Symbolic bones and interethnic violence in a frontier zone, northwest Mexico, ca. 500–900 C.E.

Significance

Persistent interethnic violence has affected some global regions for centuries. Recent research reveals that major outbreaks are often prevented or limited by creative social action. In the prehispanic Northern Frontier of Mesoamerica, approximately 500–900 C.E., people of different ethnic backgrounds struggled for standing in a shifting sociopolitical landscape. Evidence is consistent with long-term social violence, but also with the use of the dead to communicate a range of symbolic messages. Complex arrays of human skeletal material commemorated past physical conflicts, possibly discouraging their repetition, while also connecting the living symbolically with a metaphysical realm inhabited by ancestors and deities. This article highlights the postmortem agency of the dead and illustrates their roles in structuring social relations.

Abstract

Although extensive deposits of disarticulated, commingled human bones are common in the prehispanic Northern Frontier of Mesoamerica, detailed bioarchaeological analyses of them are not. To our knowledge, this article provides the first such analysis of bone from a full residential-ceremonial complex and evaluates multiple hypotheses about its significance, concluding that the bones actively represented interethnic violence as well as other relationships among persons living and dead. Description of these practices is important to the discussion of multiethnic societies because the frontier was a context where urbanism and complexity were emerging and groups with the potential to form multiethnic societies were interacting, possibly in the same ways that groups did before the formation of larger multiethnic city-states in the core of Mesoamerica.

The Epiclassic period (600–900 C.E.) was a time of rapid social and demographic change throughout Mesoamerica that began with the collapse of the massive city of Teotihuacan (1–3). Styles of architecture, costume, and iconography merged across old boundaries. New, smaller-scale polities formed in an expanding Northern Frontier (Fig. 1), including La Quemada (4). Whether or not the legendary migrants of the Teotihuacan diaspora actually reached the Northern Frontier (5), the character of this region changed dramatically during the Epiclassic.

Fig. 1.

Sites and areas mentioned in text. Shaded portion is Northern Frontier region. Sites: 1, La Quemada; 2, Teotihuacan; 3, Tzintzuntzan; 4, Alta Vista; 5, Cerro del Huistle; 6, Tingambato; 7, Cumupa. Areas: a, Atemajac Valley; b, Wixárika; c, Caxcan; d, Acaxee.

Sixteenth century accounts characterize the Northern Frontier region as multiethnic (6). These ethnic divisions must be very old; historical linguists date language fission at approximately 4,500–6,000 y B.P. (7, 8). The regions’ peoples spoke ∼30 languages (9); the relation between any pair of regional centers probably was interethnic. The formation of the Purépecha (Tarascan) state ∼1300 C.E. involved numerous instances of interethnic conflict and the suppression of former ethnic identities (10). That state, centered on Tzintzuntzan, was three- to seven-times larger than La Quemada (11–13). Still, such smaller polities could not have avoided interethnic external relations.

Two prominent characteristics of La Quemada suggest social violence: defensible architecture and extensive deposits of human skeletal material. The builders augmented natural cliffs with thick enclosing walls; they placed buildings on terraces with steep walls and narrow access from below, disadvantaging attackers. Cross-culturally, such architecture may defend ritually important places as well as residents (14).

Human bone is a powerful material that the living manipulate in many ways (15, 16). Previously described human remains at La Quemada included an immense deposit of disarticulated bone in the Hall of Columns (17) and another massive disarticulated bone assemblage at the foot of a small pyramid (18, 19). Individual inhumations were beneath the floor of the main ball court (20) and in a pyramid (21). Previous analyses questioned whether residents kept human bones in La Quemada’s living spaces in veneration of ancestors, to represent violent encounters with enemies, or both (22, 23). The present study resolves this question.

