Paleoepidemiolgical patterns of trauma in a prehistoric population from central California.код для вставкиСкачать
AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 115:13–23 (2001) Paleoepidemiolgical Patterns of Trauma in a Prehistoric Population From Central California Robert Jurmain* Department of Anthropology, San Jose State University, San Jose, California 95192-0113 KEY WORDS fracture; cranial injury; projectile wound ABSTRACT Skeletal trauma was investigated in a large collection of human remains from central California (N ⫽ 162 aged and sexed adults). Lesions investigated included cranial and long bone fractures, projectile wounds, and dislocation. Long bone fractures were found in 10.5% of individuals; overall, incidence by element was 2.3%. In addition, cranial injuries were found in 4.4% of complete adult crania. Projectile wounds were seen unambiguously in four individuals (with embedded obsidian fragments) and strongly suggested in two other individuals with partially healed lesions. Finally, one case of traumatic hip dislocation was also observed. In both incidence and patterning of injuries, this population is similar to other archeological groups from California. This evidence further supports earlier reports indicating that interpersonal aggression was quite common in prehistoric California. Am J Phys Anthropol 115:13–23, 2001. Traumatic injuries produce some of the most dramatic skeletal lesions seen in paleopathology. As reported in case studies, they can sometimes inform us of significant events in the lives of individuals. While intriguing as this information is, even more useful is the understanding of how various types of traumatic injuries are patterned within and among human populations. This report describes patterns of traumatic injuries in a large well-preserved collection of human remains from CA-SCl-038, a site located in the Santa Clara Valley, central California. Lesions resulting from cranial and postcranial fractures, weapon wounds, and dislocations were found in these remains, and their epidemiological patterns are generally congruent with other similar reports from prehistoric California. Moreover, in conjunction with these other data from California, evidence from SCl-038 further documents a distinctive epidemiological pattern from this region of North America. Because they are probably the most easily diagnosed and among the most common types of skeletal trauma, fractures have traditionally inspired the most interest from paleopathological investigators. Although individual case reports predominated for much of the twentieth century, some early work was quite systematic. Detailed data on fracture prevalence by individual element as well as comprehensive clinical comparisons were presented by Wood Jones (1910) in his famous analysis of trauma in Lower Nubian skeletal remains. Such systematic analysis of skeletal trauma, however, was not equaled for some time, but the innovative analysis by Lovejoy and Heiple (1981) again stimulated wider interest in population studies of fractures. Other investigations which adopted such an approach and also used some of the same methodological rigor advocated by Lovejoy and Heiple (1981) included those by Bennike (1985), Jurmain (1991), Grauer and Roberts (1996), Kilgore et al. (1997), Judd and Roberts (1999), and Neves et al. (1999). These paleoepidemiological studies focused primarily on postcranial fractures, and there have been, to date, fewer investigations of cranial trauma. However, some notable exceptions do exist (e.g., Bennike, 1985; Walker, 1989, 1997; Webb, 1995; Jurmain and Bellifemine, 1997; Lambert, 1997). Penetrating weapon wounds are another form of serious injury which has been described from various geographic locales. Most notably, from Europe, metal weapon injury from swords or axes has been quite frequently reported (e.g., Brothwell, 1961; Wells, 1964; Bennike, 1985; Waldron, 1994; Anderson, 1996). Similar types of injuries have also been described in the New World by Hutchinson (1996) for a postcontact site in Florida. In prehistoric contexts from North America, weapon wounds most commonly result from lithic projectiles. Primarily anecdotal reports of such projectile injury have come from a variety of locations, including Ohio (Hooton, 1920), Missouri (Black, 1979), New York (Pfeiffer, © 2001 WILEY-LISS, INC. © 2001 Wiley-Liss, Inc. *Correspondence to: Robert Jurmain, Department of Anthropology, San Jose State University, San Jose, CA 95192-0113. E-mail: email@example.com Received 16 May 2000; accepted 31 January 2001. 14 R. JURMAIN Fig. 1. Location of site, CA-SCl-038. 1985), Arizona (Merbs, 1989), South Dakota (Willey, 1990), Illinois (Milner et al., 1991), and Tennessee (Smith, 1996, 1997). In the New World, the most frequent occurrence of such projectile lesions, however, has been observed at sites in California (Tenney, 1986; Jurmain, 1991; Lambert and Walker, 1991; Lambert, 1994, 1995). Indeed, especially from sites in both central and southern California, the incidence of such lesions is as high as for any region in the world. A final category of injury which has only been rarely reported from archaeological contexts is traumatic dislocation. Diagnostic skeletal lesions resulting from dislocation are never common and are usually subtle in manifestation. Thus, descriptions of such lesions in the literature are mostly anecdotal in nature (Wood Jones, 1910; Ortner and Putschar, 1981; Jurmain, 1991; Kilgore et al., 1997). MATERIALS AND METHODS SCl-038, the “Yukisma site,” is a large cemetery located in the northern Santa Clara Valley, approximately 6 miles south of the current edge of San Francisco Bay (Fig. 1). The site was excavated in 1994 as part of an archaeological mitigation program resulting from expansion of the Santa Clara County Elmwood Correctional Facility and was carried out by the Ohlone Families Consulting Services. All osteological analyses were conducted under the supervision of the author in the Physical Anthropology and Archaeology Laboratory at San Jose State University. Radiometric dating of 21 organic samples placed the site’s occupation between ca. 240 BC and AD 1770 (Bellifemine, 1997). Distribution of dated burials as well as sequencing of temporally sensitive artifacts suggest that the site was fairly continuously occupied during this 