Human-Camelid Interactions in the Wari Hinterlands: A stable isotopic analysis of camelid remains from Auquimarca (600-1000 CE), Huancayo, Peru

doi:10.21203/rs.3.rs-8099994/v1

Download PDF

Research Article

Human-Camelid Interactions in the Wari Hinterlands: A stable isotopic analysis of camelid remains from Auquimarca (600-1000 CE), Huancayo, Peru

https://doi.org/10.21203/rs.3.rs-8099994/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Human-camelid interactions and pastoral practices have helped shape social, political, and economic lifeways throughout the history of the Central Andes. However, the methods by which humans managed and utilized camelids in the central highlands during the Middle Horizon (600-1000 CE)–and how those subsistence strategies reflect negotiations between Andean vertical ecologies, local communities, and the expanding Wari Empire–remain unclear. This study employed a life history approach using sequential stable carbon and oxygen isotope analysis from camelid enamel samples recovered from Auquimarca, a highlands burial site in the Lower Mantaro River Valley, Department of Junín, Peru, to understand the role camelids had in society and the subsistence strategies of people living amid Wari imperial influence in the region. Among the 32 camelids (n=235 enamel carbonate samples) studied, the observed δ13C and δ18O values indicate that the camelids were likely managed using different flexible pastoral strategies, including grazing on naturally-available C3 plants in the puna (3850-4700 masl), receiving maize fodder closer to human settlements in the Mantaro River Valley’s fertile kichwa (3100 – 3850 masl), or a seasonal rotation between these environments to utilize resources available at different points during the year. δ13C and δ18O variation within the lives of certain camelids suggest involvement in long-distance trade networks or camelid caravans, possibly related to broader trade systems in the region and Wari sphere of influence. These results are important for understanding how Wari influence impacted communities in distant and remote parts on the edge of an Andean empire.

Wari Empire

zooarchaeology

stable isotope analysis

camelids

central Andes

Humans have benefitted from their interactions with camelids for millennia in the central Andes. With the domestication of llamas (Lama glama) and alpacas (Vicugna pacos), humans have sourced meat, fat, bones, and wool from camelids (Bonavia, 1996; Weber, 2019), used them as pack animals (Capriles and Tripcevich 2016), and integrated them into sacred rituals, among other activities. Camelid pastoralism is an effective way to produce and manage these animal products within the Andean vertical archipelago (Murra 1972): where communities utilize distinct Andean ecozones that vary drastically in elevation and agricultural potential to produce resources that are moved throughout the Andes (Browman 1970, 1978; Nielsen 2000; Dufour and Goepfert 2020). These human-camelid interactions have therefore shaped the political economy, migration patterns of people, and fostered socio-political interactions between different people and communities (ibid).

Beyond their economic purpose, camelids are embedded in the human social system and imbued with their own cultural and social meaning in these pastoral societies (Flannery et al. 1989; Dransart 2002; Duche-Pérez and Mamani-Daza 2024). The human-animal interactions fostered through camelid pastoralism generate a powerful ideology on the real and symbolic importance of these animals and their connection to human sustenance and well-being. But decisions about the care and maintenance of camelid herds, their movement between ecozones to access additional sources of food and water, best subsistence strategies (i.e., foraging, agriculture, or some hybrid of the two) to supplement pastoral practices, and the use of camelid herds in ritual contexts are also informed by socio-political systems (Browman, 1974; Flannery et al., 1989; Takigami et al., 2020). These systems range from community forces that negotiate pasture use to imperial states that require forms of labor, ritual, or material tribute (ibid).

In this study, we combine zooarchaeological and stable isotope analysis (δ¹³C, δ¹⁸O) from camelid enamel carbonates to analyze human-camelid relationships and camelid management decisions among people from a fertile highland region in the Andes. They were recovered from the archaeological site of Auquimarca, located in the contemporary province of Huancayo in Peru. Given archaeological evidence of Wari imperial influence at the site, we assess whether and how larger imperial structures of the Wari Empire influenced those human-animal practices.

1.1 Cultural Influence of the Wari Empire

The Middle Horizon (600–1100 CE) was the first time in Andean prehistory when an imperial polity emerged, as exemplified by the Wari Empire, whose hegemony spanned regions in the coastal and highland Peruvian Andes (Fig. 1) (Schreiber 1992; Isbell 1997; Tung 2012). The heartland of the Wari Empire was at the site of Huari, in the Ayacucho Basin (Isbell 1997). Out of the heartland, the Wari Empire created a mosaic of control across much of the Peruvian Andes (Schreiber 1992), establishing administrative centers that helped to support the widespread adoption and imposition of Wari symbols and practices outside of the heartland, such as D-shaped temples, accessible subterranean tombs, consumption of chica de molle, and slip painted faceneck vessels and tumblers (Isbell and McEwan 1991; Schreiber 1992; Earle and Jennings 2012; Stone 2012; Williams and Nash 2021; Grávalos et al. 2023). The distribution of these administrative centers helps to reveal the spatial patterning and varied level of Wari influence in different Andean regions; hinterland sites adopted and modified Wari architectural and artistic features, adjusted subsistence strategies, and integrated Wari and local ritual-religious practices to fit within their existing lifestyles and cultures (Grávalos et al. 2023).

1.1.1 Wari influence in the Mantaro River Valley

While much is known about Wari influence in the Ayacucho Basin and in parts of the southern and northern hinterlands, Wari presence in the central Peruvian highlands is not as clear (Schreiber 1992; McEwan 1996; Williams 2001; Tung 2012; Watanabe 2012; Giersz 2016). The Mantaro River Valley (MRV) is in this region, and archaeological studies in the Upper Mantaro Valley, the northern section of the MRV, have reported little evidence of Wari influence (D’Altroy and Hastorf 2001). In the Lower MRV, evidence of Wari influence is clearer. Stone architecture and tall, enclosed spaces, subterranean tombs, and Wari-style ceramics have been found at several sites in this region, including at Calpish, Ñahuiraupukio, and Wari Willka (Browman 1970; Jennings and Craig 2001; Earle and Jennings 2012). No large Wari administrative centers have been identified here; however, Wari Willka has been classified as an important ritual temple or pilgrimage site in the Wari Empire (Browman 1970; D’Altroy and Hastorf 2001). The site of Auquimarca, analyzed here, is located around 3.7 kilometers away from Wari Willka in the Lower MRV.

2.1 Ecology and Agriculture of the Mantaro River Valley

The site of Auquimarca is in the city of Huancayo (department of Junín), located in the Mantaro River Valley (MRV) at 3220 masl (Fig. 2). The MRV ranges in altitude from 3100 masl to 4500 masl, with mountain peaks up to 5500 masl flanking the sides (Heikkinen 2021). Compared to other river valleys in the highland Andes, the MRV is relatively wide–around 2 km at its narrowest and 24 km at its widest–and very fertile (Browman 1970; Mayer 1979).

Today, agricultural fields on the Mantaro River plain and lower mountain slopes (between 3100–3850 masl) produce crops such as potatoes, corn, beans, barley, and ulluku which supply many nearby cities (Parsons et al. 2000; Heikkinen 2021; Aramburú Venegas 2022). Maize agriculture is now especially prevalent in the southern stretch of the MRV, where Auquimarca is located, because this region tends to be warmer than areas to the north (Mayer 1979). This kind of agricultural zone at this elevation is known as the kichwa zone (Flannery et al. 1989).

Above the kichwa on higher slopes of the MRV is the puna ecological zone (between 3850–4700 masl): an alpine meadow region with average temperatures between 0–7°C (Browman 1974; Parsons et al. 2000; Vidal 2014; Takigami et al. 2020). The lower puna (3850–4300 masl) is suitable for the growth of some tubers and is also where chunõ (freeze-dried potatoes) is often produced (Mayer 1979; Parsons et al. 2000). The upper puna (4300–4700 masl), however, is not suitable for agriculture; rather, this is the primary location of camelid grazing during summer months for modern pastoralists in the area (Flannery et al. 1989; Parsons et al. 2000). In contemporary times, farmers in the MRV commonly live in the larger towns and cities of the kichwa, and herders live in much smaller settlements in the puna. The two groups trade maize, grains, tubers, wool, dried meat, chunõ, and other products (ibid).

Finally, above the puna is what many Quechua-speaking people refer to as the urqu (over 4700 masl): snowy, glacial mountain peaks with few-to-no residents (Flannery et al. 1989; Takigami et al. 2020). In the MRV, this specifically refers to the Huaytapallana mountain range on the eastern side of the valley, with peaks up to 5500 masl (Mark et al. 2017; Heikkinen 2021).

2.2 The Site of Auquimarca

The site of Auquimarca (Fig. 3) was uncovered beneath a road during a water and drainage system infrastructure project and was excavated in 2021 and 2022 as part of a rescue excavation directed by the Peruvian Ministry of Culture (Aramburú Venegas 2022). Given its proximity to the city center of Huancayo, Auquimarca is directly surrounded by modern residential and commercial infrastructure, so the excavation team was limited in their ability to excavate the entire site. Of the six hectares they were able to excavate, much of it was badly damaged by the modern construction project. That project inadvertently uncovered tombs with human and animal remains and associated artifacts. Previous infrastructure projects, shown by water pipes bisecting some of the tombs, also damaged the burials in the years prior (ibid).

2.2.1 Human interments at Auquimarca

In total, one hundred and twenty-six (126) tombs and their surrounding artifacts and remains were excavated from Auquimarca. However, it is estimated that another twenty tombs remain under the surrounding, contemporary buildings. Archaeologists (led by Palacios) identified at least 96 individuals at Auquimarca. Further analyses by Krause and colleagues examined these human individuals to document demographic and health profiles; they also selected human samples for a dietary isotope study (Krause 2025). Food offerings, slip-painted ceramic vessels, tupu pins, and spondylus shell beads, among other artifacts, were found within the tombs (Aramburú Venegas 2022). Accelerated mass spectrometry (AMS) dates (N = 6) obtained from the bones of six human individuals at Auquimarca range from 600 to 1000 CE, suggesting that this is a Wari-era Middle Horizon site (Krause and Tung 2025). The style of the tombs and artifacts recovered from Auquimarca suggest a merging of Wari artistic styles with more local ones during the Middle Horizon. The site’s burial structures resemble those found at Conchopata, Muyo Orqo, and Aqo Wayqo in the Ayacucho Valley (Isbell 2004). Yet, given Auquimarca’s proximity to the Wari heartland, these similarities may represent wider regional traditions rather than direct imperial control. The ceramics associated with tombs are predominantly local, with only a small proportion displaying Wari-style features (Maldonado et al. 2022). As most Wari-style ceramics across the Andes were locally produced (Williams et al. 2019; Grávalos et al. 2023), it is more accurate to distinguish between local and Wari-associated forms rather than assume imported goods or centralized production. No domestic structures were found, indicating that Auquimarca was a mortuary site.

2.2.2 Middle Horizon sediment layer at Auquimarca

A 20–80 cm thick layer of sediment was found at an activity level directly above the tombs and contained faunal remains, tools, ceramic fragments, and other artifacts. Many stone hoe heads associated with agricultural work were found in this layer. Artifacts associated with textile production (stone awls and bobbins), butchering (stone scrapers), and hunting (stone points) were also uncovered at the site, but in much fewer quantities.

