(2002) Ancient Injury Recidivism: an Example from the Kerma Period of Ancient Nubia

International Journal of Osteoarchaeology Int. J. Osteoarchaeol. 12: 89–106 (2002) Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/oa.587 Ancient Injury Recidivism: an Example from the Kerma Period of Ancient Nubia MARGARET JUDD* Department of Egyptian Antiquities, The British Museum, London WC1B 3DG ABSTRACT A topical trend in clinical research has been the study of repeat trauma, referred to by clinicians as ‘‘injury recidivism,’’ which lends itself to the assessment of accumulated injuries among ancient people. The present investigation examined the healed injuries among two archaeological skeletal samples from the Kerma period (ca. 2500–1500 BC) of Sudanese Nubia. Both groups were known to have a high prevalence of multiple trauma—80% of 54 adults from the rural sites (O16 and P37) located near Dongola and 42% of 212 adults from the urban site of Kerma sustained nonfatal injuries. It was observed that a higher frequency of multi-injured adults displayed one or more violence-associated injury (cranial trauma, parry fracture). When all injuries were considered 38% of individuals with violence-related injuries had other traumatic lesions in contrast to 22% of individuals who experienced injuries associated with accidental falls (e.g., Colles’, Smiths’, Galeazzi, and paired forearm fractures), although this difference was not significant. When only the skulls and long bones were evaluated 81% of adults with multiple injuries to these major bones bore one or more violence-related injuries, while 60% of adults with single injuries sustained violence-related injuries. Most individuals with multiple injuries were male and less than 35 years of age; there was no significant difference in the frequency of violence- or accident-related multiple injury between the rural and urban communities. Although it cannot be established whether or not some of an individual’s injuries were experienced during simultaneous or independent incidents, the pattern of multiple injury among these two ancient Nubian skeletal samples reflected the profile of injury recidivism observed by modern clinicians cross-culturally. Copyright  2002 John Wiley & Sons, Ltd. Key words: trauma; injury; Nubia; Kerma; fracture; violence; Sudan Introduction Trauma, long regarded as an outcome of chance, now ranks as ‘‘a chronic and recurrent disease’’ (Sims et al., 1989) and clinicians have developed a profile of the people who have experienced repeat injury—the injury recidivists (Hedges et al., 1995; Kaufmann et al., 1998; Poole et al., 1993; Reiner et al., 1990; Smith et al., 1992; Williams et al., 1997). As a result of repeat injury, these individuals accumulate an aggregate of traumatic lesions over their lifetime, which is precisely what bioarchaeologists observe in archaeological skeletal collections. In the past, however, bioarchaeological interpretation has ∗ Correspondence to: M. Judd, Department of Egyptian Antiquities, The British Museum, London WC1B 3DG. e-mail: margaretjudd@hotmail.com Copyright  2002 John Wiley & Sons, Ltd. been limited by the absence of clinical literature on accumulative injury, with which to assess multiple trauma at the populational level. The clinical studies Sims et al. (1989) challenged the traditional notion that injury transpired because of chance circumstances, and considered how an individual’s lifestyle might influence their vulnerability to trauma over time. They found that 44% out of 263 consecutive assault admissions to an inner city Detroit (United States) hospital were repetitive casualties and victims of impoverished lifestyles involving substance abuse, unemployment, poverty, and crime. In a less volatile North American urban environment (Newark, Received 4 December 2000 Accepted 12 January 2001 90 New Jersey) Reiner et al. (1990) observed that 23% (n = 138) of consecutively admitted assault patients were injury recidivists, and of these, 53% were traumatised by the same mechanism as before (most frequently a penetrating injury). Similar results were found among male assault admissions in Washington, DC, where 49% of the cases were injury recidivists and unemployment was revealed as the foremost factor (Goins et al., 1992). The injury recidivist was profiled as being typically male (97%), an ethnic minority, and younger than the average trauma patient (mean age was 26 years); having suffered the first injury, on average, at 20 years of age; and reappearing in the trauma unit within five years. Curiously, the injuries received were no more life threatening than those of the first time presenters of trauma. Recidivism in Oakland (California), though similar in profile, produced a dramatically lower frequency of repeat trauma—3,442 out of 10,525 (3%) individuals were involved (Smith et al., 1992). The investigators attributed the lower incidence to their inclusion of all trauma patients, rather than those suffering from serious assault only (i.e., Goins et al., 1992; Reiner et al., 1990; Sims et al., 1989) and, therefore, these results produced a more realistic image of recidivism susceptibility to all types of injury mechanisms. An investigation in Jackson (Mississippi) (Poole et al., 1993) initiated a new dimension into injury recidivism research. The sample differed in that Mississippi was predominantly rural and Poole’s sample consisted of 200 cases of all types of injury (e.g., falls, burns, motor vehicle accidents, and assault) that required hospitalisation. The results were startling—first, the rural injury recidivist rate was 40%, nearly equal to that of the inner city of Detroit and Washington, DC, and second, there was no difference in injury recidivism for accidental or intentional injury victims, although the injury recidivist profile was identical to that found by previous investigators. Hedges et al. (1995) also included all types of trauma in their analysis of 22,213 injury admissions in San Diego (California) and calculated a repeat injury rate of only 1%. Assault ranked highest of all repeat injury mechanisms (38%) and falls accounted for 9% of the repeat injury mechanisms, but the injury recidivist characteristics supported the original model. Madden et al. (1997) tested the injury Copyright  2002 John Wiley & Sons, Ltd. M. Judd recidivist model to determine if injury recidivists in Raleigh (North Carolina) were most likely to be young black males. They discovered that injury recidivism was indiscriminate of race, gender, and age, but a previous assault injury within the past year was highly related to a return visit. This study was influential to injury recidivism research in that an individual’s inclusion in the study did not require hospital admission and, therefore minor injuries, such as fractured fingers, were included. Also unique was that all people who presented with injury during a one year period were tracked for injury for the year following their initial visit. The injury recidivism rate was 22%, and of this group, 78% returned one time, 16% returned two times, and 6% returned more than three times in the period of one year. A year long study of a rural sample from West Virginia found that 12% of injured individuals were injury recidivists (Williams et al., 1997). Repeat injury was not associated with sex, however, age decreased with increasing repeat trauma, while increased returns by an individual resulted in more overexertionand violence-related injuries. A statewide survey of injury recidivism in Nevada revealed that 2% of 10,137 injured patients were injury recidivists, and those suffering from interpersonal violence were at a greater risk of repeat injury (Kaufmann et al., 1998). Little injury recidivism research exists outside of the United States, with the exception of a nationwide New Zealand study (Dowd et al., 1996), where no demographic pattern was found among injury recidivists with repetitive assault injuries, and an investigation from Israel (Sayfan and Berlin, 1997). Sayfan and Berlin’s (1997) study of injury recidivism in a traditional rural Israeli society not only diversified the geographical scope of this new area of inquiry, but perhaps provided a more appropriate model with which to assess multiple trauma among ancient people. Thirty percent of 100 injured adults were injury recidivists and most were young males; socioeconomic and cultural factors were inconsequential. The profile of the injury recidivist was similar in all studies, although there were some exceptions. Typically, the individual was male, younger than the average age of the nonrecidivist trauma patient, of low socioeconomic status, unemployed, and frequently involved in Int. J. Osteoarchaeol. 