‘Sons of athelings given to the earth’:¹ Infant Mortality within Anglo-Saxon Mortuary Geography

Abstract

FOR 20 OR MORE YEARS early Anglo-Saxon archaeologists have believed children are under-represented in the cemetery evidence. They conclude that excavation misses small bones, that previous attitudes to reporting overlook the very young, or that infants and children were buried elsewhere. This is all well and good, but we must be careful of oversimplifying compound social and cultural responses to childhood and infant mortality. Previous approaches have offered methodological quandaries in the face of this under-representation. However, proportionally more infants were placed in large cemeteries and sometimes in specific zones. This trend is statistically significant and is therefore unlikely to result entirely from preservation or excavation problems. Early medieval cemeteries were part of regional mortuary geographies and provided places to stage events that promoted social cohesion across kinship systems extending over tribal territories. This paper argues that patterns in early Anglo-Saxon infant burial were the result of female mobility. Many women probably travelled locally to marry in a union which reinforced existing social networks. For an expectant mother, however, the safest place to give birth was with experienced women in her maternal home. Infant identities were affected by personal and legal association with their mother’s parental kindred, so when an infant died in childbirth or months and years later, it was their mother’s identity which dictated burial location. As a result, cemeteries central to tribal identities became places to bury the sons and daughters of a regional tribal aristocracy.

Résumé

Pendant une vingtaine d’années, les archéologues s’intéressant au début de la période anglo-saxonne ont cru que les enfants étaient sous-représentés dans les fouilles de cimetières. Ils en concluaient que des ossements si petits n’étaient pas trouvés lors des fouilles, que les pratiques antérieures n’avaient pas pour habitude de mentionner les jeunes enfants, ou que l’on enterrait ailleurs les bébés et les enfants. Tout ceci semblerait logique, mais il faut se méfier de ne pas simplifier outre mesure les réponses sociales et culturelles complexes face à la mortalité infantile et enfantine. Vu cette sous-représentation, les approches antérieures n’ont pu offrir que des dilemmes méthodo­logiques. Or, une proportion plus grande de nourrissons était ensevelie dans les grands cimetières, parfois dans des zones spécifiques. Cette tendance est statistiquement significative et ne résulte probablement pas seulement de problèmes de préservation ou d’excavation. Les premiers cimetières médiévaux, faisant partie de géographies mortuaires régionales, étaient des lieux pour mettre en scène des événements propices à la cohésion sociale à travers divers systèmes de parenté et les territoires tribaux. Cet article fait valoir que les tendances en matière de sépulture infantile au début de la période anglo-saxonne résultaient de la mobilité des femmes. Nul doute que les femmes se déplaçaient localement pour se marier, une union qui consolidait les réseaux sociaux existants. Or, le plus sûr pour une femme enceinte était d’accoucher en compagnie de femmes expérimentées, dans son foyer maternel. Les identités des nourrissons étaient marquées par les associations personnelles et légales avec les parents côté maternel, de sorte que, quand un enfant mourrait à la naissance ou quelques mois/années plus tard, c’était l’identité maternelle qui dictait le lieu de sépulture. En conséquence, les cimetières les plus importants pour l’identité tribale sont devenus les lieux de sépulture des fils et filles d’une aristocratie tribale régionale.

Zusammenfassung

Seit zwanzig oder mehr Jahren glauben die Archäologen der frühen angelsächsischen Periode, dass Kinder in den Funden auf Friedhöfen unterrepräsentiert sind. Sie kommen zu dem Schluss, dass bei Ausgrabungen kleine Knochen übersehen werden, dass frühere Einstellungen zum Berichtswesen die sehr jungen Menschen übersehen haben oder dass Kinder und Säuglinge anderswo bestattet wurden. Das ist alles schön und gut, doch wir müssen sorgfältig darauf achten, die komplexen sozialen und kulturellen Reaktionen auf die Kinder- und Säuglingssterblichkeit zu sehr zu vereinfachen. Bisherige Ansätze haben zu methodologischen Dilemmas angesichts dieser Unterrepräsentation geführt. Es wurden jedoch in bestimmten Zonen proportional mehr Säuglinge auf großen Friedhöfen beigesetzt. Dieser Trend ist statistisch signifikant und rührt also wahrscheinlich nicht von Erhaltungs- oder Ausgrabungsproblemen her. Frühmittelalterliche Friedhöfe waren Teil der regionalen Totengeografie und boten Orte, an denen man Veranstaltungen durchführen konnte, die den sozialen Zusammenhalt in Verwandtschaftssystemen über die Stammesgrenzen hinweg förderten. Dieser Artikel argumentiert, dass die Muster der Säuglingsbestattungen in der frühen angelsächsischen Zeit das Ergebnis der Mobilität der weiblichen Bevölkerung waren. Viele Frauen reisten vielleicht in der näheren Umgebung, um eine Heirat einzugehen, die bestehende soziale Netzwerke verstärkte. Für eine werdende Mutter war jedoch der sicherste Ort für die Geburt eine Umgebung von erfahrenen Frauen in ihrem mütterlichen Zuhause. Die Identität der Säuglinge wurde durch die persönliche und juristische Verbindung mit der elterlichen Verwandtschaft der Mutter geprägt, und wenn ein Säugling bei der Geburt oder Monate und Jahre später starb, bestimmte folglich die Identität der Mutter den Ort des Begräbnisses. Deswegen waren Friedhöfe, die für die Identität eines Stammes von zentraler Bedeutung waren, Orte, an denen die Söhne und Töchter der regionalen Stammesaristokratie beerdigt wurden.

Riassunto

Da vent’anni o più gli archeologi del primo periodo anglosassone ritengono che nei reperti cimiteriali i bambini non sono sufficientemente rappresentati. Concludono che durante gli scavi vanno perdute piccole ossa, che con l’atteggiamento passato in fatto di resoconti i bambini più piccoli venivano trascurati, oppure che neonati e bambini erano sepolti altrove.

Tutto questo sta bene, ma dobbiamo stare attenti a non semplificare eccessivamente le reazioni complesse, sociali e culturali, riguardo alla mortalità infantile. Di fronte all’insufficiente documentazione gli approcci precedenti hanno posto dilemmi metodologici. Tuttavia nei cimiteri più grandi erano stati sepolti proporzionalmente più neonati e a volte in zone particolari. Questa tendenza è statisticamente rilevante ed è quindi improbabile che derivi interamente da problemi di conservazione o di scavi. I cimiteri altomedievali facevano parte della geografia mortuaria regionale e rappresentavano luoghi in cui organizzare eventi che producevano coesione sociale tra tutti i sistemi di parentela, estendendosi sui territori tribali. In questo saggio si sostiene che la distribuzione delle sepolture di neonati durante il primo periodo anglosassone sia una conseguenza della mobilità femminile. È probabile che molte donne si spostassero localmente per sposarsi e formare un’unione che rafforzava i vincoli sociali esistenti. Tuttavia per una gestante il luogo più sicuro per partorire era nella propria casa materna con donne esperte. L’identità infantile era influenzata dai legami personali e legali con i parenti della madre, cosicché quando un bambino moriva durante il parto, o mesi e anni dopo, era l’identità della madre a dettare il luogo della sepoltura. Per questo i cimiteri che si trovavano al centro di identità tribali divennero i luoghi in cui seppellire la progenie delle aristocrazie tribali regionali.