2.2.3 Faunal remains recovered at Auquimarca

The faunal remains at Auquimarca were found in several contexts: 1) as offerings within tombs, 2) in association with specific tombs, but not in the tomb chamber, 3) in the sediment layer directly above the tombs at activity level, and 4) in a final layer of sediment closest to the surface. In total, seven thousand and ninety-nine (7,099) fragments from animal remains were recovered from excavations and analyzed in the field (Alaica and González 2022). Eight species of mammals and three species of birds were identified, including camelids, guinea pigs, dogs, geese, and deer. A small number of cow, pig, and horse bone fragments were identified, all post-colonial domesticates that indicate some modern contamination in the top layer of sediment near the surface (ibid). Only indigenous animal species were found at the Middle Horizon activity level (ibid).

2.2.4 Zooarchaeological Analysis of Camelid Remains

Zooarchaeological analysis of the faunal assemblage from Auquimarca recorded 7098 specimens of fragmented vertebrate remains (Table 2). Camelid remains (Lama sp./Vicugna sp.) (Number of Identified Specimens-NISP = 40.3%, Weight = 66.7%) were most common by count (NISP) and weight. Numerous guinea pig remains (Cavia porcellus) were recovered (NISP = 12.9%, Weight = 0.3%) with several found in tomb contexts on plates buried alongside the interred human burials. While several dozen fragments of Eurasian domesticates were identified at Auquimarca (Bos taurus, Sus scrofa, and Equus sp. NISP = 0.6%, Weight = 2.2%), these specimens were recovered from disturbed areas of the upper fill layer that attests to contemporary activities.

Table 2
Composition of faunal assemblage recovered from Auquimarca
	Total
Taxon	NISP	%NISP	Weight (g)	%Weight
Anas sp.	1	0.0	0.5	0.00
Anser sp.	8	0.1	5.7	0.01
Characriiformes	1	0.0	0.8	0.002
Psittaciformes	1	0.0	5.2	0.01
Ave pequeño	6	0.1	1.6	0.003
Ave mediana	6	0.1	4.7	0.01
Phyllotis sp.	4	0.1	0.2	0.0005
Cavia porcellus	918	12.9	130.2	0.03
Muridae	21	0.3	0.7	0.001
Canis sp.	1	0.0	6.4	0.01
Canis familiaris	167	2.4	378.0	0.8
Odocoileus virginianus/ Hippocamelus bisulcus	64	0.9	698.3	1.5
Lama sp./Vicugna sp.	3862	40.3	31717.9	66.7
Sus scrofa	9	0.1	200.6	0.4
Bos taurus	38	0.5	797.1	1.7
Artiodáctilo	1395	19.7	7415.9	15.6
Equus sp.	1	0.0	29.2	0.1
Simiiformes	2	0.0	5.8	0.01
Homo sapiens	247	3.5	790.2	1.7
Mamifero micro	36	0.5	1.2	0.002
Mamifero pequeño	28	0.4	35.7	0.1
Mamifero mediana	605	8.5	2626.8	5.5
Mamifero grande	673	9.5	2713.5	5.7
Mamifero	4	0.1	1.0	0.002
Total	7098	100.0	47567.3	100.0

Among the camelids, the thorax region (including the vertebrae and ribs) was the most abundant portion of the skeletons found at Auquimarca (NISP = 30.9%, Weight = 24.5%). The upper portion of the front limb (including the humerus) was also prevalent in the assemblage (NISP = 17.8, Weight = 25.6%) (Fig. 4). Both portions of the skeleton have large quantities of meat that attest to the butchery of camelids for their use in communal gatherings and possible feasts.

Finally, comparing the tomb contexts with the upper fill layer, reveals interesting pattern in the use of camelids at Auquimarca for their prevalence in the upper fill layers relative to their abundance in tombs. Guinea pigs are much more common in tombs contexts when compared with camelids, with 75% of the assemblage in these contexts consisting of guinea pigs and 25% of camelids. Furthermore, camelid specimens recovered from tomb contexts (Fig. 5) are more likely to be unfused, while in the upper fill layers, camelid remains were more often identified as fused. These patterns indicate that young camelids may have been preferred as offerings in burial contexts, while more mature animals were shared in communal events and feasts.

2.3 Camelid Pastoralism in the Mantaro River Valley

Semi-nomadic camelid pastoralism–where communities relied heavily on pastoralism in the puna, but still utilized some level of horticulture at a lower elevation at different points in the year–was the primary source of subsistence for communities in the MRV from the Formative Period (BCE 1800 − 900) through the Early Intermediate Period (1-550/600 CE) (Browman 1976; Matos Mendieta 1978). Small-scale settlements at this time were scattered throughout the kichwa and the puna, many of which contained evidence of canchas (corrals for the camelids) and middens containing camelid and plant remains (Matos Mendieta 1978).

Browman (1970, 1974, 1976) suggested a shift in MRV sites from being reliant on semi-nomadic camelid pastoralism to utilizing more agriculture in the kichwa with the advent of the Middle Horizon. This transition toward sedentary lifestyles may reflect the Wari Empire’s strategy of exploiting the region’s fertile lands for agricultural use (Matos Mendieta 1978); however others have suggested that natural population growth may have been pushing the populations towards an increased reliance on agriculture regardless of Wari influence (Browman 1976). Using stable isotope analysis to document the diets and migration patterns of camelids at the MRV site of Auquimarca provides new empirical data to explore other aspects of subsistence strategies among a community in this region. In turn, these zooarchaeological isotope data can help illuminate food management and exploitation and factors that may have contributed to those subsistence strategy decisions.

Stable isotope analysis is used within zooarchaeology to understand the role animals had in past human civilizations and environments (Pilaar Birch 2013). Here, we compare the ratio of ¹³C/¹²C and ¹⁸O/¹⁶O found in camelid enamel carbonate samples to a standard (Vienna Pee Dee Belemnite, VPDB) to produce a value expressed as δ¹³C or δ¹⁸O parts per mil (‰) used in stable isotopic comparisons and critical analyses (Coplen et al. 1983; Coplen 1996).

3.1 Carbon Stable Isotope Analysis

Landscapes of the central Andes are home to a diversity of plant species, leading to a range of δ¹³C values in the environment and in the tissues of organisms that consume these plants (Powell and Still 2009). C₄ plants (Hatch-Slack photosynthetic pathway) typically grow naturally or as crops in coastal, lowland, or kichwa regions, as these plants thrive in drier, hotter areas while C₃ plants (Calvin-Benson photosynthetic pathway) grow at all elevations throughout the Andean ecological zones (Cavagnaro 1988; Thornton et al. 2011). These photosynthetic pathways fixate carbon differently, leading to distinct δ¹³C signatures between C₃ and C₄ plants, and some variation between species within these plant groups (Smith and Epstein 1971).

From ethnographic accounts, C₄ plants that are often available to camelids include amaranth (Amaranthus caudatus), saltgrass (Distichlis spicata), matted grama (Bouteloua simplexi), and muhly (Muhlenbergia sp.) (Cadwallader et al. 2012; Dufour et al. 2014; Melton et al. 2023). Wild Andean C₄ plants have an average δ¹³C of – 13.5 ± 1‰ (Szpak et al. 2013). Maize, however, is widely accepted as a primary source of C₄ plants in the diets of camelids in the past, and in this region, has an average δ¹³C of – 11.8 ± 0.4‰ (ibid). C₃ plants available to camelids in the central Andes that are often available to camelids include cultivated quinoa (Chenopodium quinoa), potatoes (Solanum sp.), and a variety of grass species (Festuca sp., Calamogrostis sp., Stipa sp., and Poa sp.) (Flannery et al. 1989; Finucane 2009; Cadwallader et al. 2012; Melton et al. 2023). Andean C₃ plants range in δ¹³C from – 31.9‰ to – 22.5‰ (Szpak et al. 2013).

Higher altitude puna environments, where camelid grazing tends to happen, are composed primarily of C₃ grasses, and therefore have lower average δ¹³C values when compared to the more C₄-heavy kichwa (Thornton et al. 2011; Dufour et al. 2018). The difference between the δ¹³C of the consumed plants and that of the camelid enamel is 14‰ because of fractionation that occurs as bioapatite carbonate is created in the enamel (Cerling and Harris 1999; Passey and Cerling 2002; Dufour et al. 2014). Therefore, animals with pure C₃ diets tend to have a δ¹³C_{enamel carb} of – 12‰ or less, while those with pure C₄ diets tend to have a δ¹³C_{enamel carb} of + 2.5‰ or greater (Dufour et al. 2014).

3.2 Oxygen Stable Isotope Analysis

Oxygen stable isotope analysis can be used to understand the movement of organisms across the Andean landscape (Katzenberg 2008). As organisms consume water derived from different sources (i.e., rivers, lakes, glaciers, and plants), the unique δ¹⁸O values of those water sources are incorporated into their tissues (Bryant et al. 1996; Barbour 2007). Differences in δ¹⁸O values between water sources are impacted by evaporation, proximity to the coast, differences in altitude, temperature, latitude, seasonality, and other environmental factors (Knudson 2009; Pederzani and Britton 2019; Zimmer-Dauphinee et al. 2020). As such, δ¹⁸O values generally get lower with higher altitudes, colder temperatures, and further distance from the coast (Yann et al. 2016; Pederzani and Britton 2019). The vast diversity of altitudes among the Andean ecological zones, including those in the MRV, make the Andes an opportune location to utilize changes in δ¹⁸O values to study the movement of organisms in the past (Knudson 2009).

Unlike humans who drink most of their water, weaned camelids in the Andes–as evaporative-sensitive taxa–receive much of their water from the vegetation they consume, tying their remains to Andean ecological zones rather than just surface water sources (Yann et al. 2016; Miranda-de la Lama and Villarroel 2023). Plants derive most of their oxygen from soil water, so the δ¹⁸O of precipitation–the major source of soil water–highly impacts the δ¹⁸O of the weaned camelids who consume those plants (Flanagan et al. 1991). Additionally, changes in the amount of precipitation during wet versus dry seasons leads to differences in δ¹⁸O within a plant; in general, high evaporation and evapotranspiration rates during periods of low rainfall lead to δ¹⁸O enrichment, and low evaporation and evapotranspiration rates during periods of high rainfall lead to more depleted δ¹⁸O values in vegetation (ibid; Yann et al., 2016). As such, δ¹⁸O values of weaned camelids reflect primarily the season, the vegetative environment in which they were grazed in, and the environment in which their fodder was grown in (ibid; Dufour et al., 2014).

3.3 Stable Isotope Analysis of Middle Horizon Camelids

Several studies have used stable isotope analysis to investigate camelid pastoralism in the Central Andes during the Middle Horizon (see Fig. 1 for site locations). At Conchopata in the Wari heartland, Finucane et al. (2006) observed two distinct groups of camelids: a ¹³C-enriched group, with some of the highest δ¹³C values of camelids recorded in the Andes, and a ¹³C-depleted group. From this, they suggest that two different kinds of camelid pastoralism were practiced in the Wari heartlands: rangestock grazing, where camelids (here, alpacas) were brought to the puna to graze on primarily C₃ plants, and maize foddering, where camelids (here, llamas) were fed the remains of maize crops in canchas or brought to the maize fields to feed on the stubble (ibid); differences between the groups on the bases of species have been contested (Tomczyk et al. 2019). Regardless, a dimorphic husbandry strategy was observed and the high presence of maize within a subsection of the Conchopata camelids’ diets supports the importance of maize within the Wari heartland.