12: 89–106 (2002) Ancient Injury Recidivism illicit behaviour (e.g., crime, substance abuse). Depending on the research design, injury recidivism ranged from 1% to 49%. Methodological problems persist in this topical research area, and are reminiscent of those encountered in palaeotrauma research, for example, the length of the study period, sample size, definition of significant injury, types of injury included in the study (assault only or all types of mechanisms), and many more (e.g., Dowd et al., 1996; Greer and Williams, 1999; Hedges et al., 1995; Smith et al., 1992). However, the findings available to date provide an alternative means of assessing multiple injury among ancient people and the injury recidivist profile, common to the majority of the investigations, offers the paradigm with which to evaluate ancient multiple trauma. Palaeotrauma, age, and multiple injury Since Lovejoy and Heiple’s (1981) elemental analysis revolutionised human palaeotrauma recording, most bioarchaeologists produce a systematic assessment of discrete fractures for a skeletal sample and explain the patterns observed in terms of demographics, culture, and environment. One problem still persists in the epidemiological evaluation of ancient trauma, and that is the inability to determine the individual’s age when the injury occurred, unless it was shortly before death (Roberts, 1988). Not only does this create difficulties in determining which age groups were at greatest risk of injury in a given society, but the ability to detect which injuries occurred simultaneously, in the case of multiple trauma, is also restricted. Some bioarchaeologists addressed the demographic problem by applying Buhr and Cooke’s (1959) analysis of years at risk of fracture for each bone to archaeological samples using the age at death of individuals with fractures (e.g., Burrell et al., 1986; Lovejoy and Heiple, 1981; Neves et al., 1999). In their study of age-related fractures among Oxford infirmary fracture patients, Buhr and Cooke (1959) observed that certain bones were prone to fracture at specific ages. The researchers generated fracture curves that characterised the most susceptible individuals; for example, injuries to the hands and feet peaked Copyright  2002 John Wiley & Sons, Ltd. 91 among young males, a pattern that was termed the ‘‘A-curve,’’ or work-related curve, while femoral head fractures increased with age and appeared as a ‘‘J-curve.’’ Lovejoy and Heiple (1981) applied this analytical method to a Late Woodland archaeological skeletal sample from the Libben site (Ohio), and found two peak periods of discrete long bone fracture—adolescence to young adult and old age. They concluded that the people were relatively peaceful and that the injuries, which affected the sexes indiscriminately, were accidental, and a factor of increasing age rather than violence. An investigation of two ancient Nubian groups discovered an age-related change in injury pattern over time (Burrell et al., 1986). The Early Christian group exhibited an A-pattern of injury, particularly among male bones, while the Late Christian sample yielded a U-curve, that is, the youngest (less than ten years old) and oldest bones bore fractures. An improvement in general health status during the later period was proposed as an explanation for the U-shaped curve. The children led healthier, more active lives and, therefore, were predisposed to fracture-inducing activities; more older people survived the fractures of youth and lived longer because of the more amenable environment (see, Wood et al., 1992, for a discussion of the ‘‘osteological paradox’’). A more recent study of prehistoric Chileans found that the long bone fractures were age-related and that young adults suffered the least from injury (Neves et al., 1999). In all cases, the total number of long bone fractures among individuals for each age cohort was evaluated at the expense of the individual’s cluster of injuries. In palaeotrauma analysis, the multiple trauma component typically consists of a multiple injury rate for the sample (number of individuals with multiple trauma per number of individuals observed), a multiple injury rate for those with trauma (number of individuals with multiple lesions per number of individuals with trauma), and possibly a mean injury rate for both of the above (number of lesions per number of individuals with and without trauma) (e.g., Alvrus, 1999; Cybulski, 1992; Judd and Roberts, 1998; 1999; Jurmain and Kilgore, 1998; Kilgore et al., 1997; Lahren and Berryman, 1984; Lovell, 1990; Robb, 1997); some researchers have reported the multiple injury rate for each sex (e.g., Jurmain and Int. J. Osteoarchaeol. 12: 89–106 (2002) 92 Kilgore, 1998; Kilgore et al., 1997). An analysis of the accumulated number and types of injuries born by an individual at the time of death may reveal greater insight into age-related activity within the society, and the clinical concept of injury recidivism, by its focus on the accumulation of injury, provides a fresh perspective on this longneglected area of palaeotrauma research. Goals of this investigation It is proposed that in samples where multiple injuries are prominent, particularly lesions indicative of nonlethal interpersonal violence (cranial and direct force isolated ulna fractures) or accidental falls (e.g., Colles’, Smiths’, Galeazzi, and paired forearm rotational fractures), that a second level of analysis should follow, which may further elucidate patterns of injury unique to the culture. While it is difficult to ascertain whether archaeological injuries were simultaneous, the panorama of injuries displayed by a person may be the result of injury recidivism rather than a single traumatic episode. This study evaluated the multiple injury patterns of two ancient Nubian samples in order to: 1. define the demographic pattern of multiple injury, 2. ascertain if there was a difference in this injury pattern between these two samples that represented rural and urban communities of the same culture, 3. assess the association of nonlethal violenceand accident-related injuries to multiple trauma. Materials and methods The archaeological context and skeletal samples The trauma data presented here derived from the trauma analyses of two archaeological skeletal samples from the Kerma period (ca. 2500–1500 BC) of ancient Sudanese Nubia (Judd, 2000). Both samples, although divergent in socioeconomic complexity, were discovered to have had comparably high prevalences of general trauma in addition to the traditional skeletal Copyright  2002 John Wiley & Sons, Ltd. M. Judd indicators of nonlethal interpersonal violence (cranial injury, forearm fracture, and multiple injury) when compared to other archaeological samples from the Upper Nubian vicinity. Kerma, the type-site for the Kerma culture, was one of the earliest cities that dominated Upper Nubia (above the Second Cataract) and monitored the Nile trade with the African interior. The ancient city was situated 20 km south of the Nile’s Third Cataract, and 70 km north of the cemeteries of the Dongola vicinity, from where the rural skeletal samples were excavated. The skeletal material from Kerma was excavated by George Reisner (1923a; 1923b) in 1916 and is currently housed in the Duckworth Laboratory of the University of Cambridge’s Bioanthropology Department. The rural skeletal remains were excavated from two cemeteries by myself, local Sudanese workers from the Dongola vicinity, and fellow members of the Sudan Archaeological Research Society’s Northern Dongola Reach Survey (NDRS) team from 1994–97 under the direction of Dr Derek Welsby (1996; 1997) of the British Museum’s Department of Egyptian Antiquities. The biological sex of the 278 skeletons was assigned by dimorphic criteria of the skull and pelvis, as recommended by Buikstra and Ubelaker (1994). Intermittent preservation necessitated that other methods be employed and therefore, measurements from long bones, as described by Olivier (1969), were utilized. Age at death was calculated from scores obtained from the degenerative changes of the pubis (Todd, 1921a; 1921b), sternal rib end modification (Loth and Iscan, 1989), and changes to the auricular surface of the innominate (Lovejoy et al., 1985). The age cohorts were broadly defined as subadult (<25 years), young adult (25–35 years), middle adult (35–50 years), old adult (50+), and ‘‘adult’’ when bones were too fragmentary to confidently estimate the age. In this investigation, only individuals for whom age and sex could be assigned were included. The rural skeletal sample, dated to the Kerma Ancien and Moyen periods (ca. 2500–1750 BC), consisted of 55 adults, 28 males and 27 females, and of these 80% met with at least one injury, while 61.8% sustained two or more lesions. The royal cemeteries and subsidiary graves of Kerma were dated to the Kerma Classique period (ca. Int. J. Osteoarchaeol. 