Spanish conquistadors in the sixteenth century left behind detailed accounts of the extensive breeding facilities managed for the Aztec ruler, Moctezuma [1, 2]. In his Segunda Carta de Relación, Hernan Cortés describes two structures dedicated to the import and breeding of diverse carnivorous beasts (wolves, pumas, jaguars, lynxes, foxes, rattlesnakes, etc.) and birds (raptors and other tropical, freshwater, and marine birds) [3]. By this time the Aztecs were successfully capturing and even breeding wild animals in the capital, which hosted three hundred specialists dedicated to their care. This captive population probably fueled the rich offering caches at the ceremonial center in the Templo Mayor (Great Temple), where these very same species were excavated [2, 4]. This paper examines another grand ceremonial center preceding the rise of the Aztec empire to understand when such specialized animal management programs developed in Mesoamerica and to reconstruct what this initial experimentation period would have looked like.

Teotihuacan (A.D. 1–550), one of the largest ancient urban centers in the New World, is located 45 km northeast of modern day Mexico City. This World Heritage site is known for its centralized city-grid layout [5, 6], particularly in the ceremonial precinct where three major pyramids testify to Teotihuacan’s role as an important political and ritual center. Recent tunnel excavations inside the Moon Pyramid [7, 8] and the Sun Pyramid [9, 10] encountered a series of dedicatory offerings that were put in place at various stages of their building sequences. The contents of these offerings were premier expressions of state ideology and militarism, containing symbolically powerful artifacts such as carved obsidian, shell and greenstone objects, human sacrificial victims, and the remains of many carnivores [8, 11].

Animals were an integral and abundant component of these ritual caches at Teotihuacan, occurring in quantities often in accord with cosmologically significant numbers related to the Mesoamerican calendric cycle. With at least 194 animals (Minimum Number of Individuals, MNI) deposited in the offerings at the Moon Pyramid and Sun Pyramid, this assemblage attests to the important role animals played in state-level ritualized activities. Furthermore it provides one of the most prominent examples of mass animal sacrifice in Mesoamerica and is only comparable to the aforementioned Late Post-Classic caches at the Aztec ceremonial center of Tenochtitlan (A.D. 1325–1521) [4, 12].

For the Teotihuacan project we reconstructed complex life-histories for the animals including means of acquisition, management, preparation, use and discard. For this, we recorded body positions, element representations, surface modifications, and performed stable isotope analyses. Both zooarchaeological and isotopic data confirm the presence of a heterogeneous population among the carnivores utilized in state rituals; some were kept in captivity for prolong periods of time, while others were pulled from a wild population. This transformation in human-predator dynamics documents a critical moment when humans began to control animals not just as alimentary sources, but also as highly symbolic beasts for specifically ritual purposes.

Four dedicatory caches from the Moon Pyramid (assigned Entierros 2, 3, 5 and 6) and one in the Sun Pyramid (Ofrenda 2) were studied. The former was explored as part of a seven year extensive excavation program by the Moon Pyramid Project directed by Saburo Sugiyama and Ruben Cabrera [7, 13]. Five burial/offering complexes, designated Entierros 2 through 6, were assigned to one of the seven construction episodes defined by the project [7, 8]. Of these, four caches contained animal remains.

The latter monument was explored as part of the Program of Investigation and Consolidation at the Architectural Complex of the Sun Pyramid, Teotihuacan (Sun Pyramid Project) from 2005 to present by Alejandro Sarabia [9, 10]. Each offering was built in dedication to a new building phase; at the Moon Pyramid these caches were associated with Buildings 4 to 6, and the Sun Pyramid Ofrenda 2 was placed as a foundation to the construction of the main corpus of the monument (Table 1).

A full zooarchaeological study of the collection was completed at the Paleozoology laboratory at the National Autonomous University of Mexico (UNAM) under the guidance of Raúl Valadez and his team. These materials were analyzed with the approval of the project directors of the Moon Pyramid Project (Saburo Sugiyama and Ruben Cabrera) and Sun Pyramid Project (Alejandro G. Sarabia). The zooarchaeological assemblage from the Moon Pyramid and Sun Pyramid dedicatory caches are currently housed at Arizona State University’s Teotihuacan Research Laboratory in San Juan Teotihuacan, Mexico. Species identification, MNI calculation, skeletal element representation and especially surface modifications (pathologies and cutmarks) were recorded following standard zooarchaeological practice [14].

