Journal of Anthropological Archaeology 36 (2014) 1-129

Contents lists available at ScienceDirect

Journal of Anthropological Archaeology

journal homepage: www.elsevier.com/locate/jaa
ELSEVIE
R

Small scale camelid husbandry on the north coast of Peru (Virú

Valley): Insight from stable isotope analysis CrossMa
rk
Paul Szpaka* Jean-Fran^ois Millairea, Christine D. Whitea, Fred J. Longstaffeb
Department of Anthropology, The University of Western Ontario, London, ON N6A 5C2, Cañada bDepartment ofEarth Sciences, The University of Western Ontario, London,
a

ON N6A 5B7, Canada

A R T I C L E I N F Q
A B S T R A C
T
This study
Article presents carbon and nitrogen isotopic compositions for bone collagen and serially sampled hair from a large collection of South American camelids from two Early Intermediate
history:
Period
Received(c.10200 BC 2013
August to AD 800) sites (Huaca Gallinazo and Huaca Santa Clara) in the Virú Valley (north coast of Peru). The iso topic compositions of these camelids are consistent with plant
Revision
isotopic
received compositions
9 August 2014 from coastal and low altitude settings, but not from high altitude environments, suggesting that at least some of these ani- mals were raised locally. We present several
methodological approaches with respect to the treatment of isotopic data from archaeological contexts, outlining quantitative approaches that can provide consider able insight into isotopic
variation (within groups, within individuals, between groups, between individ- uals), as well as temporal variation in isotopic compositions in incrementally growing tissues. We contend that
focusing
Keywords: explicitly on variation in animal life histories has the greatest potential with respect to better understanding human-animal interactions in the past. The results demonstrate a large
amount of isotopic variability among individuals and an inconsistent amount of within-individual vari- ation, with no consistent shift in the diet leading up to the time of death for a group of
Stable isotopes
animals
Isotopic from a single ritual event. This result suggests that camelid husbandry in the Virú Valley was a small-scale activ- ity, with groups of camelids being managed by families or other small
variation
social units. Animals were likely kept primarily in close association to human habitation sites and provided with a diverse array of fodder. These prolonged interactions, occurring at a limited
Camelids
spatial scale, would have allowed a high degree of mutual familiarity to develop between humans and animals. Isotopic compositions for late Middle Hori- zon (c. AD 1100) sacrificed llamas
Human-animal interactions
from Huaca Santa Clara are consistent with Early Intermediate Period camelids, suggesting temporal stability in this small-scale camelid management strategy on the coast, which was
Animal management Animal
fundamentally different from camelid herding in the pastures of the Andean highlands. Isoto- pic analysis of prehistoric livestock has great potential with respect to better understanding
husbandry Peru
animal hus- bandry practices and human-animal interactions in the broadest sense because the data provide insight into the ways in which animals lived, rather than the manner in which
Zooarchaeology
they died. The variation-centered methodologies outlined in this paper provide a framework with which to approach some of these issues, highlighting the significance of understanding
variability in livestock life histories.
© 2014 Elsevier Inc. All rights reserved.

1. Introduction

The interaction between humans and non-human animals is a topic of

immense importance in anthropology. Shipman (2010) has proposed that the
manner in which humans interact with other animals (the ‘animal connection’)
can be placed alongside tool making, symbolism, and language as behaviors that
define humans as a species. Until very recently, anthropologists and
archaeologists
have viewed animals primarily through two lenses: subsistence and symbolism
(Mullin, 1999; Shanklin, 1985). In recent years, there has been increased
interest in the study of human-animal interactions that attempt to move beyond
the symbolic importance of animals wherein they are passive reflections of
human society (that animals are good to think with, following Lévi-Strauss,
1962), and instead sees animals as active agents that are part of human society
http://dx.doi.org/10.1016/
(Knight, 2005) — in other words, that animals are good to live with (following
jjaa.2014.08.005 0278-4165/© 2014
Haraway, 2008). As part of this turn in human-animal studies, there has been a
Elsevier Inc. All rights reserved.
call to move away from strictly anthropocentric studies of animals, and towards
what Kohn (2007, p. 4) has called an “anthropology of life’’, one that is

* Corresponding author. Present address: Department of Anthropology, University of British Columbia, Vancouver, BC V 6T 1Z1, Canada.
E-mail address: paul.szpak@gmail.com (P. Szpak).
 2 P. Szpak et al./Journal of Anthropological Archaeology 36 (2014) 110-129

explicitly concerned with human entanglements with other living beings, how sizes are robust, they offer a tangible means with which to reconstruct the
they interact with, shape, and are shaped by one another, within larger cultural, nature of and variation in how individual animals lived and how their lives were
economic, and political contexts (Kirksey and Helmreich, 2010). shaped by the humans with whom their lives were entangled. There has been
An important aspect of human-animal studies relates to scale. Most focus on some recognition of individual variation in companion animals in archaeological
various aggregations of animals (herds, entire spe- cies, or other folk/taxonomic contexts, with a number of studies examining the differential treatment of dogs
classifications), although several authors have highlighted the importance of in mortuary contexts (e.g. Byrd et al., 2013; Losey et al., 2011; Prummel, 2006).
accounting for the lived experiences of individual animals (Alger and Alger, 2003; A number of studies have examined the importance of livestock, particularly in
Argent, 2010; Bear, 2011). The recognition of the very individualized nature of ritual and mortuary settings (Goepfert, 2012; Russell and During, 2006; Salmi et
human-animal relationships has primarily been discussed within the context of al., 2011), although this research tends to be largely qual- itative and focused on
companion (Haraway, 2003, 2006, 2008; Power, 2008) and working animals symbolic aspects of animals (but see Whittle, 2003). Relatively little attention has
(Hart, 2005; Lawrence, 1985). These approaches have not been extensively been paid to the importance and meaning of individual variation in livestock life-
applied to livestock (but see Abbink, 2003; Dwyer and Minnegal, 2005). ways in archaeological contexts and what this might mean in terms of human-
Regarding live- stock, however, Knight (2005, p. 5) points out that ‘‘a animal interactions (but see deFrance (2010) for a paleopathological example).
preoccupation with the outcome of the relationship (slaughter for meat) is apt to
conceal the protracted relationship of nurturance and care that precedes it.’’
Traditional zooarchaeological research tends to empha- size this aspect of 2. South American camelids
human-animal relationships: slaughter, butchery, and the incorporation of the
The South American camelids (hereafter simply camelids) include two
carcass into the archaeolog- ical record. These activities represent a very limited
domestic (llama and alpaca) and two wild species (vicuña and guanaco). The
number of interactions that occur within the context of a much larger and more
differentiation of camelid species on the basis of postcranial skeletal morphology
complex relationship between humans and livestock. This is not to suggest that
is very difficult; accord- ingly, throughout this paper, discussion focuses
the economic role of animals be discounted, but it must be recognized that there
generally on ‘camelids’, except in cases where dental and/or soft tissue preser-
may be a significant ontological distinction between a living animal and an
vation allow for the assignation of individual animals to a particular species
animal carcass, whereby the treatment of the latter does not necessarily reflect
(Wheeler et al., 1995).
the human- animal inter-subjectivity prior to the animal’s death (Herva and
It is widely recognized that camelids were of tremendous eco- nomic, social,
Salmi, 2010). If the ultimate goal of zooarchaeological research is to better
political, and ritual significance to various groups throughout the prehispanic
understand the interactions between humans and other animal species (Reitz
Andes (Bonavia, 2008). Views of camelids and camelid herding in the region have
and Wing, 2008), and we accept the premise that the best manner in which to do
been largely shaped by ethnographic and ethnohistoric accounts of camelid
so for agropastoral societies is to focus on lived interactions between humans
pastoralism, which have been limited to high altitude pasturelands (e.g. Flores-
and animals (Knight, 2012), a focus on slaughter, butchery, and carcass disposal
Ochoa, 1979; Murra, 1965; Tomka, 1992). The introduction of European
falls short. It is, however, not sufficient to simply layer theoretical perspectives
domesticates (e.g. cattle, sheep, pig, donkey) begin- ning in the sixteenth century
concerning human-animal interactions atop data that do not speak directly to
drastically reduced the geographic range and number of camelids in the Andes
these issues. Instead, methodologies that provide insight into the interactions
(Bonavia, 2008) and several authors have suggested that camelids were herded
between living animals and humans are necessary to complement traditional
in a much wider variety of environments prior to the arrival of the Spanish
lines of evidence and begin to move beyond the study of the roles of animals in
(Dufour et al., 2014; Goepfert, 2012; Goepfert et al., 2013; Thornton et al.,
pre- historic subsistence economies (economic reductionism) and their discursive
2011), including the arid coastal region of northern Peru (Shimada and Shimada,
representations (symbolic reductionism) to one that focuses on the nature of
1985). Many questions remain, however, regarding the nature of camelid herding
lived interspecies entanglements.
outside of the high altitude zones of the Andes.
Difficulties arise in archaeological contexts, however, where human-animal
Traditional accounts of camelid husbandry in the Andean region have
interactions cannot be observed directly, and the basis for interpretation
discussed fairly large herds that graze on high altitude pas- turelands in the
necessarily starts with the remains of the animal carcass and the context
puna (3800-4800 masl) or high sierra (25003800 masl) (Flannery et al., 1989;
associated with its disposal. To cir- cumvent this problem, we must look to
Flores-Ochoa, 1979). Today, herds are of mixed composition, typically consisting
indirect evidence or ‘traces’ to better understand animal lives (for a modern, non-
of alpacas, sheep, goats, and in some cases cattle (at lower altitudes). Herding
isotopic example see Hinchliffe et al., 2005). Stable isotope analysis of animal
tissues is one of several techniques that provide insight into various aspects of
animal life histories. Because certain tissues (teeth, hair, nail, whisker) grow at
discrete intervals or continu- ously, diachronic isotopic analyses of these tissues
can provide high-resolution life histories of individuals, reflecting temporal shifts
in diet, residence, and potentially health (Balasse et al., 2001; Knudson et al.,
2007; White et al., 2009). Within the context of bioarchaeology, most analyses
have been concerned with issues at the population or regional level, but a more
detailed under- standing of larger social processes may be reached through a
con- certed focus on individuals as well as populations (Knudson and
Stojanowski, 2008). The same logic applies in zooarchaeological studies that seek
to better understand the interactions between humans and domestic animals,
and it is important to recognize that relationships formed between humans and
animals may take on a very individualized nature (Alger and Alger, 2003; Argent,
2010; Haraway, 2003, 2006, 2008; Power, 2008). Somewhat analo- gously, there
has been a recent trend in ecological research to rec- ognize the importance of
variation at the individual level (e.g. foraging specializations), which has been
addressed via isotopic analysis in a number of studies (Cherel et al., 2007;
Matich et al., 2011; Newsome et al., 2009; Szpak et al., 2012c). This has led to
the development of several interpretive techniques, which have not been
employed in archaeological contexts but can be used to assess and compare
variation both between and within groups and/or individuals (e.g. Jackson et al.,
2011; Layman et al., 2007; Martínez del Rio et al., 2009). In the context of social
zooarchaeol- ogy (Marciniak, 2005; Oma and Hedeager, 2010; Russell, 2012)
these techniques have tremendous potential because, provided that samples
P. Szpak et al./Journal of Anthropological Archaeology 36 (2014) 110-129
 4 P. Szpak et al./Journal of Anthropological Archaeology 36 (2014) 110-129
Valley (Fig. 1A). The culture history of the Peruvian north coast was extremely
dynamic, particularly during the EIP, and is only begin- ning to be understood in
light of recent excavations, re-evaluations of ceramic chronologies, and
radiocarbon dating programs (Millaire and Morlion, 2009; Millaire, 2010b).
During the EIP in Virú, there were significant increases in population size and
agri- cultural productivity, a four-tiered settlement hierarchy, expansion of large-
scale irrigation networks, the emergence of urban settle- ments, and a unified
political command over the entire Virú Valley (Fogel, 1993; Millaire, 2010b;
Willey, 1953) with its capital at the Gallinazo Group (Millaire and Eastaugh,
2011). On these bases, the Virú polity appears to have developed the political
structure asso- ciated with archaic states (Fogel, 1993; Millaire, 2010b).
The Virú polity is associated with a distinctive resist-painted ceramic style
(previously known as Gallinazo Negative), which is largely restricted to the Virú
Valley (Fig. 2). Previously, this style was treated as one part of a larger ceramic
tradition (including Gallinazo domestic wares that were very widely distributed
on the Peruvian north coast) and was considered to be characteristic of the
‘Gallinazo culture’, approximately contemporary with Moche material culture
(Bennett, 1950; Fogel, 1993; Strong and Evans, 1952). In light of recent
reappraisals of the nature of this material
Castillo Incised and Modeled Vessels (Gallinazo)
Moche-style Vessels
Viru Negative Vessels
Fig. 2. Representations of ceramics typical of the styles discussed in the text. Photographs courtesy of Museo Larco, Catalog Numbers from left to right: top row (ML016110, ML016251,
ML018016), middle row (ML000566, ML000002, ML002335), bottom row (ML018888, ML010467, ML016321).
culture (Millaire and Morlion, 2009), we use the term Virú to refer to
the political entity associated with the Virú Negative material
culture (sensu Larco Hoyle, 1945). In this schema, Virú is one of
several EIP political entities on the north coast roughly
contempo- raneous with Moche, and the Gallinazo incised
and appliqued domestic wares do not correspond with any
single political entity or cultural group, but are simply a
regional domestic ceramic tradition.
The first site from which material was collected, Huaca Gallinazo (V-59), is
the largest of the central group of mounds that comprise the Gallinazo Group,
covering over 40 ha and located c. 5 km from the coast (Fig. 1A). It is estimated
that the population of the Gallinazo Group during the EIP may have been
between 10,000 and 14,400 (Millaire and Eastaugh, 2011). Several authors have
proposed that the Gallinazo Group functioned as the capital or central
administrative center for a political entity of varying sizes - depending largely
upon the interpretation of ceramic styles (Bennett, 1950; Fogel, 1993; Millaire,
2010b). Intensive excava- tions were carried out at the Gallinazo Group in the
middle part of the twentieth century (Bennett, 1950; Strong and Evans, 1952),
revealing a long history of occupation. More recent excava- tions and associated
radiometric dating have demonstrated that
 P. Szpak et al./Journal of Anthropological Archaeology 36 (2014) 110-129
the occupation of the main residential sector at the site dates back to at least 50 tissues reflect a weighted average of the foods consumed during the period when
BC and may have continued for seven centuries (Millaire, 2010b). The most the tissues formed (DeNiro and Epstein, 1978, 1981). In terrestrial ecosystems,
salient structure at Huaca Gallinazo is a very large civic building, which was d13C and d15N values at the base of the food web (in plants and soils) vary with
likely constructed around AD 50 (Millaire, 2010b). Camelid material sampled respect to a number of environmental parameters (reviewed by Craine et al.,
from this site is derived from fills associated with two periods, the first (n = 13) 2009; Hobbie and Ouimette, 2009; Hobbie and Hogberg, 2012; Hogberg,
corresponding to c. 50 BC to AD 250 and the second (n = 43) corre- sponding to
the ultimate and penultimate occupation levels, dated to between c. AD 250 and
450.
The second site (Huaca Santa Clara, V-67), located about 15 km inland from
Huaca Gallinazo, has been interpreted as a regional administrative center
(Millaire, 2010a). A mid-sized site, Huaca Santa Clara consists of a number of
adobe buildings constructed on a natural hill. On the basis of a series of
radiocarbon dates, the occupation of the site dates from 160 BC to AD 780
(Millaire, 2010a) and it is contemporaneous with the main period of occupation
at Huaca Gallinazo. The administrative nature of the site was suggested on the
basis of substantial agricultural storage facilities (Sectors II, IV, VI), as well as by
the elite character of the material culture associated with the storage facilities
and the large adobe platform atop the hill (Sector I). The storage facilities were
built on the lower terraces along the edges of the mound; small residential
structures were constructed at the base of the mound (Sectors III, V, VII; Fig.
1B). Camelid material sampled from EIP contexts at this site (n = 43) is derived
from fills associated with the ultimate and penultimate occupation levels, which
date to between c. AD 400 and 600.
Following the abandonment of Huaca Santa Clara towards the end of the EIP,
the site was later reused for the purpose of a ritual event (c. AD 1100), a practice
that appears to have been a some- what widespread phenomenon on the north
coast of Peru in the late Middle Horizon to early Late Intermediate Period (c. AD
800-1200) (Millaire and Surette, 2011; Millaire, in press). This ritual event
involved the burial of a young female wrapped in an elaborate textile bundle,
accompanied by five other young individ- uals, as well as 28 immature llamas. In
this case, because the remains of the animals were (in most cases) complete or
nearly complete (Fig. 3), it is possible to identify them as llamas. Of these 28
llamas, 25 were sampled for isotopic analysis of bone collagen and hair samples
were taken from 9 individuals (Table 1). Similar ritual events have been
described at other sites in the Virú and
Specimen
ID Estimated age Tissue(s) sampled3 Notes
Llama 2 <3 months Bone
Llama 3 <3 months Bone
Llama 4 <3 months Bone, Hair (6)

