Pottery production, regional exchange, and state collapse during the Middle Horizon (A.D....
Transcript of Pottery production, regional exchange, and state collapse during the Middle Horizon (A.D....
Pottery production, regional exchange, andstate collapse during the Middle Horizon(A.D. 500–1000): LA-ICP-MS analyses ofTiwanaku pottery in the Moquegua Valley, Peru
Nicola Sharratt1, Mark Golitko2, P. Ryan Williams2
1Georgia State University, Atlanta, Georgia, 2Field Museum of Natural History, Chicago, Illinois
During the Middle Horizon (A.D. 500–1000), the Tiwanaku state dominated the south central Andes. Theproduction and circulation of goods were important components of statecraft. To date, studies of themovement of pottery vessels across the Tiwanaku realm have relied on stylistic analyses. This paperpresents results of Laser Ablation Inductively Coupled Plasma Mass Spectrometry (LA-ICP-MS) analysesof ceramics from the largest Tiwanaku province in the Moquegua Valley, Peru. Comparison of the derivedcompositional data with an existing chemical database of Moquegua Valley clays demonstrates that inaddition to local production, non-local ceramic vessels were being brought into the valley during theheight of Tiwanaku authority. A lower percentage of imported ceramics was identified in ceramicassemblages dating to the wake of Tiwanaku state collapse (ca. A.D. 1000). Long-distance exchangeendured despite political breakdown but there were alterations in the particular networks in whichpost-collapse communities participated.
Keywords: production, exchange, LA-ICP-MS, pottery, state collapse, Andes, Tiwanaku
IntroductionThe development of widespread, complex exchange
networks is apparent in ancient states and empires
around the world (Brumfiel and Earle 1987;
D’Altroy and Earle 1985; Earle and Ericson 1977;
Oka and Kusimba 2008). While exchange networks
are not necessarily dependent on imperial infrastruc-
ture and often thrive outside the auspices of state
authority (Nakassis et al. 2011; Parkinson 2010;
Parkinson and Galaty 2009), the movement of
goods across states and empires was an important
component of ancient political organizations, serving
as the basis for elaborate tribute systems, providing a
currency for rewarding loyal individuals, and acting
as a vehicle for materializing and spreading ideology
(D’Altroy et al. 1994; DeMarrais et al. 1996).
Conversely, the breakdown of regional exchange
networks and the concomitant decline in the long-
distance movement of goods are commonly cited as
consequences of state collapse (Renfrew 1979;
Schwartz and Nichols 2006; Tainter 1988; Yoffee
and Cowgill 1988).
The Tiwanaku state (A.D. 500–1000) developed in
the Lake Titicaca Basin in South America and its
far-reaching cultural influence is evident in the pre-
sence of Tiwanaku-style materials in pockets across
the south central Andes. The production and move-
ment of resources and craft goods were central to
the functioning of the Tiwanaku state, provisioning
the capital with crops that were important to reli-
gious and political practices but that did not grow
in the high-altitude homeland, acting as a means
through which hinterland and provincial Tiwanaku
communities asserted their affiliation with the state
center, and mediating ties with non-Tiwanaku elites
on the edges of the state’s sphere of influence (Ber-
mann 1994; Goldstein 1985, 2005; Janusek 2002;
Kolata 1993a; Torres-Rouff 2008; Torres and Con-
klin 1995; Stanish et al. 2010).
Ceramic vessels were a critical component of
Tiwanaku economy and ideology (Janusek 2002).
Central to the feasting and drinking at the heart of
Tiwanaku politics and ritual, pottery was also an
important canvas for state affiliated motifs (Janusek
2003b). As Tiwanaku economic, social and religious
sway extended across the south central Andes, the
geographical range of Tiwanaku-style ceramic
forms and decorative repertoires also increased
Correspondence to: Nicola Sharratt, Department of Anthropology,Georgia State University, P.O. Box 3998, Atlanta, G.A., 30302. Email:[email protected]
� Trustees of Boston University 2015MORE OpenChoice articles are open access and distributed under the terms of theCreative Commons Attribution Non-Commercial License 3.0DOI 10.1179/2042458214Y.0000000001 Journal of Field Archaeology 2015 VOL. 00 NO. 00 1
(Anderson 2013; Goldstein 1985, 2005; Kolata
1993a; Korpisaari and Parssinen 2011; Stovel 2001).
Although stylistic analyses of pottery are commonly
used to chart the spread of Tiwanaku artistic
canons across the region, chemical data provide an
empirical means for determining the movement of
ceramic material around this ancient Andean state.
Laser Ablation Inductively Coupled Plasma Mass
Spectrometry (LA-ICP-MS) analysis was conducted
on pottery from the Moquegua Valley, the largest
Tiwanaku province, located approximately 300 km
from the state capital. The results indicate that
during the height of Tiwanaku state authority in
Moquegua (A.D. 725–1000), potters in the valley lar-
gely relied on locally available clays to replicate craft-
ing traditions from the heartland. However, state
period ceramic assemblages from the Moquegua
Valley also contain a few imported vessels. These
non-local vessels are chemically distinct from one
another raising the possibility that the Moquegua
Valley colony had access to material from a range
of foreign locations. Although almost all of the
post-collapse (post A.D. 1000) ceramics analyzed
during the study were determined to be locally pro-
duced, one non-local sherd was identified in a post-
collapse assemblage. Interestingly, this sherd was
chemically distinct from the imported ceramics ident-
ified for the state period, indicating that the disinte-
gration of Tiwanaku political authority in the south
central Andes did not lead to the complete cessation
of long-distance movement of goods. Instead the
data suggest that the networks of exchange in
which this provincial Tiwanaku community partici-
pated changed.
The TiwanakuBeginning around A.D. 500, the Tiwanaku emerged
as the dominant polity in the altiplano, a cold,
windy, high altitude region in the Lake Titicaca
Basin (Bauer and Stanish 2001; Janusek 2008;
Kolata 1993a; Stanish 2003). The nature and work-
ings of the Tiwanaku state have been subject to con-
siderable scholarly debate since the late 19th century
but increasingly researchers agree that the Tiwanaku
state was both powerful and hierarchical, and that
elites manipulated existing Andean concepts of reci-
procity and social organization to draw together
local groups and assert authority over the region
(Albarracin-Jordan 1996; Goldstein 2005; Janusek
2004a, 2004b, 2013; Kolata 1993b, 2003; McAndrews
et al. 1997; Squier 1877; Stanish 2003, 2013).
The state capital, also called Tiwanaku, was
located near the shores of Lake Titicaca. By A.D.
800, in addition to monumental religious and elite
structures, the city of Tiwanaku was home to thou-
sands of residents (Couture 2003; Couture and
Sampeck 2003; Janusek 1999, 2003a, 2004a, 2008;
Vranich 1999, 2001, 2006). As Tiwanaku ideology
spread, communities in the immediate hinterland
were incorporated into the state (Albarracin-Jordan
1996; Albarracin-Jordan and Mathews 1990; Bandy
2001; Bermann 1994, 1997; Janusek and Kolata
2003; Stanish 2002, 2003; Stanish et al. 2005). Over
time, incorporation spread across the south central
Andes as migrants from the Tiwanaku heartland
founded distant colonies in agriculturally productive
regions (Anderson 2013; Blom et al. 1998; Goldstein
1989b, 1993a, 1993b, 2005; Knudson et al. 2004;
Torres-Rouff 2008) (FIG. 1).
The largest Tiwanaku settlement outside the Titi-
caca Basin was 300 km away, in the Moquegua
Valley (Goldstein 1989a, 1989b, 2005, 2013)
(FIG. 2). The Moquegua Valley offered economic
advantages to the state. With irrigation, the valley
is highly productive and culturally important crops
that do not grow in the altiplano, particularly
maize, thrive in Moquegua. Early Tiwanaku arrivals
to the valley were pastoralists whose presence is
archaeologically visible in temporary, tent-like struc-
tures that date to A.D. 525–700 (Goldstein 2005).
It was with a second wave of immigrants who
began arriving around A.D. 725 that permanent,
long lasting Tiwanaku settlements were established
Figure 2 The Moquegua Valley. Major Tiwanaku state
installations were located on the low hills rising above the
valley floor.
Figure 1 Location of the Tiwanaku provinces in the south
central Andes during the Middle Horizon (A.D. 500–1000).
Sharratt et al. Pottery production, regional exchange, and state collapse during the Middle Horizon (A.D. 500– 1000)
2 Journal of Field Archaeology 2015 VOL. 00 NO. 00
in Moquegua (Goldstein 1985, 1989b, 1993a, 1993b,
2005). Strontium isotope analyses of teeth as well
as analyses of non-metric cranial and dental traits
support the claim that the Moquegua Valley was
colonized by immigrants from the altiplano (Blom
1999; Knudson et al. 2004; Sutter 2000; Sutter and
Sharratt 2010).
These immigrants and their descendants occupied
the Tiwanaku towns, principally Chen Chen and
Omo M10, for the following three centuries. They
served the agricultural needs of the state, supplying
the heartland with maize and other desired crops.