Upon completing the zooarchaeological investigation, stable isotope analysis was implemented to further distinguish wild versus captive populations through paleodietary reconstructions. Carbon stable isotope ratios reflect different plant photosynthetic pathways (C₃, C₄ and CAM), with higher δ¹³C_collagen and δ¹³C_apatite values indicating a larger contribution of C₄ or CAM plants to the diet [15, 16]. While there are several C4 and CAM plants present in arid highland environments of Mexico [17], the main C₄ plant exploited in this region was agricultural maize (Zea mays). Nitrogen stable isotope ratios are strongly influenced by the tropic position of the organism [18, 19], and generally speaking, elevated δ¹⁵N_collagen values suggest a higher trophic position. Thus, monitoring carbon and nitrogen isotopic composition can illuminate degrees of human intervention in the animal’s diet. In the New World, animals raised in captivity, like macaws penned at the pre-Hispanic site of Paquimé in northern Mexico [20] or turkeys in the ancient Southwestern United States [21], were identified isotopically because of their unnaturally high reliance on C₄ based resources. At the ancient Maya site of Colha, isotopic analysis of dog bones showed that increased reliance on C₄ based resources correlated with a decrease in nitrogen values, suggesting lower levels of carnivory [22].

Destructive analysis for the isotopic study was authorized by the INAH (National Institute of Anthropology and History, Mexico), document number 401-3-10136. Most of the samples were destroyed at the time of isotopic analysis or altered into pure collagen or apatite. The left-over fragments continue to be housed at the Paleodiet laboratory, Anthropology Department in the University of California, San Diego. See a full list of specimen numbers and the isotopic results in the Supporting Online Materials (S1 Table).

Only bones of eagles, canids (wolves), felids (jaguars and pumas) and leporidae (rabbits and hares) were included for the isotopic study. When feasible, all species were sampled from each context including both primary (whole) and secondary (prepared faunal artifacts) deposits. In addition samples that capture changes in dietary patterns between early (tooth) and later (bone) periods of the same organism were taken on 43 individuals. The total number of individuals included in the study (n = 83) and the quantity of bones and teeth selected (n = 142) provide a representative sample across both species and burial contexts.

Strict criteria to document diagenesis for both collagen (% collagen yield greater than 1% and C/N ratios between 2.9 and 3.6) [23] and apatite (C/P ratios above 0.1 and IR-SF below 4.0 for bone apatite) [24–26] were set following pertinent literature (S2 Text). Fifty bone apatite, 55 tooth apatite, and 51 bone collagen results are included in the study (S2 Table). Unfortunately this resulted in no bone (apatite and collagen) samples from Entierro 5 and only two bone apatite and one bone collagen samples from Entierro 3. Tooth enamel apatite provides some of the only specimens from these two contexts.

Isotope samples were processed at the Paleodiet Laboratory in the Anthropology Department at the University of California, San Diego and analyzed at the Scripps Institute of Oceanography (SIO) Analytical Facility. Interior and exterior surface and all spongy bone was cleaned with a diamond point Dremel bit or stainless steel engraving bit. Ultrasonic baths in double distilled H₂O (two times) and acetone proceeded this initial surface scraping, at which point most consolidants and glues were no longer visible. If consolidants were still apparent, extra washes in double distilled H₂O and acetone were added. In the case of teeth, ultrasonic baths were carried out first to loosen all surface consolidants on the enamel, and once dried a section of the enamel interior and exterior surface was ablated with the drill. For teeth samples, only enamel apatite was analyzed for the present study. Based on research assessing the effects of B-72 [27] and our own experimental project on Mexican consolidants (Reconos¹¹⁰ and Reconos²²⁰) and conventional Mexican glue (Resistol 850) applied to this collection, it was determined that surface cleaning methods applied were sufficient to rid the samples of their respective consolidants.

Mineral apatite samples were processed using procedures similar to Koch et al. [28]. A small fragment of cleaned and dried bone or tooth was removed and ground into a fine powder using an agate mortar and pestle, and 35mg of this powder was put into micro-centrifuge tubes. All organics were then dissolved by soaking powder in 2% sodium hypochlorite (bleach) (NaClO) solution for 24 hours. Samples were rinsed three times with double-distilled/deionized water, and were then reacted with 0.1M acetic acid (C₂H₄O₂) for 24 hours. These were then rinsed with double-distilled/deionized water three times and dried at 50°C for 24 hours [24]. Two mg of powder were transferred to glass tubes, reacted with phosphoric acid (H₃PO₄), and analyzed on a Gas Bench Thermo MAT 253 connected to a Thermo-Finnigan Delta XP Plus mass spectrometer. An internal CaCO₃ standard calibrated to NBS-18 and NBS-19 was run over the course of six months resulting in a reproducibility of ±0.2σ for δ¹³C. δ¹³C_apatite results are presented relative to the Pee Dee Belemnite (PDB) international standard.