In the Wari hinterlands, other methods of camelid pastoralism have been observed. At Castillo de Huarmey, a Wari administrative center on the northern Peruvian coast, some camelids grazed on almost exclusively on C₃ plants, while others were provided a greater mix of C₃ and C₄ plants and moved between higher and lower pastures in the coastal region (Tomczyk et al. 2019). In this dimorphic husbandry model, most camelids were kept within proximity to the site, but some stayed in similar pastures at higher elevations while others moved between pastures at higher elevation and closer to the coast. However, stable strontium and lead isotope analysis revealed several non-local individuals who were likely part of trade networks connecting the Wari Empire (ibid).

Multiple husbandry strategies were observed at Cerro Baúl in the Moquegua River Valley as well (Thornton et al. 2011; deFrance 2014). The low average δ¹³C values among the majority of the sampled camelids were consistent with the δ¹³C values of modern camelids rangestock grazed on C₃ plants in the puna; however, three of the camelids had elevated average δ¹³C values relative to the others, who Thornton et al. (2011) suggest were llamas that were kept close to human settlements, fed maize instead of grazed, and were likely used in caravan trade networks.

On the southern coast, δ¹³C values from various camelid tissues at Uraca, Beringa, and Quilcapampa revealed differences in the food made available to camelids at these sites (Alaica et al. 2022; Melton et al. 2023). At Quilcapampa, differences were also detected in fodder strategies between llama-sized and alpaca-sized camelids based on first phalanx morphometrics in which llamas were likely moving between the highlands and coast more regularly due to their higher δ¹³C isotopic compositions (Alaica et al. 2021; Le Neün et al. 2023). Camelids from Early Intermediate Period-Middle Horizon contexts at Uraca may have had greater access to C₄ plants than camelids from Middle Horizon contexts at Beringa and Quilcapampa, likely because of close proximity to agricultural lands and greater autonomy over their foddering practices. Beringa and Quilcapampa had greater demands from the Wari heartland that required greater flexibility of camelid diets, resulting in a more mixed C₃ and C₄ diet (Alaica et al. 2022). Further analysis of δ¹³C_{enamel carb}, δ¹⁸O_{enamel carb}, and dental calculus microbotanicals from several Quilcapampa camelids suggests a variable incorporation of C₃ (likely potatoes) and C₄ (likely maize) plants; mobility appears to increase throughout life, consistent with camelids participating in caravans after weaning (Melton et al. 2023).

In comparing δ¹³C values from camelid bone collagen between coastal and highland central Andean sites and between time periods, Noe et al. (Noe et al. 2024) found differences in the subsistence strategies and acceptance of maize agriculture between coastal and highland sites during the Middle Horizon, with a broader range of C₄-enriched δ¹³C values observed among coastal sites compared to highland sites. Middle Horizon highland sites did have more outliers than coastal sites, which they argued meant camelids in the highlands were more commonly a part of camelid trade routes connecting portions of the Wari Empire (ibid). These differences in camelid management strategies–and the importance of maize for these strategies–seen across the Wari Empire suggest that ecological location, agricultural suitability, and connectedness to the Wari heartlands informed human-camelid relationships in Middle Horizon communities. As Auquimarca is one of very few Middle Horizon sites in the fertile MRV, the novel analysis of the δ¹³C and δ¹⁸O in camelid enamel carbonates from this site offers a unique perspective on the role camelids had in society and the subsistence strategies of people living amid Wari imperial influence in this region.

Dental elements, representing thirty-two (32) individual camelids, were exported from Auquimarca to the US with the permission of the Peruvian Ministry of Culture. Seven of the individuals have multiple teeth included in this study, for a total of 44 camelid teeth represented in the dataset (Table 1). The teeth chosen for stable isotope analysis were deciduous third premolars (dP3; n = 3), deciduous fourth premolars (dP4; n = 4), permanent first molars (M1, n = 14), permanent second molars (M2; 15/44), and permanent third molars (M3; n = 7). Comparing the stable isotope ratios from these different teeth provides an overview of any changes in diet and/or environment that occurred during an organism’s early life (Dufour et al. 2014). The stable isotopes of dP3 and dP4 represent the diet and environment of the mother, since these teeth mineralize in-utero (Takigami et al. 2020). Stable isotopes from M1, M2, and M3 represent later diets and environments, erupting around 6 to 9 months, 1 year 5 months to 2 years, and 2 years 9 months to 3 years 8 months, respectively (Wheeler 1982). One tooth (AM-143) was classified as an M1 or M2 because of its ambiguous traits and was not included in any analyses of tooth-based differences. In total, 235 enamel carbonate samples were prepared from the 32 camelid individuals.

The Auquimarca camelid enamel carbonate samples were prepared for stable carbon and oxygen isotope analysis in the Bioarchaeology and Stable Isotope Research Lab (BSIRL) at Vanderbilt University. Sample preparation followed a modified version of the bleach-based protocol outlined in Garvie-Lok et al. (2004). The outer enamel surface of the teeth were cleaned with sterile brushes and 18 meg-ohm ultrapure water (D2). All dental samples were left to air dry overnight before they were lightly abraded with a Dremel rotary tool to remove surface contaminants. Gloves were changed and laboratory surfaces and Dremel tools and bits were cleaned with bleach, water, and ethanol between each sample to avoid potential cross-contamination.

The enamel carbonate samples were collected sequentially from the teeth to capture more precise changes in an individual’s diet and environment throughout life (Passey and Cerling 2002; Dufour et al. 2014; Takigami et al. 2020; Alaica et al. 2022; Pilaar Birch et al. 2025). Using a Dremel tool, horizontal lines were drilled across one side of the tooth starting at the crown of the tooth (occlusal surface, OCC) moving towards the tooth root (cementoenamel junction, CEJ). Since enamel mineralizes from the OCC to the CEJ and does not turnover, the stable isotope ratios of camelid enamel carbonate samples closer to the OCC represent an individual’s earlier diet and environment than those of samples closer to the CEJ, which represent their present circumstances (Passey and Cerling 2002; Takigami et al. 2020; Alaica et al. 2022). Lines were drilled approximately 1 mm apart, to ensure definitive sampling. Amount of wear varied between teeth making the number of samples taken per tooth variable as well. An average of 5.5 sequential samples were taken from each tooth. When sequential sampling was not possible (due to size of the tooth, chipping, or poor preservation), a bulk sample was taken by drilling a vertical patch on the tooth from the OCC to the CEJ (Fig. 6).

For every sequential or bulk sample, 2 mg of enamel powder was collected and added to a 1.5 ml plastic vial. One ml of 1-1.5% NaOCl was added to each vial for 48 hours to remove organic contaminants. Samples were rinsed with D2 and 1 ml of 0.1 normality (N) CH₃COOH (acetic acid) was added for 3.5 hours to remove exogenous carbonates. Samples were rinsed with D2. One ml of CH₃OH (methanol) was added to each vial, samples centrifuged, and the methanol decanted. Samples were then left to air dry in a desiccator for 48 hours.

Samples were analyzed with a Kiel IV Carbonate Device coupled to a Thermo DeltaPlus XP IRMS at the Yale Analytical Stable Isotope Center (YASIC). All presented δ¹³C and δ¹⁸O values are reported on the VPDB scale and were calibrated on a two-point curve using reference materials PX (δ¹³C = + 2.01‰ VPDB; δ¹⁸O = – 1.91‰ VPDB)) and MERC (δ¹³C = – 48.96‰ VPDB; δ¹⁸O = – 16.48‰ VPDB). An internal standard (COW) was used to correct for drift (δ¹³C = – 7.85‰ VPDB; δ¹⁸O = – 4.47‰ VPDB).

Data were analyzed using Microsoft Excel and R version 4.3.2 (R Core Team 2023) with the psych (Revelle 2025), FSA (Ogle et al. 2025), and readr (Hester and Bryan 2024) packages. These same programs, along with the tidyverse (Wickham et al. 2023) and RColorBrewer (Neuwirth 2022) R packages, were used to develop scatterplots, line graphs, and boxplots to visualize the data. All statistical analyses were performed at an alpha level of 0.05 for two-tailed tests. The dataset was not normally distributed (Shapiro-Wilks test: δ¹³C p < 0.001, δ¹⁸O p = 0.13, n = 235), therefore we used non-parametric tests to analyze variation within the site and across other Wari-affiliated sites. Specifically, a Kruskal-Wallis test and subsequent Dunn’s post-hoc test were used to determine statistical differences between individual camelids and Mann-Whitney U tests were used to analyze differences within and between teeth and between Auquimarca and other sites. Mann-Whitney U tests were also used to assess differences between camelids recovered from tombs or in the superimposed sediment layer at Auquimarca.

The approximate percentage of C₄ plants in each individual camelid’s diet at Auquimarca and at other Wari-affiliated sites was calculated using δ¹³C_{enamel carb} and δ¹³C_collagen estimations for individuals consuming 100% C₃ diets (δ¹³C_{3 enamel carb} = – 12‰, δ¹³C_{3 collagen} = – 21‰) and 100% C₄ diets (δC_{4 enamel carb} = + 2.5‰, δC_{4 collagen} = – 6.5‰) from Dufour et al. (2014). We acknowledge that δ¹³C values vary within C₃ (~ 9.4‰) and C₄ (~ 4.6‰) plants across the Andean landscape (Szpak et al. 2013). However, the %C₄ plants in diet calculation from Dufour et al. (2014) enables us to have a standardized, general idea of varying plant consumption within the Auquimarca camelid dataset and between other Wari-affiliated sites. Few studies (Tomczyk et al. 2019; Alaica et al. 2022; Melton et al. 2023) have analyzed isotopic values from Wari-affiliated camelid enamel carbonates, so this method allows us to generally compare our dataset to a broader Wari scholarship that performed stable isotopic analyses on different materials (e.g., collagen and keratin).

The results of δ¹³C and δ¹⁸O values for the 235 camelid enamel samples included in this study are plotted in Fig. 7, with summary statistics for the data shown in Table 3 (original values found in Online Resource 1).

Table 3
Summary statistics of δ¹³C_{carb VPDB} and δ¹⁸O_{carb VPDB} for camelid enamel carbonate samples (n = 235) from Auquimarca.
	δ¹³C_{carb VPDB} (n = 235)	δ¹⁸O_{carb VPDB} (n = 235)
Mean (‰)	-8.35	-6.50
Median (‰)	-8.93	-6.80
Range (‰)	-15.10 to -0.94	-12.24 to 0.66
Standard Deviation (‰)	2.72	2.20

5.1 δ¹³C and δ¹⁸O values by individual camelid

The mean δ¹³C value of each individual camelid ranges from – 1.26‰ to – 12.32‰, while the mean δ¹⁸O value of each individual ranges from – 1.71‰ to – 9.12‰ (Table 4). The individual camelids consumed on average 28.3% C₄ plants in their diets, with predicted values ranging from – 2.2% C₄ plants in the diet (exceeding predicted pure C₃ plant consumption δ¹³C values from Dufour et al. (2014)) to 74.06% C₄ plants in the diet. 9/32 individuals had < 15% predicted C₄ plants in their diet; 9/32 individuals had between 15–30% predicted C₄ plants in their diet; and 14/32 individuals had > 30% predicted C₄ plants in their diet. Figure 8 and Fig. 9 show the range of δ¹³C and δ¹⁸O values of samples associated with each individual, respectively. There were no significant differences in average δ¹³C or δ¹⁸O between the camelids recovered within tombs at Auquimarca or in the sediment layer on top of the tombs.