Llama 5 <3 months Bone, Hair (4)

Llama 6 3-6 months Bone, Hair (5)
Llama 7 <3 months Bone
Llama 8 <3 months Bone, Hair (4)
Llama 9 <3 months
Fig. Bone
3. Examples of sacrificed llamas from Huaca Santa Clara.
Llama 10 <3 months Bone
Llama 12 <3 months Bone
Llama 13 Juvenile Bone Cranium missing
Llama 14 Juvenile Bone Cranium missing
Llama 15 <3 months Bone

Llama 16 <3 months Bone

Llama 17 Juvenile Bone, Hair (7) Cranium missing
Llama 18 <3 months Bone

Llama 19 <3 months Bone, Hair (9)

Llama 20 3 to 6 Bone
Llama 21 months
<3 months Bone

Llama 22 <3 months Bone, Hair (8)

Llama 23 <3 months Bone
Llama 24 <3 months Bone

Llama 25 3 to 6 Bone, Hair (7)

months
Llama 26 <3 months Bone, Hair (10)
Llama 27 <3 months Bone

Table 1
Estimated ages for sacrificed camelids from Huaca Santa Clara (Millaire, in press).

Moche Valleys where immature llamas were sacrificed at previ- ously abandoned
settlements (Millaire, in press).

5. Isotopic ecology of the northern Peruvian Andes

The carbon (d13C) and nitrogen (d 15N) isotopic compositions of an animal’s

a
Numbers in parentheses indicate the number of 1 cm hair segments that were sampled from
a given llama.
 P. Szpak et al./Journal of Anthropological Archaeology 36 (2014) 110-129
1997; Kohn, 2010; Petersonand Fry, 1987; Szpak, 2014; Tieszenand
1989), several of which are particularly relevant to the western slope of the
Andes (Szpak et al., 2013b). Moving from the Pacific coast into the Andean
highlands, large changes in mean annual tem- perature and rainfall occur with
annual precipitation increasing and temperature decreasing with altitude. These
environmental changes have important isotopic consequences.
The d13C values of the two main types of terrestrial plants (C 3 and C
distinct, with mean d 13C values of c. -12%c for C 4 plants (mainly grasses adapted
to hot, arid conditions) and c. -26%c for C 3 plants (the majority of terrestrial
plants) (Smith and Epstein, 1971). In the Andean region of South America, C
plants are rela- tively rare at high altitude locations (Boom et al., 2001; Szpak
al., 2013b), which tend to be cooler and wetter than the low alti- tude and
coastal regions. C3 plant carbon isotopic compositions tend to become less
depleted of 13C in arid environments (Vitousek et al., 1990), although the effects
of altitude on plant leaf morphology and associated changes in photosynthetic
capacity counter this effect in the Andes (Szpak et al., 2013b).
The trend of increasing temperature and precipitation with alti- tude in the
region influences the nitrogen isotopic composition of soils and plants (
al., 2013b). Hot and arid ecosystems are characterized by relatively high d
values due to various bio- geochemical processes that favor the loss of the lighter
isotope (14N), driven by the ‘openness’ of these ecosystems (Amundson
2003; Craine et al., 2009; Handley et al., 1999; Szpak, 2014). Several authors
have suggested that arid conditions directly cause relatively high d 15N values in
animal tissues due to various metabolic processes associated with N excretion
and water conser- vation (following Ambrose and DeNiro, 1987). More recently,
how- ever, a much clearer link between plant (and presumably soil) nitrogen
isotopic compositions and rainfall has been found (Hartman, 2011; Murphy and
Bowman, 2006), suggesting that the 15N-enriched herbivore tissues that have
been observed are lar- gely, or wholly, driven by isotopic variation in the diet.
Overall then, there is a general pattern of higher carbon and nitrogen iso- topic
compositions in plant tissues (on average) at coastal and low altitude sites
relative to high altitude sites (Szpak et al., 2013b).
Agricultural plants are less sensitive to these environmental effects on tissue
isotopic compositions. Specifically, because agri- cultural plants may be supplied
with supplemental water through irrigation (especially on the coast of Peru), the
effects of water- availability on plant d 13C and d15N values may be minimized. For
example, wild plants (leaves) sampled from coastal sites in the Moche River
Valley had d15N values as high as +17%c (Szpak et al., 2013b), while the highest
d15N value observed for unfertil- ized maize leaves grown under irrigation
agriculture in the nearby Virú Valley was +6.4%c (Szpak et al., 2012b
Additionally, agricul- tural plants may be treated with a number of organic
fertilizers, which can have moderate (+2%c to +4%c for camelid dung) to extreme
(>+20%c for seabird guano) impacts on plant nitrogen isotopic compositions
(Szpak et al., 2012a, 2012b, 2014). The nitrogen isotopic compositions of
agricultural plants grown on the coast are, therefore, difficult to predict with a
high degree of precision in the absence of isotopic measurements of contextually
associated archaeobotanical material. Nevertheless, the range of possible iso-
topic compositions for these plants, at the very least, added to the isotopic
complexity of the local food web.
6. Materials and methods
6.1. Materials
For comparative purposes, modern camelid specimens (hair and bone) were
collected from five high altitude locations in northern
Peru (3182-3595 masl, Supplementary .kml file). These data allow us to test the
assumption that camelids raised at high altitudes in northern Peru would be
characterized by diets dominated by C3 plants. Archaeological camelids were
sampled from EIP contexts at Huaca Santa Clara (AD 400-600, n = 43) and
Huaca Gallinazo (c. 50 BC-AD 250, n = 13; c. AD 250-450, n = 43), and the late
Mid- dle Horizon sacrificial event at Huaca Santa Clara (c. AD 1100, n = 25). For
these sacrificed animals (n = 9), as well as a smaller number of butchered
animals (n = 5), high levels of organic preser- vation facilitated the collection of
hair samples with adhering skin, allowing for the diachronic analysis of isotopic
compositions.
6.2. Isotopic methodology
Bone collagen was extracted using a modified Longin (1971)
described in detail elsewhere (Szpak, 2013). Hair sam- ples were
adhering particulate matter and loose hairs with fine forceps and a d
Hair was sampled at 1 cm increments wherever possible - hair that
too short (<2 cm), or could not be definitively associated with sk
sampled incrementally, but was sampled in bulk. Hair samples (incre
bulk) were placed into glass tubes and sonicated in deionized water fo
remove any additional adhering particulate matter. Samples were
centrifugation and air-dried at 60 °C, and treated with 2:1 chloroform
Samples were air-dried at 60 °C and then minced as finely as possible
Isotopic and elemental compositions were determined using eithe
Finnigan DeltaPLUS XL or Thermo Scientific Delta V Plus continuous flo
ratio mass spectrometer coupled to a Costech Elemental Analy
Laboratory for Stable Isotope Science (The University of Western Onta
for sam- ple calibrations, analytical accuracy and precision are sum
Table S1.
6.3. Data treatment
Isotopic data that are used primarily in ecological and archa
studies (d13C, d15N) are most commonly plotted as points (x,y) in biva
because the relationships between isotopic data (relative to one
relative to food sources) are qual- itatively meaningful when illustrate
Several metrics uti- lized in this paper take advantage of this
qualitatively meaningful way of plotting bivariate isotopic data (also
and Pearson (2013) for a quantitative example). We further employ a
quantitative measures that follow the same logic.
1. Ranges for carbon (CR) and nitrogen (NR
compositions.
2. Convex hull area (CHA), which is the minimum a
polygon that encloses all of the data in bivariate space (Lay
2007). CHA reflects total bivariate isotopic variability, but is
sensitive to small sample size and can be drastically changed
values or outliers.
3. Mean neighbor distance (MND), which is the me
ian distance between each individual and every other indivi
the group. MND reflects how evenly distributed isotopic data
bivariate space.
4. Standard ellipse area (SEA), which is the area o
iate ellipse with its centroid at the mean x (d13C) and y (d
(Jackson et al., 2011). The standard ellipse is defined in the
sense on the basis of the mean and standard deviation of the
data (in this case d13C and d15N) as well as their correlation
(Batschelet, 1981, pp. 129-158). Related to the basic SEA ar
SEAb (Jackson et al., 2011). The standard ellipse area correct
sample size (SEAc) has the same proportions as SEA, but the
ellipse is adjusted for sample size as described by Jackson e
The standard ellipse area Bayesian (SEAb) is defined as the a
bivariate ellipses generated using a Bayesian framework, whe
number of iterative draws (typically between 10 4 and 106) from
chain Monte Carlo simulation are used to generate a robust
areas. The series of SEAb values generated for a given data
easily compared to other similar datasets, and Jackson et
suggest that for comparative pur- poses, SEAb is the most ro
and in fact better accounts for small sample sizes than does S
For camelids analyzed in this study, within-group variation was
between groups using the SEAb metric (106 iterations). Similarity betw
was assessed using the extent of over- lap between the CHA
respectively. All analyses described in this section were performed usi
source statistical computing package R (R Development Core Team
Mac OS X, utilizing the SIBER (stable isotope Bayesian ellipses in
(Jackson et al., 2011) within the SIAR (stable isotope analy- sis in
(Parnell et al., 2010).
 P. Szpak et al./Journal of Anthropological Archaeology 36 (2014) 110-129
Tabl
e3
Fig. 4. SEAc for modern camelids. Shaded circles represent data for bone collagen: (1)
northern Peruvian highlands (this study), (2) unknown Peruvian highlands (Schoeninger and
DeNiro, 1984) and (3) Upper Mantoro Valley (Thornton et al., 2011). Open ellipses (SEAc) with
solid lines each represent serially sampled hair for individual camelids from the northern
Peruvian highlands. The open ellipse with the broken line represents bulk hair sampled from
three camelid species (llama, alpaca, vicuña) from the northern Peruvian highlands. Hair
(keratin) ¿13C values have been adjusted upwards by +1.3%« as described in the text. Carbon
isotopic compositions for modern animals have been adjusted upwards by +1.5%« to account
for the Suess Effect (Yakir, 2011).
To compare within-individual isotopic variation (differences
between long-term [bone collagen] and short-term [hair]) dietary
