- Email: [email protected]
A demographic analysis of Maiden Castle hillfort: Evidence for conflict in the late Iron Age and early Roman period
International Journal of Paleopathology 1 (2011) 68–73
Contents lists available at ScienceDirect
International Journal of Paleopathology journal homepage: www.elsevier.com/locate/ijpp
Case Study
A demographic analysis of Maiden Castle hillfort: Evidence for conflict in the late Iron Age and early Roman period Rebecca C. Redfern a,∗ , Andrew T. Chamberlain b a b
Centre for Human Bioarchaeology, Department of Archaeological Collections and Archive, Museum of London, 150 London Wall, ECY 5HN, England, United Kingdom Department of Archaeology, University of Sheffield, Northgate House, West Street, Sheffield S1 4ET, England, United Kingdom
a r t i c l e
i n f o
Article history: Received 13 December 2010 Received in revised form 19 February 2011 Accepted 21 February 2011 Keywords: Demography Mortality Warfare Roman Empire Late Iron Age Britain
a b s t r a c t The late Iron Age human remains from the British hillfort of Maiden Castle are frequently cited within the archaeological and bioarchaeological literature as providing evidence for conflict. This interpretation is based on osteological work undertaken in the late 1930s. In order to test the validity of using this sample in conflict research, the authors undertook a detailed analysis of the site’s demography in comparison with contemporary late Iron Age attritional cemeteries from Dorset (England) and additional conflict mortality data. These results showed that the 1st century BC to 1st century AD samples conformed to a catastrophic profile, as greater numbers of young adult males had been buried during this phase of occupation. In combination with new bioarchaeological findings and the identification of an embedded Roman projectile weapon, we conclude that individuals had died during an episode of warfare, one of which probably included the Roman conquest of 43AD. © 2011 Elsevier Inc. All rights reserved.
1. Introduction
2. Archaeology of the Iron Age hillfort
Skeletal evidence for violent trauma provides one of the most important categories of evidence in the study of ancient warfare (Thorpe, 2003). Maiden Castle hillfort in Dorset, England, is frequently cited in archaeological literature as providing evidence for warfare (Armit, 2007; Bishop and Knüsel, 2005; Renfrew and Bahn, 2008; Sharples, 1991a,b,c) and potentially, the Roman invasion of Britain in AD 43 (Redfern, 2011; Sharples, 2010). This interpretation is primarily based on the osteological work undertaken by Morant and Goodman in the late 1930s, which identified instances of perimortem violent trauma, and Wheeler’s (1943) reconstruction of events at the hillfort. In order to test the validity of using this sample in conflict research, the authors examined the site’s demography and undertook a detailed analysis of adult age-at-death using the pubic symphysis and compared these results to contemporary late Iron Age attritional cemeteries from Dorset and additional conflict mortality data.
Maiden Castle hillfort (Fig. 1) is located 1 km from the modern County town of Dorchester. In the late Iron Age (100 BC to AD 100), based on pottery distributions and the minting of coinage, this area of southwest England was inhabited by the Durotriges tribe which was formed by a confederacy of smaller communities (Gale, 2003, 98). Within the tribal territory a number of hillforts have been identified, and similarly to other areas of southern Britain, these appear to have fulfilled a variety of functions beyond purely defensive purposes, including domestic settlements and the production of material culture (Armit, 2007; Cunliffe, 1971, 2005; Gale, 2003, 108–117; Sharples, 2010). For the Durotriges, Maiden Castle appears to have been an important centre of power and social organisation and under-went several phases of development, including the construction of defensive earthworks and the complex gateways at the eastern and western ends of the hillfort. Archaeological evidence also shows that it was used for settlement, with the identification of roundhouses and middens; by the late Iron Age, occupation also existed in the outworks of the eastern entrance and this area was also used for metalworking (Sharples, 1991b, 2010). Sharples (1991a, 123–124) considers that these changes are indicative of wider changes within the Durotriges’ territory, particularly the emergence of centralised power in the region, and the expansion of ceramic, salt, shale, stone and metal working industries.
∗ Corresponding author. Tel.: +44 20 7814 5649. E-mail addresses: rredfern@museumoflondon.org.uk (R.C. Redfern), [email protected] (A.T. Chamberlain). 1879-9817/$ – see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.ijpp.2011.02.004
R.C. Redfern, A.T. Chamberlain / International Journal of Paleopathology 1 (2011) 68–73
69
Fig. 1. Location map of Maiden Castle, Dorset. The figure was drawn by G. Rapson.
3. Interpretation of the Iron Age hillfort As outlined in 2, hillforts were used for many different purposes during the late Iron Age. However, at Maiden Castle, the most popular and long-standing interpretation has been its defensive role in warfare, particularly during the AD 43 invasion of Britain by the Roman Emperor Claudius. The first to promote this hypothesis was Sir Mortimer Wheeler (1943, 1955), who following his excavations at the hillfort in the 1930s, envisioned a vicious, bloody battle between the late Iron Age inhabitants of the fort and the Roman legions commanded by Vespasian. Wheeler (1943) considered that the evidence for burning at the hillfort’s complex eastern entrance, the vast numbers of sling-stones recovered from the entrance, and the skeletons with evidence of fractures and weapon trauma provided unequivocal evidence of this battle, an interpretation which has remained popular amongst the public and a number of scholars (see Wainwright and Cunliffe, 1985). Wheeler (1943) assigned the burial and occupation phases of the pre-Roman hillfort into four groups in accordance with Hawkes’ (1939) cultural (‘ABC’) divisions of the British Iron Age. Wheeler’s phases were: Iron Age A and B including burials in pits/features within the hillfort and not directly associated with any gravegoods; Iron Age C, burials in the ramparts and ditches of the eastern entrance, though only a small minority of this group could be directly associated with any grave-goods; and the contemporary
Belgic war cemetery (first century BC to first century AD)—the latter included interments found behind the hillfort’s eastern entrance and were buried with grave-goods. Fig. 2 illustrates the location of the burials in the eastern entrance; location information for all the other burials are described by Wheeler (1943). Re-examination of the archaeological, funerary and material culture evidence shows that these groupings artificially divide the sample (Hamlin, 2007; Sharples, 1991b). Based on contemporary phasing of this period: the Iron Age A group (300–100 BC) dates from the middle Iron Age, and groups B, C and the Belgic war cemetery date to the late Iron Age (first century BC to first century AD); unfortunately, it has not been possible to verify these archaeological phases with direct carbon 14 dating. Therefore, based on a more accurate phasing of the hillfort, Sharples’ (1991a,b, 2010, pers. comm.) proposes that the eastern entrance of the hillfort was used for burial over many hundreds of years and that these graves should be considered to reflect a single burial area. Sharples’ (1991a,b) further revised several aspects of Wheeler’s (1943) previous interpretation through the close study of the data collected during his and Wheeler’s excavations at the site. Firstly, Wheeler (1943, 62) argues that the burials were, ‘carried out anxiously and hastily without order’. A closer examination of the funerary evidence shows that the burials follow standard Durotrigian funerary practices, with respect to body positioning and selection of grave goods and were inhumed in an existing cemetery
70
R.C. Redfern, A.T. Chamberlain / International Journal of Paleopathology 1 (2011) 68–73
Fig. 2. Location of burials (by period) in the eastern entrance (after Sharples, 1991a, Fig 90. Redrawn by Rapson).
