AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 136:93–99 (2008) Investigating Health at Kerma: Sacrificial Versus Nonsacrificial Individuals Michele R. Buzon1* and Margaret A. Judd2 1 2 Department of Sociology and Anthropology, Purdue University, West Lafayette IN 47907-2059 Department of Anthropology, University of Pittsburgh, Pittsburgh, PA 15260 KEY WORDS Nubia; stress; osteological paradox; bioarchaeology; Sudan ABSTRACT This analysis examines heterogeneity in risks by assessing the health status of individuals in two distinct burial contexts from the Nubian site of Kerma: sacrificial (n 5 100) and nonsacrificial (n 5 190) burial areas dated to the classic Kerma period (1750–1500 BC). Indicators of physiological stress that were examined include cribra orbitalia, dental enamel hypoplasia, tibial osteoperiostitis, and femur length. The analysis presented here shows that the people interred in the sacrificial and nonsacrificial burial contexts at Kerma in Upper Nubia had similar health pro- files that were comparable with other contemporaneous samples from the region. If sacrificial individuals did not experience the same risk of death as nonsacrificial individuals, it was not evident in the frequencies of nonspecific stress indicators. However, this differential risk of death may be blurred by our inability to examine nonadults for childhood disease. This research demonstrates the complexities involved in understanding the multiple factors that result in heterogeneity in skeletal samples. Am J Phys Anthropol 136:93–99, 2008. V 2008 Wiley-Liss, Inc. In recent years, bioarchaeologists have addressed the concerns raised by the ‘‘osteological paradox’’ (Wood et al., 1992; Wright and Yoder, 2003). A central issue is that individuals may vary dramatically in their ‘‘frailty’’ or susceptibility to illness due to a ‘‘hidden heterogeneity’’ in risks to disease and death. The difficulties in identifying heterogeneity in risks using archaeological and historical sources make interpreting paleopathological data challenging as it is not always possible to recognize differences in susceptibility due to various genetic, biological, environmental, and/or cultural factors. This article examines the varying risks of death by assessing the health status of individuals in light of a particular factor that may have affected variation in the skeletal sample: burial in sacrificial or nonsacrificial contexts at the Nubian site of Kerma. The unique context of the Kerma cemetery provides an excellent opportunity to explore how health profiles may have varied in relation to social factors as reflected in the burials as well as to test the idea of the reality of archaeologically-defined subgroups (sacrificial and nonsacrificial). The purpose of this article is to document and interpret health profiles of sacrificial and nonsacrificial individuals at Kerma. Specifically, we address the hypothesis that the sacrificial individuals would have been healthier because they did not die naturally. It is predicted that better health would be reflected by a lower frequency of stress indicators in sacrificial individuals. Using a bioarchaeological approach (Buzon et al., 2005), skeletal data are combined with archaeological and historical information to explore heterogeneity in the Kerma skeletal sample. Located just south of the Nile’s Third Cataract, this city held a strategic position for monitoring trade along the Nile between Egypt, Central Africa, and the Red Sea (Adams, 1977). Archaeological and Egyptian historical records document a long history of contact and conflict between Nubia and Egypt and although Kerma fought hard, it was eventually annihilated in 1520 BC by Pharaoh Thutmoses I of Egypt (Adams, 1977; O’Connor, 1993; Watterson, 1997; Morkot, 2000; Edwards, 2004). The power of Kerma and its rulers was manifest in the domestic and funerary architecture and ritual during the classic Kerma period (1750–1500 BC) (Reisner 1923a; Bonnet, 1990, 1994, 2004). The Kerma cemetery emulated the town site situated 3 km west along the Nile river. A large, mud brick temple (defuffa) dominated each landscape. The defuffa towered over domestic structures, storage pits, a central palace, a bronze foundry, and cattle pens, while the cemetery temple presided over seven massive circular tumuli (grave mounds) that were each cradled by thousands of cattle skulls and surrounded by acres of smaller tumuli. The first excavation at Kerma was conducted by Reisner (1923a,b) in 