Farming Regions in Medieval England: The Archaeobotanical and Zooarchaeological Evidence

Abstract

REGIONAL VARIATION IN LANDSCAPE CHARACTER has in the past been studied by archaeologists in terms of its physical manifestations such as different settlement patterns and field systems. Local and regional distinctiveness in landscape character also results from how rural communities practised different agricultural regimes, and historians have long recognised the extent to which these varied across the country. Archaeologists, in contrast, have compared the animal bones and cereal remains from sites of different socio-economic status, but have not previously focused on the extent to which they vary across different geologies. This paper therefore presents an analysis of the animal bones of the three main domesticates (cattle, sheep/goat and pig), and the charred grains of the four main cereal crops (bread wheat, barley, oats and rye), across a series of different surface geologies within a study area extending from East Anglia down to the South-West Peninsula. It shows that, first, patterns of animal husbandry and cereal cultivation varied considerably across different surface geologies; secondly that, while farming practices do appear to have been influenced by surface geologies, they were also affected by cultural factors, particularly as human communities responded to the opportunities of a growing market economy; and thirdly that, while archaeobotanical and zooarchaeological patterns evident in the mid-11th–mid-14th centuries conform with what documentary sources tell us, the particular importance of this archaeological dataset is that it allows us to reconstruct farming regimes back into the undocumented early medieval (and indeed earlier) periods.

Résumé

Régions agricoles de l’Angleterre médiévale: éléments archéobotaniques et zooarchéolo­giques par Stephen Rippon, Adam Wainwright et Chris Smart

Les archéologues se sont déjà penchés sur les variations régionales du caractère du paysage en termes de manifestations physiques, comme par exemple les différences au niveau des populations qui s’installent et de la délimitation des champs. Le caractère distinctif local et régional du paysage résulte également des différents régimes d’agriculture pratiqués par les communautés rurales, dont les historiens reconnaissent depuis longtemps l’ampleur des variations régionales à travers le pays. Les archéologues, de leur côté, ont comparé les vestiges d’ossements d’animaux et de céréales provenant de sites de divers statuts socio-économiques, mais sans se pencher sur l’étendue des variations en fonction des géologies. Cet article présente donc une analyse des ossements des trois principaux groupes d’animaux domestiques (bétail, moutons/chèvres, et porcs) et des graines carbonisées de quatre principales cultures céréalières (blé tendre, orge, avoine et seigle), sur toute une série de géologies de surface au sein d’une zone d’étude allant de la région East Anglia jusqu’à la péninsule sud-ouest (Cornouailles et Devon). Ceci montre d’abord que les tendances en matière d’élevage et de culture céréalière ont varié considérablement selon les géologies de surface; ensuite, si les pratiques agricoles semblent bien avoir été influencées par les géologies de surface, elles ont également été marquées par des facteurs culturels, particulièrement à mesure que les communautés humaines réagissaient face aux opportunités d’une économie de marché en plein essor; enfin, si les modèles archéobotaniques et zooarchéologiques mis en évidence du milieu du 11^ème au milieu du 14^ème siècle correspondent à ce que nous disent les sources documentaires, cet ensemble de données archéologiques est d’intérêt crucial car il nous permet de reconstituer des régimes agricoles en remontant jusqu’au début du Moyen Âge (voire à des périodes plus anciennes), pour lequel on n’a aucune référence manuscrite.

Zusammenfassung

Landwirtschaftliche Gebiete im mittelalterlichen England: archäobotanische und zooarchäologische Belege von Stephen Rippon, Adam Wainwright und Chris Smart

