions in which medieval cereals were within and beyond the core area of open grown by subjecting plant and animal fields and nucleated villages, sometimes remains from archaeological contexts to a referred to as the ‘central province’ (Roberts range of scientific analyses, allowing the asso- & Wrathmell, 2000). The purpose of this ciated farming regimes to be reconstructed. article is to present one of these case FeedSax is deploying a range of analyses studies, that of Stafford, as a ‘worked in conjunction to trace the development of example’ to illustrate how results from farming regimes involving crop rotation, use these various analyses can be integrated. of the mouldboard plough and low-input cultivation. Analysis of stable isotopes in preserved cereal grains is being used to ANGLO-SAXON AND MEDIEVAL STAFFORD establish the conditions in which crops were grown, whether in light, dry soils or heavy, Stafford provides a particularly useful case wetter soils, the degree to which study: modern, well recorded excavations

Downloaded from https://www.cambridge.org/core. IP address: 170.106.35.76, on 30 Sep 2021 at 05:08:41, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/eaa.2020.6 Hamerow et al. – An Integrated Bioarchaeological Approach to the Medieval ‘Agricultural Revolution’ 587

at several locations in the present-day The place-name Stafford, meaning town have produced plant remains span- ‘landing-place ford’, refers to the town’s ning the late ninth to thirteenth centuries, location at the intersection of land-based and local pollen data are relatively abun- and riverine transport routes, on what is dant. The faunal remains, on the other effectively a peninsula surrounded by hand, are poorly preserved; it has, there- marshy land (Dodd et al., 2014). It also fore, not been possible to conduct statistic- lies at the interface of light soils associated ally meaningful analysis of lower limb with the river terraces and heavier clays. bones and other pathologies. Stafford’s Stafford’s immediate hinterland is thus early history is comparatively well docu- well suited for crop husbandry and there is mented. According to the Anglo-Saxon good evidence for its history as an arable Chronicle,aburh (fortified settlement) was landscape. Several sites of (undated) ridge- established at Stafford in AD 913 by and-furrow earthworks formed by mould- Aethelflaed of Mercia. Stafford’s earlier board plough cultivation and marking the history is not recorded and excavations in location of medieval fields have been iden- the 1980s at several sites within the medi- tified in aerial photographs (Cotswold eval town revealed no conclusive evidence Archaeology, 2010)(Figure 1). As one for Mid-Saxon (seventh- to ninth-century) would expect, given the likely role of occupation (Carver, 2010: 31, 39). ridge-and-furrow in improving drainage, Evidence was similarly lacking for post- nearly all are on heavier soils. A concen- Conquest activity, and it was argued that tration of ridge-and-furrow lay less than 2 Stafford must have been virtually aban- km north of the burh, near the deserted doned until a revival of activity in the late medieval village of Marston (the place- twelfth century (Carver, 2010: 107–08). name suggesting a settlement near wet This view was revised following excava- ground; Cotswold Archaeology, 2010). tions at Tipping Street in the south- Indeed, Speed’s 1610 map of Stafford eastern part of the medieval town in 2009. depicts a farmer and plough team at work These uncovered three pottery kilns that in fields apparently to the north of the produced substantial quantities of charred town (Figure 2). The ready availability of plant remains, probably representing crop chaff, seemingly used as fuel, suggests fur- processing waste used as fuel (Dodd et al., thermore that crops were threshed and 2014: 72). Six radiocarbon dates from winnowed in the immediate vicinity of the these kilns indicate that they probably town; these activities are usually under- came into use during the ninth century, taken close to arable fields to reduce the suggesting that Stafford was well estab- bulk and weight of the crop prior to trans- lished before Aethelflaed fortified the site port. Most of the charred plant remains in AD 913, although the chronology of analysed here are therefore likely to have this earliest phase of activity has had to been grown close to Stafford (Moffett, be revised again in light of a series of 2010; Dodd et al., 2014: 73). new dates obtained by the FeedSax project (see below; Dodd et al., 2014:79–85, 100– 05). It is therefore possible that Stafford THE CHARRED PLANT REMAINS FROM was already a centre, perhaps established by ANGLO-SAXON AND MEDIEVAL STAFFORD Mercian royal initiative, where agricultural surpluses were being collected and pro- The archaeobotanical analysis presented cessed at the time of its fortification in 913 here is based on forty-six soil samples exca- (Dodd et al., 2014; Blair, 2018:263–64). vated at Stafford whose archaeobotanical

Figure 1. Location map of Stafford and the surrounding region showing key sites, soil types, and preserved ridge-and-furrow.

contents are dominated by charred cereal SELECTION AND DATING OF SAMPLES FOR remains (Supplementary Material 1). BIOCHEMICAL ANALYSIS These samples derive from three excavations in central Stafford: the St Mary’s Of the forty-six samples included in this Grove and Bath Street sites, excavated study, seventeen were selected for stable as part of the campaign led by Carver, isotope analysis of charred cereal grains Cane and Cane between 1975 and 1985 (Supplementary Material 1). These samples, (Carver, 2010); and the Tipping Street all containing large quantities of well-pre- South site which was excavated by served grains, were chosen to ensure that Oxford Archaeology in 2009–2010 both Late Saxon and post-Conquest con- (Dodd et al., 2014). The quantitative texts were represented, and to allow the work presented here is based on the inclusion of grains from all four of the original archaeobotanical data produced cereals found at Stafford, i.e. hulled barley, by Lisa Moffett (1987) and Denise Druce free-threshing wheat, oats and rye. (2014). Additionally, grains from each of the same

Figure 2. John Speed’s town plan of Stafford, 1610 (published in 1676 by Bassett and Chiswell) Image copyright © 2000 by Cartography Associates (https://www.davidrumsey.com/luna/servlet/ detail/RUMSEY∼8∼1∼285263∼90057936:-Staffordshire—Stafford-Countie-#, accessed 31 January 2020). Licensed under Creative Commons Licence (https://creativecommons.org/licenses/by-nc-sa/ 3.0/, accessed 31 January 2020).

seventeen samples were selected for radiocar- Street began before AD 850 (Dodd et al., bon dating. These new radiocarbon dates, in 2014:92–93), this suggestion is not sup- combination with those obtained during ported by the model produced here. The previous excavations, were used to construct discrepancy is due partly to the inclusion a new chronological model (summary and of new radiocarbon dates, and partly to further detail in Supplementary Material 2). the application of an outlier model to Six samples belong to Phase 1, a period allow for the possibility of an old wood of activity at Tipping Street South dated effect in the dates derived from charcoal to between the late ninth and early tenth (Supplementary Material 2). Nonetheless, centuries, centred on three pottery kilns; these adjustments do not contradict substantial quantities of charred plant Oxford Archaeology’s essential conclusion remains associated with the kilns and that significant activity, including pottery nearby pits suggest that cereals were also production and crop processing, took place processed nearby (Dodd et al., 2014:10– at Stafford before the historically attested 31, 73). While the chronometric study establishment of the burh in AD 913. undertaken by Oxford Archaeology postu- Fifteen samples belong to Phase 2, lated that Late Saxon activity at Tipping which encompasses the mid-tenth to mid-

