Bone fractures as indicators of intentional ... - Wiley Online Library

AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 149:26–38 (2012)

Bone Fractures as Indicators of Intentional Violence in the Eastern Adriatic From the Antique to the Late Medieval Period (2nd–16th Century AD) Mario Sˇlaus,1,2 Mario Novak,1* Zˇeljka Bedic´,1 and Davor Strinovic´2 1 2

Department of Archaeology, Croatian Academy of Sciences and Arts, 10 000 Zagreb, Croatia Department of Forensic Medicine and Criminology, Medical School, University of Zagreb, 10000 Zagreb, Croatia KEY WORDS

bioarchaeology; trauma; perimortem injury; Croatia

ABSTRACT To test the historically documented hypothesis of a general increase in deliberate violence in the eastern Adriatic from the antique (AN; 2nd–6th c.) through the early medieval (EM; 7th–11th c.) to the latemedieval period (LM; 12th–16th c.), an analysis of the frequency and patterning of bone trauma was conducted in three skeletal series from these time periods. A total of 1,125 adult skeletons—346 from the AN, 313 from the EM, and 466 from the LM series—were analyzed. To differentiate between intentional violence and accidental injuries, data for trauma frequencies were collected for the complete skeleton, individual long bones, and the craniofacial region as well as by type of injury (perimortem vs. antemortem). The results of our analyses show a significant temporal increase in total fracture frequencies when calcu-

lated by skeleton as well as of individuals exhibiting one skeletal indicator of deliberate violence (sharp force lesions, craniofacial injuries, ‘‘parry’’ fractures, or perimortem trauma). No significant temporal increases were, however, noted in the frequencies of craniofacial trauma, ‘‘parry’’ fractures, perimortem injuries, or of individuals exhibiting multiple skeletal indicators of intentional violence. Cumulatively, these data suggest that the temporal increase in total fracture frequencies recorded in the eastern Adriatic was caused by a combination of factors that included not only an increase of intentional violence but also a significant change in lifestyle that accompanied the transition from a relatively affluent AN urban lifestyle to a more primitive rural medieval way of life. Am J Phys Anthropol 149:26–38, 2012. V 2012 Wiley Periodicals, Inc.

It has now been over a century since Smith and Jones (1910) carried out the first systematic analysis of bone traumas in archaeological populations. Initially, their work failed to provoke significant scientific interest, and bone trauma observed in osteological material from archaeological sites continued to be presented at the level of descriptive portrayals of individual skeletons that accompanied various, primarily archaeological publications. This state of affairs continued up to the publication of Lovejoy and Heiple’s (1981) comprehensive analysis of bone injuries from the North American Libben site that used a methodology that was later adopted by numerous authors (e.g., Walker, 1989; Bridges, 1996; Grauer and Roberts, 1996). Since then, bone fractures have become, together with dental diseases, vertebral pathologies, and osteological and dental indicators of subadult stress, one of the most studied pathological conditions in archaeological samples (e.g., Domett and Tayles, 2006; Djuric´ et al., 2006; Jurmain et al., 2009; Scott and Buckley, 2010). In this work, we analyze three large archaeologically derived skeletal series from the antique (AN), early medieval (EM), and late medieval (LM) periods in Croatia to test whether they conform to available historical data that suggest a temporal increase in deliberate violence from the AN to the LM period in the eastern Adriatic. To accomplish this, we analyze and compare fracture frequencies and patterning in the following three temporally distinct archaeological series—the AN sample (2nd–6th c.) represented by urban communities settled on the Adriatic coast, the EM sample (7th–11th c.) represented by small rural communities settled in the mountainous hinterland of Dalmatia, and the LM sample

(12th–16th c.) represented by mixed urban and rural communities occupying both the coast and the hinterland of Dalmatia. Available historic documents describe the AN period as peaceful, with no significant episodes of intentional violence except for the brief Visigoth invasion in 378. The EM period is characterized by the struggle of Croatian rulers against Venice and the Byzantine Empire for control of the eastern Adriatic coast, particularly the wealthy Dalmatian cities. During the LM period, numerous large-scale outbreaks of violence occurred as a result of a gradual weakening of centralized royal authority, and the resulting feudal anarchy continued fighting against Venice for control of the Dalmatian cities and intrusion by the Mongols in the 13th and Ottoman Turks in the 15th and 16th centuries [for more details, see Novak (2001)]. An alternative cause for differences in bone fracture frequencies in the analyzed series may be the significant change in lifestyles that accompanied the transition

C 2012 V

WILEY PERIODICALS, INC.

C

Grant sponsor: Ministry of Science, Education and Sports of the Republic of Croatia; Grant number: 101-197-0677-0670. *Correspondence to: Mario Novak, Department of Archaeology, Croatian Academy of Sciences and Arts, Ante Kovacˇicá 5, 10 000 Zagreb, Croatia. E-mail: [email protected] Received 6 February 2012; accepted 30 March 2012 DOI 10.1002/ajpa.22083 Published online 3 May 2012 in Wiley Online Library (wileyonlinelibrary.com).

INTENTIONAL VIOLENCE IN THE EASTERN ADRIATIC

27

TABLE 1. Archaeological sites analyzed in this study Site Antique series Zadar-Relja, urban Vid, urban Total Early medieval series Dubravice, rural Sˇibenik, rural Glavice, rural Velim, rural Total Late medieval series Zadar-Jazine, urban Nin, urban Lisˇtani, rural Dugopolje, rural Total Grand total

Dating

Number of skeletons

AN (2nd–6th c.) AN (5th–6th c.)

274 72 346

EM EM EM EM

(7th–9th c.) (9th–11th c.) (8th–9th c.) (8th–9th c.)

69 65 78 101 313

LM LM LM LM

(13th c.) (12th–15th c.) (13th–15th c.) (14th–16th c.)

50 110 130 176 466 1125

from the AN to the EM and LM periods. The analyzed samples originate from very different settings—urban for the AN and (predominantly) rural for the medieval series. In relation to this, Judd and Roberts (1999) noted that two medieval British rural communities (Raunds and Jarrow Abbey) exhibited significantly higher fracture frequencies (or values bordering on significance) than contemporaneous urban communities (St. Helen-on-theWalls, St. Nicholas Shambles, and Blackfriars). They interpreted these differences as the result of a greater risk of injury associated with farming. They concluded that the locations and types of fractures recorded in their rural samples reflected segregation in activity, probably associated with labor and that the recorded traumas reflected the high risk of injury caused by everyday activities in the household, fields, and while working with large draft animals (Judd and Roberts, 1999). Typical injuries in these samples included distal oblique fractures to the forearm, which are associated with indirect forces due to tripping or short falls, and clavicular and midshaft transverse breaks of the lower limbs caused by working with large draft animals—for instance, from falls from wagons or horses or being caught under overturned vehicles (Judd and Roberts, 1999). Analyses of trauma frequencies in the Croatian series were also carried out by sex in order to test whether males and females exhibited different fracture frequencies. Higher trauma frequencies in males have been recorded in numerous archaeological populations of various chronological and geographical determinations (e.g., Robb, 1997; Judd, 2004; Djuric´ et al., 2006). According to Standen and Arriaza (2000), males are expected to exhibit higher overall frequencies of trauma, because they tend to get involved in more physically demanding and risky jobs than females. Similarly, Djuric´ et al. (2006) suggested that males may have been more exposed to injuries from various aspects of their life, such as interpersonal conflict, horse riding, hunting, and agricultural work. Robb (1997) suggests that the higher trauma frequencies recorded in males living in prehistoric Italy was not a direct result of violence, but was related to the development of gender roles that prescribed violent behavior for males and reinforced a sexual division of labor in which women were not expected to perform activities considered heavy or dangerous, including war-

Fig. 1. Map of Croatia and Bosnia and Herzegovina with the geographical locations of the analyzed sites.

fare. Similarly, modern clinical evidence suggests that most fractures result, not from unusual events or interpersonal violence, but from daily activities. Consequently, in contemporary settings, most fractures in females occur in the home, while most fractures in males occur at work or during sports (Lovell, 1997). High-fracture risks exist in occupations that have been generally restricted to men, such as agriculture, mining, forestry, and construction (Lovell, 1997). To summarize, the purpose of this study is to analyze trauma frequencies and distributions in three archaeological series from the eastern Adriatic and test whether they conform to available historical data that suggest a temporal increase in deliberate violence caused by social deterioration, civil wars, and military intrusions by the Mongols and Ottoman Turks. If these data are correct, we expect a clear temporal increase in the frequencies of bone trauma (by skeleton, long bone, and cranium) and, more importantly, a temporal increase in the frequencies of skeletal indicators of deliberate violence: perimortem trauma, sharp force injuries, ‘‘parry’’ fractures, and craniofacial fractures. If, on the other hand, historical sources exaggerated the risk of incurring injuries caused by deliberate violence in the early, and particularly LM period, trauma frequencies in the analyzed series may still be affected by the change from an urban way of life in the AN period to the rural type of existence that characterized the two medieval periods. In this scenario, we expect changes similar to those noted by Judd and Roberts (1999), with a possibly less well-defined temporal component and, most importantly, no significant temporal increases in the frequencies of skeletal indicators of deliberate violence.