2,000-year span; however, temporal distinctions of burials within the site are not precise enough to allow sytematic partitioning of the sample. Systematic archaeological excavation retrieved a large number of burials, many of which were excellently preserved. A total of 239 field-designated grave lots contained an estimated 228 individuals. Aging and sexing were done using standard osteological techniques (Ubelaker, 1989; Bass, 1991), with sex assessment augmented by metrical evaluation (Dittrick and Suchey, 1986) as well as determination of mandibular morphology (Loth, 1996). Age determination for adults in this sample was based on pubic symphysis (Katz and Suchey, 1986) and auricular surface (Lovejoy et al., 1985) remodeling. Adults are utilized in this investigation, with adult status based on fusion of all major appendicular epiphyses and/or fusion of the spheno-occipital suture. A total of 162 adult individuals (aged ⬎15 years) could also be reliably sexed. Of these, 102 are male (63%) and 60 (37%) are female (Table 1). The average age at death for males (32.2 years) was significantly less than that estimated for females (37.2 years) (t-test, two-tailed probability ⫽ 0.003; t-value ⫽ ⫺3.07; df ⫽ 144). For all aged individuals in the sample (including subadults; N ⫽ 198), the average age at death was 27.8. Evaluation of fractures and other traumatic lesions was based on gross analysis, supplemented by plain radiography. All elements with suggestive lesions visible on gross examination were radiographed; in addition, a radiographic survey was also done, utilizing a representative sample of long bones from all individuals. Unless otherwise indicated, statistical comparisons were done using chi-square analysis, and, where appropriate, a Yates correction was employed. In the postcranial skeleton, the presence of fracture was determined for the long bones (clavicle, humerus, radius, ulna, femur, tibia, and fibula). Following the methodology proposed by Lovejoy and Heiple (1981), only complete long bones were included in tabulating prevalence; a complete long bone is here defined as being at least two-thirds present, with all major articular areas preserved. Evidence of long bone fracture was confirmed macroscopically through assessment of bilateral asymmetry (including shortening of affected side), angular deformation, and presence of fracture callus. Moreover, radiographic evidence of fracture callus or remnant of original cortex further corroborated diagnosis. Craniofacial injury was assessed similarly, but also included evaluation of cranial vault depression fractures. Likewise, presence of dislocation was diagnosed through gross observation of joint remodeling changes, especially those relating to attachment of the joint capsule and alterations to the articular surface(s). Projectile/perforating wounds were established by the presence of embedded projectile fragments (in 15 PATTERNS OF TRAUMA IN CENTRAL CALIFORNIA TABLE 1. Sex and age distribution of SCl-038 sample Male Female Indeterminate Totals 0–10 11–20 21–30 31–40 41–50 51⫹ “Adult” 0 0 29 29 11 8 19 38 37 3 3 43 34 11 0 45 16 23 1 40 0 4 0 4 4 11 0 15 four cases) or perforations diagnosed on gross examination and radiographically and strongly indicated in two other cases. In no case, however, was wound trauma diagnosed solely on the basis of archaeological context. Although such diagnosis has been suggested by others (e.g., Wendorf, 1968; Lambert, 1994, 1997), the vagaries of archaeological preservation and recovery make such data almost impossible to use in sample comparisons (see further discussion below). RESULTS Long bone fractures A total of 23 long bone fractures was found in 17 different individuals. Among the affected individuals, 5 were female (8.3% prevalence) and 12 were male (11.8% prevalence); however, the difference is not statistically significant. For the combined sample of sexed adults (N ⫽ 162 individuals), prevalence of long bone fracture was 10.5%. Thirteen individuals display one lesion, two individuals have two fractures each, and another two individuals have three long bone fractures. Thus, the proportion of affected individuals with multiple lesions (4/17) is 23.5%. A more explicit way to see the pattern of fracture involvement is to evaluate individual elements. A total of 1,018 complete long bones from sexed adults was available for analysis. Thus, the combined incidence of involvement was 23/1,018 (2.3%). In addition, another 32 complete long bones were found in adult burials of individuals designated as “indeterminate sex.” In the entire sample of adult complete long bones, the combined fracture prevalence was 23/1,050 (2.2%). Table 2 shows the data for long bone fractures by element, side, and sex. The most commonly affected element is the ulna (10/151 ⫽ 6.6%), followed by the radius (3.6%) and humerus (2.1%). The forearm accounts for 69.6% of all long bone fractures, and, overall, the upper appendage for 82.6% of involvement. There is little difference in left (2.1%) as compared to right (2.4%) side involvement. However, for the ulna, there is a predilection for more left side involvement (80% of cases), and this pattern is seen to the same degree in both males and females. Still, given the relatively low prevalence here, this side difference did not reach statistical significance (P ⫽ 0.10). Regarding sex differences, there are also no statistically significant differences for any element, although the strongest predilection is found in the radius, where 5 of the 6 affected cases are among males (P ⫽ 0.50). Patterns of sex differences can also be assessed by body segment (i.e., upper vs. lower TABLE 2. Distribution of long bone fractures by side and sex1 Left Clavicle Male Female Humerus Male Female Radius Male Female Ulna Male Female Femur Male Female Tibia Male Female Fibula Male Female Totals Right N n N n 46 32 0 0 49 32 0 0 50 27 0 0 47 18 2 1 51 34 2 1 49 27 3 0 43 29 4 4 47 25 1 1 41 16 0 0 43 19 2 0 55 25 0 0 53 31 1 0 40 28 517 0 0 11 43 18 501 1 0 12 (2.1%) 1 (2.3%) n, number of fractured elements. appendage). In males, 75% (12/16) of long bone fractures occur in the upper appendage, while females have all (N ⫽ 7) fractures confined to the upper limb. However, this difference is not statistically significant (P ⫽ 0.39). The overall pattern for both sexes, with almost 