Children and infants are under-represented in archaeological discovery, and it would be unwise to believe otherwise. However, if we do not look beyond this situation we are in danger of oversimplifying a complex social, personal and cultural response to childhood mortality. For the last 20 years children have been considered important subjects for investigation and childhood has been recognised as a socially constructed and historically contingent step in an individual’s life course. However, infants and children do not exist in isolation; they are part of community networks and kinship groups which extend beyond the boundaries of one cemetery and one community. Kinship networks are route­d in landscapes because people travel to maintain them. Marriage, birth and funerals, then, are important rites of passage and their celebration helped to create and reinforce key social relationships within and beyond the immediate community. As a result, it is important to consider graves within their regional context.

Later Anglo-Saxon church and minster cemeteries contained proportionally more children than contemporary field cemeteries.³ Infants made up between 21% and 37% of churchyard burials, but between 5% and 12% of field cemeteries.⁴ Consequently, we might consider that the Church fulfilled a special function in the burial of infants and probably associated sacred spaces with baptism.⁵ Unfortunately, such obvious differences do not exist in early Anglo-Saxon cemeteries. Pre-Christian communities did not employ baptism, so social factors may have influenced their character. For example, a recent study of maternal mortality explored the double burial of a woman and an infant who died in childbirth.⁶ Notably, all similar graves have been found in large cemeteries with over 100 burials, and so female residence, mobility and affinity may have been important to some communities.

Hakenbeck⁷ investigated female migration using the Hunnic modified skull (cranium) and its appearance across Europe between the 3rd and 6th centuries ad. This distinctive body modification provided a neat insight into individual migration patterns because the physical transformation must have been applied to infants, when their skull was malleable. Examples of women with modified heads have been found as far west as France, hinting at the extent of an exogamous marriage network practiced by a pan-European elite. Hakenbeck also investigated skeletal isotopes in Bavarian cemeteries, arguing that differences seen between men and women meant that ‘a proportion of women did not always live locally, but moved there from other regions where they had access to a different diet’.⁸ These interpretations describe a society in which female mobility acted to reinforce political cohesion and was used to create or maintain elite associations.⁹

Comparative studies like these have been a particular feature of early medieval mortuary archaeology, and cemetery reports list similarities in brooch types and mortuary populations, but it is open area commercial archaeology which has demonstrated that cemeteries were not singular bounded sites but part of complex landscapes. Saltwood in Kent is a good example¹⁰ because it was not one cemetery but two or more separate, but broadly contemporary, groups of graves (Fig 1). Each group of graves was adjacent to a different Iron-Age/Roman track, which remained in use during the Anglo-Saxon period.¹¹ Saltwood was not a single cemetery but part of a mortuary landscape where different communities accessed their dead in different ways. These separate groups placed their dead in close proximity for a reason. They may have been part of a cohesive network of hamlets or villages bound by kinship ties. Individuals will have permeated these spatial boundaries and personal mobility was probably not restricted to exogamy because intermarriage would also have operated across corporate groups extending over sub-regional and regional networks.

FIG 1. Saltwood, Kent, Note the large clusters of graves which focus on Bronze-Age ring ditches. The trackway has prehistoric origins and remained in use until the 19th century, so the early Anglo-Saxons would have accessed these graves from different places. Redrawn after Brooks and Reynolds 2011. © D Sayer.

Recent excavation has made a significant contribution to early medieval archaeology and so too has the increasing number of published sites. In the early 1990s Crawford¹² published an influential study of Anglo-Saxon children and based her assessments of infant and child mortality on an analysis of 1141 burials; 11% were under five. Crawford used the available published material which consisted principally of small cemeteries, but there were exceptions; Lechlade, Gloucestershire, for example, was a large cemetery with 20% infants.¹³ Since then, and since Buckberry’s review of the under-representation of children,¹⁴ several large sites have been published, for example Blacknall Fields, Wiltshire, or Great Chesterford, Essex, which like Lechlade contained infant populations over 20%.¹⁵

Some early Anglo-Saxon cemeteries appear to have had large infant populations, others contained just a few. This study will investigate this difference and show that large cemeteries contained proportionally higher numbers of infants, and that these sites were important places within the sub-regional tribal landscape. Using a corpus of 41 early Anglo-Saxon cemeteries and statistical patterns corresponding to the size of the sites demonstrates how the data can be divided into two groups; large and small cemeteries. However, quantitative investigations can create arbitrary divisions so this study will also explore the spatial arrangement of cemeteries and the mortuary landscapes within which they were situated.

RETHINKING INFANT BURIAL

It has been accepted that children are under-represented in Anglo-Saxon archaeological excavation because the proportion of children found in cemeteries is low compared with pre-industrial societies.¹⁶ The reasoning is that children’s bones do not survive, excavation methods may not recover the very young, or there is a different location for the disposal of infants and children.¹⁷ However, child, infant and foetal remains survive in most burial conditions depending on age at death, which determines calcium levels within the bone structure so that children’s bones will survive where there is good preservation of adults.¹⁸ Excavation method cannot be universally responsible; as early as 1855 Akerman was able to identify 11 infants at Harnham Hill, Salisbury.¹⁹ By contrast, many commercial excavations are worked on quickly; small foetus bones could be missed and children buried in shallow graves may be lost to a mechanical digger. Despite this, recovery is considered better today than it was 150 years ago because of changes in attitude, collection strategies, research priorities and improvements in reporting.²⁰

The most notable protagonist for early medieval childhood studies has been Sally Crawford.²¹ Crawford identified a number of methodological problems which frustrated previous comparisons; the most notable of these rests with reporting. Crawford showed that the age categories designated for skeletons in published reports have not always been presented comparably, leading to mistakes in the way the data are considered.²² For instance, when comparing Polhill and Dover Buckland, both in Kent, Evison compared the figures presented with an upper threshold of 15 for juveniles at Dover, and 18 at Polhill because this is how the data were organised in each report.²³ As a result, Evison concluded that the proportion of children in these two sites was similar, but they are in fact quite different. Consequently, ages at death for infants and children must be tabulated according to a uniform standard that can allow reliable comparison. For this study, 3412 skeletons from 41 well-reported sites found across England have been tabulated (Fig 2). The numbers of infants and children were considered on the basis of equivalent cultural age categories.