eleventh centuries, broadly spanning the sample belonging to each cereal genus early decades of the burh to the Norman (Figure 3; Supplementary Material 1). Conquest. The most substantial deposits Hulled barley occurs in negligible pro- of charred cereal remains from this phase portions at Stafford before Sub-Phase 2.2. were excavated from a group of ovens at This suggests that barley only made a St Mary’s Grove variously interpreted as notable contribution to the local cereal crop-driers and baking ovens (Moffett, economy from the latter part of the tenth 1994). century onwards (Figure 3). Most of the Twenty samples belong to Phase 3, barley remains are likely to represent a six- which spans the mid-eleventh to early row variety, but Moffett also identified thirteenth centuries. Most derive from St two-row barley amongst the rachis seg- Mary’s Grove, chiefly characterized in this ments (1987:5–6). Remains of two-row period by a sunken-featured structure con- barley are found only from the latter part taining rich deposits of charred cereal of the tenth century onwards, when pro- remains (Carver, 2010: 103). Other rich portions of barley increase more generally; samples derive from an extensive spread of together these trends suggest that barley charred material in the backfill of nearby cultivation in Stafford’s hinterland became large quarry pits (Moffett, 1987: 14). more productive and more diverse during Four samples belong to Phase 4, span- that period (Moffett, 1987:5–6). ning the thirteenth to sixteenth centuries, The lowest proportions of free-thresh- when Stafford became established as a ing wheat grain are found in the samples market town. These samples come from dating from between the late tenth and pits, ditches, and a kiln at St Mary’s mid-twelfth centuries (Sub-Phases 2.3 to Grove, the site of a medieval residence 3.1). Rye exhibits a converse pattern, with (Carver, 2010: 122). The one remaining notably high proportions of rye grain prin- sample studied here belongs to the post- cipally in the samples belonging to Sub- medieval Phase 5 and derives from a pit Phase 3.1 (Figure 3). The situation is near a putative brewery (Carver, 2010:35). reversed in Sub-Phase 3.2, when wheat grains tend to be present in higher proportions than rye grains. PATTERNS AMONG THE CEREAL CROPS It is during this same period (Sub- Phase 3.1), when wheat was apparently The crop contents of all forty-six samples less important than rye, that some diversi- are dominated by four cereals: hulled barley fication appears amongst the charred (Hordeum vulgare L.), wheat (Triticum wheat remains. Moffett’s analysis reveals L. free-threshing type), oat (Avena L.), that two species of free-threshing wheat and rye (Secale cereale L.). The predomin- were present at Stafford from around the ance of these cereals at Stafford in the mid-eleventh century onwards: bread-type Anglo-Saxon and medieval periods is con- wheat (Triticum aestivum L.) and rivet- sistent with national patterns (Moffett, type wheat (Triticum turgidum L.). 2011). The fact that not only grains but Whereas bread-type wheat appears to be also charred chaff items from all four ubiquitous in the Stafford assemblages, cereals occur in the assemblage implies rivet wheat chaff has been identified in that all were most probably grown locally. only six of the forty-six samples, all dating A broad picture of changes in crop choices from Phases 3 and 4. Its relative scarcity over time can be obtained by calculating at Stafford is consistent with the wider the relative proportions of grains in each national picture for this period (Moffett,

Figure 3. Percentage of cereal grains in samples grouped by phase.

2006: 49; Druce, 2014: 72). In addition, harsh for other crops to thrive, and there- Moffett has observed that some of the fore suggests that its occurrence at wheat grains dating from between the Stafford—in a lowland region of relatively mid-eleventh and mid-twelfth centuries hospitable farmland—is most probably were of a distinctive but archaeobotanically unintentional (Moffett, 1994: 63). nondescript ‘bullet-shaped’ form (Moffett, In summary, the relative proportions of 1987: 5). It may be that these grains charred cereal grain in the archaeobotani- represent not a different species or subspe- cal samples from Stafford suggest that, cies of free-threshing wheat, but a wheat between the mid-tenth and mid-twelfth landrace: a variety specifically adapted to centuries, the cultivation regimes feeding thrive in a particular local environment. the town underwent some diversification More work is needed to establish whether in crop choices, while rye increased in this is indeed the case. importance relative to wheat. Such pat- The relative proportions of oat grains terns provide an essential framework suggest that oats were a significant crop within which to consider the pollen, weed from the late ninth century onwards ecological, and biomolecular evidence for (Figure 3). Moffett’s analysis (1987:6)of key developments in crop husbandry strat- the preserved chaff indicates that two egies, notably land use, extensification, domesticated species of oat were present: and crop rotation. common oat (Avena sativa L.) and bristle oat (Avena strigosa Schreb.). Bristle oat has been identified only in samples dating LANDSCAPE AND LAND USE from between the late tenth and mid- twelfth centuries, whereas common oat is There are many pollen records from the present from the late ninth century Stafford region, but relatively few dated onwards. Moffett notes that bristle oat is sequences covering the period from AD usually only grown in environments too 450 to 1450. New, high resolution data

Figure 4. Pollen site locations. Map produced in ArcGIS Pro (ESRI) and ASTER GDEM (elevation).

from Stafford West and Cranberry Bed chronologies have been applied to allow in South Yorkshire are presented here, comparison between sites, but, as the together with data from the European number of radiocarbon dates for each site Pollen Database (IMBE, 2007), British and especially those falling within the Pollen Database (BPOL, M. Grant, pers. medieval period is variable, the error comm. 2019) and published diagrams margins within the age-depth models are (Hicks, 1971; Tallis & Switsur, 1973; equally variable. For example, four of the Beales, 1980; Brown, 1988; Bartley & seven dates for Featherbed Moss Morgan, 1990). Alluvial and enclosed (Derbyshire) are Anglo-Saxon to medieval, valley sites are expected to have small, while the Hipper Sick (also in Derbyshire) localized catchments, while lakes and chronology is extrapolated from a single upland bogs should represent broader, Bronze Age date (Supplementary Material regional trends (Figure 4; see Jacobson & 3). Radiocarbon dates rejected as too early Bradshaw, 1981). The published records or late by the original researchers have referred to here are some of the most been excluded from the models. detailed for the West Midlands and Establishing the changing proportions Southern Pennines yet few have more of broad vegetation types over time allows than one sample per century for the early us to identify large-scale changes in land- medieval period, making it difficult to cover, e.g. from woodland to open grass- establish exactly when changes in land use land. However, recognizing changes in occurred. For FeedSax, we are sampling land use within open landscapes, such as a cores at higher resolution for pollen and shift from pastoral to arable farming, radiocarbon dating, in order to understand requires closer examination of the herb- changes over shorter timescales. ‘Best fit’ aceous taxa present. Turner’s(1964)

Figure 5. Cumulative pollen percentage diagrams for the Stafford region showing changes in proportions of key vegetation over time.