Table 4
Summary statistics of δ¹³C _{carb VPDB} and δ¹⁸O _{carb VPDB} by individual camelid (n = 32). %C₄ value based on equation in Dufour et al. .
	δ¹³C_{carb VPDB} (‰)				δ¹⁸O_{carb VPDB} (‰)
	Mean (SD)	Median	Range	%C₄	Mean (SD)	Median	Range
AM-122	-1.26	-	-	74.1	-7.07	-	-
AM-123	-12.11 (1.94)	-11.33	-15.10 to -9.79	-0.7	-8.31 (1.39)	-8.20	-10.77 to -5.43
AM-124	-4.74 (1.73)	-4.87	-6.74 to -1.93	50.1	-4.33 (0.77)	-4.39	-5.33 to -3.08
AM-125	-7.38 (2.52)	-8.10	-9.36 to -2.03	31.9	-7.87 (2.06)	-7.79	-10.62 to -5.03
AM-126	-9.55 (0.36)	-9.47	-10.25 to -9.19	16.9	-6.68 (0.82)	-6.35	-7.90 to -5.84
AM-127	-10.87 (0.38)	-10.82	-11.34 to -10.50	7.8	-7.55 (1.10)	-7.33	-9.02 to -6.52
AM-128	-7.26 (0.75)	-6.99	-8.11 to -6.69	32.7	-7.10 (0.36)	-6.93	-7.51 to -6.85
AM-129	-5.01 (1.43)	-5.30	-6.61 to -2.72	48.2	-6.01 (1.34)	-6.24	-7.33 to -4.49
AM-130	-8.46 (0.18)	-8.46	-8.75 to -8.27	24.4	-7.38 (0.33)	-7.43	-7.75 to -6.88
AM-131	-10.08 (0.73)	-10.36	-10.89 to -8.89	13.2	-3.58 (1.24)	-3.64	-5.17 to -1.50
AM-132	-5.65 (0.44)	-5.50	-6.15 to -5.30	43.8	-5.66 (0.30)	-5.68	-5.96 to -5.36
AM-133	-10.58 (0.58)	-10.72	-11.09 to -9.78	9.8	-7.04 (1.00)	-6.86	-8.32 to -6.12
AM-134	-10.25 (0.55)	-10.21	-11.27 to -9.61	12.0	-5.85 (1.41)	-4.97	-8.45 to -4.77
AM-135	-5.34 (0.85)	-5.40	-6.27 to -4.27	46.0	-4.07 (1.45)	-4.00	-5.84 to -2.43
AM-136	-7.18 (0.63)	-7.04	-7.94 to -6.39	33.3	-4.76 (0.87)	-4.82	-5.86 to -3.43
AM-137	-5.08 (0.78)	-5.13	-5.94 to -4.03	47.7	-5.06 (0.83)	-4.67	-6.49 to -4.43
AM-138	-8.69 (1.93)	-8.93	-11.31 to -4.93	22.8	-8.87 (2.14)	-8.41	-12.24 to -7.06
AM-139	-6.81 (2.37)	-5.77	-9.48 to -3.80	35.8	-7.52 (2.49)	-9.30	-10.11 to -3.61
AM-140	-9.25 (0.55)	-9.33	-10.05 to -8.18	18.9	-9.12 (1.42)	-8.80	-12.18 to -7.39
AM-141	-9.55 (1.09)	-9.23	-11.73 to -7.86	16.9	-7.98 (0.73)	-8.32	-9.10 to -6.74
AM-142	-2.80 (0.16)	-2.77	-3.07 to -2.67	63.4	-6.34 (0.95)	-6.80	-7.36 to -5.14
AM-143	-8.50 (1.79)	-8.63	-12.80 to -4.81	24.1	-8.67 (0.62)	-8.68	-10.03 to -7.81
AM-144	-12.32 (0.62)	-12.37	-13.07 to -11.52	-2.2	-5.44 (0.60)	-5.44	-6.14 to -4.54
AM-145	-7.76 (0.18)	-7.75	-7.80 to -7.56	29.3	-3.97 (0.71)	-3.78	-5.60 to -3.41
AM-146	-3.22 (1.79)	-2.44	-6.15 to -0.94	60.6	-1.71 (1.86)	-1.70	-4.75 to 0.66
AM-147	-6.07 (0.26)	-6.05	-6.43 to -5.65	40.9	-4.91 (0.63)	-4.88	-5.77 to -4.25
AM-148	-9.77 (0.20)	-9.78	-10.09 to -9.41	15.4	-5.99 (1.09)	-5.81	-7.49 to 4.79
AM-149	-10.52 (0.54)	-10.52	-11.35 to -9.73	10.2	-5.40 (1.88)	-5.31	-8.73 to -3.36
AM-150	-5.95 (0.84)	-5.92	-7.25 to -4.93	41.7	-5.91 (1.78)	-5.93	-8.81 to -3.33
AM-151	-10.90 (1.11)	-11.21	-12.03 to -8.82	7.6	-8.04 (1.54)	-8.04	-10.40 to -5.72
AM-152	-10.40 (0.57)	-10.32	-11.30 to -9.79	11.0	-6.34 (1.51)	-5.96	-8.93 to -5.00
AM-153	-9.17 (1.11)	-9.55	-10.46 to -7.41	19.6	-5.67 (1.79)	-6.45	-7.28 to -2.42

Kruskal-Wallis tests of the δ¹³C values (χ²(31) = 191.12, p < 0.001) and the δ¹⁸O values (χ²(31) = 157.09, p < 0.001) of each individual camelid revealed significant differences in the means of the individuals for both carbon and oxygen. Dunn post-hoc tests further showed significant differences in the δ¹³C and δ¹⁸O value means between pairs of the individuals. From these post-hoc pairs, three individuals (AM-124, AM-142, AM-146) had mean δ¹³C values higher and two individuals (AM-123, AM-144) had mean δ¹³C values lower than 25% or more other individuals in the sample; for oxygen stable isotope analyses, one individual (AM-146) had a mean δ¹⁸O value higher and two individuals (AM-140, AM-143) had mean δ¹⁸O values lower than 25% or more other individuals in the sample. Full results of these post hoc tests can be found in Online Resource 2.

5.2 δ¹³C and δ¹⁸O values by tooth type

The δ¹³C and δ¹⁸O data were organized by tooth type (dP3, dP4, M1, M2, M3) to analyze intra-population differences at particular life stages. Summary statistics of the δ¹³C and δ¹⁸O data by tooth type are presented in Table 5. Plots showing the individual values associated with each tooth, separated by tooth type, are shown in Fig. 10 and Fig. 11.

We combined the samples associated with dP3 and dP4 teeth (n = 7 teeth) and M2 and M3 teeth (n = 22 teeth) to analyze differences in diet or location between pregnant females or nursing crias and adult camelids. The average δ¹³C and δ¹⁸O values from dP3 and dP4 teeth (n = 22 enamel powder samples) were – 6.6‰ and – 7.8‰, respectively. For M2 and M3 teeth (n = 125 enamel powder samples), the average δ¹³C and δ¹⁸O values were – 8.4‰ and – 6.2‰. A Mann-Whitney U test revealed significant differences in δ¹³C (U = 1815.5, p = 0.012) and δ¹⁸O (U = 799, p = 0.001) values between the dP3/dP4 and M2/M3 groups.

Table 5
Summary statistics of δ¹³C_{carb VPDB} and δ¹⁸O_{carb VPDB} by tooth type: dP3 (3/32, n = 7), dP4 (4/32, n = 15), M1 (14/32, n = 71 samples), M2 (15/32, n = 86), M3 (7/32, n = 39).
		dP3	dP4	M1	M2	M3
δ¹³C_{carb VPDB} (‰)	Mean	-4.09	-7.79	-8.69	-8.37	-8.61
	Median	-2.78	-8.61	-9.35	-9.24	-8.93
	Range	-9.21 to -2.67	-13.85 to -1.93	-15.10 to -3.62	-15.08 to -0.94	-11.31 to -2.03
	Standard Deviation	2.46	3.71	2.75	2.67	2.15
δ¹⁸O_{carb VPDB} (‰)	Mean	-6.97	-8.19	-6.15	-5.93	-6.71
	Median	-6.8	-8.84	-5.96	-5.99	-7.04
	Range	-10.1 to -5.1	-10.77 to -3.08	-10.62 to -2.42	-10.40 to 0.66	-12.24 to -1.50
	Standard Deviation	1.62	2.22	1.72	2.22	2.64

5.3 Intra-tooth variation in δ¹³C and δ¹⁸O values

To understand changes in diet and migration between the camelids on a smaller time scale, we analyzed differences in δ¹³C and δ¹⁸O values between the sequential samples taken from each tooth. Variation in δ¹³C and δ¹⁸O values across the development of each tooth type can be seen in Fig. 10 and Fig. 11. We found no significant differences between the mean of the two samples closest to the OCC and the mean of the one to two samples closest to the CEJ for M1, M2, or M3 teeth with three or more samples for either δ¹³C or δ¹⁸O values.

5.4 Inter-site comparisons in the Wari world

We analyzed the Auquimarca camelid enamel carbonate δ¹³C and δ¹⁸O values against those of other Wari-affiliated sites–Castillo de Huarmey (Tomczyk et al. 2019), Uraca (Alaica et al. 2022), and Quilcapampa (Melton et al. 2023)–as shown in Fig. 12. Mann-Whitney U tests between the complete δ¹³C datasets from Auquimarca and these other sites revealed a significant difference (U = 5506.5, p < 0.001) between Auquimarca (n = 235) and Uraca (n = 65, mean = – 9.5‰); the difference between Auquimarca and Quilcapampa (n = 10) approached significance (mean = – 9.91‰, U = 772.5, p = 0.07). For oxygen, the δ¹⁸O values from camelids at Auquimarca significantly differed from those at both Quilcapampa (n = 10, mean = – 1.48‰, U = 2074.5, p < 0.001) and Castillo de Huarmey (n = 18, mean = – 8.86‰, U = 3933.5, p < 0.001). Uraca δ¹⁸O_{enamel carb} values were not reported and therefore are not analyzed here.

We also compared the %C₄ plants in camelid diets calculated from δ¹³C values (Dufour et al. 2014) across all Wari-affiliated sites with camelid stable isotope analyses and modern camelid samples. The Wari-affiliated sites included in this comparison are Beringa (only Middle Horizon camelids, Alaica et al. (2022)), Conchopata (Finucane et al., 2006), Quilcapampa (Melton et al. 2023), Cerro Baúl (Thornton et al. 2011), Castillo de Huarmey (Tomczyk et al. 2019), and Uraca (Alaica et al. 2022). The modern camelid samples are from puna environments in Chilligua and Tocra (southern Peruvian coast) and Quiruvilca (northern Peruvian coast), as reported in Thornton et al. (2011) and Dufour et al. (2014). The results of Mann-Whitney U tests between %C₄ plants of camelids at Auquimarca and those of other sites can be seen in Table 6.