FIG 2. Cemeteries included in this study: 1 Apple Down, West Sussex, 2 Abingdon, Oxfordshire, 3 Barrington, Cambridgeshire, 4 Beckford B, Hereford and Worcester, 5 Berinsfield, Oxfordshire, 6 Blacknall Field, Wiltshire, 7 Broughton Lodge, Nottinghamshire, 8 Castledyke South, Lincolnshire, 9 Dover Buckland, Kent, 10 Empingham II, Rutland, 11 Finglesham, Kent, 12 Great Chesterford, Essex, 13 Kingsworthy, Hampshire, 14 Lechlade, Oxfordshire, 15 Norton, Cleveland, 16 West Heslerton, North Yorkshire, 17 Alton, Hampshire, 18 Alwalton, Cambridgeshire, 19 Beckford A, Hereford and Worcester, 20 Bergh Apton, Norfolk, 21 Coddenham, Suffolk, 22 Deal, Kent, 23 Didcot, Oxfordshire, 24 Dinton, Buckinghamshire, 25 Droxford, Hampshire, 26 Empingham I, Rutland, 27 Gunthorpe, Peterborough, 28 Holborough Hill, Kent, 29 Lyminge, Kent, 30 Marina Drive, Dunstable Bedfordshire, 31 Market Lavington, Wiltshire, 32 Meonstoke, Hampshire, 33 Melbourn, Cambridgeshire, 34 Morning-Thorpe, Norfolk, 35 Portway Andover, Hampshire, 36 Ports Down, Hampshire, 37 Orpington, Kent, 38 Sewerby, East Yorkshire, 39 Wakerley, Northamptonshire, 40 Westgarth Gardens, Suffolk, 41 Winterbourne Gunner, Wiltshire. Drawn by D Sayer ©.

Gender archaeologies accept that sex is corporeal but gender is socially constructed.²⁴ The same is true for childhood; physical ages can be understood from skeletal changes²⁵ but social age is independent and punctuated by rites of passage.²⁶ For this study it is important to understand culturally constructed categories because it is these which might determine how an individual was considered by living contemporaries who dictated burial practice.²⁷ Skeletal age assessment is most accurate for younger individuals because of incremental skeletal changes²⁸ and can be combined with other multidisciplinary approaches to investigate social questions. Crawford²⁹ studied grave goods alongside historical and legal sources to described cultural age thresholds at 0–5, 5–11 and 11–15. These are similar to those described by Stoodley³⁰ and Härke³¹ who, based on the provision of grave goods also considered that the transition from infancy to childhood was at five. To facilitate comparison in this study, physical age was derived from excavation reports to allow a study of different cultural age categories; infants were considered 0–5 and children 6–12 years of age.

Tables 1 and 2 list 41 cemeteries selected because skeletal preservation was good enough to determine age at death. Table 1 lists sixteen cemeteries with over 100 inhumations. Many of these sites included high proportions of infants, notably Great Chesterford, with 40% infants, and Blacknall Fields with a mortuary population including 25% infants. In each case there were fewer children aged 6–12 with 7.2% and 8.9% respectively, but the total numbers of children aged 12 and below remained proportionally high. Other sites such as Castledyke South, Lincolnshire, and Kingsworthy, Hampshire, showed a range between 2.6% and 17.4% infant populations with 6% to 17.1% younger children, making an average of 23.5% aged 12 and below. There are some exceptions, notably Dover Buckland and Finglesham, Kent, as well as Broughton Lodge, Nottinghamshire, where the numbers of children varied between 11.4% and 16.4%. Notably, there is a difference in the total number of infants, 13.5%, versus the number of children, 11.2%, in these sites (Fig 3). However, we must be cautious of loading interpretation on to this variation. Great Chesterford may be an outlier in this dataset, with 67 infant graves, and if Great Chesterford is removed from the sample it consists of 15 cemeteries with 241 infants, thus 11.4%, and 243 children, thus overall 11.4% of 2119 graves (where infants and children are both 11% of the whole sample of 3412).

FIG 3. Pie chart showing the proportions of children, infants and adults found in large cemeteries with over 100 graves. Drawn by D Sayer ©.

Table 1. Sixteen early Anglo-Saxon inhumation cemeteries with over 100 graves.

Cemetery	Infants aged 0–5	Children aged 6–12	Total under 12	Total skeletons	Extent of cemetery excavated	Skeletal preservation	Reference
Apple Down, West Sussex	16 (13.2%)	12 (9.9%)	28 (23%)	121		good	Down and Welch 1990: 188
Abingdon, Oxfordshire	20 (16.8%)	19 (16%)	39 (33%)	119	50%–75%	unknown	Leeds and Harden 1936
Barrington, Cambridgeshire	15 (10.1%)	14 (9.5%)	29 (19.5%)	148		good	Malim and Hines 1998
Beckford B, Hereford and Worcester	9 (8.3%)	17 (15.7%)	26 (24%)	108			Evison and Hill 1996: 52
Berinsfield, Oxfordshire	12 (10.5%)	17 (14.9%)	29 (25.4%)	114		medium	Boyle et al 1995:113
Blacknall Field, Wiltshire	26 (25.7%)	9 (8.9%)	35 (34.7%)	101		poor	Annable and Eagles 2010
Broughton Lodge, Nottinghamshire	5 (4.8%)	4 (3.8%)	12 (11.4%)	105			Kinsley 1993
Castledyke South, Lincolnshire	14 (6.7%)	12 (5.3%)	26 (14%)	227	55%	poor	Drinkhall and Foreman 1998
Dover Buckland, Kent	20 (5%)	25 (6.3%)	45 (11.5%)	394		poor	Parfitt and Anderson 2012; Evison 1987
Empingham II, Rutland	11 (7.3%)	13 (8.7%)	24 (16%)	150		good	Timby 1996
Finglesham, Kent	17 (8.5%)	16 (8 %)	33 (16.4%)	201		poor	Hawkes and Grainger 2006: 314
Great Chesterford, Essex	67 (40.1%)	12 (7.2%)	79 (46.7%)	167	75%?	poor	Evison 1994
Kingsworthy, Hampshire	14 (14%)	11 (11%)	25 (25%)	100			Hawkes and Grainger 2003, 154
Lechlade, Kent	38 (17.4%)	21 (9.6%)	59 (26.9%)	219			Boyle et al 1998: 48
Norton, Cleveland	3 (2.6%)	20 (17.1%)	23 (19.7%)	117		poor	Sherlock and Welch 1992
West Heslerton, North Yorkshire	21 (11%)	33 (17.1%)	22 (28.1%)	195		poor	Haughton and Powlesland 1999, 177
Total	308 (13.5%)	255 (11.2%)	534 (23.5%)	2286	-	-	-
Average	19.3	14	33.4	142.8	-	-	-

Table 2. Twenty-five early Anglo-Saxon cemeteries with less than 100 graves.