(see also Hamlin, 2007), which argues for a lengthy funerary ritual; the ash, interpreted by Wheeler (1943) as evidence of burning, is most likely to have been created by late Iron Age metal-working in the eastern entrance; the arrows identified by Wheeler probably pre-date the Roman conquest, but the slighting of the defences probably did occur after the conquest (Sharples, 1991a, 124–125). These data, combined with the osteological evidence presented by Morant and Goodman (1943) in which about one quarter of their sample had evidence for trauma, was not considered persuasive enough by Sharples (1991a) to support Wheeler’s interpretation (see Section 3.1). Recently, the first author has collaborated with Sharples and Gilliver to re-analyse two males buried in a double grave in the Belgic war cemetery. Examination of the embedded projectile weapon in the vertebra of P7A (Wheeler, 1943, Plate LVIIIA) has identified this to be a Roman ballista bolt (Gilliver, pers. comm.)—confirming Wheeler’s (1943, 63) original identification, and the square perforation in the left temporal bone of P7 (see Redfern, 2009) is most likely to have been caused by a Roman weapon (Gilliver, pers. comm.), as the only parallel that could be identified in late Iron Age and early Roman samples was from a Gallo-Roman grave dating from AD 200 to 300 (Billard, 1991). Although the burials at Maiden Castle hillfort are comparable to contemporaneous cemeteries in terms of funerary practices, it is the only cemetery of this period in Dorset to contain the multiple burials of adults (Hamlin, 2007). Six double and one triple burial were identified in the Belgic war cemetery area (Morant and Goodman, 1943, 352–355), all containing individuals with peri-mortem trauma (Redfern, 2011). This evidence again supports Sharples’ (1991a,b, 2010) hypothesis that the Belgic war cemetery burials were part of, or associated with, an existing burial ground (see Fig. 2).
3.1. Bioarchaeological analysis of the hillfort’s Iron Age samples The first author used modern osteological methods to study the extant Wheeler sample (see Byers, 2006; Galloway, 1999; Knüsel, 2005; Johansen et al., 2008; Schmidt and Pollak, 2006; Shepherd et al., 1990; Ubelaker and Adams, 1995 for reviews of current methods). This established that the crude prevalence of individuals with peri- and ante-mortem injuries was far higher (74.2%) (Redfern, 2011) than previously reported by Morant and Goodman (1943) (27.7%). The Iron Age A group did not have any evidence of trauma present, the Iron Age B group had 54.4%, and the highest prevalence rates were observed in the Iron Age C group (75%) and the Belgic war cemetery (87.5%). The difference in the number of individuals with trauma between the latter two burial-groups was not quite statistically significant (X2 = 3.575, d.f. = 1, p = 0.0586) (Redfern, 2011). The re-analysis identified individuals of both sexes (male N = 30, female N = 24 data from Redfern, 2011) and one 15-year-old individual with over-whelming peri-mortem evidence for targeted blows to the head and body, assault injuries and over-kill, as evidenced by the presence of multiple trauma, sharp-force weapon injuries to the head, torso, upper and lower limbs, and extremities, sharp and blunt projectile injuries to the cranium, fractures caused by assault to the head and limbs, and sharp-force weapon injuries consistent with attempted decapitation. Examination of the distribution of fractures and injury patterns in males and females shows a considerable degree of concordance, with both sexes sustaining peri-mortem assault injuries to the face, torso and forearms (e.g. parry fracture), and sharp-force and projectile injuries to the skull suggesting that aggressors did not differentiate between the sexes. Additionally, many individuals of both sexes displayed healed antemortem trauma created by direct blows to the body, projectile and
R.C. Redfern, A.T. Chamberlain / International Journal of Paleopathology 1 (2011) 68–73 Table 1 Comparison of numbers and percentages of individuals in age and sex categories in the aggregated sample from the Iron Age settlement sites and the combined sample from the Maiden Castle cemeteries.
Settlement sites Maiden Castle
Subadults
Adult males
Adult females
48 (55%) 23 (31%)
23 (26%) 35 (45%)
17 (19%) 17 (24%)
sharp-force weapon injuries, suggesting that they had survived previous episodes of interpersonal violence. These findings are the first to demonstrate that, in addition to males, females and an adolescent were exposed to episodes of violence during this period. Nevertheless, it should be noted that the extant evidence from Dorset for female and adolescent involvement has only been identified in this sample, underscoring its uniqueness (see Redfern, 2007, 2011). Focusing on the peri-mortem evidence, the range and severity of the trauma observed in the sample can be illustrated by a 30–45 year male, who had sustained two fractures to his mandible, which also shows sharp-force weapon injury evidence for decapitation. In addition to which, he also suffered sharp and blunt injuries to his cranium, and fractures to the ribs and bilaterally to the femora. These injuries show that they had been sustained during a violent assault (Mohanty et al., 2007; Schmidt and Pollak, 2006). 4. Materials and methods The extant sample differed slightly from Wheeler’s (1943) published inventory (N = 83). Study of the published report, site notebooks and documentation, and Morant and Goodman’s notes showed that two individuals were no longer present for study from the late Iron Age samples excavated from the eastern gateway, identified by Wheeler as the Iron Age C and Belgic War cemetery samples. Twenty-one individuals whose context information recorded in the site notebooks showed them to be of Iron Age date but who could not be tied to a burial context were included in this study in order to provide a true representation of the extant sample (cf Redfern, 2011). It should be noted that Wheeler’s (1943) burial catalogue is incomplete and in the site report, Morant and Goodman (1943, 357) state that they are only able to provide data for 83 skeletons who could give ‘records of value’ (see also Goodman and Morant, 1940). As this present study focuses on the demographic composition of the sample, not all of the extant sample could be included because the necessary areas of the skeleton were not present. Therefore, a total of 23 subadults for whom age-at-death could be established from dental and/or skeletal methodologies and 52 adult individuals who had observable pubic symphyseal faces were used (Table 1). The inter-cemetery comparison data were collected by the first author and are derived from seven late Iron Age (1st century BC to 1st century AD) attritional burial locations in the county of Dorset, Kimmeridge, Tarrant Hinton, Flagstones, Alington Avenue, Gussage All Saints, Whitcombe Farm and Poundbury Camp (Table 2). Adult individuals from these cemeteries were only included in this study if one or both pubic symphyseal faces could be scored for age-estimation. 4.1. Sample used for analysis As stated above, this study used 23 subadults and 52 adults from the total sample examined by the first author. All individuals were recorded using the protocols published by Buikstra and Ubelaker (1994); with the sex of adult individuals (≥20 years old) determined using the morphology of the skull and pelvis following the methods described by Buikstra and Ubelaker (1994). As this study
71
Table 2 Late Iron Age comparison data: site type and location. Site name
Type
Location
Alington Avenue
Formal cemetery
Flagstones
Formal cemetery
Gussage All Saints Kimmeridge
Settlement Burial associated with a cemetery Cemetery associated with a hillfort Burials associated with a settlement Formal cemetery
South-east of Dorchester South-west of Dorchester North-east Dorset South Dorset
Poundbury Camp Tarrant Hinton Whitcombe Farm
North-west of Dorchester North-east Dorset South-east Dorset
focuses on the demographic composition of the assemblage, we have employed the pubic symphysis method of age determination (Brooks and Suchey, 1990). In order to enable direct comparison of demographic profiles, we decided to analyse the frequency distributions of phases of the pubic symphysis, a commonly used adult age indicator for which recording standards are widely available and consistently applied. Throughout adult life, the medial surface of the pubic symphysis undergoes progressive, irreversible morphological changes involving loss of transverse ridging, thickening of the joint margins and the development of pitting and unevenness of the joint surface. In the recording system devised by Brooks and Suchey (1990) six ordinal phases are defined, with Phase 1 characterising the youngest adults and Phase 6 the oldest. Each adult individual was scored into one of the six phases and, in individuals where the left and right side pubic symphysis exhibited different phases, the scores were averaged. The biological ages of subadults (<20 years old) were determined using a combination of dental development (Ubelaker, 1989), diaphyseal length (Scheuer and Black, 2000), and epiphyseal fusion methodologies (Scheuer and Black, 2000). When both dental and skeletal estimations were available, the dental age was used, as this has the most reliable correlation to chronological age (Lewis and Garn, 1960). No attempt was made to determine sex of subadults based upon skeletal morphology, as a number of studies have shown these methods to have inconsistent results (Mays and Cox, 2006). 5. Demographic analysis In Table 1, the distribution of the numbers and percentages of subadults versus adult males and adult females in the aggregated skeletal sample from the non-Maiden Castle cemetery and settlement sites were compared with the combined sample from Maiden Castle (three unsexed adult skeletons at Maiden Castle were assigned proportionately to the sexed adult categories). At the cemetery and settlement sites, more than half of the individuals are subadults, and there are approximately equal numbers of adult males and females. This corresponds to the expected pattern of attritional mortality in prehistoric populations which are dominated by subadult deaths (Chamberlain, 2006). In contrast, the Maiden Castle sample has a substantially smaller proportion of subadults and the majority of adults are male. The differences between the distributions of subadults and adult males and females are statistically significant (X2 = 10.31, d.f. = 2, p = 0.006). This difference is attributable to the predominance of adult males and deficit of subadult individuals in the Iron Age C and Belgic War cemetery component of the Maiden Castle sample. The samples of human remains from the Iron Age AB and Iron Age unspecified period contexts have a demographic profile that does not differ significantly from the non-Maiden Castle cemetery and settlement sites (X2 = 2.77, d.f. = 2, p = 0.25).