Regionale Unterschiede im Charakter von Landschaften wurden von den Archäologen in der Vergangenheit hinsichtlich ihres physische­n Ausdrucks wie zum Beispiel in verschiedenen Siedlungsmustern und Feldsystemen untersucht. Die lokale und regionale Unverwechselbarkeit eines Landschaftscharakters entsteht auch daraus, wie ländliche Gemeinschaften verschiedene landwirtschaftliche Praktiken durchführten, und Historiker haben längst erkannt, wie sehr sich diese im Land unterschieden. Archäologen haben dagegen Tierknochen und Zerealienüberreste von Fundstätten mit unterschiedlichem sozio-ökonomischem Status verglichen, aber bisher den Fokus nicht darauf gelegt, wie sehr sich diese unter verschiedenen geologischen Begebenheiten unterscheiden können. Dieser Artikel stellt daher eine Analyse von Tierknochen der drei domestizierten Haupttierarten (Rinder, Schafe/Ziegen und Schweine) vor, sowie der verkohlten Überreste der vier wichtigsten geernteten Zerealien (Brotweizen, Gerste, Hafer und Roggen) vor, und zwar für eine Reihe verschiedener Oberflächengeologien innerhalb eines Untersuchungsgebiets, das sich von East Anglia bis hinunter zur Halbinsel im Südwesten (Cornwall und Devon) erstreckt. Es wird gezeigt, dass erstens sich die Muster der Tierhaltung und des Zerealienanbaus in den verschiedenen Oberflächengeologien erheblich unterschieden; dass zweitens zwar die landwirtschaftlichen Praktiken anscheinend von der Oberflächengeologie beeinflusst wurden, aber auch von kulturellen Faktoren, insbesondere als die menschlichen Gemeinschaften auf die Möglichkeiten einer wachsenden Marktwirtschaft reagierten; und dass sich drittens zwar die archäobotanischen und zooarchäologischen Muster, die sich zwischen der Mitte des 11 und der Mitte des 14 Jahrhunderts zeigten, mit dem decken, was uns die Dokumentenquellen sagen, es aber die besondere Bedeutung dieses archäologischen Datensatzes ist, dass er uns erlaubt, landwirtschaftliche Systeme bis in die nicht dokumentierte frühmittelalterliche (und sogar noch früherer) Zeit hinein zu rekonstruieren.

Riassunto

Le regioni agrarie dell’Inghilterra medievale: le prove archeobotaniche e zooarcheologiche di Stephen Rippon, Adam Wainwright e Chris Smart

In passato le variazioni regionali del carattere del paesaggio sono state studiate dagli archeologi in termini di manifestazioni fisiche, quali i diversi sistemi di stanziamento e di sistemi agrari. I caratteri distintivi del paesaggio, locali e regionali, derivano inoltre dal modo in cui le comunità rurali hanno praticato regimi agricoli diversi e gli storici hanno riconosciuto da lungo tempo la portata di queste variazioni in tutto il paese. Gli archeologi, invece, hanno messo a confronto resti ossei di animali e resti di cereali provenienti da siti con condizioni socioeconomiche diverse, ma finora non hanno concentrato la loro attenzione su quanto ampie siano queste variazioni a seconda dei diversi caratteri geologici.

Perciò questo studio presenta l’analisi delle ossa animali delle tre specie domestiche principali (bovini, ovini e suini) e dei resti carbonizzati dei quattro cereali principali coltivati (frumento, orzo, avena e segale) attraverso una serie di geologie di superficie diverse comprese in un’area di studio che si estende dall’East Anglia fino alla penisola del sudovest (Cornovaglia e Devon). In primo luogo si dimostra che i sistemi di allevamento degli animali e di coltivazione dei cereali variavano in modo considerevole secondo le diverse geologie di superficie. In secondo luogo si rileva che se da un lato le pratiche agricole sembrano essere dipese dalla geologia di superficie, dall’altro esse erano influenzate anche da fattori culturali, in particolare quando le comunità umane rispondevano alle opportunità derivanti dalla crescente economia di mercato. E, in terzo luogo, che, mentre gli schemi archeobotanici e zooarcheologici evidenti tra la metà dell’XI e la metà del XIV secolo sono conformi a quanto ci viene riportato dalle fonti documentarie, l’importanza particolare di questo set di dati archeologici consiste nel permetterci di ricostruire sistemi agrari risalendo fino al periodo altomedievale non documentato (e addirittura a periodi più antichi).

ACKNOWLEDGEMENTS

The research contained in this paper was partly supported by the Leverhulme Trust as part of the Fields of Britannia Project (award F/00 144/BI), and partly by the University of Exeter. We also wish to thank the numerous archaeological units, HERs, and specialists who have supplied unpublished data. We are grateful to Tom Williamson and two anonymous referees for their constructive comments on an earlier draft of this paper. We also wish to thank Trevor Bailey and Theodoros Economou for their help with the statistical analysis.

Notes

1 Professor of Landscape Archaeology, Archaeology Department, University of Exeter, Laver Building, North Park Road, Exeter, Devon EX4 4ET, UK. [email protected]

2 Archaeology Department, University of Exeter, Laver Building, North Park Road, Exeter, Devon EX4 4ET, UK. [email protected]

3 Archaeology Department, University of Exeter, Laver Building, North Park Road, Exeter, Devon EX4 4ET, UK. [email protected]