1907 and focused on the seven large tumuli that ranged in KERMA ARCHEOLOGICAL CONTEXT The ancient Nubian site of Kerma, situated in modern Sudan, was the location of the earliest state power that dominated Upper Nubia (Fig. 1) from 2500 to 1500 BC. C 2008 V WILEY-LISS, INC. C Grant sponsor: National Science Foundation Dissertation Improvement; Grant numbers: 0313247. Grant sponsor: Social Sciences and Humanities Research Council of Canada; Grant number: 752-96-1319. *Correspondence to: Michele R. Buzon, Department of Sociology and Anthropology, Purdue University, 700 W. State St., West Lafayette IN 47907-2059, USA. E-mail: mbuzon@purdue.edu Received 11 June 2007; accepted 20 November 2007 DOI 10.1002/ajpa.20781 Published online 10 January 2008 in Wiley InterScience (www.interscience.wiley.com). 94 M.R. BUZON AND M.A. JUDD ficial groups with no perimortem trauma or patterns of repetitive physical abuse for either group (Judd, 2002; Judd et al., 2006). However, it is possible that the means by which the sacrificial individuals died, such as suffocation or poison, did not leave any marks on the skeleton. Second, the task of correlating each skeleton with Reisner’s burial context was daunting as the context assigned to and recorded on each individual varied with the whim of each original recorder’s placement of the letters and numbers that made up Reisner’s recording scheme. Once the correct contexts were determined, it was found that the skeletal sample interred within the corridor was not composed primarily of young females as Reisner reported, but consisted of males and females of all ages, as well as some children (Judd, 2002). Finally, the biological and ethnic identity of these individuals was recently assessed to determine whether the sacrificial group had been drawn from a different ethnic group. Craniometric analysis, the presence of distinct Nubian artifacts and burial configurations confirmed that these individuals were not foreign (Judd et al. 2006). In the absence of any perimortem physical skeletal evidence, we are left with the assumption of sacrificial status based on researchers’ interpretations of burial placement and timing as the only evidence for sacrificial status (Reisner, 1923a; Kendall, 1997). We are unable to ascertain whether these individuals went willingly to their death or what that manner of death was, but we may be able to extract information from the skeletal remains that may further refine our understanding of the sacrificial group. RESEARCH QUESTIONS AND EXPECTATIONS Health profiles of sacrificial and nonsacrificial subsamples Fig. 1. Location of Kerma. size from 151.9 to 200 m2. Each tumulus infrastructure was a network of corridors bisected by a wide, perpendicular corridor that contained as many as 322 individuals on its floor. The tumuli surfaces and the area surrounding the tumuli were cut by small pit burials that contained single individuals or more complex arrangements where the deceased was laid on a wooden cot and accompanied by one or more additional bodies. Reisner (1923a) speculated that the individuals interred in the corridor were predominantly female members of the king’s retinue, while the subsidiary burials were attributed to elite males who were accompanied by their wives upon the death of the male. This sacrificial status and social identity of the individuals from the Kerma cemetery is considered fact by many researchers based on the archeological evidence (Gadd, 1960; Adams, 1977; Simon, 1989; Bonnet, 1990; O’Connor, 1993; Kendall, 1997; Edwards, 2004). Bioarchaeological research has questioned these assumptions. First, skeletal examinations have not found evidence for any perimortem trauma attributable to human sacrifice. While a relatively high rate of healed traumatic injuries due to interpersonal violence was found among the Kerma sample (Filer, 1992; Judd, 2004; Buzon and Richman, 2007) further studies revealed a comparable injury profile for the sacrificial and nonsacriAmerican Journal of Physical Anthropology Using this assumption of sacrificial status, do the health profiles of the sacrificial and nonsacrificial groups differ? These two subgroups may reflect differences in risks to death. While the nonsacrificial individuals presumably died from a number of causes, including infection, epidemics, nutritional deficiency, as well as traumatic injury, sacrificial victims did not die from natural causes. These burials were not individuals who, for instance, succumbed to chronic