Arable/Pastoral (AP) index, while prob- Low-lying areas around Stafford appear lematic in some ways, provides a rough to have been remarkably open throughout indication of land use trends. AP indices the Anglo-Saxon and medieval periods, are used here together with a thorough and in most cases by the later Roman assessment of the diversity and relative period (Figure 5); whereas, at the lowland proportions of pollen types associated with lake site of Crose Mere in Shropshire, tree grazing and cultivation. Cereal pollen iden- cover is higher, suggesting a more wooded tification requires measurement, and desig- landscape. Nevertheless, considering over- nations beyond ‘Cereal type/Cerealia’ were representation of tree pollen and site- not available for any of the published data- catchment biases, the latter landscape was sets. For Stafford West and Cranberry Bed still relatively open (see Tauber, 1965). AP (South Yorkshire), large grasses were cate- indices for the lowland sites are all gorized as: Hordeum type, which represents strongly arable by the mid-Roman period, barley but also large wetland grasses, if not earlier: all three cereal types are Avena-Triticum type (oats and wheats), or present at Stafford West, and cereals, Secale cereale (rye) (after Andersen, 1979; Cannabis/Humulus (hemp/hops), and a Tweddle et al., 2005). Methodologies for wide range of arable indicators are present sample preparation and critical identifica- at Crose Mere at this time (Figure 6 and tions are provided online, together with Supplementary Material 3). further information about the pollen sites Interestingly, by c. AD 550–650 there is (Supplementary Material 3). a small increase in the range and quantity

Figure 6. Relative proportions of arable and pastoral land use over time in the Stafford region. Inferred land use is based on the number and variety of indicator taxa present and AP (arable pastoral) indices (after Turner, 1964).

of arable indicators and/or cereals at most (e.g. Rippon et al., 2015). It is worth of the lowland sites (Figure 5). At Crose noting that both Stafford sites and Wilden Mere in this period, a slight increase in Marsh are likely to have had small catch- pasture is suggested by the AP index, but ments, and would strongly reflect local this is accompanied by an increase in cultivation, even if there was a decline in Cannabis/Humulus and the first occurrence the wider landscape. In contrast, Crose of Centaurea cyanus (cornflower), a weed Mere, a lake site, should represent a larger associated with Anglo-Saxon farming. catchment; there is a slight increase in Although there are error margins in the pasture here, but no arable decline. age-depth models, the fact that Stafford In the Mid-Saxon period (c. AD 650– West, King’s Pool, and Wilden Marsh 850), a dramatic increase in the variety (Worcestershire) show continuation or and percentages of cereals and arable indi- even expansion of arable farming cators occurs in Stafford, with Cannabis/ (Figure 6) suggests it is a real pattern, in Humulus present at both Stafford West contrast to evidence for an arable decline, and King’s Pool by the mid to late eighth or a major shift towards pastoral farming, century. These developments are consist- in eastern England at this time (e.g. ent with an increase in the scale of arable Banham & Faith, 2014). This adds farming, together with a change in weight to the conclusion that not all methods of cultivation and the crops being regions experienced a post-Roman decline grown. Significantly, this activity predates

Phase 1 at Stafford and the foundation of Hipper Sick is on Beeley Moor, which, the burh. None of the Wilden Marsh alongside copious Bronze Age remains, (a floodplain site in Worcestershire) has vestiges of medieval occupation samples date to this period, but by c. AD (Pastscape, 2019; e.g. Monuments 31104, 900 a range of arable indicators, including 311113, 311119, 1342500, 311053). It is Centaurea cyanus, is present. All the probable that the nearby lowlands were lowland sites reflect broadly arable land- farmed and that the changes in the pollen scapes throughout the Late Saxon period. record reflect the same Mid-Saxon At Crose Mere in Shropshire, arable changes seen around Stafford, though farming remains dominant until the late with more emphasis on pasture. eleventh century, when pastoral land use The ninth to tenth centuries see a increases; this may be linked to the con- marked decline in tree cover at Cranberry struction of a nearby Norman motte of Bed and Featherbed Moss, followed by an Stockett at Cockshutt (Pastscape, 2019: increase in arable and pastoral indicators. Monument 68643). By c. AD 1200, there is All three cereal types and Cannabis/ also a decline in arable farming around Humulus are present at Cranberry Bed. Wilden Marsh, although the wide spacing This indicates a Late Saxon woodland of samples makes it difficult to establish clearance, and an expansion and change in when this happens (Figures 5 and 6). methods of cultivation similar to that seen Cultivation remains important at Stafford in the Mid-Saxon period elsewhere. The West and King’s Pool until the thirteenth spread of arable farming into, or near to, to fourteenth centuries, when arable indica- upland areas suggests larger scale cultiva- tors decline and the AP index for the latter tion, while the marked increase in pastoral becomes more pastoral (Figure 6). These land use could be the result of rough changes might result from agricultural land grazing of livestock on heathland. shifting further away from the sampling A short-lived twelfth-century birch sites as the town increased in size. regeneration at Cranberry Bed ends with a The upland sites reflect much more substantial clearance event in c. AD 1160, varied landcover and a greater emphasis on with evidence for ongoing arable and pas- pastoral land use, as might be expected toral farming. Grazing and cultivation given the topography and prevalence of continue into the central Middle Ages peat/acid soils. The upland valley bog of despite a fourteenth-century expansion of Cranberry Bed (South Yorkshire) shows a heathland (Figures 5 and 6). A post-medi- relatively open landscape with Hordeum eval farmstead and pasture existed nearby type, Secale cereale and a wide range of (Bevan, 2003: 280), but the pollen data arable indicators present in the early show a much longer record of local Roman period, giving way to heathland farming. The landscape remains broadly expansion by c. AD 250, perhaps owing to pastoral in later periods at Hipper Sick farming-induced soil deterioration and and Featherbed Moss, though evidence for cooler/wetter conditions (see Hicks, 1971: cultivation persists (Figure 6). 666). There is little or no evidence of farming at Cranberry Bed or Featherbed Moss (Derbyshire) for the fifth to ninth EXTENSIFICATION centuries (Figure 6), but at Hipper Sick (also in Derbyshire) there is an increase in A functional ecological study of weed flora pasture and a slight rise in arable indica- developed under traditional agricultural tors in the seventh to eighth centuries. regimes in Asturias in Spain and Haute

Figure 7. a) The relationship of fields in Haute Provence (open circles) and fields in Asturias (filled circles) to the discriminant function extracted to distinguish these two groups (larger symbols indicate group centroids). b–e) The relationship of archaeobotanical samples from Stafford to the discriminant function (larger symbols indicate centroids for the modern groups). f) Correlations between the functional attribute scores used as discriminating variables and the discriminant function.