4 Eg Lewis et al 1997; Roberts and Wrathmell 2002; Williamson 2003; 2013; Rippon 2008.

5 Some different approaches include Thirsk 1967; 1987; Postan 1972; Fox 1991; Overton 1996; Campbell 2000.

6 Grant 1988.

7 Kerridge 1967; 1973.

8 Thirsk 1967.

9 Campbell 2000.

10 Bartley and Campbell 1997.

11 Phythian-Adams 1986; Tranter et al 1999.

12 Fox 1989, 56.

13 Phythian-Adams 1986, xiii.

14 Fox 1989, 56.

15 Thirsk 1967; and see Overton 1996, 46–62.

16 Everitt 1986, map 2.

17 Thirsk 1987; 2000; and see Phythian-Adams 1999, 119 and Overton 1996, 46–62.

18 Although the early use of this OE place name weald, walda implies ‘woodland’, it later came to refer to areas of wood pasture with isolated stands of trees, and thereafter landscapes that were devoid of woodland: Fox 1989, 81.

19 Roberts and Wrathmell 2000, fig 1.

20 Everitt 1986, 43.

21 Everitt 1986, map 1.

22 See Rippon 2012, 151–204 for a discussion of the development of early medieval territorial structures.

23 Sawyer 1968, no 1784.

24 Ibid, no 1787; Hart 1971, no 7.

25 Rippon 2008, 95–102; 2012, chapter 8.

26 Rippon 2000; 2013a.

27 Sawyer 1968, no 111.

28 Swanton 1996, 57, 62.

29 Rudkin 1955, 389.

30 Eg Lincolnshire: Bennett and Bennett 1993, 9; Wessex: Fowler 2000; Yorkshire: Wrathmell 2012, figs 73–4.

31 Kerridge 1967, 41–180.

32 Rippon 2012, tab 11.2.

33 Williamson 2003; 2006; 2013; Williamson et al 2013.

34 Those covering eastern Devon and south-western Somerset are discussed in Rippon 2012.

35 Risdon c 1630, 5–6.

36 Ibid, 4.

37 Postan 1972, 42.

38 Astill and Grant 1988, 216.

39 Campbell 2000, 411–12.

40 Kerridge 1992, 128.

41 Eg the studies of animal bones by Grant (1988, fig 8.2) and Albarella and Davis (1994, figs 39–41); for a similar approach, based upon comparing sites of different socio-economic status, in the Roman period, see King (1999).

42 Sykes 2007.

43 Van der Veen et al 2013.

44 The Fields of Britannia Project was generously supported by the Leverhulme Trust (grant F/00 144/BI) and is published in Rippon et al forthcoming.

45 Within these broad regions there was considerable variation in the natural landscape, and so a second tier of spatial units was identified that reflects the predominant surface geology (ie drift and/or solid geology, whichever determines the soil characteristics), topography and soils. Although these unique districts each has a dominant type of natural landscape there is always some variety such as the well drained valleys that provide areas of better drained soils within the heavy interfluvial plateaus of the Boulder Clay. Although these pays mostly represent very obvious coherent districts, it must be remembered that they have been constructed for the purposes of modern scholarship: there is no map showing how local and regional variation in landscape character was perceived in the past, and a different group of scholars using different sets of data could draw up a slightly different set of pays boundaries, although it is remarkable that when a wide variety of landscape characteristics, such as place names, field names, dialects, and architectural styles are mapped, they all correspond to the same boundaries in landscape character (eg for the eastern boundary of the East Midlands see Fox 2009 and 2012; for the southern boundary of East Anglia, see Williamson 2006; Rippon 2007 and 2008; for the eastern boundary of the South-West, see Rippon 2008 and 2012).