Mortuary patterns similar to those of La Quemada occur at other contemporaneous regional centers, although detailed discussions of taphonomy are lacking for most (24). At Alta Vista, masses of disarticulated bone rested in several structures (25), and groups of skulls and long bones hung in the Temple of the Skulls (26). At Cerro del Huistle, skulls hung outdoors on racks in the main plaza (27). Mortuary practices at La Quemada differed from those at smaller surrounding sites, where most burials were subfloor inhumations in articulated position.

La Quemada was the centerpiece of a landscape made up of roads, villages, hamlets, shrines, irrigation canals, agricultural terraces, and natural landforms. Based on analogies with the Wixárika (Huichol), architecture and landforms were scenes of pilgrimages, dances, and processions that focused on ancestor veneration (28, 29), practices in which we suggest human bone played an active role.

In and Out of Doors

Terrace 18 (Fig. 2) was a temple location, residential complex, and outdoor arena. Activities would have approximated those in the modern Wixárika tukipa, a term that encompasses a temple, houses, and shrines surrounding a patio (30). Wixárika ceremonial officials and their families stay in the houses during feast periods. Terrace 18 had five houses, surrounding Patios B, C, D, E, and F. Tukipas also include small house-like shrines that were paralleled by small structures on Platforms 1–4. Two formal tombs were found, in Patio Groups B and D. Outdoor spaces included the causeway, patios and walkways, and middens. Passersby would have seen bone suspended in front of the temple. Occupants probably used the main patio for dances, ball games, and human sacrifice. Trash Middens 1 and 7 were directly associated with Terrace 18, and Midden 6 was nearby on a spur of the causeway. After abandonment, decomposition of buildings exposed the bones to puddling and erosion.

Fig. 2.

Bone deposits on Terrace 18 of La Quemada.

Bones Indoors

The temple (Room 1) contained a cache of manos (grinding stones), a large cooking pot, several mirror stones, and an incense burner that was suspended from the ceiling or in a wall niche. Three bone concentrations rested directly on the floor and evidently had been in bundles suspended from the ceiling, probably wrapped in organic material, and fell when the wrapping or cord disintegrated. Two, analytical units (AUs) 43 and 44, were intact and the other, AU 45, seemed to be rearranged by water movement across the floor. The bone in the temple is described elsewhere (22), but new analysis of the breakage patterns, cutmarks, and modification provides additional information (Table 1 and Table S1). The bones exhibited water damage, root etching, exfoliation, and much recent breakage. Some bones exhibited cutmarks and spiral fractures that may have occurred upon disarticulation from the body (Fig. S1). The majority of the proximal and distal parts of the long bones were still present; some femora were complete. Of the 130 fully analyzable bones, ∼20% had cutmarks, mainly on the cranium, femur, fibula, and humerus. A total of 67 cutmarks on cranial fragments were analyzed for shape (86% were V-shaped) and length (1.8 mm to 28.3 mm). On the long bones, almost all of the cutmarks were near the distal or proximal ends. Long bones generally had multiple cutmarks (10 bones had 171 cutmarks), ranging in length from 0.6 mm to 14.6 mm. All of the cutmarks were V-shaped, and their direction indicated that the goals were dismembering and defleshing. However, the shallowness of the cutmarks suggests that these activities occurred after the bodies were already either desiccated or partially skeletonized (31).

Table 1.

Bone summary by analytical unit

Variable	AU
	43	44	45	166	74A	74B	74C	74D	87	88	111	67
MNI	6	2	6	1	5	1	3	2	7	1	1	3
Adult	3		5	1	5	1	3	2	7	1	1	3
Male	3		5	1	5	1	3	2	7	1	1	3
Female	1		3		2
Subadult	3		1
Burning	N	N	N	N	N	N	Y	Y	N			Y
Perimortem	Y	Y	Y	Y	Y	Y		Y	Y			Y
Cutmarks	Y	Y	Y	Y	Y	Y			Y			Y

MNI, minimum number of individuals; N, no; Y, yes.

