Towards Software Component Assembly Language Enhanced with Workflows and Skeletons
Chapter 4: The Human Skeletons IN: Anglo-Saxon Graves and Grave Goods of the 6th and 7th Centuries...
Transcript of Chapter 4: The Human Skeletons IN: Anglo-Saxon Graves and Grave Goods of the 6th and 7th Centuries...
Anglo-SAxon grAveS And grAve goodS of the 6th And 7th CenturieS Ad:
A ChronologiCAl frAmework
byAlex Bayliss, John Hines, Karen Høilund Nielsen,
Gerry McCormac and Christopher Scull
with contributions byMarion Archibald, Nancy Beavan, Mads K. Holst, Simon Mays and
Michelle Thompson
THE SOCIETY FOR MEDIEVAL ARCHAEOLOGY MONOGRAPH 33
Illustrations byIan Dennis
Edited byJohn Hines and Alex Bayliss
© 2013 The Society for Medieval Archaeology and authors
ISSN 0583-9106ISBN
Edited for the Society of Medieval Archaeology by John HinesPublished by The Society for Medieval Archaeology, LondonPrinted and bound by Charlesworth Press, Wakefield, UK
This publication has been made possible by a grant from English Heritage
The Society for Medieval Archaeologywww.medievalarchaeology.org/publications
The Society for Medieval Archaeology Monographs are available from Maney Publishing, www.maney.co.uk
Hines, J and Bayliss, A (eds) Anglo-Saxon Graves and Grave Goods of the 6th and 7th Centuries AD: A Chronological Framework, The Society for Medieval Archaeology Monograph 33, London
Cover: a 7th-century necklace from Desborough, Northants. Photograph by John Hines, reproduced by kind permission of the Trustees of the British Museum; design and layout by John Hines and Kirsty Harding.
iii
ContentS
iii
List of figures and tables ..............................................viSummary ...................................................................xviiRésumé (Français) ....................................................xviiZusammenfassung ................................................... xviiiRiassunto ................................................................. xviiiResumé (Dansk) .........................................................xix
foreword ......................................................................1Acknowledgements .....................................................11
1 the Archaeological Study of early Anglo-Saxon Cemeteries (by John Hines) .................13
1.1 Graves, Grave Goods and Chronology ...........131.2 Artefacts and Typology ...................................201.2.1 Typology as mode of thought and method of
analysis .............................................................201.2.2 Artefact typology in Early Anglo-Saxon
archaeology ......................................................221.3 Periodization ...................................................251.3.1 Concepts and principles ...................................251.3.2 Periodization in Anglo-Saxon archaeology: a
brief history ......................................................271.3.3 The present situation ........................................30
2 dating methods and their modelling (by Alex Bayliss, Gerry McCormac, Michelle Thompson and John Hines) ................................................33
2.1 Methodological Introduction ............................332.2 Radiocarbon Dating of the 6th and 7th Centuries
AD (by Alex Bayliss, Gerry McCormac, Michelle Thompson) ........................................35
2.2.1 High-precision radiocarbon dating ...................372.2.2 Calibration ........................................................442.2.3 Quality assurance .............................................492.2.4 Fractionation, reservoir effects and other
potential age-offsets .........................................552.3 Seriation and Correspondence Analysis (by John
Hines) ..............................................................602.3.1 Introduction .....................................................602.3.2 Seriation by correspondence analysis: an
introduction ......................................................612.3.3 How does correspondence analysis work? .......632.3.4 The presentation of the results .........................732.4 Bayesian Chronological Modelling (by Alex Bayliss) .............................................732.4.1 The Bayesian approach ....................................742.4.2 The components of a Bayesian chronological
model ................................................................75
2.4.3 Model construction...........................................782.4.4 Model calculation .............................................842.4.5 Model validation ..............................................842.4.6 Model comparison ............................................862.4.7 Model outputs...................................................872.5 Methodological Summary ...............................87
3 the Project (by Christopher Scull) .................893.1 Background and Context ..................................893.2 Project Approval, Planning and Management ..903.2.1 Aims .................................................................913.2.2 Methods ............................................................913.2.3 Working practice ..............................................913.2.4 Timescale, project stages, review and re-design ...........................................................923.2.5 Project management and assurance ..................923.3 Project History ................................................933.3.1 Stage 1 (MAIN) : 1998–2006 ..........................933.3.2 Stage 2 (ARCV): 2005–6 .................................953.3.3 Stages 3 and 4 (ANL): 2006–13 ......................953.4 Issues Encountered and Lessons Learned .......963.4.1 Communication ................................................963.4.2 Ways of working ..............................................973.4.3 Equipment capacity and performance .............973.4.4 Programming and personnel capacity ..............983.4.5 Availability and suitability of human skeletal
remains .............................................................98
4 the human Skeletons (by Nancy Beavan and Simon Mays) .................................................101
4.1 A Note on the Osteology of the Human Burials ............................................................1014.1.1 Methods .........................................................1014.1.2 Results ...........................................................1024.2 Carbon and Nitrogen Stable-isotope and Amino-
acid Screening for Quality Assurance ............1074.2.1 Carbon/nitrogen (C:N) ratios .........................1074.2.2 Amino-acid analysis ......................................1084.2.3 Discussion of stable-isotope and amino-acid
results for quality assurance ...........................1084.3 Stable Isotopes and Diet ................................1164.3.1 Dietary analysis ............................................1174.4 Anglo-Saxon Diets: The Effects of Sex, Age and
Geography ......................................................1284.5 Conclusions ....................................................131
iv
5 typology (by Karen Høilund Nielsen) ..........1335.1 Introduction ....................................................1335.2 BU: Buckles and belt-fittings .........................1365.3 SB: Shield Bosses ..........................................1485.4 SP: Spearheads ...............................................1635.5 SW: Swords and Scabbard-mounts ................1825.6 Seaxes, Seax-pommels, and Sheath-fittings ...1905.7 BE: Beads .......................................................2035.8 PE: Pendants ..................................................2105.9 WR: Wire Rings .............................................2165.10 BR: Brooches .................................................2185.11 PI: Dress Pins .................................................2245.12 Accessories, tools, personal equipment and
vessels ............................................................227
6 interpretative Chronologies for the male graves (by Alex Bayliss, John Hines and Karen Høilund Nielsen) ...........................................231
6.1 Radiocarbon Dates .........................................2326.2 An Archaeological Model ..............................2376.3 Male Artefact-types ........................................2416.3.1 Currencies ......................................................2426.3.2 Phyletic seriation ...........................................2486.4 Seriation by Correspondence Analysis...........2516.4.1 Constructing and phasing a sequence ............2516.4.2 Incorporating further chronological data within
the framework ................................................2896.5 Sensitivity Analyses .......................................2966.5.1 Radiocarbon calibration .................................2996.5.2 ‘Uninformative’ prior information .................3006.5.3 Continental parallels with calendar dates ......3016.5.4 Dietary effects ................................................3096.6 Dating of Typological Series ..........................3286.6.1 Buckles and belt-fittings ...............................3286.6.2 Swords and scabbard-mounts ........................3326.6.3 Seaxes and seax-fittings .................................3336.6.4 Shield bosses..................................................3346.6.5 Spearheads .....................................................3346.7 Conclusions ....................................................336
7 interpretative Chronologies for the female graves (by Alex Bayliss, John Hines and Karen Høilund Nielsen) ............................................339
7.1 Radiocarbon Dates .........................................3397.2 An Archaeological Model ..............................3397.3 Female Artefact-types ....................................3567.3.1 Currencies ......................................................3567.3.2 Phyletic seriation ...........................................3717.4 Seriation by Correspondence Analysis...........3727.4.1 Selecting the starting point of the
correspondence analysis ................................3727.4.2 Further exploration of the initial incidence
matrix .............................................................376
7.4.3 Adding further artefact-types and grave-assemblages to the correspondence analysis ..383
7.4.4 The final seriation and chronological model ..4087.4.5 The frequency distributions of the female
grave ...............................................................4157.4.6 Assessing the accuracy of the female seriation ..........................................................4167.5 Sensitivity Analyses .......................................4227.5.1 Regionality.....................................................4247.5.2 Radiocarbon calibration ................................4297.5.3 ‘Uninformative’ prior information .................4327.5.4 Continental parallels with calendar dates ......4327.5.5 Associated coins with historical dates from a
further English grave ......................................4367.5.6 Associated English coins with ‘numismatic’
dates................................................................4387.5.7 Dietary effects ................................................4407.6 Dating of Typological Series ..........................4497.6.1 Beads .............................................................4497.6.2 Pendants .........................................................4507.6.3 Wire-rings ......................................................4527.6.4 Brooches ........................................................4527.6.5 Buckles and belt-fittings ................................4527.6.6 Pins ................................................................4547.6.7 Accessories, personal equipment and vessels 4547.7 Conclusions ....................................................454
8 integrations and Comparisons (by Alex Bayliss, John Hines and Karen Høilund Nielsen) .................................459
8.1 The Male and Female Sequences ..................4598.2 Integrated and Comparative Modelling of Anglo-
Saxon Burials (by Alex Bayliss and John Hines) 462
8.2.1 Comparison of phase-boundaries in the male and female sequences ............................................462
8.2.2 The end of furnished burial in England..........4648.3 A Comparison of Artefact-types in Male and
Female Graves (by John Hines) .....................4738.4 The Frequency of Furnished Burial through
Time (by Alex Bayliss and John Hines) .........4768.5 The Anglo-Saxon and Overseas Chronological
Schemes Compared (by John Hines and Karen Høilund Nielsen) ............................................479
8.5.1 Systematic comparison of the chronologies ...4848.5.2 The male graves .............................................4848.5.3 The female graves ..........................................4898.6 Summary and Conclusions (by John Hines) ..491
9 numismatics and the Chronological models (by Marion Archibald, with John Hines and Christopher Scull) ..........................................493
9.1 Introduction ....................................................4939.2 A Review of the Evidence: The Numismatic
Perspective .....................................................493
v
9.2.1 Numismatic methodology ..............................4939.2.2 Anglo-Saxon and Merovingian numismatic
chronology......................................................4979.2.3 Graves dated by this project which contain coins ..............................................................5009.2.4 Conclusions from the review .........................5069.3 The Chronological Models and the Numismatic
Chronology: A Discussion..............................5099.3.1 Comments (by John Hines and Christopher
Scull) ..............................................................5099.3.2 Reply and acknowledgements (by Marion
Archibald) ......................................................512
10 the results and their implications (by John Hines, Christopher Scull and Alex Bayliss) ..517
10.1 The Archaeology of the Early Anglo-Saxon Period .............................................................517
10.1.1 The dating of Early Anglo-Saxon graves and grave goods (by John Hines) .........................517
10.1.2 The dating and study of Early Anglo-Saxon artefact-types (by John Hines) .......................519
10.1.3 Anglo-Saxon settlement archaeology (by Christopher Scull) ....................................52110.2 Burial Practice in Anglo-Saxon England (by Christopher Scull) ....................................52310.3 Implications for Anglo-Saxon Social History
(by John Hines) .............................................52910.3.1 Society and gender .........................................52910.3.2 Society and politics ........................................54310.4 Implications for Anglo-Saxon Economic History (by Christopher Scull) ....................................54510.5 Religion and Anglo-Saxon Burial Practices (by John Hines) ..............................................54810.6 Methodological Insights and Agenda for the
Future .............................................................55410.6.1 Scientific and Statistical Issues and the Further ..
Use of Chronological Modelling in the Anglo-Saxon Period (by Alex Bayliss) .....................554
10.6.2 Approaches to Anglo-Saxon Archaeology and ... History (by John Hines) .................................558
Bibliography ............................................................573
index of Sites and grave-assemblages ..................593
4
The human skeleTons
As described in Chapter 2.2, high-precision radiocarbon dating requires the destruction of substantial samples of human bone. Consequently, prior to destructive sampling for radiocarbon dating, the sampled skeletons were subject to osteological recording. Although many of the burials already had extant osteological reports, many of these are rather old, and the aim of the current work was to record all the skeletons to a common standard (Section 4.1). The skeletal material studied comes from 90 burials from 17 sites (Tab. 4.1). Since the current sample of burials was selected on the basis of possession of chronologically significant grave goods (most males have weapons, most females items of personal adornment), any patterning in the osteological data must be interpreted in the light of this selection process.
Carbon and nitrogen stable isotopes were measured and amino-acid analysis was undertaken on each of the dated skeletons as part of the essential quality assurance measures for the radiocarbon dating programme. This chapter considers these quality assurance procedures, utilizing amino-acid and stable-isotope analyses to assess the preservation state of the material and its suitability for radiocarbon dating (Section
4.2). We then examine the suite of nitrogen and carbon stable-isotope values for the skeletons that provide information about the likely diets of the dated individuals, including the potential presence of marine and freshwater components in diet and the consequent affect upon radiocarbon calibration (Section 4.3).
4.1 A note on the oSteologY of the humAn BuriAlS
4.1.1 methods
For each burial, bone preservation was scored subjectively as ‘good’, ‘moderate’ or ‘poor’ based on the degree of superficial erosion of bone surfaces. The degree of skeletal completeness was also assessed. In adults, sex was determined using pelvic and cranial morphology (Brothwell 1981). It is not feasible to determine the sex of immature individuals reliably using osteological indicators. For juveniles, age was estimated using dental development (Gustafson & Koch 1974) and epiphysial fusion (Flecker 1942). Adult age was estimated from dental wear (Brothwell 1981; Mays et al 1995). Stature in adults was estimated from
Nancy Beavan and Simon Mays
Site Site Code n original osteological reportApple Down, West Sussex ApD 3 Harman (1990)Aston Clinton, Tring Hill, Buckinghamshire AstCli 1 —Edix Hill, Barrington, Cambridgeshire EH 9 Duhig (1998)Berinsfield, Oxfordshire Ber 5 Harman (1995)Castledyke South, South Humberside CasD 9 Boylston et al (1998)Coddenham, Suffolk Cod 1 —Dover Buckland, Kent BuD 6 —Dunstable, Marina Drive, Bedfordshire MaD 6 Brothwell (1962)Ford, Laverstock, Wiltshire Fo 1 Harcourt (1969)Gally Hills, Banstead Down, Surrey GaH 1 Lutton & James (1976)Lakenheath, Eriswell, Suffolk Eris 1 —Lechlade, Butler’s Field, Gloucestershire Lec 12 Harman (1998)Melbourn, Water Lane, Cambridgeshire Mel 9 Duhig (2003)Mill Hill, Deal, Kent MH 9 Anderson & Andrews (1997)St Peter’s Tip, Broadstairs, Kent SPTip 13 —West Heslerton, North Yorkshire WHes 1 Cox (1999)Westgarth Gardens, Suffolk WG 3 West (1988)
Table 4.1The burials studied osteologically (N = number of burials analysed)
ANGLO-SAxON GRAVES AND GRAVE GOODS102
long-bone lengths. The formulae of Trotter and Gleser for American White males (Trotter & Gleser 1958, tab. 12) and American White females (Trotter & Gleser 1952, tab. 13) were used, without age correction. Cranial and post-cranial measurements were taken according to Brothwell (1981), and the non-metric variants of Berry and Berry (1967) and Finnegan (1978) were scored. Pathological conditions were also recorded.
4.1.2 results
The major observations are listed in Table 4.2. The remaining osteological data are held in archive by English Heritage at Portsmouth. The remainder of this osteological note concentrates on two aspects: the analysis of stature and the discovery of an early case of leprosy.
Stature
The summary statistics for stature are shown in Table 4.3. The mean stature for females resembles the figure of 161 cm given as a global mean for female skeletons in Early-medieval Britain by Roberts and Cox (2003, tab. 8.1). However, the male figure differs significantly from the male mean of 172 cm presented for Early-medieval Britain by the same authors (ibid.) (t = 4.42, p = <0.001). One possible explanation for the high male stature observed is that the males measured for stature in the current group simply happened to come from populations where male stature was greater than average for the Early-medieval period. Thirteen male burials measured for stature come from sites with large numbers of burials and for which the extant osteological report provides mean male stature values estimated using identical methods to those in the current work (this applies to Berinsfield, Castledyke South, Edix Hill, Lechlade (Butler’s Field), and Melbourn (Water Lane)). Of these 13 individuals, 12 show stature figures greater than the male mean for the site from which they come, a statistically significant imbalance (chi-square = 9.3, p = <0.005). This would seem to suggest that they are not only tall compared with the average stature for the period but also in comparison with other males from their communities.