83% of lesions found in the upper limb, is significant (chi-square ⫽ 4.27; df ⫽ 1; P ⫽ 0.04). The age distribution of affected individuals is shown in Table 3. There is little indication in these data of a significant increase in involvement with advancing age, except that no young adults (⬍21 years) are involved. When looking at age and sex involvement simultaneously, it can be seen that in individuals less than 41 years of age, 91% of fractures are among males; conversely, for older adults (41⫹), 67% of involvement is found in females. However, this apparent differential fracture involvement is largely a product of the age distribution of the sample (see Table 1). Indeed, chi-square analysis of the age-partitioned sample (21– 40; 40⫹) by sex does not reveal a significant sex difference in either age group. Of the 17 involved individuals, 15 could be aged accurately (the remaining two, one male and one female, were each aged as “35⫹”). In these subsamples of affected individuals, males (N ⫽ 11) averaged 35.63 years of age, while females (N ⫽ 4) averaged 43.5 years. When compared to the overall 16 R. JURMAIN TABLE 3. Prevalence of long bone fractures by age and sex1 16–20 21–30 31–40 41⫹ Age N n N n N n N n Male Female Totals 11 8 19 0 0 0 37 3 40 5 0 5 34 11 45 5 1 6 16 27 43 2 4 6 (0%) 1 (12.5%) (13.3%) (14.0%) n, number of fractured elements. Fig. 2. Ununited fracture, midshaft, left ulna. Female, aged 30 –50 years. age data, affected males are slightly older (by 3.4 years), and involved females are somewhat older than the overall female mean (by 6.3 years). A further interesting feature of long bone fracture involvement is that two of the lesions are ununited (in two individuals, involving a left ulnar shaft and a right humeral head). Such lesions have only rarely been reported by paleopathologists (e.g., Stewart (1974) found fewer than 10 such lesions among several thousand burials). Nevertheless, more recent reports of skeletal samples from Nubia (Kilgore et al., 1997) and California (Jurmain 1991) found several such lesions (four from Nubia and five from California). In all these other cases, the affected elements were from the forearm, as was one of the examples from SCl-038 (Fig. 2). However, the other ununited element, a right humeral head, is a most unusual example. The failure to join suggests that the limb was not immobilized during healing, which would be not unusual for the forearm, but most rare for injuries to the arm. In addition to the 23 long bone fractures, another 10 individuals displayed other postcranial fractures. The most serious of these was a healed pelvic fracture (ischio-pubic ramus) in an adult male. Other lesions included two hand fractures (both metacarpals), three rib fractures (two of them in one individual), two vertebral compression fractures, and three foot fractures (two phalanges, one metatarsal). Cranial injuries Apart from the long bone involvement, seven individuals at SCl-038 displayed cranial fractures (one of whom, an adult male, also had a fractured left ulna). From the sample, 159 adult crania were fairly complete (at least 50% present); thus, the prevalence of cranial injury at SCl-038 is 7/159 (4.4%). Five of the seven cases of cranial injury are restricted to the vault, and one to the face, and one individual has both facial and vault involvement. The vault injuries are to the frontal (two cases), left parietal (two cases), and right parietal (two cases). The facial injuries are to the nasal bones (in an individual also with vault injury) and to the malar. The most severe case of cranial injury was to a young female, aged 18 –20 years, who displays a large lytic lesion to her left parietal. The lesion is located near lambda, just left of the midline, and measures approximately 18.3 mm in diameter (Fig. 3). The injury perforates both inner and outer tables, and the edges are very well-rounded (although a small remnant of inner table is still present). Diagnosis suggests perforating trauma, approximately 2 years prior to death; moreover, the individual most likely suffered from severe behavioral/neurological symptoms (Bruce Ragsdale, personal communication, 1995). In the cranium, as in the postcranium, there is not a strong indication of sex bias among injured individuals. Of the 7 affected individuals, 4 are male and 3 are female. Perforating wounds Embedded projectiles are direct evidence of weapon wounds, and four cases of such injury are present at SCl-038 (for adult postcrania yielding a prevalence of 2.5%). All embedded fragments are made of obsidian, and involved individuals include three males and one individual of indeterminate sex. Affected elements include a hand phalanx, a lumbar vertebra (Fig. 4), a rib, and a femur. The embedded fragments are generally small (⬍5 mm), with the exception of a fairly large obsidian projectile (7.9 ⫻ 4.8 mm) in the femur of a young male (age ⬃16 PATTERNS OF TRAUMA IN CENTRAL CALIFORNIA 17 Fig. 5. Healed traumatic hip dislocation, proximal left femur. The head is slightly mushroomed, and the muscle insertions on the greater trochanter are heavily remodeled. Fig. 3. Healed lytic lesion of left parietal, suggestive of a perforating wound. Female, aged 18 –20 years. ing a partially healed lytic lesion most likely resulting from perforating wounds. The first of these, in the cranium of a young female, produced a massive cranial fracture and is discussed above (Fig. 3). The second probable injury was found in the distal right femur of a middle-aged male, aged 30 – 45 years. There are three lytic lesions evident, the largest on the medial condyle measuring 9.6 ⫻ 13.5 mm (the lesion is very round, circumscribed, and tapered). There are also two smaller apertures on the distal surface, and on x-ray, all three lesions communicate. The suggested etiology explaining these changes is probable projectile injury with likely secondary osteomyelitis. Dislocation Fig. 4. Embedded obsidian projectile, L5 centrum, right side. The lesion is well-healed. Male, aged 30 – 45 years. years). In three cases, on gross examination and radiographically, bony reaction and healing are evident, but there is no such accommodation in the large femoral wound. The age range of involved individuals varies from 16 –35 years, with an average of 23.8 years for affected cases (compared to 32.2 for all males). Somewhat less direct evidence of weapon trauma comes from two additional individuals, each display- One case of a likely traumatic hip dislocation was identified in this sample. The affected individual, a male aged 25⫹, displays remodeling changes to both his left acetabulum and proximal femur. The acetabular surface is much enlarged, with a maximum diameter exceeding the contralateral by approximately 14%. The superior rim of the affected side is extended approximately 30 mm, thus most likely producing a pseudarthrosis. The left femoral head is slightly mushroomed, with moderate marginal lipping, and the greater trochanter displays hypertrophic remodeling, probably resulting from a muscle pull (m. gluteus minimus) (Fig. 5). DISCUSSION The types and patterning of traumatic injuries at SCl-038 are quite similar to observations from other prehistoric California localities. 