In AU 43 the long bones were partially broken, but all indications suggest that these bones were whole when placed in the temple. Most of the breakage on the bone ends was recent and many of the bones had partial proximal or distal ends intact. Cutmarks on the cranial pieces were shallow and likely made with bifacially flaked stone tools or retouched opportunistic blades (31). AU 44 had the fewest bones, all from the upper body, primarily arms. In AU 45, a larger cluster, all elements were represented except for hands, feet, and vertebrae. There was a great deal of breakage, exfoliation, and water destruction of all bone surfaces. A few had perimortem spiral fractures. Cutmarks near the proximal and distal ends of several of the long bones also pointed to defleshing.

The skeletal material in the Temple represented all age groups and both sexes. The majority of the bones were unmodified long bones with ample evidence of cutmarks. However, the placement, width, and depth of the cutmarks imply processing when the bodies were already desiccated or partially skeletonized. The deposits could represent three distinct episodes of bone cleaning for hanging or display; in that case they constitute a minimum of 14 individuals: 10 adults (two males, four females, and four unknowns) and 4 children (an infant and three juveniles). Alternatively, the three deposits may be part of a single preparatory process, in which case the minimum number of individuals is reduced (because of age and sex overlap) to eight individuals comprised of five adults (two males and three females) and three children (aged 2–12 y).

Given either scenario, this bone exhibition could represent the sequential deposition of community members as they died. To address the contrasting possibility that the individuals were defeated enemies, future isotopic or DNA analyses may help determine succession, relatedness, and birthplace. Currently we reject the foe explanation for two reasons. First, it is incongruent with the indoor-outdoor differences in bone modification identified here, indoor bone being modified postmortem and less destructively. Second, keeping enemy bodies in a temple would be inconsistent with practices in descendant Cora and Wixárika communities, where temples are places of ancestor worship and bundled bones of venerated ancestors were kept in temples as recently as the 17th and 18th centuries (29, 32), the latter in a traditional handmade seat called an equipal. In the Wixárika community of Keuruwit+a (Las Latas), B.A.N. asked why miniature equipals were suspended from the ceiling of the temple; a community member answered, “They are the ancestors.” We suggest that this custom echoes the inferred ancient practice of suspending ancestral remains.

A 1.2 × 1.2-m cobble-lined tomb (AU 101) was found beneath a wall in Patio Group B; only one bone fragment remained. Someone had looted the grave or the inhabitants had removed the bone to take it elsewhere. Contents included a pegged figurine that may have adorned a staff or baton of office. The figure was carved in a soft green stone and depicted a human head emerging from the mouth of a coyote, probably representing a shaman or warrior.

The second tomb, an individual interment in a substructural chamber (AU 166), was beneath the east end of Patio Group D. The chamber was faced with cut stone; it contained several broken, incomplete, decorated pots and 11 polished mirror stones. The tomb cut through Patio D, implying a late place in the construction sequence of the terrace. The femur of this very large and robust male bore cutmarks and perimortem fractures. The bone appeared to have been hollowed out by reaming. The cranial fragments were all from one individual and were partially refitted; these had cutmarks distributed across the parietal and occipital bones. There was a large perimortem trauma on the occipital with multiple fine striations (Fig. S2), suggestive of White’s percussion striae (33). The individual was over the age of 50, possibly a warrior who was killed by blunt force trauma applied to the back of the head. The cutmarks on the cranium were not diagnostic of a particular motivation, but scalping, defleshing, or removal of facial/ear tissues as trophies may explain them. The breakage on the ends of the long bones, with some cutmarks, could also be from defleshing and reduction as part of a sacrifice ritual or for marrow extraction.

Bones Outdoors

The main patio, 26 × 36 m, was delineated by raised banquettes; at the center was a small ball court. Human remains were displayed or deposited in the patio, on the front wall of the temple, and on the adjacent walkway on the west banquette. In contrast to the bones kept indoors, those that were kept outdoors had a strong tendency to exhibit deep, wide, V-shaped cutmarks (Fig. 3 and Fig. S3), indicative of perimortem processing (31).

Fig. 3.

