Abstract
The red fox (Vulpes vulpes) is a generalist, omnivorous predator that is often drawn to human environments, exploiting anthropogenic refuse. Foxes may have had little or significant economic importance for prehistoric human foragers, depending on the environmental, economic, and cultural context. Here we investigate human-fox interaction at the Late Holocene Uyak site (KOD-145) on Kodiak Island, Alaska. We apply zooarchaeological, taphonomic, and stable isotope analyses to the fox remains and find that complete animals were processed for meat and pelts and then discarded. Stable isotope results support foxes as omnivores eating in both the terrestrial and marine environments, and a comparison of archaeological and modern foxes show more dietary variability in ancient foxes. Together, these data suggest that the Uyak foxes were drawn to the village as a stable source of food subsidies, eating discarded marine and terrestrial resources, and consequently were embedded in human subsistence as sources of meat and raw materials.
The red fox (Vulpes vulpes) is a generalist, omnivorous predator that is often drawn to human environments to take advantage of their high food availability. As a result, there has been significant interest in the relationship between red foxes and humans in a variety of contexts, including urban areas (Contesse et al. 2003), rural farming communities (Jankowiak et al. 2008), fox farms (Rekiläa et al. 1997), and the process of fox domestication (Belyaev et al. 1985; Trut 1999). In general, this work demonstrates that foxes are well adapted to humanized environments, where they scavenge from human food waste and are frequently human prey.
Foxes may have had little or significant economic importance to prehistoric human foragers, depending on the environmental, economic, and cultural context (i.e., Yeshurun et al. 2009; Yeshurun and Zeder n.d.). In North America, foxes have a long history of exploitation and analysis suggests they were used for a variety of purposes. Red fox remains have been recovered in archaeo-logical sites from the Terminal Pleistocene (Lyman 2012) to the historical period (Etnier et al. 2016). Monchot and Gendron (2011) find that Dorset people in the eastern Arctic butchered and skinned arctic foxes (Vulpes lagopus), while Collins (1991) observes that the island fox (Urocyon littoralis) may have been used for ritual purposes in California’s Channel Islands. Modern studies suggest there is often a close dietary relationship between people and foxes that scavenge in human waste, and this relationship has also been suggested for ancient foxes (Hofman et al. 2016; West and France 2015).
The Kodiak Archipelago of Alaska (Fig. 1) is home to a substantial population of red foxes today, and they are found in large numbers in the archaeological record (Clark 1974; Hrdlicka 1944; Michael Etnier, personal communication); however, very little is known about their economic importance. Foxes are thought to be indigenous to the island, having survived in a glacial refugium or settling the islands in the early postglacial period (Clark 1958; Rausch 1969). The earliest evidence of foxes in the archaeological record is in the lowest levels of the Rice Ridge site, which date to before 5,000 years ago (Kopperl 2003, 2012). Despite the lack of older zooarchaeological evidence, it is likely that foxes have lived near or with people for at least 7,500 years when humans first colonized the Kodiak Archipelago. Today, red foxes are commonly harvested for their pelts and fox farming has a long history across the Gulf of Alaska (Itso 2012). However, while fox remains have been identified throughout the culture history sequence and across the archipelago (Clark 1970, 1974, 1982, 1998; Etnier et al. 2016; Hrdlička 1944; Kopperl 2003; Partlow 2003; West 2014; Yesner 1989), there has been relatively little detailed study of this mammal in a prehistoric context (but see West and France 2015).
Map of the Kodiak Archipelago, showing locations mentioned in the text. Map courtesy of the Alutiiq Museum and Archaeological Repository.
In this paper, we study fox remains from the Uyak site (KOD-145) using a combination of dietary reconstruction and zooarchaeological analyses. Using stable isotope analysis of skeletal remains (carbon [δ13C] and nitrogen [δ15N]), we compare ancient and modern fox diets to examine whether ancient foxes were potentially consuming human refuse. Given the massive middens present on the ancient landscape, we predict the fox dietary reconstruction will reflect more stable access to human food waste in the past. Second, we examine the taphonomic history of the skeletal remains to assess if and how these foxes were being used in an economic context, whether for food or fur. Together, these lines of evidence have the potential to illuminate the prehistoric economic role of the red fox on Kodiak Island.