18 R. JURMAIN TABLE 4. Prevalence of long bone fractures in various skeletal populations1 SCl-038 (U.S.) N Clavicle 159 Humerus 142 Radius 161 Ulna 144 Femur 119 Tibia 164 Fibula 129 Totals 1,018 Danish (Denmark) Ala-329 (U.S.) Libben (U.S.) St. Helens (U.K.) Raunds (U.K.) Kulubnarti (Sudan) n % N n % N n % N n % N n % N n % N 0 3 6 10 2 1 1 23 0.0 2.1 3.7 6.9 1.7 0.6 0.8 2.3 291 300 301 290 313 315 237 2,047 2 1 13 15 0 5 0 36 0.7 0.3 4.3 5.2 0.0 1.6 0.0 1.8 386 703 608 607 998 852 364 4,518 5 1 9 13 0 6 2 36 1.3 0.1 1.5 2.1 0.0 0.7 0.5 0.8 260 450 369 351 347 349 257 2,383 15 3 20 11 9 5 9 72 5.8 0.7 5.4 3.1 2.6 1.4 3.5 3.0 891 770 752 937 864 725 4,938 17 10 11 1 6 6 41 0.8 1.3 1.5 0.1 0.7 0.8 0.8 171 178 167 164 186 163 86 1,115 12 2 8 6 2 3 6 39 7.0 1.1 4.8 3.7 1.1 1.8 7.0 3.5 262 276 259 260 281 232 218 1,788 n % 1 0.4 10 3.6 16 6.2 34 13.1 3 1.1 0 0.0 3 1.4 67 3.7 Statistical comparisons, overall long bone fracture prevalence, SCl-038 compared with Ala-329 Libben Danish St. Helen-on-the-Walls Raunds Kulubnarti 1 Chi-square P df 0.90 1.53 16.85 16.21 2.89 4.63 0.34 0.22 0.0000 0.0001 0.09 0.03 1 1 1 1 1 1 n, number of fractured elements. Long bone fractures Long bone fracture prevalence is 2.3%, compared to 1.8% from another site in the same region previously analyzed by the author (Jurmain, 1991). This other site is Ala-329, on the eastern side of San Francisco Bay, approximately 14 miles north of SCl038. Employing a broader geographic basis for comparison (Table 4), long bone involvement at SCl-038 is generally intermediate in prevalence. The other North American site (Libben) from Ohio (Lovejoy and Heiple, 1981) shows a very similar overall fracture prevalence, although there is a greater involvement of the clavicle in that sample. Of the three European samples, the Danish (prehistoric/historic) group (Bennike, 1985) is significantly less involved (0.8%; P ⫽ 0.0000), as is the British urban medieval sample (Grauer and Roberts, 1996) from St. Helenon-the Walls (0.8%; P ⫽ 0.0001). However, the British medieval sample from Raunds, recently reported by Judd and Roberts (1999), shows a higher overall involvement (3.5%) than seen at SCl-038; the difference is not statistically significant. Finally, a similar overall involvement (3.7%) to that at Raunds was also found by Kilgore et al. (1997) for a Nubian medieval group; the difference in prevalence in the Nubian sample from that at SCl-038 is significant (P ⫽ 0.03). Some general conclusions can be drawn from these overall fracture data. Prehistoric North American samples tend to be moderately involved, while most European groups appear much less commonly affected. An exception to this latter pattern is apparent in the rural (Raunds) materials, and Judd and Roberts (1999) argue persuasively this elevated incidence of fractures relates to injury risks associated with farming. The Kulubnarti sample of Nubian farmers is the farthest outlier, and its high fracture prevalence has been interpreted by Kilgore et al. (1997) to reflect risks of falling on an extremely rugged terrain. Interestingly, the highest rates of fracture are found not among hunter-gatherers or urban dwellers, but among rural agriculturists. The nature of the increased risks in the groups that practiced farming, however, might vary considerably. In Britain, many of the injuries were probably caused by close work with domestic animals; however, in Nubia, large domestic animals were not a major part of the farm economy. The pattern of long bone fracture involvement also differs considerably among the comparative population samples. At Libben and Raunds, the clavicle is the most frequently fractured element (at Raunds the fibula is equally involved), but at other localities clavicular fracture is rare. At Raunds, 75% of the clavicular fractures were among males, and Judd and Roberts (1999) suggest that equestrian accidents (as well as injuries from close work with other large animals, such as oxen) could have been important risk factors. In the Libben materials, Lovejoy and Heiple (1981) diagnosed many of the lesions solely from radiographic evidence; thus, the method and ultimate quality of ascertainment may differ from that obtained in the other samples. Forearm fractures strongly dominate at SCl-038 (70% of long bone involvement) as well as at Ala-329 (78% of all long bone fractures). Likewise, in the Danish (61%), St.-Helen-on-the-Walls (51%), and Kulubnarti (75%) materials, injuries of radii and ulnae make up the majority of fractures. At Libben, such trauma constitutes exactly 50% of fracture involvement, and it is least common at Raunds (36% of fractures). Within the upper appendage, those groups with the highest relative forearm involvement (SCl-038, Ala-329, Danish, and Kulubnarti), the ulna is more commonly involved than the radius. Conversely, in the other three groups, the forearm elements are about equally involved (at St. Helen- PATTERNS OF TRAUMA IN CENTRAL CALIFORNIA on-the-Walls, Raunds), or the radius predominates (at Libben). At SCl-038, the strongest degree of asymmetric involvement is in the ulna, although, as mentioned above, this difference is not statistically significant. Of some note, however, it is the left side that predominates (80% of ulnar lesions). It would perhaps be tempting to conclude such lesions were due to parrying blows (so-called “parry fractures”) from a right-handed attacker. However, such an hypothesis is ill-founded, on several counts. Firstly, such