The Tiwanaku population in the Moquegua Valley
maintained close cultural as well as economic ties
with the altiplano. They built houses similar to
those in the altiplano (Goldstein 1989a, 1989b,
1993a, 2005), constructed provincial versions of alti-
plano religious structures, and they carried out the
same kinds of rituals as heartland communities
(Goldstein 1993b, 2005). Replication included porta-
ble material culture. Ceramic vessels recovered from
Tiwanaku sites in the Moquegua Valley were crafted
in heartland forms. They were decorated with tra-
ditional Tiwanaku motifs, including the Staff God,
pumas, trophy heads and geometric designs (FIG. 3).
The Moquegua Valley is of particular interest to
archaeologists because during the Middle Horizon,
in addition to the Tiwanaku colony, it was home to
the southernmost Wari outpost. Wari sites are
located at slightly higher altitudes than Tiwanaku
settlements, with the majority found in the narrow
upper valley where Wari agriculturalists cultivated
the steep hills using irrigation techniques imported
from the Wari heartland in the Ayacucho Valley
(Moseley et al. 2005; Nash 2002; Nash and Williams
2009; Williams 1997, 2001, 2002). Despite the recent
discovery of a Tiwanaku temple at the principal Wari
site of Cerro Baul (Williams 2013), considerable
debate remains about the nature and degree of inter-
action between the two immigrant populations
(Goldstein 2013).
Tiwanaku political authority over the region went
into decline beginning around A.D. 1000. Tiwanaku
state fragmentation was a violent, drawn-out process
that had far reaching and long lasting impacts
(Bermann et al. 1989; Graffam 1992; Janusek 2005,
2008; Kolata and Ortloff 2003; Ortloff and Kolata
1993; Owen 2005). Although scholars continue to
debate about whether state collapse was the result
of drought (Kolata and Ortloff 2003), the actions
of competing polities (Williams 2002), or internal fac-
tionalism (Janusek 2005, 2008), it is clear that state
collapse radically altered the political and material
landscape of the capital, the heartland, and the pro-
vinces. At the city of Tiwanaku, large-scale construc-
tion ceased, the monumental and residential core was
mostly deserted, and elite complexes were razed to
the ground (Couture and Sampeck 2003; Vranich
2006). Hinterland populations declined and the
four-tier settlement hierarchy of the state period
was replaced by a pattern of dispersed small villages
(Albarracin-Jordan 1996; Bandy 2001; Bermann
1994; Janusek and Kolata 2003; Stanish 2003).
Abandonment, destruction, and dispersal occurred
across the Tiwanaku realm. In the Moquegua Valley,
violence was particularly directed toward manifes-
tations of state power and symbols of religious ideol-
ogy. Monumental architecture and corporate storage
Figure 3 A Tiwanaku kero (A) and pitcher (B) recovered
from cemeteries at the site of Chen Chen in the Moquegua
Valley.
Sharratt et al. Pottery production, regional exchange, and state collapse during the Middle Horizon (A.D. 500– 1000)
Journal of Field Archaeology 2015 VOL. 00 NO. 00 3
facilities were torn apart at the administrative centers
of Omo M10 and Chen Chen. Elite burials were ran-
sacked and idols were smashed (Goldstein 1993b,
2005). The immigrant population of Tiwanaku
towns fled the large colonial settlements in the
middle section of the Moquegua Valley and estab-
lished new, smaller villages along the coast and
up-valley from the colony’s traditional territory
(Bermann et al. 1989; Owen 2005; Sims 2006; Stanish
1989) (FIG. 4).
Tiwanaku CeramicsTiwanaku ceramic assemblages have been the subject
of considerable visual analysis, in both the state
heartland and the provinces (Anderson 2013; Bur-
kholder 1997; Goldstein 1985; Isbell 2013; Janusek
1999, 2003b; Korpisaari and Parssinen 2011). These
studies have identified a Tiwanaku-wide style,
characterized by particular forms and decorative tra-
ditions. Heartland forms include keros, vase shaped
drinking vessels that held the chicha, or maize beer,
that was central to political and religious feasting;
escudillas (elaborate serving vessels); sahumadors
(ceremonial burners) and zoomorphic incensarios
used for burning; portrait vessels that depict male
individuals; tazones (flaring-sided bowls); one-
handled pitchers, as well as cooking and storage
vessels (Janusek 2003b). Among common decorative
motifs are felines and camelids as well as geometric
imagery. The inclusion of images, in particular the
Staff God, also found on the monumental stone
architecture and sculptures that constituted the
built religious and political structures of the capital
city, indicates the importance of ceramic vessels as
portable canvases for the depiction of ideologically
potent iconography. Conversely, the inclusion of
keros in the carved imagery of those same sculptures
highlights the significance of ceramic vessels in visible
manifestations of elite power.
Ceramic vessels were also an important means for
hinterland and provincial communities to carve out
and assert their own particular identities in the
wider Tiwanaku realm. Although the principal heart-
land forms and decorative traditions are found at
Tiwanaku sites across the South Central Andes, pot-
ters in different regions crafted local forms or var-
iants as well. These include challadors (vessels with
narrow bases and tapered bodies), which are charac-
teristic of Cochabamba assemblages, and ‘‘coca-cola
glass’’ keros which are particular to the Moquegua
Valley (Goldstein 1985; Janusek 2003b). Within the
heartland, differences in ceramic assemblages dis-
tinguish communities in the Titicaca Basin from
one another, and even more locally, they distinguish
neighborhoods within the state capital (Bermann
1994; Janusek 1999, 2002).
Particularities of form and decoration have been
utilized in visual identification of non-local styles in
ceramic assemblages. At the site of Tiwanaku non-
local ceramics are predominantly from the eastern
valleys and include Cochabamba vessels in the
Ch’iji Jawira neighborhood (Rivera Casanovas
2003) and Yampara and Omereque styles in
Akapana East 2, where non-local style wares comprise
between 5 and 10% of analyzed ceramics (Janusek
2004a). In the Moquegua Valley, Goldstein (2005)
Figure 4 Maps showing the location of Tiwanaku sites before (left) and after (right) state collapse ca. A.D. 1000.
Sharratt et al. Pottery production, regional exchange, and state collapse during the Middle Horizon (A.D. 500– 1000)
4 Journal of Field Archaeology 2015 VOL. 00 NO. 00
argues for the presence of imported ceramics based on
their similarity with those from the Titicaca Basin.
The movement of prestige ceramic vessels makes
sense in the broader context of resource exchange in
the Tiwanaku state, in which low-lying provincial
communities supplied the state center with agricultural
produce, as well as feathers, coca leaves, snuff and
other hallucinogenic drugs (Kolata 1993a). However,
these visual identifications of non-local ceramics in
varied Tiwanaku locales lack the empirical evidence
necessary for demonstrating that vessels, and not
just styles, were imported to a community.
Geochemical Studies of Ceramic Exchange inthe Archaeological RecordChemical characterization of archaeological
materials is now routinely used to investigate the pro-
duction and movement of goods in the past, and has
been particularly effectively applied to studies of
obsidian and pottery as well as to metals and glass
(Arnold et al. 1991; Burger et al. 1994; Burger et al.
1998a, 1998b, 1998c; Burger 2000; Burger and Glas-
cock 2000; Cecil 2004; Glascock et al. 2007; Gliozzo
and Memmi Turbanti 2004; Gratuze 1999; Kennett
et al. 2004; Kennett et al. 2001; Mallory-Greenough
et al. 1998; Mirti et al. 2004; Niedershlag et al.
2003; Perez-Arantegui et al. 1996; Shortland 2002;
Tykot 1997; Vaughn and Neff 2004). Studies identi-
fying the presence of non-local goods contribute to
interpretations of long-distance trade and interaction
in a wide geographical and temporal range of archae-
ological contexts.
Several different techniques are utilized in these
studies, among them X-ray Fluorescence (XRF),
Instrumental Neutron Activation Analysis (INAA),
and Inductively Coupled Plasma Optimal/Atomic
Emission Spectrometry (ICP-OES/AES). However,
ICP-MS (Inductively Coupled Plasma Mass Spec-
trometry) has particular advantages. Compared
with INAA, ICP-MS is cost effective and requires
minimal sample preparation, yet still allows for the
rapid collection of multi-element chemical data and
has lower detection limits (ppm for solid samples)
than XRF (Pollard et al. 2007). Further, when con-
ducted using laser ablation (LA-ICP-MS), the tech-
nique leaves a mark almost invisible to the human
eye. INAA, in contrast, requires destructive sampling
and powdering a few grams of a sample (Kennett
et al. 2001).
Compositional analysis of large samples of pottery
can identify the presence of vessels that appear
chemically distinct from the principal compositional
group. Using the ‘‘Criterion of Abundance’’ those
ceramics that are chemically distinct from the
majority are assumed to be imports (Bishop et al.
1982: 300–301). However, several studies have
sought to empirically connect finished ceramic vessels
with the area in which they were produced by
attempting to locate clays and compare their chemi-
cal signatures with those of archaeological ceramics
(Adan-Bayewitz and Perlman 1985; Bartlett et al.