infectious disease, but rather they died at a time when they may or may not have been healthy. We expect that this difference between those who died from natural causes and those who were sacrificed will be reflected in frequencies of skeletal indicators of stress. Individuals who were not sacrificed would have more evidence of ill health that eventually may have contributed to their death, such as infection, manifested by the presence of systemic osteoperiostitis. MATERIALS The individuals examined here date to the classic Kerma period (1750–1550 BC) only and were excavated by Reisner (1923a,b). Of the hundreds of individuals excavated by Reisner the remains of 362 individuals are curated at the Duckworth Laboratory at the University of Cambridge Leverhulme Centre for Human Evolutionary Studies. Because the distinction between individuals in sacrificial contexts and those who were not sacrificed is important in this study, only individuals whose context could be determined from Reisner’s (1923a) origi- INVESTIGATING HEALTH AT KERMA nal site report were included. Because of the paucity of juvenile skeletons recovered (n 5 9), only individuals at least 18 years of age were included here. It should be noted that no perimortem trauma was observed on these nine juvenile individuals. Reisner (1923a) recorded infants and juveniles in both burial areas, but collection practices at the time favored adults, particularly their skulls, which resulted in the underrepresentation of nonadult skeletons. Only individuals with a cranium were included in this study. This sample then examined consisted of 290 individuals: 100 from the sacrificial corridor context and 190 from the nonsacrificial context, which in this analysis includes individuals interred outside of the seven large tumuli corridors. Approximately one-third of the 290 individuals had long bones available for analysis. Tibiae and femora from 98 and 75 individuals, respectively, were assessed for long bone length. METHODS Age and sex To investigate the patterns of physiological stress indicators in the total sample and subgroups, age and sex data were collected. Age and sex determination methods followed standard protocols (Buikstra and Ubelaker, 1994). Sex was determined from dimorphic features of the os coxae (Phenice, 1969; Buikstra and Mielke, 1985; Milner, 1992) and cranial morphology (Acśadi and Nemeskéri, 1970). Individuals were aged using the degenerative changes of the pubic symphysis and auricular surface when available (Todd, 1921a,b; Bedford et al., 1989; Meindl and Lovejoy, 1989; Brooks and Suchey, 1990). Individuals without pelvic remains were aged by cranial suture closure (Meindl and Lovejoy, 1985) and tooth wear (Walker et al., 1991). Indications of physiological stress To investigate whether the differences between the sacrificial and nonsacrificial subgroups were reflected in health, the skeletal remains from Kerma were examined for indications of physiological stress. Unhealthy environmental conditions, whether cultural, societal, climatic, or biological, can cause stress resulting in physical disruption. Although not all stress will result in skeletal lesions (Wood et al., 1992; Goodman, 1993), prolonged exposure to unhealthy conditions, particularly during childhood, can produce skeletal manifestations of ill health. For this study, physiological stress was explored through the documentation of cribra orbitalia, dental enamel hypoplasia, tibial osteoperiostitis, and femur length. The recording of these pathological conditions followed the protocol described by Steckel and Rose (2002) and Buikstra and Ubelaker (1994). Cribra orbitalia, identified as porosity with coalescing foramina with or without thickening in the orbital surface, has been associated with nutritional deficiencies such as iron, folic acid, and vitamin C, as well as parasitic infection (e.g., Walker, 1985; Kent, 1986; Stuart-Macadam, 1992; Ortner et al., 1999; Wapler et al., 2004). Observed macroscopically, remodeled and active lesions were noted as well as degree of severity and extent of surface affected. Porotic hyperostosis, manifested as lesions on the cranial vault, was absent at Kerma and is extremely rare among other con- 95 temporaneous Nile Valley samples (Buzon, 2006); thus, this condition was not included in the analysis. The presence of systemic metabolic stress has been associated with dental enamel hypoplasia, a deficiency in enamel thickness resulting from a disruption in amelogenesis (Goodman et al., 1980; Goodman and Armelagos, 1985; Dobney and Goodman, 1991). Hypoplastic lesions were identified by the presence of a continuous linear horizontal groove or series of pits deep enough to be detected with the fingernail, and were recorded by tooth and location. Incisors and canines were examined, given that anterior teeth are the favored location of these defects (Hillson, 1996, p 197). However, premolars were also included due to the high frequency of missing anterior teeth in the sample. An individual was included in the study if at least one observable tooth (incisors, canines, premolars) was present. Osteoperiostitis, inflammation on the bone surface identified by the presence of fine pitting, longitudinal striation, and plaque-like new bone formation on the cortical surface, is associated with infection. Active (woven bone) and remodeled (sclerotic bone) lesions were distinguished to assess if the disease was actively affecting the individual at the time of death and the location and extent of involvement were recorded. Because many of the individuals were incomplete, the examination of osteoperiostitis was restricted to the tibia, the bone on which it is most commonly observed (Roberts and Manchester, 2005). Osteoperiostitis is often associated with nonspecific infections, such as those caused by Staphylococcus and Streptococcus, as well as specific infectious diseases, such as treponematosis and tuberculosis (Ortner, 2003). Because of the incomplete postcrania, recognizing patterns suggestive of a specific condition was difficult. It should also be noted that lesions on the tibia may also be the result of traumatic injury rather than infectious disease, although in this case lesions tend to be unilateral (Ortner, 2003). Full growth potential may not be reached if individuals are subjected to inadequate nutrition, disease or other unfavorable environmental conditions during childhood (Stini, 1969; Goodman, 1991; Steckel, 1995; Bogin, 1999). The femur has been shown to be a very reliable indicator of growth (Israeljohn, 1960) and its maximum length correlates more closely with stature than that of any other bone (Krogman and Iscan, 1986). Thus, as a more direct measure of growth stunting and because the presence of long bones in the sample was highly variable, the maximum length of the femur, rather than a stature estimate using various elements, was used to assess growth disruption. These physiological stress indicators were compared between the sacrificial and nonsacrificial individuals to evaluate the relative health of each subgroup. Patterns based on sex were also examined. The number of recorded cases of these pathological conditions is compared between the groups using the Fisher’s exact test (when cell counts were  5) or chi-squared test for cribra orbitalia, enamel hypoplasia, and osteoperiostitis frequencies, and t tests for the comparison of mean long bone length. RESULTS Age and sex determination The Kerma sample included in this study consisted of 30% adults aged 18–29, 49% adults aged 30–45, and 21% American Journal of Physical Anthropology 96 M.R. BUZON AND M.A. JUDD TABLE 1. Age and sex distribution of Kerma sample (N 5 290)a 18–29 Female Male Total a 30–45 461 Total Sacrificial Nonsacrificial Sacrificial Nonsacrificial Sacrificial Nonsacrificial Sacrificial Nonsacrificial 18 (29)a 8 (21) 26 (26) 42 (37) 18 (24) 60 (31) 28 (45) 23 (61) 51 (51) 52 (46) 40 (52) 92 (48) 16 (26) 7 (18) 23 (23) 20 (17) 18 (24) 38 (20) 62 (62) 38 (38) 100 114 (60) 76 (40) 190 The number in parentheses is the % of the total for each group. TABLE 2. Individuals displaying physiological stress indicatorsa Sacrificial Female Cribra orbitalia Enamel hypoplasia (all anterior teeth) Incisors Canines Premolars Tibial osteoperiostitis Femur length in cm (n) 6/62 5/14 0/4 3/7 2/11 17/28 43.31 (10) (36) (0) (43) (22) (61) (19) Male 5/38 0/8 0/0 0/1 0/8 7/16 44.91 (13) (0) (0) (0) (0) (44) (12) Nonsacrificial Total 11/100 5/22 0/4 3/8 2/19 24/44 (11) (22) (0) (36) (11) (54) Female 21/114 3/24 2/7 1/8 3/24 9/27 42.25 (18) (12) (29) (13) (13) (33) (19)e Male 6/62 6/15 1/6 4/9 4/16 12/27 47.04 (10) (40) (17) (44) (25) (48) (25)f Total 27/189 9/39 3/13 5/17 7/40 21/54 (14)b (23)c (23) (29) (18) (39)d a The number in the parentheses is the % frequency for cribra orbitalia, enamel hypoplasia and tibial osteoperiostis and the number examined for femur length. Comparison between sacrificial and nonsacrificial results (test performed, P