72
R.C. Redfern, A.T. Chamberlain / International Journal of Paleopathology 1 (2011) 68–73
Fig. 4. Frequency distribution of pubic symphysis stages in the Dorset Iron Age settlement-associated cemeteries (n = 25) compared to the Newcastle Blackgate Anglo-Saxon cemetery (n = 78) and the medieval cemetery of All Saints Fishergate in York (n = 119).
Fig. 3. (a) Frequency distributions of pubic symphysis stages in the Maiden Castle Iron Age and Belgic cemeteries compared to the Dorset Iron Age cemeteries. (b) Frequency distributions of pubic symphysis stages in the Maiden Castle Iron Age and Belgic cemeteries (n = 12, 25 respectively) compared to military mass grave assemblages from Towton (n = 22) and York (n = 69).
Twelve individuals from the Iron Age ABC groups at Maiden Castle and 25 individuals from the Belgic War cemetery had observable symphyseal faces—these two contextualised samples from Maiden Castle are analysed separately. Pubic symphysis data were also obtained from 25 adult skeletons from the seven late Iron Age attritional cemeteries. Comparable data for the pubic symphysis were collated for two large attritional cemetery samples: the AngloSaxon cemetery at Blackgate, Newcastle-upon-Tyne (Gowland and Chamberlain, 2005) and the medieval cemetery of All Saints, Fishergate, York (McIntyre and Bruce, 2010; McIntyre et al., in prep.). These cemeteries have distributions of age-at-death reflecting normal attritional mortality. Pubic symphysis data were also obtained from two military mass grave samples dating to the Wars of the Roses (Towton, Yorkshire: Boylston et al., 2000) and the English Civil War (York Barbican site: McIntyre and Bruce, 2010; McIntyre et al., in prep.). The latter two assemblages are comprised mainly of young adult males and exhibit age profiles characteristic of catastrophic patterns of mortality (Chamberlain, 2006). Fig. 3a shows that the Maiden Castle adult samples have a distinctive younger profile of age at death compared to the Dorset attritional cemeteries. Fig. 3b shows that the Maiden Castle ABC and Belgic war cemetery samples both resemble the historical-period military assemblages in that the majority of individuals are grouped in the earlier pubic symphysis phases 1–4, and very few of the individuals are in phases 5 and 6 which are associated with survival into older adulthood—it should also be noted that the Towton sample only contains males. This pattern corresponds to that expected of a catastrophic mortality profile, in which the large proportion of younger adults in the living population is directly transposed into the age distribution of the mortality profile (Gowland and Chamberlain, 2005). In contrast, Fig. 4 shows that the distribution
of pubic symphysis stages in the skeletal remains from the Dorset cemeteries closely corresponds to that found in the adults from historical-period cemeteries, in which the attritional mortality profile results in a majority of older adults whose pubic symphysis morphology is mainly in phases 4–6. The pubic symphysis distributions in the different skeletal samples were compared non-parametrically using the Mann–Whitney U test (which compares the ranks of observations across the samples) and the Kolmogorov–Smirnov Z test (which compares the shapes of the distributions between the samples). No statistically significant differences were found for comparisons amongst the catastrophic mortality samples (Maiden Castle ABC and Belgic war cemetery, Towton, York Civil War) or amongst the attritional mortality samples (Dorset Iron Age, Newcastle and York Medieval). However, all pairwise comparisons between catastrophic and attritional samples were statistically significant (p < 0.05), except for the comparisons between Towton and the Dorset Iron Age cemeteries (p = 0.09) and between Towton and York Medieval (p = 0.08). 6. Discussion and conclusions Both the overall demographic profile and the specific age pattern of adult mortality, particularly the Belgic war cemetery sample, provides demographic evidence for catastrophic mortality at the hillfort in the late Iron Age—first century BC to first century AD. These data correspond to the overwhelming bioarchaeological evidence for peri-mortem trauma, and lends further support to the hypothesis that many of the individuals buried in the eastern entrance of the hillfort had died during an episode of warfare, one of which is most likely to have included the Roman conquest (Redfern, 2011; Wheeler, 1943), and demonstrates that this sample can be used in conflict research. Our study has shown that the direct comparative analysis of the phase distribution of skeletal age indicators obviates the need to convert these indicators into actual estimates of chronological age, allowing the incorporation of data from multiple observers and making statistical testing between samples more robust. An important implication of this study is that victims of conflict can be independently identified using skeletal and paleodemographic analyses. Furthermore, these results show that episodes of conflict can be recognised in cemeteries where typical funerary strategies are employed. Acknowledgements We would like to thank the Editor, Associate Editor and Reviewers for their positive and helpful comments on this research. RR
R.C. Redfern, A.T. Chamberlain / International Journal of Paleopathology 1 (2011) 68–73