Cemetery	Infants aged 0–5	Children aged 6–12	Total under 12	Total skeletons	Reference
Alton, Hampshire	7 (14.3%)	4 (8.2%)	11 (22.4%)	49	Evison 1988
Alwalton, Cambridgeshire	1 (2.7%)	1 (2.7%)	2 (5.6%)	36	Gibson 2007, 280
Beckford A, Hereford and Worcester	1 (4.2%)	4 (16.7)	5 (20.8%)	24	Evison and Hill 1996, 47
Bergh Apton, Norfolk	-	12 (19%)	12 (19%)	63	Green and Rogerson 1978
Coddenham, Suffolk	2 (5.9%)	3 (8.8%)	5 (14.7%)	34	Penn 2011
Deal, Kent	5 (6.4%)	7 (9%)	12 (15.4%)	78	Parfitt and Brugman 1997, 215
Didcot, Oxfordshire	1 (5.9)	2 (11.8%)	3 (17.6%)	17	Boyle and Mudd 1995
Dinton, Buckinghamshire	1 (5%)	3 (15%)	4 (20%)	20	Hunn et al 1994
Droxford, Hampshire	3(7.3%)	1 (2.4%)	4 (9.8%)	41	Aldsworth & Welch 1979
Empingham I, Rutland	-	1 (7%)	1 (7%)	14	Liddle et al 2000
Gunthorpe, Peterborough	1 (3.1%)	3 (9.4%)	4 (12.5%)	32	Patrick et al 2007, 209
Holborough Hill, Kent	3 (7.7%)	1 (2.6%)	4 (10.3%)	39	Evison 1956
Lyminge, Kent	3 (6.8%)	7 (15.9%)	10 (22.7%)	44	Warhurst 1955
Marina Drive Dunstable, Bedfordshire	2 (4.1%)	10 (20.4%)	12 (24.5%)	49	Matthews 1962
Market Lavington, Wiltshire	3 (7%)	4 (9.8%)	7 (17%)	41	Williams and Newman 2006
Meonstoke, Hampshire	4 (19%)	0	4 (19%)	21	Stoodley and Stedman 2001
Melbourn, Cambridgeshire	4 (6.8%)	1 (1.7%)	5 (8.5%)	59	Duncan et al 2003
Morning Thorpe, Norfolk	2 (2.1%)	7 (7.4%)	9 (8.5%)	94	Green et al 1987
Portway Andover, Hampshire	7 (10.1%)	12 (17.4%)	19 (27.5%)	69	Cook and Dacre 1985, 61
Ports Down, Hampshire	1 (4.5%)	1 (4.5%)	2 (9.1%)	22	Corney et al 1967
Orpington, Kent	1 (1.6%)	3 (4.8%)	4 (6.3%)	63	Richardson 2005
Sewerby, East Yorkshire	1 (1.8%)	3 (5.3%)	4 (7%)	57	Hirst 1985
Wakerley, Northamptonshire	1 (1.2%)	10 (11.8%)	11 (13.9%)	85	Adams and Jackson 1988–9
Westgarth Gardens, Suffolk	6 (9.2%)	3 (4.6%)	9 (13.8%)	65	West 1988, 5
Winterbourne Gunner, Wiltshire	1	1	2	10	Musty and Stratton 1964
Total	61 (5.4%)	104 (9.2%)	165 (14.7%)	1126	-
Average	2.4	4.16	6.6	45	-

Table 2 lists 25 sites with less than 100 burials. Of these 25 sites just three are comparable with those listed in Table 1. At Portway, Andover, Hampshire, 10.1% of 69, just 7 graves, were those of infants, whereas 17.4% were young children, thus 27.5% of this cemetery were under 12. At Alton, Hampshire, 14.3% of the graves were infants (7 of the 49 burials) and 4, 8.2%, were aged 6–12 years. Meonstoke, Hampshire, is an outlier with 19% (4) of its graves containing infants. However, excavation has revealed just 10–20% of the site; similarly at Portway and Alton only fractions of much larger cemeteries have been identified. The range of infants and children found within the remaining cemeteries varies between 0% and 9.2% for infants with an average of 5.4%. The range varies between 0% and 20.4% for children and 5.6% to 27.4% for all children aged 12 and under. What is interesting is that the numbers of infants are consistently lower than children, with an average of 5.4% infant burials compared with 9.2% children (Fig 4). Many sites, like Wakerley, Northamptonshire, Sewerby, East Yorkshire, Gunthorpe, Peterborough, Dinton, Buckinghamshire, Didcot, Oxfordshire, Beckford A, Hereford and Worchester, and Alwalton, Cambridgeshire, contained just one infant grave. Only two sites, Alwalton and Empingham I, Rutland, contained a single child. When the data from both categories are viewed as a histogram arranged by the proportion of infants within the population, there is trend for large cemeteries to show comparatively higher percentages of infants (Fig 5). This difference between infants and children is not random, the result of preservation or excavation problems, nor would we see such a difference between sites if it was the product of infant mortality alone. Comparing the statistics for infants (Tab 3) and children (Tab 4), the difference between the number of infants in large and small cemeteries was deliberate,³² even with numerical outliers like Great Chesterford removed from the sample (Tab 5).

Table 3. Chi-Square investigation of infants in early Anglo-Saxon cemeteries

Burial type	Observed	Expected 11%
Infants Large Cemeteries	308	251.46
Infants Small Cemeteries	61	123.86
P-value (significance)	0.00000000002

Table 4. Chi-Square investigation of children in early Anglo-Saxon cemeteries

Burial type	Observed	Expected 11%
Children Large Cemeteries	255	251.46
Children Small Cemeteries	104	123.86
P-value (significance)	0.07

Table 5. Chi-Square infants in early Anglo-Saxon cemeteries with numerical outliers removed¹

Burial type	Observed	Expected 11%
Infants Large Cemeteries	241	233.09
Infants Small Cemeteries	52	109.78
P-value (significance)	0.00000003

¹ This calculation has been carried out with Great Chesterford and all cemeteries with fewer than 30 graves removed from the data. Even with these removed, infants and children are still each 11% of the total cemetery population.

FIG 4. Pie chart showing the proportions of children, infants and adults found in small cemeteries with less than 100 graves. Drawn by D Sayer ©.

FIG 5. Proportions of infants and children in 41 cemeteries displayed as a histogram and arranged according to the number of infants. Cemeteries with over 100 graves are highlighted in red. Notice the increased number of these cemeteries where there are greater numbers of infants. Drawn by D Sayer ©.

The difference between the two sets of cemetery data is striking. In large cemeteries with over 100 burials, 23.5% were those aged 12 and below, but for the smaller cemeteries the same group is only 14.7%. Underpinning this statistic is the significantly greater proportions of infants in large cemeteries; this difference was not the product of chance but deliberated burial practice.

INFANTS AND CHILDREN WITHIN CEMETERY SPACE

Few studies have looked at the distribution of infants and children within burial grounds. Most simply state that children are under-represented in early Anglo-Saxon cemeteries.³³ Penn and Brugmann warn against making interpretations based on the distribution of graves alone.³⁴ Stoodley disagreed; he investigated the ages of people with furnished³⁵ graves at Deal, Kent, Pewsey, Wiltshire, and Norton, Cleveland, but observed just one pattern at Norton: ‘in each of the main plots the burials of children, but also youths, are on the outer edges, with the adult burials making up the core’.³⁶ Other scholars have suggested children’s body positions deliberately mimicked adults in close proximity.³⁷ However, Crawford and Lee recognised a pairing of infants and adult women with disabilities.³⁸ Lee described Beckford, Hereford and Worcester, Barrington, Cambridgeshire and Apple Down, West Sussex, where women with skeletal trauma, leprosy and swollen limb bones, were buried with infants or near children. Although not specifically looking for patterns, she argued that this presented a ‘liminal’ location with children and impaired individuals unable to carry out the duties of adults; both, she proposed, were the recipients of a diminished legal status.³⁹ Crawford argued that the physically impaired, richly furnished females from Barrington and Castledyke were accompanied by infants to add value to their grave assemblages and that those infants had no agency.⁴⁰ Crawford explored funerary motives because in her examples infants were buried by adult agents not influenced by infant agency. This is an interesting point which needs to be expanded, because every Anglo-Saxon cemetery was the product of a plurality of motives with different local and regional agents influencing it; every grave was the cumulative result of the actions of different funeral participants meaning every cemetery looks different. Not all cemeteries contained patterns in the placement of particular graves, and even among those that did, the arrangements of graves varied considerably. Four examples of cemeteries with evident patterns include Apple Down, Berinsfield, Westgarth Gardens and Great Chesterford.