12: 89–106 (2002) Ancient Injury Recidivism 93 1750–1500 BC) and yielded 93 male and 130 female skeletons, 39.5% of whom suffered at least one injury, while 18% of the sexed sample bore two or more injuries. Method of analysis In clinical investigation, injury recidivism is determined by the number of times that an individual sought medical treatment for injury, but the exact numbers of injuries presented on each occasion are often unstated. On the contrary, in palaeotrauma analysis, bioarchaeologists have no means of knowing the number of episodes during which an individual received injuries, while the total number of injuries displayed by each person is reported. In this investigation each adult was grouped according to sex, age, and site as exhibiting no injury, one injury only, or two or more injuries, and all injuries were included (e.g., skull, long bones, hands, feet, torso). For each site, the frequencies of adults who presented solitary or multiple trauma with violence- and accidentrelated injuries as proposed in the earlier trauma analysis (Judd, 2000) were compared to the frequencies of adults with solitary or multiple trauma where no specific injury mechanism was observed. Violence-related trauma included cranial fractures and direct force forearm fractures (e.g., the parry fracture), while accident-related injuries consisted of indirect force fractures of the forearm (e.g., Colles’, Smiths, Galeazzi, and paired forearm fractures). These results were subsequently compared between the sites. The samples were then assessed using skull and long bone injuries only. Chisquare tests determined whether any differences were statistically significant, the Yate’s correction for continuity (χc2 ) was applied to small samples, and degrees of freedom was ‘‘1’’ unless otherwise stated; the significance level chosen was 0.05. Results The demographic distributions for both samples are presented in Figure 1, while Table 1 inventories the raw counts of lesions sustained by the sexes for each age cohort for both sites. When age was factored into the analysis, 80% of 54 aged and sexed rural individuals, suffered one or more injuries, which was significantly greater than their urban neighbours where 42% of 212 aged and sexed individuals incurred some form of injury 45% 40% Percent of adults 35% 30% 25% 20% 15% 10% 5% 0% <25 25−35 35−50 50+ Age cohort (years) NDRS Kerma Figure 1. Demographic distribution of the NDRS and Kerma adults. Copyright  2002 John Wiley & Sons, Ltd. Int. J. Osteoarchaeol. 12: 89–106 (2002) M. Judd 94 Table 1. Distribution of all injuries (raw counts) among aged and sexed cohorts from the NDRS and Kerma samples Sample Age cohort (years of age) No Fractures One Fracture Multiple Fractures Sample Total M F T M F T M F T M F T NDRS <25 25–35 35–50 50+ Total 0 0 1 0 1 3 2 5 0 10 3 2 6 0 11 2 2 2 0 5 0 1 2 0 3 2 3 4 0 9 2 9 7 2 20 1 6 5 2 14 3 15 12 4 34 4 11 10 2 27 4 9 12 2 27 8 20 22 4 54 Kerma <25 25–35 35–50 50+ Total 2 18 22 7 49 15 31 20 9 75 17 49 42 16 124 1 5 11 0 17 4 19 4 4 31 5 24 15 4 48 0 8 12 2 22 3 7 6 2 18 3 15 18 4 40 3 31 45 9 88 22 57 30 15 124 25 88 75 24 212 M = Number of males; F = Number of females; T = Total number of individuals. Copyright  2002 John Wiley & Sons, Ltd. 100% 90% 80% Percent of injured adults (χ 2 = 25.02, p < 0.000). When multiple injuries were considered for the two samples, 63% of the rural sample sustained multiple injuries compared to only 19% of the urban sample (χ 2 = 110.13, p < 0.000). The modal distributions revealed that males presented slightly more multiple injuries (NDRS: 59%, Kerma 55%) than the females in both groups. When the individual counts of multiple injury (number of injured individuals per total individuals) were examined for each sex, more rural males (20 out of 27, 74%) experienced multiple injuries than the rural females (52%), and significantly more rural males had multiple injuries than the urban males (25%, χ 2 = 21.46, p < 0.000). Significantly more urban males bore multiple injuries than the urban females (15%, χ 2 = 3.69, p = 0.055), as did the rural females (χ 2 = 18.51, p < 0.000). Among individuals under 35 years of age, more rural adults (53%) met with multiple trauma in comparison to the urban group (45%). Figures 2 and 3 depict the frequency distribution of multiple injury for the pooled sexes by age cohort for NDRS and Kerma samples respectively. The frequency of rural young adults displaying lesions varied directly with the increase in the number of lesions, but the frequency of injured urban young adults decreased with the number of visible lesions. Fewer rural middle-aged adults experienced multiple injuries than the younger group, although the majority (55%) did suffer from multiple lesions. The opposite trend was observed among the middle-aged urbanites who 70% 60% 50% 40% 30% 20% 10% 0% <25 25−35 25−35 50+ Age cohort (years) None One Two+ Figure 2. Distribution of injured adults by number of injuries and Age cohort for the NDRS sample. were increasingly represented as the number of injuries increased, however, the majority (56%) experienced no injuries. All of the older rural adults sustained two or more injuries, while 83% of urban elders did not suffer from multiple trauma. In the present analysis, it was observed that 34 out of 43 (79%) of the rural injured adults and 40 out of 88 (46%) of the urban injured adults met with multiple trauma (χ 2 = 11.65, p = 0.001). Table 2 tallies the frequency of multiple injury for Int. J. Osteoarchaeol. 12: 89–106 (2002) Ancient Injury Recidivism 95 100% 90% Percent of injured adults 80% 70% 60% 50% 40% 30% 20% 10% 0% <25 25−35 50+ 35−50 Age cohort (years) None One Two + Figure 3. Distribution of injured adults by number of injuries and age cohort for the Kerma sample. Table 2. Distribution of multi-injured adults for the sexed age cohorts of the NDRS and Kerma samples Sex Male Female Total Age <25 25–35 35–50 50+ <25 25–35 35–50 50+ NDRS Kerma n′ m % 4 11 10 2 4 9 12 2 2 9 7 2 1 6 5 2 54 34 P-value n′ m % 50 82 70 100 25 67 42 100 3 31 45 9 22 57 30 15 0 8 12 2 3 7 6 2 0 26 27 22 14 12 20 13 0.147 0.001∗ 0.009∗ 0.039∗ 0.562 0.000∗ 0.149 0.007∗ 59 212 40 19 0.000∗ n′ = number of individuals with injuries; m = number of individuals with multiple injuries; % = n′ /m X 100%; ∗ significant at P = 0.05. each sex by age cohort for the two samples; the p-values that resulted when the two samples were compared accompany these results. The pattern of multiple injury was comparable between the sexes for each sample, and in all but one case, the males showed a greater or equal prevalence of multiple injury than the females, although the differences were insignificant. Multiple injury was higher among the rural adults when compared to the urban group for all cohorts and was most frequent Copyright  2002 John Wiley & Sons, Ltd. in the young and old adult categories, while multiple injuries were highest among Kerma’s middle-aged adults for both sexes. A significant difference existed between all sexed age groups when the samples were compared, except for the subadults of both sexes and middle-aged females. There was no significant difference in the age distribution of multiple injuries when the sexes were examined separately for each sample (df = 3). Table 3 presents the raw counts of adults with single or multiple injuries by mechanism (violence-related injury, accident-related injury, and unknown mechanism) as determined in the previous analyses (Judd, 2000). If injuries attributed to both accidental and violent mechanisms were identified, the individual was classed as having one or more violence-related injuries. The modal distributions of the frequencies are depicted graphically in Figure 4. Individuals with injuries diagnostic of a fall on an