is most grateful to: Niall Sharples for all his help understanding the hillfort and the Editor of the Proceedings of the Prehistoric Society and The Society of Antiquaries (London). RR would also like to acknowledge the kind help and advice of Peter Woodward, Dorchester County Museum and Dorset Natural History and Archaeological Society; Gabby Rapson and of support by the CHB & DACA (MoL). AC acknowledges the assistance of Rebecca Gowland, who collected the skeletal age indicator data for the Blackgate Anglo-Saxon cemetery, and Lauren McIntyre, who collected the data for the York medieval and post-medieval skeletal assemblages. Parts of this research were supported by the AHRC through grants B/IA/AN866/APN13721, AH/G500142/1 and AH/H003037/1 to AC. References Armit, I., 2007. ‘Hillforts at war’: from Maiden Castle to Taniwaha Pa. Proceedings of the Prehistoric Society 73, 25–37. Billard, M., 1991. Violent traumatic injuries on human skeletal remains buried with horses in a Gallo-Roman collective grave (Lyon-Vaise, France, AD 200–300). International Journal of Osteoarchaeology 1, 259–264. Bishop, N.A., Knüsel, C.J., 2005. A paleodemographic investigation of warfare in prehistory. In: Parker Pearson, M., Thorpe, I.J.N. (Eds.), Warfare, Violence and Slavery in Prehistory. British Archaeological Reports International Series 1374. Archaeopress, Oxford, pp. 201–216. Boylston, A., Holst, M., Coughlan, J., 2000. Physical anthropology. In: Fiorato, V., Boylston, A., Knüsel, C. (Eds.), Blood Red Roses. The Archaeology of a Mass Grave from the Battle of Towton AD 1461. Oxbow Books, Oxford, pp. 45–59. Brooks, S., Suchey, J.M., 1990. Skeletal age determination based on the os pubis: a comparison of the Acsádi-Nemeskéri and Suchey-Brooks methods. Human Evolution 5, 227–238. Buikstra, J.E., Ubelaker, D.H., 1994. Standards for data collection from human skeletal remains. In: Proceedings of a Seminar at the Field Museum of Natural History organized by Jonathan Haas. Arkansas Archaeological Survey Research Series No 44 , Fayetteville, AR. Byers, S.N., 2006. Introduction to Forensic Anthropology. Allyn and Bacon, New York. Chamberlain, A., 2006. Demography in Archaeology. Cambridge University Press, Cambridge. Cunliffe, B., 1971. Some aspects of hill-forts and their cultural environments. In: Hill, D., Jesson, M. (Eds.), The Iron Age and its hill-forts. Papers presented to Sir Mortimer Wheeler on the occasion of his eightieth year at a conference held by the Southampton Archaeological Society 5th–7th March, 1971. Southampton University Archaeological Society, Southampton, pp. 53–83. Cunliffe, B., 2005. Iron Age Communities in Britain. An Account of England, Scotland and Wales from the Seventh Century BC until the Roman Conquest. Routledge, London. Gale, J., 2003. Prehistoric Dorset. Tempus Publishing Ltd., Stroud. Galloway, A. (Ed.), 1999. Broken Bones: Anthropological Analysis of Blunt Force Trauma. Charles C Thomas, Springfield, IL. Goodman, C.N., Morant, G.M., 1940. The human remains of the Iron Age and other periods from Maiden Castle, Dorset. Biometrika XXXI, 295–312. Gowland, R.L., Chamberlain, A.T., 2005. Detecting plague: palaeodemographic characterisation of a catastrophic death assemblage. Antiquity 79, 146–157. Hamlin, C., 2007. The material expression of social change: the mortuary correlates of gender and age in late Pre-Roman Iron Age and Roman Dorset. Unpublished Ph.D. Dissertation, University of Wisconsin-Milwaukee.
73
Hawkes, C.F.C., 1939. The ABC of the British Iron Age. Antiquity 33, 170–182. Johansen, V.A., Wahl, A.K., Weisaeth, L., 2008. Assaulted victims of nondomestic violence in Norway—injury, crime characteristics and emotions during the assault. Scandinavian Journal of Caring Sciences 22, 445–454. Knüsel, C.J., 2005. The physical evidence for warfare: a subtle stigmata? In: Parker Pearson, M., Thorpe, I.J.N. (Eds.), Warfare, Violence and Slavery in Prehistory. British Archaeological Reports International Series 1374. Archaeopress, Oxford, pp. 49–66. Lewis, A.B., Garn, S.M., 1960. The relationship between tooth formation and other maturational factors. The Angle Orthodontist 30, 70–77. Mays, S., Cox, M., 2006. Sex determination in skeletal remains. In: Cox, M., Mays, S. (Eds.), Human Osteology in Archaeology and Forensic Science. Cambridge University Press, Cambridge, pp. 117–130. McIntyre, L., Bruce, G., 2010. Excavating All Saint’s, a medieval church rediscovered. Current Archaeology 245, 30–37. Mohanty, M.K., Panigrahi, M.K., Mohanty, S., Dash, J.K., Dash, S.K., 2007. Selfdefense injuries in homicidal deaths. Journal of Forensic Legal Medicine 14, 213–215. Morant, G.M., Goodman, C., 1943. Human bones. In: Wheeler, R.E.M. (Ed.), Maiden Castle, Dorset. Research Report for the Committee of the Society of Antiquaries Number 12, London. pp. 337–360. Redfern, R.C., 2007. A bioarchaeological analysis of violence in Iron Age females: a perspective from Dorset England (mid to late seventh century BC to the first century AD). In: Davis, O.P., Sharples, N.M., Waddington, K.E. (Eds.), Changing Perspectives on the First Millennium B.C. Oxbow Books, Oxford, pp. 139–160. Redfern, R.C., 2009. Does cranial trauma provide evidence for projectile weaponry in late Iron Age Dorset? Oxford Journal of Archaeology 28.4, 399–424. Redfern, R.C. 2011, A re-appraisal of the evidence for violence in the late Iron Age human remains from Maiden Castle hillfort, Dorset, England. Proceedings of the Prehistoric Society 77, in press. Renfrew, C., Bahn, P., 2008. Archaeology. Theories, Methods and Practice. Thames and Hudson, London. Scheuer, L., Black, S., 2000. Developmental Juvenile Osteology. Academic Press, London. Schmidt, U., Pollak, S., 2006. Sharp force injuries in clinical forensic medicine: findings in victims and perpetrators. Forensic Science International 159, 113– 118. Sharples, N., 1991a. English Heritage Book of Maiden Castle. BT Batsford Ltd., London. Sharples, N., 1991b. Maiden Castle. Excavations and Field Survey 1985–6. English Heritage Archaeological Reports No 19. English Heritage, London. Sharples, N., 1991c. Warfare in the Iron Age of Wessex. Scottish Archaeological Review 8, 79–89. Sharples, N., 2010. Social Relations in Later Prehistory. Wessex in the First Millennium BC. Oxford University Press, Oxford. Shepherd, J.P., Shapland, M., Pearce, N.X., Scully, C., 1990. Pattern, severity and aetiology of injuries in victims of assault. Journal of the Royal Society of Medicine 83, 75–78. Thorpe, I.J.N., 2003. Anthropology, archaeology and the origin of warfare. World Archaeology 35, 145–165. Ubelaker, D.H., 1989. Human Skeletal Remains: Excavation, Analysis and Interpretation. Taraxacum, Washington. Ubelaker, D.H., Adams, B.J., 1995. Differentiation of perimortem and postmortem trauma using taphonomic indicators. Journal of Forensic Sciences 40, 509–512. Wainwright, G.J., Cunliffe, B.W., 1985. Maiden Castle: excavation, education, entertainment? Antiquity 59, 97–100. Wheeler, R.E.M., 1943. Maiden Castle, Dorset. Research Report for the Committee of the Society of Antiquaries London Number 12, London. Wheeler, R.E.M., 1955. Still Digging. Michael Joseph Ltd., London.