The early Anglo-Saxon cemetery at Apple Down, West Sussex, was discovered in 1981 and excavated between 1982 and 1987. The site contained 121 inhumation burials and 64 cremations with the majority dating to the 6th century ad. The excavators⁴¹ suggest that the cemetery was arranged around a core which was part an initial phase in the late 5th and early 6th century; however, these burials were in use throughout the 6th century. Interestingly, there is a pattern in the location of older adults and infants (Fig 6). Infant and child graves are distributed throughout the cemetery with a particular grouping of seven infant graves on the west of the site, adjacent to a tight group of seven W/E oriented older adults at the very centre.

FIG 6. Top: Apple Down, West Sussex. This illustration shows the majority of inhumation graves, cremations are not included because age determination is impossible. Note the concentration of older adults and infants marked by the arrow. Bottom: Berinsfield, Oxfordshire. Berinsfield has two groups of graves oriented S/N and W/E and it also has two distinct concentration of infants and children marked with the arrows. Drawn by D Sayer ©.

At Berinsfield, too, there is a concentration of infant and children’s graves. Berinsfield is located in the Upper Thames Valley, Oxfordshire. The cemetery was excavated between 1974 and 1975 and comprised of 100 graves containing 114 burials and four cremations dating between the mid-5th and early 7th centuries. The excavators believe that 75% of the cemetery was recorded and the rest was lost to quarrying.⁴² The site was organised into at least two parts, a northern half where graves are predominantly S/N oriented and a southern half where the graves are predominantly W/E.⁴³ As with Apple Down, infants and children were distributed around the cemetery; however, there is a concentration of seven children and two infants to the east within the northern half of the site (Fig 6). These graves were part of a tight cluster of 16 inhumations (56% under 12) located within an earlier ditch system used to define their extent. A second cluster of five infants and two children was located on the eastern side of the southern half of the site, but the boundaries of this group are harder to define. It consisted of about 18 burials of which 28% were infants.⁴⁴ This high proportion was created by deliberately concentrating infant graves in a particular zone.

A similar pattern was visible at Westgarth Gardens, Suffolk, which was excavated in 1972. The cemetery dates to between the mid-5th century and the early 7th century ad.⁴⁵ Westgarth Gardens consisted of 65 inhumations in 61 graves. However, this site was not completely excavated and restricted intervention discovered a further four inhumations. The published site probably represents some 30%–50% of the cemetery. Westgarth Gardens was a larger cemetery and has been studied repeatedly because its concentration of women to the north and men to the south led Pader⁴⁶ to identify separate plots of graves allocated by gender. Recently Penn and Brugmann⁴⁷ showed that this was improbable; men, women and children were tightly clustered, so it is more likely that these graves were part of a comprehensive and internally sub-divided plot (like at Apple Down). However, Pader was not wrong to see a pattern in the graves. As at Apple Down and Berinsfield, infants and children were found accompanying adults across the site, but there was also a concentration of five infants and three children placed in between the male and female groups.⁴⁸ As with Apple Down, the infants were intentionally positioned adjacent to the oldest individuals in the cemetery (Fig 7).

FIG 7. Top: Great Chesterford, Essex. About 75% of this cemetery has been excavated; note the two zones of children and infants each with concentrations of infant burials. Bottom: Westgarth Gardens, Suffolk. This site has single focus of infants and children in the centre adjacent to the older adults. Drawn by D Sayer ©.

Great Chesterford, Essex, was excavated between 1953 and 1955 and revealed 161 inhumations and 33 cremations,⁴⁹ but only 75% survived.⁵⁰ The site shows a recurring pattern based on the orientation of graves with tight clusters of richly furnished S/N graves surrounded by less richly furnished burials on a W/E orientation. Furthermore infant and child graves were found in particular zones (Fig 3). To the north a cluster of 35 children occupied an area of approximately 35 sq m. These graves were interspersed with adults, but within this area there were two concentrations of infants without adults. The northern section consists of about seven infants, and a single child, with a later adult burial cutting several of the graves. To the south, but within the 35 sq m zone, there was a concentration of eight infant graves unaccompanied by adults.

Great Chesterford had a second cluster of 20 infants and children to the south interspersed with adults over a 25 sq m area. There were also clusters of infant graves within this zone; the largest to the north contained seven graves. What defines these two zones of infant burial was the comparative absence of infants and children in the cemetery’s central area. This centre consisted of a group of adult graves on a W/E orientation and the two largest, also the two most richly furnished graves, were positioned together and on a contrasting orientation. This strategy signalled their position and defined an important place for generations of successive cemetery goers.

The early Anglo-Saxons did not just use grave goods to differentiate their dead; they also used burial location. But the nature of each cemetery varied because its multigenerational architects were local to particular sites and because different agents constructed each grave from a plurality of experiences and motivations. At Apple Down and Westgarth Gardens, infants were intentionally placed adjacent to older adults. At Berinsfield, patterns were less obvious, but some infants and children were placed in one of two zones in the cemetery and at Great Chesterford infants were placed away from a central area, sometimes in clusters of graves. This disparity is important because the archaeological distinction between large or small sites cannot be absolute; some large cemeteries and some small sites have no patterns,⁵¹ and other small sites like Westgarth Gardens do. It is this variation between sites which helps to support the statistical evidence. If the grouping of infants and children had been a ubiquitous phenomenon, then we might argue that smaller excavations simply missed the areas in which infants were buried. However, not all cemetery architects employed groups of graves and not all groups of graves were positioned on the edges of a cemetery and so would have been easily missed. In fact, the funerary parties used a variety of locations to bury infants and children; these were locally meaningful positions, depending on whom was being buried and who attended the funeral.

MORTUARY GEOGRAPHIES: CEMETERIES IN THEIR SUB-REGIONAL CONTEXT

To investigate cemeteries above and below 100 graves is to study an arbitrary distinction; the earliest architects of early Anglo-Saxon cemeteries would not know the final number of graves in a site. Nor would they know who would be buried in a cemetery over its 150- or 200-year lifespan. Equally, many cemeteries have not been fully excavated. As a result, the difference between a large site with over 100 inhumations and a smaller one with less than 100 inhumations is an artificial and archaeological distinction which allows the comparative analysis of cemeteries. Nonetheless, it is a useful division and the statistical contrast between cemeteries points to differences in how sites were used and consequently how they were understood. In 2012 Dickinson noted that the largest cemeteries in the Avon Valley, Warwickshire, occupied a cultural borderland.⁵² She presented a distribution map, based on one produced by Scheschkewitz, to show the distribution of cemeteries around the persistent late Roman and early Anglo-Saxon cemetery at Wasperton.⁵³ On Dickinson’s plan the larger sites, Beckford B, Bridford, Alverton, Wasperton and Baginton were located along the Avon, close to a Roman road (Fig 8). As a group they appear to trace the boundary between the south-east with early Anglo-Saxon inhumation cemeteries and the north-west without them. This she argued was a cultural boundary and the larger cemeteries marked a conspicuous demarcation of space.⁵⁴ Large cemeteries were part of an expression of sub-regional identity that transcended the local community; they were part of a mutable regional identity network that probably allowed the utilisation of mortuary spaces for a variety of purposes.