2000; Dorais et al. 2004; Hein et al. 2004; Neff
et al. 1992; Phillips and Morgenstein 2002; Sharratt
et al. 2009; Sherriff et al. 2002; Strazicich 1998;
Vaughn and Neff 2004). Such comparisons are not
straightforward. The chemical composition of cer-
amics is affected by mixing with other clays, by the
removal of inclusions in natural clays and by the
addition of materials as temper (Arnold 2005;
Arnold et al. 1991; Pollard et al. 2007; Sillar 1996).
Nonetheless, comparison of the chemical signature
of finished pottery with that of natural materials rep-
resents a means of establishing whether or not all cer-
amic vessels in an assemblage were produced using
local clays.
One of the principal advantages of undertaking a
study of ceramic production and importation
among Tiwanaku assemblages in the Moquegua
Valley was that a geochemical study of clays in the
Moquegua Valley had already been carried out
(Sharratt et al. 2009). In this study, clays were col-
lected throughout the Moquegua Valley and five
chemically distinguishable clay groups were identified
using LA-ICP-MS (FIG. 5). Most of the clays were
collected from Quaternary alluvium and the differ-
ences in their chemical signatures are explained by
the five distinct geological formations from which
they eroded as follows. The Moquegua Formation,
which is an upper Tertiary formation of sedimentary
origin and the primary geological formation present
in the middle Moquegua Valley between 1000 and
1500 masl (Bellido and Landa 1998) is the likely
parent material for clays we called the ‘‘Moquegua
Valley Group.’’ Located on inter-valley formations
above 3000 masl, the Huallilas Formation is also an
Figure 5 Chemically distinguishable clay groups identified
for the Moquegua Valley (after Sharratt et al. 2009).
Sharratt et al. Pottery production, regional exchange, and state collapse during the Middle Horizon (A.D. 500– 1000)
Journal of Field Archaeology 2015 VOL. 00 NO. 00 5
upper Tertiary formation and the likely parent
material for the clays labeled ‘‘Otora Valley 2.’’
Dating to the Upper Cretaceous are the Inagoya
and the Paralaque Formations, which are both
found between 2500 and 3000 masl. Of the two, Ina-
goya is of volcanic origin and is the likely source for
the ‘‘Torata Valley Group’’ clays, while Paralaque is
of sedimentary origin and likely the source for
‘‘Otora Valley Group 1’’ clays. Finally, the Toque-
pala Group Intrusive rocks are of early Tertiary
age and clays attributed to material eroded from
these rocks are in the ‘‘Tumilaca Valley Group.’’
Study SamplesThe present study was intended to compositionally
compare state period and post-collapse Tiwanaku
ceramics from Moquegua with locally available
clays in order to determine the presence of imported
pottery and to investigate changing patterns of
resource procurement in the valley. This necessitated
using the same methodology, LA-ICP-MS, as that
utilized in the earlier clay study (Sharratt et al.
2009), and therefore required exporting ceramic
sherds to the U.S.A. Other analytical methodologies,
particularly portable X-ray fluorescence (p-XRF),
have the advantage that they can be conducted at
the location of curation, meaning that large samples,
including entire vessels which would be impossible to
export from Peru, can be analyzed. P-XRF analyses
were undertaken of Tiwanaku pottery in Moquegua
from the sites of Chen Chen and Tumilaca la
Chimba (Schur 2011; Sharratt 2011a). In one
(Schur 2011), 229 vessels ascribed to the height of
state authority were analyzed with p-XRF. In the
other (Sharratt 2011a), 192 different state period
vessels and 78 post-collapse sherds were analyzed
using p-XRF. However, because p-XRF measures
fewer elements and has higher detection limits than
ICP-MS, it proved impossible to differentiate the
five locally available clay groups in the Moquegua
Valley using p-XRF. Given the importance of associ-
ating ceramics with the particular clay group used in
their production, the chemical compositions of cer-
amic materials from the same two Moquegua Valley
sites included in the p-XRF studies (Chen Chen and
Tumilaca la Chimba) were analyzed using LA-ICP-
MS (FIG. 4).
Chen Chen dates to the height of Tiwanaku state
authority in the Moquegua Valley with AMS dates
falling between A.D. 700 and 1030 (Goldstein 2005;
Sharratt 2011a). It was the largest Tiwanaku site in
the valley. As well as extensive residential sectors,
the site consists of agricultural fields and substantial
storage facilities (Goldstein 2005; Williams 1997,
2002). The site likely represents a major state instal-
lation that played a significant role in the production
and distribution of the agricultural goods sought by
the Tiwanaku state. The large cemeteries may have
included as many as 13,000 individuals (Owen
1997). The mortuary population was biologically
related to populations in the Tiwanaku heartland
and included immigrants from the altiplano (Blom
1999; Knudson et al. 2004; Sutter and Sharratt
2010). In life, the inhabitants of Chen Chen mirrored
altiplano heartland daily and ritual practices, con-
structing and using public and private spaces as
well as material culture very similar to that found
in the Titicaca Basin.
The ceramic assemblage from Chen Chen is
characteristic of state period Tiwanaku pottery in
the Moquegua Valley, which has been most exten-
sively studied by Goldstein (1985; 1989a, 1989b;
2005). Ceramic vessels from the height of the Tiwa-
naku occupation include a number of forms includ-
ing keros, tazones, one-handled pitchers and ollas
(cooking and storage vessels). Fine-ware was typi-
cally red-slipped with polychrome decoration in
blocky geometric motifs, as well as anthropomorphic
and zoomorphic images, particularly birds, felines
and camelids. Goldstein identifies considerable stan-
dardization in ceramic production at Tiwanaku
sites in the Moquegua Valley, noting that vessels
cluster into particular ranges of size and volume
and were produced in forms that were easy to stack
for storage and transport. He argues that standardiz-
ation in form and decoration are suggestive of cer-
amic workshops (Goldstein 2005). Wari style
ceramics have been recovered from graves at Chen
Chen, adding weight to debates about interactions
between Tiwanaku and Wari in Moquegua, but
they are rare and none were included in the sample
analyzed in this study (Garcia Marquez 1990;
Sharratt 2011a).
Ceramic material was also analyzed from the site
of Tumilaca la Chimba. This site was one of several
established following the abandonment of Tiwanaku
state administrative centers in Moquegua. Located
15 km up-valley from Chen Chen, Tumilaca la
Chimba is a much smaller site, although it is one of
the larger post-collapse settlements in the valley.
It consists of a residential sector and four cemeteries.
The collapse phase Tiwanaku occupation of the site
lacks large, public ritual spaces and central adminis-
trative or storage facilities, in common with other
post-collapse sites in the Moquegua Valley (Bawden
1993; Bermann et al. 1989; Goldstein 2005).
Radiocarbon dates indicate that Tumilaca la
Chimba was occupied between A.D. 950 and 1250
(Sharratt 2011a). Analyses of non-metric dental
traits from individuals buried at Chen Chen and
Tumilaca la Chimba demonstrate that the two popu-
lations shared an ancestor-descendent relationship,
Sharratt et al. Pottery production, regional exchange, and state collapse during the Middle Horizon (A.D. 500– 1000)
6 Journal of Field Archaeology 2015 VOL. 00 NO. 00
supporting the longstanding hypothesis that inhabi-
tants of Tumilaca la Chimba were refugees and
their descendants, who abandoned state centers
when the Tiwanaku state began to collapse (Sutter
and Sharratt 2010).
Distinctions between state period and collapse
phase pottery in the Moquegua Valley are often
expressed as differences in quality (Bermann et al.
1989; Goldstein 1985, 2005) and comparison of the
Tumilaca la Chimba and Chen Chen assemblages
meets with these generalizations (Sharratt 2011a).
Ceramic material recovered from excavations in the
cemeteries and in domestic structures at Tumilaca
la Chimba is stylistically similar to that at
Chen Chen (FIG. 6). However, although state period
forms (such as keros and tazones) were produced
and used, there is a greater range in size and
volume and keros are on average larger than their
state period predecessors. Similar decorative reper-
tories were utilized. Decorated pottery is typically
red-slipped, although reds tend to be darker than
those at Chen Chen. However, surface treatments
appear more hurried and fewer colors were used in
decoration than in the Chen Chen assemblage. Ico-
nographic repertories at Tumilaca la Chimba include
both the maintenance and rejection of state period
motifs. Pottery from Tumilaca la Chimba displays
more geometric motifs relative to anthropomor-
phized imagery. Significantly although many motifs
were maintained in modified forms, other state
period images, most notably the Staff God, are
absent from post-collapse ceramic assemblages in
the Moquegua Valley.
The context of production for collapse phase cer-
amics remains uncertain. Although pigments have
been tentatively identified on the patio of one exca-
vated residence at Tumilaca la Chimba, to date no
clear evidence for pottery production has been
found during fieldwork at the site, or during the
smaller excavations undertaken at other collapse
phase sites. However, the greater internal variation
in collapse phase assemblages with regard to form,
decoration, slip color and surface treatment, coupled
with imprecise execution have been taken as evidence
for a shift from workshop to domestic production
in the wake of state breakdown (Bawden 1989;
Bermann et al. 1989).