signals, circular statistics were employed to quantify the magni- tude
and direction of dietary change (Schmidt et al., 2007). This approach
generates metrics both for the direction of change (the mean angle of
the change, a) and the dispersion of that change (the mean vector
length for all paired data, r, and the angular devi- ation, s). The latter
two metrics (r and s) provide information with respect to the
consistency of change, with larger values of r and smaller values of s
being associated with consistent change.
For direct comparisons of isotopic compositions derived from
different tissues, it is important to account for potential changes in
tissue-diet fractionation (denoted with a capital delta, A). For
mammalian bone collagen, the following tissue-diet fractionations
were used: +3.7%c for A13C and +3.6%c for A15N (Szpak et al.,
2012c). For mammalian hair keratin, a review of published litera-
ture yielded the following tissue-diet fractionation factors: +2.4%c for
A13C (n = 14) and +3.5%c for A15N (n = 9) (Szpak, 2013). On this basis
A15Nkeratin-diet and A15Nconagen-diet were assumed to be the same,
but the higher value for A13Ccollagen-diet, meant that an adjustment
was necessary to directly compare car- bon isotopic compositions of
bone collagen and hair. To this end, when hair and bone collagen
isotopic compositions were plotted together, ó13Ckeratin values were
increased by 1.3%c.
7. Results
7.1. Modern highland camelids
In order to determine whether or not the isotopic compositions of
archaeological coastal camelids are distinct from those of mod- ern
camelids herded in the high-altitude regions of the Andes, it is first
necessary to establish what is expected isotopically of came- lids
herded at high altitudes. Carbon and nitrogen isotopic compo-
sitions of modern highland animals are presented in Fig. 4 as
standard ellipse areas corrected for sample size (SEA c). Isotopic and
elemental data for all modern camelids are presented in Table S2.
These data are derived from bone collagen, bulk hair, and serially
sampled hair. The modern camelids are characterized by low d 13C
and d15N values, consistent with a diet comprised prin- cipally of C3
vegetation. There is also relatively little isotopic vari- ability both
between and within individuals. Variability metrics (Jackson et al.,
2011; Layman et al., 2007) for the camelids with serially sampled
hair are presented in Table 2. On this basis, ani- mals raised in the
Andean highlands should be characterized by low carbon and
nitrogen isotopic compositions, and a relatively low amount of inter-
and intra-individual isotopic variability. The low isotopic variation in
these camelids is driven by two factors. First, there are few C 4 plants
in high altitude pasturelands and there is, therefore, a relatively low
amount of baseline variation in plant d 13C (Szpak et al., 2013b).
Second, based on ethnographic and ethnohistoric accounts, animals
tend to graze together in
(continued on
nextpage)
 8 P. Szpak et al./Journal of Anthropological Archaeology 36 (2014) 110-129

Isotopic and elemental compositions for bone collagen, collagen yields, and associated contextual information for all camelids analyzed in this study.

Sample ID Sector ACa Room Unit d13C (fe, VPDB) d15N (fe, AIR) %C %N C:N Collagen yield

AIS 405 large groups,

relatively VI 110 groups may consist
even though these -17.05
of Camelid+6.8diets therefore 46.8
tend to17.2be fairly
3.17 consistent
17.7 between

camelids
AIS 407owned byVInumerous families (Murra,
117 1965; Tomka, 1992).
-15.39 individuals.
+5.8 41.3 14.8 3.24 18.1
AIS 393 VI 97 -15.16 +6.1 44.9 16.2 3.23 24.8
AIS 396 VI 97 -17.59 +7.5 45.4 15.8 3.35 18.3
AIS 397 VI 97 -15.43 +5.7 40.7 15.0 3.16 20.7
AIS 375 I 93 -16.72 +8.4 44.7 16.0 3.25 14.9
 P. Szpak et al./Journal of Anthropological Archaeology 36 (2014) 110-129

7.2. Archaeological camelid bone collagen

Tabl
e 3Carbon and nitrogen isotopic compositions, as well as associ-
ated elemental data and preservation criteria (%C, %N, C:N ratio,
collagen yield) for all archaeological camelids sampled are pre-
sented in Table 3. Only collagen samples that met all of the follow-
ing criteria were included in summary plots and statistical analyses
(Ambrose, 1990; DeNiro, 1985): collagen yield >1%, %C p 13%, %N p
4.8%, 2.9 < C:N ratio <3.6. In addition to carbon and nitrogen
isotopic compositions, convex hull areas (CHA) are presented in Fig.
5A and standard ellipse areas corrected for sam- ple size (SEA c) are
presented in Fig. 5B. For the sake of clarity, the SEA c (without
individual isotopic data) for these archaeological camelids from Virú
are plotted alongside the SEA c for the modern highland camelids in
Fig. 6A. To generate the modern camelid ellipse in Fig. 6A, mean
d13C and d15N values were used for individ- uals with serially
sampled hair.
Variability metrics (Jackson et al., 2011; Layman et al., 2007) for
the four Virú camelid groups, as well as camelids from other
archaeological sites, are presented in Table 4. The Virú camelids are
characterized by a large range of carbon and nitrogen isotopic
compositions, with a more-or-less continuous distribution of points
in bivariate space between -20%c and -10%c for d 13C; the EIP
camelids from Huaca Gallinazo and Huaca Santa Clara have lar- ger
d13C ranges (CR) than any other camelid group analyzed to date

(continued on
nextpage)
 10 P. Szpak et al./Journal of Anthropological Archaeology 36 (2014) 110-129

(Table 4). Because of this, the Virú camelids do not lend themselves camelid groups from coastal sites are characterized by greater iso-
toFig.
being characterized
5. Carbón and nitrogen as consuming
isotopic predominantly
compositions C 3 or
for archaeological Virú C4 plants.
camelid bone collagentopic variability
according to site and(CR, NR, CHA,
time period: SEA
(A) Huaca c) relative
Gallinazo (EIP, 50toBCcamelids
to AD 250),from high
(B) Huaca
Gallinazo
While (EIP,
it is AD 250
true thatto 450),
there (B)are
Huaca Santaindividuals
more Clara (EIP, AD with
400 todiets
600), Huaca Santa Clara (Late
consisting Middlesites.
altitude HorizonBecause
sacrifices, AD
of 1100). In each panel of
the sensitivity the the
irregular polygons
metrics enclosing allin
presented
data points represent the convex hull area (CHA) for each group and the shaded ellipses are standard ellipses corrected for sample size (SEA c).
of a high proportion of C 4 plants relative to modern highland animals Table 4 to sample size, however, it is very difficult to compare the
(Fig. 6A), this does not adequately reflect the complexity and extreme camelids from this study to other archaeological camelid groups
variability in these data. Generally, analyzed to date in any statistically meaningful way. Syvaranta
■20 -18 -16 -14 -12
<S13C (%„, VPDB)
Fig. 6. (A) SEAt for archaeological camelids from Virú compared to modern highland camelids (d 13C corrected by +1.5%« to
account for the Suess Effect). (B) Density plot showing the credible intervals of the standard ellipse areas for Virú and
etmodern
al. (2013) suggested
camelids. a minimum
Thickest boxes number of samples
(50% CI), medium-thickness boxes (75% for
CI), reason-
thinnest boxes (95%The
CI), ellipses
horizontal (SEA
line c) for the Virú camelids are of comparable sizes
able
(meancomparison
standard ellipseof these
area), circlemetrics to (standard
within boxes be aroundellipsenarea
= 30. Withforrespect
corrected to SEA
small sample size onec). another, but much larger than the modern camelid ellipse
to the comparisons generated on the basis of SEA b, Jackson et al. (even though this modern dataset is derived from llamas, alpacas,
(2011) recommend a minimum of n = 10 for each group being com- and vicuñas drawn from both northern and southern Peru). The
pared. Two of the four camelid groups analyzed in this study exceed Huaca Santa Clara sacrificed camelids have slightly higher carbon
these minimum numbers (n = 43 for Huaca Gallinazo AD 250-400 and nitrogen isotopic compositions than the EIP camelids, although
and n = 43 for Huaca Santa Clara AD 400-600), while the Huaca there is considerable overlap among the groups. This is likely caused
Santa Clara sacrifice group approaches this number (n = 25). Most by the very young age of these camelids, which were still largely or
other previously studied archaeological camelid groups are smaller wholly dependent on their mother’s milk (Brown,
than n = 10 (Table 4). At these low sample sizes, high values for 2000) , creating a slight nursing-induced offset (Balasse et al.,
many of the metrics (e.g. SEA) are as much (if not more) a reflection 2001) . The posterior probability distributions of SEAb for the Virú
of the uncertainty in generating these metrics as they are a reflection camelids are plotted in Fig. 6B. The ellipse areas of the four Virú
of the real variation expected in a given group (cf. Jackson et al., camelids groups are generally very similar, although the Huaca
2011). This feature makes large-scale com- parisons with the Santa Clara sacrifice group is smaller than the three EIP groups,
camelid groups reported in the present study problematic. Also, it is which is likely driven by: (1) the smaller sample size in the sacri- fice
not meaningful to group individuals from the same site together group, and (2) the extremely restricted age range of these ani- mals.
when the contexts from which they were derived span several While the ellipse for the earlier camelids at Huaca Gallinazo is
thousand years. While archaeological contexts will always be approximately the same size as the other two EIP groups (Fig. 6A),
associated with comparatively poor temporal resolution, at the very the effect of small sample size on the level of uncertainty in the
least, an attempt can be made to restrict comparative groups to some estimates is clear from the large range in the posterior prob- ability
kind of meaningful spatial and temporal context. In other words, distributions in Fig. 6B. This underscores the importance of robust
large groups at the site level that span several chronological periods sample sizes for comparative purposes.
should be avoided even if this means that a sufficiently large sample The extent of dietary similarity between these groups can be best
cannot otherwise be produced. assessed by comparing the SEA and CHA metrics, which are broadly
 P. Szpak et al./Journal of Anthropological Archaeology 36 (2014) 110-129 11