From a review of several cemetery reports from Britain, Härke (1990; 1992a) found that in the 6th and 7th centuries, male individuals accompanied in the grave by weapons were on average 2 cm taller than those without (although he did not test this difference for statistical significance). The majority of the male burials in the current study material (33 out of 37) were buried with weaponry (one or more of the following: sword, seax, shield and spear). When the stature comparisons are confined to those with weapons (N = 17), the mean stature (176.7 cm) remains significantly different from Roberts
and Cox’s (2003) mean figure for Early-medieval males in Britain (t = 3.88, p = 0.001). The current male stature results would seem to be in line with Härke’s findings.
One explanation of the data (termed hypothesis 1) is that these weapon burials were of incomers from a taller population. Other possibilities involve differences in nutrition and/or disease experience in the study populations during the growth period. At a population level, adult stature is an indicator of conditions during the growth period (albeit a less sensitive one than height for age in children), poor conditions being associated with reduced stature (Tanner 1989, 161). It seems likely that the presence of weapons as grave goods indicates an elevated social status because weapon burials are generally richer in grave goods than non-weapon interments (Härke 1990). If this social status was ascribed at birth, then the social sub-group signified by burial with weapons may have had better living conditions in infancy and childhood and hence have attained higher than average stature. This is termed hypothesis 2. If the presence of weapons as grave goods indicated a warrior, or some other social role for which an imposing physique might be advantageous, then taller adults might be selected for this social position or may more often have achieved this, or taller juveniles might preferentially be selected as candidates for such a role in future (hypothesis 3). These taller members of society may be so because of genetic factors (they come from tall families) and/or because they enjoyed better conditions during childhood. Hypothesis 2 implies a significant role for environmental factors (better nutrition and/or less disease) in the causation of the tall stature of the weapon burials, as may hypothesis 3.
In his work, Härke (1990; 1992a) favoured hypothesis 1. Drawing on various strands of evidence, he argued that rather than signifying warrior status, the inclusion of weapons as grave furnishings served as identifiers of individuals of Germanic origin, whereas those without were, he thought, of native British stock. For this explanation to work, the Germanic migrants would need to have been taller than the native British. Härke notes that there was a stature increase between the Romano-British and the Early-medieval periods, which might be consistent with this. However, subsequent large-scale surveys of stature in early populations in Britain (Roberts & Cox 2003), whilst confirming the stature increase from the Roman to Early-medieval periods, also demonstrate fluctations in mean values between other archaeological periods which it would be naïve to ascribe to influxes of taller or shorter invaders. The stature increase between Roman and Anglo-Saxon Periods in Britain is not therefore compelling evidence of an influx from a taller foreign population.
THE HUMAN SKELETONS 103
grave Sex Age Stature Preservation Completeness deh Cribra weaponsApD107 F 30–49 M 60–80 PApD117 F 50+ 155.1 M 40–60 1 0ApD134 F 18–29 M 60–80 0 0AstCli12 F 50+ 159.4 M 60–80 0EH007 M 30–49 171.0 G 80+ 0 0 1EH012 M 30–49 183.0 G 80+ 0 1EH014 F 50+ 167.8 M 80+EH033 M 18–29 174.0 M 40–60 0 0 1EH048 M 30–49 176.0 G 80+ 0 P 1EH079 F 30–49 160.8 M 80+ 0 0EH083 F 18–29 169.4 G 80+ 0 PEH090 F 50+ 156.0 G 80+ 0EH091 U 18–29 M 20–40 0Ber022 F? 18–29 G 40–60 0Ber028 M 30–49 179.8 M 80+ 0 0 1Ber102 F 18–29 170.4 M 80+ 1 0Ber107 F 18–29 158.8 M 80+ 0 0Ber134/1 F 50+ M 60–80 0CasD013 F 18–29 157.1 M 60–80 1 0CasD053 F 30–49 161.3 M 60–80CasD088 F 18–29 167.8 M 60–80 0 PCasD094 M 30–49 179.8 P 40–60 1CasD096 F 18–29 159.6 M 60–80 0CasD120 ?F 13–17 M 60–80CasD134 ?F 13–17 160.8 M 60–80 0 0CasD182 M 18–29 172.3 M 40–60 1 CasD183 F 18–29 154.5 M 60–80 0Cod30 F 30–49 179.6 P 60–80BuD222 F 50+ P <20 0BuD250 F 30–49 M 80+ 0 0BuD264 M 30–49 P 20–40 0 0 1BuD323 M 30–49 P 20–40 0 0 1BuD339 F 18–29 P 40–60 0BuD414 M 30–49 180.6 P 20–40 0 P 1MaDC7 F 18–29 156.2 G 80+ 1 CMaDD10 F 18–29 169.3 G 80+ 0 PMaDE1 J 7–12MaDE2 J 13–17 M 80+ 0 TMaDE3 J 7–12 P 60–80 0 0MaDF2 J 0–6 M 60–80
Table 4.2. Catalogue of principal results by burial. Site: codes as in Table 4.1. Sex: M = male, F = female, U = unsexed adult (aged 18+ years), J = unsexed juvenile (under 18 years), ? used as a suffix or prefix indicates probable sex in adults and juveniles respectively. Age: age class in years unless stated. Stature: in cm. Preservation = gross bone preservation, scored on a subjective basis according to the degree of post-depositional erosion shown by the bones: G = good, M = moderate, P = poor. Completeness: approximate skeletal completeness as a percentage. DEH (dental enamel hypoplasia): 0 = absent, blank denotes anterior dentition missing so condition could not be scored, 1 denotes condition present. Cribra (orbitalia): 0 = absent, blank denotes orbits missing so condition could not be scored, letter entries indicate condition present, letter indicates type of lesion (P = porotic, C = cribriotic, T = trabecular (scored following Brothwell 1981)). Weapons: 1 = adult male weapon burial (i.e. with one or more of the following grave goods: seax, sword, spear, shield).
ANGLO-SAxON GRAVES AND GRAVE GOODS104
grave Sex Age Stature Preservation Completeness deh Cribra weaponsFo M 30–49 180.8 M 80+ 0 0 1GaH M ADULT 188.1 M 20–40 1Eris104 M 18–29 177.2 G 80+ 0 0 1Lec014 ?F 13–17 M 80+ 0Lec018 F ADULT 160.0 M 80+Lec040 M 30–49 172.3 M 80+ 0 0 1Lec136 F 18–29 163.2 G 80+ 0 0Lec138 F 18–29 155.4 M 60–80 0Lec148 J 0–6 M 40–60 0 PLec155 M 18–29 173.6 G 80+ 0 0 1Lec172/1 M 30–49 172.9 G 60–80 0 0 1Lec172/2 J 0–6 G 80+ PLec179 F 30–49 154.4 M 60–80 2 0Lec183 M 50+ 172.3 M 80+ 0 0 1Lec187 F 30–49 162.9 M 80+ 0 0MelSG069 U 30–49 P 80+ 0MelSG075 F 50+ P 60–80 0MelSG077 M 18–29 G 80+ 4 P 1MelSG078 F 30–49 158.8 G 80+ 0 0MelSG079 M 50+ 180.9 G 80+ 0MelSG080 M 50+ 177.8 M 80+ 0MelSG082 F 18–29 157.6 G 80+ 0 0MelSG089 F 30–49 163.0 M <20 0 CMelSG095 F 30–49 P 40–60 0MH040 M 30–49 P 20–40 0 1MH064 F? 18–29 P <20 0MH068 J 13–17 P <20 0 0MH079 U 30–49 P <20 0MH081 M ADULT P 20–40 1MH093 M 30–49 P 40–60 0 1MH094 M 50+ P 20–40 0 0MH095 F 50+ 163.3 P 60–80 0MH105c F 30–49 151.9 M 60–80 0 0SPTip008 M 50+ P <20 1 0 1SPTip042 M 18–29 P 20–40 0 0 1SPTip068 M 30–49 179.7 P 20–40 1SPTip073a ?F 13–17 P 40–60 0 0SPTip113 M 18–29 P 40–60 0 P 1SPTip194 M 30–49 171.3 P 60–80 1SPTip196 M 30–49 171.0 P 60–80 0 0 1SPTip208 F 50+ P 40–60 0 PSPTip212 M 30–49 P 40–60 0 0 1SPTip250 M 50+ P 60–80 0 0 1SPTip263 M 30–49 P 40–60 0 0 1SPTip318 M 30–49 P 40–60 0 1SPTip360 M 30–49 P 40–60 0 0 1WHes113 ?F 13–17 M 60–80 0WG11 M 30–49 P 20–40 0 1WG27 U ADULT P <20WG66 M 30–49 P 40–60 0 0 1
THE HUMAN SKELETONS 105
It is also unclear whether mean male stature was greater in Germanic populations than those from England during the Early-medieval period. For example, Bruchhaus and Finke (1997) present a mean figure of 172.1 cm collated from various cemeteries in Saxony and Thüringia, and Huber (1968) cites a mean of 173.7 cm for a large cemetery in Lower Saxony. Broadly similar figures have been obtained from Early-medieval sites from south-west Germany (Czartnetzki et al 1985; Siegmund 2010, 41–8). Reviewing the osteological evidence up to 1968, Huber (1968) concluded that Early-medieval populations from northern Europe should be regarded as a homogeneous group as far as stature was concerned; there seems little reason to revise that conclusion.
Härke (1990) evaluated a possible environmental contribution to the tall stature in his weapon burials by studying data on dental enamel hypoplasias (DEH). DEH are transverse lines on the dentition caused by interruptions of growth of the tooth crowns. They are indications of disease or poor nutrition during the period in which the crowns were forming (generally early-middle childhood) (Goodman & Rose 1990). Härke found that the frequencies of DEH reported for weapon burials was no less than for other burials, and argued that this counted against an important environmental factor in the stature difference. DEH are often rather faint. This means that there is a large element of subjectivity in their scoring. DEH data gathered by different authors cannot reliably be compared, as a defect that may be scored by one worker may be disregarded by another (Mays 2010). The fact that Härke relied on DEH data in osteological reports prepared by a variety of different authors makes his data problematic.
In this study, Mays recorded DEH in individuals who had one or more anterior teeth available for examination. He also scored the presence or absence of cribra orbitalia. This is a form of pitting in the roofs of the orbits, which probably reflects anaemia suffered in childhood (Mays 2010, 209–11). There are problems when comparing
frequencies of cribra orbitalia with published data because most data are presented with number of individuals in the cemetery as the denominator rather than number of individuals with orbital roofs intact for observations to be made. With the problems of inter-observer comparisons in both DEH and cribra orbitalia, it is difficult to compare current results with those published in the literature by other authors for Early-medieval populations. With this in mind, these data are instead compared with the later medieval population from Wharram Percy (N. Yorks), for which Mays (2007) has also scored DEH and cribra orbitalia in an identical fashion to the current material. Wharram Percy is a low-status rural site, and the burials date mainly to the 11th to 14th centuries AD. The suggestion is that if our weapon burials are high-status from birth and had a privileged upbringing: hypothesis 2), or were selected from individuals of tall stature (who would on average have likely had more favourable conditions for growth than those of shorter stature, assuming an environmental component in stature differences: hypothesis 3) then they might be expected to show lower rates of DEH and cribra orbitalia than the medieval peasants from Wharram Percy. The results are shown in Table 4.4.
DEH show a lower frequency among weapon burials than in Wharram Percy males (chi-square = 10.1, p = <0.005), consistent with the hypothesis of favourable growth conditions for the former. However there is no difference in the prevalence of cribra orbitalia. There are differences in the nature of these stress indicators. Cribra orbitalia is not a direct indicator of growth interruption, but DEH are a direct consequence of interruption of the growth of the dentition, and have been shown to correlate with interruptions in long-bone growth (Mays 1995). In summary, the pathology data provide at least some evidence to support the idea that the tall stature of the weapon burials is due to favourable conditions during the growth period. This tends to favour hypothesis 2 or 3 over hypothesis 1 but does not shed light on which of the former is more likely.
n mean sdMales 20 176.7 4.8Females 26 161.3 6.3
Table 4.3Stature of the skeletons analysed (in cm).
deh Cribra orbitaliaAnglo-Saxons Wharram Percy Anglo-Saxons Wharram Percy
Males* 2/28 (7%) 47/122 (38%) 4/23 (17%) 23/143 (16%)Females 5/26 (19%) 22/92 (24%) 7/26 (27%) 26/109 (24%)
Table 4. 4Frequency of dental enamel hypoplasia and cribra orbitalia in the current study material and at Wharram Percy (North Yorkshire). *Weapon burials only for the male Anglo-Saxons.
ANGLO-SAxON GRAVES AND GRAVE GOODS106
Some studies of skeletal material from Early-medieval Germany also show differences in stature among males according to whether they were buried with weapons or not, or else according to social rank. Straub (1956), Czarnetzki et al (1985) and Siegmund (2010) report stature differences by social rank in some cemeteries in southern Germany, those of higher rank being taller. Also looking at southern German material, Huber (1968) divided male burials into ‘heavily armed’, ‘lightly armed’ and ‘unarmed’, and found that the heavily armed group were the tallest, the unarmed the shortest, with the lightly armed occupying an intermediate position. It could be that the weapon burials or other high-status individuals are, in each case, incomers from some unspecified but taller population(s). However, the observation that northern Europeans are at this time rather homogeneous in terms of stature weakens this explanation. There seems no compelling reason to accept that weapon or other high-status burials were migrants from taller populations. It may be that in both continental Europe and in Britain, wealthier social groups (or at least their male members) simply had better living conditions.
If the males in the current study group had particularly good nutrition during the growth
period, then one might potentially expect to see dietary differences between them and the females revealed in stable isotope ratios. The stable isotope data did produce some evidence of this. In the males, there was a higher δ15N in the 30–49 year age class than in the 18–29 year group, the former being enriched in 15N compared to the latter (mean 9.7‰ vs 9.1‰). This suggests a minor dietary change during the life-cycle of the males, which did not have a parallel in females. Cortical bone collagen turnover rates are only about 1.5–3% per year in adult males (Hedges et al 2007), so this finding must relate to a dietary shift which occurred well before 30 years of age, probably during the latter part of the growth period. Although a multiplicity of factors may affect bone collagen δ15N (Section 4.3, below), one explanation would be a change during childhood to a diet where a greater part of the protein came from animal products.
In summary, the current data show a higher than average stature among male weapon burials. Environmental factors, perhaps in the form of better nutrition, may have been important causes of this pattern.
Härke (1990; 1992a) noted that weapon burials showed no greater frequency of blade injury than other male burials, and he used this (amongst other
Figure 4.1Superior view of the maxillae from burial E1, Marina Drive, Dunstable (Herts). There is resorption of the anterior nasal spine (long arrow), rounding of the pyriform aperture (short arrows), and well- remodelled new bone formation on the superior surface of the hard palate.
THE HUMAN SKELETONS 107
evidence) to suggest that current or past warrior status was not a decisive factor in influencing whether or not an individual was buried with weapons. The current data are consistent with this. With the possible exception of EH048 (Duhig 1998, 182–4), none shows evidence for blade injury. In addition, the overall prevalence of healed fractures among the weapon burials was similar to that in the female burials: 4/33 (12%) of weapon burials showed a fracture (5/2105 bones examined) compared to 6/37 (16%) of women (7/3363 bones examined). It would seem that whatever social roles were signified by burial with weapons, they conferred no additional risk of skeletal injury.
An early case of leprosy
MaDE1 showed facial signs of leprosy (Fig. 4.1). The hard palate was thickened by deposition of new bone upon the superior surface. The new bone was well-remodelled and slightly pitted. There is widening of the pyriform aperture, and its margins lack their normal sharpness, being thickened and rounded. There is resorption of the anterior nasal spine. This triad of changes is strongly suggestive of leprosy (Møller-Christensen 1961; Andersen & Manchester 1992). No other parts of this rather fragmentary skeleton showed abnormalities. There was another, less certain case of leprosy from the same site (burial E3). Like E1, E3 showed pitting of the superior surface of the hard palate, indica-tive of long-term inflammation in the nasal cavity. However the nasal aperture and anterior nasal spine were normal. The hand and foot bones were miss-ing.