MethodsForty-five ceramic sherds from Chen Chen and 49
from Tumilaca la Chimba were exported to the
U.S.A. for compositional analysis. All analyzed
material from Chen Chen was excavated during
rescue excavations at the site’s cemeteries by Bruce
Owen in 1995 and by Romulo Pari Flores in 2002
(Owen 1997; Pari Flores et al. 2002). All ceramic
material from Tumilaca la Chimba analyzed during
this study was excavated from mortuary and residen-
tial contexts under the auspices of Proyecto Arqueo-
logico Cerro Baul in 2006/2007 (Sharratt 2011a;
Sharratt et al. 2012; Williams 2008). At both sites,
ceramic sherds as well as complete vessels were
placed in graves (Sharratt 2011a). The samples ana-
lyzed in this study therefore included small fragments
that had already broken off entire vessels and sherds
unattributed to larger vessels from the cemeteries.
The samples from both sites also included decorated
and undecorated sherds.
Sherds were subjected to LA-ICP-MS at the
Elemental Analysis Facility (EAF) at the Field
Museum of Natural History in Chicago, the same
facility used to analyze clays from the Moquegua
Valley (Sharratt et al. 2009). Protocols established
for the Field Museum’s facility were used for the
LA-ICP-MS analysis (Dussubieux et al. 2007; Golitko
and Terrell 2012; Niziolek 2013; Vaughn et al. 2011).
A Varian ICP-MS instrument equivalent to the
Varian 810 instrument was used. The Varian is a quad-
ropole mass spectrometer. Quadropole mass filters are
appropriate for trace element measurement because
they rapidly scan a wide mass range (Pollard et al.
2007). In the Field Museum machine, the ion beam
is bent 90uu by a series of lenses before it enters the
quadropole, increasing the sensitivity of the instru-
ment 200 times (Elliot et al. 2004). The facility at the
Field Museum uses a New Wave UP213 (Helium car-
rier gas, 213 nm laser operated at 0.2 mJ and a pulse
frequency of 15 Hz) laser in conjunction with the
ICP-MS to introduce solid samples.Figure 6 A kero excavated from a burial at Tumilaca la
Chimba.
Sharratt et al. Pottery production, regional exchange, and state collapse during the Middle Horizon (A.D. 500– 1000)
Journal of Field Archaeology 2015 VOL. 00 NO. 00 7
The ceramic samples were ablated using the laser,
with a spot size of 150 microns, and a dwell time of
90 seconds. The laser was focused on a broken
edge of a sherd, rather than the exterior or interior
surface, so that analysis concentrated on pastes, not
slips or paints. It was positioned to avoid large
temper grains so that analysis focused on the chemi-
cal signature of the clays used in ceramic production.
Each sample was ablated 10 times, with the laser
repositioned in a new place on the sherd each time,
and a total of 55 elements were measured, using29Si as an internal standard to control for variability
in time efficiency and resulting signal strength.
Chemical concentrations were calculated using
National Institute of Standards and Technology
(NIST) standards n610, n612 and Brick Clay
(n679), in line with established protocols (Gratuze
et al. 2001). Error values were established using ana-
lyses of New Ohio Red Clay, which was subjected to
the same protocols as the ceramic samples. Error esti-
mates were similar to those reported by Sharratt and
colleagues (2009).
Statistical ProceduresPrior to analysis, elements which are known tomeasure
poorly on the EAF ICP-MS instrument due to oxide
interferences or high ionization energies were removed
from consideration. These included Cl and As.
Additionally, several elements that measured close to
instrumental detection limits with poor precision—
Ag, In, and Bi—were also removed from the analysis.
Finally, Cu measurements displayed consistent differ-
ences across analyses, possibly due to memory effects
associated with other projects undertaken at the EAF
during the duration of analysis, and consequently, Cu
was also removed from consideration.
All remaining chemical measurements were first
converted to base-10 logarithms to normalize their
distribution and eliminate scaling differences between
high and low abundance elements. Initial pattern rec-
ognition was achieved by first performing an R-Q
mode factor analysis on the correlation matrix
(Baxter 1992; Neff 1994). Factor scores for all
samples were then subjected to a hierarchical cluster
analysis using the average linkage method on
squared Euclidean distances between ceramic sherds
(Shennan 1997). Both Hierarchical Cluster Analysis
(HCA) and visual inspection of Principal Component
Analysis (PCA) plots and bivariate elemental plots
resulted in the identification of two primary chemical
groups, here referred to as Groups 1 and 2, and three
distinct outlier chemical profiles, referred to as Out-
lier Types 1–3 (FIGS. 7, 8, 9). Relative to Group 1
sherds, Group 2 ceramics exhibit higher Al, Nb,
Ba, and light Rare Earth Elements (REE) concen-
trations. The single sherd included in Outlier Type
1 is distinguished by very high concentrations
of Fe, Sb (and As, although this was not included
in formal statistical analysis), and U content.
Outlier Type 2 sherds are characterized by very
Figure 8 Bivariate plot of logged (base 10) antimony and
aluminum concentrations in measured ceramics and clays.
Ellipses delimit 90% confidence intervals for Groups 1 and 2.
Figure 7 R-Q mode biplot of factor 1 and 2 scores for
measured ceramics and clays showing two group structure
in the data and three identified outlier chemical types.
Ellipses delimit 90% confidence intervals for Groups 1 and 2.
Figure 9 Bivariate plot of logged (base 10) zinc and cesium
concentrations in measured ceramics and clays. Ellipses
delimit 90% confidence intervals for Groups 1 and 2.
Sharratt et al. Pottery production, regional exchange, and state collapse during the Middle Horizon (A.D. 500– 1000)
8 Journal of Field Archaeology 2015 VOL. 00 NO. 00
high Sb and Cs content, and Outlier Type 3 sherds by
high Zn content. Refinement of Groups 1 and 2 pro-
ceeded by calculation of jackknifed Mahalanobis dis-
tance-based probabilities of group membership (Neff
2002), which resulted in a number of sherds remain-
ing unassigned. These all appear to be outliers of
Groups 1 and 2, rather than reflecting the presence
of additional distinct chemical profiles. Additionally,
Outlier Types 1–3 were compared to Groups 1 and 2
and found to have negligible probabilities of mem-
bership (v1%) in either group.
In a prior publication (Sharratt et al. 2009), we
argued that most ceramics analyzed from Chen Chen
could be best matched to clays locally available in the
Moquegua Valley, although statistical overlap with
clays from upland sources in the Otora valley were
noted. Group 1, or the ‘‘Tiwanaku’’ chemical group
we identified in that earlier study, displays the same
associations, overlapping statistically primarily with
clays collected in the middle Moquegua Valley and
included in the ‘‘Moquegua Valley’’ chemical group,
but also with some clays in the Otora Valley 1 clay
chemical group (TABLE 1). With the expanded
number of sherds now included in Group 1, there is
also a small, but non-negligible, overlap with clays
included in the Torata Valley chemical group, also
located in the uplands above the middle valley. As we
previously suggested, this may reflect use of clay
sources near Chen Chen, where eroded material from
the uplands mixes with material eroding from the
slopes of the middle Moquegua Valley.
Group 2 ceramics, present only at Tumilaca la
Chimba, are more statistically similar to clays available
fromupland sources in theOtora,Torata, andTumilaca
valleys, although the small number of samples included
inGroup 2make statistical associations somewhat tenu-
ous (TABLE 1). To constrain the impact of small sample
size on estimates of confidence intervals forGroup 2 cer-
amics (Baxter 2001: 135; Harbottle 1976: 58), compari-
sons were made using only the first four factor scores.
However, these ceramic sherds chemically match some
ceramics we had previously analyzed from the nearby
Wari settlements of Cerro Baul and Cerro Mejia,
suggesting that post-collapse potters at Tumilaca la
Chimba may in some cases have utilized upland sources
similar to those used by Moquegua-area Wari potters.
While this suggests the occasional use of clays located
closer toTumilaca laChimba, it is clear that themajority
of ceramics deposited there were produced from clays
indistinguishable fromthoseusedby state-periodpotters
residing at Chen Chen.
The outlier type sherds do not match concentrations
found in any Moquegua area clays studied to date.
While Tumilaca Valley clays are also characterized
by high Cs content, they are otherwise very dissimilar
to Outlier Type 2 sherds, and do not display the high
Sb contents observed in those ceramics. This suggests
that the outlier type sherds were all produced in areas
of distinctive geochemistry and subsequently trans-
ported into the Moquegua area.
ResultsChen ChenEarlier p-XRF analyses of large samples from
Chen Chen indicated that ceramics recovered from
Table 1 Group membership probabilities for ceramicgroups 1 and 2 relative to clay chemical groups identifiedfor the Moquegua Valley.