Materials and Methods
Larsen Bay and the Uyak Site
Alutiiq people have lived in the Uyak Bay region of Kodiak Island for thousands of years, where they built villages at the coast, throughout the bay, and inland along rivers, streams, and lakeshores. These villages were substantial and numerous, and people stored food and dumped trash around their settlements in middens (Steffian and Saltonstall 2015). Despite its protected location, people living at the Uyak site would have had access to marine resources in the larger Uyak Bay, where marine mammals, marine and anadromous fish, birds, and shellfish are available. In contrast to this busy prehistoric landscape, today the majority of Kodiak’s population is concentrated in Kodiak City to the northeast and several smaller villages around the island. One of these is the village of Larsen Bay, which is located on the Larsen Bay inlet of larger Uyak Bay. Larsen Bay has a population of approximately 60 people, one of the oldest salmon canneries in the archipelago, and a number of sport fishing and hunting lodges. Unlike exposed prehistoric middens, garbage disposal is generally limited to town facilities and dumpsters to prevent scavenging by foxes and bears (Ursus arctos middendorffi).
The Uyak archaeological site is located at the mouth of Larsen Bay, where its residents had close access to the larger Uyak Bay and the open waters of Shelikof Strait. This site was excavated by physical anthropologist Aleš Hrdlička from 1933–1936, where he found a late Holocene village and a massive midden that formed a landmass in Larsen Bay (Hrdlička 1944). The excavated deposits were at least three meters deep, and these were divided into three layers (lower, middle, and upper) based on gross stratigraphic changes and artifact types. Hrdlička’s main goal was to collect human remains, and his methods were lacking: there was very little stratigraphic control, the material was not screened, and only whole or mostly whole animal bones were collected (Hrdlička 1944; Hrdlička in Allen 1939).
Hrdlička (1944:325) estimated the deposits to be approximately 2,000 years old, which has been supported by later reanalysis of his work (Heizer 1956) and new excavation (Steffian 1992), as well as a series of radiocarbon dates taken directly on dog bone (West and Jarvis 2012). These dates confirm the site was occupied in the period from 2000 to 500 cal BP. The three occupation layers defined by Hrdlička (1944) have been assigned to the Kachemak (2700–900 cal BP) and Koniag (900–200 cal BP) phases of the Kodiak culture history chronology, based on artifacts and the recent radiocarbon dates. Later excavations by Amy Steffian (1992) revealed a series of late Kachemak house structures with food storage both inside and outside in the form of pits.
Zooarchaeology and Taphonomy
The faunal material collected by Hrdlička in the 1930s is curated at the Smithsonian Institution and includes osteological remains of marine mammals, dogs, and foxes, with small numbers of other terrestrial carnivores. Using the Smithsonian’s modern mammal collections as reference, we separated all fox specimens and identified them to body part. The cranium and mandible of small domestic dogs are easily differentiated from foxes: fox crania have a longer and more slender appearance, while dog crania are deeper and broader and show tooth crowding (Wayne 1986a). In addition, limb bones of small dogs are broader than fox bones, even when they are the same length (Wayne 1986b).
Given the field-collection biases, we focused our analyses on the cranium, mandible, humerus, radius, ulna, femur, and tibia. These are the largest and most conspicuous bones in the fox skeleton, making them less susceptible to collection biases compared to smaller elements (e.g., carpals, tarsals, and phalanges) or more fragile ones (e.g., vertebrae). Most of our analyses use straightforward specimen counts (Number of Identified Specimens [NISP]). The skeletal-element profile was constructed using the Minimum Animal Unit (MAU) to standardize skeletal elements that have different abundances in the complete body (Binford 1981).