Deep cutmarks on the right proximal femoral neck region circling the head of the femur (AU 45).

The most prominent outdoor display was a presumed skull rack (tzompantli) along the west wall of main patio. Four concentrations of human bones fell along the West Banquette Walkway at the edge of Patio A (AUs 74 A–D). These bones came to rest on or just above the patio floor and appeared to have been suspended from above the edge of the patio walkway. Unlike the temple, the deposit included only skulls and long bones. AU 74D was associated with tangible evidence of a rack; four courses of fallen cobbles in the patio marked a freestanding feature about 1-m wide. The bones were resting underneath the fallen cobbles, implying that they had hung on the side facing the main patio. We infer that the feature had a rock base and wooden superstructure, and that before it collapsed the bones fell at its base. The pattern of bones along the edge of the Patio A is isomorphic to that in Plaza A at nearby Cerro del Huistle, where the better preserved skulls exhibited drilling (27).

Beginning at the northern end of the western wall of Patio A, AU 74A consisted of fragmentary adult long bone shafts with postmortem damage and erosion of all bone surfaces. The bones appeared to have been purposefully modified so that only the shafts of each arm and leg bone remained. This treatment could have aided their drying if they were suspended. If primarily a display, it was constructed largely of arm and leg bones from five adults, of which three (based on dimorphic features) may be males.

AU 74B included highly fragmentary and exfoliated long bones and cranial fragments. One humerus shaft had perimortem equidistant excisions on the lateral and medial sides, corresponding to musical instruments known as raspers (34). The bone deposit midway along the patio wall (AU 74C) was comprised of poorly preserved long bone shafts very similar to those in AU 74A. The southernmost bone deposit consisted of adult, probably male, long bone shafts (AU 74D).

Two drilled crania were included in the human remains that apparently hung from the front wall of the Temple (AU 87), paralleling those along the patio edge 3-m away. The only complete specimen was very fragile and exfoliated (Fig. 4). The cranium was fully adult but quite small and delicate with no discernable brow ridge or occipital margins, suggesting a female. At the top of the cranium, along the sagittal suture, was a drilled circular hole, 8 × 8 mm. The margins of the hole were even, with a slight beveling inward in the cranial walls. A second cranium composed of a frontal bone with parts of the left and right parietal bones attached also had an identically shaped hole, 10 × 10 mm. The completely fused sutures may indicate an adult over the age of 50. This frontal bone had two distinctive cutmarks along the left side of the temple. The long bones appeared to have been manually reduced to approximately the same size.

Fig. 4.

Skull perforated at apex for suspension (AU 87).

Each house included a sunken patio (tragaluz) of approximately 2 × 2 m. In Patio B (AU 111) were 73 very poorly preserved human bone fragments that apparently hung from a portico over the walkway. All of the bone was concentrated along the west edge of the patio; had it been on a wall, it would have fallen at the base of the wall as happened on the walkway in front of the temple (AU 87). Four relatively large cranial fragments were clearly from adults. The fragmentary and splintered shafts of all of the major long bones were represented, but the state of preservation did not permit analysis of breakage patterns. All of the postcranial material appeared to be from adults. The long bones consisted of shafts; virtually no long bone ends were present. No minimum number of individuals beyond at least one individual can be assessed.

Several mandibles rested on a stairway that led downward from Platform 4 to Platform 3 (AU 67). The mandibles had perimortem breaks, including adherent flakes and dental ablation (teeth broken off in one case), and portions were burned black, indicating high heat and fresh bone (Fig. S5).

Ten of 25 known middens were excavated, including two associated with Terrace 18, Middens 1 and 7. Nearby is Midden 6, also discussed here. Midden 1 contained over 300 pieces of human bone, mostly fragmentary pieces of the ends of long bones, ribs, hands, feet, and bone splinters. Many of these bones were burned to a dark and blackened state. Most of the bone appeared to be from adults. Cutmarks (10%), spiral fractures (80%), and burning (20%) suggest that these human remains resulted from ritual production of modified bone not found here (long bones and crania). These elements may represent the initial preparation of human remains for display, trophies, or for cooking and consumption.