value). b 2 v , P 5 0.43. c Fisher’s, P 5 0.61. d 2 v , P 5 0.12. e T test, P 5 0.12. f T test, P 5 0.08. aged 46 and older. Sixty-one percent were female, 29% were male. When demographic profiles of the sacrificial and nonsacrificial subgroups were compared (Table 1), the samples were quite similar. A Kolmogorov-Smirnov test found that the age distributions were not statistically different. The largest age cohort for both groups was about 50% for the middle-aged (30–45) group with younger (18–29) and older adults (461) accounting for 20–32% each. The sex distribution was also comparable with more females than males in each subgroup; chisquared tests did not show any statistical differences between the subgroups in regard to sex distribution. Indications of physiological stress Physiological stress was examined through indications of macroscopic nonspecific disease processes visible on the skeletal remains (Table 2). Thirteen percent (37/289) of the total observable sample displayed cribra orbitalia, with 11% of the sacrificial subgroup and 14% of the nonsacrificial subgroup affected. All lesions present were remodeled. Twenty-three percent (14/61) of those with observable teeth displayed enamel hypoplasia, with 22% of the sacrificial subgroup and 23% of the nonsacrificial subgroup affected. The frequency of enamel hypoplasia by tooth type is also presented. Tibial osteoperiostitis was present on 46% (45/98) of the total observable sample, with 54% of the sacrificial subgroup and 39% of the nonsacrificial group affected. Three sacrificial individuals and one nonsacrificial individual had active osteoperiostitis; all of the other cases were remodeled. For these three disease processes, none of the differences in condition frequency between the sacrificial and nonsacrificial subgroups were statistically significant. Also, differences were not significant when the sexes were examined separately (Table 2). Within the subgroups, males and females were also compared. Generally, the results were similar, though within the nonsacrificial subgroup, American Journal of Physical Anthropology females had a higher rate of cribra orbitalia than males and males had a higher rate of enamel hypoplasia than females; these differences were not statistically significant when using the Fisher’s exact test (see Table 2), which is more appropriate for the small sample sizes. Mean femur length is an additional means to assess stress. For the total measurable sample, the mean female femoral length was 42.78 cm (n 5 38) and the mean for males measured 46.35 cm (n 5 37). Sacrificial females and males had mean femoral lengths of 43.31 and 44.91 cm, respectively; nonsacrificial females had a mean femur length of 42.25 cm and the nonsacrificial male mean was 47.04 cm (Table 2). When comparing the subgroups, mean femur length did not differ significantly between the groups for males or females. DISCUSSION This discussion examines the patterns of nonspecific stress at Kerma. Specific attention is given to lack of differences between the sacrificial and nonsacrificial subgroups. The frequencies of paleopathological conditions are assessed in comparison with other regional samples and possible explanations for the observed patterns are presented. Overall health of Kerma sample What do the patterns of nonspecific stress at Kerma tell us about the people buried there? Overall, the frequencies of paleopathological conditions at Kerma are very similar to the frequencies observed among other relatively contemporaneous samples (Table 3) from Nubia, such as the adult skeletal sample from Tombos, located just north of Kerma, and C-Group and Pharaonic samples in Lower Nubia excavated by the Scandinavian Joint Expedition to Sudanese Nubia (Buzon, 2006). In comparison with the 13% frequency of cribra orbitalia at 97 INVESTIGATING HEALTH AT KERMA a TABLE 3. Comparison in stress indicators between Kerma and other regional samples Kerma Sacrificial Cribra orbitalia Enamel hypoplasia Tibial osteoperiostis Femur length in cm Female Male Nonsacrificial Tombosb c C-Groupb Pharaonicb 11/100 (11) 5/22 (22) 24/44 (54) 27/189 (14) 9/39 (23) 21/54 (39) 3/69 (4) 12/52 (23) 15/33 (45) 20/178 (11) 10/60 (17) 30/67 (45) 13/62 (21) 3/10 (30) 11/29 (38) 43.31 (19) 44.91 (12) 42.25 (19) 47.04 (25) 41.93 (12) 43.82 (7) 42.63 (24) 45.61 (14) 43.47 (9) 44.36 (9) a The number in the parentheses is the % frequency for cribra orbitalia, enamel hypoplasia and tibial osteoperiostis and the number of