46 Williamson 2003; Rippon 2008; 2012.

47 Rippon et al forthcoming, tab 3.2.

48 Ibid, tab 6.2.

49 Ibid, tabs 4.2, 5.2, 6.2, 7.2, 8.2 and 9.2.

50 Ie where the boundaries of the strips and furlongs are preserved within the post-enclosure landscape.

51 See Rippon et al forthcoming, including tab 3.7.

52 By the 1960s, a view prevailed that open fields were created in a piecemeal fashion, for example through gradual clearance and improvement of former rough pasture (eg Orwin and Orwin 1938; Beresford 1957, 44, 69; Thirsk 1964, 8–9), although subsequent scholars argued instead for a ‘great replanning’ of the landscape and a ‘village moment’: a large-scale, planned, re-organisation of settlements and field systems that created long furlongs that over time became sub-divided into the more complex pattern that survived into the post-medieval period to be recorded on maps and as earthworks. The results of the Fields of Britannia Project, however, must call into question the extent of this ‘great replanning’ as within the Central Zone 72% of the excavated Romano-British field systems in historic landscapes characteristic of former open fields (ie in areas of enclosure by agreement where the pattern of strips and furlongs survives within the historic landscape of today) have a common orientation or alignment (in East Anglia the figure is 73%, and the North-East Lowlands 67%). This suggests that in many cases open fields were created within a pre-existing framework of boundaries that dates back to the Roman period, rather than through a ‘great replanning’ of the landscape. Setting aside her dating of the origins of open fields to the high Middle Ages, we can therefore return to Thirsk’s (1964) seminal paper in which she argues for them having had an evolutionary origin through the refashioning of pre-existing, irregular, multi-field arrangements within which the arable land may already have been sub-divided into strips. There are some examples of open fields that clearly were created in a planned fashion covering entire parishes or townships in Yorkshire (eg Holderness: Harvey 1978; 1981; 1983), Northamptonshire (Partida et al 2013; Williamson et al 2013), and areas of wetland reclamation in Fenland (D Hall 1996) but these would appear to be the exception rather than the rule.

53 Eg Bishop’s Cleeve in Gloucestershire (King 1993; Barber and Walker 1998; Lovell et al 2007); Buildings Farm in Great Dunmow, Essex (Lavender 1997); Grange Farm, Snetterton, in Norfolk (Robertson 2004); Latton Lands, in Wiltshire (Powell et al 2009, 57); Little Oakley in Essex (Barford 2002, 55); Main Street, in Market Overton, Leicestershire (Shore 2007); Melford Meadows in Brettenham, Norfolk (Mudd 2002, 27); Monk Field Farm and The Grange, Camborne, Cambridgeshire (Wright et al 2009; North Shoebury in Essex (Wymer and Brown 1995); Water End East, Great Barford, Bedfordshire (Timby et al 2007b, 201); Weldon Gap, Rose Lane, Great Chesterford, Essex (Rees 2008).

54 Rippon 2012.

55 Rippon 2013b; Rippon et al forthcoming.

56 Sheep and goat bones can usually not be distinguished.

57 Sykes 2007 provides a general discussion across sites of different social status. On rural sites of the 5th–mid-9th centuries across England, for example, just 0.005% of the bones are from deer, which fell to 0.003% in the mid-9th–mid-11th centuries, and although they increased thereafter to 0.008% in the mid-11th–mid-12th centuries and 0.024% in the mid-12th–mid-14th centuries, deer were clearly not a significant part of the diet of medieval rural communities (Sykes 2007, appendix 1a). For specific sites across different regions and pays: Great Linford and Westbury (Boulder Clay) in Buckinghamshire: Mynard and Zeepvat 1992, tab 11; Ivens et al 1995, tab 54; Caldecote (chalk) in Hertfordshire: Beresford 2009, fig 16.1; Raunds and West Cotton (Boulder Clay) in Northamptonshire: Audouy and Chapman 2009, tab 8.1; Chapman 2010, tab 13.1; Puxton (reclaimed marshland) and Shapwick (mixed geology) in Somerset: Rippon 2006, tabs 10.17–10.18; Gerrard with Aston 2007, fig 22.4; West Stow (heathland): West 1985; Wharram Percy (chalk) in Yorkshire: Rahtz and Watts 2004, tab 63; Stamper and Croft 2000, tab 28.

58 Many excavation reports do not specify which type of wheat grains are present, but where they do bread wheat is always dominant, for example 100% at 6th-century Cowick Lane in Exeter (Caine and Valentine 2011), 97% middle Saxon Lundenwic (Cowie and Blackmore 2012), and 99% in late-Saxon and high medieval West Cotton (Chapman 2010).

59 This has long been recognised (eg Green 1981; Hillman 1981; Jones 1981; Van der Veen 1992), and is borne out by recent work.

60 Based on the revised dating of Ipswich Ware (Blinkhorn 2012).

61 Of which 881 could be identified as either barley, oats, rye, or wheat (the rest being indeterminate).

62 West 1985, 104.

63 On the Somerset Levels, Pevensey Levels, Romney Marsh, the East Kent Marshes, the Norfolk Broads, and Fenland.

64 Rippon 2001, tabs 1 and 2.

65 At Westbury (Boulder Clay) in Buckinghamshire ‘middle Saxon’ contexts produced 739 cereal grains and just 3 legumes, and high medieval contexts produced 1876 cereals and 24 legumes (Ivens et al 1995, tab 56); examples of sites on other geologies include Caldecote (chalk) in Hertfordshire: Beresford 2009, fig 16.12; Raunds and West Cotton (Boulder Clay) in Northamptonshire: Audouy and Chapman 2009, tabs 9.2–9.11; Chapman 2010, tabs 12.2–12.20; Shapwick (mixed geology) in Somerset: Gerrard with Aston 2007, figs 21.16–21.17; West Stow (heathland): West 1985; Wharram Percy (chalk) in Yorkshire: Stamper and Croft 2000, tabs 33–5.