To obtain detailed demographic data we carried out a series of measurements and qualitative observations of the fox remains. The greatest length (GL), proximal breadth (BP), and distal breadth (BD) measurements were taken for all limb bones according to Driesch (1976). We took the following skull measurements (numbers in parentheses refer to Driesch’s [1976] measurements for the canid skull): condylobasal length (#2), length of cheek tooth row (#15), zygomatic breadth (#30), greatest neurocranium breadth (#29), breadth at postorbital constriction (least breadth of skull) (#31), skull height, basioccipital to saggital crest (#38), and length (L) and breadth (B) measurements of the M2, M1, and P4 teeth.
To determine the age of each animal, we noted bone fusion on every limb bone and interpreted these observations using Harris (1978). We recorded skull age by the state of closure of three sutures, following Churcher (1960): the presphenoid-basisphenoid (PSBS), basioccipital-basisphenoid (BOBS), and the lateral palatal portion of the premaxillary-maxillary suture (PMMS) lying between the anterior palatine foramen and the lingual margin of the alveolus of the upper canine. Churcher (1960) provides the timing of suture closure, based on a large sample of known-age farm foxes. Considering this suture-closure sequence, we were able to assign a year-class or a range of years to each skull.
A detailed study of bone-surface modifications was conducted on a sample of the first thirty specimens of each skeletal element (humerus, femur, radius, tibia, mandible, and skull). During the analysis, specimens were picked from the collection and recorded in the datasheet in random order, and hence selecting the first 30 specimens constituted a random and representative sampling of the very large assemblage. Additionally, all pathological specimens were examined. The specimens were systematically examined using a stereoscopic microscope (Olympus SZX7) with a high-intensity oblique light source, at 7–40× magnification, following the procedure described in Blumenschine et al. (1996; see Yeshurun et al. 2014 for details and references). We searched for cut marks and interpreted them according to modern butchery experiments that recorded cut marks on fox carcasses (Val and Mallye 2011, and personal observation). Butchery marks corresponding to skinning activities usually occurred on the extremities (head and sometimes lower limb elements) while disarticulation and filleting marks correspond with various bone-ends and with muscle-bearing locations of the upper limb bones, respectively. We looked for burning, bone-working, carnivore punctures, scoring, and digestion marks (Binford 1981), as well as rodent gnaw marks (Brain, 1981) and biochemical (root) marks (Dominguez-Rodrigo and Barba 2006). Trampling striations (Dominguez-Rodrigo et al. 2009) and abrasion of bone edges (Shipman and Rose, 1988) were sought, and weathering was noted (Behrens-meyer 1978).
Stable Isotopes
The stable isotopes of carbon (12C, 13C) and nitrogen (14N, 15N) are useful for dietary reconstruction because they fractionate at each level of the food web, and animal body tissues become enriched in the heavier isotope at higher trophic levels. The δ15N and δ13C values of bone collagen are particularly useful for distinguishing between terrestrial and marine contributions to the diet in coastal environments. In marine ecosystems, the food web is long and complex, and marine consumers show significantly more positive δ15N and δ13C values than consumers in a terrestrial food web (Schoeninger and DeNiro 1984). Animals whose diets derive from marine protein generally produce δ15Ncollagen values between +10 and +20 ‰ and δ13Ccollagen values between –20 and –10 ‰, while terrestrial diets produce δ15Ncollagen values between 0 and +10 ‰ and δ13Ccollagen values between –30 and –17 ‰ (DeNiro and Epstein 1981; Schoeninger and DeNiro 1984). In general, these values are enriched at each trophic level: carbon enriches by ~1 ‰ and nitrogen by ~3–4 ‰. Results are reported in standard delta notation relating the ratio of heavy to light isotopes compared to an international standard:
δX = [(Rsample – Rstandard)/Rstandard] * 1000
where X is the system of interest (i.e., 13C or 15N), R is the isotope ratio (i.e., 13C/12C or 15N/14N), units are per mil (‰), and the standards are V-PDB and atmospheric air for carbon and nitrogen respectively.