MATERIALS AND METHODS Geographic and archaeological context The skeletal material analyzed in this study consists of three composite series: the AN (2nd–6th c.), EM (7th– 11th c.), and LM (12th–16th c.) samples (Table 1). All of the analyzed sites are located on the eastern Adriatic coast or its hinterland (contemporary Croatia and Bosnia and Herzegovina) (Fig. 1). American Journal of Physical Anthropology

28

M. SˇLAUS ET AL.

Although the sites analyzed in this study come from a relatively small area, important climatic differences resulting in different subsistence strategies are present between the coastal area (Adriatic Croatia), and the mountainous hinterland that represents the western outskirts of the Dinarid mountain range. Adriatic Croatia is characterized by a Mediterranean climate, with short, wet winters during which the temperature rarely falls below 58C. The region has few arable fields that lie between the Adriatic Sea and the large Dinarid Mountain range. Olives and grapes are the main agricultural products, while additional sources of nutrition derive from fishing and sheep and goat pastoralism (Goldstein, 1995). The mountainous Adriatic hinterland has a continental climate with harsh, very cold winters, and hot summers. The region is characterized by almost impassable macchia shrubbery and thickets composed of mostly thorny species (Rogosˇic´, 2001). During the medieval and the modern age periods, the inhabitants of this region were engaged in transhumant pastoralism and various forms of agriculture, with a diet based on meat and animal products and cereals such as barley, rye, and oats (Jurin-Starcˇevic´, 2008; Sˇaric´, 2008). The AN sample consists of 346 adult skeletons from two archaeological sites: Zadar-Relja (Brusic´ and Glusˇcˇevic´, 1990; Fadic´, 2007) and Vid (Marin, 2002; Mardesˇic´, 2004). Both sites were large, well-developed, urban centers with the status of Roman colonies. All skeletons belonging to this sample are dated between the 2nd and 6th centuries based on burial rites, grave goods, and horizontal stratigraphy. The osteological material was recovered from simple inhumations in plain ground, stone tombs, and graves covered with tegulae (roof tiles) or fragments of amphorae. Grave goods mostly composed of glass bottles, oil lamps (lucernae), coins, fibulae, and some silver and gold jewelry. The vast majority of burials in this sample are represented by ‘‘poor’’ graves. Of potential interest is the fact that comparison between these individuals and individuals recovered from ‘‘wealthy’’ graves (burials containing three or more artifacts or artifacts made from precious metals) revealed no differences in trauma frequencies. The EM sample consists of 313 adult skeletons from four sites: Dubravice (Krncˇevic´, 1998), Sˇibenik (Krncˇevic´, 1998), Glavice (Petrinec, 2002), and Velim (Juric´, 2006). All skeletons from this series were recovered from cemeteries associated with small rural settlements. Grave goods, burial rites, and radiocarbon dating show that these cemeteries were in use between the 7th and the 11th centuries. Most of the burials were simple inhumations aligned in rows, dug deep into plain soil or bedrock, and covered and encased with stone slabs. Grave finds were scarce and uniform: pottery, bronze and silver jewelry, knives, and coins. The LM sample consists of 466 skeletons recovered from four archaeological sites: Zadar-Jazine (Vujevic´, personal communication), Nin (Kolega, 2001, 2002), Lisˇtani (Maric´, 2005), and Dugopolje (Gjurasˇin, 2007). Two of the sites, Zadar-Jazine and Nin, represent urban centers located on the eastern Adriatic coast, while the remaining two, Dugopolje and Lisˇtani, represent small rural communities situated in the mountainous hinterland. Based on recovered archaeological finds, horizontal stratigraphy, and historical sources, these cemeteries were in use between the 12th and 16th centuries. The recovered grave finds were modest and mostly consisted of coins, rings, earrings, belt buckles, buttons, and iron American Journal of Physical Anthropology

pins. Most of the graves were simple inhumations dug into plain soil or bedrock and covered and encased with stone slabs, although some graves were covered with stecći: monumental tombstones depicting various decorative motifs. In contrast to the other two samples, the LM series consists of both a rural component—representing the majority (n 5 306) and an urban component representing a minority (n 5 160) of the recovered individuals. Several previously conducted analyses dealing with alveo–dental pathologies, skeletal and dental indicators of subadult stress, and infectious diseases (Novak and Sˇlaus, 2007; Martincˇic´, 2009) have, however, demonstrated that there are no significant differences in the frequencies of these pathologies between these two components suggesting similar life conditions in both subsamples. Consequently, both components of this sample are treated as a single entity that was, in terms of living conditions, more similar to the rural communities of the EM sample than to the developed urban communities of the AN sample.

Description of the methods used in the analysis Only adult skeletons (individuals over 15 years of age) were analyzed. Subadults were excluded from the study, because modern clinical studies have shown that subadults sustain only a minor percent of trauma in developing countries (Judd, 2004), and the same trend has been noted in many bioarchaeological studies (e.g., Sˇlaus, 2008; Scott and Buckley, 2010). Sex and the age-at-death of the recovered individuals were determined using methods described in Buikstra and Ubelaker (1994). Only individuals with unambiguous determinations of sex were included in the analysis. Adults were grouped into one of three composite age categories: young (between 15 and 30 years), middle-aged (between 31 and 45 years), and old (451 years). All skeletons were examined macroscopically for the possible presence of trauma using methods proposed by Maples (1986) and Lovell (1997). The location of the injury on the skeletal element was recorded, as well as the shape, dimensions, and any possible complications. For more complicated or ambiguous fractures, additional analyses, such as radiographic imaging and CT scans, were used. A distinction was made between ante- and perimortem skeletal injuries. Antemortem injuries were identified by evidence of healing and bone remodeling (Sauer, 1998), while perimortem injuries were identified with the following criteria: absence of healing and formation of new bone (Sauer, 1998), fragments remaining attached to one another (Sauer, 1998), internal beveling (Fachini et al., 2007), defined or sharp edges (Wheatley, 2008), and flat or polished surfaces with macroscopically visible striations (Wakely, 1997). Considering that ground pressure and postmortem damage can mimic perimortem injuries (Murphy et al., 2010), the differential texture and color of the lesions were used to differentiate between perimortem and postmortem trauma (Sauer, 1998; Facchini et al., 2007). Because of the varying degrees of preservation of the skeletons, fracture frequencies were calculated by skeleton, long bone, and cranium. Trauma frequencies by skeleton were calculated from the number of skeletons exhibiting fractures (including small bone fractures such as fractures of the bones of hand and feet, and the ribs) divided by the total number of skeletons, regardless of the degree of preservation.

29

INTENTIONAL VIOLENCE IN THE EASTERN ADRIATIC TABLE 2. The sex and age distributions in the analyzed series AN 15-30 31-44 451 Total

EM

LM

M

F

M

F

M

F

29 112 54 195

37 81 33 151

22 92 50 164

32 66 51 149

47 160 76 283

41 90 52 183

M, males; F, females.

The presence of fractures in long bones was assessed on the clavicles, humerii, radii, ulnae, femora, tibiae, and fibulae. Only long bones preserved to an extent of at least two-thirds of their surface, and with all major articular surfaces preserved, were analyzed. The presence of trauma was determined by checking for bilateral bone asymmetry, angular deformities, and the presence of bone calluses. Analysis of cranial injuries was carried out on crania with all major bones (frontal, both parietal, both temporal, occipital, facial bones, and mandible) preserved to at least two-thirds of their extent. Cranial bones were analyzed for the presence of three types of trauma: penetrating injuries, blunt force trauma, and sharp force trauma. Penetrating wounds exhibit internal and external beveling (Berryman and Symes, 1998), while partially penetrating wounds have depressed, comminuted fractures caused by the crushing of bone at the impact site (Milner et al., 1991). Blunt force trauma is usually caused by a relatively low-velocity impact over a relatively large surface and is typically produced through contact with a blunt instrument or during falls (Galloway et al., 1999). It is identified by the presence of plastic deformation, fractures radiating from an impact site, or concentric fractures exhibiting an internal bevel (Berryman and Symes, 1998). Sharp force trauma is identified by the presence of linear lesions with well-defined sharp edges that have flat, smooth, and polished cut surfaces. These lesions exhibit a V-shaped cross-section and macroscopically or microscopically visible parallel striations perpendicular to the kerf floor (Houck, 1998; Reichs, 1998; Symes et al., 1998; Kjellstro¨m, 2005). The term ‘‘accidental injuries’’ indicates injuries caused by unplanned events that happen unexpectedly, while the term ‘‘violence’’ is used to imply a harmful interaction between people (i.e., interpersonal violence) (Walker, 2001). According to Resolution 48/104 of the United Nations (1993), violent behavior is defined as ways of behavior between people that are likely to cause personal harm or injury. The Centers for Disease Control and Prevention of the United States Department of Health and Human Services (2011) define physical violence as ‘‘the intentional use of physical force with the potential for causing death, disability, injury, or harm.’’ Jurmain et al. (2009) argue that patterns of interpersonal aggression in archaeological series can best be understood through the analysis of correlations between different types of injuries in different osseous elements. In this analysis, we therefore use similar criteria to those suggested by Jurmain et al. (2009) and record the presence and co-occurrence of four skeletal indicators of deliberate violence: craniofacial injuries (facial and frontal regions combined), sharp force lesions, ‘‘parry’’ fractures, and perimortem trauma. In cases where a skeleton exhibited a craniofacial fracture that was the result of a sharp force lesion, or a perimortem trauma, only