The first cases of leprosy in Britain showing diagnostic facial signs date from the early Anglo-Saxon period (Table 4.5). Cases are also known from Early-medieval continental Europe (e.g. Lauchheim, Baden-Württemberg, Germany, AD 450–680: Boldsen 2008).
4.2 CArBon And nitrogen StABle-iSotoPe And Amino-ACid SCreening for QuAlitY ASSurAnCe
The accuracy of radiocarbon dates on human bone is affected by several factors, of which we consider two in this section: the preservation state of bone protein, and contamination with exogenous carbon. Bone protein may be too poorly preserved for accurate dating (Gillespie et al 1984; Masters 1987; van Klinken 1999) due to the loss of native nitrogen during diagenesis in burial environments that are either too arid or, alternatively, are waterlogged. Secondly, exogenous carbon from the burial environment or post-excavation conservation methods can contaminate the protein, resulting in incorrect dates due to the linking of carbon from burial contaminates such as humic acids with native bone protein. We have employed two methods by which to screen bone protein as part of the quality assurance procedures for this project: carbon to nitrogen ratios (atomic C:N ratios) and amino-acid analysis.
We first provide a background to understand the basics of both C:N and amino-acid analysis, and then present an overview of the results of these quality assurance procedures. Full accounts of the methods of physical and chemical preparation and measurement at the Queen’s University, Belfast and Rafter Radiocarbon New Zealand laboratories, and details of the amino-acid analysis methodology are presented by McCormac et al (2011) and Beavan et al (2011).
4.2.1 Carbon/nitrogen (C:n) ratios
Stable-isotope analysis of carbon and nitrogen provides a screening method to determine the survival of protein and the extent of exogenous carbon contamination in archaeological bone. The atomic C:N ratio of a given sample is compared to the expected carbon and nitrogen percentage range found in modern and well-preserved bone protein.
Site Burial Approximate dating Sex & age referenceEdix Hill, Barrington, Cambridgeshire
Grave 18, skeleton 42B AD 575–650 M, 17–25 Duhig 1998
Broughton Lodge, Nottinghamshire 62/63/64E AD 475–600 U, 50+ Roberts 1993
Burwell, Cambridgeshire Grave 111 7th century AD M, Adult Møller-Christensen and Hughes 1962
Cannington, Somerset Burial 159 5th century AD? F, ca. 22 Brothwell et al 2000
Marina Drive, Dunstable, Bedfordshire Burial E1
cal AD 650–670 (95% probability; UB-4552 (MaDE3); Fig. 7.65)
J, 7–8
Eccles, Kent 7th century AD M, 25–30 Manchester 1981
Table 4.5 Early Anglo-Saxon cases of leprosy from Britain. Sex: M = adult male, F = adult female, U = unsexed adult, J = juvenile. All approximate datings, except for Marina Drive, Dunstable, are based on artefact typologies.
ANGLO-SAxON GRAVES AND GRAVE GOODS108
Atomic C:N ratios (cf. DeNiro 1985) are calculated using the formula:
Atomic C:N = (%C/%N) x weight percentage C:N
Where weight percentage C:N = 1.1666
The optimum C:N ratio range for well-preserved bone protein is 2.9–3.6, established by DeNiro (1985) in a series of analyses on archaeological bone of different preservation states. Bone protein is on average 35% to 40% carbon and 15% to 20% nitrogen. However, combinations of a range of carbon from 30% to 50%, and nitrogen of 7% to 22%, can produce C:N ratios in the acceptable range. Lower nitrogen percentages indicate protein degradation, and higher carbon percentages can indicate exogenous carbon contamination. Therefore, the percentages of carbon and nitrogen are themselves indicative of protein survival or of exogenous carbon contamination.
Table 4.6 lists the stable-isotope results from the 94 skeletons and 14 replicate samples that were measured. The entire Anglo-Saxon data-set has a mean C:N ratio of 3.2, and is within the expected C:N range for modern or well-preserved prehistoric collagen (DeNiro 1985), with one significant outlier of 4.2 (Grubbs test, Z-score 6.985, p<0.05) for UB-6042 (CasD088). The result for UB-6042, with 50.8% carbon and 14% nitrogen, suggests the possible presence of an exogenous carbon contaminant; the C:N outside of the DeNiro acceptance range is due to a nitrogen value of 14%, although this is in itself not indicative of badly preserved protein. This observation can be compared with UB-6040 (CasD053), which has 57.3% carbon, but a C:N ratio of 2.9; this ratio falls within the acceptable DeNiro range, due to the 23.4% nitrogen present .
4.2.2 Amino-acid analysis
The amino-acid analysis screening method compares a suite of seven key amino-acids: hydroxyproline (Hyp), aspartic acid (Asp), glutamic acid (Glu), proline (Pro), glycine (Gly), alanine (Ala) and arginine (Arg) (Stafford et al 1988) against expected amounts of each amino in the profile composition of modern, unaltered bone protein (Veit et al 2006; van Klinken & Mook 1990).
The relative percentages of amino-acids in fossil collagen can be altered by exogenous amino-acids in the burial environment, or oxidation/deamination due to the differential solubility of amino-acids (Tuross 2002; van Klinken & Mook 1990; Turban-Just & Schramm 1998; DeNiro & Weiner 1988; Long et al 1989; Law & Hedges 1989; Weiner & Bar-Yosef 1990). Degraded collagen has a relative increase in smaller amino-acids, including alanine
and glycine (Turban-Just & Schramm 1998, 111), with Gly/Ala ratios that deviate from the ideal 2.8 expected for non-degraded bone. Glycine residues comprise approximately one-third of the amino-residues in collagen. Low glycine concentrations in fossil bones indicate deterioration of this amino-acid; higher concentrations than expected for pristine bone could indicate contamination with exogenous glycine (van Klinken & Mook 1990, 157). Aspartic acid is abundant in non-collagenous bone proteins as well as in the environment (i.e. bacterial protein), and Gly/Asp ratios are frequently used to test for contamination (DeNiro & Weiner 1988; Long et al 1989; Law & Hedges 1989; Weiner & Bar-Yosef 1990). Hydroxyproline is not bone-specific, as exogenous hydroxyproline occurs in soil as a free amino-acid and can originate from decayed plant remains and fungi (Radhakrishnan & Giri 1954; Greenstein & Winitz 1961) or diatom cell walls (Hardy 1985). Therefore, hydroxyproline values less than 101±5 residues/1000 are suggestive of protein deterioration, whereas values greater than 101 residues/1000 would indicate the presence of exogenous hydroxyproline. Glutamic acid is also abundant in the burial environment and as an outcome of diagenesis likely to contribute to a fossil bone profile (van Klinken & Mook 1990).
The amino-acid profiles from the Anglo-Saxon skeletons have been compared to the expected values for amino-acid profiles in non-degraded bone (van Klinken & Mook 1990, 156), the C:N range for modern and good prehistoric collagen (DeNiro 1985), and ideal Gly/Ala and Gly/Asp ratios based upon a modern collagen profile of van Klinken and Mook (1990, 156). Gly/Asp ratios of greater than 6–7 are expected for pure to well-preserved collagen (Weiner and Bar-Yosef 1990). Together these indicators provide a better assessment of the robustness of the bone protein than any one type of analysis alone.
4.2.3 discussion of stable-isotope and amino-acid results for quality assurance
For the 80 skeletons plus six replicate samples, for which amino-profiles were completed (Tab. 4.7), hydroxyproline is consistently below the expected 101 residues/1000 for well-preserved collagen (mean 75 residues/1000, min. = 39 residues/1000, max. = 119 residues/1000, SD = 13). Glutamic acid in most burials is 1 to 25 residues/1000 above the expected value, and glycine values in approximately half of the samples are 38 to 45 residues/1000 higher than the expected values. In particular, a combination of low hydroxyproline, increased glycine, and alanine, and variation from an ideal Gly/Ala ratio of 2.8 occurs in burials from Edix Hill (UB-4509, UB-4511 and UB-4512), Marina Drive (UB-4550 and UB-4553), Melbourn (UB-4883, UB-4884, UB-4887, UB-4890), Mill
THE HUMAN SKELETONS 109
Table 4.6. Stable-isotope results from 94 skeletons and 16 replicate samples. Carbon and nitrogen analysis of gelatin pre-pared by Rafter Radiocarbon 1) Error on stable isotope analysis is determined from mean error on EDTA standards within the stable isotope analysis run; 2) Total error reported for 15N and 13C includes analytical error and vari-ation in stable-isotope results associated with chemistry preparation of the bone protein; 3) Atomic C:N ratio = 1.1666(%C/%N).
laboratory number Site
Carbon and nitrogen analysis of rafter gelatin preparations
%C δ13C(‰)
error (±)1
total error (±)2
%n δ15n(‰)
error (±)1
total error (±)2
atomic C:n3
UB-5208 Apple Down: grave 107 42.3 −20.3 0.2 0.36 15.0 8.0 0.2 0.32 3.3UB-4965 Apple Down: grave 117 45.6 −20.6 0.1 0.32 16.7 7.9 0.3 0.39 3.2
UB-6344 Apple Down: grave 117 (replicate) 38.5 −20.2 0.2 0.36 13.8 8.2 0.2 0.32 3.3
UB-4835 Apple Down: grave 134 39.5 −20.2 0.2 0.36 14.4 8.7 0.2 0.32 3.2UB-4975 Aston Clinton: grave 12 39.3 −20.8 0.1 0.32 14.0 9.4 0.3 0.39 3.3UB-4735 Berinsfield: grave 22 50.7 −19.9 0.12 0.32 18.4 10.3 0.26 0.36 3.2UB-4736 Berinsfield: grave 28 31.4 −20.0 0.12 0.32 11.2 9.5 0.26 0.36 3.3UB-4739 Berinsfield: grave 134/1 39.9 −20.3 0.12 0.32 14.6 8.8 0.26 0.36 3.2UB-4077 Buttermarket: grave 4275 42.4 −20.2 0.07 0.31 15.6 10.7 0.18 0.31 3.2UB-4046 Buttermarket: grave 4344 44.2 −20.1 0.07 0.31 16.4 9.4 0.18 0.31 3.2UB-6036 Castledyke South: grave 13 42.2 −20.9 0.1 0.32 18.3 11.5 0.3 0.39 2.7UB-6040 Castledyke South: grave 53 57.3 −21.1 0.1 0.32 23.4 10.2 0.3 0.39 2.9UB-6042 Castledyke South: grave 88 50.8 −20.3 0.1 0.32 14.0 10.1 0.3 0.39 4.2
UB-6039 Castledyke South: grave 94, skeleton 1452 46.5 −20.5 0.1 0.32 17.8 10.1 0.3 0.39 3.0
UB-6035 Castledyke South: grave 96 43.1 −21.3 0.1 0.32 18.1 11.0 0.3 0.39 2.8
UB-6034 Castledyke South: grave 120 45.4 −20.9 0.1 0.32 14.7 9.7 0.3 0.39 3.6
UB-6037 Castledyke South: grave 134 48.5 −20.6 0.1 0.32 18.5 10.0 0.3 0.39 3.1
UB-6041 Castledyke South: grave 182 48.4 −20.5 0.1 0.32 17.6 8.9 0.3 0.39 3.2
UB-6038 Castledyke South: grave 183 44.7 −21.0 0.1 0.32 16.7 9.9 0.3 0.39 3.1
UB-4964 Coddenham: grave 30 45.9 −20.4 0.1 0.32 17.1 10.3 0.3 0.39 3.1UB-6472 Dover Buckland: grave 222 43.5 −19.8 0.2 0.36 16.1 9.8 0.2 0.32 3.2
UB-6472 Dover Buckland: grave 222 (replicate) 38.4 −20.1 0.1 0.32 14.1 9.6 0.2 0.32 3.2
UB-6473 Dover Buckland: grave 250 39.5 −19.8 0.2 0.36 14.4 10.4 0.2 0.32 3.2UB-6474 Dover Buckland: grave 264 39.2 −19.7 0.2 0.36 14.4 9.3 0.2 0.32 3.2
UB-6474 Dover Buckland: grave 264 (replicate) 42.6 −20.0 0.1 0.32 15.9 9.5 0.2 0.32 3.1
UB-6475 Dover Buckland: grave 323 37.4 −19.8 0.2 0.36 13.8 9.8 0.2 0.32 3.2UB-6476 Dover Buckland: grave 339 41.9 −19.7 0.2 0.36 15.6 8.9 0.2 0.32 3.1UB-4958 Dover Buckland: grave 375 43.0 −19.8 0.12 0.32 15.9 9.0 0.26 0.36 3.2