outstretched hand were the minority for all categories. People with multiple injuries experienced a higher frequency of skeletal indicators of violence in comparison to those who had an isolated violencerelated injury (NDRS: χ 2 = 1.22, p = 0.036; Kerma: χ 2 = 7.04, p = 0.008). Forearm injuries, diagnostic of an accident, were not significantly associated with multiple trauma in either group, although they occurred more frequently in an aggregate of injuries than in isolation. No significant difference was observed for violenceor accident-related injury and the presence or absence of other trauma when the samples were compared. When the involvement of minor bones (hands, feet, and ribs) with violence-related multiple injuries was assessed, hand and/or foot injuries occurred more often among the rural group than the urban group (χc2 = 4.06, p = 0.011). Among individuals with accident-related multiple injuries, only the hands were involved more frequently among the rural people in comparison to the urban people (χc2 = 2.93, p = 0.029). There was no significant difference between the sites in the involvement of the ribs with injuries of violence or accident. Table 4 tabulates the raw counts of individuals with skull and long bone injuries by mechanism, and Figure 5 displays the modal distributions of their frequencies. Although substantially more elements were associated with violence for most of Int. J. Osteoarchaeol. 12: 89–106 (2002) M. Judd 96 100% 90% 80% Percent of injured adults 70% 60% 50% 40% 30% 20% 10% m a rie ju in le tip M M ul ul tip O le ne in in ju ju rie ry K sN sK er er m RS D RS D N ry ju in ne O a 0% Injury distribution by mechanism Violence Accident Other Figure 4. Distribution of injury by mechanism (all injuries) for the NDRS and Kerma samples. Table 3. Distribution of injured adults with any injuries (raw counts) by presence of possible injury mechanism among the NDRS and Kerma samples Sample Injury Mechanism One Injury Multiple injuries NDRS Violence-related injuries Accident-related injuries Unknown injury mechanisms 0 0 9 12 2 20 Total individuals Kerma 9 34 Violence-related injuries Accident-related injuries Uncertain forearm injuries Unknown injury mechanisms 12 3 2 31 21 4 1 14 Total individuals 48 40 Copyright  2002 John Wiley & Sons, Ltd. the injury cohorts, no significant differences were noted between the number of injuries incurred and the mechanisms for either site, nor were significant differences observed between the sites when mechanism cohorts were contrasted by the numbers of injuries observed. Discussion Bioarchaeologists have made great strides in recording and reporting palaeotrauma, but little attention, if any, has been given to the distribution of multiple trauma at the populational level perhaps for the lack of a comparative clinical model. Int. J. Osteoarchaeol. 12: 89–106 (2002) Ancient Injury Recidivism 97 100% 90% Percent of injured adults 80% 70% 60% 50% 40% 30% 20% 10% 0% One injury NDRS Multiple injuries NDRS One injury Kerma Multiple injuries Kerma Injury distribution by mechanism Violence Accident Other Figure 5. Distribution of injury by mechanism for skull and long bones for the NDRS and Kerma samples. Table 4. Distribution of injured adults with skull and long bone injuries (raw counts) by presence of possible injury mechanism for the NDRS and Kerma samples Sample Injury Mechanism One injury Multiple injuries NDRS Violence-related injuries Accident-related injuries Unknown injury mechanisms Total individuals 6 1 6 13 6 1 1 8 Kerma Violence-related injuries Accident-related injuries Uncertain forearm injuries Unknown injury mechanisms Total individuals 22 6 2 4 34 11 1 1 0 13 Clinical investigations do not assess the accumulation of trauma in the sense that bioarchaeologists do, but within the last decade a topical trend in clinical research has been the study of chronic, recurrent trauma, referred to by clinicians as ‘‘injury recidivism,’’ and this concept lends itself to the assessment of accumulated injuries among Copyright  2002 John Wiley & Sons, Ltd. ancient people. Clinicians developed a profile of the injury recidivist, and few exceptions exist. The habitual injury recidivist was male, received the first injury by about 20 years of age and the second injury before the age of 30 years, and a similar mechanism was involved. Socially, the injury recidivists came from a lower socioeconomic background and were unemployed ethnic minorities; when injury recidivists who suffered from assault only were examined, substance abuse and criminal activity were also influential factors. Location of residence was inconsequential, and although some researchers found that injury recidivists were predisposed to both accident and assault injuries, most observed that one occurrence of violence-related trauma predisposed the individual to later injuries, which in some cases for very active injury recidivists, proved fatal. This investigation examined three aspects of the clinical injury recidivist profile among two ancient skeletal samples: demographic (age and sex distribution of injury), residential (rural or urban), Int. J. Osteoarchaeol. 12: 89–106 (2002) 98 and finally, involvement of palaeopathological indicators of violence (cranial fracture and direct force forearm trauma) and accident (indirect force forearm trauma). The demographic pattern The pattern of age-related multiple trauma followed identical paths for the sexes within each sample, and all rural age groups consistently displayed more multiple injuries regardless of sex. Males comprised the greatest proportion of multiinjured adults for both samples, which adhered to the clinical pattern for injury recidivism. The lower prevalence of multiple injury among the subadults for both samples concurred with the clinical model as most would not yet have encountered their first or second injury. The rural group sustained more injuries, generally to the extremities, while injuries among the urban youth were diverse and involved the skull and long bones. In the urban sample, three individuals under 25 years, all female, suffered violence-related injuries, but in only one case were the injuries multiple. It is after this age that significantly more rural adults exhibited multiple trauma. Because the demographic distributions and local landscapes were identical for each sample, this variation must be due to behaviour or changing social roles with increased age (e.g., Mays, 2000), which may have influenced the individual’s environmental exposure and activity, rendering the rural group more vulnerable to injury, while being comparatively more ‘‘protective’’ of the urbanites. The rural and urban residence pattern The multiple injury pattern observed at Kerma contrasted noticeably to that of the rural group. Not only was there a greater prevalence of multiple injury among the rural cohorts, but where the rural sample exhibited a bimodal distribution of multiple injury prevalence, only one peak was observed at Kerma, and that was among the middle-aged adults. The injury distribution displayed among the rural group was similar to some clinical findings, where two peak periods of increased trauma were observed, and also Copyright  2002 John Wiley & Sons, Ltd. M. Judd reflected general trauma trends (e.g., Buhr and Cooke, 1959). The first peak occurred among the economically active and adventurous young adults, while a resurgence in multiple trauma occurred among the oldest cohort. Adults under 35 years of age composed 52% of the NDRS skeletal sample and sustained 53% of the multiple injuries, which was similar to the Kerma group where a nearly identical number of young adults (53%) accounted for 45% of the multiple injuries. While this may reflect the supposedly lower life expectancy of past peoples, it may also be a factor of injury recidivism, possibly leading to a premature death in some cases, particularly for those involved in military activity. The Nubians were renowned for their archery ability and skill with a dagger, and in many cases, people from Kerma were interred with their weapons (Bonnet, 1990; Fischer, 1961; Reisner, 1923a, 1923b, 187–194). The low prevalence of accumulated injury among the urban elders was in sharp contrast to the accumulated injuries (none of which were age-related) born by all NDRS adults over 50 years and, in fact, 66% of the older urbanites experienced no trauma at all. This contradicts the clinical expectation that injuries increase in frequency as the years accumulate and the individual survives (e.g., Buhr and Cooke, 1959; Zylke, 1990). In clinical practice, the increase in trauma among elderly people is recognized as a consequence of age-related sensory deterioration (e.g., loss of sight and hearing) that renders them more vulnerable to accident, and age-related loss of bone mass that predisposes the weakened bone to fracture during a fall (Buhr and Cooke, 1959; Matkovic et al., 1993; Stini, 1990). All of the rural adults over 50 years old sustained multiple lesions, however, the fractures that are most typical of age-related falls, for example, femoral head and distal radius fractures, were absent. In clinical research, the absence of age-related injury among older rural adults has been attributed to their more physically demanding rural lifestyle, as high levels of activity throughout an individual’s lifetime are thought to enhance bone strength, and thus increase resistance to the fractures associated with increasing age (e.g., Agarwal, 1980; Jónsson et al., 1992; Sernbo et al., 1988; Stini, 1990). Mays (2000) suggested that once old age was attained, Int. J. Osteoarchaeol. 