TABLE 3. The sex distributions of fracture frequencies calculated by skeleton

Male Female Total

AN n/N (%)

EM n/N (%)

LM n/N (%)

32/195 (16.4) 25/151 (16.6) 57/346 (16.5)

53/164 (32.3) 20/149 (13.4) 73/313 (23.3)

85/283 (30.0) 31/183 (16.9) 116/466 (24.9)

n 5 number of skeletons exhibiting a fracture; N 5 number of examined skeletons; and % 5 percentage of skeletons exhibiting a fracture.

one aspect of the injury was recorded. In cases where a skeleton exhibited a perimortem cranial injury and an antemortem cranial fracture, both traumas were recorded. All other recorded injuries were deemed to have been caused by accidents. The analyzed skeletal samples were compared by chronological period, age, and sex. Data gathered in this study were analyzed using SPSS 14.0 for Windows. The observed differences were evaluated with the v2 test using Yates correction when appropriate, and statistical significance was defined by probability levels of P 0.05. At this point, it is important to note that inferential statistical tests are only valid when considering frequencies with respect to individuals, rather than with respect to number of bones. This is because bones in an individual skeleton are not independent observations as required by standard statistical test, and, therefore, inferential tests carried out on frequencies expressed by total number of bones carry a high risk of type 1 error. With this in mind, we will not test potentially significant differences in long bone fracture frequencies, but will simply present the collected data to help illustrate possible trends and patterns.

RESULTS Sex and age distribution The sex and age distributions of the analyzed skeletons are shown in Table 2. The AN sample consists of 346 skeletons (151 females and 195 males), the EM of 313 skeletons (149 females and 164 males), and the LM sample of 466 skeletons (183 females and 283 males). The sex ratios (females vs. males) within the samples are as follows: for the AN sample 1:1.29, for the EM sample 1:1.1, and for the LM sample 1:1.55. The sex ratios between the samples are generally similar although a significant difference is noted between the male/female ratios in the EM and LM series (v2 5 4.982, df 5 1, P 5 0.026). Total age distributions in all three samples are, again, similar. No significant differences were noted in the male/female age distributions in the analyzed samples.

Fracture frequencies analyzed by skeleton Fracture frequencies per skeleton are shown in Table 3. In the AN sample, 16.5% of skeletons exhibit one or more fractures; in the EM sample, these values increase to 23.3%; while in the LM series, 24.9% of all skeletons display skeletal injury. In the LM series, the rural subsample exhibits somewhat higher fracture frequencies compared to the urban (26.8% vs. 21.2%) without, however, achieving statistical significance. Both medieval samples exhibit significantly higher total frequencies American Journal of Physical Anthropology

M. SˇLAUS ET AL.

30

TABLE 4. Long bone fracture frequencies by bone element Clavicle n/N (%)

Humerus n/N (%)

Radius n/N (%)

Males Females Total

4/199 (2.0) 2/162 (1.2) 6/361 (1.7)

2/201 (1.0) 2/155 (1.3) 4/356 (1.1)

2/160 (1.2) 6/127 (4.7) 8/287 (2.8)

Males Females Total

3/261 (1.1) 0/240 (0.0) 3/501 (0.6)

3/364 (0.8) 1/137 (0.7) 4/501 (0.8)

11/238 (4.6) 3/197 (1.5) 14/435 (3.2)

Males Females Total

7/443 (1.6) 2/134 (1.5) 9/577 (1.6)

3/408 (0.7) 0/277 (0.0) 3/685 (0.4)

11/393 (2.8) 5/237 (2.1) 16/630 (2.5)

Ulna n/N (%) AN 2/181 (1.1) 2/131 (1.5) 4/312 (1.3) EM 11/244 (4.5) 4/204 (2.0) 15/448 (3.3) LM 15/394 (3.8) 2/233 (0.9) 17/627 (2.7)

Femur n/N (%)

Tibia n/N (%)

Fibula n/N (%)

Total n/N (%)

0/258 (0.0) 1/190 (0.5) 1/448 (0.2)

5/196 (2.5) 0/138 (0.0) 5/334 (1.5)

0/172 (0.0) 0/121 (0.0) 0/293 (0.0)

15/1367 (1.1) 13/1024 (1.3) 28/2391 (1.2)

1/283 (0.3) 0/257 (0.0) 1/540 (0.2)

7/271 (2.6) 0/229 (0.0) 7/500 (1.4)

3/228 (1.3) 0/185 (0.0) 3/413 (0.7)

39/1789 (2.2) 8/1549 (0.5) 47/3338 (1.4)

0/432 (0.0) 0/293 (0.0) 0/725 (0.0)

14/403 (3.5) 3/274 (1.1) 17/677 (2.5)

7/399 (1.7) 1/223 (0.4) 8/622 (1.3)

57/2772 (2.1) 13/1771 (0.7) 70/4543 (1.5)

n 5 number of long bones exhibiting a fracture; N 5 number of examined long bones; and % 5 percentage of long bones exhibiting a fracture.

than the AN sample (EM vs. AN v2 5 4.445, df 5 1, P 5 0.035; LM vs. AN v2 5 7.899, df 5 1, P 5 0.005). The difference between the EM and LM series is not significant. When sexes are compared within the samples, males from both medieval series exhibit significantly higher fracture frequencies than females (for the EM v2 5 14.546, df 5 1, P \ 0.001; for the LM v2 5 9.505, df 5 1, P 5 0.002). No significant differences were noted between males and females in the AN series. Sex comparisons between the samples reveal significantly higher fracture frequencies in males from both medieval series when compared with AN males (EM vs. AN v2 5 11.608, df 5 1, P \ 0.001; LM vs. AN v2 5 10.869, df 5 1, P \ 0.001). No significant differences were noted between the female subsamples.

Long bone fractures Long bone fracture frequencies are shown in Table 4. The AN sample exhibits the lowest values (1.2%), intermediate values are recorded in the EM sample (1.4%), while the LM series exhibits the highest frequencies (1.5%). In the LM series, the rural subsample again exhibits a slightly higher fracture frequency when compared with the urban (1.6% vs. 1.2%). Females from the AN sample exhibit slightly higher fracture prevalence than males (1.3% vs. 1.1%), while in both medieval series, males exhibit much higher frequencies than females (in the EM series, 2.2% males vs. 0.5% females; in the LM, 2.1% vs. 0.7%). Sex comparisons between the samples show that males from both medieval series exhibit much higher fracture frequencies when compared with AN males (EM 2.2% vs. AN 1.1%; LM 2.1% vs. AN 1.1%). Concerning side distribution of the fractures, the AN sample exhibits identical total long bone fracture frequencies (upper and lower limbs combined) on the left and right sides (1.2% or 14/1192 on the left side vs. 1.2% or 14/1199 on the right), while both medieval series exhibit higher total long bone fracture frequencies on the left side (EM 1.8% or 30/1667 on the left side vs. 1.0% or 17/1661on the right; LM 1.6% or 37/2268 vs. 1.4% or 33/2275). In the AN sample, long bone fractures are most frequent on the radius (2.8%), followed by the clavicle (1.7%) and tibia (1.5%). In the EM sample, the highest trauma prevalence is recorded in the ulna (3.3%), followed by the radius (3.2%) and tibia (1.4%). The same basic pattern is recorded in the LM sample with the highest frequency observed in the ulna (2.7%), while the American Journal of Physical Anthropology

radius and tibia exhibit identical values (2.5% each). As is evident from Table 5, all samples share a common characteristic: an increase in long bone trauma frequencies with advanced age. Because ulnar fractures are often used as indicators of intentional violence, this type of skeletal injury is of special interest in trauma analyses. However, it is important to note that not all ulnar fractures result from deliberate violence. When all ulnar fractures are recorded, the lowest frequency is noted in the AN sample (1.3%); in the EM sample, these values increase to 3.3%; while in the LM series, they decrease to 2.7%. When, however, only ‘‘parry’’ fractures as defined by Judd (2008) are analyzed, a constant increase of fracture frequencies between the periods is evident. Two ‘‘parry’’ fractures are recorded in the AN sample (2/312 or 0.6%, both in males), five in the EM sample (5/448 or 1.1%, four in males and one in a female), and eight in the LM series (8/627 or 1.3%, all in males).