Provence in France successfully differen- and weeding) from those, like medieval tiated between high- and low-input open fields, receiving low inputs (low/no farming methods (Bogaard et al., 2016). manuring and weeding). This was The aim of the study was to use discrim- achieved on the basis of five functional inant analysis to develop a model for dif- traits that predict the response of weed ferentiating fields managed with high species to soil fertility and/or disturbance inputs per unit area (intensive manuring due to tillage and weeding: specific leaf

area (leaf area/leaf dry weight), canopy on levels of mechanical soil disturbance height and diameter, the ratio of leaf area and what this might reveal about plough- per node to fresh leaf thickness, and flow- ing regimes (Orwin & Orwin, 1938). At ering duration (Figure 7a). This model is Laxton we conducted a botanical survey of suitable for archaeobotanical studies since twenty sites, ranging from unploughed but it is based on the presence or absence of periodically grazed and annually cut ‘sykes’ weed species in different farming regimes, (hay meadow areas, unsprayed by herbi- rather than their ubiquity within individual cides, between or on the edges of the open fields. fields) to weed flora in arable fields, either Samples from Stafford containing at along cereal field edges without herbicide least ten seeds of weed taxa identified to spraying or in fallow fields (i.e. in the species level were entered into the classifi- third, fallow year of the rotation scheme) cation phase of the discriminant analysis that had not recently been sprayed. as unknown cases, in order to assess their Figure 8a shows the results of a discrimin- similarity to the modern high- versus low- ant analysis that separates (with 100 per input regimes. The results (Figure 7) show cent success) the grassland flora of the that late ninth- to early tenth-century ‘sykes’ from the weed flora of the arable samples (Phase 1) exhibit variable scores fields, using two attributes relating to soil on the discriminant function, albeit mostly disturbance as discriminating variables: within the low-input spectrum. From the flowering duration and (for perennials mid-tenth century onwards, however, only) vegetative propagation (i.e. the ability there is a clear tendency for the Stafford to regenerate from fragments of root/ weed data to conform to ‘low-input’ stolon/rhizome). Figure 8b–e shows, in growing conditions, resembling present- comparison, the discriminant scores of day extensive farming of cereals in Haute the Stafford samples, by phase. There is Provence. It should be acknowledged that an increasing tendency through time for within each phase there is some variation, the Stafford weed data to conform to the as observed among present-day fields more disturbed conditions of arable fields; (Figure 7a), but the results clearly point to none of the samples dating to the twelfth extensification. This indicates that, from century or later resembles the poorly dis- the mid-tenth century onwards, the great turbed grassland of the ‘sykes’.Thiscon- majority of cereals appear to have been trasts with earlier phases where samples grown with little or no manuring and resembling those from the sykes could hand-weeding. This evidence, combined represent the intermittent cultivation of with palynological evidence for largely land normally used for pasture (see open, arable landscapes around Stafford in Hooke, 1981:207)ortheinterfaceof this period (see above), implies that cereal arable and grassland. The implication is cultivation featured low-input manage- that, within the extensification process, ment within an extensive, large-scale arable fields at Stafford were more con- system in which land (rather than labour sistently and comprehensively disturbed inputs per unit area) was the limiting from the twelfth century onwards than in factor of production (see Bogaard et al., earlier periods. This observation is con- in press). sistent with a more pervasive use of the A study of the present-day arable and hay mouldboard plough in this final phase, meadow flora at Laxton, Nottinghamshire, when it appears that effectively all cereal known for its extant open field landscape, farming took place in heavily disturbed offers a complementary perspective focused conditions.

Figure 8. a) The relationship of Laxton arable fields (grey and black squares) and hay meadow ‘sykes’ (open squares) to the discriminant function extracted to distinguish these two groups (larger symbols indicate group centroids). b–e) The relationship of archaeobotanical samples from Stafford to the discriminant function (larger symbols indicate centroids for the modern groups). f) Correlations between the functional attribute scores used as discriminating variables and the discriminant function.

CROP ROTATION (Bogaard et al., 2001). A recent study of Anglo-Saxon charred plant remains found A functional ecological study of field that the application of correspondence surveys in modern Germany successfully analysis to quantified weed seed data, differentiated the weed flora of autumn- using the same functional attributes as the and spring-sown crops, using flowering German study, was effective at detecting onset and duration of arable weeds as a subtle patterns in sowing seasonality functional trait in discriminant analyses (McKerracher, 2019:96–124). The same

Figure 9. Distribution of cereal items and weed seeds in correspondence analysis output, coded by cereal type and seasonality association.

methods and criteria detailed in that study tended to be sown in the spring (Banham have been applied to the Stafford dataset. & Faith, 2014: 58). For Stafford, this is Correspondence analysis was performed most clearly the case between the mid- using Canoco for Windows 4.51 and tenth and early thirteenth centuries; the CanoDraw for Windows 4.1 programs evidence for earlier and later phases is less (after Braak & Smilauer, 2002). compelling because of the relatively small Two complementary trends emerge number of samples. However, these results from this analysis (Figure 9). On the x- cannot exclude the possibility that the axis, weed species associated with spring- cereals were simply sown at different times sowing tend to occur at the positive end, in different fields, rather than grown in along with barley grain and rachis. By rotation in the same fields. To establish contrast, weed species associated with with greater certainty whether rotation was autumn-sowing tend towards the negative practised, the evidence of crop stable iso- end of the x-axis, as do wheat and, espe- topes was considered. cially, rye; oat lies roughly between the Ratios of stable carbon isotopes in two groups. The patterns detected are, cereals are influenced by soil moisture thus, consistent with the idea, familiar during growth (Wallace et al., 2013): from later medieval history, that the wheat more positive δ13C values generally reflect and rye consumed at Stafford were typic- drier conditions, and negative values ally autumn-sown crops, whereas barley wetter conditions. Other environmental

Figure 10. a) Cereal grain δ13C values shown by the whole assemblage and b) by phase; c) cereal grain δ15N values shown by the whole assemblage and d) by phase.

factors affecting crop δ13Cvaluesinclude conditions, and hence that the expected slope, canopy cover, light intensity, and physiological offsets between crops would temperature (Heaton, 1999;Bogaard be expressed. et al., 2016). Beyond environmental The δ13C results from Stafford reflect factors, interpretation of crop δ13Cvalues the overall physiological differences expected must also take account of physiological between the four crops—rye, oat, barley, differences among species. The δ13C and free-threshing wheat—suggesting that values of mixed cereals grown under the they grew in similar conditions, i.e. in a same conditions show that oat is statistic- similar landscape with a similar amount of ally significantly lower than rye, while rye rainfall (Figure 10a). The δ13Cvaluesof appearstobecomparabletofree-thresh- each species remain relatively stable over ing wheat (see Supplementary Materials 4 time,withtheexceptionoffree-threshing and 5 for details). Other research shows wheat (see below). There is no significant that barley tends to be ∼1–2‰ lower difference between oat and barley in any of than wheat due to physiological differ- the periods, and oat and barley are always ences (Wallace et al., 2013; Styring et al., ∼1‰ or more offset from rye. 2017). If crops were grown systematically There is, however, a significant differ- in rotation, it is expected that their δ13C ence between wheat and rye during Phase values would reflect similar growing 3 (broadly, the mid-eleventh to early

Table 1. Results of the statistical tests conducted on stable isotope data from Stafford. Significant Tukey post hoc P-values shown in bold. ANOVA Material Isotopic data F-ratio Degrees of freedom P-value Tukey post hoc test