analogous to the dietary niche of a particular group (Fink et al.,

Tabl
2012). The extent of overlap, expressed as a percentage of
e3
overlapping area between the two polygons or ellipses, for the four
groups of archaeological camelids from Virú is summarized in Table
5. Additionally, direct comparisons can made between each of the
areas generated using the bootstrapping procedure in the calculation
of SEAb (106), which generates a proportion of the areas of Group 1
that are larger than the areas of Group 2, for which 0.5 indicates
complete similarity, and values closer to 0 (no areas of Group 1 > any
areas of Group 2) or 1 (any area of Group 1 > all areas of Group 2)
indicate increasing dissimilarity in elliptical areas. These metrics are
presented in Table 5 as SEA1>2. There is considerable overlap
between each of these camelid groups, sug- gesting very similar
diets. There tends to be less overlap between CHAs relative to SEAs,
which is due to the fact that CHA is more sensitive to extreme
values. Based on the various assessments of similarity discussed
above, both with respect to the extent of iso- topic variation, and the
extent of overlap between polygons or ellipses in bivariate space,
these four groups do not significantly differ from one another.

7.3. Archaeological camelid hair

Carbon and nitrogen isotopic compositions and associated ele-

mental data (%C, %N, C:N ratio) for archaeological camelid hair are
presented in Table S3 and summarized in Table 6. Incremen- tally
sampled hair carbon and nitrogen isotopic compositions are plotted
in Fig. 7A; SEAc for each of these individual camelids are presented
in Fig. 7B. Variability metrics (Jackson et al., 2011; Layman et al.,
2007) for individual camelids with incrementally sampled hair are
presented in Table 6.
Some camelids had hair ó13C values that varied by as much as
5%c, suggesting substantial diachronic dietary variation. By way of
comparison, in a C3/C4 diet switching experiment (14%c
Table 6
Variability metrics and circular statistics for archaeological camelids (carbon and nitrogen isotopic compositions of serially sampled hair). Individual hair segment data for these camelids are
presented in Table S3.
C N MN CH SE
Sample ID Segments d13C Mean ± SD (fe, VPDB) d15N Mean ± SD (fe, AIR) R R D A Ac aa rb sc

Huaca Santa Clara (EIP, AD 400-600) AIS 374 1 -20.14 +8.6

AIS 389 1 -15.19 +9.1 - - - - - - - -
AIS 395 5 -19.15 ±0.15 +7.8 ± 0.1 0. 0 0.2 <0. <0. - - -
4 1 1
AIS 579 11 -13.53 ±0.64 +10.7 ±0.7 2 .2 1.0 0. 0.5 - - -
AIS 578 9 -19.73 ±1.25 +7.8 ± 1.0 .
3. .
2.
2.0 8
5.8 4.6 - - -
8 7
Huaca Santa Clara (Late MH Sacrifices, AD 1100) 2
3. . 331 0.8
Llama 4 6 -15.71±1.21 +7.3 ± 0.9 1
2 2.0 4.5 4.1 ” 9 27”
Llama 5 4 -14.90±0.20 +10.0±0.2 0. 0. 0.3 0. 0. 213 0.4 62”
4 4 ” 2
Llama 6 5 -16.93 ±0.76 +7.1 ± 0.2 2 0 1.0 1
0. 2
0. 359 0.8 31”
” 5
Llama 8 4 -17.20±2.19 +6.8 ± 0.4 .4. .
1 2.7 6
1.9 8
4.0
354 0. 72”
7 ”
Llama 17 7 -15.51 ±1.78 +8.2 ± 1.7 4. .4. 2.8 5.2 4.6 63” 22
0. 46”
5 7
Llama 19 9 -17.95 ±1.22 +9.2 ± 0.8 3. 2 1.7 1. 1. 305 68
0.9 13”
8 ” 7
Llama 22 8 -18.19 ±0.28 +7.2 ± 0.5 0. .
1. 0.7 6
0. 0
0.4 330 0.9 10
9 3 ” 8
Llama 25 7 -18.32 ±0.59 +7.2 ± 0.4 1 1. 0.9 6
0.7 0. 358 0.9 15”
”
3 ” 7
Llama 26
10
-14.26±0.66 +9.6 ± 1.5 .3. 1. 1.3 2. 6
1.9 259 0.9 9”
0 4 ” 9
8
a
Mean angle of change between hair segments and bone collagen (a).
b
Mean vector length for all hair segments relative to bone collagen (r). Larger values of r reflect more consistent change.
c
Angular dispersion for all hair segments relative to bone collagen (s) roughly analogous to standard deviation. Maximum value for angular dispersion is 81”.

difference in d13C between two diets) that measured d 13C in cattle tail
hairs, there was a gradual equilibration over a period of 6075 days
before a pure C3 or C4 signal was recorded in the hair (Jones et al.,
1981). Given that many of the sacrificed Huaca Santa Clara camelids
lived less than three months, this implies significant temporal
variation in the diet of some of these individual animals during their
short lives. The sacrificed animals with the greatest changes in diet
tend to have lower d13C values leading up to the time of death,
discounting the possibility that these animals were ceremonially fed
maize or maize beer in preparation for sacrifice for any appreciable
amount of time. Conversely, other animals are characterized by little
or no change over the length of hair sam- pled. Overall, there is no
clear trend among these individuals in carbon or nitrogen isotopic
compositions leading up to the time of death.
The carbon and nitrogen isotopic compositions of bone and hair
(with adhering skin) from individual camelids for which both tis-
sues were available, were compared directly using the angular/cir-
cular approaches discussed previously (Fig. 7C and Table 6). Some
individuals have consistent differences between hair and bone col-
lagen isotopic compositions (reflected in high values for on
(continued r and low
nextpage)
 12 P. Szpak et al./Journal of Anthropological Archaeology 36 (2014) 110-129

values for s) while others do not (reflected in low values for r and
high values for s). Additionally, there is no consistent trend in the
direction (measured as mean angle of change, a) of isotopic differ-
ence between the hair and the collagen (Table 6). At the individual
level, Virú camelids are characterized by an extremely wide range of
isotopic variation (measured as SEAc) and evenness (measured as
MND) (Fig. 8); comparatively, modern highland animals are
characterized by consistently low isotopic variation and high even-
ness (Fig. 8).

8. Discussion

8.1. Local/Coastal camelid herding

Previously, relatively high d 13C and d15N values have been inter-
preted as evidence of camelid herding in the low sierra or coastal
regions, possibly due to a high reliance on marine algae (DeNiro,
1988) or lomas plants (Thornton et al., 2011). In this study, relatively
high d13C and to a lesser extent d 15N values were observed for
camelids recovered from coastal sites in relation to animals herded in
high altitude areas of the Andes (Fig. 6). When the cam- elid isotopic
data are considered in relation to plants sampled along an altitudinal
transect in northern Peru, the variation observed in camelid isotopic
compositions can be satisfactorily explained by the consumption of
local terrestrial vegetation. Fig. 9 depicts the convex hulls formed by
the carbon and nitrogen isotopic compositions for modern plants
sampled in five major ecological zones along an altitudinal transect
in northern Peru. Transposed onto these polygons are the convex
hulls for the three groups of camelids analyzed in this study.
Following the basic prin- ciples of isotopic mixing, the isotopic
composition of a consumer in bivariate space must fall within the
polygon formed by the sources (foods); if this condition is not met,
these sources cannot explain
 P. Szpak et al./Journal of Anthropological Archaeology 36 (2014) 110-129 13

the consumer isotopic composition. The area of overlap between the
Tabl
camelid (consumer) and plant (source) polygons represents the
e3
relative percentage of consumer isotopic data that can be pro- duced
by the consumption of the plants making up the source polygon
(upper left of each panel in Fig. 9). In this case, the coastal polygon
explains 100% of the isotopic variation observed for each of the
consumer (camelid) polygons, while the source (plant) poly- gons in
each of the other ecological zones never explains more than 73% of
this variation. This is true for all four groups of Virú came- lids. On
this basis, it is reasonable to suggest that many of these animals
were raised on local vegetation, and were not imported from high
altitude contexts. This does not preclude the possibility
that some of these camelids may have been imported from locations
in the highlands or in the nearby yungas (c. 10002500 masl).
Nevertheless, some of these camelids have tissue isotopic
compositions that are not consistent with the isotopic variation
observed in non-coastal vegetation.
The extremely restricted age range of the sacrificed camelids from
the late Middle Horizon context at Huaca Santa Clara is wor- thy of
additional discussion within the context of coastal herding. The
variation in carbon and nitrogen isotopic compositions in these
sacrificed animals is very similar to that observed in the EIP came-
lids from Huaca Santa Clara and Huaca Gallinazo (Fig. 5), with a
high degree of overlap between these groups (Table 5). There is

Fig. 7. (A) Carbón and nitrogen isotopic compositions of serially sampled camelid hair. Hair segments formed closest to the time of the animal's
death are towards the left of the x-axis. (B) Standard ellipse areas corrected for small sample size. (C) Circular diagrams depicting changes in
carbon and nitrogen isotopic compositions between hair and bulk bone collagen for individual camelids. The center of each diagram represents the
bone collagen isotopic composition. Lines emanating from the center represent the mean angle of change between the collagen and keratin (hair
segment) (r) and shaded areas around this broken line represent the angular deviation (s). 0” = increasing ¿ 15N, 180” = decreasing ¿ 15N; 90” =
increasing ¿13C, 270” = decreasing ¿13C. Numbers (e.g. 395, 578, 579 in the top panels) correspond to sample IDs as outlined in Table 3.