Craniofacial injuries The types of fractures recorded in all three series include small circular and/or larger irregular depressions, sharp force lesions, penetrating injuries, nasal fractures, and mandibular fractures (Table 6). The AN sample is represented by 173 complete crania of which 29 (or 16.8%) exhibit trauma (Table 7). In the EM sample, 39 of 245 or 15.9% of crania exhibit evidence of fractures, while in the LM sample, 52/258 or 20.1% of crania show signs of injury. In the LM series, the rural subsample exhibits somewhat higher fracture frequencies when compared with the urban (21.6% vs. 16.4%) again, without achieving statistical significance. When comparisons are made between the sexes, males exhibit higher frequencies of craniofacial injuries than females in all three series, but the difference achieves significance only in the LM sample (v2 5 13.413, df 5 1, P \ 0.001). Comparisons between the series at the level of total samples, as well as by sex, show no significant differences. All samples share a common characteristic: a positive correlation between craniofacial trauma frequencies and advanced age that achieves significance in the AN and LM samples (AN v2 5 6.418, df 5 2, P 5 0.04; LM v2 5 8.382, df 5 2, P 5 0.015). A total of 38 cranial injuries are recorded in the AN sample: 23 skulls exhibit one trauma, 4 skulls display

31

(0.2) (1.3) (3.1) (1.5) 2/926 30/2374 38/1,243 70/4,543

TABLE 6. Frequencies of craniofacial injuries by type Depression n (%)

(0.0) (0.8) (1.2) (0.7)

AN EM LM

32 (84.2) 42 (80.8) 56 (82.3)

Sharp force Piercing Nasal Mandible n (%) n (%) n (%) n (%) 4 (10.6) 2 (3.8) 8 (11.8)

1 (2.6) 0 (0.0) 1 (1.5)

1 (2.6) 8 (15.4) 2 (2.9)

0 (0.0) 0 (0.0) 1 (1.5)

Total N (%) 38 (100.0) 52 (100.0) 68 (100.0)

n 5 number of recorded craniofacial injuries by type; % 5 percentage of total number of recorded craniofacial injuries; and N 5 total number of recorded craniofacial injuries.

n 5 number of long bones exhibiting a fracture; N 5 number of examined long bones; and % 5 percentage of long bones exhibiting a fracture.

0/440 7/851 6/480 13/1,771 (0.4) (1.5) (4.2) (2.1) 2/486 23/1523 32/763 57/2,772 (0.4) (1.1) (2.5) (1.4) 2/561 18/1702 27/1,075 47/3,338 (0.0) (0.0) (1.5) (0.5) 0/327 0/689 8/533 8/1,549 (0.8) (1.8) (3.5) (2.2) 2/234 18/1013 19/542 39/1,789 (0.4) (1.2) (1.8) (1.2) 2/514 15/1264 11/613 28/2,391 (0.3) (1.2) (2.7) (1.3) 1/305 6/494 6/225 13/1,024 (0.5) (1.2) (1.3) (1.1) 1/209 9/770 5/388 15/1,367 16–30 31–45 451 Total

Total n/N (%) EM Females n/N (%) Males n/N (%) Total n/N (%) Males n/N (%)

AN Females n/N (%)

TABLE 5. Long bone fracture frequencies in the three series by sex and age

Males n/N (%)

LM Females n/N (%)

Total n/N (%)


two, 1 skull exhibits three, and 1 skull displays four traumas. In the EM sample, a total of 52 skull injuries are noted: 32 crania exhibit one trauma, 3 crania have two, and 3 crania also display three traumas, while 1 cranium displays five injuries. The LM series exhibit a total of 68 cranial fractures: 40 crania exhibit one trauma, 9 crania exhibit two, and 1 crania exhibit three, while 1 exhibits four traumas. The distribution of craniofacial injuries by cranial element is shown in Table 8. The frontal bone is the most affected cranial bone in all three series (in the AN 5 47.4%, EM 5 32.7%, and LM 5 38.3%; Figs. 2–4) with both parietal bones being the second most frequently affected bones (values for the left parietal range from 20.6 to 34.2%, for the right from 13.2 to 29.4%). Analysis by sex shows that the largest difference in the AN sample is noted for the right parietal bone, in the EM sample in the facial region, and in the LM sample for the left parietal bone. None of these differences are, however, significant. A gradual temporal shift of cranial trauma from the left to the right side of the skull is also noted. In the AN and EM series, most cranial injuries are located on the left side (AN 63.2% on the left side vs. 36.8% on the right, v2 5 4.263, df 5 1, P 5 0.04; in the EM 63.5% vs. 36.5%, v2 5 6.5, df 5 1, P 5 0.01), while in the LM sample, cranial trauma are predominantly located on the right side (66.2% on the right side vs. 33.8% on the left, v2 5 12.971, df 5 1, P \ 0.001).

Perimortem trauma Perimortem traumas are recorded in seven individuals, two (both males) from the AN sample, two (again males) from the EM sample, and three (two males and a female) from the LM sample. The total frequencies of perimortem trauma calculated by the skeleton are identical and very low (0.6%) in all three series (Table 9). Four of the seven skeletons that exhibit perimortem trauma have one lesion, two exhibit three perimortem injuries, and one exhibits four (Table 10). All the injuries are located in the upper third of the skeleton—primarily in the cranium, while additional elements include the clavicle, second thoracic vertebra, first rib, and left scapula (twice). The perimortem injuries can be distinguished by the weapons that were used to inflict them: five skeletons exhibit injuries caused by sharp bladed weapons (most likely swords or battle knives; Fig. 5), while two skeletons exhibit penetrating fractures probably caused by arrows (Fig. 6).

The frequency and co-occurrence of skeletal indicators of intentional violence The frequency of skeletal indicators of deliberate violence is lowest in the AN sample where 30 individuals American Journal of Physical Anthropology

M. SˇLAUS ET AL.

32

TABLE 7. Craniofacial fracture frequencies in the three series by sex and age AN Males n/N (%)

Females n/N (%)

16–30 1/12 (8.3) 1/24 31–45 10/46 (21.7) 5/46 451 7/29 (24.1) 5/16 Total 18/87 (20.7) 11/86

EM Total n/N (%)

(4.2) 2/36 (10.9) 15/92 (31.2) 12/45 (12.8) 29/173

Males n/N (%)

(5.6) 3/19 (16.3) 11/72 (26.7) 11/41 (16.8) 25/132

Females n/N (%)

(15.8) 1/20 (15.3) 6/49 (26.8) 7/44 (18.9) 14/113

LM Total n/N (%)

Males n/N (%)

(5.0) 4/39 (10.3) 4/20 (12.2) 17/121 (14.0) 14/72 (15.9) 18/85 (21.2) 24/56 (12.4) 39/245 (15.9) 42/148

Females n/N (%)

(20.0) 2/24 (19.4) 5/52 (42.9) 3/34 (28.4) 10/110

Total n/N (%)

(8.3) 6/44 (13.6) (9.6) 19/124 (15.3) (8.8) 27/90 (30.0) (9.1) 52/258 (20.1)

n 5 number of crania exhibiting a fracture; N 5 number of examined crania; % 5 percentage of crania exhibiting a fracture. TABLE 8. Distribution of craniofacial fractures by cranial element FR n (%)

LP n (%)

RP n (%)

Males Females Total

12 (48.0) 6 (46.1) 18 (47.4)

10 (40.0) 3 (23.1) 13 (34.2)

1 (4.0) 4 (30.8) 5 (13.2)

Males Females Total

12 (33.3) 5 (31.2) 17 (32.7)

14 (38.9) 3 (18.8) 17 (32.7)

6 (16.7) 3 (18.8) 9 (17.3)

Males Females Total

22 (39.3) 4 (33.3) 26 (38.3)

13 (23.2) 1 (8.3) 14 (20.6)

14 (25.0) 6 (50.0) 20 (29.4)

LT n (%) AN 0 (0.0) 0 (0.0) 0 (0.0) EM 0 (0.0) 0 (0.0) 0 (0.0) LM 0 (0.0) 0 (0.0) 0 (0.0)

RT n (%)

OCC n (%)

FA n (%)

TOT N (%)

0 (0.0) 0 (0.0) 0 (0.0)

1 (4.0) 0 (0.0) 1 (2.6)

1 (4.0) 0 (0.0) 1 (2.6)

25 (100.0) 13 (100.0) 38 (100.0)

0 (0.0) 0 (0.0) 0 (0.0)

0 (0.0) 1 (6.2) 1 (1.9)

4 (11.1) 4 (25.0) 8 (15.4)

36 (100.0) 16 (100.0) 52 (100.0)

2 (3.6) 0 (0.0) 2 (2.9)

2 (3.6) 0 (0.0) 2 (2.9)

3 (5.3) 1 (8.3) 4 (5.9)