femurs measured for femur length. b Data from Buzon (2006). c Statistically different from nonsacrificial subgroup, v2, P 5 0.03. Kerma, the frequency at these three sites ranges from 4 to 21%. The 23% frequency of dental enamel hypoplasia at Kerma is comparable with these samples as well with a range of 17–30%. Finally, the 46% frequency of osteoperiostitis at Kerma is also similar to the range of 38–45% in these Nubian samples. The only statistically significant difference is the higher frequency of cribra orbitalia in the nonsacrificial Kerma group in comparison with Tombos. Generally, these similarities suggest that the groups experienced similar environmental and social factors that affected health; other research found that diet and exposure to infectious agents was also similar for these groups, although social status might have varied (Buzon, 2006). Interpretations of health profiles and implications for research questions Because the sacrificed individuals at Kerma did not die under the same conditions as the nonsacrificial individuals, differences in paleopathological patterns were expected. However, as presented above, the sacrificial and nonsacrificial groups did not show any statistical differences in skeletal indicators of stress. The patterns of stress indicators showed a remarkable degree of similarity in regard to cribra orbitalia, dental enamel hypoplasia, and osteoperiostitis. Femur length also revealed no significant differences. How can this overall similarity in health patterns between these subgroups be explained? One question that arises involves our ability to observe variations between these groups. Possible differences between the sacrificial and nonsacrificial individuals may have been obscured by a number of factors. For instance, the lack of juveniles limited our ability to examine critically the patterns of conditions primarily associated with childhood stress such as cribra orbitalia, dental enamel hypoplasia, and stunted femur length. Researchers have observed that active cribra orbitalia is found mostly in young children less than 5 years of age (e.g., Walker, 1986; Mittler and Van Gerven, 1994; Van Gerven et al., 1995; Fairgrieve and Molto, 2000). Thus, adult individuals in the Kerma sample represent only those who were able to withstand and live through the stress that caused the lesions. We cannot evaluate the differences between the groups regarding those who died during childhood. Enamel hypoplasia and stunted growth provide an indication of childhood circumstances, but again we only can examine the individuals who were able to survive a period (or periods) of stress. In addi- tion, the children who were not able to withstand the stress long enough to produce skeletal lesions and died are not included in this sample. The only condition that reflects adult conditions is osteoperiostitis, with no evident differences between subgroups. Along with the lack of evidence for differences in the frequency of traumatic injuries (Judd et al., 2006) these results again serve to highlight the similarities between the sacrificial and nonsacrificial subgroups. One possible interpretation of these findings is that the assumption of sacrificial victims from the archaeological remains is indeed incorrect. However, we must not understate the difficulties in sorting out the many factors that contribute to heterogeneity in a skeletal sample. At times, skeletal indicators of stress may not reflect subgroup differences that we assume exist based on archaeological evidence due to the limitations of skeletal analysis. Despite this, a bioarchaeological approach that combines health data from the skeleton with archaeological materials, historical records and cultural ideas remains the best approach to assist researchers in exploring the possible factors affecting health, as well as assist in assessing assumptions made about the people buried at an archaeological site. CONCLUSIONS This research has demonstrated the complexity involved in understanding the multiple factors that result in heterogeneity in skeletal samples. This analysis reveals no differences between individuals from the sacrificial corridors and those from the nonsacrificial burial areas at Kerma in Upper Nubia. In addition, the data presented here indicate that individuals from Kerma had similar health profiles to other contemporaneous samples in the region. While sacrificial individuals may not have experienced the same risk of death as nonsacrificial