12: 89–106 (2002) Ancient Injury Recidivism exposure to risky activities was reduced, and that the individual’s injury pattern observed reflected the accumulated fractures acquired during the activities of youth and middle age. While this may have been true among Mays’ urban skeletal sample from Victorian London, individuals that live and work in an agricultural environment continue to labour well into old age in modern society (e.g., Purschwitz and Field, 1990), and likely older adults continued to contribute to the rural family’s livelihood in the past. Many of the injuries sustained by the rural people may reflect the hazardous nature of the ancient rural environment and its increased potential for accident (Alvrus, 1999; Judd and Roberts, 1999; Molleson, 1992). That many urban people (59%), did not suffer from any visible osseous injury attests to the less demanding lifestyle—for most. Because Kerma was the state centre, it was also the focus of less physically exerting administrative, religious, and craft activities, which may explain the lower prevalence of injury among the older urban adults. Seventy percent of the urban individuals were excavated from the ‘‘royal tombs’’ and therefore, likely experienced a longer, more gentile lifestyle in keeping with their status as members of the royal family, administrators, court personnel, business people, or religious specialists, although some had military affiliations (Bonnet, 1990), which may have predisposed them to an earlier death. A feature of the royal burials was the ‘‘sacrificial corridor’’ deemed to be the final resting place of an assortment of the ruler’s retainers or slaves (Bonnet, 1990; Kendall, 1997; Reisner, 1923a), and although the social role of these individuals is not fully understood, they may have enjoyed a less strenuous daily routine, particularly in their old age, relative to that of their rural neighbours (see, Owsley et al., 1987, for differences in urban and rural slave labour in historic North America). 99 proximate mechanisms (direct, indirect, stress, and pathological forces) and can only attempt to determine if healed injuries were due to violence or accident, at best. Forearm and skull (Figure 6) injuries are the most useful lesions for diagnosing the ultimate mechanism in broad terms (e.g., violence or accident) for archaeological samples, but even then we are restricted to speculation, since any injury may be due to either intentional or nonintentional causes. The Colles’ and Galeazzi (Figure 7) fractures of the forearm, for example, are universally associated with a fall on an outstretched hand (e.g., Adams and Hamblen, 1992; Buhr and Cooke, 1959; Loder and Mayhew, 1988; Rogers, 1992, and many others), but whether the fall was catalysed by a third party during a confrontation or was the result of a fall due to clumsiness eludes the investigator (e.g., Ebong, 1978). Likewise, the parry fracture (Figure 8), notorious for its misuse in bioarchaeological interpretation (Jurmain, 1999; Lovell, 1997), may indeed have been the result of blocking a blow from an attack or the result of having an offensive blow deflected, but alternatively it may have been obtained from a defensive gesture to protect the head from a falling object or when falling against a protruding object. The consistent location and configuration of the ‘‘possible’’ parry fractures in these two samples, however, suggested that the injuries were intentional and the result of a similar mechanism; a more random distribution of injuries would implicate an accidental proximate injury mechanism (Galloway, 1999a). Injuries of nonlethal violence In the clinical analysis of injury recidivism, injuries are classified as being due to assault, falls, motor vehicle accident, burns, and so on. In palaeotrauma analysis we are limited to Copyright  2002 John Wiley & Sons, Ltd. Figure 6. Depressed skull fracture of the left frontal bone. Int. J. Osteoarchaeol. 12: 89–106 (2002) 100 Figure 7. Indirect force fracture of the left radial midshaft (anterior view). The skull and forearm injuries The previous trauma analyses (Judd, 2000, Table 6.9) revealed no significant differences in the prevalence of skull or direct force forearm injuries among individuals between the NDRS (skull = 22.9%, forearm = 10.9%) and Kerma (skull = 13.9%, forearm = 7.6%) samples. When multiple injury patterns that involved all bones were scrutinized, however, a different portrait of the past behaviour emerged, which revealed that the NDRS people bore a significantly higher frequency of multiple trauma (61.8%) than the Kerma group (17.9%) (χ 2 = 43.49, p < 0.000). The rural adults with skull or direct force forearm fractures invariably had other lesions, while among the Kerma group, only 64% (21 out of 33) of individuals with violence-related injuries incurred multiple trauma. A greater percentage of Copyright  2002 John Wiley & Sons, Ltd. M. Judd Figure 8. Direct force fracture (parry fracture) of the left ulna (anterior view). multi-injured rural people, therefore, were prone to more extensive injuries from a single incident than the urban group based on these indicators. Alternatively, from a recidivistic perspective, the presence of violence-related fractures among the rural group signalled the presence of other injuries, which were acquired at some point in their lives, whether accidental or intentional, as predicted from clinical models. The presence of an accident-related forearm injury was not significantly associated with multiple trauma at either site. Some bioarchaeologists argue that the presence of skull injuries are reliable indicators of interpersonal violence, while parry lesions are questionable and may be due to any number of causes (e.g., Jurmain, 1999; Jurmain and Kilgore, 1998; Smith, 1996, 1997). Interestingly, injuries involving isolated ulna shaft fractures are nearly always due to fending off a blow in clinical Int. J. Osteoarchaeol. 12: 89–106 (2002) Ancient Injury Recidivism cases (e.g., DuToit and Gräbe, 1979; Heppenstall, 1980, 496; Rogers, 1992, 816, 828), and no correlation between skull fracture and forearm injury within a sample appears with regularity in the clinical literature of assault injury. In this investigation, direct force ulna fractures (parry fracture) and skull injuries were, in most cases, mutually exclusive (except for two out of ten cases in the NDRS sample and five out of 21 urban cases). This subset of seven individuals with both skull and parry fractures accounted for 23% of all adults with a violence-associated fracture, and all were male except for two urban females aged 25 to 35 years old. Among four of the middle-aged males, the accumulated injuries were quite extensive and involved other long bones as well. The dual presence of these injuries does not necessarily mean that they, along with other minor lesions, occurred simultaneously, although the possibility of a particularly devastating incident cannot be ignored. Perhaps one of the best examples of an injury complex that may have occurred during a single episode was that born by a middleaged female who suffered a Colles’ injury on the right radius with an associated ulna fracture to the right forearm, in addition to oblique injuries of the right clavicle and fourth metacarpal shaft—an injury cluster