Contents lists available at ScienceDirect
International Journal of Paleopathology journal homepage: www.elsevier.com/locate/ijpp
Case Study
A demographic analysis of Maiden Castle hillfort: Evidence for conflict in the late Iron Age and early Roman period Rebecca C. Redfern a,∗ , Andrew T. Chamberlain b a b
Centre for Human Bioarchaeology, Department of Archaeological Collections and Archive, Museum of London, 150 London Wall, ECY 5HN, England, United Kingdom Department of Archaeology, University of Sheffield, Northgate House, West Street, Sheffield S1 4ET, England, United Kingdom
a r t i c l e
i n f o
Article history: Received 13 December 2010 Received in revised form 19 February 2011 Accepted 21 February 2011 Keywords: Demography Mortality Warfare Roman Empire Late Iron Age Britain
a b s t r a c t The late Iron Age human remains from the British hillfort of Maiden Castle are frequently cited within the archaeological and bioarchaeological literature as providing evidence for conflict. This interpretation is based on osteological work undertaken in the late 1930s. In order to test the validity of using this sample in conflict research, the authors undertook a detailed analysis of the site’s demography in comparison with contemporary late Iron Age attritional cemeteries from Dorset (England) and additional conflict mortality data. These results showed that the 1st century BC to 1st century AD samples conformed to a catastrophic profile, as greater numbers of young adult males had been buried during this phase of occupation. In combination with new bioarchaeological findings and the identification of an embedded Roman projectile weapon, we conclude that individuals had died during an episode of warfare, one of which probably included the Roman conquest of 43AD. © 2011 Elsevier Inc. All rights reserved.
1. Introduction
2. Archaeology of the Iron Age hillfort
Skeletal evidence for violent trauma provides one of the most important categories of evidence in the study of ancient warfare (Thorpe, 2003). Maiden Castle hillfort in Dorset, England, is frequently cited in archaeological literature as providing evidence for warfare (Armit, 2007; Bishop and Knüsel, 2005; Renfrew and Bahn, 2008; Sharples, 1991a,b,c) and potentially, the Roman invasion of Britain in AD 43 (Redfern, 2011; Sharples, 2010). This interpretation is primarily based on the osteological work undertaken by Morant and Goodman in the late 1930s, which identified instances of perimortem violent trauma, and Wheeler’s (1943) reconstruction of events at the hillfort. In order to test the validity of using this sample in conflict research, the authors examined the site’s demography and undertook a detailed analysis of adult age-at-death using the pubic symphysis and compared these results to contemporary late Iron Age attritional cemeteries from Dorset and additional conflict mortality data.
Maiden Castle hillfort (Fig. 1) is located 1 km from the modern County town of Dorchester. In the late Iron Age (100 BC to AD 100), based on pottery distributions and the minting of coinage, this area of southwest England was inhabited by the Durotriges tribe which was formed by a confederacy of smaller communities (Gale, 2003, 98). Within the tribal territory a number of hillforts have been identified, and similarly to other areas of southern Britain, these appear to have fulfilled a variety of functions beyond purely defensive purposes, including domestic settlements and the production of material culture (Armit, 2007; Cunliffe, 1971, 2005; Gale, 2003, 108–117; Sharples, 2010). For the Durotriges, Maiden Castle appears to have been an important centre of power and social organisation and under-went several phases of development, including the construction of defensive earthworks and the complex gateways at the eastern and western ends of the hillfort. Archaeological evidence also shows that it was used for settlement, with the identification of roundhouses and middens; by the late Iron Age, occupation also existed in the outworks of the eastern entrance and this area was also used for metalworking (Sharples, 1991b, 2010). Sharples (1991a, 123–124) considers that these changes are indicative of wider changes within the Durotriges’ territory, particularly the emergence of centralised power in the region, and the expansion of ceramic, salt, shale, stone and metal working industries.
∗ Corresponding author. Tel.: +44 20 7814 5649. E-mail addresses: rredfern@museumoflondon.org.uk (R.C. Redfern), [email protected] (A.T. Chamberlain). 1879-9817/$ – see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.ijpp.2011.02.004
R.C. Redfern, A.T. Chamberlain / International Journal of Paleopathology 1 (2011) 68–73
69
Fig. 1. Location map of Maiden Castle, Dorset. The figure was drawn by G. Rapson.
3. Interpretation of the Iron Age hillfort As outlined in 2, hillforts were used for many different purposes during the late Iron Age. However, at Maiden Castle, the most popular and long-standing interpretation has been its defensive role in warfare, particularly during the AD 43 invasion of Britain by the Roman Emperor Claudius. The first to promote this hypothesis was Sir Mortimer Wheeler (1943, 1955), who following his excavations at the hillfort in the 1930s, envisioned a vicious, bloody battle between the late Iron Age inhabitants of the fort and the Roman legions commanded by Vespasian. Wheeler (1943) considered that the evidence for burning at the hillfort’s complex eastern entrance, the vast numbers of sling-stones recovered from the entrance, and the skeletons with evidence of fractures and weapon trauma provided unequivocal evidence of this battle, an interpretation which has remained popular amongst the public and a number of scholars (see Wainwright and Cunliffe, 1985). Wheeler (1943) assigned the burial and occupation phases of the pre-Roman hillfort into four groups in accordance with Hawkes’ (1939) cultural (‘ABC’) divisions of the British Iron Age. Wheeler’s phases were: Iron Age A and B including burials in pits/features within the hillfort and not directly associated with any gravegoods; Iron Age C, burials in the ramparts and ditches of the eastern entrance, though only a small minority of this group could be directly associated with any grave-goods; and the contemporary
Belgic war cemetery (first century BC to first century AD)—the latter included interments found behind the hillfort’s eastern entrance and were buried with grave-goods. Fig. 2 illustrates the location of the burials in the eastern entrance; location information for all the other burials are described by Wheeler (1943). Re-examination of the archaeological, funerary and material culture evidence shows that these groupings artificially divide the sample (Hamlin, 2007; Sharples, 1991b). Based on contemporary phasing of this period: the Iron Age A group (300–100 BC) dates from the middle Iron Age, and groups B, C and the Belgic war cemetery date to the late Iron Age (first century BC to first century AD); unfortunately, it has not been possible to verify these archaeological phases with direct carbon 14 dating. Therefore, based on a more accurate phasing of the hillfort, Sharples’ (1991a,b, 2010, pers. comm.) proposes that the eastern entrance of the hillfort was used for burial over many hundreds of years and that these graves should be considered to reflect a single burial area. Sharples’ (1991a,b) further revised several aspects of Wheeler’s (1943) previous interpretation through the close study of the data collected during his and Wheeler’s excavations at the site. Firstly, Wheeler (1943, 62) argues that the burials were, ‘carried out anxiously and hastily without order’. A closer examination of the funerary evidence shows that the burials follow standard Durotrigian funerary practices, with respect to body positioning and selection of grave goods and were inhumed in an existing cemetery
70
R.C. Redfern, A.T. Chamberlain / International Journal of Paleopathology 1 (2011) 68–73
Fig. 2. Location of burials (by period) in the eastern entrance (after Sharples, 1991a, Fig 90. Redrawn by Rapson).