MembershipProbability
Sample Clay chemical group Group 1* Group 2**
MVC001 Moquegua Valley 0.00 0.16MVC002 Moquegua Valley 0.01 0.20MVC003 Moquegua Valley 0.00 0.26MVC004 Moquegua Valley 0.03 2.70MVC005 Moquegua Valley 0.12 0.39MVC006 Moquegua Valley 0.00 0.13MVC007 Moquegua Valley 0.00 0.04MVC008 Moquegua Valley 0.06 0.19MVC009 Moquegua Valley 0.18 0.18MVC010 Moquegua Valley 10.83 0.09MVC011 Moquegua Valley 0.29 0.18MVC012 Moquegua Valley 0.01 0.12MVC013 Moquegua Valley 0.59 0.12MVC014 Moquegua Valley 0.00 0.03MVC016 Moquegua Valley 0.10 0.11MVC017 Moquegua Valley 0.00 0.04MVC018 Moquegua Valley 0.11 0.33MVC019 Moquegua Valley 4.92 0.53MVC020 Moquegua Valley 2.65 0.59MVC021 Moquegua Valley 0.06 0.99MVC024 Moquegua Valley 0.00 0.51MVC025 Moquegua Valley 0.05 0.46MVC026 Moquegua Valley 22.84 0.27MVC027 Moquegua Valley 0.41 0.31MVC028 Moquegua Valley 0.44 0.70MVC029 Moquegua Valley 0.23 0.44MVC031 Moquegua Valley 0.32 0.47MVC032 Moquegua Valley 0.03 0.49MVC033 Moquegua Valley 0.00 0.09MVC045 Moquegua Valley 0.20 0.35MVC039 Otora Valley 1 0.13 1.59MVC040 Otora Valley 1 22.05 0.78MVC041 Otora Valley 1 0.11 0.63MVC042 Otora Valley 1 0.00 2.05MVC046 Otora Valley 2 0.02 3.37MVC047 Otora Valley 2 0.00 0.70MVC048 Otora Valley 2 0.04 0.63MVC049 Otora Valley 2 0.06 1.67MVC034 Torata Valley 1.13 1.07MVC035 Torata Valley 0.00 1.19MVC036 Torata Valley 1.92 0.34MVC037 Torata Valley 2.30 0.16MVC043 Torata Valley 0.03 0.31MVC044 Torata Valley 3.37 0.30MVC038 Tumilaca Valley 0.00 0.75MVC050 Tumilaca Valley 0.00 0.46MVC051 Tumilaca Valley 0.00 2.34MVC052 Tumilaca Valley 0.00 0.20MVC015 Unassigned 0.00 0.03MVC030 Unassigned 0.04 0.14MVC031A Unassigned 0.00 0.18
*calculating using the first 12 factor scores ** calculated using
the first 4 factor scores
Sharratt et al. Pottery production, regional exchange, and state collapse during the Middle Horizon (A.D. 500– 1000)
Journal of Field Archaeology 2015 VOL. 00 NO. 00 9
the site were compositionally homogeneous, with the
majority grouping together chemically (Schur 2011;
Sharratt 2011a). Comparison of compositional data
derived from LA-ICP-MS of the Chen Chen ceramic
sample reported here with the five clay groups ident-
ified in the earlier clay survey confirms that the
majority (91.1%) of these sherds were produced using
locally available clays. Specifically, the vast majority
of analyzed sherds were likely crafted using clays
from the Moquegua Valley Group. Although clays
from this group are found throughout the alluvial
deposits in the middle Moquegua Valley, some clay
samples identified as this group were collected from
locations within 5 km of Chen Chen. Therefore,
Moquegua Valley clays were easily accessible to pot-
ters at the site of Chen Chen.
However, four sherds from Chen Chen were ident-
ified as outliers and were not made using locally
available clays. These four sherds were recovered
from a total of three different graves. Burials at
Chen Chen were arranged in spatially discrete ceme-
teries. There are as many as 35 different cemeteries.
They vary considerably in size and are thought to
represent intra-community groups, likely extended
family-or-kin based (Blom 1999; Sharratt 2011a).
The three graves containing non-local ceramic
material were in two of the six cemeteries included
in the sample (TABLE 2). From cemetery 33, a non-
local sherd (Outlier Type 3) was identified from a
broken tazon that was also included in the pit
burial of an adult female, aged 36–45 years old.
The tomb also contained a wooden spoon and a
corn cob.
Three non-local sherds from two different graves
were identified from cemetery 30. One fragment (Out-
lier Type 2)was recovered fromone of the raremultiple
interments at Chen Chen. This stone-lined cist
included the remains of an infant, a middle aged
adult of undetermined sex and a child aged 3–4 years
old. The grave also included a red-slipped tazon with
a simple black geometric design. However, the non-
local sherd was not from this vessel.
The final two non-local sherds were recovered
from the same grave, that of an adult male (36–45
years old). The grave was a stone-lined cist. The indi-
vidual was buried with a red-slipped tazon decorated
with a black painted zig-zag motif and with a
red-slipped kero with an image of a camelid repeated
four times. The complete kero in this grave is notable
because its form is particular to the Moquegua
Valley, known as a ‘‘coca-cola glass’’ kero. It was
included in the analyzed sample, and was determined
to have been made from locally available clays. The
grave also included ceramic sherds in addition to
the complete vessels. LA-ICP-MS analysis revealed
that although some of these were produced with
local clays, two were non-local sherds (Outlier Type
2 and Outlier Type 3). One of these is likely from a
modeled incensario. Thus, the grave contained com-
plete vessels produced with local clays, including a
kero that is stylistically particular to the Moquegua
Valley colony, but also non-local sherds.
Although it is not known where these non-local
sherds were imported from, they do not share the
same chemical signatures. This raises the possibility
that they have different points of origin. Of the
sherds described above, the sherd in cemetery 33
shared a chemical signature with one of those from
the male grave in cemetery 30. The other sherd
from the male grave in cemetery 30 shared a chemical
signature with the sherd from the multiple burial in
cemetery 30.
Tumilaca la ChimbaDespite up-rooting and settling some 15 km from
Chen Chen and other Tiwanaku towns, potters at
Tumilaca la Chimba largely continued to use clays
that are chemically indistinguishable from those
used by their state period predecessors. This is
notable because chemically distinct clays are found
within 5 km of Tumilaca la Chimba in the Tumilaca
drainage (Tumilaca Valley Group). However, given
that the site appears to be located on the very limit
of the Moquegua Valley Group’s geographical
range, the continued use of clays from that group
may have been the result of access and availability,
rather than a cultural choice to continue using the
same materials as their ceramicist forerunners.
Although made from the same clay group as Chen
Chen ceramics, the Tumilaca la Chimba sherds are
more chemically diverse. This was noted also
during p-XRF analyses (Schur 2011; Sharratt
2011a). Although this larger diversity mirrors the
greater range seen in visual analyses of collapse
Table 2 Burial contexts of outlier ceramic fragments from Chen Chen.
Specimen No. Vessel type Outlier type Cemetery Tomb type Human remains Other cultural inclusions
M1-303011A Undetermined 2 30 Stone-lined cist Adult Male (36–45) Coca-cola glass kero, tazonM1-303011E Incensario? 3M1-303037 Undetermined 2 30 Stone-lined cist Infant Tazon
Child (2–3)Adult (undetermined sex)
M1-331007 Tazon 3 33 Pit Adult Female (36–45) Wooden spoon, corn cob
Sharratt et al. Pottery production, regional exchange, and state collapse during the Middle Horizon (A.D. 500– 1000)
10 Journal of Field Archaeology 2015 VOL. 00 NO. 00
phase pottery, this chemical diversity could be
explained by the likely location of clay acquisition
by potters at Tumilaca la Chimba. The site is located
further up the Moquegua Valley than Chen Chen,
where clays, despite being in the Moquegua Valley
Group, may include a greater contribution from
clays eroded from the highlands, including those
characterized as the Otora 1 Group.
One analyzed sherd was determined to be non-
local (Outlier Type 1). This sherd was recovered
from the second largest of the four cemeteries at
Tumilaca la Chimba. It was included in the stone-
lined cist grave of an adolescent (12–15 years old)
of undetermined sex. No complete ceramic vessels
were recovered from the grave, and the non-local
sherd is a red-slipped body sherd from an unknown
vessel form. The grave also included two cactus
spines, commonly used as needles in Tiwanaku sites
in the Moquegua Valley, a small corn cob, a
wooden box with four compartments, an incomplete
wooden spoon, two incomplete gourd vessels, and
the worked shaft of the long bone from a large
mammal. This bone had been worked in such a
way that it resembled a weaving tool. However, it
was too small to have been functional as such, and
perhaps is instead a model or toy weaving tool.
Although it is not known from where this non-local
sherd was imported, it was chemically distinct from
the non-local sherds identified at Chen Chen.
DiscussionAs in many other ancient polities, ceramic vessels ful-
filled significant economic, political, and ritual func-
tions in the Tiwanaku state. They were utilized
extensively in the feasts and conspicuous drinking
that were vital tools in Tiwanaku statecraft. They
were also important portable media for the spread
of heartland iconographic repertoires that materia-
lized elite ideology. Visual analyses of ceramic
material have identified the presence of non-local
variants of Tiwanaku pottery in communities across
the south central Andes. Yet, to date, differences in
style have been the principal means of inferring the
movement of vessels around the state’s territory
(Goldstein 2005; Janusek 2004a; Kolata 1993a;
Korpisaari 2006).
The existing compositional data on clays found in
the Moquegua Valley provide the necessary basis for
determining the presence of ceramic imports in
assemblages from this major Tiwanaku province
and for examining patterns in provincial resource
procurement. The results presented in this study con-
firm that potters in the Moquegua Valley colony lar-
gely used locally available clays, those found in the
vicinity of the state’s administrative centers at Omo
and Chen Chen, to craft replicas of heartland
pottery. This process of replication was a means
through which provincial craft producers and consu-
mers asserted their ongoing cultural affiliation with
the state center (Goldstein 2005).