associated with a fall on an outstretched hand (Loder and Mayhew, 1988). It is essential that bioarchaeologists acknowledge that fractures account for at most 40% of injuries, whether accidental or intentional, while the remainder of injuries observed in emergency and trauma units affect the soft tissue only, and the amount of discernible skeletal injuries may be distorted by the weapon preference of the culture (e.g., Butchart and Brown, 1991; Chalmers et al., 1995; Geldermalsen, 1993; Geldermalsen and Stuyft, 1993; Khalil and Shaladi, 1981; Matthew et al., 1996; Shepherd et al., 1987, 1988, 1990). Likewise, soft tissue injuries may be responsible for death, but will not be visible in the skeletal record, a point that has been widely overlooked in bioarchaeological research. Therefore, the total number of injuries observed and the role of injury in the individual’s death will always be underestimated unless the soft tissue is preserved with the skeletal remains. Copyright  2002 John Wiley & Sons, Ltd. 101 The role of minor injuries An examination of minor injuries in the injury suite displayed by the individual may provide further insight into the individual’s lifestyle and susceptibility to accident or aggressive incidents. The problems associated with differentiating between ulna stress and specific parry fracture configurations were discussed in the earlier trauma analysis (Judd, 2000). In clinical practice these two lesions are identical and the only trait that distinguishes them is their injury aetiology, which is reported by the patient. Because ulna stress injuries are associated with heavy lifting (Evans, 1955; Hamilton, 1984; Kitchin, 1948; Koskinen et al., 1997), the coexistence of other liftingrelated injuries, such as spondylolysis of the fifth lumbar vertebrae (Kennedy, 1989; Merbs, 1989; Morris and Blickenstaff, 1967, 184–185), suggested that the ulna parry fractures of three individuals may have been due to labour rather than defence. Among rural individuals, more multiple injuries were representative of minor skeletal indicators of violence. The role of the extremities in palaeotrauma interpretation cannot be ignored as they also have been shown to be predictors of injury recidivism due to violence, notably the fifth metacarpal neck fracture (known as the ‘‘boxer’s fracture’’) (Greer and Williams, 1999), although the mechanism of this fracture is ambiguous. While only one of many punchrelated hand injuries (Adams and Hamblen, 1992; Cailliet, 1975; Jonge et al., 1994; Kraemer and Gilula, 1992a, 1992b; Rogers, 1992), 61% of the ‘‘boxer’s fractures’’ in Greer and Williams’ (1999) investigation were attributed to punching a person or object, while 24% were due to falls or sports. Among the cases that they observed, 27% were injury recidivists, but no difference was detected between recidivism due to intentional or unintentional injuries. Fractures to any metacarpal neck region that result in the dorsal inclination of head are the most common metacarpal injuries in emergency units and are caused by hitting a surface with the fist (e.g., Bora, 1980, 588–589; Galloway, 1999b, 154), but like the ‘‘boxer’s fracture’’ these injuries may be contracted from a fall or sports. More importantly, the second or third metacarpal necks are the Int. J. Osteoarchaeol. 12: 89–106 (2002) M. Judd 102 more frequent locations for punch-related injuries among professional fighters (Bora, 1980, 589; Galloway, 1999b, 154) and, therefore, should also be considered particularly when the profession of the individual was unknown. Among the ancient Nubians, 14 people had multiple injuries with metacarpal involvement. Of these, three males with other violence-related injuries displayed broadly defined ‘‘boxer’s fractures’’ (a lesion on the neck of one or more of metacarpals 2 to 5), while four other individuals (including one female) without violence-related injuries exhibited these fractures. In these samples, the presence of the ‘‘boxer’s fracture’’ did not aid in differentiating between mechanisms, and to assign interpersonal violence as the injury mechanism based on the ‘‘boxer’s fracture’’ would be highly speculative in palaeotrauma analysis. Other seemingly minor injuries were those of the ribs, which are often implicated in cases of abuse (Walker et al., 1997; Wladis et al., 1999), but may also result from falls, accident, stress due to coughing or activity, or even birth (e.g., DeMaeseneer et al., 2000; Galloway, 1999b, 106–109; Sinha et al., 1999; Walker et al., 1997). The angle of the fracture line and location of the lesions, however, aids in identifying the proximate injury mechanism (Galloway, 1999b, 107). The oblique fracture line caused by an indirect force, such as a fall, typically occurs at the rib’s posterior angle, and if the lesions are bilateral the injury may be the result of crushing. Transverse lesions are more often due to localized blows to the chest or coughing, and may involve one or more ribs. In these two samples, four males, three urban and one rural, with violence-related injuries had rib lesions, but all were angled. One urban male sustained angled rib lesions in addition to a distally fractured fibula and metatarsal stress fracture, both of which are locomotor injuries (Adams and Hamblen, 1992, 246; Black, 1983; Linenger and Shwayhat, 1992; Rogers, 1992, 1340–1341). Among females, one urbanite sustained posterior angled rib lesions as well as clavicular and two extremity injuries, while six other females, of which five were urban, experienced minor, but transverse lesions to the sternal rib ends; their associated injuries were minor as well (e.g., impacted phalangeal articular surfaces). Copyright  2002 John Wiley & Sons, Ltd. Ancient injury recidivists? When involvement of the minor bones was considered, the hands and feet were implicated in interpersonal violence among the NDRS group, which supports the findings of the previous general trauma analysis that the extremities may have been involved more frequently in interpersonal confrontations among the rural group in contrast to the urban group (Judd, 2000). However, when only the skull and long bones were considered, as is common in most populational studies of palaeotrauma, the occurrence of another injured major bone accompanying a skull or direct force ulna fracture was similar for both samples. Among the NDRS group six out of 12 (50%) individuals with one injury of violence had a second lesion on the same bone or had one or more additional injured major bones, while among the Kerma group 11 out of 33 (33%) people with one injury of violence had an additional lesion on a major bone. That any of these other injured bones occurred preceding to, subsequent to, or simultaneously with the violence-related injury remains unknown, but two conclusions can be drawn from this investigation. First, the majority of skull or parry fractures did not have another major fractured bone present among the skeletal remains, although the frequency was high, particularly for the rural group. Second, had the injuries occurred independently, the range of 33% to 50% of individuals with multiple major bone injuries rests within clinical injury recidivism rates. This higher range is reasonable, as the archaeological skeleton presents a lifetime of accumulated injury rather than injuries collected over a short study period, for example, one to five years, as is typical among clinical research. When forearm injuries that were diagnostic of falls were considered, all of the rural individuals (n = 2) and four out of seven urban people (57%) had additional injuries to minor bones, but when only the major bones were assessed, the range dropped to 50% (one out of two) for the rural people and 14% (one out of seven) for the urban group. This variation between the samples was likely the result of the rural lifestyle and subsistence activities. Int. J. Osteoarchaeol. 