parry fractures (clinically sometimes called “nightstick fractures”) are thought mostly to manifest near the midshaft. At SCl-038, only two ulnar fractures occur in the middle third of the diaphysis, while the remaining six (75%) are found at the distal end (where falls are a more likely etiology). Secondly, even when ulnar fractures are located midshaft, the cause is not clearly a result of aggression, a point well emphasized by others (Smith, 1996; Kilgore et al., 1997; Judd and Roberts, 1999). Isolated midshaft fractures of the ulna probably indicate impact by a sharp object to a pronated appendage. However, such injuries can occur when, for example, an individual falls against a sharp edge of a boulder or a piece of farm equipment. It is thus not justifiable to assume that all (or even most) ulnar midshaft fractures relate to interpersonal aggression, especially in the absence of other corroboratory skeletal evidence (see below). Cranial injuries The prevalence of craniofacial fractures in the SCl-038 sample (7/159 ⫽ 4.4%) is moderately high by worldwide standards, but not remarkable when compared with other groups from prehistoric California. For example, Bennike (1985), in her Danish series, found generally low rates of cranial injury, rising to a peak of 4.6% for the medieval period (an even higher prevalence was found in the Mesolithic sample, but here only 12 crania were complete enough for analysis). In a systematic overview of varied skeletal samples from Siberia, Spain, the UK, and the US, Walker (1997) found the following (combined) prevalence rates: nasal bones (7.0%), frontal (4.6%), and parietal (3.9%). Robb (1997) also reported the prevalence of cranial fractures at 8.9% from an Italian Iron Age site (N ⫽ 56). Owsley et al. (1994) examined a large sample (⬃500 skeletons) from Easter Island and found a 2.5% prevalence of craniofacial injury. From Australia, Webb (1995) reported extremely high rates of cranial injury from most regions (almost always greater than 15%, and in one area rising to 26.2% of female crania and 27.7% of male crania). In the New World, outside California, several investigators reported on cranial injury (e.g., Hooton, 1930; Snow, 1948; Stewart and Quade, 1969; Morse, 1969; Miles, 1975; Ferguson, 1980). In all these studies, prevalence rates of cranial injury varied from 2–5% of individuals. There are also several 19 contexts from various regions of North America where large-scale violence produced very high rates of perimortem cranial trauma. These cases of probable massacre and dismemberment have been described for Crow Creek, South Dakota (Willey, 1990); Norris Farms, Illinois (Milner et al., 1991); and Koger’s Island, Alabama (Bridges, 1996). In addition, widespread perimortem cranial injury associated with cannibalism has been reported in North America, especially from the southwestern US (White, 1992; Turner, 1993). Indeed, a recent report documented a similar pattern at the 100,000-yearold Moula-Guerey Neandertal site in southern France (DeFleur et al., 1999). Excepting these unusual circumstances of mass violence, in North America, generally, the highest incidence of craniofacial trauma has been reported from sites in California (including Baja). One exception is the investigation by Newman (1957), where in an all-male sample (N ⫽ 121), only 2.1% of individuals were involved. In contrast, Walker (1989) found in his highly systematic study of 744 crania from the Santa Barbara Channel region 19.4% of individuals with cranial vault injuries (for adults only, the frequency translates to 19.9%; N ⫽ 717). Compared to the 4.4% prevalence at SCl-038, this difference is highly significant (chi-square ⫽ 22.15; df ⫽ 1; P ⫽ 0.000). Similarly, in a smaller series (N ⫽ 49) from Baja California, Tyson (1977) reported a remarkable 30% of individuals with traumatic cranial involvement. Again, the difference from SCl-038 is statistically significant (chi-square ⫽ 27.20; df ⫽ 1; P ⫽ 0.000). Lastly, from the neighboring Ala-329 site, 2.7% of adult crania (N ⫽ 260) were involved (Jurmain and Bellifemine, 1997). Although showing somewhat less prevalent craniofacial injury than at SCl-038, this difference is not statistically significant (P ⫽ 0.51). All the investigators involved in the above research relating to California sites have commented on the likelihood of interpersonal aggression contributing to the relatively high levels of craniofacial injury. Tyson (1977), in addition, postulated that self-mutilation may also have played a significant role, a factor also considered by Webb (1995) as relating to his Australian series. Perforating wounds That aggression was fairly common in prehistoric California is indicated by more direct evidence. At SCl-038, four individuals have embedded fragments of obsidian projectiles, and another two individuals have partially healed lytic lesions strongly indicative of perforating wounds. There are further suggestive clues, as well, from archaeological context, including the finding in situ of a large obsidian point in the vertebral column of a 31– 40-year-old male. The artifact was found between T1 and T2, but there was no bone alteration whatsoever on either vertebra. It is thus difficult to see how including such 20 R. JURMAIN indirect evidence of skeletal trauma could be used systematically as paleoepidemiological evidence. Nevertheless, some researchers have included such data in their tabulation of projectile trauma incidence. For example, Wendorf (1968), in an analysis of the late Paleolithic Jebel Sahaba site in Nubia, found a very high incidence, leading him to conclude that “almost half the population probably died violently” (p. 993). However, of the 24 individual involved cases counted by Wendorf (1968), only four actually displayed embedded projectile fragments. The remaining 20 cases were included solely on the basis of archaeological association. From the data of Wendorf (1968), the revised frequency of relatively complete adults showing such bone involvement translates to 9.8% (4/41), rather than more than 50%. Similarly, Lambert (1994, 1997) reported in a large southern California sample a prevalence of 2.2% of individuals (38/1,744) actually showing embedded projectiles. She also argued, from suggestive nicks and archaeological context, that an additional 20 individuals were