UB-4959 Dover Buckland: grave 391A 28.5 −20.3 0.12 0.32 10.3 9.8 0.26 0.36 3.2
UB-6477 Dover Buckland: grave 414 40.0 −19.8 0.2 0.36 14.8 9.4 0.2 0.32 3.2UB-4923 Edix Hill: grave 7 39.5 −20.3 0.1 0.32 14.3 10.4 0.3 0.39 3.2
UB-4923 Edix Hill: grave 7 (replicate) 38.2 −20.4 0.1 0.32 14.1 10.7 0.2 0.32 3.2
UB-4508 Edix Hill: grave 12 34.3 −20.0 0.12 0.32 12.5 10.2 0.26 0.36 3.2UB-4709 Edix Hill: grave 14 34.9 −20.4 0.12 0.32 12.5 10.4 0.26 0.36 3.3UB-4509 Edix Hill: grave 33 37.5 −20.2 0.12 0.32 13.5 9.8 0.26 0.36 3.2UB-4510 Edix Hill: grave 48 nd nd nd nd nd nd nd nd nd
UB-4922 Edix Hill: grave 48 (replicate) 37.4 −20.6 0.1 0.32 13.4 9.5 0.3 0.39 3.3
ANGLO-SAxON GRAVES AND GRAVE GOODS110
laboratory number Site
Carbon and nitrogen analysis of rafter gelatin preparations
%C δ13C(‰)
error (±)1
total error (±)2
%n δ15n(‰)
error (±)1
total error (±)2
atomic C:n3
UB-4707 Edix Hill: grave 79 28.2 −20.3 0.12 0.32 10.2 8.8 0.26 0.36 3.2UB-4708 Edix Hill: grave 83 38.8 −20.2 0.12 0.32 14.3 10.9 0.26 0.36 3.2UB-4511 Edix Hill: grave 90 36.5 −20.1 0.12 0.32 13.1 9.6 0.26 0.36 3.3
UB-4511 Edix Hill: grave 90 (replicate) 39.6 −20.7 0.10 0.32 14.5 9.9 0.2 0.32 3.2
UB-4512 Edix Hill: grave 91 51.6 −20.2 0.12 0.32 18.4 10.7 0.26 0.36 3.3UB-4976 Ford, Laverstock: barrow 2 40.4 −20.4 0.1 0.32 14.4 9.0 0.3 0.39 3.3
UB-4920 Gally Hills: replicate (post PVA extraction) 42.5 −20.2 0.1 0.32 15.0 10.4 0.3 0.39 3.3
UB-4727 Gally Hills: primary burial (PVA) Not analysed due to PVA contamination
UB-6347 Lakenheath: ERL 104 4222 38.8 −19.8 0.2 0.36 13.5 9.2 0.2 0.32 3.4UB-4501 Lechlade: grave 14 40.4 −19.9 0.12 0.32 14.8 10.1 0.26 0.36 3.2UB-4984 Lechlade: grave 18 39.0 −20.4 0.1 0.32 13.9 8.0 0.3 0.39 3.3
UB-4984 Lechlade: grave 18 (replicate) 37.3 −20.6 0.1 0.32 13.5 9.1 0.2 0.32 3.2
UB-4683 Lechlade: grave 40 41.6 −19.7 0.2 0.36 15.3 10.5 0.2 0.32 3.2
UB-4683 Lechlade: grave 40 (replicate) 42.8 −20.5 0.1 0.32 15.8 11.1 0.2 0.32 3.1
UB-4983 Lechlade: grave 136 41.8 −20.7 0.1 0.32 14.6 9.8 0.3 0.39 3.3UB-4502 Lechlade: grave 138 43.6 −20.1 0.12 0.32 15.9 8.9 0.26 0.36 3.2UB-4503 Lechlade: grave 148 42.0 −20.1 0.12 0.32 15.2 10.9 0.26 0.36 3.2UB-4982 Lechlade: grave 155 44.8 −20.7 0.1 0.32 15.7 9.7 0.3 0.39 3.3UB-4505 Lechlade: grave 172/1 44.1 −20.3 0.12 0.32 16.2 9.2 0.26 0.36 3.2UB-4506 Lechlade: grave 172/2 39.9 −20.0 0.12 0.32 14.5 8.9 0.26 0.36 3.2UB-4504 Lechlade: grave 179 43.9 −20.2 0.12 0.32 16.1 11.4 0.26 0.36 3.2UB-4981 Lechlade: grave 183 39.8 −20.3 0.1 0.32 14.1 8.0 0.3 0.39 3.3UB-4507 Lechlade: grave 187 44.0 −19.6 0.12 0.32 16.4 9.4 0.26 0.36 3.1UB-4549 Marina Drive: grave C7 33.2 −20.4 0.12 0.32 12.0 9.8 0.26 0.36 3.2UB-4553 Marina Drive: grave D10 30.3 −20.3 0.12 0.32 11.0 11.1 0.26 0.36 3.2UB-4550 Marina Drive: grave E1 36.1 −19.9 0.12 0.32 13.0 9.6 0.26 0.36 3.2UB-4551 Marina Drive: grave E2 54.8 −20.2 0.12 0.32 20.2 10.4 0.26 0.36 3.2UB-4552 Marina Drive: grave E3 45.2 −20.1 0.12 0.32 16.1 9.1 0.26 0.36 3.3UB-4554 Marina Drive: grave F2 48.2 −19.9 0.12 0.32 17.3 9.4 0.26 0.36 3.3UB-4889 Melbourn: SK1293, SG69 34.6 −20.2 0.12 0.32 12.4 9.4 0.26 0.36 3.3UB-4890 Melbourn: SK1307 SG75 36.2 −20.3 0.12 0.32 12.8 9.6 0.26 0.36 3.3UB-4886 Melbourn: SK1204 SG77 32.1 −20.0 0.12 0.32 11.4 8.9 0.26 0.36 3.3
UB-6345 Melbourn: SK1204 SG77 (replicate) 38.3 −19.5 0.2 0.36 13.8 8.7 0.2 0.32 3.2
UB-4885 Melbourn: InL1189 SG78 36.6 −20.1 0.12 0.32 13.3 9.5 0.26 0.36 3.2
UB-4885 Melbourn: InL1189 SG78 (replicate) 41.3 −20.5 0.1 0.32 15.3 9.2 0.2 0.32 3.2
UB-4884 Melbourn: SK1188 SG79 37.4 −20.4 0.12 0.32 13.3 10.4 0.26 0.36 3.3
UB-4884 Melbourn: SK1188 SG79 (replicate) 32.5 −20.4 0.1 0.32 11.9 10.2 0.2 0.32 3.2
UB-4882 Melbourn: SK1187 SG80 40.0 −20.1 0.12 0.32 14.3 9.3 0.26 0.36 3.3UB-4887 Melbourn: SK 1229 SG82 34.8 −20.3 0.12 0.32 12.7 9.3 0.26 0.36 3.2UB-4888 Melbourn: SK1271 SG89 38.5 −20.3 0.12 0.32 13.5 9.1 0.26 0.36 3.3UB-4883 Melbourn: SK1038, SG95 35.8 −20.1 0.12 0.32 13.0 10.3 0.26 0.36 3.2UB-6479 Mill Hill: grave 40 39.9 −19.5 0.2 0.36 14.7 8.8 0.2 0.32 3.2UB-4728 Mill Hill: grave 64 45.6 −19.5 0.12 0.32 17.0 9.4 0.26 0.36 3.1
THE HUMAN SKELETONS 111
laboratory number Site
Carbon and nitrogen analysis of rafter gelatin preparations
%C δ13C(‰)
error (±)1
total error (±)2
%n δ15n(‰)
error (±)1
total error (±)2
atomic C:n3
UB-4729 Mill Hill: grave 68 47.4 −19.5 0.12 0.32 17.5 10.2 0.26 0.36 3.2UB-4730 Mill Hill: grave 79 31.0 −19.1 0.12 0.32 11.4 10.1 0.26 0.36 3.2UB-4921 Mill Hill: grave 81 21.4 −20.3 0.1 0.32 7.7 9.3 0.3 0.39 3.3UB-4731 Mill Hill: grave 93 44.1 −19.8 0.12 0.32 16.2 10.1 0.26 0.36 3.2UB-4732 Mill Hill: grave 94 47.1 −19.7 0.12 0.32 17.0 9.9 0.26 0.36 3.2UB-4733 Mill Hill: grave 95 36.0 −19.9 0.12 0.32 13.3 10.0 0.26 0.36 3.1UB-4734 Mill Hill: grave 105C 60.7 −20.3 0.12 0.32 22.1 9.6 0.26 0.36 3.2UB-4961 St Peter’s Tip: grave 8 45.5 −19.5 0.1 0.32 16.5 9.5 0.3 0.39 3.2UB-4930 St Peter’s Tip: grave 42 44.5 −19.4 0.1 0.32 16.1 8.7 0.3 0.39 3.2
UB-6346 St Peter's Tip: grave 42 (replicate) 41.3 −19.0 0.2 0.36 15.0 8.1 0.2 0.32 3.2
UB-4925 St Peter’s Tip: grave 68 40.0 −19.5 0.1 0.32 14.5 10.2 0.3 0.39 3.2
UB-4925 St Peter’s Tip: grave 68 (replicate) 40.0 −19.3 0.1 0.32 15.0 10.3 0.2 0.32 3.1
UB-6032 St Peter's Tip: grave 73A 58.8 −20.9 0.1 0.32 21.2 8.5 0.3 0.39 3.2UB-4924 St Peter’s Tip: grave 113 41.4 −19.7 0.1 0.32 15.0 8.3 0.3 0.39 3.2
UB-6534 St Peter's Tip: grave 113 (replicate) 39.8 −19.4 0.2 0.36 14.4 8.8 0.2 0.32 3.2
UB-4929 St Peter’s Tip: grave 194 41.2 −20.0 0.1 0.32 15.0 9.9 0.3 0.39 3.2UB-4962 St Peter’s Tip: grave 196 44.9 −19.9 0.1 0.32 16.3 9.1 0.3 0.39 3.2UB-4963 St Peter’s Tip: grave 208 43.4 −19.8 0.1 0.32 15.4 9.5 0.3 0.39 3.3UB-4926 St Peter’s Tip: grave 212 37.7 −20.0 0.1 0.32 13.5 10.4 0.3 0.39 3.3UB-4928 St Peter’s Tip: grave 250 39.8 −20.1 0.1 0.32 14.5 9.5 0.3 0.39 3.2UB-4927 St Peter’s Tip: grave 263 38.6 −19.8 0.1 0.32 14.2 9.3 0.3 0.39 3.2
UB-4927 St Peter’s Tip: grave 263 (replicate) 40.7 −19.7 0.1 0.32 15.2 9.7 0.2 0.32 3.1
UB-4931 St Peter’s Tip: grave 318 39.7 −20.0 0.1 0.32 14.4 9.7 0.3 0.39 3.2UB-6478 St Peter's Tip: grave 360 40.2 −20.1 0.2 0.36 14.5 10.6 0.2 0.32 3.2UB-6033 West Heslerton: grave 113 59.9 −20.8 0.1 0.32 20.3 10.6 0.3 0.39 3.4
UB-4706 West Heslerton: 002BA 00536 47.1 −19.9 0.12 0.32 17.2 9.0 0.26 0.36 3.2
UB-4705 West Heslerton: 002BA 00606 27.2 −20.3 0.12 0.32 9.9 8.9 0.26 0.36 3.2
UB-4985 Westgarth Gardens: grave 11 42.9 −20.8 0.1 0.32 15.3 8.4 0.3 0.39 3.3
UB-4836 Westgarth Gardens: grave 27 44.5 −19.8 0.2 0.36 16.1 9.6 0.2 0.32 3.2
UB-4682 Westgarth Gardens: grave 66 41.1 −19.9 0.2 0.36 15.2 9.9 0.2 0.32 3.2
112
lab
orat
ory
num
ber
Site
/Con
text
res
idue
s per
100
0 ±
5%at
omic
C
:ng
ly/A
lag
ly/A
sp
hyd
roxy
prol
ine
(hyp
)A
spar
tic a
cid
(Asp
)g
luta
mic
aci
d(g
lu)
Prol
ine
(Pro
)g
lyci
ne(g
ly)
Ala
nine
(Ala
)A
rgin
ine
(Arg
)id
eal A
min
o A
cid
and
C:n
rat
ios
101
5175
116
317
113
492.
9–3.
62.
86.
2
UB
-520
8A
pple
Dow
n: g
rave
107
7458
9410
334
811
647
3.3
3.0
6.0
UB
-496
5A
pple
Dow
n: g
rave
117
8056
8912
931
811
247
3.2
2.8
5.7
UB
-634
4A
pple
D
own:
gr
ave
117
(rep
licat
e)10
450
9710
229
912
354
3.3
2.4
6.0
UB
-483
5A
pple
Dow
n: g
rave
134
7454
9010
035
911
450
3.2
3.1
6.6
UB
-497
5A
ston
Clin
ton:
gra
ve 1
279
6097
137
321
112
473.
32.
95.
3
UB
-473
5B
erin
sfiel
d: g
rave
22
8249
7012
035
212
244
3.2
2.9
7.2
UB
-473
6B
erin
sfiel
d: g
rave
28
8549
7111
734
612
245
3.3
2.8
7.1
UB
-473
9B
erin
sfiel
d: g
rave
134
/173
5595
101
332
119
473.
22.
86.
1
UB
-496
4C
odde
nham
: gra
ve 3
079
5490
130
321
117
483.
12.
75.
9
UB
-647
2D
over
Buc
klan
d: g
rave
222
7753
9510
533
911
750
3.2
2.9
6.3
UB
-647
3D
over
Buc
klan
d: g
rave
250
7556
9998
340
114
503.
23.
06.
0
UB
-647
4D
over
Buc
klan
d: g
rave
264
7549
8910
634
711
651
3.2
3.0
7.1
UB
-647
5D
over
Buc
klan
d: g
rave
323
7553
9510
132
211
652
3.2
2.8
6.0
UB
-647
6D
over
Buc
klan
d: g
rave
339
7549
8810
435
411
551
3.1
3.1
7.3
UB
-647
7D
over
Buc
klan
d: g
rave
414
7552
9410
533
311
751
3.2
2.8
6.5
UB
-492
3Ed
ix H
ill: g
rave
776
4483
106
364
114
513.
23.
28.
2
UB
-450
8Ed
ix H
ill: g
rave
12
4272
8912
331
413
757
3.2
2.3
4.4
UB
-470
9Ed
ix H
ill: g
rave
14
8251
8314
232
295
513.
33.
46.
3
UB
-450
9Ed
ix H
ill: g
rave
33
4860
8413
533
913
756
3.2
2.5
5.6
UB
-451
0Ed
ix H
ill: g
rave
48
3973
9411
830
912
956
3.3
2.4
4.2
Tab
le 4
.7
Amin
o-ac
id a
nd C
:N re
sults
for f
reez
e-dr
ied
gela
tin o
n 80
bur
ials
and
six
repl
icat
e sa
mpl
es. A
min
o-ac
id a
naly
sis o
f the
skel
eton
s com
pare
d w
ith e
xpec
ted
valu
es fo
r am
ino-
acid
pr
ofile
s in
non-
degr
aded
bon
e (v
an K
linke
n &
Moo
k 19
90,1
56)
, the
C:N
rang
e fo
r mod
ern
and
good
pre
hist
oric
col
lage
n (D
eNiro
198
5), a
nd id
eal G
ly/A
la a
nd G
ly/A
sp ra
tios
base
d up
on a
mod
ern
colla
gen
profi
le (v
an K
linke
n &
Moo
k 19
90, 1
56).
113
lab
orat
ory
num
ber
Site
/Con
text
res
idue
s per
100
0 ±
5%at
omic
C
:ng
ly/A
lag
ly/A
sp
hyd
roxy
prol
ine
(hyp
)A
spar
tic a
cid
(Asp
)g
luta
mic
aci
d(g
lu)
Prol
ine
(Pro
)g
lyci
ne(g
ly)
Ala
nine
(Ala
)A
rgin
ine
(Arg
)U
B-4
922
Edix
Hill
: gra
ve 4
8 (r
eplic
ate)
7357
8911
031
612
650
3.3
2.5
5.5
UB
-470
7Ed
ix H
ill: g
rave
79
8250
8111
631
312
254
3.2
2.6
6.3
UB
-470
8Ed
ix H
ill: g
rave
83
7958
8312
431
411
052
3.2
2.8
5.4
UB
-451
1Ed
ix H
ill: g
rave
90
4761
8313
533
413
753
3.3
2.4
5.5
UB
-451
2Ed
ix H
ill: g
rave
91
4760
8313
434
113
653
3.3
2.5
5.7
UB
-497
6Fo
rd, L
aver
stoc
k: b
arro
w 2
7753
9313
530
811
748
3.3
2.6
5.8
UB
-472
7G
ally
Hill
s: p
rimar
y bu
rial
7771
9111
130
611
653
3.3
2.6
4.3
UB
-492
0G
ally
H
ills:
pr
imar
y bu
rial
(rep
licat
e)77
5295
101
362
114
423.
33.
27.
0
UB
-634
7La
kenh
eath
: ER
L 10
4 42
2210
051
9510
031
211
951
3.3
2.6
6.1
UB
-450
1Le
chla
de: g
rave
14
6547
8813
931
012
354
3.2
2.5
6.7
UB
-498
4Le
chla
de: g
rave
18
7661
9512
830
611
247
3.3
2.7
5.1
UB
-468
3Le
chla
de: g
rave
40
7649
8999
355
118
503.
23.
07.
2
UB
-450
2Le
chla
de: g
rave
138
6455
9113
330
812
153
3.2
2.5
5.6
UB
-450
3Le
chla
de: g
rave
148
6647
8813
630
912
355
3.2
2.5
6.6
UB
-498
2Le
chla
de: g
rave
155
8238
9513
632
211
448
3.3
2.8
8.5
UB
-450
5Le
chla
de: g
rave
172
/181
5070
114
346
121
453.
22.
96.
9
UB
-450
6Le
chla
de: g
rave
172
/286
4872
119
351
121
443.
22.
97.
3
UB
-450
4Le
chla
de: g
rave
179
9048
7111
434
711
846
3.2
2.9
7.2
UB
-498
1Le
chla
de: g
rave
183
8058
9313
131
111
548
3.3
2.7
5.4
UB
-450
7Le
chla
de: g
rave
187
8147
6911
734
712
245
3.1
2.8
7.4
UB
-454
9M
arin
a D
rive:
gra
ve C
753
6487
117
346
123
523.
22.
85.
4
UB
-455
3M
arin
a D
rive:
gra
ve D
1056
5383
124
354
119
513.
23.
06.
7
UB
-455
0M
arin
a D
rive:
gra
ve E
160
6610
013
339
212
759
3.2
3.1
5.9
UB
-455
1M
arin
a D
rive:
gra
ve E
256
6892
118
339
122
553.
22.
85.
0
114
lab
orat
ory
num
ber
Site
/Con
text
res
idue
s per
100
0 ±
5%at
omic
C
:ng
ly/A
lag
ly/A
sp
hyd
roxy
prol
ine
(hyp
)A
spar
tic a
cid
(Asp
)g
luta
mic
aci
d(g
lu)
Prol
ine
(Pro
)g
lyci
ne(g
ly)
Ala
nine
(Ala
)A
rgin
ine
(Arg
)U
B-4
552
Mar
ina
Driv
e: g
rave
E3
5359
9011
334
812
151
3.3
2.9
5.9
UB
-455
4M
arin
a D
rive:
gra
ve F
255
6694
114
349
121
533.