12: 89–106 (2002) Ancient Injury Recidivism Among the pooled sample, more individuals with violence-related injuries experienced additional trauma to other major bones (38%, 17 out of 45 adults), than did those with accident-related trauma (22%, two out of nine adults), although this difference was not significant. Seventeen out of 21 (81%) adults with multiple injuries to major bones bore one or more violence-related injuries, while 60% (28 out of 47) of the adults with single injuries displayed a violence-related injury, a difference that also was insignificant. Nevertheless, in these two societies, individuals with major violence-related injuries were more likely to have additional long bone or minor bone injuries than people with accident-related lesions, and a segment of this group were likely injury recidivists. Conclusions In the clinical arena, individuals that continually present trauma are referred to as injury recidivists, and are profiled as young ethnic minority males who are unemployed, suffer from socioeconomic problems, and have at least one injury caused by violence. By analogy, multiple injuries sustained by ancient people also may be the result of repeat injuries rather than a single event, and in some cases, may account for a premature death. This investigation evaluated the distribution of multiple injuries between two ancient Nubian skeletal samples from the Kerma period (ca. 2500–1500 BC), one rural and one urban, to determine if characteristics of the injury recidivism profile, developed by modern clinical researchers, existed in past societies. While injury recidivism cannot be established unconditionally among ancient societies, the pattern of multiple injury among the Kerma culture corresponded to the clinical profile of injury recidivism in many respects: 1. most individuals with multiple injuries were male and less than 35 years of age at the time of death, 2. no significant difference in violence- or accident-related multiple injury was apparent between the rural and urban communities, 3. a high frequency of multi-injured adults displayed one or more skeletal indicators of nonlethal violence. Copyright  2002 John Wiley & Sons, Ltd. 103 The topical area of injury recidivism among clinical researchers promises to be a viable resource for palaeotrauma analysis. Acknowledgements This research received financial support from the Social Sciences and Humanities Research Council of Canada (SSHRC Award 752-96-1319), the Boise Fund (Institute of Biological Anthropology, University of Oxford), the Sudan Archaeological Research Society, the Faculty of Graduate Studies and Research (University of Alberta) and Department of Anthropology (University of Alberta). Director General Hassan Hussein Idriss of the National Corporation for Antiquities and Museums in Khartoum, Sudan is thanked for his support and assistance with the North Dongola Reach Survey Project. Mr Vivian Davies of The British Museum’s Department of Egyptian Antiquities generously permitted access to the NDRS skeletal collection. Dr Robert A. Foley allowed me to examine the Kerma skeletal remains at the Department of Biological Anthropology at the University of Cambridge and Maggie Bellatti, the osteological technician, was most accommodating during my time in Cambridge. Mr Cyril Chan (Office of the Chief Medical Examiner, Edmonton, Alberta) gave his time and skills in x-raying the NDRS skeletal material. I am most grateful to Dr Derek Welsby of the Department of Egyptian Antiquities of The British Museum for his discussions on Nubian culture history and for offering me with the opportunity to excavate in Sudan. Members of the NDRS team (1994–97) and the people of the Dongola vicinity are thanked for excavating the skeletal remains for this project. Drs Nancy Lovell, Owen Beattie, M. Anne Katzenberg, Charles Merbs, and Pamela Willoughby provided helpful advice and discussion of this manuscript. References Adams J, Hamblen D. 1992. Outline of Fractures Including Joint Injuries, 10th edition. Churchill Livingstone: New York. Int. J. Osteoarchaeol. 12: 89–106 (2002) 104 Agarwal RP. 1980. Pattern of bone injuries in a hill area (Pauri-Garhwal). Journal of the Indian Medical Association 74: 65–66. Alvrus A. 1999. Fracture patterns among the Nubians of Semna South, Sudanese Nubia. International Journal of Osteoarchaeology 9: 417–429. Black J. 1983. Stress fractures of the foot in male soldiers. Journal of the American Podiatry Association 73: 633–634. Bonnet C. 1990. Sépultures et coutumes funéraires. In Kerma Royaume de Nubie: L’Antiquité Africaine au Temps des Pharaons, Bonnet C (ed.). Mission archéologique de l’Université de Genève au Soudan: Genève; 69–86. Bora FW. 1980. Injuries of the hand. In Fracture Treatment and Healing, Heppenstall RB (ed.). W B Saunders: Toronto; 558–610. Buhr AJ, Cooke AM. 1959. Fracture patterns. The Lancet March 14 531–536. Buikstra JE, Ubelaker DH. 1994. Standards for Data Collection from Human Skeletal Remains. Arkansas Archaeological Survey Research Series, Volume 44. Arkansas Archaeological Survey: Fayetteville. Burrell LL, Maas MC, Van Gerven DP. 1986. Patterns of long-bone fracture in two Nubian cemeteries. Human Evolution 1: 495–506. Butchart A, Brown D. 1991. Non-fatal injuries due to interpersonal violence in Johannesburg-Soweto: incidence, determinants and consequences. Forensic Science International 52: 35–51. Cailliet R. 1975. Hand Pain and Impairment. F A Davis Company: Philadelphia. Chalmers DJ, Fanslow JL, Langley JD. 1995. Injury from assault in New Zealand: an increasing public health problem. Australian Journal of Public Health 19: 149–154. Cybulski JS. 1992. A Greenville Burial Ground. Canadian Museum of Civilization: Hull, Quebec. DeMaeseneer M, DeMey J, Debaere C, Meysman M, Osteaux M. 2000. Rib fractures induced by coughing: an unusual cause of acute chest pain. American Journal of Emergency Medicine 18: 194–197. Dowd MD, Langley J, Koepsell T, Soderberg R, Rivara FP. 1996. Hospitalizations for injury in New Zealand: prior injury as a risk factor for assaultive injury. American Journal of Public Health 86: 929–934. DuToit FP, Gräbe RP. 1979. Isolated fractures of the shaft of the ulna. South African Medical Journal 56: 21–25. Ebong W. 1978. The pattern of fractures and dislocations in Western Nigeria. Injury 9: 221–224. Evans DL. 1955. Fatigue fractures of the ulna. Journal of Bone and Joint Surgery (British) 37: 618–621. Fischer HG. 1961. The Nubian mercenaries of Gebelein during the First Intermediate Period. Kush 9: 44–80. Copyright  2002 John Wiley & Sons, Ltd. M. Judd Galloway A. 1999a. The circumstances of blunt force trauma. In Broken Bones, Galloway A (ed.). Charles C. Thomas: Springfield, IL. Galloway A. 1999b. Fracture patterns and skeletal morphology. In Broken Bones: Anthropological Analysis of Blunt Force Trauma, Galloway A (ed.). Charles C Thomas: Springfield, IL. Geldermalsen AAV. 1993. Assault, a burden for a rural hospital in Lesotho. Tropical and Geographical Medicine 45: 30–32. Geldermalsen AAV, Stuyft PVD. 1993. Interpersonal violence: patterns in a Basotho community. Journal of Tropical Medicine and Hygiene 96: 93–99. Goins WA, Thompson J, Simpkins C. 1992. Recurrent intentional injury. Journal of the National Medical Association 84: 431–435. Greer SE, Williams JM. 1999. Boxer’s fracture: an indicator of intentional and recurrent injury. American Journal of Emergency Medicine 17: 357–360. Hamilton HK. 1984. Stress fracture of the diaphysis of the ulna in a body builder. American Journal of Sports Medicine 12: 405. Hedges BE, Dimsdale JE, Hoyt DB, Berry C, Leitz K. 1995. Characteristics of repeat trauma patients, San Diego County. American Journal of Public Health 85: 1008–1010. Heppenstall RB. 1980. Fractures of the forearm. In Fracture Treatment and Healing, Heppenstall RB (ed.). W B Saunders: Philadelphia; 481–503. Jonge JD, Kingma J, Lei BVD, Klasen H. 1994. Phalangeal fractures of the hand. Journal of Hand Surgery (British and European Volume) 19B: 168–170. Jónsson B, Gärdsell P, Johnell O, Redlund-Johnell I, Sernbo I. 1992. Differences in fracture pattern between an urban and a rural population: a comparative population-based study in Southern Sweden. Osteoporosis International 2: 269–273. Judd MA. 2000Trauma and Interpersonal Violence in Ancient Nubia during the Kerma Period (ca. 2500–1500 BC).. PhD Dissertation, Department of Anthropology, University of Alberta. Judd MA, Roberts CA. 1998. Fracture patterns at the medieval leper hospital in Chichester. American Journal of Physical Anthropology 105: 43–55. Judd MA, Roberts CA. 1999. Fracture trauma in a medieval British farming village. American Journal of Physical Anthropology 109: 229–243. Jurmain R. 1999. Stories from the Skeleton. Behavioral Reconstruction in Human Osteology. Gordon and Breach: Amsterdam. Jurmain R, Kilgore L. 1998. Sex-related patterns of trauma in humans and African apes. In Sex and Gender in Paleopathological Perspective, Grauer AL, StuartMacadam P (eds). Cambridge University Press: Cambridge; 11–26. Int. J. Osteoarchaeol. 