56 (100.0) 12 (100.0) 68 (100.0)

n 5 number of recorded craniofacial fractures by cranial element; % 5 percentage of total number of recorded craniofacial fractures; N 5 total number of recorded craniofacial fractures; FR 5 frontal bone; LP 5 left parietal bone; RP 5 right parietal bone; LT 5 left temporal bone; RT 5 right temporal bone; OCC 5 occipital bone; FA 5 facial region; TOT 5 total.

exhibit one indicator of deliberate violence (30/346 or 8.7% of the total sample). It then significantly increases during the EM period (44/313 or 14.1%; v2 5 4.259, df 5 1, P 5 0.039) and then slightly decreases in the LM sample where 58 skeletons exhibit at least one indicator of deliberate violence (58/466 or 12.4%). When multiple indicators of deliberate violence are recorded, a slight temporal increase of the frequency of such cases from the AN to the LM sample is noted (Table 11). In the AN sample, two skeletons display multiple indicators of deliberate violence (2/346 or 0.6% of the total sample); in the EM sample, there are four such skeletons (4/313 or 1.3%); and in the LM sample, eight skeletons exhibit multiple indicators of deliberate violence (8/466 or 1.7%). None of these differences are, however, significant. Because of the fact that 13 of 14 individuals with multiple skeletal indicators of intentional violence are males, the trend is slightly more pronounced in the male subsamples (1.0% in the AN series, 2.4% in the EM, and 2.8% in the LM series), but again without achieving significance. At the level of the complete analyzed sample (all three series), males are 10 times more likely to exhibit multiple indicators of violence than females (13/642 or 2.0% in males compared to 1/483 or 0.2% in females).

DISCUSSION As previously noted, historical sources suggest a gradual temporal increase in violence from the AN through the EM to the LM periods in the eastern Adriatic coast and interpret this as a result of recurrent armed conflicts related to fighting over control of the wealthy Dalmatian cities, feudal anarchy, and military intrusions by the Mongols or Ottoman Turks. During the same periAmerican Journal of Physical Anthropology

ods, people living in this area underwent a significant change of lifestyle—from an urban to a predominantly rural type of existence related to the destruction and abandoning of large urban centers caused by the disintegration of the Western Roman Empire, and large-scale immigration by various Germanic and Slavic tribes. The purpose of our investigation was to test whether trauma frequencies and distributions in a large number of skeletons dated to these periods, and from this region, confirm historical data that suggest a temporal increase in deliberate violence or, alternatively, show no significant increase in deliberate violence, but document changes related to the switch from an urban to a rural lifestyle. Before we discuss the results of our analyses, it is germane to point out some ambiguities related to the use of different methods for calculating bone trauma frequencies. As Domett and Tayles (2006) noted, the use of different methods for calculating bone trauma frequencies can produce different results. Calculating trauma frequencies by skeleton can result in ambiguous results as skeletons may, because of numerous reasons, differ in their degree of preservation. As archaeological skeletal remains are often incomplete, with missing or broken elements, using only complete skeletons to calculate fracture frequencies will underestimate the number of fractures (Pietrusewsky and Douglas, 2002; 104). As a consequence, most paleopathologists advocate calculating trauma frequencies by skeletal elements (e.g., Jurmain, 1999; Lovell, 2008). Although the two methods can produce different results, in this study, both show a clear trend of increased bone trauma frequencies from the AN to the LM period. The results of our analyses suggest, however, that the underlying cause responsible for this increase is ambiguous. It may, as suggested by historic sources, be


33

Fig. 2. Distribution of craniofacial injuries in the AN series.

Fig. 3. Distribution of craniofacial injuries in the EM series.

the result of increased intentional violence during the EM and LM periods caused by increasingly common armed conflicts related to feudal anarchy or conflicts between early Croats and either: Byzantium, Venice, Mongols, Ottoman Turks, but it may also be related to different and greater risks associated with significant differences in lifestyles between the early and LM Croats and the AN populations that inhabited Croatia from the 2nd to the 6th centuries AD. Skeletal material from the AN sample belongs to the inhabitants of two prosperous urban communities—modern Zadar and Vid—both of which achieved the status of Roman colonies very early, during Caesar’s rule in the 1st century BC (Colonia Iulia Iader and Colonia Iulia Narona). Skeletons from the EM sample, on the other hand, derive from small rural communities, while those from the LM series from both urban and rural communities although it is fairly certain that the urban communities of the LM period were more similar to small rural communities, than to the large, sophisticated urban centers of the Roman Empire. It is, therefore, fairly certain that individuals from these samples led significantly different lifestyles. The inhabitants of the Roman colonies from the eastern Adriatic coast were most likely engaged

in crafts, commerce, and public administration (Suic´, 2003). The number of people living in these urban communities, their comfort, security, and sanitation levels were, subsequent to the fall of the Western Roman Empire, not reached until the 19th century (Suic´, 2003). In contrast, the inhabitants of the EM villages from the hinterland lived in small rural communities and were predominantly engaged in transhumant pastoralism and various forms of agriculture (Goldstein, 1995). The rural component of the LM series very likely lived in almost identical conditions as the individuals from the EM sample while the urban component of this series is represented by individuals from significantly smaller and less prosperous communities than those that existed during the AN period. Additionally, the infrastructure of these urban centers was, at best, rudimental enabling living conditions that were probably considerably more similar to those in the rural medieval communities than to the preceding AN urban centers (Goldstein, 1995). Analyzing trauma frequencies and distributions by sex can potentially clarify the reasons for the increase in trauma frequencies during the medieval period in Croatia. When the sexes are compared within the samples, males in both medieval samples exhibit significantly American Journal of Physical Anthropology

M. SˇLAUS ET AL.

34

higher fracture frequencies when calculated by skeleton and higher fracture frequencies calculated by long bone than females. Although historic sources detailing the life of EM rural populations that inhabited the eastern Adriatic are nonexistent, written records for later periods (the LM and early modern period) are more abundant and can be used as broad indicators of the way of life during the EM period. Historical and ethnographic data indicate a strict sexual division of labor in rural communities from

the Adriatic hinterland during the LM and the early modern periods. Males performed physically more demanding tasks: plowing, hoeing, mowing and clearing land, carpentry, preparation and treatment of hides, woodcutting, loading carts, work with large domestic animals, and transhumant pastoralism (Ivanisˇevic´, 1987; Trazˇivuk, 2001). Women performed tasks related with milk and textile processing (milking, production of cheese, washing, screening and wool spinning, weaving and production of clothes, and other textile items), all activities related to the house and yard such as gardening, fetching water, washing and cleaning, cooking, and animal feeding, as well as taking care of children (Muraj, 2004; Sˇestan, 2008). These data clearly document higher fracture risks resulting from everyday activities in males and are, therefore, consistent with the results of trauma analyses that document significantly higher male fracture frequencies calculated by skeleton and much higher male fracture frequencies calculated by long bone during the medieval periods. In contrast to both medieval periods, males and females from the AN urban series exhibit almost identical fracture frequencies, once again regardless of the methods used to calculate them. This distribution is not specific to AN urban populations from the eastern Adriatic. Similar distributions have been observed in AN urban populations from Hungary (Gilmour, 2009), Italy (Paine et al., 2009), and other parts of Croatia (Novak et al., 2009). The reason may be that, as Berdowski (2007) noted, despite legal, ideological, and cultural limitations, women played an important role in the Roman economy. This was particularly true for brick and tile production, production of pottery and amphorae, food processing, cloth production, and some areas of commerce. Women participated in this not only as part of the workforce, but also as business managers. In some sectors, such as brick and tile production, the proportions of men and women could almost be equal (Berdowski, 2007). It, therefore, appears that the sexual division of labor in AN urban communities was less pronounced than in the TABLE 9. Sex and age distribution of perimortem injuries by skeleton

Male Female Total Fig. 4. Distribution of craniofacial injuries in the LM series.

AN

EM

LM

n/N (%)

n/N (%)

n/N (%)

2/195 (1.0) 0/151 (0.0) 2/346 (0.6)

2/164 (1.2) 0/149 (0.0) 2/313 (0.6)

2/283 (0.7) 1/183 (0.5) 3/466 (0.6)

n 5 number of skeletons exhibiting perimortem injuries; N 5 number of examined skeletons; % 5 percentage of skeletons exhibiting perimortem injuries.