32.
95.
3
UB
-488
9M
elbo
urn:
SK
1293
, SG
6975
5387
121
336
129
503.
32.
66.
4
UB
-489
0M
elbo
urn:
SK
1307
SG
7573
5594
117
325
129
503.
32.
55.
9
UB
-488
6M
elbo
urn:
SK
1204
SG
7772
5590
118
316
125
513.
32.
55.
8
UB
-634
5M
elbo
urn:
SK
1204
SG
77
(rep
licat
e)98
5092
9932
411
651
3.2
2.8
6.5
UB
-488
5M
elbo
urn:
InL1
189
SG78
7350
8912
232
012
951
3.2
2.5
6.4
UB
-488
4M
elbo
urn:
SK
1188
SG
7976
4590
118
327
130
533.
32.
57.
2
UB
-488
2M
elbo
urn:
SK
1187
SG
8081
4390
122
314
128
493.
32.
47.
3
UB
-488
7M
elbo
urn:
SK
122
9 SG
8274
4387
120
332
130
513.
22.
57.
7
UB
-488
8M
elbo
urn:
SK
1271
SG
8972
5286
112
326
128
523.
32.
56.
2
UB
-488
3M
elbo
urn:
SK
1038
, SG
9576
4989
118
328
130
523.
22.
56.
7
UB
-647
9M
ill H
ill: g
rave
40
7255
9011
432
512
949
3.2
2.5
5.9
UB
-472
8M
ill H
ill: g
rave
64
7669
9011
231
011
053
3.1
2.8
4.5
UB
-472
9M
ill H
ill: g
rave
68
7965
8711
531
811
252
3.2
2.8
4.9
UB
-473
0M
ill H
ill: g
rave
79
8061
8611
732
011
253
3.2
2.8
5.2
UB
-492
1M
ill H
ill: g
rave
81
7351
9111
532
212
747
3.3
2.5
6.3
UB
-473
1M
ill H
ill: g
rave
93
8162
8812
031
810
653
3.2
3.0
5.1
UB
-473
2M
ill H
ill: g
rave
94
8161
8512
131
811
151
3.2
2.9
5.2
UB
-473
3M
ill H
ill: g
rave
95
7861
8812
131
211
051
3.1
2.8
5.1
UB
-473
4M
ill H
ill: g
rave
105
C81
6386
121
314
112
533.
22.
85.
0
UB
-496
1St
Pet
er’s
Tip
: gra
ve 8
8358
9313
031
411
545
3.2
2.7
5.5
UB
-493
0St
Pet
er’s
Tip
: gra
ve 4
269
5397
119
339
126
513.
22.
76.
4
UB
-634
6St
Pe
ter's
Ti
p:
grav
e 42
(r
eplic
ate
)11
953
112
113
218
141
564.
21.
64.
1
115
lab
orat
ory
num
ber
Site
/Con
text
res
idue
s per
100
0 ±
5%at
omic
C
:ng
ly/A
lag
ly/A
sp
hyd
roxy
prol
ine
(hyp
)A
spar
tic a
cid
(Asp
)g
luta
mic
aci
d(g
lu)
Prol
ine
(Pro
)g
lyci
ne(g
ly)
Ala
nine
(Ala
)A
rgin
ine
(Arg
)U
B-4
925
St P
eter
’s T
ip: g
rave
68
6051
9811
834
612
552
3.2
2.8
6.8
UB
-603
2St
Pet
er's
Tip:
gra
ve 7
3A99
5190
102
328
121
503.
22.
76.
4
UB
-492
4St
Pet
er’s
Tip
: gra
ve 1
1371
4698
120
344
128
513.
22.
77.
4
UB
-653
4St
Pe
ter's
Ti
p:
grav
e 11
3 (r
eplic
ate
)75
4793
103
358
119
503.
23.
07.
6
UB
-492
9St
Pet
er’s
Tip
: gra
ve 1
9470
5097
116
345
125
523.
22.
87.
0
UB
-496
2St
Pet
er’s
Tip
: gra
ve 1
9677
5692
135
311
120
473.
22.
65.
6
UB
-496
3St
Pet
er’s
Tip
: gra
ve 2
0882
6997
137
327
123
483.
32.
74.
7
UB
-492
6St
Pet
er’s
Tip
: gra
ve 2
1270
4998
118
341
126
533.
32.
77.
0
UB
-492
8St
Pet
er’s
Tip
: gra
ve 2
5069
5399
119
334
124
523.
22.
76.
2
UB
-492
7St
Pet
er’s
Tip
: gra
ve 2
6370
5199
118
341
125
523.
22.
76.
7
UB
-493
1St
Pet
er’s
Tip
: gra
ve 3
1869
5398
120
340
124
493.
22.
76.
4
UB
-647
8St
Pet
er's
Tip:
gra
ve 3
6067
5598
122
336
125
513.
22.
76.
1
UB
-603
3W
est H
esle
rton:
gra
ve 1
1310
249
9210
434
212
652
3.4
2.7
7.0
UB
-470
6W
est H
esle
rton:
002
BA
005
3680
5683
125
314
111
493.
22.
85.
7
UB
-470
5W
est H
esle
rton:
002
BA
006
0683
5180
133
308
109
523.
22.
86.
0
UB
-498
5W
estg
arth
Gar
dens
: gra
ve 1
179
5693
131
321
116
483.
32.
85.
7
UB
-483
6W
estg
arth
Gar
dens
: gra
ve 2
771
5396
119
340
123
523.
22.
86.
4
UB
-468
2W
estg
arth
Gar
dens
: gra
ve 6
673
5199
118
342
125
533.
22.
76.
7
ANGLO-SAxON GRAVES AND GRAVE GOODS116
Hill (UB-6479) and St Peter’s Tip (UB-4926, UB-4927, UB-4930, UB-4931, UB-4963 and UB-6478). Thirty-five of the 86 burials fall below the ideal Gly/Asp ratio of 6.2 for well-preserved collagen based on van Klinken and Mook (1990). Low values for hydroxyproline can suggest protein deterioration and exclude exogenous contaminants, which would have increased hydroxyproline values relative to the expected amino-acid profiles for well-preserved collagen.
In contrast, C:N ratios for these and other burials are uniformly within the expected range for well-preserved collagen (2.9–3.6; DeNiro 1985), with the exception of CasD088 (UB-6042) with a C:N ratio of 4.2. Despite the variability of the collagen amino-acid profiles discussed above, C:N ratios may indeed be the best indicator of protein preservation. A similar finding was observed by Brock et al (2007) who determined that, from a number of pre-screening criteria tested, %N, %C and C:N ratio may return the best correlation to protein preservation. Amino-acid profiles would, however, indicate the presence of exogenous contaminants in higher hydroxyproline and glycine values. Additionally, while significant variation in amino-acids may alter total carbon and nitrogen stable-isotope values, degraded protein or the possible presence of exogenous amino-acids may not necessarily prohibit obtaining reliable radiocarbon dates for individuals if the exogenous materials are of similar age to the burials.
4.3 StABle iSotoPeS And diet
The elements that make up bone collagen are sourced from diet. These foods, and their associated radiocarbon and stable isotope signatures, may come from a variety of reservoirs, not all of which may be in equilibrium with the terrestrial biosphere (Tauber 1984; Lanting & van der Plicht 1998; Arneborg et al 1999). If the reservoir source has an inherent radiocarbon offset — for example, foods coming from depleted reservoirs such as the marine environment, or the ‘hardwater effects’ passed up the food-chain in certain terrestrial aquatic environments — then the bone will take on some proportion of radiocarbon that is not in equilibrium with the terrestrial environment. These radiocarbon offsets affect the calibration of our radiocarbon ages, as the basis of our project-specific calibration curve is tree-ring dated wood which fixed carbon from the atmosphere by photosynthesis (above, Section 2.2). If the dated Anglo-Saxons consumed foods which derive from a mix of radiocarbon reservoirs, then this project-specific calibration curve may not provide the most appropriate method of calibration for these individuals. Identification of dietary offsets relies mainly upon stable-isotope analysis of either the mineral or protein (collagen) fraction of the bone.
Our main concern for the Anglo-Saxon chronology is the potential for diet-induced radiocarbon offsets, which must be assessed for accurate calibration of the radiocarbon ages on the Anglo-Saxon human bone samples. The most common isotopes used in stable-isotope analysis of diet are δ13C and δ15N. Isotope-value ranges for foods from a particular source environment can be assumed from known global averages (cf. DeNiro & Epstein 1978; 1981) but one can also use region-specific faunal databases (Britton et al 2008; Beavan Athfield et al 2008; Jay & Richards 2006).
The relationship between stable isotopes and diet is complex and as yet incompletely understood (Bayliss et al 2004b; Petchey & Green 2005; Hedges & Reynard 2007) but better understanding of regional isotope affects and possible ways to undertake dietary analysis offer ways to improve these calculations (Beavan Athfield et al 2008).
Animal feeding experiments (e.g. Ambrose & Norr 1993) indicate that δ15N of bone protein tracks dietary protein sources, rather than the whole diet, and essential amino-acids from these protein sources influence a consumer’s overall δ15N values. Nitrogen isotopes (δ15N) generally reflect the biome source (i.e. terrestrial protein versus marine foods) and the trophic level (e.g. plant versus animal food sources, due to the increasingly enriched values of δ15N as one progresses up the food chain: the ‘trophic level effect’: DeNiro & Epstein 1981). This means, for example, that terrestrial herbivores have higher δ15N values than the terrestrial plants upon which they feed, and that in turn terrestrial carnivores have higher δ15N values than terrestrial herbivores (marine and freshwater organisms tend to have more elevated δ15N values still as food chains in aquatic environments tend to be longer). Some non-dietary factors, such as nutritional stress, can also influence δ15N values (Fuller et al 2005).
Diets that are wholly terrestrial generally result in human value mean of about −20‰ for δ13C and δ15N of +5‰ to +12‰, depending upon the type and amount of animal protein in the diet. Diets of marine fish and shellfish typically produce values in human consumers of δ13C −12‰ and δ15N of +12‰ to +22‰ (DeNiro & Epstein 1978; 1981; Schoeninger et al 1983). Diets which are not derived solely from one biome (e.g. a combination of terrestrial and marine food sources) result in consumer isotopic signatures which are determined by the proportion of each type of food source in the diet; whether this relationship is linear, and what appropriate endpoints are for foods from different environments, is a point of debate (Ambrose & Norr 1993; Müldner & Richards 2005; Hedges & Reynard 2007). The δ13C of consumer collagen principally reflects the protein portion of the diet but the spacing between δ13C in collagen and diet may vary, depending on whether the carbon isotope value of the protein equals that of the whole diet
THE HUMAN SKELETONS 117
(Harrison & Katzenburg 2003).Estimating the amount of fish consumed by the
Anglo-Saxon populations from different biomes is important in assessing how far each 14C reservoir (i.e. terrestrial, marine, and freshwater) may affect the radiocarbon age in our study (see esp. Section 2.2.4). The marine environment contains less radiocarbon than the contemporary atmosphere, and so bone from individuals who derive some of their diet from this reservoir — such as humans eating marine fish or shellfish — will have an apparently older radiocarbon age than a contemporary sample derived purely from the atmosphere. The scale of the marine offset varies regionally, but is in the order of 400 years for English coastal waters (Harkness 1983). Some terrestrial aquatic sources may have the potential to introduce radiocarbon offsets because some of the carbon they contain derives from dissolved carbonate (Ascough et al 2007; Culleton 2006). This carbon component would be radiocarbon ‘dead’ with the potential for imparting significant and anomalously old ages as the dissolved inorganic carbon is synthesized by aquatic plants and becomes part of the aquatic food-chain. However, there is no information on freshwater radiocarbon offsets for fish in the overall region in which these sites are situated.
To estimate the potential diet mixtures of the Anglo-Saxon populations, we undertake an analysis of the stable-isotope data for the skeletons using two dietary mixing model methods. We first consider the overall position of the Anglo-Saxon
population relative to likely sources of food. We then discuss how observed variation within or between sites could be indicative of a status-based dietary difference, as, for instance, previously seen in a 5th- to 6th-century Anglo-Saxon cemetery at Berinsfield, Oxon (Privat et al 2002) and status differences as determined by burial-type in a Roman-period cemetery at Poundbury, Dorset (Richards et al 1998).
4.3.1 dietary analysis
Details of all the stable-isotope analysis performed on freeze-dried gelatin from the 96 skeletons and 14 replicate samples are provided in Table 4.6 and illustrated in Figure 4.2. A full report on the statistical treatment of the entire data-set (including the weighted means of the 14 replicate analyses) can be found in Beavan et al (2011).
Over the entire Anglo-Saxon data-set, stable isotopic data indicate that protein sources in human diets at all sites studied were overwhelmingly terrestrial. The mean isotopic values for the Anglo-Saxons (δ 13C −20.15‰, SD 0.4, and δ 15N +9.66‰, SD 0.75) are not significantly different from the −20.0‰ and +10.3‰ reported by Müldner and Richards (2007) for late 7th- and 8th-century human remains from York.
We first present a graphic representation of where Anglo-Saxon isotopes sit in relation to regional dietary sources. We have constructed diet-to-collagen value boxes based upon reported
7
8
9
10
11
12
-22 -21 -20 -19 -18
δ3C ‰ (VPDB, )
δ15N
‰ (A
IR)
Apple Down
Aston Clinton
Berinsfield
Buttermarket
Castledyke South
Coddenham
Dover Buckland
Edix Hill
Ford, Laverstock
Gally Hills
Lechlade
Marina Drive
Melbourn
Mill Hill
St Peter's Tip
West Heslerton
Westgarth Gardens
Lakenheath
Figure 4.2δ13C and δ15N values for all Anglo-Saxon skeletons, distinguished by site.
ANGLO-SAxON GRAVES AND GRAVE GOODS118
δ15N and δ13C values for archaeological fauna from various sites in England (Müldner 2005; Birchall 2002; Richards et al 2006; Jay & Richards 2006; Privat et al 2002; M. Richards 2000; O’Connell & Lawler 2009; Müldner & Richards 2005; 2007) and freshwater, anadromous, and marine fish (Müldner 2005; Richards et al 2006). For a trophic enrichment factor — the enrichment of δ15N and δ13C between diet source bone-values and consumer bone-values — we take into consideration previous discussions concerning the per mille enrichment factor for both carbon and nitrogen (van der Merwe & Vogel 1978; Peterson & Fry 1987; Hedges & Reynard 2007). Here we use a trophic enrichment factor of 1‰ for 13C, based on enrichment in herbivores (Jay & Richards 2006) and 4‰ for 15N, based on evidence for trophic enrichment between nursing women and breastfeeding children (Richards et al 2002).
In Figure 4.3, the ‘terrestrial vegetation’ box uses minimum and maximum cattle values from data cited above as a proxy for where human collagen values would be with a wholly vegetarian diet. We similarly use minimum and maximum values from these sources for the ‘terrestrial animal protein’ (domestic fowl, sheep, and cattle), ‘freshwater fish’, ‘salmonid’ and ‘eel’ boxed ranges, but to approximate where human consumers would sit in relation to these food sources we have added
the trophic enrichment factors of 1‰ for δ13C and 4‰ for δ15N. Individual Anglo-Saxon isotope signatures were then plotted against the dietary isotope-range boxes. The plot indicates that all skeletons are enriched in δ13C and δ15N to points well above pure ‘terrestrial vegetation’ ranges, and the human values sit well within the range of ‘terrestrial animal protein’ sources yet they do not necessarily exclude overlaps of freshwater fish and eel isotope ranges.
We next present two methods of calculating the proportion of these food sources for the Anglo-Saxons. The first method to estimate dietary proportions is ISOSOURCE, a mixing model which derives probable diet-source partitions from up to seven diet sources. While the mixing model was first developed for source-proportional diet determinations in ecology (see Phillips et al 2005), it has also previously been successfully applied in archaeology (Newsome et al 2004; Beavan Athfield et al 2008).