66 Hillman 1981; Jones 1981; Kerridge 1992; Van der Veen 1992, 145; Campbell 2000, 214–18.

67 Fox 1991.

68 Eg Van der Veen 2008; Van der Veen et al 2007; 2013.

69 Replaced by Planning Policy Statement 5: Planning for the Historic Environment (2010) and then the National Planning Policy Framework (2012).

70 Thomas 2013, 12.

71 ADS, http://archaeologydataservice.ac.uk/; Albarella and Pirnie 2008; http://archaeologydataservice.ac.uk/archives/view/animalbone_eh_2007; Oxford Archaeology, http://library.thehumanjourney.net; Wessex Archaeolog­y, <http://www.wessexarch.co.uk/reports/>; TVAS <http://www.tvas.co.uk/reports/>; University of Leicester Research Archive, <https://lra.le.ac.uk/handle/>.

72 Mudd forthcoming.

73 Grant 1988; Albarella and Davis 1994; Sykes 2007.

74 Woodward and Leach 1993.

75 Rogerson and Dallas 1984; Dallas 1993; Wallis 2004; Atkins and Connor 2010.

76 The reclamation of marshland involved constructing an embankment to prevent tidal flooding, and the digging of ditches to drain the land, so changing the ecology from saline saltmarshes to freshwater.

77 Excluding the sites at Bantham (Devon), Tintagel (Cornwall), and Cadbury Congresbury (Somerset) where the presence of pottery imported from the Mediterranean suggests a high social status.

78 There are 21 assemblages (7201 frags) with assemblages phased in the excavations reports as ‘early to middle Saxon’, on which cattle bones amounted to 47%, sheep/goat 36% and pig 17%: as these figures are so similar to the 46 assemblages (44,145 fragments) dated specifically to the 5th to 7th centuries (on which cattle were 48%, sheep/goat 36% and pig 16%), the ‘early Saxon’ and ‘early to middle Saxon’ datasets are combined in order to allow for greater sample sizes when examining individual geologies.

79 Across the Boulder Clay as a whole, cattle declined from 64% in the Roman period (compared to the average on rural settlements across all geologies of 48%) (Rippon et al forthcoming, tab 3.3.) to 56% in the 5th–7th centuries (when the average for rural settlements was 48%), to 53% in 8th–mid-9th centuries cattle (cf the average of 49%), and 42% in the mid-9th to mid-11th centuries (cf the average of 40%). By the high Middle Ages cattle were just 35% on the Boulder Clay (cf the average of 40% for all rural settlements). This steady decline in cattle on the Boulder Clay is matched by a steady rise in the significance of sheep/goat from 33% in the 5th to 7th centuries, to 43% in the 8th to mid-9th, 41% in the mid-9th to mid-11th, and 47% in the high Middle Ages.

80 In the 5th to 7th centuries cattle were 44% in the South-East Midlands, 47% in the East Midlands, but 64% in East Anglia/Essex; in the 8th to mid-9th century the figures are 37% in the South-east Midlands and 82% in Essex, in the mid-9th to mid-11th century there is only data for the East/South-East Midlands (42%).

81 Ingleborough in West Walton, Hay Green in Terrington St Clement, and Rose Hall Farm in Walpole St Andrew: Crowson et al 2005.

82 Van Zeist 1974; van Zeist et al 1976; Bottema et al 1980; Behre and Jacomet 1991; Rhoades et al 1992.

83 West Fen Rd, Consortium site, in Ely (Cambridgeshire): Mudd and Webster 2011; Baltmore Wall, in East Lyng (Somerset): Watts and Scaife 2008; Church Field, in Shapwick (Somerset): Gerrard with Aston 2007; Bestwall Quarry, in Wareham (Dorset): Carruthers 2006.

84 50% of the 947 grains: Carruthers 2006.

85 72% of the 215 grains: Cuttler et al 2011.

86 80% of the 126 grains: Kenyon and Watts 2006.

87 Eg Campbell’s (2000, 418) correlation between his Type 3 land use (arable with limited but valuable grassland) and heavy soils, and Williamson’s (2003) argument that open fields developed on certain types of soils that were prone to compaction and puddling if ploughed when too wet.

88 Rippon 2006; 2008.

89 Grant 1988; Campbell 2000, 156–65.

90 Cuttler et al 2011.

91 Thirsk 1967; 1987; Campbell 2000.