Wheat in Phases 1, 2, and 3 δ13C 22.9 2,8 <0.001 Phase 1 – Phase 3 <0.001 Phase 2 – Phase 3 0.002 Phase 1 – Phase 2 0.49 δ15N 0.814 2,8 0.476 Rye in Phases 1, 2, and 3 δ13C 0.133 2,10 0.877 δ15N 5.68 2,10 0.023 Phase 1 – Phase 3 0.018 Phase 2 – Phase 3 0.83 Phase 1 – Phase 2 0.096 Barley in Phases 2 and 3 δ13C t(1.4)* 1,4 0.235 δ15N t(-2.8)* 1,4 0.05 Oat in Phases 1, 2, and 3 δ13C 1.89 2,13 0.19 δ15N 3.322 2,13 0.068 Phase 1, all species δ13C 20.14 2,6 0.022 Oat – Wheat 0.003 Rye – Wheat 0.979 Rye – Oat 0.004 δ15N 0.377 2,6 0.701 Phase 2, all species δ13C 15.94 3,12 <0.001 Wheat – Barley 0.045 Oat – Barley 0.204 Rye – Barley 0.059 Oat – Wheat <0.001 Rye – Wheat 1 Rye – Oat <0.001 δ15N 3.325 3,12 0.057 Phase 3, all species δ13C 45.25 3,17 <0.001 Wheat – Barley 0.063 Oat – Barley 0.301 Rye – Barley <0.001 Oat – Wheat <0.001 Rye – Wheat 0.002 Rye – Oat <0.001 δ15N 2.819 3,17 0.07 Sub-Phase 3.1, all species δ15N 9.078 3,14 0.001 Wheat – Barley 0.001 Oat – Barley 0.004 Rye – Barley 0.015 Oat – Wheat 0.513 Rye – Wheat 0.186 Rye – Oat 0.799 All species by phase δ15N 9.342 2,43 <0.001 Phase 1 – Phase 3 <0.001 (not separated by species) Phase 2 – Phase 3 0.044 Phase 1 – Phase 2 0.115 All species by phase, δ15N 8.397 2,37 <0.001 Phase 1 – Phase 3 0.001 except barley Phase 2 – Phase 3 0.035 Phase 1 – Phase 2 0.3

* T-test conduct instead of ANOVA

thirteenth centuries), a departure from the becomes less enriched in 13C, and is pattern of similar values for the earlier statistically significantly different from rye, periods (Table 1; Figure 10b). In Phase 2, although it should be noted that there there is no statistical difference between are only four wheat samples from Phase 3. wheat and rye; whereas, in Phase 3, wheat Several causes for this change are possible.

Figure 11. Cereal grain δ15N results from Stafford shown by phase, regardless of species.

The samples analysed may represent including environmental conditions, such wetter years, or wetter soils; the low as de-nitrification in waterlogged or peri- number of samples analysed could also odically wet conditions, as well as have produced a misleading picture. anthropogenic causes, notably manuring Alternatively, the isotopic difference could (Handley et al., 1999; Bogaard et al., represent a change in the type of wheat 2007; Fraser et al., 2011). Overall mean cultivated; as mentioned above, the δ15N values indicate similar soil 15N for archaeobotanical remains from Phase 3 oat, rye, and wheat (Figure 10c), while include the chaff of rivet wheat and barley is consistently elevated in 15N com- ‘bullet-shaped’ grains of an otherwise pared to the other crops in Phases 2 and 3 unidentified variety, possibly a local land- (no barley samples were available for race. It is therefore possible that the Phase 1) (Figure 10d). However, there is change in wheat isotopic values between no statistically significant difference Phases 2 and 3 is the consequence of a between the species’ δ15N values within switch to a variety with lower δ13C, either Phases 2 and 3, potentially because of the for physiological reasons or because it was uneven distribution of samples between cultivated in wetter soils. It is of course species (Table 1). If Sub-Phase 3.1, with also possible that wheat was no longer the highest number of samples per species, grown in rotation with the other species is examined, barley is statistically signifi- in Phase 3, although this seems unlikely cantly different from the other three crop given that the δ15N values for wheat species (Table 1). Though the limited remain consistent with rotation through- number of samples per species per phase out all periods (see below). must be borne in mind, it appears that Ratios of the stable nitrogen isotopes barley was cultivated in soil more enriched in cereals reflect soil N composition. in 15N than soil where the other three Multiple factors can affect this ratio, crops were cultivated, and therefore was

not grown in rotation with them. The of the eleventh century, to cultivate previ- limited differences between rye, oat, and ously saturated soils, as lower rainfall and/ free-threshing wheat, on the other hand, or higher temperatures dried out previ- suggest that these were grown in rotation. ously saturated land (Xoplaki et al., 2011). A general trend of increasing 15N It is also likely that use of the mouldboard enrichment over time can be seen plough to create ridge-and-furrow (Figure 11). Rye δ15N values change sig- improved drainage in wetter soils. nificantly between Phase 1 (5.3‰) and Phase 3 (8‰), while barley shows a significant difference between Phase 2 CONCLUSION (8.7‰) and Phase 3 (10.6‰)(Table 1). There is no statistical difference among By bringing together the results of the the means of oat or wheat from any phase various analyses presented above, it has but the mean values of oat and wheat do been possible to draw certain conclusions increase over time (for oat, Phase 1: regarding the farming regimes that fed the 5.4‰, Phase 3: 8.0‰; for wheat, Phase 1: population of Stafford from the late ninth 5.9‰, Phase 3: 7.1‰). When all sample to thirteenth centuries. values are averaged irrespective of species, The strongly arable signal in the pollen an increase in δ15N values and ranges for data from Stafford and the immediately the three phases is evident with Phases 1 surrounding region points to an emphasis and 2 statistically different from Phase 3 on cereal farming beginning well before (Table 1). Phases 1 and 2 remain statistic- the foundation of the burh in AD 913. In ally significantly different from Phase 3 the Mid-Saxon period, a broader range of even when barley is removed. crops and associated weeds appears, indi- The increasing enrichment of 15N over cating an increase in the scale of arable time could be due to several factors. The farming and possibly changes in farming lack of wild fauna to create a ‘herbivore practices. This trend mirrors changes baseline’ (e.g. Styring et al., 2016) means evident more widely across England, not that understanding whether this enrich- only in the archaeobotanical record but ment is due to arable practices (e.g. also in the structure and character of Mid- manuring) or to wider environmental Saxon settlements, with growing evidence factors relies on other strands of evidence, of enclosures such as paddocks and pens particularly functional weed ecology. The as well as droveways pointing to new ways weed results, as discussed above, indicate of managing the movement of livestock, that fertility did not increase over time, while the appearance of barns, corn driers, which means that the increasing 15N and watermills suggests that cereal sur- enrichment is unlikely to be a consequence pluses were being mobilized in new ways of manuring. Other possibilities include (see Hamerow, 2012; McKerracher, seasonal wetting and drying and/or water- 2018). The arable signal remains domin- logging of heavier, poorly draining clays ant in the pollen record until the thir- and seasonally wet soils surrounding teenth century; the decline at this time Stafford (Figure 1). This is not, however, could reflect the growth of Stafford supported by the δ13C values which, apart pushing agricultural activity further away from wheat, show only limited change from the town. during the later phases. It is possible that An expansion or shift onto heavier, less changes due to the Medieval Climate well-drained soils is suggested by the Anomaly enabled farmers, by the middle increasing enrichment of soil N values