Table 6
Comparison of the average %C₄ plants in the diets of individuals from Auquimarca to those of individuals from other Wari-affiliated sites and modern herders (from Tocra, Quiruvilca, and Chilligua) using Mann-Whitney U tests. δ¹³C values used to calculate %C₄ plants in diet values come from different materials: collagen (col), enamel carbonates (ec), and keratin (ker). *asterisks indicate significance.
Middle Horizon Wari-Affiliated Sites	%C₄ plants in diet per individual (Dufour et al. 2014)		Mann-Whitney U test vs. Auquimarca
Middle Horizon Wari-Affiliated Sites	Average	Range	p value	U statistic
Beringa (ker, n = 5) (Alaica et al. 2022)	5.18%	-8.22% to 35.7%	0.008*	20
Quilcapampa (ec, n = 4) (Melton et al. 2023)	11.6%	0% to 18.6%	0.113	32
Conchopata (col, n = 17) (Finucane et al. 2006)	54.8%	10.3% to 88.3%	0.005*	407
Cerro Baúl (col, n = 11) (Thornton et al. 2011)	31.4%	10.1% to 74.3%	0.687	191
Castillo de Huarmey (ec, n = 18) (Tomczyk et al. 2019)	21.6%	-1.72% to 46.1%	0.229	228
Uraca (ec, n = 14) (Alaica et al. 2022)	17.3%	2.5% to 44.0%	0.064	146
Modern sites (col, n = 9) (Thornton et al. 2011; Dufour et al. 2014)	8.69%	1.38% to 13.9%	0.003*	49

6.1 Dietary breadth among Auquimarca camelids

δ¹³C values from the Auquimarca camelid enamel carbonates samples (n = 235) ranged from – 15.10‰ (− 2.2% C₄ plants) to – 0.94‰ (74.1% C₄ plants), with an average of – 8.35‰. If all of the camelids at Auquimarca had been raised and rangestock grazed in the puna, like many modern camelids in the MRV and other highland Andean regions, δ¹³C values would have been more negative, likely between – 17.9‰ and – 8.5‰ (Szpak et al. 2013; Dufour et al. 2014). These lower values would reflect a primarily C₃ diet consistent with the puna grasslands and ecosystem (Mayer 1979; Flannery et al. 1989; Dransart 2002; Thornton et al. 2011; Dufour et al. 2014). Conversely, if the camelids had been primarily given maize fodder and allowed to graze on maize stubble after the harvest, similar to the feeding practices for camelids from Middle Horizon coastal sites, the camelids would have had an overall higher δ¹³C average, likely between – 1.6‰ to + 2.5‰ (Dufour et al., 2014; Finucane et al., 2006; Noe et al., 2024; Szpak et al., 2013; Thornton et al., 2011; Tomczyk et al., 2019). But the broad range of δ¹³C values observed in the Auquimarca samples suggests that camelids at Auquimarca had access to both C₃ and C₄ plants, with some variation in the proportions of these kinds of plants in their diets.

There was major variation in the diets of the 32 individual camelids at Auquimarca. When the δ¹³C values of the samples associated with a particular individual were averaged, there were significant differences between the means of the individuals from each other and from the population mean, suggesting that multiple husbandry strategies were utilized by those managing the Auquimarca camelid herds. Three individuals (AM-124, AM-142, AM-146) had mean δ¹³C values greater than 25% of other individuals. They consumed, on average, approximately 58.3% C₄ plants in their diets. This is similar to the δ¹³C enriched group of camelids observed at Conchopata, the urban Wari heartland site where camelids were likely kept in canchas and fed primarily maize (Finucane et al., 2006). Eleven additional camelids in the study sample consumed between 30% to 75% C₄ plants (average = 46.4%). These δ¹³C values are more consistent with Central Andean coastal sites with greater incorporation of maize into camelid diets (average range of 40–70% C₄ plants in diet) (Szpak et al. 2019). Auquimarca’s MRV location provides a more suitable location for maize agriculture on wide fields, much like those seen in some coastal areas and lowland areas. Therefore, where other highland kichwa sites have more limited maize influence on human and camelid diets, it appears that many of the camelids at Auquimarca had diets influenced by this environment and possibly by the maize subsistence priorities of the Wari Empire.

In contrast, two individuals (AM-123, AM-144) had mean δ¹³C values lower than 25% of the study population. They consumed virtually no C₄ plants based on Dufour et al’s (2014) method, with an average of – 1.45% C₄ plants in their diets. This is more on par with the very low δ¹³C values of camelids in modern pastoral communities who tend to rangestock graze their livestock in the puna almost exclusively. For comparison, modern camelids from highland areas on the northern and southern Peruvian coasts, analyzed by Thornton et al. (2011) and Dufour et al. (2014), had an average of 8.69% C₄ plants in their diets (Table 6). Sixteen (16) additional individuals (18/32) also had low levels of maize consumption, consuming approximately less than 30% C₄ plants in their diets (average = 14.3%). This is consistent with other studies of camelids in the Andean highlands, where closer proximity to the puna grasslands and desire to diversify subsistence strategies often led to large levels of rangestock grazing (Browman, 1974; Finucane et al., 2006). At sites such as Beringa and Quilcapampa, camelids were likely primarily pastured on C₃ grasses. The minimal incorporation of maize in the diet may reflect Wari imperial influence (ibid; Alaica et al., 2022; Melton et al., 2023).

Visual analysis of the δ¹³C boxplots for each individual (Fig. 8) supports the existence of camelid groups with varying proportions of C₃ to C₄ plants in their diets. In these boxplots, the interquartile ranges scarcely overlap with the population mean δ¹³C value, showing that some of the camelids were eating diets more heavily maize-based while others consumed primarily C₃ plants. The variety of diets suggests that those managing the camelids at Auquimarca were not engaged in purely maize foddering or purely rangestock grazing. More likely, separate or flexible pastoral strategies were used that allowed the agropastoralists to lean more heavily on maize foddering or rangestock grazing, while oftentimes incorporating elements of both (Alaica et al., 2022; Eerkens et al., 2014; Finucane et al., 2006; Thornton et al., 2011; Tomczyk et al., 2019). This kind of dimorphic husbandry model, as has been described by Tomczyk et al. (2019), has been observed at Wari-affiliated sites like Conchopata, Castillo de Huarmey, and Cerro Baúl to varying degrees (ibid; Finucane et al., 2006; Thornton et al., 2011).

To compare the flexible husbandry model observed at Auquimarca to subsistence practiced in the MRV, we turn to Browman’s (1970; 1974; 1976) work in the region. He argued that prior to the Middle Horizon, people in this region were primarily engaged in rangestock grazing practices in the puna, with limited horticulture at lower elevations close to the river. With the onset of the Middle Horizon and influence by the Wari Empire, people in the MRV relied more heavily on maize agriculture in the kichwa and camelid meat for protein. Combined with other archaeological evidence from this site, like stone hoes for tilling and butchery marks on camelid remains, the camelids with higher δ¹³C values seem to support this hypothesis. This suggests that maize was a significant part of livelihoods during the Middle Horizon in the MRV. The lower %C₄ in camelid diets from Auquimarca suggest that rangestock grazing methods likely used during earlier periods in the MRV did not disappear during the Middle Horizon. Other studies have suggested that dimorphic and flexible husbandry models like this were used as risk management strategies to withstand maize crop failures (Finucane et al., 2006; Noe et al., 2024; Tomczyk et al., 2019). Furthermore, maintaining some level of rangestock grazing could also have been a cultural practice to keep connections to the mountain spirits, as seen in modern herding communities (Flannery et al. 1989; Dransart 2002; Duche-Pérez and Mamani-Daza 2024).

Analysis of δ¹³C values within the individual camelids at Auquimarca revealed additional variation in their diets on a smaller time scale. The mean δ¹³C values of dP3 and dP4 teeth (n = 7 teeth, mean = – 6.6‰) were significantly higher than those of M2 and M3 teeth (n = 22 teeth, mean = – 8.4‰) (Wheeler 1982; Takigami et al. 2020). This reinforces ethnohistoric accounts of camelids staying in the same place for up to two years for training before joining caravans (Flannery et al. 1989). Likewise, visual analysis of the sequential line graphs for dP3/dP4 teeth and M2/M3 teeth (Fig. 10) reveals stark differences in the diets between individual teeth. This supports the idea that multiple flexible pastoral strategies were being practiced by those at Auquimarca. Some pregnant and adult camelids had a fairly constant, primarily C₃ diets while others seemed to have a more constant source of C₄ plants, like maize, in their diets.

The dP3/dP4 and M1, M2, and M3 teeth differ in the amount of δ¹³C change within the development of those teeth. δ¹³C values remain fairly constant from the OCC to the CEJ in dP3 and dP4 teeth, whereas many of the molars exhibited steep increases (AM-123 [M1], AM-125 [M2]), decreases (AM-123 [M2]), or more cyclical patterns (AM-129 [M2], AM-146 [M2], AM-151 [M2], and AM-138 [M3]) of δ¹³C values between sequential samples (Fig. 10). Although we did not observe significant differences between the intra-tooth samples associated with the earliest and latest periods of development in the molars analyzed, visual analysis of the line plots for these teeth reveals the entire extent of δ¹³C change across the development of those teeth. Increases or decreases in δ¹³C values within the molars could reflect movement across the Andean environment throughout life as part of trade networks that connected hinterland communities to each other and to the Wari heartlands: an aspect of the vertical archipelago (Murra 1972; Alaica and González La Rosa 2019; Tomczyk et al. 2019; Alaica et al. 2022; Melton et al. 2023). Camelids then would have had access to different kinds of plants more readily available in different ecological zones. The camelids with more cyclical δ¹³C change within their teeth could have traveled across microclimates at different elevations, rotating between rangestock grazing in the puna to grazing on the stubble of maize fields after a harvest in the kichwa as part of a seasonal migration, as has been suggested at Castillo de Huarmey and other Andean sites (Dransart 2002; Tomczyk et al. 2019). Alternatively, these patterns could reflect movement of the camelids between ecological zones as part of camelid caravans, consuming different amounts of C₃ and C₄ plants based on their availability in that environment (Takigami et al. 2020).

Although camelid remains analyzed in this study were found at activity level either within the tombs or in the sediment layer above them, we do not have specific AMS dates from the camelids themselves. The AMS dates from the human burials indicate that Auquimarca was occupied during the Middle Horizon; however, the AMS dates are not entirely homogenous (Krause and Tung 2025). Therefore, it is possible (and likely) that the camelids in this sample were deposited at Auquimarca at different times throughout the Middle Horizon and potentially at periods with varying impacts from imperial Wari hegemony. As such, the multiple pastoral strategies that we have suggested could have been used at differing times during the Middle Horizon rather than occurring concurrently. However, no significant differences were observed between the camelids recovered from inside tombs versus those in the sediment layer at activity level, so if different time periods were represented in the layers, they likely did not have a significant impact on the pastoral activities of those at Auquimarca. Further analysis of these camelid remains, including AMS dates and strontium isotope analysis (⁸⁷Sr/⁸⁶Sr), is needed to get a better sense of the depositional timeline and local environment of these camelids.