(continued on
nextpage)
 14 P. Szpak et al./Journal of Anthropological Archaeology 36 (2014) 110-129

less ovariation in these sacrificed animals, but this is likely a product

■ tion relative to the EIP camelids. Because the
of the extremely restricted Decreasingage Dietaryrange
Evennessand Over Timesmaller sample size for this
group. Moreover, there
-------------------- is a high
1------------------- degree
1------------------ of within-individual variation
1----------- ------- 1------------------- 1
>>
---------------
sacrificed group represents an extremely restricted
in some of the# Huaca © Huaca Santa Clara (EIP, AD 400-600)
sacrificed camelids, as well as the EIP camelids,
Santa Clara (Sacrificas, AD 1100) 5
temporal window this suggests that the variation
although for
•
both groups the general pattern is an incon- "o sistent
” O Modern
n=5 © n> observed in the EIP camelids is not simply an
amount of variation within individuals (Fig.O 7). On thisO o basis, to it is
0)
artificial product reflecting the combination of a
_
# n=10 ^
reasonable to suggest that the pattern of animal husbandry 5‘
IQ
temporally disparate sample.
n=15
associated with these sacrificed camelids was very similar |3to ; that
5'
associated with the earlier EIP camelids. 3 The high between-individual and inconsistent within-individ- ual
LD
<
(/> Alternatively, it is possible that the type of isotopic ‘signature'
o.
< isotopic variation for the sacrificed llamas from Huaca Santa Clara
31
observed for these camelids is the product of a group of camelids c are inconsistent with these animals being kept together for a
derived from disparate geographic areas, some being imported o <. from significant period of time. Although the notion that these animals
long-distance, possibly from high altitude locations. Aside from the were imported from distant, ecologically distinct locations is not
incompatibility_ of this • explanation 0 o with the isotopic var- iation in
ó
supported by the isotopic data, this does not imply that these ani-
e ©q*©
o
plants previously discussed, there are several biological aspects of mals were herded in the immediate vicinities of the sites from which
II I I
camelid reproductive 0. biology that make this suggestion 2.5 unlikely. they were eventually recovered. These animals may have been herded
5 1.0 1.5 MND
Taking into account the ages of these sacrificed camelids (Table 1), if 2.0 in the vicinity of any of the numerous settlements in the Virú Valley
(%o)
they were ‘imported' from elsewhere but killed at <6 months of age (Willey, 1953), but the dearth of vegetation at a regional scale,
(and in Comparison
Fig. 8. most cases <3 months
of variation (SEAc) of andage), evenness they (MND)would metrics need within to individual
traverse combined with the isotopic variability observed within these camelid
modern (high sierra) and archaeological (coastal) camelids generated from serially sampled
the distance to the site as very young animals. There is no record of groups points to a scale and mode of hus- bandry that differs from
hair. Shape sizes are scaled to sample size (number of hair segments sampled).
caravans including pregnant female camelids, or juveniles that have what has traditionally been associated with highland environments
not been fully weaned, which usually occurs between seven and nine in modern and historic accounts - herds moving between large
months (Brown, 2000). Caravan llamas begin training after two years highland pastures.
of age (Browman, 1990), with cara- vans consisting of castrated
males between two and eight years old (Nielsen, 2001). For the first 8.2. The scale of camelid husbandry and human-camelid interactions
few months of life, camelids consume 10% of their body weight in
milk every day, suckling for several minutes between every two hours Aside from the isotopic compositions of camelid groups and their
and several times per hour throughout the day (Fowler, 1998). relation to source (plant) isotopic compositions discussed previously,
Moreover, neonatal and pre- weaned camelids are associated with it is important to consider the isotopic variation both within groups
high morbidity and mortality rates (as high as 70% mortality in the and within individuals. The quantification of dietary (isotopic)
Andes), with the highest mor- tality rates associated with animals <6 variation rather than an attempt at a precise reconstruc- tion of diet
months of age (Davis et al., 1998). Combined, these factors make it composition has excellent potential to provide insight into the nature
unlikely that these juvenile animals would have traversed any of animal husbandry practices and human-animal interactions in
significant distances prior to being killed as part of this ritual event. prehistoric contexts.
The improbability that such young animals were imported from long Camelids are generally very adaptable in terms of their diet. Since
distances, and the similar- ity in isotopic compositions and variation llamas and alpacas have become more common on farms outside of
between the sacrificed and EIP camelids, supports the assertion that the Andes in the last fifty years, they have been raised in a great
these animals were locally raised. It could be argued that the large diversity of environments, and on nearly every type of pasture in
isotopic variation observed in the four archaeological Virú camelid North America, Europe, Australia, and New Zealand (Fowler, 1998).
groups reflects some unknown temporal variation caused by the Moreover, isotopic evidence from archaeological camelids suggests
coarseness of the archaeological record, but the sacrificed camelids that foddering with agricultural products or byproducts (maize
from Huaca Santa Clara are characterized by similar amounts of specifically) occurred outside of the puna (Dufour et al., 2014;
isotopic varia- Finucane et al., 2006; Verano and DeNiro, 1993). This practice has
also been recorded ethnographically (Godoy, 1984; McCorkle, 1987;
Nielsen, 2001) and ethnohistorical- ly (Nielsen, 2009), with maize
being the only cultigen specifically mentioned. Accordingly, that
camelids would have varied diets is not surprising, but in most
cases, the large dietary variation that has been observed occurs at
the group level over time, rather than between individuals within
groups. For instance, the grazing of field stubble after harvest is a
seasonal activity that is often part of transhumant movement of
pastoral groups, but all camelids within these groups feed on the
same kinds of food throughout the year. Although the proportions of
these different foods may vary somewhat between individuals, this
scale of variation is unli- kely to manifest itself to a significant degree
in the isotopic compo- sition of a tissue like bone collagen. These
smaller variations will most likely be dampened by the slow turnover
rate of collagen, which reflects dietary intake over at least the last
several years of an animal's life (Hedges et al., 2007) and inter-
individual variation would therefore be low.
The pattern of very high between-individual variation and var-
iable within-individual variation for the Virú camelids is inconsis-
tent with the type of camelid husbandry that has been described for
puna and high sierra environments in the Andes (Flannery et al.,
1989; Flores-Ochoa, 1979; Kuznar, 1990). Instead of the rel- atively
large herds associated with this classic model of Andean pastoralism,
the patterns observed for these archaeological came- lids suggest a
 P. Szpak et al./Journal of Anthropological Archaeology 36 (2014) 110-129 15

smaller-scale variety of camelid husbandry in the

Tabl
e3

(continued on
nextpage)
 16 P. Szpak et al./Journal of Anthropological Archaeology 36 (2014) 110-129

low altitude and coastal areas of northern Peru - small groups of
camelids may have been kept by families or other small social units.
In this case, the large amount of isotopic variation for these Virú
camelids may have been driven by differences in management
strategies by individual herders for any number of reasons: (1)
access to particular kinds of forage (either agricultural (by)prod- ucts
or wild pasture) may have varied between kin groups or set-
tlements, (2) notions of the ‘proper way’ to raise one’s camelid, and
what it should be fed were highly variable, and/or (3) certain
individual camelids or groups of camelids exhibited strong prefer-
ences for particular types of forage. In support of the first, there is
large isotopic variation in both wild and domestic local plants, and
considerably more variation with respect to other ecological zones
along the western slope of the Andes (Fig. 9). With respect to the
second, there is a great diversity in the attitudes of modern owners
in terms of the appropriate foddering method (Fowler, 1998). The
tending of small herds provided with distinct types of forage may
have been further necessitated by the unique nature of the
environment in the coastal valleys, with sparse vegetation outside of
the lush cultivated zones in close proximity to rivers and irriga- tion
networks.
The practice of grazing camelids on field stubble in the context of
low altitude husbandry merits further discussion. In general, plant
organs that serve as nitrogen sources (leaves and stalks) tend to be
enriched in 15N relative to plant organs that are nitrogen sinks
(grains) (Szpak, 2014). These effects have been observed in maize
specifically, and in some cases, may be very large (>10% o; Szpak et
al., 2012a; Szpak et al., 2012b). If particular groups of