West and France (2015) performed the stable isotope analysis of the archaeological sample (n=20), which includes left mandibles from all three strata of the Uyak site, or the period from approximately 2000–500 cal BP. The modern red fox (n=11) were harvested for this project by the Pingree family at the Quartz Creek Lodge in Uganik Bay in the fall of 2010 (Fig. 1). To be consistent with the ancient samples, we selected left mandibles, which were cleaned in the field and defleshed at the Smithsonian Institution’s Dermestid beetle colony. In both cases, fully adult animals were selected based on full tooth eruption following Hillson (2005) to avoid the effects of nursing or weaning on stable isotope values in juvenile specimens (Fogel et al. 1989; Katzenberg et al. 1996).
Sampling methods for the modern foxes follow West and France’s (2015:527) sampling strategy for the ancient foxes: bone samples (~500 mg) were removed from the mandibles using a rotary tool and coring bit. Prior to collagen extraction, the samples were soaked in a 1:1 methanol-chloroform mixture to remove lipids. Collagen was extracted using a modified Longin (1971) method. Samples were sonicated in ultrapure water to remove sediments and labile salts, and rinsed. Each sample was decalcified in 0.6M HCl at 4°C for 24-hour increments (acid replaced daily) until reaction ceased, then rinsed to neutrality and soaked in 0.125M NaOH for 24 hours at room temperature to remove humic and fulvic contaminants. The resulting crude gelatin was rinsed and reacted in 0.03M HCl at 95°C for 24 hours to denature the protein. The resulting supernatant was freeze-dried to produce a purified collagen extract.
The extracted collagen samples were analyzed at the Smithsonian Institution OUSS/MCI Stable Isotope Mass Spectrometry Facility. Samples were weighed into tin cups and combusted in a Costech 4010 Elemental Analyzer (EA) with a zero-blank autosampler. The resulting N2 and CO 2gases were analyzed for δ15N and δ13C values in a Thermo Delta V Advantage mass spectrometer coupled to a Conflo IV interface. Raw data were linearly corrected to a calibrated acetanilide and urea standard. All δ15N and δ13C values are reported with an error of ±0.2 ‰ (1σ), which is based on the reproducibility of repeated standard and sample measurements. The atomic C:N ratios of the collagen were calculated based on calibration to a homogeneous acetanilide standard. C:N ratios between 2.8 and 3.6 indicate good preservation for isotopic analysis (Ambrose 1990; DeNiro 1985). Samples with values outside of this range were eliminated from the analysis.
Finally, the δ13C of an animal is influenced by the δ13C of the atmospheric CO2, which has changed significantly in the last 150 years (Francey et al. 1999). Therefore, to compare temporal changes in fox diet, the δ13C values of each modern fox specimen were adjusted by +1.1 ‰ following Newsome et al. (2007) for the North Pacific. Adjusted values are presented in Fig. 5, and both unadjusted raw data and adjusted values are presented in Table 4.
Results
Zooarchaeology and Taphonomy
Hrdlička’s Uyak site collection yielded 1,647 fox skeletal elements, representing a minimum number of 189 individual animals based on the number of complete right mandibles. The majority of the specimens (n=1,583, 96%) are complete or nearly so (Fig. 2). The results of the osteometry study point to a large-bodied fox population and confirm the taxonomic identification as red fox. Kodiak is within the known range of the red fox but is very close to the present-day distribution of the arctic fox (Vulpes lagopus) (IUCN 2014). According to Monchot and Gendron (2010), the humerus and tibia GL measurements are significantly different for red and arctic foxes, so we compared our archaeological samples to modern female and male red and arctic foxes using Student’s t-test (Table 1). The Uyak humeri and tibiae lengths are significantly larger than modern arctic fox measurements. They are also larger than modern female red foxes, while they are similar to (humerus) or smaller than (tibia) the modern red fox males. Additionally, our dental measurements indicate large-bodied foxes: the lower M1 maximum mesial-distal length ranges between 14.0–18.5 mm (n=144, mean=16.46, SD=0.84) and the maximum buccal-lingual width ranges between 5.2–8.1 mm (n=144, mean=6.91, SD=0.47). These measurements mostly fall within the range for red fox and do not overlap with the arctic fox range in Lyman’s (2012:Fig. 4) study. Thus, while the presence of some arctic foxes in the Uyak sample cannot be ruled out completely, it seems that the overwhelming majority of the remains belong to red fox.