6.2 Andean mobility among Auquimarca camelids

δ¹⁸O values from the enamel carbonate samples (n = 235) ranged from – 12.24‰ to + 0.66‰ VPDB. Dufour et al’s (2014) approximations for δ¹⁸O_{enamel carb} by Andean ecological zone, which account for evaporative and meteoric water sources, are the following: lowland and coastal regions range from – 5.8‰ to + 0.5‰, kichwa regions range from – 6.2‰ to – 2.3‰, and puna regions range from – 15.0‰ to – 5.5‰. Under this model, the camelids at Auquimarca, when looked at collectively, likely inhabited a variety of environments ranging from the coast to the puna, with the average δ¹⁸O value (– 6.50‰) falling in the lower puna region. Additionally, the broad range of δ¹⁸O values in the sample and the intra-population and intra-individual variation in these values suggests that there was human management of the camelids at Auquimarca, helping determine the kinds of water and plants they had access to and their movement across the Andean landscape (Tomczyk et al. 2019; Carrasco et al. 2022). However, other factors tied to seasonality, like temperature and humidity levels, and the kinds of plants consumed can impact δ¹⁸O_{enamel carb} values, since weaned camelids obtain most of their water from vegetal sources post-weaning and little water from evaporative water sources, so caution must be taken when directly comparing the δ¹⁸O_{enamel carb} values of Auquimarca camelids to general environmental estimations (Kohn et al. 1996; Wright and Schwarcz 1998; Katzenberg 2008; Yann et al. 2016; Carrasco et al. 2022; Enke et al. 2022).

There was less variation between the mean δ¹⁸O values of the 32 individual camelids than of those seen in the δ¹³C dataset. This can be observed in the higher overlap of interquartile ranges with the entire δ¹⁸O sample mean among the Auquimarca δ¹⁸O boxplots, an indicator that many of the camelids were local to Auquimarca (Fig. 9). However, there were still significant differences between the means of certain camelids and others and many of the individuals have large ranges of δ¹⁸O values. This can suggest that a large proportion of the camelids samples from Auquimarca were traversing the Andean landscape throughout their lives, exposing them to a variety of δ¹⁸O values in their foods and thus water sources (Weber 2019; Takigami et al. 2020; Melton et al. 2023).

The individual camelids at Auquimarca exhibited significantly lower δ¹⁸O values than the camelids sampled at Quilcapampa and Castillo de Huarmey. This is to be expected because these sites are located at lower elevations closer to the coast (Quilcapampa = 800 masl, Castillo de Huarmey = 50 masl) compared to Auquimarca, which is located in the high kichwa region (3220 masl) (Katzenberg 2008; Knudson 2009; Dufour et al. 2014; Tomczyk et al. 2019; Alaica et al. 2022; Noe et al. 2024). Within the full camelid populations at these three sites, there is a much broader range of δ¹⁸O values at Auquimarca (n = 235, range = 12.9‰) than at Quilcapampa (n = 10, range = 8.43‰) or Castillo de Huarmey (n = 18, range = 7.32‰). The larger range of δ¹⁸O in the population suggests that the camelids at Auquimarca had varied diets and varied migration histories (ibid). It also attests to the lengths people went to - to bring their camelids to this site to engage in mortuary activities and to offer their animals in feasts and as burial offerings in tombs. This is supported by the previously discussed differences in the proportions of C₃ and C₄ plants between the camelids at Auquimarca. But the sample size at Auquimarca is much larger than those of Quilcapampa or Castillo de Huarmey, so these range values alone do not necessarily represent the entire extent of diet and mobility variation at Quilcapampa or Castillo de Huarmey.

When the δ¹⁸O distribution of the entire sample population and the individual camelids are considered alongside intra-individual changes in δ¹⁸O values, the extent of movement and dietary diversity within the population is revealed. The mean values of dP3 and dP4 teeth (n = 7 teeth, mean = – 7.8‰) and M2 and M3 teeth (n = 22 teeth, mean = – 6.2‰) significantly differed from each other. But as seen in our carbon data, looking at the line graphs of these teeth show differences in δ¹⁸O between teeth in these categories. dP3/dP4 teeth have relatively linear δ¹⁸O trends throughout the development of those teeth that range between – 10.8‰ and – 3.08‰ (Fig. 11). From this, the pregnant camelids included in the sample were probably kept in the same region with similar diets for at least the later part of their pregnancy when these teeth would have mineralized (Takigami et al. 2020). But the region and diet in question varied, suggesting that there was no one setting where camelids were kept during pregnancy.

Differences between the M2/M3 teeth are similar to those of the dP3/dP4 teeth. Some individuals had consistently and significantly high δ¹⁸O values throughout their enamel development while others had consistently low δ¹⁸O values; the range of δ¹⁸O from M2 and M3 teeth are – 12.24 to + 0.66‰. So even in later life, the camelids at Auquimarca likely inhabited different environments, further supporting the idea of multiple flexible pastoral strategies at the site (Tomczyk et al. 2019; Takigami et al. 2020; Melton et al. 2023). Some camelids likely inhabited the puna region while others might have been kept closer to human settlements in the kichwa where they consumed greater proportions of C₄ plants that could further enrich δ¹⁸O values (Kohn et al. 1996; Wright and Schwarcz 1998; Katzenberg 2008; Yann et al. 2016; Carrasco et al. 2022; Enke et al. 2022). Finucane et al. (2006)and others have suggested that differences like this may represent alpacas versus llamas, where alpacas were kept in the puna and llamas were kept closer to human settlements for easier access to be used in trading (Alaica and González La Rosa 2019). However, this assumption has been disputed because llamas and alpacas can live in a single herd and can crossbreed (Shimada and Shimada 1985; Tomczyk et al. 2019). Regardless, this cannot be assessed here because of the fragmented nature of the Auquimarca camelid samples and our inability to tell whether the camelids were alpacas or llamas.

Among the camelid molars sampled, we also see many individuals with large differences in δ¹⁸O values between the sequential samples. Once again, there was no significant difference between the δ¹⁸O values of the samples associated with earlier and later periods of enamel development, but we do see other patterns between those samples. Visual analysis of the line graphs for δ¹⁸O (Fig. 11) shows steady δ¹⁸O value increases, decreases, or cyclical patterns of movement across the vertical archipelago in many of the molars studied. Variation in soil water δ¹⁸O within an ecological zone due to seasonal changes in humidity, temperature, and precipitation could have also contributed to the observed cyclical patterns of δ¹⁸O increase and decrease among the camelids’ sequential δ¹⁸O values (Yann et al. 2016). But because of the range of δ¹⁸O values between the sequential samples, these camelids were likely moving between coastal/lowland and kichwa or kichwa and puna ecological zones. It is possible that the movement between these regions were the result of the agropastoralists negotiating with their Andean vertical landscape, preventing overgrazing while still having the resources available to produce agricultural products by moving the camelids between ecological zones to protect puna grasslands and provide access to maize stubble after harvests (Tomczyk et al. 2019). This model fits in well with the mixed C₃/C₄ plant diets observed in many of the individuals. Or, these increases and decreases in δ¹⁸O values between the sequential samples could be the result of the camelids moving between ecological zones in order to trade products with other parts of the Wari Empire, as seen among camelids at Castillo de Huarmey and Cerro Baúl (ibid; Thornton et al. 2011). The MRV is such a suitable environment for maize cultivation. It could be that some of the camelids were used as pack animals in caravans to transport this important crop to other parts of the empire in exchange for other goods; unfortunately, this cannot be shown from these isotope analyses alone. Further work must be done to untangle the trade relationships between those at Auquimarca and the rest of the Wari world.

Our study contributes significant new, fine-grained, intra-individual camelid stable carbon and oxygen isotope data from the highland Middle Horizon mortuary site of Auquimarca. The work provides the first zooarchaeological biochemical data to study subsistence practices in the MRV, a region where the extent of Wari influence has been relatively unknown. These results are important for understanding how Wari influence impacted communities in distant and remote parts on the edge of an Andean empire. We contribute a life history study of diet and migration histories of camelids at Auquimarca to the existing literature on Wari-associated subsistence strategies. As such, we elucidate a crucial account of human-animal interactions and management practices in the central highlands and build upon the limited research about camelid pastoralism in the Wari Empire.

The reported δ¹³C and δ¹⁸O data revealed that the camelids were likely managed using different flexible pastoral strategies attuned to the constraints and affordances of pregnant and juvenile camelids and were utilized for different purposes within the agropastoral system. Some camelids were likely rangestock grazed on C₃ plants in the puna, some were maize foddered in the kichwa, and others fell somewhere in between, possibly kept in close proximity to human settlements but rotated between C₃ grass grazing and consuming maize stubble to ensure long-term subsistence success. Certain camelids had differences in δ¹³C and δ¹⁸O values throughout their lives. This suggests camelid involvement in long distance trade networks or camelid caravans, possibly related to broader trade and exchange systems in the MRV and the Wari Empire. The agropastoral system described here simultaneously considers the importance of maize in the Middle Horizon, and the maintenance of puna grazing practices for risk management and possibly cultural retention amid social change brought about by Wari hegemony (Finucane et al. 2006; Finucane 2009; Tung 2012; Tomczyk et al. 2019; Noe et al. 2024).

Author Contribution

L.B. wrote the main manuscript text, analyzed the data, and prepared all figures and tables not mentioned elsewhere. A.A. wrote the zooarchaeological analysis section and created table 2 and figure 4. M.K. helped write the Auquimarca site context and human interments at Auquimarca sections. A.A., A.L., and T.T. provided editing support to L.B. on the manuscript text. D.A. and J.P.G. oversaw and managed the rescue excavations of Auquimarca. A.A. and L.M.G. conducted the zooarchaeological inventory and analysis. L.B. and A.L. prepared the enamel carbonate samples. T.T. organized the export of the camelid remains to Vanderbilt University, obtained funding for analysis and writing this manuscript, and provided conceptual support. All authors reviewed the manuscript.

Acknowledgement

This work was supported by the Vanderbilt University College of Arts & Science and the Vanderbilt University Office of the Provost. Funding from the Killam Trust and the Department of Anthropology at the University of Alberta supported the zooarchaeological analysis. The Department of Anthropology at The University of British Columbia supported some of the writing in this manuscript. We also thank the Peruvian Ministry of Culture for supporting the rescue excavation at Auquimarca in 2021 and 2022 and for allowing the camelid samples analyzed here to be exported to Vanderbilt University for isotopic analysis in the Bioarchaeology and Stable Isotope Research Lab. We thank everyone who supported the excavation of Auquimarca. Finally, we would like to thank Alyssa Sklar for providing GIS technical support.