Fig. 9. Mixing polygons for plants sampled from five major ecological zones along an altitudinal gradient in northern Peru ( Szpak et al., 2013b) arranged by columns
compared to convex hull areas for the three groups of Virú camelids (bone collagen) discussed in the text (adjusted for trophic level fractionation for bone collagen)
arranged by rows. Numbers in the top left of each panel are the percentage overlap between the camelid and plant polygons. Sampling locations for the plants used to
generate these polygons are presented in the Supplementary .kml file.
 P. Szpak et al./Journal of Anthropological Archaeology 36 (2014) 110-129
animals grazed consistently on harvested fields
Tabl
and others did not (perhaps because camelids herded
e3
by certain families had more limited access to
agricultural fields and/or were situated on the valley
margin with limited access to water), this may have
served as a further source of variation in source
isotopic compositions, contributing to the high inter-
individual variation observed for the Virú camelids.
Related to this point is the potential importance of
the fertilization of agricultural fields with camelid
dung, either purposefully and/or coincidentally
during the grazing of field stub- ble, which has been
shown to increase plant tissue d15N values by 2-4%c
with a single season of application at a relatively low
rate (Szpak et al., 2012b); such effects may be more
pronounced with consistent manuring over the long
term (Fraser et al., 2011). While there may be no
predictable and consistent means with which to
model the relative importance of these effects, the
significant point here is that they may serve to
increase the source isotopic variation (increase the
area of the mixing polygon) and contribute to high
levels of isotopic variation among consumers where
cultigens con- stitute significant portions of the diet.
Differences in ownership (or at least management responsibili-
ties) of camelids combined with the lack of large local pastures at
coastal and low altitude locations may have necessitated the spa- tial
segregation of small groups of animals. In the highlands, individual
families typically own two to several hundred animals, although all of
the animals owned by a community will graze together in a much
larger aggregation with ownership being phys- ically ascribed to the
animals (often on their ears) (Murra, 1965; Tomka, 1992). For the
Virú camelids, the isotopic data are consis- tent with some degree of
physical separation of the animals throughout their lives. The
extreme variation, but fairly even spac- ing of isotopic compositions
in bivariate space for the Virú camelids (Fig. 5) is indicative of a
diverse range of animal management strategies and human-animal
interactions. Given the relative dearth of wild vegetation, the
maintenance of local camelid herds would almost certainly have
required camelid diets be supplemented by agricultural products
and/or byproducts. In this case, rather than animals being brought
to a particular pasture, fodder would have been brought to the
animals, and their diet would have been further constrained by the
choices of the herders. Given the fact that both wild and domestic
vegetation was concen- trated in the vicinity of rivers and associated
irrigation networks, it stands to reason that these animals would
have been kept for a significant portion of time in the vicinity of
settlements, which were similarly concentrated in these areas. Some
evidence for small cor- rals exists along the coast (Bawden, 1982;
Shimada, 1994; Wilson, 1988). In Virú, many sites contain
significant deposits of camelid dung in association with small stone
enclosures, which could have housed small numbers of animals
(Supplementary .kml file). Recent excavations at Huaca Gallinazo
(Millaire and La Torre, 2014) demonstrate that at least some animals
were kept inside habitations. In Architectural Compound 3, the floor
of a small room (3 x 4 m) located in a food production area was
covered with camelid dung (15 cm thick), a deposit that was
eventually sealed with a beaten clay floor. In this scenario, where
animals are kept close to human habitation sites and food is
17
gathered and brought to ani- mals in corrals, there is a greater
chance for close, face-to-face interactions between people and
camelids, much more than would occur if animals spent significant
periods of time away from human settlements, grazing on distant
pastures. These conditions (limited spatial and extended temporal
contexts) are precisely the type that allow for complex relationships
to develop between individuals of each species (Knight, 2012).
Thinking back to Kohn’s (2007) notion of an ‘anthropology of life’
in which the entanglements between humans and other spe- cies are
examined, a much different mode of interaction existed between
coastal and highland camelid herders and their animals, at least on
the basis of their diets. There was not, however, a characteristic
‘coastal’ pattern of camelid diet, suggesting that the practice of
camelid husbandry was not highly standardized on the coast.
Focusing solely on the circumstances surrounding the deaths of
these animals obscures this variation. For instance, the large group
of juvenile camelids that were part of the ritual sacrifice at Huaca
Santa Clara died under generally similar circum- stances - they were
part of the same ritual sacrifice - but the iso- topic data demonstrate
that these animals lived very separate lives prior to this event.
It is entirely possible that additional differences exist between the
traditional high altitude mode of Andean pastoralism and that which
occurred along the prehispanic north coast of Peru although it is not
possible to speculate on what these might be based on the isotopic
data presented here. Additional lines of evidence that are able to
provide insight into how animals lived will shed further light on these
issues, but zooarchaeologists and their collaborators must prioritize
their research accordingly. Such lines of evidence include, but are
not limited to: paleopathology, ancient DNA, dental wear, molecular
and microscopic coproscopy, and additional isotopic or elemental
markers. Primary zooarchaeological data remain of immense
importance, and none of the above-mentioned techniques would be
meaningful and/or feasible without these data.
8.3. Regional significance
The management of livestock and, accordingly, the interactions
that occur between humans and animals are strongly influenced by
a range of social and political factors (Comaroff and Comaroff, 1990;
Holloway, 2001; Kirksey and Helmreich, 2010). The organi- zation
and scale of north coast polities during the EIP and Middle Horizon
are widely debated, and a clear picture of political organi- zation in
the region is only beginning to emerge. The importance of the local
ruling polity in Virú in directing large-scale public works projects
(most notably the expansion of irrigation networks) and the presence
of large storage facilities at Huaca Santa Clara suggest the collection
of some form of tribute from the local populace (Millaire, 2010a).
However, it is not clear to what extent this polity may have exerted
control over the management of domestic ani- mals. The most
extensive comparative data available come from Murra
(1965,1968,1980), who summarizes a number of ethnohis- toric
accounts to develop a detailed picture of how camelid hus- bandry
operated (primarily in southern Peru) under Inka control and
immediately following the Spanish conquest. In these accounts,
herding took place primarily in the puna. Generally, herds that were
controlled either by the state or by the church were very large, while
community herds were of extremely variable size. There are
numerous instances in these accounts in which camelids were
segregated into distinct groups, not only with respect to ownership
and management, but also by physical segre- gation of particular
herds. For instance, there were herds con- trolled by the state or
‘crown’ and the ‘church’, both of which were large, and physically
separated from one another. Addition- ally, state herds in different
districts were assigned to particular pastures, with animals from one
area being restricted from grazing in the pastures assigned to
another (Polo de Ondegardo, 1571). Conflicting accounts exist as to
whether or not community animals were provided(continued
as tribute
on
to the
nextpage)
 18 P. Szpak et al./Journal of Anthropological Archaeology 36 (2014) 110-129
state, and whether or not state or community animals were destined
for sacrificial events. Murra (1965) notes that in the southeastern
portion of the Inka Empire, despite frequent wars and the
appearance of a series of ruling pol- ities with different sociopolitical
structures, the activity of camelid herding itself remained largely
organized at the kin group level. The similarities in isotopic
compositions between the EIP and late Middle Horizon Virú camelids
may also reflect continuity in camelid herding practices. Specifically,
it is possible that the tradi- tional north coast mode of camelid
management involved highly differentiated herding of small groups of
animals by kin groups, and in Virú at least, continued unchanged
with the rise and fall of the EIP polities, and with the introduction of
new ritual practices during the Middle Horizon. The occurrence of
llama sacrifices on previously abandoned settlements may represent
one such ritual practice ‘imported’ at this time (Millaire, in press).
The isotopic compositions of these sacrificed camelids, however,
suggest that these rituals involved camelids that were raised locally
in a man- ner that dates back to at least the early part of the EIP.
The high lev- els of between-individual isotopic variation demonstrate
the inclusion of animals from several different herds, possibly
signify- ing that these ritual events involved the ‘donation’ of
camelids from many different groups of people. This may be because
these events were coordinated by local elites who required donations
of camelids as ‘substitutes’ for particular kin groups (Küchler,
2002) or because these events had a unifying function for the local
community in absence of any top-down control (see also Szpak et al.,
in press).
That herding appears to have continued at this small scale fol-
lowing the collapse of the ruling EIP polity in Virú does not imply
that local elites did not impact camelid husbandry in a broad sense.
They may have required new levels of tribute in the form of whole
camelids for ritual or of camelid products (e.g. wool). They may also
have altered the nature of the ownership of camelid herds. There is,
however, no evidence of the expansion of the scale of camelid herding
to anything equivalent to what is observed in the highlands, or of the
large-scale importation of animals raised in high altitude contexts to
meet demand that could not be sup- ported by the smaller coastal
herds. It is possible that local envi- ronmental conditions did not
allow for the maintenance of the very large herds that were
associated with later Andean polities, such as the Inka (Murra,
1965). Even when comparing herds in the puna and sierra today,
herd sizes are often on the order of 100-300 animals in the former,
but typically much smaller (15100) in the latter (Dransart, 2002;
Kuznar, 1990). Accordingly, the small herds of coastal settlements
may represent a further extension of this pattern. Additional isotopic
studies of camelids from other periods and regions on the coast of
Peru will aid in clar- ifying these issues.
9. Conclusions
The camelids from Huaca Gallinazo and Huaca Santa Clara are
characterized by high levels of inter-individual isotopic variation, as
well as inconsistent levels of within-individual isotopic variation.
Taking into account the isotopic compositions of sources from
different ecological zones in the Andes, only plants growing in the low
altitude and coastal regions adequately explain the isotopic variation
observed among these individuals, suggesting that at least some of
these animals were raised locally. Virú camelids were characterized
by a large amount of variation in carbon and nitrogen iso- topic
compositions, suggesting variable management practices, fitting with
a small-scale type of husbandry where small groups of animals were
kept relatively close to human habitation sites and fed a highly
variable diet. Such practices may be due to a com- bination of
environmental (availability of adequate forage or differ- ential access
to different types of forage) and social (different perceptions about
‘proper’ camelid husbandry practices) factors. The ranges of carbon
and nitrogen isotopic compositions were similar for the EIP and late
Middle Horizon camelids, suggesting temporal continuity in these
small-scale management practices.
The methodologies for quantifying isotopic variation outlined in
this paper have great potential in the broadest sense for
archaeological data. They offer quantitative and robust means with
which to compare groups across space and time, although, the
problem of small sample size must be considered. In a more general
sense, it is important for isotopic anthropologists and archae-
ologists to explore such quantitatively grounded methodologies in
their interpretations. The basic principles of most of these methods
(Euclidean distance in bivariate space) make use of the same logi- cal
spatial principles that anthropologists and archaeologists pres- ently
use in a strictly descriptive or qualitative sense to interpret their
data. Finally, considering the extreme uncertainty associated with
source isotopic compositions in archaeological contexts, more
attention must be paid to interpretive methods that do not only or
primarily seek to reconstruct diet composition, but also consider the
importance of diet variation. All of this is not to suggest that the
methods outlined take the place of any other that may be applied to
isotopic data. Instead, they represent one tool in an ever-expanding
interpretive toolkit that may serve to provide new insight into isotopic
data.
Isotopic analysis of faunal material has been used in a diverse
range of contexts to address the nature of human-animal interac-
tions, in both wild (Barrett et al., 2011; Szpak et al., 2012c, 2013a)
and domestic species (Balasse et al., 2002; Cannon et al., 1999;
Finucane et al., 2006; Pearson et al., 2007). Further research along
these lines will undoubtedly continue to shed light on a number of
significant archaeological and anthropological issues and we sug-
gest that these and other types of data that reflect animal life his-
tories hold the greatest potential for gaining insight into past human-
animal interactions. This is especially true for domestic species, and
the consideration of variation (isotopic and otherwise) among
individuals and what this might imply about the nature and scale of
these interactions should be explored in other contexts. The methods
outlined in this paper offer one means with which to address isotopic
data in this way and investigators should con- tinue to explore
alternative interpretive strategies to assess animal life histories.
Acknowledgments
We thank George R. Milner, John M. O’Shea, and three anony-
mous reviewers for their insightful comments that greatly improved
the quality of this manuscript. Estuardo La Torre coordinated sample
collection. Kim Law and Li Huang provided technical assistance.
Sharon Buck, Rebecca Dillon, and Rebecca Parry assisted with
sample preparation. Isotopic research was conducted under
Resolutción Vicesministerial No. 014-2013-VMPCIC-MC. This pro-