Collagen samples were prepared following protocols described by Schoeninger et al. [29]. Only bone samples underwent collagen extraction, whereby cleaned bones were demineralized with 0.25M hydrochloric acid (HCl) over a course of several weeks at room temperature. The HCl was changed every two days. Upon demineralization, samples were neutralized through five rinses of double-distilled/deionized water. Humic acids were removed through soaking the collagen in 0.125M sodium hydroxide (NaOH) for 24 hours. Subsequently samples were rinsed an additional five times with double distilled/deionized water. Collagen was then solubilized by placing samples in pH3 HCl, heating them to 75°C, and transferring the liquid into Teflon beakers to dry over the course of several days. Samples were then resolibilized with pH3 HCl, frozen, and lyophilized in scintillation vials. Once weighed, collagen samples were analyzed with a Conflow and Costech 4010 Elemental Analyzer coupled to a Thermo-Finnigan Delta XP Plus Mass Spectrometer at the SIO stable isotope analytical facility through a double continuous-flow (CF) IRMS combustion. Peedee belemnite (PDB) marine limestone standard for carbon and atmospheric N₂ (AIR) standard for nitrogen were utilized with a reproducibility of ±0.2σ for δ¹³C and δ¹⁵N.

A couple of data corrections compensate for material offsets as well as to account for the few modern comparative samples included in the study. Warriner and Tuross [30] have demonstrated that enamel apatite tends to be enriched over bone by 2.3‰ in carbon and 1.7‰ in oxygen values. Thus tooth isotopic ratios were corrected to match bone samples in all of the figures presented in the study (note S1 Table presents raw data prior to data corrections). When comparing modern skeletal isotopic ratios to archaeological ones, we must account for depletion in atmospheric ¹³CO₂ in modern versus pre-industrial periods as a result of burning fossil fuels [31]. Thus all modern δ¹³C bone values were adjusted positively by 1.5‰ in the text and figures.

Zooarchaeological and isotopic evidence from Teotihuacan demonstrate the carnivores utilized in state rituals at Teotihuacan were a heterogeneous group: some of the sacrificed individuals were kept in captivity, others were brought in from the wilderness, and some were used to produce ritual paraphernalia. Canid and felid skeletons demonstrate that sacrificial victims were often kept in captivity while faunal products used as ritual paraphernalia originated from a wild population. In contrast, both sacrificed and prepared skeletons of eagles demonstrate some degree of captive management. Compared to the well-established breeding programs recorded at the Aztec capital, the animals found at Teotihuacan probably reflect initial experimentation in manipulating dangerous and specialized carnivores. This process probably involved capturing the animals alive while they were young to be raised within the city confines on small animals, most likely rabbits and hares that ate an exclusively C₄ based diet. In some instances, rabbits and hares may have been supplemented by other omnivores, including dogs and possibly even humans.

Specimens like the eagle, Element 2246, attest to the hardships in learning to manipulate highly specialized carnivores. Keeping a captive population alive must have been difficult, sometimes resulting in accidental deaths of their captive eagles that were then expediently prepared to keep the body intact; these stuffed cadavers would have participated alongside live eagles as if they were also “sacrificed”. Caring for and manipulating the region’s most dangerous apex predators sometimes required the use of brute force as evidenced by an unnaturally high frequency of healed fractures, violent injuries, bone deformity and disease.

The broader implications of these finds are apparent; they not only provide the earliest direct evidence for ritualistic management of carnivores but also demonstrate a fundamental shift in past human-carnivore interactions that was intrinsically tied to the development of one of the most empowered ceremonial landscapes in Mesoamerica. As both the jaguar and Mexican grey wolf are currently on the endangered species lists, the archaeological and isotopic data presented provide a window into understanding human-animal encounters of these endangered species through the long dureé, beginning with the earliest documentation of how these animals were captured and managed. These data also speak to a direct connection between carnivore manipulation and state power. There was an impetus to bring live carnivores into the city as cubs/chicks to be raised within the urban metropolis as sacrificial victims par excellence. It is not a coincidence that such active animal management programs coincided with the development of large monumental structures where animals were embedded into pyramids as key symbols of the Teotihuacan state.