ISOSOURCE
ISOSOURCE produces a set of solutions derived from the combination of source proportions that satisfy an isotopic mass balance mixing model, with descriptive statistics to characterize the
0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
-28 -26 -24 -22 -20 -18 -16 -14 -12 -10
δC ‰ (VPDB)
δ15N
‰ (A
IR)
Apple Down
Aston Clinton
Berinsfield
Buttermarket
Castledyke South
Coddenham
Dover Buckland
Edix Hill
Ford, Laverstock
Gally Hills
Lechlade
Marina Drive
Melbourn
Mill Hill
St Peter's Tip
West Heslerton
Westgarth Gardens
Lakenheath
freshwater fish
eel
terrestrial animal protein
salmonid
marine
terrestrial vegetation
Figure 4.3Stable-isotope values for all the Anglo-Saxon skeletons analysed fit well within parameters for a mainly terrestrial animal protein diet. The boxes are created from graphing minimum and maximum stable-iso-tope values of δ13C and δ15N for ancient food sources (vegetarian, terrestrial animal protein, eel, fresh-water fish and marine fish) from published values for British sites (Richards et al 2006; Jay & Richards 2006; Müldner 2005; Müldner & Richards 2005; Privat et al 2002; Birchall 2002; O'Connell & Lawler 2009). Error bars on boxes from Standard Deviation on isotopes for food-types; errors on Anglo-Saxon skeletons from analysis error for element.
119
lab
orat
ory
num
ber
Site
iSo
Sou
rC
em
ays (
1997
) (δ
13C
)
# so
lutio
ns%
veg
etat
ion
%
terr
estr
ial
anim
al
prot
ein
%
eel
% fr
eshw
ater
fis
h%
salm
onid
s%
mar
ine
fish
%
mar
ine
fish
UB
-520
8A
pple
Dow
n: g
rave
107
133,
813
65.4
±6.0
12.2
±9.8
4.0±
3.2
3.5±
3.0
8.2±
5.4
6.6±
4.3
12.6
±7.5
UB
-496
5A
pple
Dow
n: g
rave
117
167,
613
63.6
±6.6
13.9
±11.
04.
7±3.
74.
2±3.
47.
5±5.
06.
0±4.
011
.6±7
.4U
B-6
344
UB
-483
5A
pple
Dow
n: g
rave
134
282,
543
56.8
±7.7
16.8
±13.
15.
8±4.
45.
1±4.
18.
6±5.
66.
9±4.
513
.7±7
.5
UB
-497
5A
ston
Clin
ton:
gra
ve 1
239
0,99
647
.8±1
0.0
21.8
±17.
312
.6±7
.89.
7±7.
14.
6±3.
63.
6±2.
97.
4±7.
4
UB
-473
5B
erin
sfiel
d: g
rave
22
844,
766
37.4
±11.
626
.3±2
0.3
10.1
±7.1
8.7±
6.5
9.7±
6.4
7.8±
5.1
16.8
±7.4
UB
-473
6B
erin
sfiel
d: g
rave
28
513,
801
47.3
±9.6
21.3
±16.
57.
6±5.
66.
7±5.
19.
5±6.
27.
6±5
15.8
±7.4
UB
-473
9B
erin
sfiel
d: g
rave
134
/132
3,17
155
.0±8
.418
.4±1
4.3
6.6±
4.9
5.8±
4.5
7.9±
5.2
6.3±
4.2
12.6
±7.4
UB
-407
7B
utte
rmar
ket:
grav
e 42
7595
8,94
532
.7±1
2.3
28.5
±21.
613
.7±8
.911
.4±8
.17.
6±5.
36.
1±4.
313
.7±7
.4
UB
-404
6B
utte
rmar
ket:
grav
e 43
4449
0,62
448
.1±9
.621
.3±1
6.5
7.8±
5.7
6.8±
5.2
8.9±
5.8
7.1±
4.7
14.7
±7.4
UB
-603
6C
astle
dyke
Sou
th: g
rave
13
580,
351
28.3
±10.
222
.2±1
7.1
26.3
±12.
616
.0±1
1.7
4.0±
3.4
3.1±
2.7
6.3±
7.4
UB
-604
0C
astle
dyke
Sou
th: g
rave
53
306,
978
41.9
±8.9
18.6
±15.
021
.1±1
0.3
12.7
±9.4
3.2±
2.8
2.5±
2.2
4.2±
7.4
UB
-604
2C
astle
dyke
Sou
th: g
rave
88
729,
487
39.0
±11.
926
.6±2
0.8
11.7
±7.9
9.8±
7.2
7.2±
5.1
5.7±
4.1
12.6
±7.4
UB
-603
9C
astle
dyke
Sou
th: g
rave
94,
sk
elet
on 1
452
661,
828
39.3
±11.
726
.1±2
0.5
13.2
±8.5
10.7
±7.7
6.0±
4.4
4.8±
3.6
10.5
±7.4
UB
-603
5C
astle
dyke
Sou
th: g
rave
96
216,
055
35.3
±8.1
15.8
±12.
931
±11.
613
.3±1
0.6
2.6±
2.4
2.0±
1.9
2.1±
7.4
Tab
le 4
.8IS
OSO
URC
E is
otop
ic m
ass b
alan
ce m
ixin
g m
odel
(Phi
llips
& G
regg
200
3) a
nd li
near
inte
rpol
atio
n of
the
mar
ine
com
pone
nt o
f die
t fro
m δ
13C
onl
y, fo
llow
ing
May
s (19
97).
Thou
sand
s of e
quat
ions
are
run
by IS
OSO
URC
E fo
r eac
h is
otop
e to
reso
lve
for t
he m
ean
and
stan
dard
dev
iatio
n of
pos
sibl
e so
lutio
ns fo
r eac
h se
t of
food
sour
ces p
rovi
ded.
Site
s whe
re re
plic
ate
isot
opic
ana
lysi
s was
com
plet
ed o
n th
e sk
elet
on sh
ow a
ll U
B id
entifi
catio
n nu
mbe
rs a
ssoc
iate
d w
ith th
e sk
elet
on. F
or
skel
eton
s with
repl
icat
e an
alys
is, t
he m
ean
isot
opic
val
ues w
ere
used
in th
e ca
lcul
atio
ns.
120
lab
orat
ory
num
ber
Site
iSo
Sou
rC
em
ays (
1997
) (δ
13C
)
# so
lutio
ns%
veg
etat
ion
%
terr
estr
ial
anim
al
prot
ein
%
eel
% fr
eshw
ater
fis
h%
salm
onid
s%
mar
ine
fish
%
mar
ine
fish
UB
-604
1C
astle
dyke
Sou
th: g
rave
182
356,
068
53.1
±9.2
20.2
±15.
88.
1±5.
76.
9±5.
26.
5±4.
55.
1±3.
610
.5±7
.4
UB
-603
8C
astle
dyke
Sou
th: g
rave
183
347,
127
44.1
±9.4
20.0
±16.
017
.7±9
.511
.9±8
.63.
6±3.
02.
8±2.
45.
3±7.
4
UB
-496
4C
odde
nham
: gra
ve 3
075
8,12
536
.9±1
2.1
27.1
±21.
113
.4±8
.610
.9±7
.96.
5±4.
75.
2±3.
811
.6±7
.4
UB
-647
2 D
over
Buc
klan
d: g
rave
222
592,
357
44.7
±10.
222
.8±1
7.6
8.4±
6.0
7.3±
5.5
9.4±
6.1
7.5±
4.9
15.8
±7.4
UB
-647
2
UB
-647
3D
over
Buc
klan
d: g
rave
250
884,
452
36.5
±11.
626
.4±2
0.2
9.9±
7.0
8.6±
6.4
10.3
±6.7
8.3±
5.4
17.9
±7.5
UB
-647
4D
over
Buc
klan
d: g
rave
264
456,
100
49.0
±8.9
19.7
±15.
26.
8±5.
16.
0±4.
710
.2±6
.68.
2±5.
316
.8±7
.4U
B-6
474
UB
-647
5D
over
Buc
klan
d: g
rave
323
599,
025
44.3
±9.8
21.9
±16.
97.
7±5.
76.
8±5.
210
.7±6
.88.
6±5.
517
.9±7
.5
UB
-647
6D
over
Buc
klan
d: g
rave
339
227,
749
56.8
±6.5
13.6
±10.
74.
3±3.
53.
9±3.
211
.9±7
.49.
7±6.
018
.9±7
.5
UB
-495
8D
over
Buc
klan
d: g
rave
375
287,
342
54.9
±7.2
15.4
±12.
05.
0±3.
94.
5±3.
611
.2±7
.09.
0±5.
717
.9±7
.5
UB
-495
9D
over
Buc
klan
d: g
rave
39
1A63
0,12
742
.5±1
1.2
25.0
±19.
510
.4±7
.28.
8±6.
57.
3±5.
15.
8±4.
112
.6±7
.5
UB
-647
7D
over
Buc
klan
d: g
rave
414
431,
610
49.6
±8.5
18.7
±14.
46.
3±4.
85.
6±4.
410
.9±6
.98.
8±5.
617
.9±7
.5
UB
-492
3 Ed
ix H
ill: g
rave
785
7,68
734
.7±1
2.3
27.9
±21.
513
.5±8
.811
.1±8
.07.
1±5.
05.
6±4.
012
.6±7
.4U
B-4
923
UB
-450
8Ed
ix H
ill: g
rave
12
802,
855
38.3
±11.
626
.3±2
0.3
10.2
±7.2
8.8±
6.6
9.1±
6.0
7.3±
4.9
15.8
±7.4
UB
-470
9Ed
ix H
ill: g
rave
14
786,
124
35.9
±12.
127
.4±2
1.2
13.8
±8.9
11.2
±8.1
6.5±
4.7
5.2±
3.8
11.6
±7.4
UB
-450
9Ed
ix H
ill: g
rave
33
639,
219
42.7
±11.
024
.7±1
9.2
9.8±
6.9
8.4±
6.3
8.0±
5.4
6.3±
4.4
13.7
±7.4
UB
-451
0Ed
ix H
ill: g
rave
48
481,
426
46.1
±10.
623
.4±1
8.4
11.3
±7.4
9.2±
6.7
5.6±
4.2
4.4±
3.3
9.5±
7.4
UB
-492
2
UB
-470
7Ed
ix H
ill: g
rave
79
323,
171
55.0
±8.4
18.4
±14.
36.
6±4.
95.
8±4.
57.
9±5.
26.
3±4.
212
.6±7
.4
121
UB
-470
8Ed
ix H
ill: g
rave
83
1,02
1,40
430
.8±1
2.2
28.6
±21.
514
.8±9
.412
.2±8
.57.
6±5.
36.
1±4.
313
.7±7
.4
UB
-451
1 Ed
ix H
ill: g
rave
90
611,
989
42.5
±11.
325
.1±1
9.7
11.1
±7.5
9.3±
6.8
6.7±
4.8
5.3±
3.8
11.6
±7.4
UB
-451
1
UB
-451
2Ed
ix H
ill: g
rave
91
958,
945
32.7
±12.
328
.5±2
1.6
13.7
±8.9
11.4
±8.1
7.6±
5.3
6.1±
4.3
13.7
±7.4
UB
-497
6Fo
rd, L
aver
stoc
k: b
arro
w 2
381,
038
52.3
±9.1
20.2
±15.
77.
6±5.
56.
6±5.
07.
4±5.
05.
9±4.
011
.6±7
.4
UB
-492
0G
ally
Hill
s (po
st P
VA
extra
ctio
n)85
7,17
435
.7±1
2.2
27.8
±21.
412
.3±8
.210
.3±7
.57.
7±5.
46.
1±4.
313
.7±7
.4
UB
-634
7La
kenh
eath
: ER
L 10
4 42
2235
6,35
052
.2±7
.817
.1±1
3.2
5.7±
4.4
5.1±
4.0
11.1
±7.0
8.9±
5.6
17.9
±7.5
UB
-450
1Le
chla
de: g
rave
14
752,
418
39.9
±11.
125
.0±1
9.3
9.3±
6.6
8.1±
6.1
9.8±
6.4
7.9±
5.2
16.8
±7.4
UB
-498
4 Le
chla
de: g
rave
18
308,
338
55.6
±8.6
19.0
±5.9
7.3±
5.3
6.3±
4.8
6.6±
4.5
5.2±
3.7
10.5
±7.4
UB
-498
4U
B-4
683
Lech
lade
: gra
ve 4
01,
026,
609
31.7
±12.
328
.7±2
1.6
13.5
±8.8
11.3
±8.1
8.2±
5.6
6.6±
4.5
14.7
±7.4
UB
-468
3
UB
-498
3Le
chla
de: g
rave
136
503,
923
43.2
±10.
823
.7±1
8.8
13.6
±8.4
10.6
±7.7
5.0±
3.8
3.9±
3.1
8.4±
7.4
UB
-450
2Le
chla
de: g
rave
138
324,
071
54.6
±8.0
17.5
±13.
66.
0±4.
65.
3±4.
29.
2±5.
97.
3±4.
814
.7±7
.4
UB
-450
3Le
chla
de: g
rave
148
1,06
0,75
330
.8±1
2.3
28.7
±21.
514
.0±9
.011
.7±8
.38.
2±5.
66.
6±4.
514
.7±7
.4
UB
-450
5Le
chla
de: g
rave
172
/143
9,11
849
.9±9
.621
.2±1
6.5
8.1±
5.8
7.0±
5.3
7.6±
5.2
6.1±
4.1
12.6
±7.4
UB
-450
6Le
chla
de: g
rave
172
/230
4,19
855
.1±7
.616
.6±1
2.9
5.6±
4.3
5.0±
3.9
9.8±
6.3
7.9±
5.1
15.8
±7.4
UB
-450
4Le
chla
de: g
rave
179
1,15
5,88
026
.7±1
1.7
28.0
±20.
417
.4±1
0.6
14.2
±9.6
7.6±
5.3
6.1±
4.3
13.7
±7.4
UB
-498
1Le
chla
de: g
rave
183
133,
813
65.4
±6.0
12.2
±9.8
4.0±
3.2
3.5±
3.0
8.2±
5.4
6.6±
4.3
12.6
±7.4
UB
-450
7Le
chla
de: g
rave
187
371,
850
50.6
±7.7
16.7
±12.
95.
4±4.
24.
9±3.
912
.3±7
.710
.0±6
.220
.0±7
.4
lab
orat
ory
num
ber
Site
iSo
Sou
rC
em
ays (
1997
) (δ
13C
)
# so
lutio
ns%
veg
etat
ion
%
terr
estr
ial
anim
al
prot
ein
%
eel
% fr
eshw
ater
fis
h%
salm
onid
s%
mar
ine
fish
%
mar
ine
fish
122
UB
-454
9M
arin
a D
rive:
gra
ve C
761
1,98
942
.5±1
1.3
25.1
±19.
711
.1±7
.59.
3±6.
86.
7±4.
85.
3±3.
811
.6±7
.4
UB
-455
3M
arin
a D
rive:
gra
ve D
101,
024,
468
29.2
±12.
028
.2±2
1.0
16.6
±10.
113
.4±9
.37±
5.0
5.6±
4.0
12.6
±7.4
UB
-455
0M
arin
a D
rive:
gra
ve E
153
5,31
546
.4±9
.521
.2±1
6.4
7.5±
5.5
6.6±
5.1
10.1
±6.5
8.1±
5.2
16.8
±7.4
UB
-455
1M
arin
a D
rive:
gra
ve E
285
7,17
435
.7±1
2.2
27.8
±21.
412
.3±8
.210
.3±7
.57.
7±5.
46.
1±4.
313
.7±7
.4
UB
-455
2M
arin
a D
rive:
gra
ve E
338
7,65
152
.0±8
.619
.1±1
4.8
6.7±
5.0
5.9±
4.6
9.0±
5.9
7.2±
4.7
14.7
±7.4
UB
-455
4M
arin
a D
rive:
gra
ve F
245
6,10
049
.0±8
.919
.7±1
5.2
6.8±
5.1
6.0±
4.7
10.2
±6.6
8.2±
5.3
16.8
±7.4
UB
-488
9M
elbo
urn:
SK
1293
, SG
6949
9,12
647
.7±9
.922
.0±1
7.1
8.3±
6.0
7.2±
5.5
8.2±
5.5
6.5±
4.4
13.7
±7.4
UB
-489
0M
elbo
urn:
SK
1307
SG
7556
4,75
245
.0±1
0.7
23.8
±18.