likely involved. Lambert (1997) concluded that to ignore these other indirect clues suggesting trauma would “likely result in a serious underrepresentation of the actual number of victims.” Clearly, Lambert (1997) is correct is this assertion, but in paleopathology we are inevitably underestimating actual morbid conditions. Given that some individuals incur bony injury from projectiles, several more would almost certainly have suffered strictly soft-tissue wounds. While we can confidently assess the former (bone involvement), we can never accurately estimate the latter. In order to conduct systematic paleoepidemiology, our population comparisons must be as equivalent as possible (i.e., controlling for degree of completeness, age, and sex). To introduce into our epidemiological comparisons a variety of further uncontrollable variables, tied to site disturbance and precision of archaeological recovery, makes any conclusions ambiguous at best. As pertaining to the present analysis, it is probable that the obsidian projectile found in the upper thoracic spine of the middle aged male was there at the time of death. However, it is only because the burial was not disturbed and was excavated with good archaeological controls that such an assertion can be made. It is empirically reasonable to mention this case, but not entirely sound to count it for comparative purposes. Thus, the most rigorous evidence of projectile trauma from SCl-038 is the four cases showing embedded objects. Moreover, in prehistoric North America, California groups show by far the highest prevalence of such weapon wounds. Research (Walker, 1989; Lambert and Walker, 1991; Lambert, 1994, 1997) in the Santa Barbara Channel region has added greatly to our knowledge of such injuries in California. Moreover, from a survey of numerous individuals (⬎2,000) from several sites in central California, Tenney (1986) found 18 embedded projectiles in 13 individuals. Lastly, from a site (Ala-329), also in central California, the highest prevalence has been reported for a single locality (12 embedded projectiles in 10 individuals; N ⫽ 248). Moreover, the prevalence of projectile injury at SCl-038 (4/162 ⫽ 2.5%) is quite high, even by California standards (while less common than at Ala-329, the difference is not statistically significant; P ⫽ 0.57). Fracture patterning From both the evidence of craniofacial injury and the more direct evidence of projectile wounds, it is evident that, as elsewhere in California, interpersonal aggression was common at SCl-038. Of interest, as well, the pattern of forearm (especially ulnar) fractures does not conform to the more direct evidence. As argued elsewhere (Kilgore et al., 1997; Jurmain, 1999), the interpretation of “parry fractures” is seen as highly ambiguous and is best avoided in paleoepidemiological interpretations of past populations. The soundest interpretation comes from a consideration of the more conclusive evidence. Individual fractures are less informative than the pattern of fracture involvement as distributed among different body segments. This last point has been much emphasized by Berger and Trinkaus (1995) in their intriguing interpretation of traumatic injuries among Neandertals. Through comparison with contemporary clinical data (including, especially, rodeo performers), Berger and Trinkaus (1985) concluded that Neandertals suffered common physical misfortune from “frequent close encounters with large ungulates unkindly disposed to the humans involved.” In the comparisons of fracture patterns, Berger and Trinkaus (1985) tabulated data for four body segments (head/neck; trunk/pelvis; upper appendage; and lower appendage). For purposes of comparison with California archaeological samples, these data were retabulated (Table 5) to include three body segments (head/neck; and upper and lower appendages). The trunk and pelvis were deleted, since recovery of ribs is not usually complete and, moreover, will vary between sites. Thus, in order to ensure comparability, only those better-represented body segments are included. The distribution of fractures seen in the various samples is intriguing and mostly contrasts in proportion of head/neck injuries. Rodeo performers have the highest proportion of such injuries, followed by Neandertals (and these groups are not statistically significantly different from one another). SCl-038 has the third highest proportion of head/neck injuries, and, interestingly, the distribution of fractures by body segment is not significantly different from that seen in the Neandertal sample (although it is significantly different from the distribution in rodeo performers). The patterning of fractures at the two central California sites (Ala-329 and SCL-038) is in many ways similar. However, of all the archaeological samples, the most distinctive is 21 PATTERNS OF TRAUMA IN CENTRAL CALIFORNIA 1 TABLE 5. Pattern of fracture involvement by body segment in various populations Upper appendage Head/neck SCl-038 Ala-329 Neandertals2 Rodeo performers2 Kulbubnarti (Sudan) Lower appendage N % N % N % Total fractures 7 8 7 71 1 23 18 41 55 1 19 31 7 47 61 63 71 41 36 90 4 5 3 11 6 13 11 18 9 9 30 44 17 129 68 Statistical comparisons SCl-038 vs. Neandertals SCl-038 vs. Rodeo Ala-329 vs. Neandertals SCl-038 vs. Ala-329 SCl-038 vs. Kulubnarti Ala-329 vs. Kulubnarti Neandertal vs. Rodeo 1 2 Chi-square P df 2.26 9.83 6.43 0.42 14.38 10.66 1.93 0.32 0.007 0.04 0.81 0.0008 0.005 0.38 2 2 2 2 2 2 2 %, proportion of fractures by body segment. Data from Berger and Trinkaus (1995). that from Kulubnarti, which shows an extremely low rate of cranial injury. Conversely, these medieval Sudanese Nubians show unusually high levels of upper appendage involvement. Another possible avenue of investigation concerns the distribution of fractures within the upper appendage. In the investigation of the medieval farming community from Raunds, Judd and Roberts (1999) found a proportionately high incidence of clavicular fracture, and they attributed this patterning to likely injury from close contact with large animals. From a preponderance of head and neck injuries, Berger and Trinkaus (1995) drew a similar conclusion. To date, no investigation has collated the evidence of both types of injuries to evaluate possible patterns. The patterning at SCl-038 (and at Ala-329) reveals fairly