TABLE 10. Description of the perimortem injuries Chronological period

Site/grave number

Sex/age

Description of perimortem trauma

AN AN EM

Zadar-Relja, 378B Zadar-Relja, 661 Sˇibenik, 48

M, 30–40 M, 40–50 M, 30–40

EM LM LM LM

Velim, 22 Zadar-Jazine, 40 Lisˇtani, 161 Dugopolje, 81

M, 30–40 M, 25–30 F, 20–30 M, 30–35

Large cut to the left parietal bone and occipital bone Piercing trauma to the left parietal bone Cut through right parietal and occipital bone, two cuts to right mastoideus, and cut to right clavicle Cut to the left scapula, T2, and 1st left rib Piercing trauma to the left parietal bone Cut to the left scapula Cut to the right clavicle, right parietal bone, and right temporal bone

American Journal of Physical Anthropology

35

INTENTIONAL VIOLENCE IN THE EASTERN ADRIATIC rural medieval populations and that as a consequence men and women in Roman cities were exposed to similar risk levels while engaged in everyday tasks. The result of this is that both sexes from the AN sample exhibit similar fracture frequencies. When analyses of fractures are made by side, higher long bone fracture frequencies (upper and lower limbs combined) are recorded on the left side in the EM and LM series, while identical frequencies on both sides are recorded in the AN sample. These results mirror those reported in several studies dealing with fracture frequencies in AN and medieval populations from Europe in the sense that these studies report mixed results: some report higher long bone fracture frequencies on the right side (e.g., Sˇlaus and Novak, 2006; Gilmour, 2009; Novak et al., 2009), while others report higher frequencies on the left side (e.g., Judd and Roberts, 1999; Djuric´ et al., 2006). At this point, it is unclear what causes are responsible for these variations in long bone fracture side predominance frequencies. In terms of injuries caused by intentional violence, the left side of the body is generally thought to bear the brunt of trauma as right-handed weapon wielders will preferentially strike their opponents on their left side in scenarios involving face-to-face combat. However, not all combat is carried out face-to-face and not all long bone fractures result from intentional violence. The final recorded frequencies

Fig. 5. Perimortem sharp force lesion (A) and antemortem blunt force trauma (B). Zadar-Relja, adult male.

are thus clearly the result of numerous biological, behavioral, and cultural factors, suggesting that further multidisciplinary analyses are necessary if we are to obtain more conclusive answers. Long bone injuries generally associated with accidents are radial fractures (especially Colles’ fractures), lower limb injuries, and clavicular fractures. In the three analyzed series, radial fractures are either the most common long bone fracture (in the AN series) or the second most common fracture (in the two medieval series with the provision that in the LM series radial and tibial fractures exhibit identical frequencies). All the recorded radial injuries are located on the distal part of the bone (mostly Colles’ fractures). This type of trauma occurs when a falling individual puts out his hand in order to break the fall (Alvrus, 1999). Of potentially greater interest, however, is the clear temporal increase in tibial and fibular fractures in the three series. The vast majority of these are either oblique or spiral midshaft fractures or fractures of the proximal and distal epiphyses. Oblique tibial fractures are usually associated with acci-

Fig. 6. Reconstruction of perimortem piercing injury. A: An antique period iron arrowhead penetrating the skull. The arrowhead was not recovered with this individual but is part of the antique weapons collection from the Archaeological museum in Zagreb. It is used in this picture to illustrate the manner in which such an injury could be sustained; B: endocranial ‘‘cone’’shaped defect. Zadar-Relja, adult male.

TABLE 11. Co-occurrence of the skeletal indicators of intentional violence Site and sex Zadar-Relja, male Zadar-Relja, male Velim, male Velim, female Sˇibenik, male Sˇibenik, male Lisˇtani, male Lisˇtani, male Lisˇtani, male Lisˇtani, male Dugopolje, male Dugopolje, male Dugopolje, male Dugopolje, male

Chronological period

Craniofacial

AN AN EM EM EM EM LM LM LM LM LM LM LM LM

X X

Perimortem

Sharp force X

X X

X

X

X

X

X

X X X X

X X X X X

X X X

‘‘Parry’’

X

X X X X X

X


36

M. SˇLAUS ET AL.

dents, generally falls from a low height (Russell et al., 1991), while spiral fractures, also accident-related, occur when the foot is stable, and the body is twisted over the foot (Trafton, 2003). Fractures of the proximal and distal epiphysis of the tibia result from repeated loading (Abel, 2004) or are caused by falls from a height (Bartlett et al., 1997). The clear temporal increase in the frequencies of these injuries may be related to the elevated risk of falls encountered while performing various agricultural activities in the rugged terrain of the Adriatic hinterland. To get a clearer idea of the underlying cause of the increased fracture frequencies in the two medieval periods, an effort was made to distinguish between long bone fractures that resulted from accidents and those that resulted from deliberate violence. A part of the skeleton that has frequently been used as an indicator of deliberate violence is the ulna, particularly the presence of ‘‘parry’’ fractures, although the problem of attributing such fractures exclusively to intentional violence has been raised and thoroughly explained by Smith (1996) and Judd and Roberts (1999). When criteria established by Judd (2008) are applied to ulnar fractures analyzed in this study, a constant increase of ‘‘parry’’ fractures from the AN to the LM period is clearly visible. The fact that the vast majority of recorded ‘‘parry’’ fractures in all three series are noted in males suggests that interpersonal violence was mostly male-oriented. Several authors (e.g., Walker, 1989; Alvrus, 1999; Standen and Arriazza, 2000; Jurmain et al., 2009) also point out that high frequencies of head and face trauma are conclusive proof of intentional violence. Walker (1997; 160) states that ‘‘from a strategic standpoint, the head and especially the face are attractive targets because injuries of this area can be very painful. Well placed blows to the head are also likely to produce bleeding and conspicuous bruises that serve as a highly visible symbol of the aggressor’s social dominance.’’ All the series analyzed in this work exhibit relatively high frequencies of craniofacial trauma (between 15.9 and 20.1%) with none of the differences, either at the level of total samples, or when analyzed by sex, achieving significance. Of interest is the fact that the AN and EM series display almost identical values, while higher frequencies are observed in the LM sample. Males exhibit higher frequencies of craniofacial injuries in all series, which mirrors the distribution recorded in numerous archaeological populations. According to Jurmain (1999), nearly all bioarchaeological studies have reported a higher frequency of cranial injuries in males, which is a result of sexual division of labor, where more difficult and hazardous activities are performed by males as well as cultural behavior that associates virility with aggressiveness (Ember and Ember, 1997; Robb, 1997). The majority of the recorded craniofacial injuries in the analyzed samples are depression fractures, again mirroring the fact that cranial vault depressed injuries are probably the most common type of head injury in archaeological settings (Roberts and Manchester, 1995). Most of the craniofacial injuries in the AN and the EM samples are located on the left side, which is expected, because the left side of the skull is the most frequent injury site in face-to-face combat with a right-handed aggressor (Bolyston, 2000; Djuric´ et al., 2006; Owens, 2007; Jimeńez-Brobeil et al., 2009). The higher frequency of cranial injuries on the right side observed in the LM series may indicate that the injuries were sustained American Journal of Physical Anthropology

while the victims were fleeing their attacker or perhaps while they were laying prone (Larsen, 1999). The distribution of craniofacial injuries by bone element shows that in all three samples, the frontal bone was the most affected part of the skull (in the AN series 47.4%, EM 32.7%, and LM 38.3%). This type of distribution is consistent with the presence of interpersonal violence. In his studies dealing with craniofacial injuries in archaeological and modern samples, Walker (1989, 1997) states that the frontal location of cranial fractures is an indicator of deliberate violence. Additional evidence for the presence of deliberate violence in all three samples is the presence of nasal fractures (one in the AN sample, eight in the EM sample, and two in the LM sample), injuries that have a high specificity for the clinical diagnosis of assault (Lovell, 2008). Perimortem injuries on the skeleton provide the most direct evidence of intentional violence (Merbs, 1989; Alvrus, 1999). In this context, it is important to note that the frequencies of skeletons exhibiting perimortem trauma are identical and very low (0.6%) in all three series. The high proportion of males (6/7) in the sample of individuals with perimortem injuries is consistent with the previously noted male predominance in ‘‘parry’’ fractures as well as the higher craniofacial fracture frequencies recorded in men. Together, these data strongly suggest that intentional violence in the eastern Adriatic samples was sex-specific and almost exclusively directed against males. The majority of the recorded perimortem injuries were caused by sharp-bladed weapons, such as swords (males from Zadar-Relja, Sˇibenik, Velim, and Dugopolje) or a battle knife (the female from Lisˇtani). Two penetrating wounds recorded in males from ZadarRelja and Zadar-Jazine were most likely caused by arrows. The distribution of perimortem injuries in the skeleton shows that the primary target of attack was the upper third of the body, particularly the head and neck region. This distribution is similar to the distribution of perimortem injuries recorded in battle or massacre related skeletal series from Aljubarrota (Cunha and Silva, 1997), Uppsala (Kjellstro¨m, 2005), Sandbjerget ˇ epin (Sˇlaus (Bennike, 1998), Towton (Novak, 2000), and C et al., 2010). Analyses of the presence of one skeletal indicator of deliberate violence in the three series show a significant temporal increase from the AN to the medieval periods, while analyses of the co-occurrence of two or more skeletal indicators of deliberate violence—injuries of the facial and frontal region of the cranium, sharp force lesions, ‘‘parry’’ fractures, and perimortem trauma, show a slight temporal increase that does not achieve significance. The predominance of male skeletons (13 of 14) exhibiting multiple indicators of deliberate violence again points to male-directed interpersonal aggression. The injuries observed in the only female with multiple signs of deliberate violence can, cautiously, be interpreted as domestic assault, because this individual exhibits both a ‘‘parry’’ fracture and a craniofacial injury, a distribution of injuries that has been attributed to domestic violence in both bioarchaeological and clinical studies (e.g., Fonseka, 1974; Walker, 1997; Novak, 2009). In conclusion, the results of our analyses show that the three analyzed Croatian series exhibit a significant temporal increase of total fracture frequencies when calculated by skeleton as well as of individuals exhibiting one skeletal indicator of deliberate violence (sharp force