ject was supported by the Wenner-Gren Foundation (Dissertation
Fieldwork Grant to PS), Social Sciences and Humanities Research
Council of Canada (Standard Research Grant to CDW, FJL, JFM;
Bombardier Doctoral CGS to PS), Natural Sciences and Engineering
Research Council Discovery Grant (FJL), Canada Foundation for
Innovation and Ontario Research Fund Infrastructure Grants (FJL),
Canada Research Chairs Program (CDW, FJL), and The University of
Western Ontario. Author Contributions: PS, JFM, CDW, FJL
designed research. PS performed research. PS, JFM, FJL, CDW inter-
preted the data. PS wrote the paper with editorial input from JFM,
CDW, FJL. This is Laboratory for Stable Isotope Science Contribution
#305.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at
 P. Szpak et al./Journal of Anthropological Archaeology 36 (2014) 110-129
http://dx.doi.org/10.1016/jjaa.2014.08.005. These data
Tabl
include
e3
Google maps of the most important areas described in this
article.
References
Abbink, J., 2003. Love and death of cattle: the
paradox in Suri attitudes toward livestock. Ethnos
68, 341-364.
Alger, J.M., Alger, S.F., 2003. Cat Culture: The
Social World of a Cat Shelter. Temple University
Press, Philadelphia.
Ambrose, S.H., 1990. Preparation and
characterization of bone and tooth collagen for
isotopic analysis. J. Archaeol. Sci. 17, 431-451.
Ambrose, S.H., DeNiro, M.J., 1987. Bone
nitrogen isotope composition and climate. Nature
325, 201.
Amundson, R., Austin, A.T., Schuur, E.A.G.,
Yoo, K., Matzek, V., Kendall, C., Uebersax, A.,
Brenner, D., Baisden, W.T., 2003. Global
patterns of the isotopic composition of soil and
plant nitrogen. Global Biogeochem. Cycles
17,1031.
Argent, G., 2010. Do the clothes make the horse?
Relationality, roles and statuses in Iron Age Inner
Asia. World Archaeol. 42,157-174.
Balasse, M., Bocherens, H., Mariotti, A.,
Ambrose, S.H., 2001. Detection of dietary
changes by intra-tooth carbon and nitrogen
isotopic analysis: an experimental study of
dentine collagen of cattle (Bos taurus). J.
Archaeol. Sci. 28, 235-245.
Balasse, M., Ambrose, S.H., Smith, A.B., Price,
T.D., 2002. The seasonal mobility model for
prehistoric herders in the south-western Cape of
South Africa assessed by isotopic analysis of
sheep tooth enamel. J. Archaeol. Sci. 29, 917932.
Barrett, J.H., Orton, D., Johnstone, C., Harland,
J., Van Neer, W., Ervynck, A., Roberts, C.,
Locker, A., Amundsen, C., Enghoff, I.B.,
Hamilton-Dyer, S., Heinrich, D., Hufthammer,
A.K., Jones, A.K.G., Jonsson, L., Makowiecki,
D., Pope, P., O'Connell, T.C., de Roo, T.,
Richards, M., 2011. Interpreting the expansion of
sea fishing in medieval Europe using stable
isotope analysis of archaeological cod bones. J.
Archaeol. Sci. 38, 1516-1524.
Batschelet, E., 1981. Circular Statistics in
Biology. Academic Press, London.
Bawden, G., 1982. Gallindo: a study in cultural
transition during the middle horizon. In: Moseley,
M.E., Day, K.C. (Eds.), Chan Chan: Andean
Desert City. University of New Mexico Press,
Albuquerque, pp. 285-320.
Bear, C., 2011. Being angelica? exploring
individual animal geographies. Area 43, 297-304.
Bennett, W.C., 1950. The Gallinazo Group, Viru
Valley. Yale University Press, New Haven, Peru.
Bonavia, D., 2008. The South American
Camelids. Cotsen Institute of Archaeology,
University of California, Los Angeles.
Boom, A., Mora, G., Cleef, A.M., Hooghiemstra,
H., 2001. High altitude C4 grasslands in the
northern Andes: relicts from glacial conditions?
Rev. Palaeobot. Palynol. 115, 147-160.
Browman, D.L., 1990. Camelid pastoralism in the
Andes: Llama caravan fleteros, and their
importance in production and distribution. In:
Salzman, P.C., Galaty, J.G. (Eds.), Nomads in a
Changing World. Instituto Universitario
Orientale, Naples, pp. 395-438.
Brown, B.W., 2000. A review on reproduction in
South American camelids. Anim. Reproduct. Sci.
58, 169-195.
Byrd, B.F., Cornellas, A., Eerkens, J.W.,
Rosenthal, J.S., Carpenter, T.R., Leventhal, A.,
Leonard, J.A., 2013. The role of canids in ritual
and domestic contexts: new ancient DNA insights
from complex hunter-gatherer sites in prehistoric
Central California. J. Archaeol. Sci. 40, 2176-
2189.
Cannon, A., Schwarcz, H.P., Knyf, M., 1999.
Marine-based subsistence trends and the stable
isotope analysis of dog bones from Namu, British
Columbia. J. Archaeol. Sci. 26, 399-407.
Cherel, Y., Hobson, K.A., Guinet, C., Vanpe, C.,
2007. Stable isotopes document seasonal changes
19
in trophic niches and winter foraging individual
specialization in diving predators from the
Southern Ocean. J. Anim. Ecol. 76, 826-836.
Comaroff, J.L., Comaroff, J., 1990. Goodly
beasts, beastly goods: cattle and commodities in a
South African context. Am. Ethnol. 17,195-216.
Craine, J.M., Elmore, A.J., Aidar, M.P.M.,
Bustamante, M., Dawson, T.E., Hobbie, E.A.,
Kahmen, A., Mack, M.C., McLauchlan, K.K.,
Michelsen, A., Nardoto, G.B., Pardo, L.H.,
Peñuelas, J., Reich, P.B., Schuur, E.A.G., Stock,
W.D., Templer, P.H., Virginia, R.A., Welker,
J.M., Wright, I.J., 2009. Global patterns of foliar
nitrogen isotopes and their relationships with
climate, mycorrhizal fungi, foliar nutrient
concentrations, and nitrogen availability. New
Phytol. 183, 980-992.
Davis, R., Keeble, E., Wright, A., Morgan, K.L.,
1998. South American camelids in the United
Kingdom: population statistics, mortality rates
and causes of death. Veter. Rec. 142, 162-166.
deFrance, S.D., 2010. Paleopathology and health
of native and introduced animals on Southern
Peruvian and Bolivian Spanish Colonial sites. Int.
J. Osteoarchaeol. 20, 508-524.
DeNiro, M.J., 1985. Postmortem preservation and
alteration of in vivo bone collagen isotope ratios
in relation to palaeodietary reconstruction. Nature
317, 806-809.
DeNiro, M.J., 1988. Marine food sources for
prehistoric coastal Peruvian camelids: isotopic
evidence and implications. British Archaeological
Reports, International Series 427. In: Wing, E.S.,
Wheeler, J.C. (Eds.), Economic Prehistory of the
Central Andes. Archaeo Press, Oxford, pp. 119-
128.
DeNiro, M.J., Epstein, S., 1978. Influence of diet
on the distribution of carbon isotopes in animals.
Geochim. Cosmochim. Acta 42, 495-506.
DeNiro, M.J., Epstein, S., 1981. Influence of diet
on the distribution of nitrogen isotopes in animals.
Geochim. Cosmochim. Acta 45, 341-351.
Dransart, P.Z., 2002. Earth, Water, Fleece, and
Fabric: An Ethnography and Archaeology of
Andean Camelid Herding. Routledge, London.
Dufour, E., Goepfert, N., Gutiérrez Léon, B.,
Chauchat, C., Franco Jordan, R., Vásquez
Sánchez, S., 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.
Dwyer, P.D., Minnegal, M., 2005. Person, place
or pig: animal attachments and human
transactions in New Guinea. In: Knight, J. (Ed.),
Animals in Person: Cultural Perspectives on
Human-Animal Intimacy. Berg, Oxford, pp. 37-
60.
Fink, P., Reichwaldt, E.S., Harrod, C., Rossberg,
A.G., 2012. Determining trophic niche width: an
experimental test of the stable isotope approach.
Oikos 121, 19851994.
Finucane, B., Agurto, P.M., Isbell, W.H., 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.
Flannery, K.V., Marcus, J., Reynolds, R.G., 1989.
The Flocks ofWamani: A Study of the Llama
Herders on the Punas of Ayacucho. Academic
Press, San Diego, Peru.
Flores-Ochoa, J.A., 1979. Pastoralists of The
Andes: The Alpaca Herders of Paratía. Institute
for the Study of Human Issues, Philadelphia.
Fogel, H.P., 1993. Settlements in Time: A Study of Social and Political Development
During the Gallinazo Occupation of the North Coast of Perú, Ph.D. Dissertation.
Department of Anthropology, Yale University, New Haven.
Fowler, M.E., 1998. Medicine and Surgery of
South American Camelids. Iowa State University
Press, Ames.
Fraser, R.A., Bogaard, A., Heaton, T., Charles,
M., Jones, G., Christensen, B.T., Halstead, P.,
Merbach, I., Poulton, P.R., Sparkes, D., Styring,
A.K., 2011. Manuring and stable nitrogen isotope
ratios in cereals and pulses: towards a new
archaeobotanical approach to the inference of
land use and dietary practices. J. Archaeol. Sci.
(continued on
nextpage)
20 P. Szpak et al./Journal of Anthropological Archaeology 36 (2014) 110-129
38, 2790-2804.
Godoy, R., 1984. Ecological degradation and
agricultural intensification in the Andean
highlands. Hum. Ecol. 12, 359-383.
Goepfert, N., 2012. New zooarchaeological and
funerary perspectives on Mochica culture (AD
100-800). Peru. J. Field Archaeol. 37, 104-120.
Goepfert, N., Dufour, E., Gutiérrez, B., Chauchat,
C., 2013. Origen geográfico de camélidos en el
periodo mochica (100-800 AD) y análisis
isotópico secuencial del esmalte dentario: enfoque
metodológico y aportes preliminares. Bulletin de
l'Institut Fran^ais d'Études Andines 42,1-24.
Grove, M., Pearson, J., 2013. Visualisation and
permutation methods for archaeological data
analysis. Archaeol. Anthropol. Sci., 1-10.
Handley, L.L., Austin, A.T., Stewart, G.R.,
Robinson, D., Scrimgeour, C.M., Raven, J.A.,
Heaton, T.H.E., Schmidt, S., 1999. The N natural abundance
15
(515N) of ecosystem samples reflects measures of
water availability. Austral. J. Plant Physiol. 26,
185-199.
Haraway, D., 2003. The Companion Species
Manifesto: Dogs, People, and Significant
Otherness. Prickly Paradigm Press, Chicago.
Haraway, D., 2006. Encounters with companion
species: entangling dogs, baboons, philosophers,
and biologists. Configurations 14, 97-114.
Haraway, D., 2008. When Species Meet.
University of Minnesota Press, Minneapolis.
Hart, L.A., 2005. The elephant-mahout
relationship in India and Nepal: a tourist
attraction. In: Knight, J. (Ed.), Animals in Person:
Cultural Perspectives on Human-Animal
Intimacy. Berg, Oxford, pp. 15 163-189.
Hartman, G., 2011. Are elevated 5 N values in herbivores
in hot and arid environments caused by diet or
animal physiology? Funct. Ecol. 25, 122-131.
Hedges, R.E.M., Clement, J.G., Thomas, D.L.,
O'Connell, T.C., 2007. Collagen turnover in the
adult femoral mid-shaft: modeled from
anthropogenic radiocarbon tracer measurements.
Am. J. Phys. Anthropol. 133, 808-816.
Herva, V.-P., Salmi, A.-K., 2010. Engaging with
sea and seals: environmental and human-animal
relations on the northern coast of the early
modern Gulf of Bothnia. Norweg. Archaeol. Rev.
43, 115-127.
Hinchliffe, S., Kearnes, M.B., Degen, M.,
Whatmore, S., 2005. Urban wild things: a
cosmopolitical experiment. Environ. Plan. D:
Soc. Space 23, 643-658.
Hobbie, E.A., Hogberg, P., 2012. Nitrogen
isotopes link mycorrhizal fungi and plants to
nitrogen dynamics. New Phytol. 196, 367-382.
Hobbie, E., Ouimette, A., 2009. Controls of
nitrogen isotope patterns in soil profiles.
Biogeochemistry 95, 355-371.
Hogberg, P., 1997. Tansley Review No. 95. 15
N natural
abundance in soil-plant systems. New Phytol.
137, 179-203.
Holloway, L., 2001. Pets and protein: placing
domestic livestock on hobby-farms in England
and Wales. J. Rural Stud. 17, 293-307.
Jackson, A.L., Inger, R., Parnell, A.C., Bearhop,
S., 2011. Comparing isotopic niche widths among
and within communities: SIBER - Stable Isotope
Bayesian Ellipses in R. J. Anim. Ecol. 80, 595-
602.
Jones, R.J., Ludlow, M.M., Troughton, J.H.,
Blunt, J.H., 1981. Changes in the natural carbon
isotope ratios of the hair from steers fed diets of
C4, C3 and C4, species in sequence. Search 12,
85-87.
Kirksey, S.E., Helmreich, S., 2010. The
emergence of multispecies ethnography. Cult.
Anthropol. 25, 545-576.
Knight, J., 2005. Introduction. In: Knight, J.
(Ed.), Animals in Person: Cultural Perspectives
on Human-Animal Intimacy. Berg, Oxford, pp. 1-
13.
Knight, J., 2012. The anonymity of the hunt: a
critique of hunting as sharing. Curr. Anthropol.
53, 334-355.
Knudson, K., Stojanowski, C., 2008. New
directions in bioarchaeology: recent contributions
to the study of human social identities. J.
Archaeol. Res. 16, 397432.
Knudson, K.J., Aufderheide, A.E., Buikstra,J.E.,
2007. Seasonality and paleodiet in the Chiribaya
polity of southern Peru. J. Archaeol. Sci. 34, 451-
462.
Kohn, E., 2007. How dogs dream: Amazonian
natures and the politics of transspecies
engagement. Am. Ethnol. 34, 3-24.
Kohn, M.J., 2010. Carbon isotope compositions of
terrestrial C3 plants as indicators of
(paleo)ecology and (paleo)climate. Proc. Natl.
Acad. Sci. 107,19691-19695.
Küchler, S., 2002. Malanggan: Art. Memory and
Sacrifice, Berg, Oxford.
Kuznar, L.A., 1990. Economic Models, Ethnoarchaeology, and Early Pastoralism in the High
Sierra of the South Central Andes, Unpublished Ph.D. thesis. Northwestern University,
Evanston.
Larco Hoyle, R., 1945. La Cultura Viru. Sociedad
Geográfica Americana, Buenos Aires.
Lawrence, E.A., 1985. Hoofbeats and Society:
Studies of Human-Horse Interactions. Indiana
University Press, Bloomington.
Layman, C.A., Arrington, D.A., Montaña, C.G.,
Post, D.M., 2007. Can stable isotope ratios
provide for community-wide measures of trophic
structure? Ecology 88, 42-48.
Lévi-Strauss, C., 1962. Totemism. Merlin Press,
London.
Longin, R.,1971. New method of collagen extraction
for radiocarbon dating. Nature 230, 241-242.
Losey, R.J., Bazaliiskii, V.I., Garvie-Lok, S.,
Germonpré, M., Leonard, J.A., Allen, A.L., Anne
Katzenberg, M., Sablin, M.V., 2011. Canids as
persons: early Neolithic dog and wolf burials, Cis-
Baikal, Siberia. J. Anthropol. Archaeol. 30, 174-
189.
Marciniak, A., 2005. Placing Animals in the
Neolithic: Social Zooarchaeology of Prehistoric
Farming. UCL Press, London.
Martínez del Rio, C., Sabat, P., Anderson-Sprecher,
R., Gonzalez, S.P., 2009. Dietary and isotopic
specialization: the isotopic niche of three
Cinclodes ovenbirds. Oecologia 161, 149-159.
Matich, P., Heithaus, M.R., Layman, C.A., 2011.
Contrasting patterns of individual specialization
and trophic coupling in two marine apex
predators. J. Anim. Ecol. 80, 294-305.
McCorkle, C.M., 1987. Punas, pastures and fields:
grazing strategies and the agropastoral dialectic in
an indigenous Andean community. In: Browman,
D.L. (Ed.), Arid Land Use Strategies and Risk
Management in the Andes: A Regional
Anthropological Perspective. Westview Press,
Boulder, pp. 57-80.
Millaire, J.-F., 2010a. Moche political expansionism
as viewed from Virú: recent archaeological work
in the close periphery ofa hegemonic city-state
system. In: Quilter, J., Castillo, L.J. (Eds.), New
Perspectives on Moche Political Organization.
Dumbarton Oaks, Washington, DC, pp. 221-249.
Millaire, J.-F., 2010b. Primary state formation in the
Virú Valley, North Coast of Peru. Proc. Natl.
Acad. Sci. 107, 6186-6191.
Millaire, J.-F., in press. The sacred character of ruins on the Peruvian north coast. In:
Fitzsimmons, J., Shimada, I. (Eds.), Living with the Dead in the Andes. University of
Arizona Press, Tuscon.
Millaire,J.-F., Eastaugh, E., 2011.Ancient urban
morphology in the Virú Valley, Peru: Remote
sensing work at the Gallinazo Group (100 BC-AD
700). J. Field Archaeol. 36, 289-297.
Millaire, J.-F., La Torre, E., 2014. Proyecto Arqueológico Virú Temporada 2014. Report
presented to Ministerio de Cultura, Lima, Peru.
Millaire, J.-F., Morlion, M. (Eds.), 2009. Gallinazo:
An Early Cultural Tradition on the Peruvian North
Coast. Cotsen Institute of Archaeology Press, Los
Angeles.
Millaire,J.-F., Surette, F., 2011. Un fardo funerario
procedente de Huaca Santa Clara, valle de Virú
(ca. 1150 AD). Bulletin de l'Institut Fran^ais
d'Études Andines 40, 1-14.
Mullin, M.H., 1999. Mirrors and windows:
sociocultural studies of human-animal
relationships. Ann. Rev. Anthropol. 28, 201-224.
Murphy, B.P., Bowman, D.M.J.S., 2006. Kangaroo
metabolism does not cause 15 the relationship
between bone collagen 5 N and water
availability. Funct. Ecol. 20, 1062-1069.
 P. Szpak et al./Journal of Anthropological Archaeology 36 (2014) 110-129 21