from around the mid-eleventh century. implying that this cereal was grown in dif- This enrichment was not the result of ferent conditions. The fact that barley was increased manuring, as the weed data also underrepresented in the archaeobota- clearly show. Such a shift or expansion nical record supports the conclusion that is supported by the distribution of the barley consumed in Stafford was not ridge-and-furrow earthworks in the vicin- grown in rotation with the other cereals. ity of Stafford. The carbon (δ13C) values It should be emphasized that no single also indicate that, from the mid-eleventh analytical method on its own would have century onwards, a different variety of made it possible to draw such a detailed free-threshing wheat, possibly rivet picture. Only by comparing the results wheat, and potentially a new landrace of from crop stable isotope analysis, pollen wheat, may have been consumed in analysis, functional weed ecology, and Stafford. quantitative archaeobotanical analysis, has A shift to extensive, low-input farming such a holistic and dynamic picture of regimes between the mid-tenth and mid- farming practices during this critically eleventh centuries is clearly indicated by a important period emerged. The case study functional ecological analysis of the weed of Stafford thus illustrates the potential of flora. The wide spectrum of regimes integrated approaches to resolve long- apparent during the later ninth to early standing debates about the spread of low- tenth century, i.e. from relatively intensive input cultivation, specifically open field to highly extensive, was replaced in this farming, in medieval Europe. later period by a greater uniformity of practice and the widespread adoption of highly extensive regimes, presumably SUPPLEMENTARY MATERIAL representing some form of open field farming. Comparison with modern weed To view supplementary material for this data from Laxton in Nottinghamshire, article, please visit https://doi.org/10.1017/ furthermore, suggests a growing tendency eaa.2020.6. over time for the Stafford crops to be grown in highly disturbed conditions, consistent with increasing use of the heavy ACKNOWLEDGEMENTS plough. The ninth and tenth centuries also see woodland clearance at upland The FeedSax project is supported by the sites, together with an increase in arable European Research Council (ERC) under indicators and cereal types, supporting the the European Union’s Horizon 2020 conclusion that a larger proportion of the research and innovation programme under landscape came under cultivation at this grant agreement no. 741751. The authors time. A corresponding increase in pastoral would like to thank the following for their indicators suggests growing use of less cul- advice and assistance, notably in providing tivable heathland areas for grazing. access to and permission to analyse archae- Crop rotation is indicated from at least obotanical material from Stafford: Lisa the mid-tenth century onwards, by the Moffett (Historic England), Peter weed flora indicating the presence of Ditchfield (School of Archaeology, autumn- and spring-sown cereals, and by University of Oxford), Mike Kimber comparable δ13C and δ15N values for dif- (Headland Archaeology), Scott Timpany ferent cereals. The δ15N values for barley, (University of the Highlands and Islands), however, were different across all phases, Michael Grant (University of

Southampton), Alison Nicholls Vegetation History and Archaeobotany, 25: (Gladstone Pottery Museum, Stoke-on- 57–73. https://doi.org/10.1007/s00334- Trent), and Joseph Perry (Potteries 015-0524-0 Bogaard, A., Fochesato, M. & Bowles, S. in Museum, Stoke-on-Trent). press. The Farming-Inequality Nexus: New Insights from Ancient Western Eurasia. Antiquity. Brown, A.G. 1988. The Palaeoecology of REFERENCES Alnus (Alder) and the Postglacial History of Floodplain Vegetation: Pollen Andersen, S.T. 1979. Identification of Wild Percentages and Influx Data from the Grass and Cereal Pollen. Danmarks West Midlands, United Kingdom. New Geologiske Undersøgelse, Årbog, 1978: 66–92. Phytologist, 110: 425–36. Banham, D. & Faith, R. 2014. Anglo-Saxon Carver, M. 2010. The Birth of a Borough: An Farms and Farming. Oxford: Oxford Archaeological Study of Anglo-Saxon University Press. Stafford. Woodbridge: Boydell Press. Bartley, D.D. & Morgan, A.V. 1990. The Cotswold Archaeology 2010. Land North of Palynological Record of the King’s Pool, Beaconside, Stafford. Archaeological Stafford, England. New Phytologist, 116: Desk-Based Assessment. Cotswold 177–94. Archaeology Report 10004, Cirencester. Beales, P.W. 1980. The Late Devensian and https://www.staffordbc.gov.uk/sites/default/ Flandrian Vegetational History of Crose files/cme/DocMan1/Programme%20Officer/ Mere, Shropshire. New Phytologist, 85: Programme%20Officer%202013/J32— 133–61. Land-North-of-Beaconside-Archaelogical- Bevan, W. 2003. The Upper Derwent: Long- Desk-Based-Assessment.pdf term Landscape Archaeology in the Peak Dodd, A., Goodwin, J., Griffiths, S., Norton, District. (unpublished PhD dissertation, A., Poole, C. & Teague, S. 2014. Department of Archaeology, University of Excavations at Tipping Street, Stafford, Sheffield). Dataset at Archaeology Data 2009–10: Possible Iron Age Roundhouses, Service (distributor), York, available at: Three Stafford-type Ware Kilns, and https://doi.org/10.5284/1000267 Medieval and Post-Medieval Urban Blair, W.J.B. 2018. Building Anglo-Saxon Remains. Staffordshire Archaeological and England. Princeton: Princeton University Historical Society Transactions, 47: 1–114. Press. Druce, D. 2014. Charred Plant Remains. In: Bogaard, A., Jones, G., Charles, M. & A. Dodd, J. Goodwin, S. Griffiths, Hodgson, J.G. 2001. On the A. Norton, C. Poole & S. Teague, Archaeobotanical Inference of Crop Possible Iron Age Roundhouses, Three Sowing Time Using the FIBS Method. Stafford-type Ware Kilns, and Medieval Journal of Archaeological Science, 28: 1171– and Post-Medieval Urban Remains. 83. https://doi.org/10.1006/jasc.2000.0621 Staffordshire Archaeological and Historical Bogaard, A., Heaton, T., Poulton, P. & Society Transactions, 47: 65–75. Merbach, I. 2007. The Impact of Duby, G. 1954. La révolution agricole Manuring on Nitrogen Isotope Ratios in médiévale. Revue de Géographie de Lyon, Cereals: Archaeological Implications for 29: 361–66. Reconstruction of Diet and Crop Dyer, C., Thoen, E. & Williamson, T. 2018. Management Practices. Journal of The Rationale of Open Fields: A Archaeological Science, 34: 335–43. https:// Collection of Essays. In: C. Dyer, E. doi.org/10.1016/j.jas.2006.04.009 Thoen & T. Williamson, eds. Peasants Bogaard, A., Hodgson, J., Nitsch, E., Jones, and Their Fields: The Rationale of Open G., Styring, A., Diffey, C., et al. 2016. Field Agriculture 700–1800. Turnhout: Combining Functional Weed Ecology and Brepols, pp. 1–4. Crop Stable Isotope Ratios to Identify Fraser, R., Bogaard, A., Heaton, T., Charles, Cultivation Intensity: A Comparison of M., Jones, G., Christensen, B., et al. Cereal Production Regimes in Haute 2011. Manuring and Stable Nitrogen Provence, France and Asturias, Spain. Isotope Ratios in Cereals and Pulses:

Towards a New Archaeobotanical the Botanical Evidence for the Pre- Approach to the Inference of Land Use Burh Economy. In: J. Rackham, ed. and Dietary Practices. Journal of Environment and Economy in Anglo-Saxon Archaeological Science, 38: 2790–804. England. York: Council for British https://doi.org/10.1016/j.jas.2011.06.024 Archaeology, pp. 55–64. Hall, D. 2014. The Open Fields of England. Moffett, L. 2006. The Archaeology of Oxford: Oxford University Press. Medieval Plant Foods. In: C.M. Woolgar, Hamerow, H. 2012. Rural Settlements and D. Serjeantson & T. Waldron, eds. Food Society in Anglo-Saxon England. Oxford: in Medieval England. Oxford: Oxford Oxford University Press. University Press, pp. 41–55. Handley, L., Austin, A., Robinson, D., Moffett, L. 2010. Plant Studies, Report 10.1. Scrimgeour, C.M.A., Raven, J.A.C., In: M. Carver, ed. The Birth of a Borough: Heaton, T., et al. 1999. The 15N Natural An Archaeological Study of Anglo-Saxon Abundance (δ15N) of Ecosystem Samples Stafford. Woodbridge: Boydell Press, Reflects Measures of Water Availability. p. 121 and online report available at Australian Journal of Plant Physiology, 26: https://archaeologydataservice.ac.uk/archive 185–99. DS/archiveDownload?t=arch-912-1/dissemi Heaton, T. 1999. Spatial, Species, and nation/pdf/Stafford_Field_Reports_10_ Temporal Variations in the 13C/12C Plants/FR_10.1_Plant_Macrofossils_Lisa_ Ratios of C3 Plants: Implications for Moffet.pdf Palaeodiet Studies. Journal of Archaeological Moffett, L. 2011. Food Plants on Science, 26: 637–49. https://doi.org/10. Archaeological Sites: The Nature of the 1006/jasc.1998.0381 Archaeobotanical Record. In: H. Hicks, S. 1971. Pollen-analytical Evidence for Hamerow, D.A. Hinton & S. Crawford, the Effect of Prehistoric Agriculture on eds. The Oxford Handbook of Anglo-Saxon the Vegetation of North Derbyshire. New Archaeology. Oxford: Oxford University Phytologist, 70: 647–67. Press, pp. 346–60. Hooke, D. 1981. Anglo-Saxon Landscapes of Orwin, C.S. & Orwin, C.S.L. 1938. The Open the West Midlands: The Charter Evidence. Fields of England. Oxford: Oxford Oxford: British Archaeological Reports. University Press. IMBE (Institut Méditerranéen de la Pastscape. 2019. PastScape [online] [accessed Biodiversité et d’Ecologie marine et 17 July 2019]. Available at: https://www. continentale, Aix-Marseille) 2007. The pastscape.org.uk/ European Pollen Database [online] Rippon, S., Smart, C. & Pears, B. 2015. The [accessed 7 January 2020]. Available at: Fields of Britannia: Continuity and Change Landscape. Oxford: Oxford University Jacobson, G.L.J. & Bradshaw, R.H.W. 1981. Press. The Selection of Sites for Roberts, B. & Wrathmell, S. 2000. An Atlas of Palaeovegetational Studies. Quaternary Rural Settlement in England. London: Research, 16: 80–96. English Heritage. McKerracher, M. 2018. Farming Transformed in Styring, A.K., Ater, M., Hmimsa, Y., Fraser, Anglo-Saxon England: Agriculture in the R., Miller, H., Neef, R., et al. 2016. Long Eighth Century.Oxford:Windgather. Disentangling the Effect of Farming Practice McKerracher, M. 2019. Anglo-Saxon Crops from Aridity on Crop Stable Isotope Values: and Weeds: A Case Study in Quantitative A Present-day Model from Morocco and its Archaeobotany. Oxford: Archaeopress. Application to Early Farming Sites in the Moffett, L. 1987. The Macro-botanical Eastern Mediterranean. The Anthropocene Evidence from Late Saxon and Early Review,3:2–22. https://doi.org/10.1177/ Medieval Stafford. Unpublished Ancient 2053019616630762 Monuments Laboratory Report 169/87. Styring, A.K., Charles, M., Fantone, F., Hald, London: Historic Buildings and M., McMahon, A., Meadow, R., et al. Monuments Commission for England. 2017. Isotope Evidence for Agricultural Moffett, L. 1994. Charred Cereals from Some Extensification Reveals How the World’s Ovens/Kilns in Late Saxon Stafford and First Cities Were Fed. Nature Plants,3:

1–11. https://doi.org/10.1038/nplants. [email: [email protected]]. 2017.76 ORCID 0000-0001-5643-5888 Tallis, J.H. & Switsur, V.R. 1973. Studies on Southern Pennine Peats VI: A Radiocarbon-dated Pollen Diagram from Featherbed Moss, Derbyshire. Journal of Amy Bogaard is Professor of Neolithic Ecology, 61: 743–51. and Bronze Age Archaeology in the Tauber, H. 1965. Differential Pollen School of Archaeology, University of Dispersion and the Interpretation of Oxford. Her research centres on the Pollen Diagrams. Denmarks Geologiske Undersøgelse, Raekke 2, 89: 1–69. ecology, evolution, and social implications ter Braak, C. & Smilauer, P. 2002. CANOCO of past farming systems in Europe and Reference Manual and CanoDraw for western Asia. Windows User’s Guide: Software for Canonical Community Ordination (version Address: Amy Bogaard, School of 4.5). Ithaca, NY, and Wageningen: – Microcomputer Power & Biometris. Archaeology, University of Oxford, 34 6 Turner, J. 1964. The Anthropogenic Factor in Beaumont Street, Oxford OX1 2PG. Vegetational History I: Tregaron and [email: [email protected]]. Whixall Mosses. New Phytologist,63:73–90. ORCID 0000-0002-6716-8890 Tweddle, J.C., Edwards, K.J. & Fieller, N.R.J. 2005. Multivariate and Other Statistical Approaches for the Separation of Cereal from Wild Grass Pollen Using a Large Mike Charles is Senior Research Fellow in Holocene Dataset. Vegetation History and Environmental Archaeology in the School Archaeobotany, 14: 15–30. of Archaeology, University of Oxford. His Wallace, M., Jones, G., Charles, M., Fraser, research is on the origins and spread of R., Halstead, P., Heaton, T. & Bogaard, agriculture and plant domestication in A. 2013. Stable Carbon Isotope Analysis as a Direct Means of Inferring Crop western Asia and Europe from c. 15,000 Water Status and Water Management BC to the modern era. Practices. World Archaeology, 45: 388–409. https://doi.org/10.1080/00438243.2013. Address: Michael Charles, School of 821671 Archaeology, University of Oxford, 34-6 White, L. 1940. Technology and Invention in the Middle Ages. Speculum, 15: 141–59. Beaumont Street, Oxford OX1 2PG. Xoplaki, E., Fleitmann, D. & Diaz, H. 2011. [email: [email protected]]. Editorial: Medieval Climate Anomaly. PAGES News, 19.1: 4. Emily Forster is a post-doctoral researcher in the School of Archaeology at the BIOGRAPHICAL NOTES University of Oxford, and a freelance pollen and diatom specialist. After her PhD on Helena Hamerow is Professor of Early palaeoecological evidence for early medieval Medieval Archaeology at the University of land use in the Lake District (University of Oxford. Her research focuses on the Southampton), she worked for English archaeology of rural communities in Heritage, then at the University of Sheffield northwest Europe c. AD 400–1000, espe- on the NERC-funded ‘Out of Asia’ project cially in the North Sea zone. and ERC-funded ‘Evolutionary Origins of Agriculture’ project. Address: Helena Hamerow, Institute of Archaeology, University of Oxford, 34–6 Address: Emily Forster, School of Beaumont Street, Oxford OX1 2PG. Archaeology, University of Oxford, 34–6