The Uyak site fox remains spread on the lab table during analysis. Note the completeness of most specimens.
Comparison of the Uyak archaeological specimens against modern female and male red (Vulpes vulpes) and arctic (Vulpes lagopus) fox using humerus (HU) and tibia (TI) Greatest Length (GL) measurements. Modern data after Monchot and Gendron (2010:Table 3).
Because the field collection was undoubtedly biased towards complete specimens, we opted for obtaining a coarse skeletal-element profile using the conspicuous cranium, mandible and larger limb bones. We divided the paired elements by two to obtain Minimum Animal Unit (MAU) values for each skeletal element; the results are shown in Table 2 and Fig. 3 and indicate the mandible and humerus display the highest values, the radius and femur are well represented, and the skull, ulna, and tibia show lower representation.
Skeletal elements represented in the Uyak fox assemblage.
Bone surfaces are very well preserved, enabling the taphonomic study of 216 specimens (Table 3). Only eight specimens (3.7%) displayed considerable weathering (Behrensmeyer’s Stage 3), and bone abrasion and rodent gnawing were low (<3%). However, a considerable number of specimens yielded trampling striations and root marks, probably because of some bone movement and biochemical activity, respectively, which occurred in the Uyak midden. Burned specimens were absent, but human butchery of foxes is strongly attested by cut marks that appear on one-quarter of the studied specimens (n=53; Fig. 4). The marks appear on all skeletal elements, but their frequency on the body extremity (skull) and meat-bearing upper limb bones (humerus and femur) are especially marked (Table 3). Also, a considerable number of specimens display carnivore-gnawing marks of some kind (n=26; 12%), including punctures, pits, and crenelated edges.
Taphonomic properties of the Uyak fox assemblage.
Juvenile foxes are rare in the assemblage. Only 25 limb bones out of 1,203 (2.0%) had both epiphyses unfused, while 15 bones (1.2%) displayed an unfused proximal end and 12 (1.0%) had an unfused distal end. Skulls, which were aged by the closure of the sutures, yielded a similar picture in that only three skulls out of 82 (3.7%) belonged to Churcher’s (1960) “Year 1” class (during which limb epiphyseal fusion also occurs). The majority of skulls were assigned to Year 2 (n=39) or Year 3 and older (n=40).
Finally, a remarkable number of specimens (n=42) displayed pathologies, defined as excess bone growth, extremely worn or absent teeth, or distorted epiphyses. While a detailed assessment of the pathologies is beyond the scope of this paper, the pathologies warrant further study to assess their cause; more comparative studies are needed to assess whether this was a common occurrence for Alaskan mesocarnivores.
Stable Isotopes
The results of stable isotope analysis are presented in Table 4 and Fig. 5, showing the difference in distribution between modern and archaeological red fox samples. As described above, C:N ratios between 2.8 and 3.6 suggest collagen is viable for isotopic analysis. Based on this criterion, ten of the 20 archaeological samples and nine of 13 modern foxes produced viable collagen. Those with good preservation had a C:N value range of 3.3–3.6. The isotope results show the following ranges: modern fox δ15N = +8.7 to +14.7 ‰; archaeological fox δ15N = +7.1 to +16.7 ‰; modern fox δ13C = –15.3 to –20.3 ‰ (adjusted to –14.2 to –19.2); archaeological fox δ13C = –11.9 to –20.9 ‰ (Table 4, Fig. 5). The error reported is ±0.2 ‰ (1σ).