12: 89–106 (2002) Ancient Injury Recidivism Kaufmann CR, Branas CC, Brawley ML. 1998. A population-based study of trauma recidivism. Journal of Trauma 45: 325–332. Kendall T. 1997. Kerma and the Kingdom of Kush 2500–1500 BC: The Archaeological Discovery of an Ancient Nubian Empire. National Museum of African Art, Smithsonian Institution: Washington, DC. Kennedy KAR. 1989. Skeletal markers of occupational stress. In Reconstruction of Life from the Skeleton, Iscan MY, Kennedy KAR (eds). Alan R Liss: New York; 129–160. Khalil A, Shaladi OA. 1981. Fractures of the facial bones in the Eastern region of Libya. British Journal of Oral Surgery 19: 300–304. Kilgore L, Jurmain R, Van Gerven D. 1997. Palaeoepidemiological patterns of trauma in a medieval Nubian skeletal population. International Journal of Osteoarchaeology 7: 103–114. Kitchin ID. 1948. Fatigue fracture of the ulna. Journal of Bone and Joint Surgery (British) 30B: 622–623. Koskinen SK, Mattila KT, Alanen AM, Aro HT. 1997. Stress fracture of the ulnar diaphysis in a recreational golfer. Clinical Journal of Sport Medicine 7: 63–65. Kraemer B, Gilula L. 1992a. Metacarpal fractures and dislocations. In The Traumatized Hand and Wrist. Radiographic and Anatomic Correlation, Gilula L (ed.). W B Saunders Company: Philadelphia; 171–219. Kraemer B, Gilula L. 1992b. Phalangeal fractures and dislocations. In The Traumatized Hand and Wrist. Radiographic and Anatomic Correlation, Gilula L (ed.). WB Saunders: Philadelphia; 105–170. Lahren CH, Berryman HE. 1984. Fracture patterns and status at Chucalissa (4OSY1): a biocultural approach. Tennessee Anthropologist 9: 15–21. Linenger J, Shwayhat A. 1992. Epidemiology of podiatric injuries in US marine recruits undergoing basic training. Journal of the American Podiatric Medical Association 82: 269–271. Loder R, Mayhew H. 1988. Common fractures from a fall on an outstretched hand. American Family Physician 37: 327–328. Loth S, Iscan M. 1989. Morphological assessment of age in the adult: the thoracic region. In Age Markers in the Human Skeleton, Iscan M (ed.). Charles C Thomas: Springfield, IL; 105–135. Lovejoy CO, Heiple KG. 1981. The analysis of fractures in skeletal populations with an example from the Libben Site, Ottowa, Ohio. American Journal of Physical Anthropology 55: 529–541. Lovejoy CO, Meindl RS, Pryzbeck TR, Mensforth RP. 1985. Chronological metamorphosis of the auricular surface of the ilium: a new method for the determination of adult skeletal age at death. American Journal of Physical Anthropology 68: 15–28. Copyright  2002 John Wiley & Sons, Ltd. 105 Lovell NC. 1990. Patterns of Injury and Illness in Great Apes: A Skeletal Analysis. Smithsonian: Washington. Lovell NC. 1997. Trauma analysis in paleopathology. Yearbook of Physical Anthropology 40: 139–170. Madden C, Garrett JM, Cole TB, Runge JW, Porter CQ. 1997. The urban epidemiology of recurrent injury: beyond age, race, and gender stereotypes. Academic Emergency Medicine 4: 772–775. Matkovic V, Ilich J, Hsieh L. 1993. Influence of age, sex and diet on bone mass and fracture rate. Osteoporosis International Supplement 1: S20–S22. Matthew PK, Kapua F, Soaki PJ, Watters KAK. 1996. Trauma admissions in the Southern Highlands of Papua New Guinea. Australian and New Zealand Surgery 66: 659–663. Mays S. 2000. Age-dependent cortical bone loss in women from 18th and early 19th century London. American Journal of Physical Anthropology 112: 349–361. Merbs CF. 1989. Spondylolysis: its nature and anthropological significance. International Journal of Anthropology 4: 163–169. Molleson T. 1992. Mortality patterns in the RomanoBritish cemetery at Poundbury Camp near Dorchester. In Death in Towns: Urban Responses to the Dying and the Dead, 100-1600, Bassett S (ed.). Leicester University Press: Leicester; 43–55. Morris JM, Blickenstaff LD. 1967. Fatigue Fractures: A Clinical Study. Charles C Thomas: Springfield, IL. Neves WA, Barros AM, Costa MA. 1999. Incidence and distribution of postcranial fractures in the prehistoric population of San Pedro de Atacama, Northern Chile. American Journal of Physical Anthropology 109: 253–258. Olivier G. 1969. Practical Anthropology. Charles C Thomas: Springfield, IL. Owsley DW, Orser CE, Mann RW, Moore-Jansen PH, Montgomery RL. 1987. Demography and pathology of an urban slave population from New Orleans. American Journal of Physical Anthropology 74: 185–1987. Poole GV, Griswold JA, Thaggard VK, Rhodes RS. 1993. Trauma is a recurrent disease. Surgery 113: 608–611. Purschwitz MA, Field WE. 1990. Scope and magnitude of injuries in the agricultural workplace. American Journal of Industrial Medicine 18: 179–192. Reiner DS, Pastena JA, Swan KG, Lindenthal JJ, Tischler CD. 1990. Trauma recidivism. American Surgeon 56: 556–560. Reisner GA. 1923a. Excavations at Kerma. Parts I-III. Harvard African Studies: Cambridge, MA. Reisner GA. 1923b. Excavations at Kerma. Parts IV-V. Harvard African Studies: Cambridge, MA. Int. J. Osteoarchaeol. 12: 89–106 (2002) 106 Robb J. 1997. Violence and gender in Early Italy. In Troubled Times: Violence and Warfare in the Past, Martin DL, Frayer DW (eds). Gordon and Breach: Amsterdam; 111–144. Roberts CA. 1988. Trauma and Its Treatment in British Antiquity. PhD Dissertation, Department of Archaeological Sciences, University of Bradford. Rogers L. 1992. Radiology of Skeletal Trauma, (2nd edn). Churchill Livingstone: New York. Sayfan J, Berlin Y. 1997. Previous trauma as a risk factor for recurrent trauma in rural northern Israel. Journal of Trauma 43: 123–125. Sernbo I, Johnell O, Anderson T. 1988. Differences in the incidence of hip fracture: comparison of an urban and rural populations in Southern Sweden. Acta Orthopaedica Scandinavica 59: 382–385. Shepherd J, Al-Kotany MY, Subadan C, Scully C. 1987. Assault and facial soft tissue injuries. British Journal of Plastic Surgery 40: 614–619. Shepherd J, Gayford J, Leslie I, Scully C. 1988. Female victims of assault. Journal of Cranio-maxillary-facial Surgery 16: 233–237. Shepherd J, Shapland M, Pearce NX, Scully C. 1990. Pattern, severity and aetiology of injuries in victims of assault. Journal of the Royal Society of Medicine 83: 75–78. Sims DW, Bivins BA, Obeid FN, Horst HM, Sorensen VJ, Fath JJ. 1989. Urban trauma: a chronic and recurrent disease. Journal of Trauma 29: 940–946. Sinha AK, Kaeding CC, Wadley GM. 1999. Upper extremity stress fractures in athletes: clinical features of 44 cases. Clinical Journal of Sports Medicine 9: 199–202. Smith MO. 1996. Parry fractures and female-directed interpersonal violence: implications from the Late Archaic Period of West Tennessee. International Journal of Osteoarchaeology 6: 84–91. Smith MO. 1997. Osteological indications of warfare in the Archaic period of the Western Tennessee Copyright  2002 John Wiley & Sons, Ltd. M. Judd Valley. In Troubled Times: Violence and Warfare in the Past, Martin DL, Frayer DW (eds). Gordon and Breach: Amsterdam; 241–266. Smith RS, Fry WR, Morabito DJ, Organ CH. 1992. Recidivism in an urban trauma center. Archives of Surgery 127: 668–670. Stini WA. 1990. Osteoporosis: etiologies, prevention and treatment. Yearbook of Physical Anthropology 33: 151–194. Todd TW. 1921a. Age changes in the pubic bone. I: The male white pubis. American Journal of Physical Anthropology 3: 285–334. Todd TW. 1921b. Age changes in the pubic bone. III: The pubis of the white female. IV: The pubis of the female white-negro hybrid. American Journal of Physical Anthropology 4: 1–70. Walker PL, Cook DC, Lambert PM. 1997. Skeletal evidence for child abuse: a physical anthropological perspective. Journal of Forensic Sciences 42: 196–207. Welsby DA. 1996. The North Dongola Reach Survey, a. the 1995/6 season. The Sudan Archaeological Research Society Newsletter 10: 2–9. Welsby DA. 1997. The North Dongola Reach Survey: the 1996/7 season. Sudan & Nubia 1: 2–10. Williams JM, Furbee PM, Hungerford DW, Prescott JE. 1997. Injury recidivism in a rural ED. Annals of Emergency Medicine 30: 176–180. Wladis A, Bostrom L, Nilsson B. 1999. Unarmed violence-related injuries requiring hospitalization in Sweden from 1987 to 1994. Journal of Trauma 47: 733–737. Wood JW, Milner GR, Harpending HC, Weiss KM. 1992. The osteological paradox. Current Anthropology 33: 343–370. Zylke JW. 1990. As nation grows older, falls become greater source of fear, injury, death. Journal of the American Medical Association 263: 2021. Int. J. Osteoarchaeol. 12: 89–106 (2002)