individuals, it is not evident in the frequency of nonspecific stress indicators. This differential risk of death may be clouded by our inability to examine juveniles for childhood disease. Alternatively, these results may suggest a reassessment of the assumption of sacrificial status based on archaeological evidence. ACKNOWLEDGMENTS Marta Lahr and Robert Foley at the University of Cambridge provided access to the Kerma sample in the American Journal of Physical Anthropology 98 M.R. BUZON AND M.A. JUDD Duckworth Collections. Maggie Bellati is thanked for her helpfulness during our visits. We thank Editor Clark Larsen and three anonymous reviewers for their helpful comments and suggestions. LITERATURE CITED Acśadi G, Nemeskéri J. 1970. History of human life span and mortality. Budapest: Akadémiai Kiadó. Adams WY. 1977. Nubia: corridor to Africa. Princeton, New Jersey: Princeton University Press. Bedford ME, Russell KF, Lovejoy CO. 1989. The utility of the auricular surface aging technique. Poster presented at the 58th Annual Meeting of the American Association of Physical Anthropologists, San Diego. Bogin B. 1999. Patterns of human growth, 2nd ed. Cambridge: Cambridge University Press. Bonnet C. 1990. Sépultures et coutumes funéraires. In: Bonnet C, editor. Kerma Royaume de Nubie: L’Antiquité Africaine au Temps des Pharaons. Genève: Mission archéologique de l’Université de Genève au Soudan. p 69–86. Bonnet C. 1994. Kerma: les apports historiques de l’archéologie. In: Bonnet C, editor. Études Nubiennes. Genève: Fonds National Suisse de la Recherche Scientifique. p 101–110. Bonnet C. 2004. The Kerma culture. In: Welsby DA, Anderson JR, editors. Sudan: Ancient Treasures. London: The British Museum Press. p 70–77. Brooks ST, Suchey JM. 1990. Skeletal age determination based on the os pubis: a comparison of the Ascádi-Nemeskéri and Suchey-Brooks methods. Hum Evol 5:227–238. Buikstra JE, Mielke JH. 1985. Demography, diet and health. In: Gilbert RI, Mielke JH, editors. The analysis of prehistoric diets. New York: Academic Press. p 359–422. Buikstra J, Ubelaker D, editors. 1994. Standards for data collection from human skeletal remains. Fayetteville: Arkansas Archaeological Survey Research Series No. 44. Buzon MR. 2006. Health of the nonelites at Tombos: nutritional and disease stress in New Kingdom Nubia. Am J Phys Anthropol 130:26–37. Buzon MR, Eng JT, Lambert PM, Walker PL. 2005. Bioarchaeological methods. In: Maschner HDG, Chippindale C, editors. Handbook of archaeological methods. Lanham, CA: Altamira Press. p 871–918. Buzon MR, Richman R. 2007. Traumatic injuries and imperialism: the effects of Egyptian colonial strategies at Tombos in Upper Nubia. Am J Phys Anthropol 133:783–791. Dobney K, Goodman A. 1991. Epidemiological studies of the dental enamel hypoplasias in Mexico and Bradford: their relevance to archaeological skeletal studies. In: Bush H, Zvelebil M, editors. Health in past societies: biocultural interpretations of human skeletal remains in archaeological contexts. Oxford: BAR. p 81–100. Edwards DN. 2004. The Nubian past: an archaeology of the Sudan. London: Routledge. Fairgrieve SI, Molto JE. 2000. Cribra orbitalia in two temporally disjunct population samples from the Dakleh Oasis, Egypt. Am J Phys Anthropol 111:319–331. Filer JM. 1992. Head injuries in Egypt and Nubia: a comparison of skulls from Giza and Kerma. J Egypt Archaeol 78:281–285. Gadd CJ. 1960. The spirit of living sacrifices in tombs. Iraq 22:51–58. Goodman A, Armelagos G. 1985. Factors affecting the distribution of enamel hypoplasias within the human permanent dentition. Am J Phys Anthropol 68:479–493. Goodman A, Armelagos G, Rose J. 1980. Enamel hypoplasia as indicators of stress in three prehistoric populations from Illinois. Hum Biol 52:515–528. Goodman AH. 1991. Health, adaptation, and maladaptation in past societies. In: Bush H, Zvelebil M, editors. Health in past societies: biocultural interpretations of human skeletal remains in archaeological contexts. British Archaeological Reports, International Series, Vol. 567. Oxford: Tempus Reparatum. p 31–38. American Journal of Physical Anthropology Goodman AH. 1993. On the interpretation of health from skeletal remains. Curr Anthropol 34:281–288. Hillson S. 1996. Dental anthropology. Cambridge: University Press. Israeljohn WJ. 1960. Description and modes of analysis of human growth. In: Tanner JM, editor. Human growth. New York: Pergamon Press. p 21–42. Judd MA. 2002. Battle-weary or just battered: the plight of the