69.
6±6.
78.
2±6.
17.
4±5.
15.
9±4.
112
.6±7
.4
UB
-488
6M
elbo
urn:
SK
1204
SG
7722
4,94
757
.5±6
.613
.8±1
0.8
4.4±
3.5
3.9±
3.3
11.3
±7.1
9.1±
5.7
17.9
±7.4
UB
-634
5
UB
-488
5 M
elbo
urn:
InL1
189
SG78
469,
759
48.7
±9.9
21.9
±17.
18.
5±6.
07.
3±5.
57.
6±5.
16.
0±4.
112
.6±7
.4U
B-4
885
UB
-488
4 M
elbo
urn:
SK
1188
SG
7975
8,12
536
.9±1
2.1
27.1
±21.
113
.4±8
.610
.9±7
.96.
5±4.
75.
2±3.
811
.6±7
.4U
B-4
884
UB
-488
2M
elbo
urn:
SK
1187
SG
8045
5,34
649
.4±9
.320
.6±1
5.9
7.4±
5.4
6.5±
5.0
8.9±
5.8
7.1±
4.7
14.7
±7.4
UB
-488
7M
elbo
urn:
SK
122
9 SG
8246
9,75
948
.7±9
.921
.9±1
7.1
8.5±
6.0
7.3±
5.5
7.6±
5.1
6.0±
4.1
12.6
±7.4
UB
-488
8M
elbo
urn:
SK
1271
SG
8940
9,08
651
.2±9
.320
.5±1
67.
7±5.
66.
7±5.
17.
7±5.
26.
1±4.
212
.6±7
.4
UB
-488
3M
elbo
urn:
SK
1038
, SG
9583
8,62
036
.9±1
2.0
27.2
±21.
011
.2±7
.79.
6±7.
08.
4±5.
76.
7±4.
614
.7±7
.4
UB
-647
9M
ill H
ill: g
rave
40
143,
349
59.2
±5.5
10.6
±8.6
3.2±
2.7
2.9±
2.5
13.3
±8.2
10.9
±6.6
21.1
±7.5
UB
-472
8M
ill H
ill: g
rave
64
338,
054
51.2
±7.3
15.7
±12.
25.
0±4.
04.
5±3.
713
.0±8
.110
.6±6
.521
.1±7
.4
lab
orat
ory
num
ber
Site
iSo
Sou
rC
em
ays (
1997
) (δ
13C
)
# so
lutio
ns%
veg
etat
ion
%
terr
estr
ial
anim
al
prot
ein
%
eel
% fr
eshw
ater
fis
h%
salm
onid
s%
mar
ine
fish
%
mar
ine
fish
123
UB
-472
9M
ill H
ill: g
rave
68
708,
639
40.5
±10.
022
.4±1
7.1
7.7±
5.7
6.8±
5.2
12.5
±7.9
10.1
±6.3
21.1
±7.4
UB
-473
0M
ill H
ill: g
rave
79
488,
430
44.1
±8.0
17.4
±13.
45.
5±4.
35.
0±4.
015
.4±9
.512
.5±7
.625
.3±7
.4
UB
-492
1M
ill H
ill: g
rave
81
469,
759
48.7
±9.9
21.9
±17.
18.
5±6.
07.
3±5.
57.
6±5.
16.
0±4.
112
.6±7
.4
UB
-473
1M
ill H
ill: g
rave
93
737,
694
40.3
±10.
824
.3±1
8.6
8.8±
6.3
7.7±
5.8
10.5
±6.8
8.4±
5.4
17.9
±7.4
UB
-473
2M
ill H
ill: g
rave
94
619,
692
43.4
±9.7
21.8
±16.
87.
6±5.
66.
7±5.
111
.3±7
.29.
1±5.
818
.9±7
.4
UB
-473
3M
ill H
ill: g
rave
95
707,
106
41.2
±10.
824
.3±1
8.7
8.9±
6.4
7.8±
5.9
9.9±
6.4
7.9±
5.2
16.8
±7.4
UB
-473
4M
ill H
ill: g
rave
105
C56
4,75
245
.0±1
0.7
23.8
±18.
69.
6±6.
78.
2±6.
17.
4±5.
15.
9±4.
112
.6±7
.4
UB
-493
0St
Pet
er’s
Tip
: gra
ve 4
211
,233
66.4
±3.2
4.3±
3.9
1.1±
1.2
1.0±
1.1
14.6
±9.1
12.7
±7.3
24.2
±7.4
UB
-634
6U
B-4
925
St P
eter
’s T
ip: g
rave
68
726,
834
39.7
±9.9
22.2
±17.
07.
5±5.
66.
7±5.
213
.2±8
.210
.6±6
.622
.1±7
.4U
B-4
925
UB
-603
2St
Pet
er's
Tip,
gra
ve 7
3A23
0,67
857
.8±8
.318
.1±1
4.4
9.1±
6.1
7.3±
5.5
4.3±
3.3
3.3±
2.7
6.3±
7.4
UB
-492
4St
Pet
er’s
Tip
: gra
ve 1
1312
0,26
161
.2±5
.210
.0±8
.23.
0±2.
62.
7±2.
412
.6±7
.810
.4±6
.320
±7.4
UB
-653
4
UB
-492
9St
Pet
er’s
Tip
: gra
ve 1
9467
4,52
242
.1±1
0.8
24.3
±18.
89.
1±6.
57.
9±5.
99.
3±6.
17.
4±4.
915
.8±7
.4
UB
-496
2St
Pet
er’s
Tip
: gra
ve 1
9634
6,84
953
.0±7
.917
.2±1
3.4
5.8±
4.4
5.2±
4.1
10.4
±6.6
8.4±
5.3
16.8
±7.4
UB
-496
3St
Pet
er’s
Tip
: gra
ve 2
0847
1,42
848
.2±8
.819
.5±1
5.1
6.7±
5.0
5.9±
4.6
10.9
±6.9
8.7±
5.6
17.9
±7.4
UB
-492
6St
Pet
er’s
Tip
: gra
ve 2
1288
9,23
835
.9±1
2.0
27.4
±21.
011
.0±7
.69.
5±7.
09.
0±6.
07.
2±4.
815
.8±7
.4
UB
-492
8St
Pet
er’s
Tip
: gra
ve 2
5052
6,78
946
.8±9
.922
.1±1
7.1
8.1±
5.9
7.1±
5.4
8.8±
5.8
7.0±
4.7
14.7
±7.4
UB
-492
7 St
Pet
er’s
Tip
: gra
ve 2
6344
3,51
948
.8±8
.418
.6±1
4.3
6.2±
4.7
5.6±
4.4
11.6
±7.3
9.3±
5.9
18.9
±7.4
UB
-492
7
lab
orat
ory
num
ber
Site
iSo
Sou
rC
em
ays (
1997
) (δ
13C
)
# so
lutio
ns%
veg
etat
ion
%
terr
estr
ial
anim
al
prot
ein
%
eel
% fr
eshw
ater
fis
h%
salm
onid
s%
mar
ine
fish
%
mar
ine
fish
ANGLO-SAxON GRAVES AND GRAVE GOODS124
lab
orat
ory
num
ber
Site
iSo
Sou
rC
em
ays (
1997
) (δ
13C
)
# so
lutio
ns%
veg
etat
ion
%
terr
estr
ial
anim
al
prot
ein
%
eel
% fr
eshw
ater
fis
h%
salm
onid
s%
mar
ine
fish
%
mar
ine
fish
UB
-647
8St
Pet
er's
Tip:
gra
ve 3
6095
4,53
833
.6±1
2.3
28.3
±21.
612
.6±8
.410
.6±7
.68.
3±5.
76.
6±4.
514
.7±7
.5
UB
-603
3W
est H
esle
rton,
gra
ve 1
1356
8,90
635
.9±1
0.9
23.7
±18.
719
.1±1
0.5
13.4
±9.6
4.4±
3.6
3.5±
2.9
7.4±
7.4
UB
-470
6W
est H
esle
rton:
002
BA
00
536
313,
197
54.3
±7.6
16.4
±12.
85.
4±4.
24.
9±3.
910
.5±6
.68.
4±5.
316
.8±7
.4
UB
-470
5W
est H
esle
rton:
002
BA
00
606
351,
066
53.8
±8.7
19.1
±14.
97.
0±5.
26.
1±4.
77.
8±5.
26.
2±4.
212
.6±7
.4
UB
-498
5W
estg
arth
Gar
dens
: gra
ve 1
123
2,36
958
.8±8
.217
.9±1
4.1
8.0±
5.5
6.6±
5.0
4.9±
3.6
3.8±
2.9
7.4±
7.4
UB
-483
6W
estg
arth
Gar
dens
: gra
ve 2
751
2,64
946
.9±9
.120
.3±1
5.7
7.0±
5.2
6.2±
4.8
10.8
±6.9
8.7±
5.5
17.9
±7.5
UB
-468
2W
estg
arth
Gar
dens
: gra
ve 6
666
2,62
942
.5±1
0.5
23.5
±18.
18.
6±6.
27.
5±5.
79.
9±6.
58.
0±5.
216
.8±7
.5
THE HUMAN SKELETONS 125
distribution of feasible solutions (Phillips & Gregg 2003). How source values and adjustments such as enrichment values are handled is left to the user. In this example, we use faunal and fish bone-collagen isotopic values for the diet sources and solve for individual human bone-collagen isotopic values. This is a simplification that excludes the additional conversions of faunal and fish bone-values to flesh, and the concomitant translation of diet to human collagen values.
Our choice of food source-values for the model also takes into account the importance of referring to or developing region-specific averages, as local endpoint values for C and N in specific food-webs can deviate from global mean values (Ambrose 1993; Doucett et al 1999; Takai & Sakamoto 1999; McCarthy & Waldron 2000). Müldner and Richards (2005) also commented on suitable end points, as well as the handling of trophic enrichment factors when stable-isotope mixing models are used for archaeological situations.
In the ISOSOURCE calculations we therefore use the mean δ13C and δ15N values for foods as collated from British archaeological sites (Richards et al 2006; Jay & Richards 2006; Privat et al 2002; Richards 2000; Müldner & Richards 2005; DeNiro & Epstein 1978). These are:
‘terrestrial vegetation’ (−21.8‰ for carbon and +4.6‰ for nitrogen)
‘terrestrial animal protein’ (−21.5‰ and +5.9‰) ‘eel’ (−23.3‰ and +10.9‰) ‘freshwater fish’ (−21.7‰ and +13.7‰)
‘salmonid’ (−15.0‰ and +11.2‰)‘marine fish’ (−13.1‰ and +13.5‰)
The mean of each food-type was then given a trophic enrichment factor of 1‰ for 13C and 4‰ for 15N, except for ‘terrestrial vegetation’, as the cattle who are proxies for human vegetarians have essentially provided the trophic enrichment factor. The ISOSOURCE calculations were then run for each of the 94 skeletons (Tab. 4.8).
ISOSOURCE results suggest that for Anglo-Saxon females, overall, dietary protein sources consisted of 43.7±11.0% terrestrial vegetation, 24.5±19.1% terrestrial animal protein, 10.0±7.0% eel, 8.5±6.3% freshwater fish, 7.4±5.1% salmonids, and 5.9±4.1% marine fish (N calculations = 597,278). For males, ISOSOURCE indicated that the protein sources were 46.0±9.9% terrestrial vegetation, 22.1±17% terrestrial animal protein, 8.0±5.8% eel, 7.0±5.3% freshwater fish, 9.4±6.1% salmonids, and 7.5±4.9% marine fish (N calculations = 552,610).
In calculations from the mean isotopic values in males and female skeletons by site, the estimated proportion of dietary protein made up by marine foods ranged from 3.1±2.7% (Castledyke South females) to 9.3±5.8% (Mill Hill males). The calculations for individuals in all sites suggest that marine fish made up from 1.4% (Castledyke South grave 94) to 12.7% (St Peter’s Tip grave 42) of dietary protein.
The most enriched mean δ15N (+10.2‰) among sites was observed at Castledyke South. However,
0
5
10
15
20
25
30
35
40
45Ap
ple D
own
Aston
Clin
tonBe
rinsfi
eld
Butte
rmar
ket
Castle
dyke
Sou
thCod
denh
amDov
er B
uckla
nd
Edix
Hill
Ford
, Lav
ersto
ckGall
y Hills
Lech
lade
Lake
nhea
th
Marina
Driv
e
Melbou
rn
Mill Hill
St P
eter
's Tip
Wes
t Hes
lerto
n
Wes
tgarth
Gar
dens
sites
perc
ent o
f die
t% marine fish% eel% freshwater fish% salmonids
Figure 4.4Percentage of eels, freshwater fish, marine fish and salmonids for individual skeletons from each site calculated by ISOSOURCE.
THE HUMAN SKELETONS 127
eel and freshwater fish than marine sources (Tab. 4.8). Geography may play a role here, in the type of aquatic biome from which food could be gathered. Castledyke South is on the south bank of the Humber estuary, while Buckland Dover, Mill Hill and St Peter’s Tip are in proximity to the coast in Kent. If small proportions of fish are contributing to the slightly elevated nitrogen values in these sites, the mean δ13C value of −20.79‰ for women at Castledyke South suggests a terrestrial aquatic source. Castledyke South’s females and males have cumulative non-marine fish proportions of 35.5±8.2% and 25.5±6.1%, respectively, represented largely by eel and freshwater fish. In the Kent coastal sites, mean non-marine fish estimates for Buckland Dover (females 24.7±5.8%; males 23.5±7.1%), Mill Hill (females 25.2±5.9%; males 23.8±5.7%), and St Peter’s Tip (females 21.8± 5.3%; males 24.7±5.9%) indicate that salmonids dominate the calculations (Tab. 4.8). In Figure 4.4 the percentage of marine, eel, freshwater fish and salmonids in dietary protein has been plotted for individual skeletons from each site. In Figure 4.5 a map shows the regional variation in the components of marine and non-marine fish.
Linear interpolation from δ13C
We now turn to a second method of estimating dietary protein-source proportions from human isotopic signature, by interpolation calculations of percent marine from δ13C. Here we use the formula of Mays (1997):
where δCO is the δ13C value of collagen, δT is the estimated value of collagen for those consuming only terrestrial resources, and δM is estimated value of collagen for those consuming only marine resources, the values for which are:
δT : −21.5±0.6‰ for δ13CδM: −12.0±1.5‰ for δ13C
For errors in the Mays (1997) method, we have calculated the percent equivalent of the per mille standard deviations of isotope values used here for δT and δM, and the associated error on each skeleton’s δCO. The percentage equivalents are then squared, summed and divided by 3, and the square root taken for the combined percent error for each skeleton.
ISOSOURCE results had indicated that, overall, protein sources for Anglo-Saxon females and males have an average of 1.4% to 12.7% marine fish, with notable inter- and intra-site variations. Mays’s (1997) δ13C back-calculation estimates of the marine proportion of dietary protein return higher estimates of percentage marine (2.1% to 25.3%) in comparison with the ISOSOURCE calculations (Tab. 4.8). The two methods compare well within their estimated errors, although that of Mays (1997) does estimate higher percentages of marine protein in the diet (Fig. 4.6). This difference may be due to the more basic assumptions of back-calculation formulae such as Mays (1997) and others (Arneborg et al 1999; Ambrose 1993), including that all variation in δ 13C is due to marine versus terrestrial foods, that δ13C varies in a linear fashion with protein composition in diet,
R2 = 0.0962
R2 = 0.0966
0.0
5.0
10.0
15.0
20.0
25.0
30.0
35.0
Dover
Buc
klan
d
Mill
Hill
St P
eter
's Ti
p
Butte
rmar
ket
Cast
ledy
ke S
outh
Appl
e Dow
nW
est H
esle
rton
Codd
enha
m
Ford
, Lav
erst
ock
Lake
nhea
thGal
ly H
ills
Wes
tgar
th G
arde
ns
Lech
lade
Edix
Hill
Mel
bour
n
Mar
ina
Drive
Berin
sfiel
dAs
ton
Clin
ton
perc
ent
mar
ine
sites
Mays (1997)
ISOSOURCE
Figure 4.6Comparison of marine percentage in diet calculated from ISOSOURCE (Phillips & Gregg 2003), and by linear interpolation from δ13C (Mays 1997). Sites are ordered from left to right as closest (0 km) to furthest (115 km) from the sea. R2 of the linear trend for ISOSOURCE and Mays 1997 calculations show that while Mays 1997 returns higher estimates of marine percentage, the overall trends of the two meth-ods are similar.