Alaica AK, González LM Analisis Zooarqueologico. In: Venegas DA (ed) Aquishmarca: Los Abuelos del Valle de Mantaro. Gobierno Regional de Junín, Lima, pp 134–148
Alaica AK, González LM (2019) A Look to the North and South: Camelid Herding Strategies in the Desert Coast of Peru. In: Alday C, Morrisset S (eds) Desert Archaeology. Cambridge, UK
Alaica AK, González LM, Muro Ynoñán LA, et al (2021) Camelid caravans and Middle Horizon exchange networks. In: Caravans in Socio-Cultural Perspective, 1st edn. Routledge, p 23
Alaica AK, Scaffidi BK, González La Rosa LM, et al (2022) Flexible agropastoral strategies during the 1st millennium CE in southern Peru: Examining yunga Arequipa camelid husbandry practices during Wari expansion through stable isotope analysis (δ13C and δ15N) in the Majes and Sihuas Valleys. Quat Int 634:48–64. https://doi.org/10.1016/j.quaint.2022.06.015
Aramburú Venegas D (2022) Auquishmarca: Los abuelos de Valle del Mantaro. National Library of Peru, Huancayo
Barbour MM (2007) Stable oxygen isotope composition of plant tissue: a review. Funct Plant Biol 34:83. https://doi.org/10.1071/FP06228
Browman DL (1970) Early Peruvian peasants: the culture history of a central highlands valley
Browman DL (1978) Advances in Andean Archaeology. De Gruyter Mouton, The Hague [Noordeinde 41]
Browman DL (1974) Pastoral Nomadism in the Andes. Curr Anthropol 15:188–196
Browman DL (1976) Demographic Correlations of the Wari Conquest of Junin. Am Antiq 41:465–477. https://doi.org/10.2307/279012
Bryant JD, Froelich PN, Showers WJ, Genna BJ (1996) Biologic and climatic signals in the oxygen isotopic composition of Eocene-Oligocene equid enamel phosphate. Palaeogeogr Palaeoclimatol Palaeoecol 126:75–89. https://doi.org/10.1016/S0031-0182(96)00071-5
Cadwallader L, Beresford-jones DG, Whaley OQ, O’connell TC (2012) The Signs of Maize? A Reconsideration of What δ13C Values Say about Palaeodiet in the Andean Region. Hum Ecol 40:487–509. https://doi.org/10.1007/s10745-012-9509-0
Capriles J, Tripcevich N (2016) The Archaeology of Andean Pastoralism
Carrasco TS, Scherer CS, Ribeiro AM, Buchmann FS (2022) Paleodiet of Lamini camelids (Mammalia: Artiodactyla) from the Pleistocene of southern Brazil: insights from stable isotope analysis (δ13C, δ18O). Paleobiology 48:513–526. https://doi.org/10.1017/pab.2022.10
Cavagnaro JB (1988) Distribution of C3 and C4 grasses at different altitudes in a temperate arid region of Argentina. Oecologia 76:273–277
Cerling TE, Harris JM (1999) Carbon isotope fractionation between diet and bioapatite in ungulate mammals and implications for ecological and paleoecological studies. Oecologia 120:347–363. https://doi.org/10.1007/s004420050868
Coplen TB (1996) Guidelines for the Reporting of Stable Hydrogen, Carbon, and Oxygen Isotope-Ratio Data. Paleoceanography 11:369–370
Coplen TB, Kendall C, Hopple J (1983) Comparison of stable isotope reference samples. Nature 302:236–238. https://doi.org/10.1038/302236a0
D’Altroy TN, Hastorf CA (2001) Empire and Domestic Economy. Kluwer Academic/Plenum Publishers, New York, NY
deFrance SD (2014) The Luxury of Variety: Animals and Social Distinction at the Wari site of Cerro Baúl, Southern Peru. In: Arbuckle BS, McCarty SA (eds) Animals and Inequality in the Ancient World. University Press of Colorado, pp 63–84
Dransart PZ (2002) Earth, Water, Fleece and Fabric: An ethnography and archaeology of Andean camelid herding. Routledge, London
Duche-Pérez AB, Mamani-Daza LJ (2024) Llamas, Barter and Travel Rituals: An Ethnographic Study on the Esquela Tusuy Dance of the Uchumiri Peasant Community, Condesuyos, Peru. Religions 15:534. https://doi.org/10.3390/rel15050534
Dufour E, Goepfert N (2020) Past Andean Pastoralism: A Reconsidered Diversity. Introduction to the Special Issue. Environ Archaeol 25:257–261. https://doi.org/10.1080/14614103.2019.1619981
Dufour E, Goepfert N, Le Neün M, et al (2018) Life History and Origin of the Camelids Provisioning a Mass Killing Sacrifice During the Chimú Period: Insight from Stable Isotopes. Environ Archaeol 25:310–324. https://doi.org/10.1080/14614103.2018.1498165
Dufour E, Goepfert N, Léon BG, et al (2014) Pastoralism in Northern Peru during Pre-Hispanic Times: Insights from the Mochica Period (100–800 AD) Based on Stable Isotopic Analysis of Domestic Camelids. PLOS ONE 9:e87559. https://doi.org/10.1371/journal.pone.0087559
Earle T, Jennings J (2012) Remodeling the Political Economy of the Wari Empire. Bol Arqueol Pucp 16:209–226
Eerkens JW, de Voogt A, Dupras TL, et al (2014) Intra- and inter-individual variation in δ 13 C and δ 15 N in human dental calculus and comparison to bone collagen and apatite isotopes. J Archaeol Sci 52:
Enke N, Brinkmann L, Runa RA, et al (2022) Drinking behaviour of llamas (Lama glama) in choice tests for fresh or saline water. Small Rumin Res 216:106806. https://doi.org/10.1016/j.smallrumres.2022.106806
Finucane B, Agurto PM, Isbell WH (2006) Human and animal diet at Conchopata, Peru: stable isotope evidence for maize agriculture and animal management practices during the Middle Horizon. J Archaeol Sci 33:1766–1776. https://doi.org/10.1016/j.jas.2006.03.012
Finucane BC (2009) Maize and Sociopolitical Complexity in the Ayacucho Valley, Peru. Curr Anthropol 50:535–545. https://doi.org/10.1086/599860
Flanagan LB, Comstock JP, Ehleringer JR (1991) Comparison of Modeled and Observed Environmental Influences on the Stable Oxygen and Hydrogen Isotope Composition of Leaf Water in Phaseolus vulgaris L. 1. Plant Physiol 96:588–596. https://doi.org/10.1104/pp.96.2.588
Flannery K, Marcus J, Reynolds R (1989) The Flocks of the Wamani: A Study of Llama Herderson the Punas of Ayacucho, Peru. Academic Press, San Diego
Garvie-Lok SJ, Varney TL, Katzenberg MA (2004) Preparation of bone carbonate for stable isotope analysis: the effects of treatment time and acid concentration. J Archaeol Sci 31:763–776. https://doi.org/10.1016/j.jas.2003.10.014
Giersz M (2016) Castillo de Huarmey: centro político wari en la costa norte del Perú. Nuevas Perspect En Organ Política Wari ANDES Bol Cent Estud Precolomb Univ Vars Vol 9 Ed Miłosz Giersz Krzysztof Makowski Pol Soc Lat Am Stud Inst Fr DÉtudes Andin Wars-Lima
Grávalos EM, Reid DA, Nash DJ, Ryan Williams P (2023) Crafting cosmopolitanism: Ceramic production and exchange during Wari imperialism (600–1000 CE). J Archaeol Sci Rep 50:103878. https://doi.org/10.1016/j.jasrep.2023.103878
Heikkinen AM (2021) Climate change, power, and vulnerabilities in the Peruvian Highlands. Reg Environ Change 21:82. https://doi.org/10.1007/s10113-021-01825-8
Hester J, Bryan J (2024) readr: Read Rectangular Text Data
Isbell WH (2004) Mortuary Preferences: A Wari Culture Case Study from Middle Horizon Peru. Lat Am Antiq 15:3–32. https://doi.org/10.2307/4141562
Isbell WH (1997) Reconstructing Huari. Plenum Press, New York, New York :
Isbell WH, McEwan GF (1991) Huari Administrative Structure: Prehistoric Monumental Architecture and State Government. Dumbarton Oaks Research Library and Collection, Washington, D.C
Jennings J (2025) Wari: Imperialism, Low Power, and Globalization in the Middle Horizon Central Andes. J Archaeol Res 33:249–296. https://doi.org/10.1007/s10814-024-09199-1
Jennings J, Craig N (2001) Politywide Analysis and Imperial Political Economy: The Relationship between Valley Political Complexity and Administrative Centers in the Wari Empire of the Central Andes. J Anthropol Archaeol 20:479–502. https://doi.org/doi:10.1006/jaar.2001.0385
Katzenberg MA (2008) Stable Isotope Analysis: A Tool for Studying Past Diet, Demography, and Life History. In: Katzenberg MA, Saunders SR (eds) Biological Anthropology of the Human Skeleton, 2nd edn. Wiley
Knudson KJ (2009) Oxygen isotope analysis in a land of environmental extremes: the complexities of isotopic work in the Andes. Int J Osteoarchaeol 19:171–191. https://doi.org/10.1002/oa.1042
Kohn MJ, Schoeninger MJ, Valley JW (1996) Herbivore tooth oxygen isotope compositions: Effects of diet and physiology. Geochim Cosmochim Acta 60:3889–3896. https://doi.org/10.1016/0016-7037(96)00248-7
Krause MB (2025) Children of the Empire: A Bioarchaeological Investigation of Childhood and Socialization in the Wari Empire (600 to 1000 CE). Vanderbilt University
Krause MB, Tung TA (2025) Infant Feeding Practices within a Bioarchaeology of Care Framework: Investigating Isotopic and Paleopathological Data from a Wari Community (600-1000 CE) in the Peruvian Andes. Denver, CO
Le Neün M, Dufour E, Goepfert N, et al (2023) Can first phalanx multivariate morphometrics help document past taxonomic diversity in South American camelids? J Archaeol Sci Rep 47:103708. https://doi.org/10.1016/j.jasrep.2022.103708
Maldonado YL, Tomaylla MP, Mendoza XS (2022) Producción de la Cerámica Utilitaria y Ritual de Auquishmarca. In: Venegas DA (ed) Auquishmarca: Los Abuelos Del Valle Del Mantar. Gobierno Regional de Junín, Lima
Mark BG, French A, Baraer M, et al (2017) Glacier loss and hydro-social risks in the Peruvian Andes. Glob Planet Change 159:61–76. https://doi.org/10.1016/j.gloplacha.2017.10.003
Matos Mendieta R (1978) The Cultural and Ecological Context of the Mantaro Valley During the Formative Period. In: Advances in Andean Archaeology. De Gruyter Mouton, The Hague [Noordeinde 41]
Mayer E (1979) Land use in the Andes: Ecology and agriculture in the Mantaro Valley of Peru with special reference to potatoes. Centro Internacional de la Papa
McEwan GF (1996) Archaeological Investigations at Pikillacta, a Wari Site in Peru. J Field Archaeol 23:169–186. https://doi.org/10.1179/009346996791973945
Melton MA, Alaica AK, Biwer ME, et al (2023) Reconstructing Middle Horizon Camelid Diets and Foddering Practices: Microbotanical and Isotope Analyses of Dental Remains from Quilcapampa, Peru. Lat Am Antiq 34:783–803. https://doi.org/10.1017/laq.2022.80
Miranda-de la Lama GC, Villarroel M (2023) Behavioural biology of South American domestic camelids: An overview from a welfare perspective. Small Rumin Res 220:106918. https://doi.org/10.1016/j.smallrumres.2023.106918
Murra JV (1972) El control vertical de un máximo de pisos ecológicos en la economía de las sociedades andinas. In: Murra JV (ed) Formaciones económicas del mundo andino. Instituto de Estudios Peruanos, Lima
Neuwirth E (2022) RColorBrewer: ColorBrewer Palettes
Nielsen AE (2000) Andean caravans: An ethnoarchaeology. Ph.D., The University of Arizona
Noe SJ, McCool WC, Wilson KM (2024) A cross-regional examination of camelid herding practices in Peru from 900 BCE to 1450 CE: Insights from stable isotopes in camelid bone collagen and fiber. PLOS ONE 19:e0306205. https://doi.org/10.1371/journal.pone.0306205
Ogle DH, Doll, Wheeler AP, Dino A (2025) FSA: Simple Fisheries Stock Assessment Methods
Parsons JR, Hastings CM, Matos RM (2000) Prehispanic Settlement Patterns in the Upper Mantaro and Tarma Drainages, Junín, Peru. University of Michigan, Ann Arbor, MI
Passey BH, Cerling TE (2002) Tooth enamel mineralization in ungulates: implications for recovering a primary isotopic time-series. Geochim Cosmochim Acta 66:3225–3234. https://doi.org/10.1016/S0016-7037(02)00933-X
Pederzani S, Britton K (2019) Oxygen isotopes in bioarchaeology: Principles and applications, challenges and opportunities. Earth-Sci Rev 188:77–107. https://doi.org/10.1016/j.earscirev.2018.11.005
Pilaar Birch SE (2013) Stable isotopes in zooarchaeology: an introduction. Archaeol Anthropol Sci 5:81–83. https://doi.org/10.1007/s12520-013-0126-7
Pilaar Birch SE, Metzger M, Ridder E, et al (2025) Domestic herbivore mobility and herd management at Bronze Age Politiko-Troullia, Cyprus. Archaeol Anthropol Sci 17:114. https://doi.org/10.1007/s12520-025-02226-2
Powell R, Still C (2009) Biogeography of C3 and C4 vegetation in South America. XIV Simpósio Bras Sensoriamento Remoto 14:
R Core Team (2023) R: A Language and Environment for Statistical Computing
Revelle W (2025) psych: Procedures for Psychological, Psychometric, and Personality Research
Schreiber KJ (1992) Wari Imperialism in Middle Horizon Peru. University of Michigan Museum of Anthropological Archaeology, Ann Arbor, MI
Shimada M, Shimada I (1985) Prehistoric Llama Breeding and Herding on the North Coast of Peru. Am Antiq 50:3–26. https://doi.org/10.2307/280631
Smith BN, Epstein S (1971) Two Categories of 13C/12C Ratios for Higher Plants 1. Plant Physiol 47:380–384. https://doi.org/10.1104/pp.47.3.380
Stone RR (2012) Art of the Andes: From Chavín to Inca, 3rd edn. Thames & Hudson, New York, NY
Szpak P, Millaire J-F, Chapdelaine C, et al (2019) An Integrated isotopic study of Early Intermediate Period camelid husbandry in the Santa Valley, Perú. Environ Archaeol 25:1–17. https://doi.org/10.1080/14614103.2019.1583302
Szpak P, White CD, Longstaffe FJ, et al (2013) Carbon and Nitrogen Isotopic Survey of Northern Peruvian Plants: Baselines for Paleodietary and Paleoecological Studies. PLoS ONE 8:e53763. https://doi.org/10.1371/journal.pone.0053763
Takigami M, Uzawa K, Seki Y, et al (2020) Isotopic Evidence for Camelid Husbandry During the Formative Period at the Pacopampa Site, Peru. Environ Archaeol 25:262–278. https://doi.org/10.1080/14614103.2019.1586091
Thornton EK, Defrance SD, Krigbaum J, Williams PR (2011) Isotopic evidence for Middle Horizon to 16th century camelid herding in the Osmore Valley, Peru. Int J Osteoarchaeol 21:544–567. https://doi.org/10.1002/oa.1157
Tomczyk W, Giersz M, Sołtysiak A, et al (2019) Patterns of camelid management in Wari Empire reconstructed using multiple stable isotope analysis: evidence from Castillo de Huarmey, northern coast of Peru. Archaeol Anthropol Sci 11:1307–1324. https://doi.org/10.1007/s12520-017-0590-6
Tung TA (2012) Violence, ritual, and the Wari Empire: A social bioarchaeology of imperialism in the Ancient Andes. University Press of Florida, Gainesville, FL
Vidal JP (2014) Las ocho regiones naturales del Perú. Terra Bras Rev Rede Bras História Geogr E Geogr Histórica. https://doi.org/10.4000/terrabrasilis.1027
Watanabe S (2012) Dinámicas sociopolíticas y continuidad cultural en la sierra norte peruana:un caso de estudio del Horizonte Medio Cajamarca. Bol Arqueol PUCP 105–130. https://doi.org/10.18800/boletindearqueologiapucp.201201.006
Weber SL (2019) Pulling Abundance Out of Thin Air: the Role of Camelid Pastoralism at 3000 B.P. Ph.D., Harvard University
Wheeler JC (1982) Aging Llamas and Alpacas by their Teeth. Llama World 12–17
Wickham H, François R, Henry L, et al (2023) dplyr: A Grammar of Data Manipulation
Williams PR (2001) Cerro Baúl: A Wari Center on the Tiwanaku Frontier. Lat Am Antiq 12:67–83. https://doi.org/10.2307/971758
Williams PR, Nash DJ (2021) Consuming Kero: Molle Beer and Wari social identity in Andean Peru. J Anthropol Archaeol 63:101327. https://doi.org/10.1016/j.jaa.2021.101327
Williams PR, Nash DJ, Cook AG, et al (2019) Wari Ceramic Production in the Heartland and Provinces. In: Glascock M, Neff H, Vaughn KJ (eds) Ceramics of the Indigenous Cultures of South America: Studies of Production and Exchange through Compositional Analysis. University of New Mexico Press, Albuquerque, pp 125–133
Wright L, Schwarcz H (1998) Stable carbon and oxygen isotopes in human tooth enamel: Identifying breastfeeding and weaning in prehistory. J Physiol Anthropol 106:1–18
Yann LT, DeSantis LRG, Koch PL, Lundelius EL (2016) Dietary ecology of Pleistocene camelids: Influences of climate, environment, and sympatric taxa. Palaeogeogr Palaeoclimatol Palaeoecol 461:389–400. https://doi.org/10.1016/j.palaeo.2016.08.036
Zimmer-Dauphinee J, Scaffidi BK, Tung TA (2020) Andean isoscapes: Creating and testing oxygen isoscape models to aid in the identification of missing persons in Peru. In: Parra RC, Zapico SC, Ubelaker DH (eds) Forensic Science and Humanitarian Action: Interacting with the Dead and the Living. Wiley, Hoboken, NJ 2019. “A Look to the North and South: Camelid Herding Strategies in the Desert Coast of Peru.” In Desert Archaeology, edited by Camila Alday and Sara Morrisset. Vol. 34.1. Archaeological Review from Cambridge. Cambridge, UK.