high craniofacial involvement, but a very low incidence of clavicular injury. Still, these results are not surprising; these central California groups were huntergatherers, but most likely did not regularly physically engage the largest fauna at close contact. These data on fracture patterning suggest some potentially informative trends. Firstly, the most distinctive component across population samples relates to the proportion of head/neck injuries. In those populations with the highest levels of such lesions, risk factors appear to derive from either dangerous contact with large animals or from interpersonal conflict. In groups exposed to neither type of risk, craniofacial injury can be expected to be infrequent. This informative relationship of elevated craniofacial injury to interpersonal aggression was noted previously (Smith, 1996; Kilgore et al., 1997). Finally, there is a cautionary note to be added: it is often not possible to distinguish behavioral patterns from traumatic injuries, even when rigorously evaluating patterns of involvement. Sometimes skeletons convey unambiguous evidence of the lives of past peoples. More often, however, the evidence is less clear, thus challenging skeletal biologists to make judicious interpretations. CONCLUSIONS Several interesting patterns of traumatic skeletal involvement at SCl-038 are apparent from this investigation: 1) The prevalence of long bone fracture (by individual ⫽ 10.5%; by element ⫽ 2.3%) is moderate when compared to other sites in North America and elsewhere. 2) As in most other archaeological samples, a majority of the long bone fractures occur in the forearm. 3) Cranial fractures are also fairly common, paralleling results from other California locales (but are significantly less than frequencies reported from southern California). 4) Embedded projectile points are seen quite frequently, again a similarity with other prehistoric California groups. 5) From the combined evidence of craniofacial injury and projectile wounds, it is apparent that interpersonal aggression was common. 6) Other less common traumatic involvement found at SCl-038 includes traumatic hip dislocation (one case), ununited fracture (two cases), and nasal fracture (one case). While relatively infrequent even today, these types of lesions have, to date, been reported extremely rarely in most paleopathological studies. More close evaluation of skeletal remains might well improve diagnosis of these types of traumatic manifestations (and, as relating to nasal injury, such an approach was also argued persuasively by Walker, 1997). 22 R. JURMAIN ACKNOWLEDGMENTS The analysis of the human remains from CA-SCl038 involved a team of researchers, and I am most indebted to these colleagues at San Jose State: Lorna Pierce, Anton Musladen, Leon Pappanostos, Sandra Weldon, David Calleri, Muriel Maverick, Susan Rodriquez, Susan Morley, Viviana Bellifemine, Victoria Wu, and Alan Leventhal. I am also deeply grateful to Rosemary Cambra, Chairwoman of the Muwekma Ohlone Tribe, as well as other members of the Muwekma Ohlone Tribe, for their continued support and encouragement. The research project was directed by Muwekma Ohlone descendants and the osteological analysis conducted at San Jose State was contracted through the Ohlone Families Consulting Services to the San Jose State University Foundation. Following completion of analysis, the skeletal remains were returned to the Muwekma Ohlone Tribe and were reinterred. Grateful acknowledgment is also due the San Jose State University Student Health Services for providing radiographs, and to the Instructional Resources Center Photography Department, and to Jean Shiota of the Almquist Center for Instructional Development for assistance with figures. I am also grateful to the two anonymous reviewers who provided useful suggestions to improve this article, and I am especially thankful to Emöke Szathmáry, editor, for her guidance and professionalism. This article is dedicated to the memory of my friend and colleague, Tony Musladen (1923–1998). LITERATURE CITED Anderson T. 1996. Cranial injuries from Anglo-Saxon Dover. Int J Osteoarchaeol 6:10 –14. Bass W. 1991. Human osteology. A laboratory and field manual. 4th ed. Columbia, MO: Missouri Archaeological Society, University of Missouri Special Publications. Bellifemine V. 1997. Mortuary variability in prehsitoric central California: a statistical study of the Yukisma site, CA-SCl-38. M.A. thesis, San Jose State University, San Jose, CA. Bennike P. 1985. Palaeopathology of Danish skeletons. Copenhagen: Akademisk Forlag. Berger TD, Trinkaus E. 1995. Patterns of trauma among the Neandertals. J Archaeol Sci 22:841– 852. Black TK. 1979. The biological and social analysis of a Mississippian cemetery from southeast Missouri: the Turner Site, 23BU21A. Ann Arbor: Museum of Anthropology, University of Michigan Anthropological Papers, no. 68. Bridges PS. 1996. Warfare and mortality at Koger’s Island, Alabama. Int J Osteoarchaeol 6:66 –75. Brothwell D. 1961. The paleopathology of early British man: an essay on the problem of diagnosis and analysis. J R Anthropol Inst Gr Br 91:318 –345. Defleur A, White T, Valensi P, Slimak L, Crégut-Bonnoure E. 1999. Neanderthal cannibalism at Moula-Guercy, Ardèche, France. Science 286:128 –131. Dittrick J, Suchey JM. 1986. Sex determination of prehistoric central California skeletal remains using discriminant analysis of the femur and humerus. Am J Phys Anthropol 70:3–9. Ferguson C. 1980. Analysis of skeletal remains. In: Cordell LS, editor. Tijeras Canyon: analysis of the past. Albuquerque: University of New Mexico. p 121–148. Grauer AL, Roberts CA. 1996. Paleoepidemiology, healing, and possible treatment of trauma in a medieval cemetery popula- tion of St. Helen-on-the-Walls, York, England. Am J Phys Anthropol 100:531–544. Hooton EA. 1920. Indian village site and cemetery, Maddisonville, Ohio. Pap Peabody Mus Am Archaeol Ethnol 8:83–104. Hooton EA. 1930. The Indians of Pecos Pueblo. New Haven: Yale University Press. Hutchinson DL. 1996. Brief encounters: Tatham Mound and the evidence for Spanish and Native American confrontation. Int J Osteoarchaeol 6:51– 65. Judd MA, Roberts CA. 1999. Fracture trauma in a medieval British farming village. Am J Phys Anthropol 109:229 –243. Jurmain RD. 1991. Paleoepidemiology of trauma in a central California