These results show some difference between archaeological and modern δ15N and δ13C values and produced values along a gradient from terrestrial to marine diets. Visually and statistically, as shown in Fig. 5, the archaeological foxes show more variability (1σ = 3.4 and 2.9 for δ15N and δ13C, respectively) than the modern foxes (1σ = 1.8 and 1.4 for δ15N and δ13C, respectively). F-tests for equality of variance are significant for both nitrogen (F9,8 = 3.63, p < 0.05) and carbon, using adjusted modern values (F9,8 = 4.48, p < 0.05).
Carbon and nitrogen isotope values for the Uyak red fox (Vulpes vulpes).
Discussion and Conclusions
Our results provide a detailed picture of human exploitation of foxes in a Late Holocene village on Kodiak Island. The Uyak midden yielded numerous red fox remains in cultural context, associated with human food and other refuse (Hrdlička 1944). This is one of the largest archaeological fox assemblages in the Gulf of Alaska and the Northwest Coast, and its magnitude is even more impressive considering the incomplete collection of the material in the field. Despite the collection methods, subjecting this legacy collection to a comprehensive zooarchaeological study allowed us to extract valuable information.
Our zooarchaeological and taphonomic analyses demonstrate that fairly complete fox carcasses were deposited in the Uyak midden. The MAU values displayed in Fig. 3 suggest that the mandible and humerus have the highest relative abundance, followed by the radius and femur; the skull, ulna, and tibia have lower relative abundances. Notwithstanding the poor collection methods, this profile points to deposition of whole carcasses consisting of both meat-poor elements (e.g., skull) and meat-rich elements (e.g., upper limbs).
Skeletal element profile of the foxes, based on Minimum Animal Unit (MAU).
Based on the analysis of cut marks, humans skinned, disarticulated and defleshed complete fox carcasses and then discarded them in the midden (Table 3 and Fig. 4). Skinning activities tend to leave marks on the head and lower limb elements, while disarticulation and filleting may leave marks on the ends of muscle-bearing limb bones. We observed cut marks on the cranium/mandible, lower forelimb (radius/ulna), and lower hind limb (tibia) that suggest the foxes were skinned (n=28), while the upper forelimb (humerus) and hind limb (femur) show evidence of both disarticulation (n=11) and filleting (n=10; four additional cut marks were equivocal with respect to function). Thus, all stages of butchery are represented in the midden, and it can be ascertained that foxes were used for their meat, and likely their furs as well. After deposition, the remains underwent some carnivore ravaging, most likely by the flourishing community of village dogs (West and Jarvis 2012). The mild degree of ravaging may indicate that foxes were discarded after processing with little or no meat left on the bones. Some postdiscard movement and exposure to the elements affected the bones to a small degree, resulting in a large and well-preserved assemblage.
Examples of butchery marks on fox humeri in the Uyak assemblage: (a, b) Sample #16, disarticulation marks on the distal end ; (c, d) Sample #30, filleting marks on the diaphysis; (e, f) Sample #29, disarticulation marks on the distal end. Black square show area of magnification.
While archaeological fox assemblages across the Gulf of Alaska and the Northwest Coast tend to be small, foxes are relatively abundant in sites on Kodiak, which suggests they were an important resource (Clark 1974). Clark (1974) made an informal analysis of fox bones from sites on Kodiak and his results are similar to this study: Clark argues the bones are generally from adult animals and many of them exhibit cut marks, which he interprets as pelt harvesting. However, he suggests burning on some of the bones may be the result of fox butchery and meat consumption. Looking to the ethno-historical, ethnographic, and oral history records of fox use in the archipelago offer little insight into these practices. Oral histories from the Kodiak Archipelago make brief references to foxes (Bergsland and Dirks 1990; Golder 1903), and it is clear that foxes were harvested for their pelts during the Russian period (Black 1977; Clark 1974). In terms of noneconomic uses, there has been a suggestion that people on Kodiak were observed keeping foxes at pets (Huggins 1981), and Hrdlička (1944) writes that he excavated foxes in association with a human burial. However, the association between the human burial and the fox remains cannot be confirmed because of Hrdlička’s lack of strati-graphic control and the absence of illustrations or photographs, but also because it was common practice for past people to mix midden material with burials in other areas of the Uyak site (Steffian 1992).