(see also Hamlin, 2007), which argues for a lengthy funerary ritual; the ash, interpreted by Wheeler (1943) as evidence of burning, is most likely to have been created by late Iron Age metal-working in the eastern entrance; the arrows identified by Wheeler probably pre-date the Roman conquest, but the slighting of the defences probably did occur after the conquest (Sharples, 1991a, 124–125). These data, combined with the osteological evidence presented by Morant and Goodman (1943) in which about one quarter of their sample had evidence for trauma, was not considered persuasive enough by Sharples (1991a) to support Wheeler’s interpretation (see Section 3.1). Recently, the first author has collaborated with Sharples and Gilliver to re-analyse two males buried in a double grave in the Belgic war cemetery. Examination of the embedded projectile weapon in the vertebra of P7A (Wheeler, 1943, Plate LVIIIA) has identified this to be a Roman ballista bolt (Gilliver, pers. comm.)—confirming Wheeler’s (1943, 63) original identification, and the square perforation in the left temporal bone of P7 (see Redfern, 2009) is most likely to have been caused by a Roman weapon (Gilliver, pers. comm.), as the only parallel that could be identified in late Iron Age and early Roman samples was from a Gallo-Roman grave dating from AD 200 to 300 (Billard, 1991). Although the burials at Maiden Castle hillfort are comparable to contemporaneous cemeteries in terms of funerary practices, it is the only cemetery of this period in Dorset to contain the multiple burials of adults (Hamlin, 2007). Six double and one triple burial were identified in the Belgic war cemetery area (Morant and Goodman, 1943, 352–355), all containing individuals with peri-mortem trauma (Redfern, 2011). This evidence again supports Sharples’ (1991a,b, 2010) hypothesis that the Belgic war cemetery burials were part of, or associated with, an existing burial ground (see Fig. 2).
3.1. Bioarchaeological analysis of the hillfort’s Iron Age samples The first author used modern osteological methods to study the extant Wheeler sample (see Byers, 2006; Galloway, 1999; Knüsel, 2005; Johansen et al., 2008; Schmidt and Pollak, 2006; Shepherd et al., 1990; Ubelaker and Adams, 1995 for reviews of current methods). This established that the crude prevalence of individuals with peri- and ante-mortem injuries was far higher (74.2%) (Redfern, 2011) than previously reported by Morant and Goodman (1943) (27.7%). The Iron Age A group did not have any evidence of trauma present, the Iron Age B group had 54.4%, and the highest prevalence rates were observed in the Iron Age C group (75%) and the Belgic war cemetery (87.5%). The difference in the number of individuals with trauma between the latter two burial-groups was not quite statistically significant (X2 = 3.575, d.f. = 1, p = 0.0586) (Redfern, 2011). The re-analysis identified individuals of both sexes (male N = 30, female N = 24 data from Redfern, 2011) and one 15-year-old individual with over-whelming peri-mortem evidence for targeted blows to the head and body, assault injuries and over-kill, as evidenced by the presence of multiple trauma, sharp-force weapon injuries to the head, torso, upper and lower limbs, and extremities, sharp and blunt projectile injuries to the cranium, fractures caused by assault to the head and limbs, and sharp-force weapon injuries consistent with attempted decapitation. Examination of the distribution of fractures and injury patterns in males and females shows a considerable degree of concordance, with both sexes sustaining peri-mortem assault injuries to the face, torso and forearms (e.g. parry fracture), and sharp-force and projectile injuries to the skull suggesting that aggressors did not differentiate between the sexes. Additionally, many individuals of both sexes displayed healed antemortem trauma created by direct blows to the body, projectile and
R.C. Redfern, A.T. Chamberlain / International Journal of Paleopathology 1 (2011) 68–73 Table 1 Comparison of numbers and percentages of individuals in age and sex categories in the aggregated sample from the Iron Age settlement sites and the combined sample from the Maiden Castle cemeteries.
Settlement sites Maiden Castle
Subadults
Adult males
Adult females
48 (55%) 23 (31%)
23 (26%) 35 (45%)
17 (19%) 17 (24%)
sharp-force weapon injuries, suggesting that they had survived previous episodes of interpersonal violence. These findings are the first to demonstrate that, in addition to males, females and an adolescent were exposed to episodes of violence during this period. Nevertheless, it should be noted that the extant evidence from Dorset for female and adolescent involvement has only been identified in this sample, underscoring its uniqueness (see Redfern, 2007, 2011). Focusing on the peri-mortem evidence, the range and severity of the trauma observed in the sample can be illustrated by a 30–45 year male, who had sustained two fractures to his mandible, which also shows sharp-force weapon injury evidence for decapitation. In addition to which, he also suffered sharp and blunt injuries to his cranium, and fractures to the ribs and bilaterally to the femora. These injuries show that they had been sustained during a violent assault (Mohanty et al., 2007; Schmidt and Pollak, 2006). 4. Materials and methods The extant sample differed slightly from Wheeler’s (1943) published inventory (N = 83). Study of the published report, site notebooks and documentation, and Morant and Goodman’s notes showed that two individuals were no longer present for study from the late Iron Age samples excavated from the eastern gateway, identified by Wheeler as the Iron Age C and Belgic War cemetery samples. Twenty-one individuals whose context information recorded in the site notebooks showed them to be of Iron Age date but who could not be tied to a burial context were included in this study in order to provide a true representation of the extant sample (cf Redfern, 2011). It should be noted that Wheeler’s (1943) burial catalogue is incomplete and in the site report, Morant and Goodman (1943, 357) state that they are only able to provide data for 83 skeletons who could give ‘records of value’ (see also Goodman and Morant, 1940). As this present study focuses on the demographic composition of the sample, not all of the extant sample could be included because the necessary areas of the skeleton were not present. Therefore, a total of 23 subadults for whom age-at-death could be established from dental and/or skeletal methodologies and 52 adult individuals who had observable pubic symphyseal faces were used (Table 1). The inter-cemetery comparison data were collected by the first author and are derived from seven late Iron Age (1st century BC to 1st century AD) attritional burial locations in the county of Dorset, Kimmeridge, Tarrant Hinton, Flagstones, Alington Avenue, Gussage All Saints, Whitcombe Farm and Poundbury Camp (Table 2). Adult individuals from these cemeteries were only included in this study if one or both pubic symphyseal faces could be scored for age-estimation. 4.1. Sample used for analysis As stated above, this study used 23 subadults and 52 adults from the total sample examined by the first author. All individuals were recorded using the protocols published by Buikstra and Ubelaker (1994); with the sex of adult individuals (≥20 years old) determined using the morphology of the skull and pelvis following the methods described by Buikstra and Ubelaker (1994). As this study
71
Table 2 Late Iron Age comparison data: site type and location. Site name
Type
Location
Alington Avenue
Formal cemetery
Flagstones
Formal cemetery
Gussage All Saints Kimmeridge
Settlement Burial associated with a cemetery Cemetery associated with a hillfort Burials associated with a settlement Formal cemetery
South-east of Dorchester South-west of Dorchester North-east Dorset South Dorset
Poundbury Camp Tarrant Hinton Whitcombe Farm
North-west of Dorchester North-east Dorset South-east Dorset
focuses on the demographic composition of the assemblage, we have employed the pubic symphysis method of age determination (Brooks and Suchey, 1990). In order to enable direct comparison of demographic profiles, we decided to analyse the frequency distributions of phases of the pubic symphysis, a commonly used adult age indicator for which recording standards are widely available and consistently applied. Throughout adult life, the medial surface of the pubic symphysis undergoes progressive, irreversible morphological changes involving loss of transverse ridging, thickening of the joint margins and the development of pitting and unevenness of the joint surface. In the recording system devised by Brooks and Suchey (1990) six ordinal phases are defined, with Phase 1 characterising the youngest adults and Phase 6 the oldest. Each adult individual was scored into one of the six phases and, in individuals where the left and right side pubic symphysis exhibited different phases, the scores were averaged. The biological ages of subadults (<20 years old) were determined using a combination of dental development (Ubelaker, 1989), diaphyseal length (Scheuer and Black, 2000), and epiphyseal fusion methodologies (Scheuer and Black, 2000). When both dental and skeletal estimations were available, the dental age was used, as this has the most reliable correlation to chronological age (Lewis and Garn, 1960). No attempt was made to determine sex of subadults based upon skeletal morphology, as a number of studies have shown these methods to have inconsistent results (Mays and Cox, 2006). 5. Demographic analysis In Table 1, the distribution of the numbers and percentages of subadults versus adult males and adult females in the aggregated skeletal sample from the non-Maiden Castle cemetery and settlement sites were compared with the combined sample from Maiden Castle (three unsexed adult skeletons at Maiden Castle were assigned proportionately to the sexed adult categories). At the cemetery and settlement sites, more than half of the individuals are subadults, and there are approximately equal numbers of adult males and females. This corresponds to the expected pattern of attritional mortality in prehistoric populations which are dominated by subadult deaths (Chamberlain, 2006). In contrast, the Maiden Castle sample has a substantially smaller proportion of subadults and the majority of adults are male. The differences between the distributions of subadults and adult males and females are statistically significant (X2 = 10.31, d.f. = 2, p = 0.006). This difference is attributable to the predominance of adult males and deficit of subadult individuals in the Iron Age C and Belgic War cemetery component of the Maiden Castle sample. The samples of human remains from the Iron Age AB and Iron Age unspecified period contexts have a demographic profile that does not differ significantly from the non-Maiden Castle cemetery and settlement sites (X2 = 2.77, d.f. = 2, p = 0.25).