Midden 7 lay at the base of the terrace, behind the temple and its associated Patio Group B, and adjacent to a cardinal entrance to the site. Wixárika pilgrims returning from treks spend time at such entrance points, feasting with those who receive them back into the community. Midden 7 also was associated with the temple. Because of these associations, Midden 7 materials may have had significance more public than private.

Midden 7 (AU 301) contained long bone fragments, almost all of which (85%) showed spiral fractures and cutmarks (15%). There were also signs of perimortem processing, including chop marks and percussion pits. Cutmarks included V-shaped cuts along the ends of long bones, and on rib and cranial fragments. The distribution and morphology of the cutmarks is consistent with dismembering (bone-from-bone). A right femur had cutmarks and carnivore gnawing on the ends, and it was exceptionally well preserved. There were cutmarks along the shaft, and carnivore gnawing removed the proximal and distal ends. This bone (Fig. S4) appears to have been a trophy bone that was handled over many years, eventually reaching the midden, possibly as part of colluvium from the disintegrating buildings above. The selective human remains found in this midden are very different from those of Midden 1 because these include a more narrow range of elements (mostly long bones). Perhaps this was the final resting place for bones used in ritual or ceremonial activities.

Midden 6 accumulated at the base of a shrine platform situated at the end of a spur off the causeway. We suspect that this shrine was an offering location for people arriving at the western cardinal entrance of La Quemada. Human remains recovered from Midden 6 include 12 fragments of bone representing long bone fragments, as well as cranial, hand, and rib pieces, all too small for analysis. However, most of the bones show spiral fractures and other signs of processing, including adherent flakes and crushing.

Conclusions

In the multiethnic Northern Mesoamerican Frontier, human bone was prominent in living spaces of ceremonial centers. Residents handled and processed human bone extensively. Their actions included killing, disarticulating, dismembering, defleshing, burning, drilling, bundling, suspending, displaying, and discarding. Nonhuman site formation processes were also at work over time, producing taphonomic forces that greatly affected the character and quantity of the human remains. These included carnivore gnawing, weathering, building collapse, water transport, and decomposition by physical and chemical forces present in the sediments and construction materials into which some human bone became incorporated after abandonment. It is important to emphasize that the skeletal remains described here are not typical of the whole population of the Malpaso Valley; occupants made ordinary subfloor inhumations in the surrounding villages (35).

Within the context analyzed here, archaeological and osteological analyses point to six different cultural contexts of use. First, three clusters of human remains in the temple represent all ages and adult males and females. These are primarily whole long bones with shallow cutmarks, suggesting defleshing after some period of postmortem desiccation and decay. This arrangement suggests ancestor veneration. Second, Patio B contained bone fragments with cutmarks, bone splinters, and burning. This area was likely a place for public feasting and anthropophagy. The Patio B bones had pot polishing and other characteristics consistent with cannibalism. Third, the walkway in front of the temple contained processed long bones and drilled crania, which suggests the public display, possibly of defeated enemies, hanging on the outside wall of the temple and on freestanding racks. Fourth, Midden 1 includes a great variety of small bone elements with some burning and processing, which may represent bones not used and discarded after processing. Fifth, Midden 7 represents bones with spiral fractures and cutmarks. This may be bone discard from creating the display along the West Banquette or within the Temple. Finally, rare individual inhumations were placed in tombs.

Two basic pathways of treatment ultimately corresponded to bones kept indoors or outdoors. Bones were kept indoors in the temple and in tombs. In the temple, sets of bone elements were wrapped in bundles and suspended from the ceiling, whereas in the subfloor tombs individuals were inhumed. It is possible that these two kinds of treatment were stages in a single, long-term processing program. Indoor bone had very little perimortem modification other than cutmarks. Subsequently and probably after the site’s abandonment, the bones kept hanging in the temple were gnawed by carnivores.