FIG 8. The Avon Valley, large cemeteries are spaced along the Avon hemmed by smaller sites along this boundary between two cultural zones. 1 Beckford B, 2 Bridford, 3 Alverton, 4 Wasperton and 5 Baginton. After Scheschkewitz 2006, 11. © D Sayer.

Other regions have similar patterns in the placement of cemeteries. Kent is a good example because a considerable amount is known about the history of the early landscape.⁵⁵ In West Kent the identified cemeteries are focused in the north, with a greater concentration along the major Roman roads and in the lowland areas (Fig 9). Brookes⁵⁶ noted that the density of burials at 2 sq km intervals corresponds to the location of the 10th-century lathes — territorial sub-divisions above the hundred and below the shire peculiar to Kent in Domesday Book. Moreover, a considerable body of scholarship identifies these territories as the 6th-century boundaries for territorial identity groups, small tribal or folk groups.⁵⁷

FIG 9. Kent with lathes, major Roman roads and rivers marked. Not the position of large cemeteries in west Kent with one or two to a lath each hemmed by small sites focus around a Roman road or river. In east Kent the large cemeteries are also hemmed by smaller ones but they also seem to mark the coast as a boundary. 1 Riseley, 2 Polhill, 3 Eccles, 4 Saltwood, 5 Lyminge, 6 Bifrons, 7 Bishopsbourne, 8 Kingston, 9 Breach Downs, 10 Brooksend, 11 Sarre, 12 Ozengell, 13 Broadstairs, 14 Peters Tip, 15 Dover Buckland, 16 Beacon Hill, 17 Finglesham, 18 Gilton, 19 Updown, 20 Sibertswold/Barfriston. After Richardson 2005; Brookes 2011. © D Sayer.

The most westerly of the lathes is Sutton, which contains 27 identified early Anglo-Saxon cemeteries; eight of these are single inhumations and two contained over 100 burials, Riseley to the north and Polhill to the south. Equally, Aylesford contains 26 identified cemeteries, ten of which were individual inhumations and one of which, Eccles, contained over 100 inhumations. Limen, the southernmost lathe, contains 17 cemeteries, six of which are individual burials and two of which, Saltwood to the south and Lyminge to the north, contain over 100 inhumations. Wye contains 22 known cemeteries, six of which are individual graves, but none are currently believed to have contained over 100 inhumations. Milton is split into two geographic areas; the northern part contains 13 cemeteries, of which five are individual inhumations. However, many of these sites have only been partl­y excavated and the majority of Kentish cemeteries were excavated between the 18th and early 20th centuries so may be incomplete. Despite this, West Kent shows a similar pattern to the Avon Valley; each large site was hemmed by smaller cemeteries which acted as central places within a wider landscape of settlements and cemeteries.

East Kent is notable for the high concentration of early Anglo-Saxon cemeteries found in Eastry and the Borough (Fig 9). The description of The Borough presented in the 11th century contains the Isle of Thanet, which was listed as a hundred, and yet both geographically and in terms of cemetery distributions Thanet appears to have been an independent territory in the 6th century. The Borough (excluding Thanet) contains 22 cemeteries, of which six were individual burials and four contained over 100 inhumations (Bifrons, Bishopsbourne, Kingston and Breach Downs). All of these cemeteries date to the 7th century, which saw the development of Canterbury as an early Christian and civil centre within the kingdom of Kent.⁵⁸ Thanet contains 26 cemeteries, five of which were individual inhumations. Four sites, however, contained over 100 burials and it is possible that Brooksend at the north of the island is also a large cemetery, but it remains unexcavated.⁵⁹ Of the four that have been excavated, Sarre is to the west and Ozengell is to the south-east. Broadstairs is on the north-east coast and St Peters Tip is a 7th-century cemetery close by. These last two sites are exceptional because of their size: Ozengell is a cemetery of over 700 inhumations and St Peters Tip has over 400 graves. Eastry is similar to the Borough and Thanet in that there are six large cemeteries, with 37 in total, ten of which were lone burials. The larger cemeteries are Dover Buckland, Beacon Hill, Finglesham and Gilton, with two 7th-century sites further inland, Updown and Sibertswold/Barfriston at the centre of Eastry.

Both Thanet and Eastry contained a larger number of 6th-century cemeteries, more than the other lathes of Kent, but as with the Avon Valley they are situated on a boundary. The Eastry cemeteries were all set back from the coast and each one was hemmed by smaller cemetery sites.⁶⁰ The Eastry cemeteries are dispersed evenly and concentrate on the east coast with a small central cluster. However, this might be misleading because Upton in the centre of this cluster is a 7th-century site, and in the 6th century Finglesham was a small single plot cemetery which expanded significantly in the 7th to become a multi-plot, multi-focal site.⁶¹ This change might be best interpreted alongside the place-name evidence. According to Hawkes⁶² the village’s name is Old English derived from Þengels-Hām meaning ‘prince’s manor’, so Finglesham like the Canterbury cemeteries was the burial place for a new 7th-century elite. The three territories, Thanet, the Borough and Eastry have a complex history including the development of new boundaries in the 7th century; within these lathes central sites, the large cemeteries, acted as foci for territorial groups like those in West Kent and the Avon Valley.

Early Anglo-Saxon cemeteries were central places and focal points for local community funerals and commemoration activities,⁶³ but the communities which prepared and used these cemeteries did not exist in isolation. Local communities were part of networks within which identities were continually negotiated and re-negotiated. In both the Avon Valley and Kent certain cemeteries were central places within sub-regional territories, and the evidence from Kent suggests that these were folk groups or small tribal units that occupied territories which transformed over time. Within each territory single or multiple large 6th-century cemeteries were spaced out and hemmed in by several smaller cemeteries.⁶⁴ The larger burial grounds seem to have acted as central places for more than one community and fulfilled a palimpsest of roles as meeting places and commemorative spaces for folk identity groups; they encoded myths about shared antecedents, shared kinship networks and shared values. In other words, these central cemetery sites were important in the construction and maintenance of tribal cohesion, connecting disparate groups of people spread around a common landscape.

SOCIAL COHESION AND RESIDENTIAL MOBILITY

Historians interested in social systems, including Seebohn, Stenton, Goody and Murray,⁶⁵ have described the tribal nature of early Anglo-Saxon society. Tribal societies are organised around kinship groups with corporate descent providing a collective character. This may be a common ancestor with biological or mythical heritage,⁶⁶ but importantly it allows social networks and intercommunity ties to be created, reinforced and revisited through social institutions which included marriage strategies rooted within lifeway routines.