72
R.C. Redfern, A.T. Chamberlain / International Journal of Paleopathology 1 (2011) 68–73
Fig. 4. Frequency distribution of pubic symphysis stages in the Dorset Iron Age settlement-associated cemeteries (n = 25) compared to the Newcastle Blackgate Anglo-Saxon cemetery (n = 78) and the medieval cemetery of All Saints Fishergate in York (n = 119).
Fig. 3. (a) Frequency distributions of pubic symphysis stages in the Maiden Castle Iron Age and Belgic cemeteries compared to the Dorset Iron Age cemeteries. (b) Frequency distributions of pubic symphysis stages in the Maiden Castle Iron Age and Belgic cemeteries (n = 12, 25 respectively) compared to military mass grave assemblages from Towton (n = 22) and York (n = 69).
Twelve individuals from the Iron Age ABC groups at Maiden Castle and 25 individuals from the Belgic War cemetery had observable symphyseal faces—these two contextualised samples from Maiden Castle are analysed separately. Pubic symphysis data were also obtained from 25 adult skeletons from the seven late Iron Age attritional cemeteries. Comparable data for the pubic symphysis were collated for two large attritional cemetery samples: the AngloSaxon cemetery at Blackgate, Newcastle-upon-Tyne (Gowland and Chamberlain, 2005) and the medieval cemetery of All Saints, Fishergate, York (McIntyre and Bruce, 2010; McIntyre et al., in prep.). These cemeteries have distributions of age-at-death reflecting normal attritional mortality. Pubic symphysis data were also obtained from two military mass grave samples dating to the Wars of the Roses (Towton, Yorkshire: Boylston et al., 2000) and the English Civil War (York Barbican site: McIntyre and Bruce, 2010; McIntyre et al., in prep.). The latter two assemblages are comprised mainly of young adult males and exhibit age profiles characteristic of catastrophic patterns of mortality (Chamberlain, 2006). Fig. 3a shows that the Maiden Castle adult samples have a distinctive younger profile of age at death compared to the Dorset attritional cemeteries. Fig. 3b shows that the Maiden Castle ABC and Belgic war cemetery samples both resemble the historical-period military assemblages in that the majority of individuals are grouped in the earlier pubic symphysis phases 1–4, and very few of the individuals are in phases 5 and 6 which are associated with survival into older adulthood—it should also be noted that the Towton sample only contains males. This pattern corresponds to that expected of a catastrophic mortality profile, in which the large proportion of younger adults in the living population is directly transposed into the age distribution of the mortality profile (Gowland and Chamberlain, 2005). In contrast, Fig. 4 shows that the distribution
of pubic symphysis stages in the skeletal remains from the Dorset cemeteries closely corresponds to that found in the adults from historical-period cemeteries, in which the attritional mortality profile results in a majority of older adults whose pubic symphysis morphology is mainly in phases 4–6. The pubic symphysis distributions in the different skeletal samples were compared non-parametrically using the Mann–Whitney U test (which compares the ranks of observations across the samples) and the Kolmogorov–Smirnov Z test (which compares the shapes of the distributions between the samples). No statistically significant differences were found for comparisons amongst the catastrophic mortality samples (Maiden Castle ABC and Belgic war cemetery, Towton, York Civil War) or amongst the attritional mortality samples (Dorset Iron Age, Newcastle and York Medieval). However, all pairwise comparisons between catastrophic and attritional samples were statistically significant (p < 0.05), except for the comparisons between Towton and the Dorset Iron Age cemeteries (p = 0.09) and between Towton and York Medieval (p = 0.08). 6. Discussion and conclusions Both the overall demographic profile and the specific age pattern of adult mortality, particularly the Belgic war cemetery sample, provides demographic evidence for catastrophic mortality at the hillfort in the late Iron Age—first century BC to first century AD. These data correspond to the overwhelming bioarchaeological evidence for peri-mortem trauma, and lends further support to the hypothesis that many of the individuals buried in the eastern entrance of the hillfort had died during an episode of warfare, one of which is most likely to have included the Roman conquest (Redfern, 2011; Wheeler, 1943), and demonstrates that this sample can be used in conflict research. Our study has shown that the direct comparative analysis of the phase distribution of skeletal age indicators obviates the need to convert these indicators into actual estimates of chronological age, allowing the incorporation of data from multiple observers and making statistical testing between samples more robust. An important implication of this study is that victims of conflict can be independently identified using skeletal and paleodemographic analyses. Furthermore, these results show that episodes of conflict can be recognised in cemeteries where typical funerary strategies are employed. Acknowledgements We would like to thank the Editor, Associate Editor and Reviewers for their positive and helpful comments on this research. RR
R.C. Redfern, A.T. Chamberlain / International Journal of Paleopathology 1 (2011) 68–73