The residents treated outdoor bone differently. The bones in such deliberately dramatic displays had far more frequent and varied perimortem and postmortem modification. The occupants suspended skulls and long bones from a rack on the western edge of Patio A, from the adjacent front wall of the temple near its entrance, and from a roof above the walkway in Patio B. They also kept a set of mandibles either on the staircase near Platform 3 or in a structure on that platform. Finally, the occupants discarded unwanted bone in middens, apparently after culling it from skeletons that were processed on the terrace.

The above analysis answers a question raised by the initial study of only the remains around the temple (22): were the human bones of Terrace 18 those of ancestors or enemies? The present study concludes, “both.” The residents of La Quemada treated the bones as animated, powerful objects, constituting two basic categories of social persons: ancestors and adversaries.

Ethnohistory attests to similar bone treatment among the ethnically diverse groups inhabiting the Northern Frontier at the time of European contact. Indigenous violence was endemic and included “extended killing,” i.e., disarticulating enemies’ bodies and further treatments. For example, the Caxcan people made trophies and war implements from them (36), the Acaxee consumed their cooked flesh (37), Cumupa residents hung the bones of defeated enemies outdoors (38), and Purépecha leaders passed the bones of defeated enemies among cooperating elite (10). Thus, contextualized regionally and ethnohistorically, the ancient practices documented here form part of a widespread pattern of persistent interethnic violence.

Findings here are also consistent with enemy skeletal displays being symbolically important as mnemonic devices to recall past episodes of violence. Skeletal displays could focus ceremonies and stories on remembering and possibly cauterizing (39) further violence. The historical evidence suggests that the enemies are likely to have been ethnic others, yet evidence from modern zones of persistent conflict shows that ethnic diversity does not predict frequency of conflict (40). To follow up on the conclusion that the violence was interethnic, the obvious next steps are stable isotopic and DNA analyses of skeletal samples from the identified contexts. If the inferences are correct, those analyses should corroborate the contrasts presented here with indications of biological distance.

Materials and Methods

Bones from seven seasons of excavation were analyzed in the field laboratory (Tables S1 and S2, and Datasets S1 and S2). We discuss the archaeological recovery strategies elsewhere (41). Here we report on bones from Terrace 18 and two associated middens. We draw on ethnographic understanding from participation in 13 Wixárika ceremonies since 1995 (30). Bioarchaeological methods conformed to the fragmentary, disarticulated, and commingled deposition. White’s (33) protocol provided a systematic approach to examining taphonomy, cultural modification, and formation processes. The analysis characterizes bone type, preservation, side, and completeness, along with taphonomic properties, particularly peri- and postmortem changes on the bones. It describes the physical evidence (such as cutmarks, breakage, reduction, and burning) as well as depositional context. Each bone is coded for the presence or absence of spiral fractures, internal vault release, scarring, pitting, flaking, peeling, cutmarks, chop marks, polishing, scraping, and burning. Each element was assessed for rodent or carnivore activity and root, water, weathering, and other postmortem damage. Biological indicators of age, sex, health status, robusticity, and trauma were recorded (42).

A total of 417 bone identifiable elements were recovered and analyzed, of which 402 were sufficiently preserved for thorough analysis. Of the remaining 2,015, 840 could be identified or partially analyzed; the rest were too fragmentary. The recovered assemblages are likely representative of their original deposition. Excavators strove for complete bone recovery, sifting all sediments with one-quarter inch (6 mm) screens. In many areas bone attrition was severe. About 75% of the total surface area of Terrace 18 was excavated, including almost all of the roofed architectural area, the ball court in the center of the main patio, and the western half of the main patio. Excavation penetrated the substructure in less than half the excavated area and may have missed substructural chambers. As a result, we cannot be sure of the exact representativeness of the skeletal bone assemblage to everyone who died, but it is clearly representative of those whose remains were intentionally placed inside and outside buildings.