Murra, J.V., 1965. Herds and herders in the Inca Strong, W.D., Evans, C., 1952. Cultural
state. In: Leeds, A., Vayda, A.P. (Eds.), Man,
Tabl stratigraphy in the Viru Valley northern Peru.
e 3Culture, and Animals: The Role of Animals in Columbia University Press, New York.
Human Ecological Adjustments. American Syvaranta, J., Lensu, A., Marjomaki, T.J.,
Association for the Advancement of Science, Oksanen, S., Jones, R.I., 2013. An empirical
Washington, DC, pp. 185-215. evaluation of the utility of convex hull and
Murra, J.V., 1968. An Aymara Kingdom in 1567. standard ellipse areas for assessing population
Ethnohistory 15, 115-151. niche widths from stable isotope data. PLoS ONE
Murra, J.V., 1980. The Economic Organization of 8, e56094.
the Inka State. JAI Press, Greenwich. Szpak, P., 2013. Stable Isotope Ecology and Human-Animal Interactions in the Northern
Newsome, S.D., Tinker, M.T., Monson, D.H., Peru, Ph.D. dissertation. Department of Anthropology, The University of Western Ontario,
Oftedal, O.T., Ralls, K., Staedler, M.M., Fogel, London, Ontario.
M.L., Estes, J.A., 2009. Using stable isotopes to Szpak, P., 2014. Complexities of nitrogen isotope
investigate individual diet specialization in biogeochemistry in plant-soil systems:
California sea otters (Enhydra lutris nereis). implications for the study of ancient agricultural
Ecology 90, 961974. and animal management practices. Front. Plant
Nielsen, A.M., 2001. Ethnoarchaeological Perspectives on Caravan Trade in the South-Central Sci. 5, 288.
Andes. In: Kuznar, L.A. (Ed.), Ethnoarchaeology of Andean South America: Contributions Szpak, P., Longstaffe, F.J., Millaire, J.-F., White,
to Archaeological Method and Theory, International Monographs in Prehistory. C.D., 2012a. Stable isotope biogeochemistry of
Ethnoarchaeological Series 4. Ann Arbor, pp. 163201. seabird guano fertilization: results from growth
Nielsen, A.E., 2009. Pastoralism and the non- chamber studies with maize (Zea mays). PLoS
pastoral world in the late Pre- Columbian history ONE 7, e33741.
of the southern Andes (1000-1535). Nomadic Szpak, P., Millaire, J.-F., White, C.D.,
Peoples 13, 17-35. Longstaffe, F.J., 2012b. Influence of seabird
Oma, K.A., Hedeager, L., 2010. Introduction. Special Issue: Humans and Animals. World guano and camelid dung fertilization on the
Archaeology 42, 155-156. nitrogen isotopic composition of field-grown
Parnell, A.C., Inger, R., Bearhop, S., Jackson, A.L., maize (Zea mays). J. Archaeol. Sci. 39, 3721-
2010. Source partitioning using stable isotopes: 3740.
coping with too much variation. PLoS ONE 5, Szpak, P., Orchard, T.J., McKechnie, I., Grocke,
e9672. D.R., 2012c. Historical ecology of late Holocene
Pearson,J.A., Buitenhuis, H., Hedges, R.E.M., sea otters (Enhydra lutris) from northern British
Martin, L., Russell, N., Twiss, K.C., 2007. New Columbia: isotopic and zooarchaeological
light on early caprine herding strategies from perspectives. J. Archaeol. Sci. 39, 1553-1571.
isotope analysis: a case study from Neolithic Szpak, P., Orchard, T.J., Salomon, A.K., Grocke,
Anatolia. J. Archaeol. Sci. 34, 2170-2179. D.R., 2013a. Regional ecological variability and
Peterson, B.J., Fry, B., 1987. Stable isotopes in impact of the maritime fur trade on nearshore
ecosystem studies. Annu. Rev. Ecol. Syst. 18, ecosystems in southern Haida Gwaii (British
293-320. Columbia, Canada): evidence from stable isotope
Polo de Ondegardo, J., 1571. Relación de los fundamentos acerca del notable daño que analysis of rockfish (Sebastes spp.) bone
resulta de no guarar a los Indios sus fueros. Coleccion de Libros y Documentos collagen. Archaeol. Anthropol. Sci. 5, 159-182.
referentes a la Historia del Peru 1, 45-188. Szpak, P., White, C.D., Longstaffe, F.J., Millaire,
Power, E., 2008. Furry families: making a J.-F., Vásquez Sánchez, V.F., 2013b. Carbon and
human-dog family through home. Soc. Cul. nitrogen isotopic survey of northern Peruvian
Geogr. 9, 535-555. plants: baselines for paleodietary and
Prummel, W., 2006. Bronze age dogs from graves paleoecological studies. PLoS ONE 8, e53763.
in Borger (Netherlands) and Dimini (Greece). In: Szpak, P., Longstaffe, F.J., Millaire,J.-F., White,
Snyder, L.M., Moore, E.A. (Eds.), Dogs and C.D., 2014. Large variation in nitrogen isotopic
People in Social, Economic or Symbolic composition of a fertilized legume. J. Archaeol.
Interaction, Working. Oxbow Books, Oxford, pp. Sci. 45, 72-79.
67-76. Szpak, P., Millaire, J.-F., White, C.D., Bourget, S., Longstaffe, F.J., in press. Life
R Development Core Team, 2007. R: A Histories of Sacrificed Camelids from Huancaco (Virú Valley). In: Klaus, H.D., Toyne,
Language and Environment for Statistical J.M. (Eds.), Reconstructing Sacrifice on the North Coast of Peru: Archaeological Studies
Computing. R Foundation for Statistical of Ritual Violence in the Ancient Andes. University of Texas Press, Austin.
Computing, Vienna, Austria. Thornton, E.K., Defrance, S.D., Krigbaum, J.,
Reitz, E.J., Wing, E.S., 2008. Zooarchaeology, Williams, P.R., 2011. Isotopic evidence for
second ed. Cambridge University Press, Middle Horizon to 16th century camelid herding
Cambridge. in the Osmore Valley, Peru. Int. J. Osteoarchaeol.
Russell, N., 2012. Social Zooarchaeology: 21, 544-567.
Humans and Animals in Prehistory. Cambridge Tieszen, L.L., Boutton, T.W., 1989. Stable carbon
University Press, Cambridge. isotopes in terrestrial ecosystem research. In:
Russell, N., During, B.S., 2006. Worthy is the Rundel, P.W., Ehleringer, J.R., Nagy, K.A.
lamb: a double burial at Neolithic ^atalhoyük (Eds.), Stable Isotopes in Ecological Research.
(Turkey). Paléorient 32, 73-84. Springer-Verlag, New York, pp. 167-195.
Salmi, A.-K., Áikas, T., Lipkin, S., 2011. Tomka, S.A., 1992. Vicuñas and llamas: parallels
Animating rituals at Sámi sacred sites in northern in behavioral ecology and implications for the
Finland. J. Soc. Archaeol. 11, 212-235. domestication of Andean Camelids. Hum. Ecol.
Schmidt, S.N., Olden, J.D., Solomon, C.T., 20, 407433.
Zanden, M.J.V., 2007. Quantitative approaches to Verano, J.W., DeNiro, M.J., 1993. Locals or
the analysis of stable isotope food web data. foreigners? Morphological, biometric and isotopic
Ecology 88, 27932802. approaches to the question of group affinity in
Schoeninger, M.J., DeNiro, M.J., 1984. Nitrogen human skeletal remains recovered from unusual
and carbon isotopic composition of bone collagen archaeological context. In: Sandford, M.K. (Ed.),
from marine and terrestrial animals. Geochim. Investigations of Ancient Human Tissue:
Cosmochim. Acta 48, 625-639. Chemical Analysis in Anthropology. Gordon and
Shanklin, E., 1985. Sustenance and symbol: Breach, Langhorne, pp. 361-386.
anthropological studies ofdomesticated animals. Vitousek, P.M., Field, C.B., Matson, P.A., 1990. Variation in foliar 5 C
13
Ann. Rev. Anthropol. 14, 375-403. in Hawaiian Metrosideros polymorpha: a case of
Shimada, I., 1994. Pampa Grande and the internal resistance? Oecologia 84, 362-370.
Mochica Culture. University of Texas Press, Wheeler, J.C., Russel, A.J.F., Redden, H., 1995.
Austin. Llamas and alpacas: pre-conquest breeds and
Shimada, M., Shimada, I., 1985. Prehistoric llama post-conquest hybrids. J. Archaeol. Sci. 22, 833-
breeding and herding on the north coast of Peru. 840.
Am. Antiq. 50, 3-26. White, C.D., Nelson, A.J., Longstaffe, F.J.,
Shipman, P., 2010. The animal connection and Grupe, G., Jung, A., 2009. Landscape bioarchaeology at
human evolution. Curr. Anthropol. 51, 13 519-538 . Pacatnamu, Peru: inferring mobility from 5 C and
13
Smith, B.N., Epstein, S., 1971.Two categories of C/ C ratios 5 N values of hair. J. Archaeol. Sci. 36,1527-
12 15
for higher plants. Plant Physiol. 47, 380-384. 1537. (continued on
nextpage)
22 P. Szpak et al./Journal of Anthropological Archaeology 36 (2014) 110-129

Whittle, A., 2003. The Archaeology of People:

Dimensions of Neolithic Life. Routledge, London.
Willey, G.R., 1953. Prehistoric Settlement
Patterns in the Virú Valley, Perú. Smithsonian
Institution, Washington, DC.
Wilson, D.J., 1988. Prehispanic settlement
patterns in the lower Santa Valley, Peru: a
regional perspective on the origins and
development of complex North Coast society.
Smithsonian Institution Press, Washington, DC.
Yakir, D., 2011. The paper trail of the C of atmospheric
13
CO2 since the industrial revolution period.
Environ. Res. Lett. 6, 034007.