Beaumont Street, Oxford OX1 2PG. Cambridge) is interested in the develop- [email: [email protected]]. ment of scientific methods in archae- ORCID 0000-0002-0934-1931 ology and their application to study changes in landscape use and dietary practice in the past. She currently works Matilda Holmes is a consultant zooarch- as a laboratory technician in stable aeologist working on British animal bone isotope analysis at the Research assemblages, but also involved in work in Laboratory for Archaeology and the Italy, Russia, Libya, and Afghanistan. History of Art, University of Oxford. Alongside commercial contracts, Matilda hasworkedasalecturerattheuniversitiesof Address: Samantha Neil, School of Birmingham, Nottingham, Leicester, and Archaeology, 1 South Parks Road, UCL. She is involved in numerous outreach University of Oxford, Oxford OX1 3TG. projects at local schools and serves as an [email: [email protected]]. advisor to the team working on the Roman ORCID 0000-0002-8850-8719 villa at Piddington, Northamptonshire.

Address: Matilda Holmes, School of Christopher Ramsey is Director of the Archaeology and Ancient History, Research Laboratory for Archaeology and University of Leicester, University Road, the History of Art, University of Oxford Leicester LE1 7RH. [email: Matty@ and of the Oxford Radiocarbon archaeozoology.co.uk]. ORCID 0000- Accelerator Unit. His research focuses on 0002-1704-5024 the application of physical sciences to archaeology and the environmental sciences and in particular in the use of Mark McKerracher is a post-doctoral radiocarbon isotope studies and numerical researcher in the School of Archaeology at modelling methods in chronology. the University of Oxford, where he com- pleted his DPhil on Mid-Saxon agriculture Address: Christopher Bronk Ramsey, School in 2014. After working in museum archiv- of Archaeology, 1 South Parks Road, ing, software development, and freelance University of Oxford, Oxford OX1 3TG. archaeobotany, he is currently researching [email: [email protected]]. medieval farming practices as part of the ORCID 0000-0002-8641-9309 ‘Feeding Anglo-Saxon England’ project. His interests include archaeobotany, database development, agricultural production, Elizabeth Stroud is a post-doctoral and Anglo-Saxon archaeology. researcher in bioarchaeology for the ‘Feeding Anglo-Saxon England’ project. Address: Mark McKerracher, School of She specializes in archaeobotany and stable Archaeology, University of Oxford, 34-6 isotope analysis at the University of Oxford. Beaumont Street, Oxford OX1 2PG. Her research has focused on understanding [email: [email protected]]. crop-based agricultural production in mul- ORCID 0000-0002-0299-6687 tiple regions of the world. Her research for her doctoral thesis used archaeobotany and stable isotope analysis to investigate Samantha Neil (PhD the University of Chalcolithic crop husbandry and social Durham, MPhil University of organization in Central Anatolia.

Address: Elizabeth Stroud, School of on the study of human-animal relations, Archaeology, University of Oxford, 34-6 primarily in the medieval and early Beaumont Street, Oxford OX1 2PG. modern period and the study of animal [email: [email protected]]. health, disease and injury. ORCID 0000-0003-4299-6638 Address: Richard Thomas, School of Archaeology and Ancient History, Richard Thomas is an archaeologist based University of Leicester, University Road, at the University of Leicester with expert- Leicester LE1 7RH. [email: Rmt12@leices- ise in zooarchaeology. His research centres ter.ac.uk]. ORCID 0000-0003-1207-793X

Une approche bioarchéologique intégrée envers la ‘révolution agricole : Stafford en Angleterre, un cas d’étude d’environ 800 à 1200 apr. J.-C.

Dans une grande partie de l’Europe, l’arrivée entre d’environ 800 et 1200 apr. J.-C de régimes d’agriculture céréalière utilisant peu de ressources a permis aux propriétaires terriens (seigneurs) de s’enrichir en accroissant considérablement la surface des terres cultivées et en exploitant le labeur des autres. L’analyse scientifiques de restes de plantes et d’animaux provenant de contextes archéologiques permet d’obtenir les premières preuves directes de l’évolution de ce genre de régime agricole. Dans cet article, les collaborateurs du projet FeedSax exposent leurs méthodes, destinées à résoudre des questions clés concer- nant la « céréalisation » de la campagne au Moyen Age, et présentent leurs résultats préliminaires provenant du cas d’étude de Stafford. Ces résultats indiquent une augmentation de la cultivation à l’époque anglo-saxonne moyenne et un passage à un régime utilisant peu de ressources et vraisemblablement une expansion vers des terres plus lourdes à l’époque anglo-saxonne tardive. L’assolement semble avoir été pratiqué depuis au moins le milieu du Xe siècle. Translation by Madeleine Hummler

Mots-clés: agriculture médiévale, Angleterre à l’époque anglo-saxonne, isotopes stables de plantes cultivées, écologie fonctionnelle des mauvaises herbes, Stafford, champs ouverts

Eine integrierte bioarchäologische Vorgehensweise zur Frage der „landwirtschaftlichen Revolution“: eine Fallstudie aus Stafford, England, von ca. 800 bis 1200 n. Chr.

In weiten Teilen Europas waren die Grundbesitzer (Landherren) in der Lage, durch die Einführung einer Getreidelandwirtschaft mit geringem Aufwand zwischen ungefähr 800 und 1200 n. Chr. reich zu werden, indem sie ihre Äcker maßgeblich erweiterten und die Arbeitskraft ausbeuteten. Die wis- senschaftliche Untersuchung von pflanzlichen Resten und Tierknochen aus archäologischen Befunden ermöglicht es, die ersten direkten Hinweise über die Entwicklung von solch einer Landwirtschaft zu erhalten. In diesem Artikel exponieren die Mitarbeiter des FeedSax Projektes ihre Methoden, um Kernfragen zur Intensivierung des Getreideanbaus (“Getreidierung”) in der mittelalterlichen Landschaft zu lösen, und stellen die vorläufigen Ergebnisse der Fallstudie von Stafford vor. Diese zeigen, dass der Ackerbau in der mittelangelsächsischen Phase zugenommen hat und dass eine Landwirtschaft mit geringem Aufwand und wahrscheinlich eine Ausdehnung auf schwereren Böden in der spätangelsächsischen Phase aufgenommen wurde. Die Wechselwirtschaft wurde scheinbar seit mindestens der Mitte des 10. Jahrhunderts betrieben. Translation by Madeleine Hummler

Stichworte: mittelalterliche Landwirtschaft, England in angelsächsischer Zeit, stabile Isotopen von Kulturpflanzen, funktionelle Ökologie von Unkräuter, Stafford, offene Felder