population. In: Ortner DJ, Aufderhiede AC, editors. Human paleopathology: current synthesis and future options. Washington, DC: Smithsonian Institution Press. p 241–248. Jurmain RD. 1999. Stories from the skeleton. Behavioral reconstruction in human osteology. Amsterdam: Gordon and Breach. Jurmain RD, Bellifemine VL. 1997. Patterns of cranial trauma in a prehistoric population from central California. Int J Osteoarchaeol 7:43–50. Katz D, Suchey JM. 1986. Age estimation of the male os pubis. Am J Phys Anthropol 69:427– 435. Kilgore L, Jurmain R, Van Gerven DP. 1997. Paleoepidemiological patterns of trauma in a medieval Nubian skeletal population. Int J Osteoarchaeol 7:103–114. Lambert PM. 1994. War and peace on the western front: a study of violent conflict and its correlates in prehistoric hunter-gatherer societies of coastal southern California. Ph.D. dissertation, University of California, Santa Barbara. Lambert PM. 1995. Projectile injuries as evidence for violent conflict in prehistoric populations [abstract]. Am J Phys Anthropol [Suppl] 20:128. Lambert PM. 1997. Patterns of violence in prehistoric hunter gatherer societies of coastal southern California. In: Frayer DW, Martin DL, editors. Troubled times: violence and warfare in the past. Amsterdam: Gordon and Breach. p 77–109. Lambert PM, Walker PL. 1991. Physical anthropological evidence for the evolution of social complexity in coastal southern California. Antiquity 65:963–973. Loth SR. 1996. Mandibular ramus flexure: a new morphological indicator of sexual dimorphism in the human skeleton. Am J Phys Anthropol 99:473– 485. Lovejoy CO, Heiple KG. 1981. Analysis of fractures in skeletal populations with an example from the Libben Site, Ottawa County, Ohio. Am J Phys Anthropol 55:529 –541. Lovejoy CO, Meindl RS, Preyzbeck TR, Mensforth RP. 1985. Chronological metamorphysis of the auricular surface of the ilium: a new method for the determination of adult skeletal age at death. Am J Phys Anthropol 68:15–28. Merbs CF. 1989. Trauma. In: Iscan MY, Kennedy KAR, editors. Reconstruction of life from the skeleton. New York: Alan R. Liss. p 161–189. Miles JS. 1975. Orthopedic problems of the Wetherill Mesa population, Mesa Verde National Park, Colorado. Washington, DC: US Department of the Interior, National Park Service. Milner GR, Anderson E, Smith VG. 1991. Warfare in late prehistoric west-central Illinois. Am Antiq 56:581– 603. Morse D. 1969. Ancient diseases in the Midwest. Springfield: Illinois State Museum, Reports of Investigations, no 15. Neves WA, Barros AM, Costa MA. 1999. Incidence and distribution of postcranial fracture in the prehistoric population of San Pedro de Atacama, northern Chile. Am J Phys Anthropol 109: 253–258. Newman RW. 1957. A comparative analysis of prehistoric skeletal remains from the Lower Sacramento Valley. Berkeley: University of California Archaeological Survey Report, 39. Ortner DJ, Putschar WGJ. 1981. Identification of pathological conditions in human skeletal remains. Washington, DC: Smithsonian Institution Press. Owsley DW, Gill GW, Owsley SD. 1994. Biological effects of European contact on Easter Island. In: Larsen CS, Milner GR, editors. In the wake of contact. Biological responses to conquest. New York: Wiley-Liss. p 161–177. PATTERNS OF TRAUMA IN CENTRAL CALIFORNIA Pfeiffer S. 1985. Paleopathology of Archaic peoples of the Great Lakes. Can Rev Phys Anthropol 4:1–7. Robb J. 1997. Violence and gender in early Italy. In: Frayer DW, Martin DL, editors. Troubled times: violence and warfare in the past. Amsterdam: Gordon and Breach. p 111–144. Smith MO. 1996. “Parry” fractures and female-directed interpersonal violence: implications from the Late Archaic Period of west Tennessee. Int J Osteoarchaeol 6:84 –91. Smith MO. 1997. Osteological indicators of warfare in the Archaic period of the Western Tennessee Valley. In: Frayer DW, Martin DL, editors. Troubled times: violence and warfare in the past. Amsterdam: Gordon and Breach. p 241–265. Snow CE. 1948. Indian Knoll skeletons of site Oh 2 Ohio County, Kentucky. Knoxville: University of Kentucky Reports in Anthropology no. 4. p 321–554. Stewart TD. 1974. Nonunion of fractures in antiquity, with description of five cases from the New World involving the forearm. Bull NY Acad Med 50:875-– 891. Stewart TD, Quade LG. 1969. Lesions of the frontal bone in American Indians. Am J Phys Anthropol 30:89 –110. Tenney J. 1986. Trauma among early California populations [abstract]. Am J Phys Anthropol 69:271. Turner C. 1993. Cannibalism in Chaco Canyon: the charnel pit excavated in 1926 at Small House by Frank H.H. Roberts Jr. Am J Phys Anthropol 91:421– 439. Tyson RA. 1977. Historical accounts as aids to physical anthropology. Examples of head injury in Baja California. Pacif Coast Archaeol Soc Q 13:52–58. Ubelaker DH. 1989 Human skeletal remains. Excavation, analysis, intepretation. 2nd ed. Washington, DC: Teraxacum Press. 23 Waldron T. 1994. The human remains. In: Evison V, editor. An Anglo-Saxon cemetery from Great Chesterford, Essex. York: Council for British Archaeology, reserach report 91. p 52– 66. Walker PL. 1989. Cranial injuries as evidence of violence in prehistoric southern California. Am J Phys Anthropol 80:313– 323. Walker PL. 1997. Wife beating, boxing, and broken noses: skeletal evidence for the cultural patterning of violence. In: Frayer DW, Martin DL, editors. Troubled times: violence and warfare in the past. Amsterdam: Gordon and Breach. p 145– 179. Webb S. 1995. Palaeopathology of aboriginal Australians. Cambridge: Cambridge University Press. Wells C. 1964. Bones, bodies, and disease. Evidence of disease and abnormality in early man. New York: Praeger. Wendorf F. 1968. Site 117: a Nubian final Paleolithic graveyard near Jebel Shaba, Sudan. In: Wendorf F, editor. The prehistory of Nubia. Volume II. Dallas: Southern Methodist University Contribution to Anthropology, no. 2. p 954 –985. White TD. 1992. Prehistoric cannibalism at Mancos 5MTUMR2346. Princeton: Princeton University Press. Willey P. 1990. Prehistoric warfare on the great plains. Skeletal analysis of the Crow Creek massacre victims. New York: Garland Publishing Co. Wood Jones F. 1910. Fractured bones and dislocations. In: ElliotSmith G, Wood Jones F, editors. The archaeological survey of Nubia, volume II, report on the human remains. Cairo: National Printing Department.