However, these data also provide the first empiri-
cal evidence for the movement of ceramic vessels
into the Moquegua Valley during the height of
Tiwanaku authority there. The importation of pot-
tery (or less likely, the importation of raw materials
used to produce this pottery) was part of the larger
exchange networks that existed during the height of
the Tiwanaku state, and that furnished the state
center with agricultural produce and valuable
resources. The ratio of local to non-local ceramics
(4 out of 45 or 8.9%) identified in this LA-ICP-MS
study of Moquegua pottery is comparable to that
of between 5 and 10% reported for other Tiwanaku
communities based on stylistic analyses (Janusek
2004a). However, we note that none of the four
sherds identified in this analysis would have necess-
arily been recognized as non-local on the basis of
style alone. Two were slipped but undecorated
sherds, a third was identified as a red-slipped tazon
with geometric decoration in black and white, and
a fourth was likely from a modeled incensario.
Both sherds whose form could be determined were
similar to locally produced vessels.
Where they were imported from is unclear. Until
compositional data from other regions in the Tiwanaku
sphere are available for comparative analysis, stylistic
assessmentwill continue toguide interpretation. Several
of these regions are currently the subject of clay surveys
and compositional study (including theArica and Lluta
valleys of northern Chile, by the authors). It is notable,
however, that the four non-local sherds do not consti-
tute a chemically homogeneous group. Instead, two dis-
tinct chemical signatures are represented by the sherds.
Chemicallydistinct sherdswere recoverednotonly from
the same cemetery, but also, in one case, from the same
grave.As already commented, determining the proveni-
ence of non-local ceramic material in the Moquegua
Valley awaits the results of ongoing surveys and ana-
lyses of clays in otherTiwanaku regions. The two differ-
ent chemical signatures may reflect different paste
recipes within a single area. However, it is also plausible
that these chemical signatures reveal the importation of
ceramic vessels from several other Tiwanaku zones.
Given that goods and products, as well as people,
were moving in multiple directions around the south
central Andes during the Middle Horizon, it is possible
that the Moquegua Valley colony’s participation in
varied networks of exchange is revealed in the distinct
chemical signatures of these non-local sherds.
When the archaeological context of these sherds is
considered, considerable equality of access is
suggested. The identification of non-local sherds in
Sharratt et al. Pottery production, regional exchange, and state collapse during the Middle Horizon (A.D. 500– 1000)
Journal of Field Archaeology 2015 VOL. 00 NO. 00 11
two different cemeteries at Chen Chen indicates that
access to imported ceramics was not restricted to
only one kin group in the community. Their inclusion
in both male and female graves demonstrates that
association with imported material, at least in
death, was not sex restricted either. Their inclusion
in tombs of various degrees of investment (stone-
lined cists and simple pit graves) suggests that
access to these vessels, or at least to the sherds, was
not restricted to those individuals whose families
could afford more elaborate tombs. The mechanism
of exchange needs to be clarified. The fact that
non-local sherds are found in a range of contexts
and are not limited to particular kin groups or
sexes, might be evidence that it was through personal
ties with other Tiwanaku regions that these vessels
were brought into the valley. Whether the individuals
buried in these particular graves were also non-local
to the Moquegua Valley is unknown, and would
merit further study, particularly given existing evi-
dence for individuals who had been raised in the
state heartland among the mortuary population at
Chen Chen (Blom 1999; Knudson et al. 2004).
It is noteworthy that in several instances, non-local
sherds accompanied ceramic vessels determined to be
locally produced, either by compositional analysis or
by stylistic analysis. Furthermore, in three cases, the
non-local sherds were not part of the intact vessels
found in graves. The practice of including ceramic
sherds as well as intact vessels has been noted in
Tiwanaku graves in the Moquegua Valley. Although
the meaning of this practice is elusive, it is striking
that in three instances, the non-local material is rep-
resented by only a single sherd, raising the possibility
that the original imported vessel had broken prior to
burial, possibly intentionally when the interred indi-
vidual died. However, sherds were also curated, per-
haps as tokens of ancestry or physical manifestations
of long-distance social ties.
Although questions about the meaning of some of
the non-local grave inclusions remain, the identifi-
cation of non-local ceramics in the Chen Chen
assemblage is significant. Scholars have previously
suggested that there were ceramic imports in the
Moquegua Valley during the Tiwanaku occupation
there (Goldstein 1993a, 2005), but these chemical
data, despite representing a tiny fraction of the pot-
tery from Chen Chen and from other state period
Tiwanaku sites in Moquegua, are the first physical
evidence for imported ceramics in the Tiwanaku
colony and the first empirical data demonstrating
that actual pots, rather than just Tiwanaku ceramic
styles, were brought into the province.
The compositional data presented above also
address pottery production and circulation in a
period relatively understudied in comparison with
the height of Tiwanaku political authority, namely
the phase immediately following the violent break-
down of the state. More variability is present in the
Tumilaca la Chimba analyzed sample, which included
Group 2 as well as Group 1 ceramics. Interestingly,
Group 2 ceramics were only identified in the samples
from two of the four cemeteries at the site.
As previously reported (Sharratt 2011a, 2011b), the
four cemeteries at Tumilaca la Chimba are distin-
guished from one another by particularities of ritual
practice and differences in grave inclusions that we
have argued are indicative of the increasing assertion
of social identities rooted inmembership of the various
intra-community groups who used each cemetery. It is
notable, then, that distinctions in crafting activity may
be another line of difference between the cemeteries.
Of the 49 sherds analyzed from Tumilaca la
Chimba, only one was determined to be non-local.
Due to the relatively small size of the samples ana-
lyzed with LA-ICP-MS, while the difference between
4/45 sherds at Chen Chen and 1/49 sherds at
Tumilaca la Chimba suggests a reduction in the over-
all presence of non-local ceramics in the wake of state
collapse, an inference supported by the p-XRF ana-
lyses conducted on larger samples (Sharratt 2011a),
the difference is not statistically significant. Interest-
ingly, this sherd is chemically distinct from the four
non-local sherds in the state period (Chen Chen)
sample, raising the possibility that not only had
long-distance exchange been reduced in degree, the
networks in which the community at Tumilaca la
Chimba participated had changed. It is also plaus-
ible, given the visual similarity of pottery at Chen
Chen and Tumilaca la Chimba (which can make it
difficult to attribute individual sherds to the state
period or collapse phase) that this sherd was curated
from the earlier state period colony.
Overall, the minimal presence of non-local material
at Tumilaca la Chimba suggests that although potters
at sites such as Tumilaca la Chimba continued to craft
vessels largely similar, if more internally variable and
of lesser quality, to those at Chen Chen, at other
state period Moquegua Valley sites and in the wider
environs of the Tiwanaku state, consumers at Tumi-
laca la Chimba did not have the same degree of or
as varied access to vessels brought from outside the
valley. In this way, as well as in others, they felt the
long term ramifications of regional political turmoil.
ConclusionsIn sum, the LA-ICP-MS study reported here demon-
strates the relevance of compositional analyses to
understandings of social and economic networks in
the Prehispanic Andes. The data derived from cer-
amic sherds excavated at Chen Chen confirm that
during the height of Tiwanaku state authority in
Sharratt et al. Pottery production, regional exchange, and state collapse during the Middle Horizon (A.D. 500– 1000)
12 Journal of Field Archaeology 2015 VOL. 00 NO. 00
the Moquegua Valley (A.D. 725–1000), local pottery
production utilized clays available in the immediate
vicinity of the Tiwanaku colonial sites. All locally pro-
duced pottery clusters in one chemical group,
suggesting considerable consistency in paste recipes.
However, in addition to local production, a small
number of ceramics were being imported into the
Moquegua Valley. Although the mechanisms that
facilitated the circulation of goods remain unclear,
and although determining where these vessels or frag-
ments came from awaits the results of ongoing field
research, it is striking that the non-local ceramics are
not homogeneous, but likely came from more than
one place.
The breakdown of long-distance exchange networks
is commonly cited as a consequence of political frag-
mentation (Renfrew 1979; Schwartz and Nichols
2006; Tainter 1988; Yoffee and Cowgill 1988). How-
ever, the collapse phase sample from Tumilaca la
Chimba did contain one non-local sherd, negating
the hypothesis that all ceramic material at the site
was locally produced. Notably this sherd has a chemi-
cal signature distinct from those of the non-local sherds
identified from Chen Chen, suggesting that even if
post-collapse communities maintained long distance
exchange networks, they were altered in the wake of
political turmoil. Interestingly, the results of the
LA-ICP-MS analysis demonstrate greater internal
variability in the chemical composition of the collapse
phase sample, with two distinct chemical groups pre-
sent, indicating greater variability in crafting activity
at Tumilaca la Chimba compared with the homogen-
eity evident in the local material from Chen Chen.
This greater variability adds weight to the argument
that the context of ceramic production was altered in
the wake of state collapse, with the demise of commu-
nity wide ceramic workshops in which production was
highly standardized.