Table 1 is available in the Supplementary Files section.

No competing interests reported.

Download PDF

Editorial decision: Revision requested
06 Jan, 2026
Reviews received at journal
03 Jan, 2026
Reviews received at journal
27 Dec, 2025
Reviews received at journal
22 Dec, 2025
Reviewers agreed at journal
30 Nov, 2025
Reviewers agreed at journal
28 Nov, 2025
Reviewers agreed at journal
28 Nov, 2025
Reviewers agreed at journal
28 Nov, 2025
Reviewers invited by journal
28 Nov, 2025
Editor assigned by journal
13 Nov, 2025
Submission checks completed at journal
12 Nov, 2025
First submitted to journal
12 Nov, 2025

You are reading this latest preprint version

Human-Camelid Interactions in the Wari Hinterlands: A stable isotopic analysis of camelid remains from Auquimarca (600-1000 CE), Huancayo, Peru

Status:

Version 1

Abstract

Figures

1. Introduction

1.1 Cultural Influence of the Wari Empire

1.1.1 Wari influence in the Mantaro River Valley

2. Site and Environmental Background

2.1 Ecology and Agriculture of the Mantaro River Valley

2.2 The Site of Auquimarca

2.2.1 Human interments at Auquimarca

2.2.2 Middle Horizon sediment layer at Auquimarca

2.2.3 Faunal remains recovered at Auquimarca

2.2.4 Zooarchaeological Analysis of Camelid Remains

3. Stable Isotope Analysis of Camelid Pastoralism

3.1 Carbon Stable Isotope Analysis

3.2 Oxygen Stable Isotope Analysis

3.3 Stable Isotope Analysis of Middle Horizon Camelids

4. Materials and Methods

5. Results

5.1 δ¹³C and δ¹⁸O values by individual camelid

5.2 δ¹³C and δ¹⁸O values by tooth type

5.3 Intra-tooth variation in δ¹³C and δ¹⁸O values

5.4 Inter-site comparisons in the Wari world

6. Discussion

6.1 Dietary breadth among Auquimarca camelids

6.2 Andean mobility among Auquimarca camelids

7. Conclusion

Declarations

Author Contribution

Acknowledgement

References

Table 1

Additional Declarations

Supplementary Files

Status:

Version 1

Human-Camelid Interactions in the Wari Hinterlands: A stable isotopic analysis of camelid remains from Auquimarca (600-1000 CE), Huancayo, Peru

Status:

Version 1

Abstract

Figures

1. Introduction

1.1 Cultural Influence of the Wari Empire

1.1.1 Wari influence in the Mantaro River Valley

2. Site and Environmental Background

2.1 Ecology and Agriculture of the Mantaro River Valley

2.2 The Site of Auquimarca

2.2.1 Human interments at Auquimarca

2.2.2 Middle Horizon sediment layer at Auquimarca

2.2.3 Faunal remains recovered at Auquimarca

2.2.4 Zooarchaeological Analysis of Camelid Remains

3. Stable Isotope Analysis of Camelid Pastoralism

3.1 Carbon Stable Isotope Analysis

3.2 Oxygen Stable Isotope Analysis

3.3 Stable Isotope Analysis of Middle Horizon Camelids

4. Materials and Methods

5. Results

5.1 δ13C and δ18O values by individual camelid

5.2 δ13C and δ18O values by tooth type

5.3 Intra-tooth variation in δ13C and δ18O values

5.4 Inter-site comparisons in the Wari world

6. Discussion

6.1 Dietary breadth among Auquimarca camelids

6.2 Andean mobility among Auquimarca camelids

7. Conclusion

Declarations

Author Contribution

Acknowledgement

References

Table 1

Additional Declarations

Supplementary Files

Status:

Version 1

5.1 δ¹³C and δ¹⁸O values by individual camelid

5.2 δ¹³C and δ¹⁸O values by tooth type

5.3 Intra-tooth variation in δ¹³C and δ¹⁸O values