INTENTIONAL VIOLENCE IN THE EASTERN ADRIATIC lesions, craniofacial injuries, ‘‘parry’’ fractures, or perimortem trauma). No significant temporal increases were, however, noted in the frequencies of craniofacial trauma, ‘‘parry’’ fractures, perimortem injuries, or of individuals exhibiting multiple skeletal indicators of intentional violence. Together with the high frequencies of accidentrelated injuries to the radius, and the temporal increase in tibial and fibular fractures, these results combine both scenarios that we proposed to test at the beginning of this analysis. The accumulated data therefore suggest that the temporal increase in fracture frequencies recorded in the eastern Adriatic was most likely caused by a combination of factors that included both the historically documented increase of intentional violence caused by recurrent armed conflicts related to fighting over control of the Dalmatian cities, feudal anarchy, and intrusions by the Mongols and Ottoman Turks, as well as a significant change in lifestyle that accompanied the transition from a relatively affluent AN urban lifestyle, to a more primitive rural medieval way of life.

LITERATURE CITED Abel SM. 2004. Biocultural variation of skeletal trauma in contemporary Greeks, Unpublished PhD thesis, University of Florida, Gainesville. Alvrus A. 1999. Fracture patterns among the Nubians of Semna South, Sudanese Nubia. Int J Osteoarchaeol 9:417–429. Bartlett CS, Weiner LS, Yang EC. 1997. Treatment of type II and type III open tibia fractures in children. J Orthop Trauma 11:357–362. Bennike P. 1998. De faldne fra Krigergraven: analyse af skeletterne fra Sandbjerget. Liv og Levn 12:14–21. Berdowski P. 2007. Some remarks on the economic activity of women in the Roman Empire: a research problem. In: Berdowski P, Blahaczek B, editors. Haec mihi in animis vestris templa. Studia classica in memory of Professor Lesław Morawiecki. Rzeszo´w: Instytut Historii UR. p 283–298. Berryman HE, Symes SA. 1998. Recognizing gunshot and blunt cranial trauma through fracture interpretation. In: Reichs KJ, editor. Forensic osteology: advances in the identification of human remains. Springfield, IL: Charles C. Thomas. p 333–352. Boylston A. 2000. Evidence for weapon related trauma in British archaeological samples. In: Cox M, Mays S, editors. Human osteology in archaeology and forensic medicine. London: Greenwich Medical Media. p 357–380. Bridges PS. 1996. Warfare and mortality at Koger’s Island, Alabama. Int J Osteoarchaeol 6:66–75. Brusic´ Z, Glusˇcˇevic´ S. 1990. Zadar—istrazˇivanje anticˇkog grada mrtvih. Zagreb: Muzejsko Galerijski Centar. Buikstra JE, Ubelaker DH. 1994. Standards for data collection from human skeletal remains. Fayetteville, AR: Arkansas Archaeological Survey. Cunha E, Silva AM. 1997. War lesions from the famous Portuguese medieval battle of Aljubarrota. Int J Osteoarchaeol 7:595–599. Djuric´ MP, Roberts CA, Rakocˇevic´ ZB, Djonic´ DD, Lesˇic´ AR. 2006. Fractures in Late Medieval skeletal populations from Serbia. Am J Phys Anthropol 130:167–178. Domett KM, Tayles N. 2006. Adult fracture patterns in prehistoric Thailand: a biocultural interpretation. Int J Osteoarchaeol 16:185–199. Ember CR, Ember M. 1997. Violence in the ethnographic record: results of cross-cultural research on war and aggression. In: Martin DL, Frayer DW, editors. Troubled times: osteological and archaeological evidence of violence. New York: Gordon and Breach. p 1–20. Facchini F, Rastelli E, Belcastro MG. 2007. Peri mortem cranial injuries from a medieval grave in Saint Peter’s Cathedral, Bologna, Italy. Int J Osteoarchaeol 18:421–430.

37

Fadic´ I. 2007. Zadar-Relja (Trgovacˇki centar). In: Mesic´ J, editor. Hrvatski arheolosˇki godisˇnjak, Vol. 3. Zagreb: Ministarstvo kulture Republike Hrvatske. p 347–349. Fonseka S. 1974. A study of wife-beating in the Camberwell area. Br J Clin Prac 28:400–402. Galloway A, Symes SA, Haglund WD, France DL. 1999. The role of forensic anthropology in trauma analysis. In: Galloway A, editor. Broken bones: anthropological analysis of blunt force trauma. Springfield, IL: Charles C. Thomas. p 5–31. Gilmour RJ. 2009. Broken Romans: long bone trauma in a civilian population at Aquincum, Budapest, Hungary, Unpublished MSc thesis, Durham University, Durham. Gjurasˇin H. 2007. Kasnosrednjovjekovna groblja opcíne Dugopolje. Split: Muzej Hrvatskih Arheolosˇkih Spomenika. Goldstein I. 1995. Hrvatski rani srednji vijek. Zagreb: Novi Liber—Zavod za hrvatsku povijest Filozofskog fakulteta Sveucˇilisˇta u Zagrebu. Grauer AL, Roberts CA. 1996. Paleoepidemiology, healing and possible treatment of trauma in a medieval cemetery population of St. Helen-on-the-Walls, York, England. Am J Phys Anthropol 100:531–544. Houck MH. 1998. Skeletal trauma and the individualization of knife marks in bone. In: Reichs KJ, editor. Forensic osteology. Advances in the identification of human remains. Springfield, IL: Charles C. Thomas. p 410–424. Ivanisˇevic´ F. 1987. Poljica—narodni zˇivot i obicˇaji. Split: Knjizˇevni krug. Jimeńez-Brobeil SA, du Souich P, Al Oumaoui I. 2009. Possible relationship of cranial traumatic injuries with violence in the south-east Iberian Peninsula from the Neolithic to the Bronze Age. Am J Phys Anthropol 140:465–475. Judd M. 2004. Trauma in the city of Kerma: ancient versus modern injury patterns. Int J Osteoarchaeol 14:34–51. Judd MA. 2008. The parry problem. J Archaeol Sci 35: 1658–1666. Judd MA, Roberts CA. 1999. Fracture trauma in a medieval British farming village. Am J Phys Anthropol 109:229–243. Juric´ R. 2006. Velim–Velisˇtak. In: Mesic´ J, editor. Hrvatski arheolosˇki godisˇnjak, vol. 2. Zagreb: Ministarstvo kulture Republike Hrvatske. p 316. Jurin-Starcˇevic´ K. 2008. Zemlja, seljasˇtvo i agrikultura u razdoblju osmanske vlasti. In: Kusin V, editor. Dalmatinska zagora—nepoznata zemlja. Zagreb: Ministarstvo kulture Republike Hrvatske. p 233–243. Jurmain R. 1999. Stories from the skeleton. Behavioral reconstruction in human osteology. Amsterdam: Gordon and Breach. Jurmain R, Bartelink EJ, Leventhal A, Bellifemine V, Nechayev I, Atwood M, DiGiuseppe D. 2009. Paleoepidemiological patterns of interpersonal aggression in a prehistoric central California population from CA-ALA-329. Am J Phys Anthropol 139:462–473. Kjellstro¨m A. 2005. A sixteenth-century warrior grave from Uppsala, Sweden: the battle of Good Friday. Int J Osteoarchaeol 15:23–50. Kolega M. 2001. Nin—arheolosˇka istrazˇivanja u sklopu zˇupne crkve sv. Asela (Anselma). Obavijesti Hrvat Arheol Drusˇ 2: 83–90. Kolega M. 2002. Nin—nadzˇupni kompleks sv. Anselma (Asela) —Istrazˇivanja godine 2001. Obavijesti Hrvat Arheol Drusˇ 2:73–78. Krncˇevic´ Zˇ. 1998. Srednjovjekovna arheolosˇka nalazisˇta na sˇibeˇ ecˇuk B, editor. Izdanja Hrvatskog nskom podrucˇju. In: C arheolosˇkog drusˇtva, Vol. 19. Zagreb: Hrvatsko arheolosˇko drusˇtvo. p 197–228. Larsen CS. 1999. Bioarchaeology: interpreting behavior from the human skeleton. Cambridge: Cambridge University Press. Lovejoy CO, Heiple KG. 1981. Analysis of fractures in skeletal populations with an example from the Libben Site, Ottowa County, Ohio. Am J Phys Anthropol 55:529–541. Lovell N. 1997. Trauma analysis in paleopathology. Yearb Phys Anthropol 40:139–170. Lovell NC. 2008. Analysis and interpretation of skeletal trauma. In: Katzenberg MA, Saunders SR, editors. Biological anthropology of the human skeleton. Hoboken, NJ: Wiley. p 341–386.