For comparison, several very large archaeological fox assemblages in cultural contexts are known in the North American Arctic from Greenland to the Mackenzie Delta (Monchot and Gendron 2011; Novecosky and Popkin 2005). Some of these assemblages contain thousands of fox remains, both red and arctic (Monchot and Gendron 2011:Table 1). A comprehensive zooarchaeological and taphonomic study of the Late Dorset site of Tayara in Nunavik found similar patterns of butchery, deposition, and ravaging by dogs, stressing the important role of foxes in the economy of some Arctic groups (Monchot and Gendron 2011).
Furthermore, our results shed light on the life history of foxes prior to their capture by humans. The discarded foxes were mainly adults, either at the end of their first year of life or older, some suffering from pathological lesions on the bones, which is an aspect which should be explored further. The isotopic values of the archaeological bones displayed a variable diet compared to recent fox remains from the island, which includes contributions from both terrestrial and marine resources (Fig. 5). West and France (2015) argue that the archaeological foxes show isotopic variability, with some animals eating exclusively terrestrial protein (samples 560028-32, 37, 42); some foxes eating almost exclusively marine protein, similar to dogs and humans (560028-35, 43); and the rest of the foxes consuming a mixed diet (Table 4). Nonetheless, while many of the modern foxes analyzed for this study show a mixed diet, there is generally less variability and overall less contribution of food from a high trophic level or the marine environment.
Carbon and nitrogen isotope values for the archaeological Vulpes vulpes samples analyzed by West and France (2015) and the modern Vulpes vulpes samples analyzed for this study.
It is possible that people hunted for or trapped foxes at some distance from the site and, therefore, captured foxes with variable dietary signatures. These animals tend to move predictably and could be captured with snares at dens, eating sites, trails, and other familiar locations, and historically were harvested in the fall when their pelts were in prime condition (Davydov 1977; Steffian et al. 2015). While capture of foxes away from the site cannot be ruled out, the complete skeletal representation and heavy exploitation of these small animals (which provide very modest meat and fat yields), coupled with their high numbers and variable diet, may indicate that a good proportion of the foxes were actually captured opportunistically by the villagers as the foxes approached to scavenge food. The dietary data support this interpretation: the marine component of the archaeological fox diet may have derived from stored or discarded marine mammal and fish remains (West and France 2015). Moreover, the terrestrial resources could have come from the population of voles that probably fed on the midden as well, and frequently dominate fox diet (Haltenorth and Roth 1968). Comparison with the modern foxes lends support to this idea, as the modern fox remains were collected across Uganik Bay, where today very little human refuse is available for consumption.
Foxes are known to be commensal and feed on human trash, which is clear from empirical studies of modern foxes (Bino et al. 2010). Similar models come from the Near Eastern Epipaleolithic-Neolithic transition, where rising fox proportions combined with taphonomic evidence for exploitation and other evidence for symbolic roles, coincide with the first sedentary settlements (Yeshurun et al. 2009; Yeshurun and Zeder n.d.). Given this broader context of fox-human interactions, the Uyak results suggest that—in the context of an island with impoverished terrestrial resources—foxes at the Uyak site may have been commensal. They would have focused on the persistent refuse dump of the village that served as a predictable and abundant food supply source, and they were consequently integrated into the human economy.
Acknowledgments
We thank M. Zeder and T. Rick for their great help in the course of this research and for facilitating our access to the collections, C. France for providing access to the Smithsonian Institution’s Stable Isotope Lab, and to C. Hofman for analytical support. A. Edmondson assisted with the osteometric data collection. We thank A. Steffian and the staff at the Alutiiq Museum for their guidance, and the people of Larsen Bay, Alaska for their continued support of archaeology. Thanks to the Pingree Family of Quartz Creek Lodge and L. VanDaele for providing the modern samples. This research was supported by the Smithsonian Institution’s Office of Fellowships funding for CFW and the National Museum of Natural History’s Peter Buck Postdoctoral Fellowship to RY.