Some of the terminology that social anthropology uses to describe kinship is rooted in locality; a society might be described as patrilocal or matrilocal. In societies described this way the relocation of individuals through endogamy (marriage within a social group) or exogamy (marriage outside a social group) resulted in new spousal homes, the location of which was governed by gender.⁶⁷ As a result, landscapes encapsulated networks of people in constant change because of the repeated creation and destruction of social bonds. Spousal locality is dependent on specific sets of social rules within which various negotiations coexisted: familial structures, kinship ties, residential and domestic structures, marriage procedures and attitudes towards family fertility and reproduction.⁶⁸

Within early medieval archaeology the narratives which described Goths, Frisians, Angles, Saxon and Jutes has given way to fluid discussions about nested identities and the flexibility of material expression.⁶⁹ It is accepted that this plurality of persona is rooted within life course,⁷⁰ but it is also embedded within landscapes across which social relationships are played out. This means that it is culturally embedded behaviour, not simply material culture, which may be studied by archaeology. For example, many continental cemeteries are the product of sub-regional population groups; Hamfelde (Holstein) contained 884 male cremations from the 1st and 2nd centuries ad and at Kasseedorf (Holstein) there were 219 female cremations from the 3rd century ad. This gender⁷¹ locality for burial is also seen in Schelswig, where Süderrarup contained only male cremations (based on grave goods) and was used between the 1st and 4th centuries ad. By contrast, Issendorf in the lower Elbe region, west of Hamburg, was a mixed gender site, but with over 3795 cremations it was, like Hamfelde, a large cemetery which provided burial space for a catchment area containing numerous interconnected communities with a shared regional identity.⁷² These cemeteries operated as nuclei for groups who travelled to a particular spot to bury their dead and rebuild social bonds lost in death. These were central places for a sub-regional tribal network or Sippe.⁷³

England does not host the same mortuary landscape in the 6th century. The large cremation cemeteries in Lincolnshire (Cleatham, Ancaster, South Elkington) and Norfolk (Caster and Spong Hill) are like the inhumation cemeteries described in west Kent because many are hemmed by smaller inhumation, cremation and mixed rite sites. There are no large exclusive regional burial places, but that does not mean that early Anglo-Saxon cemeteries did not have a role within the construction and expression of regional identities.⁷⁴ Specific cemeteries provided centres for kinship networks over generations. These lifeway routines would have involved marriage, birth and death. In this way female mobility may account for differences in the proportions of infants and children. For a pregnant woman the safest place to give birth emotionally and perhaps physically was their maternal home.⁷⁵ It is unlikely that small hamlet-sized communities could all have experienced ‘midwives’ and so pregnant women may have travelled away from their marriage household to their mother’s home to be surrounded by experience and reassurance at a vulnerable time. But not all infants died in childbirth. Female mobility and a reliance on the maternal kindred for infant protection are attested by historical evidence:

Law of Ethelbert,⁷⁶ ad 600:

[74] If a person buys a maiden with a [bride-]price, let the bargain be [valid], if there is no deception.

[74.2] 76.2 If she bears a living child, let her obtain half the goods belonging to the household if the husband dies first.

[74.3] 76.3 If she should wish to dwell with the children, let her obtain half the goods [of the household].

[74.4] 76.4 If she should wish to take a man [ie, another husband], provision as for one child [ie, the inheritance is split equally between the mother and each of the children].

[74.5] 76.5 If she does not bear a child, her paternal kin should obtain [her] property and the morning-gift.

[75] 77. If a person takes a maiden by force: to the owner [of her protection] 50 shillings, and afterwards let him buy from the owner his consent [to marry her].

Law of Holthere and Eadric,⁷⁷ ad 673–85:

[4]. If a freeman should die with a living wife and child, it is right that it, that child, should be under the protection of the mother, and one should give for him one among his father’s kin who willingly gives surety to maintain his property, until he should be 10 years old.

Law of Ine,⁷⁸ ad 688–94:

38. If a husband and wife have a child together, and the husband dies, the mother is to have her child and rear it; she is to be given six shillings for its maintenance, a cow in summer, and ox in winter; the kinsmen are to take charge of the paternal home, until the child is grown up.

These 7th-century law codes are partially contemporary with the early Anglo-Saxon cemeteries. They are important for two main reasons. First, they identify a patrilocal residence pattern where children were ‘under the protection of the mother’ or when the paternal kinsmen take charge of a marital home the mother is to ‘have her child and rear it’. In all three examples woman moved for marriage and returned ‘home’ in the event of spousal death, or if they wished to leave. Secondly, they highlight the responsibility of the paternal kindred. Two similar codes in the laws of Holthere and Eadric and the laws of Ine indicate that a child stayed with its mother after paternal death until it was ‘grown up’, when it became entitled to the paternal inheritance.

These historical sources indicate that the rite of passage from infancy to childhood emphasised changing responsibilities and protection, although the age categories used in archaeological assessment may not have been so fixed in practice, and were probably subject to localised interpretation and fluidity. For infants, the responsibility for care was with the maternal kindred, changing with age (probably around five years) to the paternal kindred with whom an adult would have a closer connection. An infant was effectively an extension of its mother’s kin and a child its father’s. The archaeological pattern with proportionally more infants in larger cemeteries may be a result of this fluctuating obligation played out over generations. Central kin groups married their daughters to the sons of contiguous communities within their sub-regional network reaffirming the tribal bond with endogamy. A woman moved to her new marriage home, but if her infant died before it reached childhood its body was transported for burial in the principal cemetery associated with its mother’s kin. It was these children, the sons and daughters of the tribal elite, who were buried in central places and on the boundaries of sub-regional territories — locations important to tribal identities with a strong maternal kinship bias, sites like Abingdon, Berinsfield, Blacknall Fields, Great Chesterford, Lechlade, Meonstoke and Westgarth Gardens. These cemeteries contributed to the maintenance of wide kinship networks because it was in these places and during funerals within them that the tribal elite reaffirmed their cooperative network following the loss of important affines.

THE SOCIAL CONTEXT OF INFANT BURIAL

Infants are under-represented in early Anglo-Saxon cemeteries; however, the context of their discovery shows statistical variation between sites. In the total sample of 3412 graves, 369 (10.8%) were infants and 359 (10.5%) were children. This means that 700 (20.5%) of the sample were aged 12 and below. This figure of 10.8% infants from the total sample is similar to Crawford’s⁷⁹ c 11%. The World Health Organisation reports that 98.1% of infant deaths (aged under five) in 1990 were in developing countries; this figure rose to 98.6% in 2011. In 2012 in sub-Saharan Africa 109 children in 1000 died before the age of five, a fall from 1990 when 178 infants were lost in every 1000 live births. In 1990 infant mortality in European countries that have over 40% rural populations varied from 3 in every 1000 live births in Albania, to 10 in every 1000 in Slovenia and 37 per 1000 in Romania. Mortality statistics from Africa are extreme, caused by mechanised warfare, AIDS and prolonged famine. The whole of the developing world lost 97 infants in every 1000 in 1990 dropping to 57 in 2011.⁸⁰ However, modern infant mortality is influenced by medicine and economic development and the decline in mortuary statistics for the developing world is the result of an international political effort. Early Anglo-Saxon England did not have modern healthcare and had no large urban centres, AIDS or mechanised war. Early Anglo-Saxon society almost certainly employed cultural strategies which controlled female fertility and reduced the chances of both female and neonatal mortality.⁸¹ As a result, it is hard to see infant mortality ranging much beyond 10% or 12% unless driven by disease. Therefore cemeteries like Lechlaide, with a 17.4% infant mortuary population, included individuals from outside of the immediate community. And sites like Great Chesterford, with 40.1% infants, or Blacknall Fields, with 25.7% infants, were the result of regional fluidity in mortuary practices.