38

M. SˇLAUS ET AL.

Maples WR. 1986. Trauma analysis by the forensic anthropologist: In: Reichs KJ, editor. Forensic osteology: advances in the identification of human remains. Springfield, IL: Charles C. Thomas. p 218–228. Mardesˇic´ J. 2004. The burials: the archaeological context and the pottery finds. In: Marin E, Vickers M, editors. The rise and fall of an imperial shrine—Roman sculpture from the Augusteum at Narona. Split: Arheolosˇki muzej Split. p 255–264. Maric´ M. 2005. Rezultati arheolosˇkih iskopavanja u Lisˇtanima kraj Livna, lokalitet Podvornice (Bosna i Hercegovina). Obavijesti Hrvat Arheol Drusˇ 1:74–84. Marin E. 2002. Eresˇove bare—villa suburbana iz 3. stoljecá i starokrsˇcánska crkva iz 7. stoljecá u Naroni. Vjesn Arheol Hist Dalm 94:9–80. Martincˇic´ O. 2009. Paleodemografska i paleopatolosˇka analiza ljudskog osteolosˇkog materijala sa srednjovjekovnog groblja uz crkvu Sv. Asela u Ninu, Unpublished MSc thesis, University of Zagreb, Zagreb. _ can MY, Kennedy KAR, editors. Merbs CF. 1989. Trauma. In: Is Reconstruction of life from the skeleton. New York: Alan R. Liss. p 161–190. Milner GR, Anderson E, Smith VG. 1991. Warfare in late prehistoric west-central Illinois. Am Antiq 56:581–603. Muraj A. 2004. Privred-ivanje: poljodjelstvo i rukotvorstvo. In: Vitez Z, editor. Hrvatska tradicijska kultura. Zagreb: Institut za Etnologiju i Folkoristiku. p 95–163. Murphy MS, Gaither C, Goycochea E, Verano JW. Cock G. 2010. Violence and weapon-related trauma at Puruchuco-Huaquerones, Peru. Am J Phys Anthropol 142:636–649. Novak G. 2001. Prosˇlost dalmacije. Zagreb: Golden Marketing. Novak M, Sˇlaus M. 2007. Ucˇestalost i distribucija cribrae orbitaliae u kasnosrednjovjekovnoj populaciji iz Dugopolja. Starohrvatska Prosvjeta 34:451–475. Novak M, Premuzˇic´ Z, Vyroubal V, Sˇlaus M. 2009. Bioarchaeology of the late Antique population from Sˇtrbinci. Arh Rad Raspr 16:265–326. Novak S. 2000. Battle related trauma. In: Fiorato V, Boylston A, Knu¨sel C, editors. Blood red rose: the archaeology of a mass grave from the Battle of Towton AD 1461. Oxford: Oxbow Press. p 90–102. Novak S. 2009. Beneath the façade: a skeletal model of domestic violence. In: Gowland R, Knu¨sel C, editors. The social archaeology of human remains. Oxford: Oxbow Press. p 238–252. Owens LS. 2007. Craniofacial trauma in the Prehispanic Canary Islands. Int J Osteoarchaeol 17:465–478. Paine RR, Vargiu R, Signoretti C, Coppa A. 2009. A health assessment for Imperial Roman burials recovered from the necropolis of San Donato and Bivio CH, Urbino, Italy. J Anthropol Sci 87:193–210. Petrinec M. 2002. Dosadasˇnji rezultati istrazˇivanja ranosrednjovjekovnog groblja u Glavicama kraj Sinja kao prilog razrjesˇavanju problema kronologije starohrvatskih grobalja. Opusc Archaeol 26:205–246. Pietrusewsky M, Douglas MT. 2002. Ban Chiang, a prehistoric village site in northeast Thailand: the human skeletal remains. Philadelphia: University of Pennsylvania, Museum of Archaeology and Anthropology. Reichs KJ. 1998. Postmortem dismemberment: recovery, analysis and interpretation. In: Reichs KJ, editor. Forensic osteology. Advances in the identification of human remains. Springfield, IL: Charles C. Thomas. p 353–388. Robb JE. 1997. Violence and gender in early Italy. In: Martin DL, Frayer DW, editors. Troubled times: osteological and archaeological evidence of violence. New York: Gordon and Breach. p 108–141. Roberts CA, Manchester K. 1995. The archaeology of disease. Ithaca, NY: Cornell University Press. Rogosˇic´ J. 2001. Prirodna obiljezˇja podrucˇja Dugopolja. In: Gulin A, editor. Zbornik radova opcíne Dugopolje. Zagreb–Dugopolje: Opcína Dugopolje. p 29–46. Russell T, Taylor J, LaVelle D. 1991. Fractures of the tibia and fibula. In: Rockwood C, Green D, Bucholz R, editors. Rockwood and Green’s fractures in adults, Vol. 2. Philadelphia: Lippincott. p 1915–1982.


Sˇaric´ M. 2008. Ekohistorijski osvrt na planine i morlacˇki svijet. In: Kusin V, editor. Dalmatinska zagora—nepoznata zemlja. Zagreb: Ministarstvo kulture Republike Hrvatske. p 221–231. Sauer NJ. 1998. The timing of injuries and manner of death: distinguishing among antemortem, perimortem and postmortem trauma. In: Reichs KJ, editor. Forensic osteology. Advances in the identification of human remains. Springfield, IL: Charles C. Thomas. p 321–332. Scott RM, Buckley HR. 2010. Biocultural interpretations of trauma in two prehistoric Pacific Island populations from Papua New Guinea and the Solomon Islands. Am J Phys Anthropol 142:509–518. Sˇestan I. 2008. Tradicijsko gospodarstvo, poljodjelstvo, vinogradarstvo i stocˇarstvo: In: Kusin V, editor. Dalmatinska zagora— nepoznata zemlja. Zagreb: Ministarstvo kulture Republike Hrvatske. p 463–476. Sˇlaus M. 2008. Osteological and dental markers of health in the transition from the Late Antique to the Early Medieval period in Croatia. Am J Phys Anthropol 136:455–469. Sˇlaus M, Novak M. 2006. An analysis of traumas in medieval samples from Klisˇkovac and Crkvari. Pril Inst Arh 23:213–228. Sˇlaus M, Novak M, Vyroubal V, Bedic´ Zˇ. 2010. The harsh life on the 15th century Croatia-Ottoman empire military border: analyzing and identifying the reasons for the massacre in ˇ epin. Am J Phys Anthropol 141:358–372. C Smith GE, Jones FW, editors. 1910. Fractured bones and dislocations. The archaeological survey of Nubia, Vol II: Report on the human remains. Cairo: National Printing Department. p 293–342. 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. Standen VG, Arriaza BT. 2000. Trauma in the preceramic coastal populations of northern Chile: violence or occupational hazards? Am J Phys Anthropol 112:239–249. Suic´ M. 2003. Anticˇki grad na istocˇnom Jadranu. Zagreb: Golden Marketing–Tehnicˇka knjiga. Symes SA, Berryman HE, Smith OC. 1998. Saw marks in bone: introduction and examination of residual kerf contour. In: Reichs KJ, editor. Forensic osteology. Advances in the identification of human remains. Springfield, IL: Charles C. Thomas. p 389–409. Trafton PG. 2003. Tibial shaft fractures. In: Browner BD, Jupiter JB, Levine AM, Trafton PG, editors. Skeletal trauma: basic science, management, and reconstruction. Philadelphia: Saunders. p 2131–2256. Trazˇivuk BV. 2001. Prilog istrazˇivanju etnografije Dugopolja. In: Gulin A, editor. Zbornik radova opcíne Dugopolje. Zagreb– Dugopolje: Opcína Dugopolje. p 461–472. United Nations. 1993. Declaration on the elimination of violence against women. Resolution 48/104. New York: UN Department of Public Information. United States Department of Health and Human Services. 2011. Centers for Disease Control and Prevention. Intimate partner violence: definitions. Retrieved on March 7, 2011, from: http://www.cdc.gov/ViolencePrevention/intimatepartnerviolence/definitions.html. Wakely J. 1997. Identification and analysis of violent and nonviolent head injuries in osteo-archaeological material. In: Carman J, editor. Material harm: archaeological studies of war and violence. Glasgow: Cruithne Press. p 24–46. 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: Martin DL, Frayer DW, editors. Troubled times: osteological and archaeological evidence of violence. New York: Gordon and Breach. p 145–180. Walker PL. 2001. A bioarchaeological perspective on the history of violence. Annu Rev Anthropol 30:573–596. Wheatley B. 2008. Perimortem or postmortem bone fractures? An experimental study of fracture patterns in deer femurs. J Forensic Sci 53:69–72.