ANGLO-SAxON GRAVES AND GRAVE GOODS128
and that only terrestrial and marine end-points are considered. Each of these assumptions may impart some error on the estimation of percentage marine component in a given diet.
4.4 Anglo-SAxon dietS: the effeCtS of Sex, Age And geogrAPhY
The aim of this section is to investigate differences in δ13C and δ15N (and by implication, differences in diet) according to age at death, sex and geographical location of the sampled skeletons. Only adult burials for which sex could be determined using standard osteological methods (Brothwell 1981) are included in this part of the study. Sub-adults are excluded as previous work on other English medieval material (Privat et al 2002; Richards et al 2002) has shown significant age-related variation in isotopic values, and by implication diet, in sub-adults. There are 76 adults for whom sex was determined by osteological methods: 38 males and 38 females (Tab. 4.2).
The archaeological sites from which the burials were excavated were classified as coastal (Buckland Dover, Mill Hill and St Peter’s Tip) or inland (Apple Down, Aston Clinton, Edix Hill, Berinsfield, Coddenham, Marina Drive, Ford, Gally Hills, Lakenheath, Lechlade, Melbourn Water Lane, West Heslerton and Westgarth Gardens). Two sites, Castledyke South and Buttermarket (Ipswich), are located in the lower reaches of major rivers, and hence are classified here as riverine. For the inland sites, distance from the sea, measured in a straight line to the nearest point on the coast, was also recorded.
There was no statistically valid correlation between δ13C and δ15N (r = 0.10, p = 0.40). Therefore the δ13C and δ15N data are analysed
separately. Lilliefors tests indicated no evidence for departure from normality for the δ15N data, permitting parametric statistics to be used in analysis. However, the δ13C data departed from normality. Therefore the carbon data were analysed using non-parametric statistics, and the median and the inter-quartile range (IQR) were used, instead of mean and standard deviation (SD), as measures of central tendency and dispersion respectively.
There was no association between δ13C and age at death. However, analysis of variance indicated an association between δ15N and age for males which just attained conventional statistical significance (F = 3.29, p = 0.05). The Tukey HSD post hoc multiple comparison test indicated that the only between-group difference that attained statistical significance was between the 18–29 and 30–49 year age classes, the former being depleted in 15N compared to the latter (mean values +9.1‰ to +9.7‰ respectively).
Summary statistics for the isotopic data, split by sex and location, are presented in Tables 4.9 and 4.10. These data indicate no difference in stable-isotope values between males and females at any of the three geographical locations. In addition, there is no significant difference in the distributions of age at death in sites from the three geographical locations. Therefore, in the analyses directed at evaluating geographical differences in stable-isotope data, individuals of all three age-groups and both sexes are combined.
A Kruskal-Wallis one-way non-parametric analysis of variance indicated significant variation in δ13C across the three geographical groups (χ2 = 38.7, p <0.0001). The post hoc non-parametric multiple comparison test of Dunn incorporating the modification for tied ranks (Dunn 1964; Zar 1999), revealed significant differences between the coastal
Table 4.9.Summary statistics for δ13C (‰); IQR = interquartile range; Mann-Whitney test used to test for sex dif-ferences in δ13C in each context. * = two identical values hence no IQR calculated.
males females All adults mann-whitney testlocation n median iQr n median iQr n median iQr z p
Inland 17 -20.2 0.4 25 -20.3 0.3 42 -20.3 0.3 0.79 0.43Riverine 2 -20.5 -* 6 -21.0 0.9 8 -20.7 0.7 0.67 0.50Coastal 19 -19.8 0.5 7 -19.8 0.2 26 -19.8 0.4 0.26 0.79
Table 4.10Summary statistics for δ15N (‰); t-test used to test for sex differences in δ15N in each context.
males females All adults t-testlocation n mean Sd n mean Sd n mean Sd t p
Inland 17 9.5 0.7 25 9.5 0.9 42 9.5 0.8 0.10 0.92Riverine 2 9.5 0.8 6 10.6 0.6 8 10.3 0.8 1.64 0.30Coastal 19 9.6 0.6 7 9.7 0.5 26 9.6 0.6 0.37 0.72
THE HUMAN SKELETONS 129
group and the other two groups (p = <0.001 in each case).
For the δ15N data, analysis of variance indicates significant variation across the three geographical groups. The Tukey HSD post hoc multiple comparison test indicated a significant difference between the riverine group and the inland group (p <0.05).
For the inland sites, there is no association between distance from the coast and δ13C (r = –0.07). However, there is a weak but statistically significant positive association between distance from the coast and δ15N (r = 0.35, p = 0.02).
As an alternative to the ISOSOURCE method and use of published values for UK archaeological fauna and fish, Mays (1997) has suggested human bone collagen δ13C endpoints of approximately −21.5‰ and −12‰ for fully terrestrial and fully marine protein components of diets in a medieval English context. In this light, the δ13C data indicate diets dominated by terrestrial resources at each geographic zone. The variation in δ13C is low. That in δ15N is somewhat greater, but this is as expected, given the multiplicity of factors that potentially contribute to variation in δ15N. For example, δ15N differs in marine versus terrestrial environments; varies with trophic level so that animal products generally have higher δ15N than plant foods (Bocherens & Drucker 2003) and freshwater foods often have elevated values (Müldner 2005); varies in different domestic animals (Müldner & Richards 2007); and differs in leguminous versus
non-leguminous plants (DeNiro & Hastorf 1985). The nitrogen balance of the organism also affects δ15N (Fuller et al 2005), and manuring practices influence δ15N in crops (Bogaard et al 2007).
Comparison of the degree of variation in stable-isotope ratios can be drawn with later medieval Barton-on-Humber, situated in the same locality as the Castledyke site in the current work (Beavan et al 2011b). The Barton-on-Humber material came from a parish church, and so to all intents and purposes represents a single community. The material was split by period into six phases, most of about one to two centuries. The adult IQRs for δ13C at Barton for the different phases range from 0.3–2.4‰; the SDs for δ15N from 0.6–2.3‰. At Wharram Percy, an inland later medieval site, the adult IQR for δ13C was 0.49‰, the SD for δ15N 1.04‰ (computed from data of Richards et al 2002). The results therefore suggest that, despite the geographical diversity of the current data, the variability in dietary protein sources between individuals was less than in a single riverine community, and was similar to that in a single inland community, in the later medieval period.
An age-pattern was found in δ15N in males whereby those over about 30 years have elevated δ15N compared with younger adults, although the difference in means is small (0.6‰). There was no interaction between age and geographical location which might explain this. This pattern echoes that found by Privat et al (2002) in a detailed study of the Berinsfield group. It would appear to suggest
R2 = 0.0189
–22
–21
–20
–19
–18
7 8 9 10 11 12
δ15N ‰ (AIR)
δ13C
‰ (V
PDB
)
Figure 4.7δ13C and δ15N values for all Anglo-Saxon skeletons analysed show no relationship between enriched nitrogen and enriched carbon, which would be characteristic of marine inputs to diet.
ANGLO-SAxON GRAVES AND GRAVE GOODS130
a minor dietary change during the male life-cycle, which did not have a parallel in females, with a slightly increased proportion of dietary protein at older ages coming from sources somewhat more enriched in 15N. Cortical bone collagen turnover rates are only about 1.5–3% per year in adult males (Hedges et al 2007), so this finding must relate to a shift that occurred well before 30 years of age, probably during the latter part of the growth period. However, identifying more closely the age range at which this might have occurred is problematic given the imprecision of current skeletal ageing techniques.
There are some very minor, but nevertheless statistically significant differences in δ13C between the three geographical zones. The coastal burials have less negative δ13C than those from the other two locations. Given the less negative δ13C of marine versus terrestrial foods, it would seem likely that, for the coastal people, marine foods tended to make up a somewhat greater proportion of protein sources than was the case at the other sites (or perhaps that in coastal locations domestic animals were fed, or allowed to forage for, marine foods such as seaweed). However, if this interpretation is correct, the question remains as to why there is no corresponding elevation of δ15N in the coastal burials. As explained above, δ15N in bone collagen is influenced by a multiplicity of factors, and our understanding of the causes of δ15N variation in past populations remains incomplete. It may be that the contribution to inter-individual variability in δ15N made by differential consumption of marine resources is eclipsed by that due to other factors: in other words, it is easier to pick out minor marine dietary components using analysis of bone collagen for δ13C because, on current understanding, it appears to be influenced by fewer additional factors than is δ15N. In support of this is the lack of significant correlation between δ15N and δ13C in the sample as a whole (Fig. 4.7). If differential consumption of marine foods were the dominant factor behind inter-individual variation in δ15N then we would expect a correlation with δ13C. That this does not seem to be the case suggests that factors other than differential seafood consumption are the principal causes of inter-individual variation in δ15N.
Mean δ13C at the coastal locations is more negative than in some later medieval English coastal settlements (e.g. Newcastle and Hartlepool: −18.6‰ and −18.2‰ respectively) and also in a later medieval urban centre (York: −18.8‰ — −19.7‰) (Mays 1997; Müldner & Richards 2007). If the above interpretation of the current δ13C data is correct, then the seafood component of the coastal Anglo-Saxon series is, whilst greater than at contemporary non-coastal sites, rather less than at later medieval coastal ports and inland trading centres. This may be consistent with the suggestion
of Barrett et al (2004a–b), based on faunal evidence, that there was an expansion of marine fishing in England around AD 1000, and that prior to this exploitation of marine resources was minor. It was anticipated that if marine resources were regularly traded inland, then one might observe an association between increasing distance from the coast and more negative δ13C. There was no indication of any such trend.
Turning to the nitrogen isotope data, the δ15N values are significantly greater in the riverine sites than the inland sites. As noted above, interpretation of δ15N data from bone collagen is somewhat complex, but the observation that freshwater resources tend to have rather elevated δ15N, suggests that increased consumption of riverine resources in communities situated in the lower reaches of major water courses may be a factor here. The data (Tab. 4.10) indicate that it is in fact the females from the riverine sites that have the elevated δ15N. However, the sex-difference in diet at riverine locations is not statistically significant, reflecting the small sample size for males (two individuals). That there are only two males from riverine sites means that we cannot adequately characterize male diets at these locations, so it would be unwise, on the currently available data, to argue that the inland/riverine dietary difference was solely restricted to females.
For inland sites, there was a positive correlation between distance from the coast and δ15N. This association was weak and, as discussed above, intra-site variation is very minor. It is difficult to explain this trend in dietary terms — for example it is the opposite of what might be expected if the variation were due to proximity of marine food resources.
There is some historical and archaeological evidence for fish consumption in Anglo-Saxon sites. Recent research into fishing and fish-consumption in Anglo-Saxon England (Barrett et al 2004a–b) has shown a dramatic growth of large-scale marine fishing and the wide distribution of the catches only around the end of the 10th century AD. The data also suggest that there had previously been a steady growth in the quantity of herring being landed from at least the 8th century to the 10th; in the 8th century, however, this marine fish is found only at sites of special character — monasteries and wīcs — located very close to the sea. The fishbone evidence from earlier sites is dominated by freshwater and estuarine species: especially cyprinids (carp family) and eel. There is also evidence of the regular collection of shellfish. At the monastery at Hartlepool (Co. Durham), however, there was clearly some off-shore catching of gaddids (cod and related species) in the late 7th and 8th centuries (Huntley & Rackham 2007, 109–10 & 120–2). At Flixborough (S. Humberside), the regular hunting and consumption of bottlenose dolphin, presumably from the Humber estuary, is
THE HUMAN SKELETONS 131
clearly in evidence both in the 8th and the 10th century: both periods for which a special status for this settlement site is therefore inferred — perhaps, again, monastic in the earlier period (Dobney et al 2007, 199–214; Loveluck 2007, 92–4 & 147–57).
Of particular importance in the present context is the fact that we do not know whether the apparently general increase in the provision of herring from the 8th to 10th centuries can (negatively) be back-projected into the 6th and 7th centuries — which could imply that fishing then was limited almost entirely to riverine and estuarine trapping. A considerable number of fish-traps dating across the period from the 5th century to the 9th have been identified, especially in the major English rivers: the Thames, Trent, and Severn. The earliest reliably dated charters, from south-eastern England, of the second half of the 7th century, include watercourses and shores amongst the resource areas granted in a regular, if formulaic, manner, implying that the value of those zones, for fishing as well as other productive uses, was fully appreciated.
4.5 ConCluSionS
Individuals were selected for inclusion in this study on the basis that they were accompanied by certain artefact-types whose chronological distribution was of interest. Their relatively rich grave goods mean that the individuals in the current study probably represent higher status members of Anglo-Saxon society. Whether our findings have a more general applicability must await further isotopic work on non-elite segments of Anglo-Saxon communities.
The women in our sample were of average stature compared to the general female population, although the men in our sample, in particular those buried with weapons, were significantly taller than average. Environmental factors, perhaps in the form of better nutrition during childhood and adolescence, may have been important causes of this difference. Weapon burials, and indeed male burials, in our sample do not have a higher prevalence of injury or fracture than the female burials. One skeleton, MaDE1, showed facial signs of leprosy.
The two screening methods for protein preservation — the atomic C:N ratios and amino-acid analysis — generally confirmed the suitability of bone for radiocarbon dating and stable-isotope analysis. C:N ratios for the burials are within the expected range for well-preserved collagen (2.9–3.6: DeNiro 1985). Results on the replicate δ13C and δ15N values, the %C and %N values, and atomic C:N ratios for fourteen of the dated skeletons were statistically consistent (at 95% confidence; Ward & Wilson 1978) and demonstrate the reproducibility of the δ13C and δ15N values quoted.
The stable-isotope data indicate that protein sources in human diets at all sites studied were
overwhelmingly terrestrial. Estimates of the proportion of various food sources by both ISOSOURCE and Mays (1997) suggests that marine fish, if available or utilized, usually made up less than 10% in these diets and would not affect radiocarbon ages significantly.
A sub-set of the data examined differences by age, sex and geographical relationships. For males, adults over about 30 years had slightly enriched δ15N compared with younger adults. There was no evidence for age differences in δ15N in females nor in δ13C for either sex. There was no evidence of any sex differences in δ13C or δ15N at sites in any of the three geographic locations. There were some differences between communities in inland, riverine, and coastal locations, consistent with seafood making a greater contribution to dietary protein in coastal communities and freshwater resources being more important in communities situated in the lower reaches of major rivers. However the differences were very minor. Overall, despite the geographical spread of the current data-set, dietary protein sources appear of similar or greater homogeneity than was the case even within single communities in later medieval times.
The slightly greater consumption of marine and freshwater resources inferred for coastal and riverine sites respectively may suggest that in times of shortage communities fell back on locally available wild resources, and that at coastal and riverine locations these included marine and freshwater resources respectively. Alternatively, marine and freshwater resources may, when available locally, have constituted a minor but regular supplement to foods generated by the agrarian economy. The current data do not permit distinction between these two scenarios.
While consumption of freshwater fish also has the potential to introduce radiocarbon offsets, there is sparse archaeological evidence for non-marine fish consumption in Anglo-Saxon sites, and little information on freshwater radiocarbon offsets for fish in the overall region in which these sites are situated. Notwithstanding the number of fish-traps which have been identified from the 5th century to the 9th centuries in major English rivers as evidence for riverine and estuarine fishing in the period, there is little evidence from site-associated fish-remains to further support extensive fish consumption. Müldner and Richards (2005) had observed that higher δ15N coupled with terrestrial range δ13C could be indicative of some freshwater fish contribution to diet; but given the overlap of some freshwater fish isotopic values with terrestrial protein, teasing out <10% proportions of this food-type in diet would be difficult by any current method. Our evaluation of the stable-isotope data for the Anglo-Saxon skeletons from furnished graves can, at best, provide informed and probable parameters for calibration of the radiocarbon ages.