people in the Kerma corridors. Paper presented at the International Society for Nubian Studies 10th International Conference, Rome. Judd MA. 2004. Trauma in the city of Kerma: ancient versus modern injury patterns. Int J Osteoarchaeol 14:34–51. Judd MA, Irish J, Froment A. 2006. Dying to serve: human sacrifice during the classic Kerma period. Am J Phys Anthropol 129:110–111. Kendall T. 1997. Kerma and the kingdom of Kush 2500-1500 BC: the archaeological discovery of an ancient Nubian empire. Washington DC: National Museum of African Art, Smithsonian Institution. Kent S. 1986. The influence of sedentism and aggregation of porotic hyperostosis and anaemia: a case study. Man 21:605– 636. Krogman WM, Iscan MY. 1986. The human skeleton in forensic medicine. Springfield: Charles C. Thomas. Meindl RS, Lovejoy CO. 1985. Ectocranial suture closure: a revised method for the determination of skeletal age at death based on the lateral-anterior sutures. Am J Phys Anthropol 68:57–66. Meindl RS, Lovejoy CO. 1989. Age changes in the pelvis: implications for paleodemography. In: Iscan MY, editor. Age markers of the human skeleton. Springfield: Charles C. Thomas. p 137–168. Milner GR. 1992. Determination of skeletal age and sex: a manual prepared for Dickson Mounds reburial team. Ms. On file. Lewiston, IL: Dickson Mounds Museum. Mittler DM, Van Gerven DP. 1994. Developmental, diachronic, and demographic analysis of cribra orbitalia in the medieval Christian populations of Kulubnarti. Am J Phys Anthropol 93:287–297. Morkot R. 2000. The black pharaohs: Egypt’s Nubian rulers. London: Rubicon Press. O’Connor D. 1993. Ancient Nubia: Egypt’s rival in Africa. Philadelphia: The University Museum, University of Philadelphia. Ortner D. 2003. Identification of pathological conditions in human skeletal remains. San Diego: Academic Press. Ortner D, Kimmerle EH, Diez M. 1999. Probable evidence of scurvy in subadults from archaeological sites in Peru. Am J Phys Anthropol 108:321–331. Phenice T. 1969. A newly developed visual method of sexing in the os pubis. Am J Phys Anthropol 30:297–301. Reisner GA. 1923a. Excavations at Kerma. Parts I–III. Cambridge, MA: Harvard African Studies. Reisner GA. 1923b. Excavations at Kerma. Parts IV–V. Cambridge, MA: Harvard African Studies. Roberts C, Manchester K. 2005. The archaeology of disease. Ithaca, NY: Cornell University Press. Simon C. 1989. Les populations kerma-evolution interne et relations historiques dans le contexte Egypto-Nubien. Archéologie du Nil Moyen 3:139–148. Steckel RH. 1995. Stature and the standard of living. J Econ Lit 33:1903–1940. Steckel RH, Rose JC, editors. 2002. The backbone of history: health and nutrition in the western hemisphere. Cambridge: Cambridge University Press. Stini WA. 1969. Nutritional stress and growth: sex difference in adaptive response. Am J Phys Anthropol 31:417–426. Stuart-Macadam P. 1992. Porotic hyperostosis: a new perspective. Am J Phys Anthropol 87:39–47. Todd TW. 1921a. Age changes in the pubic bone. II. The pubis of the white female. IV. The pubis of the female white-negro hybrid. Am J Phys Anthropol 4:1–70. Todd TW. 1921b. Age changes in the pubic bone I. The male white pubis. Am J Phys Anthropol 3:285–334. INVESTIGATING HEALTH AT KERMA Van Gerven DP, Sheridan SG, Adams WY. 1995. The health and nutrition of a medieval Nubian population—the impact of political and economic change. Am Anthropol 97:468–480. Walker PL. 1985. Anemia among prehistoric Indians of the American Southwest. In: Merbs CF, Miller RJ, editors. Health and disease in the prehistoric Southwest. Tempe: Arizona State University. p 139–164. Walker PL. 1986. Porotic hyperostosis in a marine-dependent California Indian population. Am J Phys Anthropol 69:345– 354. Walker PL, Dean G, Shapiro P. 1991. Estimating age from tooth wear in archaeological populations. In: Kelley M, Larsen CS, 99 editors. Advances in dental anthropology. New York: Alan R. Liss. p 169–178. Wapler U, Crubezy E, Schultz M. 2004. Is cribra orbitalia synonymous with anemia? Analysis and interpretation of cranial pathology in Sudan. Am J Phys Anthropol 123:333– 339. Watterson B. 1997. The Egyptians. Oxford: Blackwell. Wood J, Milner G, Harpending H, Weiss K. 1992. The osteological paradox: problems of inferring prehistoric health from skeletal samples. Curr Anthropol 33:343–370. Wright LE, Yoder CJ. 2003. Recent progress in bioarchaeology: approaches to the osteological paradox. J Archaeol Res 11:43–70. American Journal of Physical Anthropology