Infant bones are small, they are easy to miss, and they do not always survive, but it would be a mistake to assume that all early Anglo-Saxon cemeteries were the same. Quite the contrary: different sites may have served different functions just as the communities which used them had different and changing roles within a wider cooperative kinship network. Archaeologically it is possible to see some of these differences statistically; large sites and small sites contained different proportions of infants and children. However, not all large cemeteries contained significant numbers of infant burials and not all sites, large and small, have been fully excavated so it is important to consider the variation in how sites were used and thus how they were perceived by a wider folk community. Sites like Great Chesterford contain extraordinary numbers of infant graves and, although exceptional, Great Chesterford is not unique. The 67 infants buried over c 150 years in zones at Great Chesterford, and the cluster of infants at Apple Down or Westgarth Gardens, found adjacent to groups of older individuals, are unlikely to have been placed there by accident. These patterns were the product of deliberate behaviours, activities focused on reaffirming kinship networks, tribal identities and endogamous sub-regional marriage networks which operated across landscapes and within corporate identity groups. Within this context, kinship obligation fluctuated and infants were assonate with female identities and the mother’s kindred, but after a rite of passage into childhood the obligation and connection passed to the father’s kin. These cemeteries were the burying places for local families and the sub-regional tribal elite, they were central places for female genealogies and the graves of unfortunate infants; the sons and daughters of athelings buried in the earth next to real and imagined antecedents.

ACKNOWLEDGMENTS

I would like to thank Jim Morris for advice about statistics and Howard Williams for reading and commenting on various drafts of this article. Students in my 2012–13 post-Roman and early medieval class were particularly helpful and enjoyed a pilot exercise counting graves from dozens of published reports. The ideas which formed this paper were presented to visitors who came to see the Oakington cemetery excavation in 2012 and 2013 where the idea started to take shape, at EAA Pilzen in 2013 and at research seminars held at the Universities of Cambridge, Nottingham and Chester. In all these places people showed enthusiasm for the work, asked questions and proposed challenges which have helped shape this published paper. My final thanks go to Meredith Carroll for support and proof reading. While much of the thinking in this paper was shaped through discussion, all of the mistakes remain my own.

Notes

1 Line 93 from the Anglo-Saxon Poem The Seafarer, The Exeter Book. This line is taken from Professor J Glenn’s (1986) translation in which he prefers ‘sons of Athelings’, whereas Whitelock (1979, 873) preferred ‘sons of nobles’.

2 School of Forensic and Investigative Science, University of Central Lancashire, Preston PR1 2HE, UK. [email protected]

3 Lee 2008; Sayer 2013.

4 Sayer 2013.

5 Thompson 2004, 71–2; Hadley 2011, 288–314, 294.

6 Sayer and Dickinson 2013a; 2013b.

7 Hakenbeck 2009.

8 Hakenbeck et al 2010, 247.

9 Van Gennep 1965; Bloch and Parry 1999.

10 Glass et al 2011.

11 Brookes and Reynolds 2011.

12 Crawford 1991; 1993.

13 Crawford 1993, 84–5; 1999.

14 Buckberry 2000.

15 Under five years old.

16 Buckberry 2000; Lucy 1994.

17 Crawford 1993; Buckberry 2000.

18 Merbs 1997; Ingvarsson-Sundström 2003; Lewis 2006, 25–6.

19 Akerman 1855.

20 Buckberry 2000.

21 Crawford 1991; 1993; 1999; 2000; 2011.

22 Crawford 1999, 22–3.

23 Crawford 1999, 24; Evison 1987, 128.

24 Lucy 1997.

25 Lewis 2006; Scheuer and Black 2000.

26 Gilchrist 2012.

27 Sayer 2010.

28 Lewis 2006; Jackes 1992; Scheuer and Black 2000.

29 Crawford 1999, 30–1.

30 Stoodley 2000.

31 Härke 1997.

32 Chi-Square is derived (Shennan 1997, 106). Compare Tables 3 and 4, a P value of 0.05 or lower is considered significant. Infants and children are each 11% of the total cemetery population of 3412 even with outliers removed.

33 Evison 1987, 128; Hawkes and Grainger 2003, 153; Parfitt and Brugmann 1997, 217.

34 Penn and Brugmann 2007, 88.

35 Furnished burials describes those with grave goods, commonly costume, vessels or weapons. These are commonly differentiated by gender and wealth with unfurnished, furnished and richly furnished graves including different quantities and qualities of artefacts.

36 Stoodley 2011, 654.

37 Pader 1982.

38 Lee 2008; Crawford 2007.

39 Lee 2008, 28–36.

40 Crawford 2007, 90.

41 Down and Welch 1990, 9.

42 Boyle et al 1995, xvi–xviii.

43 Sayer 2010; Sayer and Wienhold 2013.

44 Berinsfield contained 10.5% infants, see Table 1.

45 West 1988.

46 Pader 1982.

47 Penn and Brugmann 2007, 86.

48 West 1988, 4.

49 Unfortunately cremations are rarely assigned to specific age groups.

50 Evison 1994.

51 There are no clusters of infant and child graves at Lechlade and at Blacknall Field these are ambiguous and smaller than those from Berinsfield. However, both Lechlade and Blacknall Field have high proportions of infants.

52 Dickinson 2012.

53 Scheschkewitz 2006, 11.

54 Dickinson 2012.

55 Brookes and Reynolds 2013; Brookes 2007; 2011; Richardson 2005.

56 Brookes 2011.

57 Ibid, 158–9; Campbell 1986, 95–6; Blair 1991; 1989, 98–103; York 1995, 39–43.

58 Lyle 2000, 48–50.

59 Perkins 1987.

60 Richardson 2005.

61 Sayer 2010.

62 Hawkes 1982, 24.

63 Härke 2001; Williams 2002.

64 Many sites have been only partially excavated and so this distribution will change as more is known about each site. Nevertheless some cemeteries acted and central places for multiple communities. It is these cemeteries, often large sites, which may have been associated with communities significant to the tribal hierarchy.

65 Seebohn 1902; Stenton 1971; Goody 1983; Murray 1983.

66 Charles-Edwards 1972.

67 Pasternack et al 1998; Goody 1976.

68 Handwerker 1977, 259.

69 Hakenbeck 2007; Williams and Sayer 2009; Lucy 2005.

70 Stoodley 2000; Gilchrist 2012.

71 These are cremation sites and so gender has been identified from grave goods.

72 Hills 1999a, 1999b; Ravn 2003; Scheschkewitz 2006, Michael Gebühr pers comm.

73 Ravn 2003.

74 Squires 2012.

75 Sayer and Dickinson 2013a; 2013b.

76 Oliver 2002, 79.

77 Ibid, 129.

78 Whitelock 1979, 391–403.

79 Crawford 1993.

80 UNICEF, WHO, World Bank, UN Population Division 2012.

81 Sayer and Dickinson 2013a; 2003b; Gilchrist 2012.