The detailed visual and stylistic analyses of the
large ceramic assemblages recovered from Tiwanaku
sites across the south central Andes have been instru-
mental in furthering our understanding of how pot-
tery vessels were used by elites in the spread and
materialization of state ideology, and by provincial
and local communities to simultaneously demon-
strate allegiance to the state and assert local identi-
ties. Identification of stylistic differences between
provinces, communities and neighborhoods has
thus far provided an important measure for examin-
ing the spread and exchange of ceramic style, and the
ratios of locally produced to imported ceramics
identified in this compositional study are comparable
to those suggested by visual analyses. We note how-
ever, that this small LA-ICP-MS study identified
sherds as non-local that would likely have been
assumed to be local to the Moquegua Valley based
on style alone, and suggest that the incorporation
of compositional analyses, such as those discussed
above, has considerable potential to further under-
standings of the movement of physical materials as
well as styles, and the ways in which those processes
of circulation were embedded in wider political struc-
tures and social change in the south central Andes.
AcknowledgmentsLA-ICP-MS analyses were funded by the National
Science Foundation (DDIG 0937303). Fieldwork at
Tumilaca la Chimba was supported by Fulbright IIE,
Dumbarton Oaks, the Graduate College and Depart-
ment of Anthropology at the University of Illinois of
Chicago as well as by the Women’s Board and the
Department of Anthropology at the Field Museum.
Excavations at Tumilaca la Chimba were conducted
with permission from the Ministerio de Cultura del
Peru, Lima (RDN 1208/INC awarded to Patrick
Ryan Williams and Maria Elena Rojas Chavez in
2006/2007).RomuloPariFlores andBruceOwen facili-
tated studyof theChenChen ceramic samples.Ceramic
sherds were exported from Peru to the U.S.A. with the
permission of the Ministerio de Cultura del Peru,
Lima (# 1659/792). Particular thanks are due Laure
Dussubieux for providing invaluable assistance in the
Elemental Analysis Facility at the Field Museum.
Nicola Sharratt (Ph.D. 2011, University of Illinois at
Chicago) is Assistant Professor of Anthropology, at
Georgia State University. Her interests include
Andean South America, state collapse, craft pro-
duction, complex societies, and archaeometry.
Mark Golitko (Ph.D. 2010, University of Illinois at
Chicago) is Regenstein Research Scientist, at the
Field Museum of Natural History. His interests
include Prehistoric social networks, trade, archaeome-
try, the western Pacific, and Europe.
Patrick Ryan Williams (Ph.D. 1997, University of
Florida) is Associate Director of Research and Associ-
ate Curator of Archaeological Science, at the Field
Museum of Natural History. His interests include the
Anthropology of imperialism and colonialism, complex
societies, agricultural dynamics, landscape ecology,
geographic information systems and remote sensing
applications, geoarchaeology, archaeometry, and
Andean South America.
ReferencesAdan-Bayewitz, D., and I. Perlman. 1985. ‘‘Local Pottery Proveni-
ence Studies: A Role for Clay Analysis,’’ Archaeometry 27(2):203–217.
Albarracin-Jordan, J. 1996. ‘‘Tiwanaku Settlement System: TheIntegration of Nested Hierarchies in the Lower TiwanakuValley,’’ Latin American Antiquity 7(3): 183–210.
Sharratt et al. Pottery production, regional exchange, and state collapse during the Middle Horizon (A.D. 500– 1000)
Journal of Field Archaeology 2015 VOL. 00 NO. 00 13
Albarracin-Jordan, J., and J. E. Mathews. 1990. AsentamientosPrehispanicos del Valle de Tiwanaku 1. La Paz: ProduccionesCIMA.
Anderson, K. 2013. ‘‘Tiwanaku Influence of the Central Valley ofCochabamba,’’ in A. Vranich and C. Stanish, eds., Visions ofTiwanaku. Los Angeles: The Cotsen Institute of Archaeology,87–112.
Arnold, D. E. 2005. ‘‘Linking Society with the CompositionalAnalyses of Pottery: A Model from Comparative Ethnogra-phy,’’ in A. Livingstone-Smith, D. Bosquet andR. Martineau, eds., Pottery Manufacturing Processes: Recon-stitution and Interpretation. B.A.R. International Series 1349.Oxford: Archaeopress, 15–21.
Arnold, D. E., H. Neff, and R. Bishop. 1991. ‘‘CompositionalAnalysis and ‘Sources’ of Pottery: An EthnoarchaeologicalApproach,’’ American Anthropologist 93(1): 70–90.
Bandy, M. 2001. Population and History in the Ancient TiticacaBasin. Ph.D. dissertation, University of California at Berkeley.Ann Arbor: University Microfilms.
Bartlett, M. L., H. Neff, and P. McAnany. 2000. ‘‘Differentiationof Clay Resources On a Limestone Plain: The Analysis of ClayUtilization During the Maya Formative at K’axob, Belize,’’Geoarchaeology 15(2): 95–133.
Bauer, B. S., and C. Stanish. 2001. Ritual and Pilgrimage in theAncient Andes. Austin: University of Texas Press.
Bawden, G. 1989. ‘‘The Tumilaca Site and Post-Tiahuanaco Occu-pational Stratigraphy in the Moquegua Drainage,’’ in D. S.Rice, C. Stanish and P. R. Scarr, eds., Ecology, Settlementand History in the Osmore Drainage, Peru. B.A.R. InternationalSeries 545. Oxford: Archaeopress, 287–302.
Bawden, G. 1993. ‘‘An Archaeological Study of Social Structureand Ethnic Replacement in Residential Architecture of theTumilaca Valley,’’ in M. S. Aldenderfer, ed., Domestic Archi-tecture, Ethnicity, and Complementarity in the South-CentralAndes. Iowa City: University of Iowa Press, 42–54.
Baxter, M. J. 1992. ‘‘Archaeological Uses of the Biplot—aNeglected Technique?’’ in G. Lock and J. Moffett, eds.,Computer Applications and Quantitative Methods in Archaeology,1991. B.A.R. International Series S577. Oxford: TempusReparatum, 141–148.
Baxter, M. J. 2001. ‘‘Statistical Modelling of Artefact Compo-sitional Data,’’ Archaeometry 43(1): 131–147.
Bellido, E., and C. Landa. 1998. Mapa geologico cuadrangulo deMoquegua. Ministerio de Fomento y Obras Publicas, Direc-cion de Mineria, Comision Carta Geologica Nacional.
Bermann, M. 1994. Lukurmata: Household Archaeology in Prehis-panic Bolivia. Princeton: Princeton University Press.
Bermann, M. 1997. ‘‘Domestic Life and Vertical Integration in theTiwanaku Heartland,’’ Latin American Antiquity 8(2): 93–112.
Bermann, M., P. S. Goldstein, C. Stanish, and L. Watanabe. 1989.‘‘The Collapse of the Tiwanaku State: A View from theOsmore Drainage,’’ in D. S. Rice, C. Stanish and P. R.Scarr, eds., Ecology, Settlement and History in the OsmoreDrainage, Peru. B.A.R. International Series 545. Oxford:Archaeopress, 269–285.
Bishop, R., R. L. Rands, and G. R. Holley. 1982. ‘‘Ceramic Com-positional Analysis in Comparative Perspective,’’ Advances inArchaeological Method and Theory 5: 275–330.
Blom, D. E. 1999. Tiwanaku Regional Interaction and Social Iden-tity: A Bioarchaeological Approach. Ph.D. disseration, Univer-sity of Chicago. Ann Arbor: University Microfilms.
Blom, D. E., B. Hallgrimsson, L. Keng, M. C. Lozada Cerna, andJ. E. Buikstra. 1998. ‘‘Tiwanaku ‘Colonization’: BiologicalImplications for Migration in the Moquegua Valley, Peru,’’World Archaeology 30(2): 238–261.
Brumfiel, E.M., andT.K. Earle. 1987.Specialization, Exchange, andComplex Societies. Cambridge: Cambridge University Press.
Burger, R. L. 2000. ‘‘Through the Glass Darkly: Prehispanic Obsi-dian Procurement and Exchange in Southern Peru and North-ern Bolivia,’’ Journal of World Prehistory 14(3): 267–362.
Burger, R. L., F. Asaro, H. V. Michel, F. Stross, and E. Salazar.1994. ‘‘An Initial Consideration of Obsidian Procurementand Exchange in Prehispanic Ecuador,’’ Latin American Anti-quity 5: 228–255.
Burger, R. L., F. Asaro, G. Salas, and F. Stross. 1998a. ‘‘TheChivay Obsidian Source and the Geological Origin of TiticacaBasin Type Obsidian Artifacts,’’ Andean Past 5: 203–224.
Burger, R. L., F. Asaro, P. Trawick, and F. Stross. 1998b. ‘‘TheAlca Obsidian Source: The Origin of Raw Material forCuzco Type Obsidian Artifacts,’’ Andean Past 5: 185–202.
Burger, R. L., and M. D. Glascock. 2000. ‘‘Locating the QuispisisaObsidian Source in the Department of Ayacucho, Peru,’’ LatinAmerican Antiquity 11(3): 258–268.
Burger, R. L., K. J. Schreiber, M. D. Glascock, and J. Ccencho.1998c. ‘‘The Jampatilla Obsidian Source: Identifying the Geo-logical Source of Pampas Type Obsidian Artifacts fromSouthern Peru,’’ Andean Past 5: 225–239.