is most grateful to: Niall Sharples for all his help understanding the hillfort and the Editor of the Proceedings of the Prehistoric Society and The Society of Antiquaries (London). RR would also like to acknowledge the kind help and advice of Peter Woodward, Dorchester County Museum and Dorset Natural History and Archaeological Society; Gabby Rapson and of support by the CHB & DACA (MoL). AC acknowledges the assistance of Rebecca Gowland, who collected the skeletal age indicator data for the Blackgate Anglo-Saxon cemetery, and Lauren McIntyre, who collected the data for the York medieval and post-medieval skeletal assemblages. Parts of this research were supported by the AHRC through grants B/IA/AN866/APN13721, AH/G500142/1 and AH/H003037/1 to AC. References Armit, I., 2007. ‘Hillforts at war’: from Maiden Castle to Taniwaha Pa. Proceedings of the Prehistoric Society 73, 25–37. Billard, M., 1991. Violent traumatic injuries on human skeletal remains buried with horses in a Gallo-Roman collective grave (Lyon-Vaise, France, AD 200–300). International Journal of Osteoarchaeology 1, 259–264. Bishop, N.A., Knüsel, C.J., 2005. A paleodemographic investigation of warfare in prehistory. In: Parker Pearson, M., Thorpe, I.J.N. (Eds.), Warfare, Violence and Slavery in Prehistory. British Archaeological Reports International Series 1374. Archaeopress, Oxford, pp. 201–216. Boylston, A., Holst, M., Coughlan, J., 2000. Physical anthropology. In: Fiorato, V., Boylston, A., Knüsel, C. (Eds.), Blood Red Roses. The Archaeology of a Mass Grave from the Battle of Towton AD 1461. Oxbow Books, Oxford, pp. 45–59. Brooks, S., Suchey, J.M., 1990. Skeletal age determination based on the os pubis: a comparison of the Acsádi-Nemeskéri and Suchey-Brooks methods. Human Evolution 5, 227–238. Buikstra, J.E., Ubelaker, D.H., 1994. Standards for data collection from human skeletal remains. In: Proceedings of a Seminar at the Field Museum of Natural History organized by Jonathan Haas. Arkansas Archaeological Survey Research Series No 44 , Fayetteville, AR. Byers, S.N., 2006. Introduction to Forensic Anthropology. Allyn and Bacon, New York. Chamberlain, A., 2006. Demography in Archaeology. Cambridge University Press, Cambridge. Cunliffe, B., 1971. Some aspects of hill-forts and their cultural environments. In: Hill, D., Jesson, M. (Eds.), The Iron Age and its hill-forts. Papers presented to Sir Mortimer Wheeler on the occasion of his eightieth year at a conference held by the Southampton Archaeological Society 5th–7th March, 1971. Southampton University Archaeological Society, Southampton, pp. 53–83. Cunliffe, B., 2005. Iron Age Communities in Britain. An Account of England, Scotland and Wales from the Seventh Century BC until the Roman Conquest. Routledge, London. Gale, J., 2003. Prehistoric Dorset. Tempus Publishing Ltd., Stroud. Galloway, A. (Ed.), 1999. Broken Bones: Anthropological Analysis of Blunt Force Trauma. Charles C Thomas, Springfield, IL. Goodman, C.N., Morant, G.M., 1940. The human remains of the Iron Age and other periods from Maiden Castle, Dorset. Biometrika XXXI, 295–312. Gowland, R.L., Chamberlain, A.T., 2005. Detecting plague: palaeodemographic characterisation of a catastrophic death assemblage. Antiquity 79, 146–157. Hamlin, C., 2007. The material expression of social change: the mortuary correlates of gender and age in late Pre-Roman Iron Age and Roman Dorset. Unpublished Ph.D. Dissertation, University of Wisconsin-Milwaukee.
73
Hawkes, C.F.C., 1939. The ABC of the British Iron Age. Antiquity 33, 170–182. Johansen, V.A., Wahl, A.K., Weisaeth, L., 2008. Assaulted victims of nondomestic violence in Norway—injury, crime characteristics and emotions during the assault. Scandinavian Journal of Caring Sciences 22, 445–454. Knüsel, C.J., 2005. The physical evidence for warfare: a subtle stigmata? In: Parker Pearson, M., Thorpe, I.J.N. (Eds.), Warfare, Violence and Slavery in Prehistory. British Archaeological Reports International Series 1374. Archaeopress, Oxford, pp. 49–66. Lewis, A.B., Garn, S.M., 1960. The relationship between tooth formation and other maturational factors. The Angle Orthodontist 30, 70–77. Mays, S., Cox, M., 2006. Sex determination in skeletal remains. In: Cox, M., Mays, S. (Eds.), Human Osteology in Archaeology and Forensic Science. Cambridge University Press, Cambridge, pp. 117–130. McIntyre, L., Bruce, G., 2010. Excavating All Saint’s, a medieval church rediscovered. Current Archaeology 245, 30–37. Mohanty, M.K., Panigrahi, M.K., Mohanty, S., Dash, J.K., Dash, S.K., 2007. Selfdefense injuries in homicidal deaths. Journal of Forensic Legal Medicine 14, 213–215. Morant, G.M., Goodman, C., 1943. Human bones. In: Wheeler, R.E.M. (Ed.), Maiden Castle, Dorset. Research Report for the Committee of the Society of Antiquaries Number 12, London. pp. 337–360. Redfern, R.C., 2007. A bioarchaeological analysis of violence in Iron Age females: a perspective from Dorset England (mid to late seventh century BC to the first century AD). In: Davis, O.P., Sharples, N.M., Waddington, K.E. (Eds.), Changing Perspectives on the First Millennium B.C. Oxbow Books, Oxford, pp. 139–160. Redfern, R.C., 2009. Does cranial trauma provide evidence for projectile weaponry in late Iron Age Dorset? Oxford Journal of Archaeology 28.4, 399–424. Redfern, R.C. 2011, A re-appraisal of the evidence for violence in the late Iron Age human remains from Maiden Castle hillfort, Dorset, England. Proceedings of the Prehistoric Society 77, in press. Renfrew, C., Bahn, P., 2008. Archaeology. Theories, Methods and Practice. Thames and Hudson, London. Scheuer, L., Black, S., 2000. Developmental Juvenile Osteology. Academic Press, London. Schmidt, U., Pollak, S., 2006. Sharp force injuries in clinical forensic medicine: findings in victims and perpetrators. Forensic Science International 159, 113– 118. Sharples, N., 1991a. English Heritage Book of Maiden Castle. BT Batsford Ltd., London. Sharples, N., 1991b. Maiden Castle. Excavations and Field Survey 1985–6. English Heritage Archaeological Reports No 19. English Heritage, London. Sharples, N., 1991c. Warfare in the Iron Age of Wessex. Scottish Archaeological Review 8, 79–89. Sharples, N., 2010. Social Relations in Later Prehistory. Wessex in the First Millennium BC. Oxford University Press, Oxford. Shepherd, J.P., Shapland, M., Pearce, N.X., Scully, C., 1990. Pattern, severity and aetiology of injuries in victims of assault. Journal of the Royal Society of Medicine 83, 75–78. Thorpe, I.J.N., 2003. Anthropology, archaeology and the origin of warfare. World Archaeology 35, 145–165. Ubelaker, D.H., 1989. Human Skeletal Remains: Excavation, Analysis and Interpretation. Taraxacum, Washington. Ubelaker, D.H., Adams, B.J., 1995. Differentiation of perimortem and postmortem trauma using taphonomic indicators. Journal of Forensic Sciences 40, 509–512. Wainwright, G.J., Cunliffe, B.W., 1985. Maiden Castle: excavation, education, entertainment? Antiquity 59, 97–100. Wheeler, R.E.M., 1943. Maiden Castle, Dorset. Research Report for the Committee of the Society of Antiquaries London Number 12, London. Wheeler, R.E.M., 1955. Still Digging. Michael Joseph Ltd., London.