Burkholder, J. E. 1997. Tiwanaku and the Anatomy of Time:A New Ceramic Chronology from the Iwawi site, Departmentof La Paz, Bolivia. Ph.D. dissertation, State University ofNew York, Binghamton. Ann Arbor: University Microfilms.
Cecil, L. G. 2004. ‘‘Inductively Coupled Plasma Emission Spec-troscopy and Post Classic Peten Slipped Pottery: An Examin-ation of Pottery Wares, Social Identity and Trade,’’Archaeometry 46(3): 385–404.
Couture, N. C. 2003. ‘‘Ritual, Monumentalism, and Residence atMollo Kontu, Tiwanaku,’’ in A. L. Kolata, ed., Tiwanakuand Its Hinterland: Archaeology and Paleocology of AnAndean Civilization Vol 2. Washington, D.C.: SmithsonianInstitution Press, 202–225.
Couture, N. C., and K. Sampeck. 2003. ‘‘Putuni: a History ofPalace Architecture in Tiwanaku,’’ in A. L. Kolata, ed.,Tiwanaku and its Hinterland: Archaeology and Paleoecologyof An Andean Civilization. Washington, D.C.: SmithsonianInstitution Press, 226–263.
D’Altroy, T. N., and T. K. Earle. 1985. ‘‘Staple Finance, WealthFinance, and Storage in the Inka Political Economy,’’ CurrentAnthropology 26(2): 187–206.
D’Altroy, T. N., A. M. Lorandi, and V. I. Williams. 1994.‘‘Produccion y Uso de Ceramica en la Economia PoliticaInka,’’ in I. Shimada, ed., Tecnologia y Organizacion de la Pro-duccion Ceramica Prehispanica en los Andes. Lima: PontificaUniversidad Catolica del Peru, 395–441.
DeMarrais, E., L. J. Castillo, and T. K. Earle. 1996. ‘‘Ideology,Materialization and Power Strategies,’’ Current Anthropology37(1): 15–31.
Dorais, M. J., M. Lindblom, and C. M. Shriner. 2004. ‘‘Evidencefor a Single Clay/Temper Source for the Manufacture ofMiddle and Late Helladic Aeginetan Pottery from Asine,Greece,’’ Geoarchaeology 19(7): 657–684.
Dussubieux, L., M. Golitko, P. R. Williams, and R. J. Speakman.2007. ‘‘Laser Ablation-Inductively Coupled Plasma-Mass Spectrometry Analysis Applied to the Characterizationof Peruvian Wari Ceramics,’’ in M. D. Glascock, R. J. Speak-man and R. S. Popelka-Filcoff, eds., Archaeological Chemistry:Analytical Techniques and Archaeological Interpretation. Amer-ican Chemical Society symposium series. Washington, D.C.:Oxford University Press, 349–363.
Earle, T. K., and J. Ericson, E. 1977. Exchange Systems in Prehis-tory. New York: Academic Press.
Elliot, S., M. Knowles, and I. Kalinitchenko. 2004. ‘‘A New Direc-tion in ICP-MS,’’ Spectroscopy 19(1): 30–38.
Garcia Marquez, M. E. 1990. ‘‘Excavacion Arqueologia en elCementerio de Chen Chen, Moquegua; una Interaccion deContextos Funerarios: Tiwanaku/Wari.’’ Lic. Unpublishedthesis, Universidad Catolica Santa Maria, Arequipa.
Glascock, M. D., R. J. Speakman, and R. L. Burger. 2007.‘‘Sources of Archaeological Obsidian in Peru: Descriptionsand Geochemistry,’’ in M. D. Glascock, R. J. Speakman andR. S. Popelka-Filcoff, eds., Archaeological Chemistry: Analyti-cal Techniques and Archaeological Interpretation. AmericanChemical Society symposium series. Washington, D.C.:Oxford University Press, 524–552.
Gliozzo, E., and I. Memmi Turbanti. 2004. ‘‘Black Gloss Pottery:Production Sites and Technology in Northern Etruria, Part I:Provenance Studies,’’ Archaeometry 46(2): 201–225.
Goldstein, P. S. 1985. ‘‘TiwanakuCeramics of theMoqueguaValley,Peru.’’ Unpublished M.A. thesis, University of Chicago.
Goldstein, P. S. 1989a. Omo: A Tiwanaku Provincial Center inMoquegua, Peru. Ph.D. disseration, University of Chicago.Ann Arbor: University Microfilms.
Goldstein, P. S. 1989b. ‘‘The Tiwanaku Occupation of Moque-gua,’’ in D. S. Rice, C. Stanish and P. R. Scarr, eds., Ecology,Settlement and History in the Osmore Drainage, Peru. Oxford:B.A.R. International Series 545, 219–255.
Goldstein, P. S. 1993a. ‘‘House,Community, andState in theEarliestTiwanaku Colony: Domestic Patterns and State Integration atOmo M12, Moquegua,’’ in M. S. Aldenderfer, ed., DomesticArchitecture, Ethnicity, and Complementarity in the South-Cen-tral Andes. Iowa City: University of Iowa Press, 25–41.
Sharratt et al. Pottery production, regional exchange, and state collapse during the Middle Horizon (A.D. 500– 1000)
14 Journal of Field Archaeology 2015 VOL. 00 NO. 00
Goldstein, P. S. 1993b. ‘‘Tiwanaku Temples and State Expansion:A Tiwanaku Sunken-Court Temple in Moquegua, Peru,’’Latin American Antiquity 4(1): 22–47.
Goldstein, P. S. 2005. Andean Diaspora: The Tiwanaku Coloniesand The Origins of South American Empire. Gainesville: Uni-versity Press of Florida.
Goldstein, P. S. 2013. ‘‘Tiwanaku and Wari State Expansion:Demographic and Outpost Colonization Compared,’’ inA. Vranich and C. Stanish eds., Visions of Tiwanaku. LosAngeles: Cotsen Institute of Archaeology, 41–63.
Golitko, M., and J. E. Terrell. 2012. ‘‘Mapping Prehistoric SocialFields on the Sepik Coast of Papua New Guinea: CeramicCompositional Analysis Using Laser Ablation-InductivelyCoupled Plasma-Mass Spectrometry,’’ Journal of Archaeologi-cal Science 39: 3568–3580.
Graffam, G. 1992. ‘‘Beyond State Collapse: Rural History, RaisedFields, and Pastoralism in the South Andes,’’ AmericanAnthropologist 94(4): 882–904.
Gratuze, B. 1999. ‘‘Obsidian Characterization by Laser AblationICP-MS and its Application to Prehistoric Trade in the Med-iterranean and the Near East: Sources and Distribution ofObsidian within the Aegean and Aeolian,’’ Journal of Archae-ological Science 26: 869–881.
Gratuze, B., M. Blet-Lemarquand, and J. N. Barrandon. 2001.‘‘Mass Spectrometry with Laser Sampling: A New Tool toCharacterize Archaeological Materials,’’ Journal of Radioana-lytical and Nuclear Chemistry 247(3): 645–656.
Harbottle, G. 1976. ‘‘Activation Analysis in Archaeology,’’ inG. W. A. Newton, ed., Radiochemistry. London: The ChemicalSociety, 33–72.
Hein, A., H. Mommsen, and G. Zender. 2004. ‘‘Pliocene Claysfrom Aegina (Greece): Reference Material for Chemical Prove-nance Studies on Bronze Age Pottery from the Island,’’Geoarchaeology 19(6): 553–564.
Isbell, W. H. 2013. ‘‘Nature of An Andean City: Tiwanaku and theProduction of Spectacle,’’ in A. Vranich and C. Stanish, eds.,Visions of Tiwanaku. Los Angeles: Cotsen Institute of Archae-ology, 167–196.
Janusek, J. W. 1999. ‘‘Craft and Local Power: Embedded Special-ization in Tiwanaku Cities,’’ Latin American Antiquity 10(2):107–131.
Janusek, J. W. 2002. ‘‘Out of Many, One: Style and Social Bound-aries in Tiwanaku,’’ Latin American Antiquity 13(1): 35–61.
Janusek, J. W. 2003a. ‘‘The Changing Face of Tiwanaku Residen-tial Life: State and Local Identity in An Andean City,’’ inA. L. Kolata, ed., Tiwanaku and its Hinterland: Archaeologyand Paleoecology of an Andean Civilization. Washington,D.C.: Smithsonian Institution Press, 264–295.
Janusek, J. W. 2003b. ‘‘Vessels, Time, and Society: Toward a Cer-amic Chronology in the Tiwanaku Heartland,’’ in A. L.Kolata, ed., Tiwanaku and its Hinterland: Archaeology andPaleoecology of An Andean Civilization. Washington, D.C.:Smithsonian Institution Press, 30–89.
Janusek, J. W. 2004a. Identity and Power in the Ancient Andes.London: Routledge.
Janusek, J. W. 2004b. ‘‘Tiwanaku and Its Precursors; RecentResearch and Emerging Perspectives,’’ Journal of Archaeologi-cal Research 12(2): 121–183.
Janusek, J. W. 2005. ‘‘Collapse As Cultural Revolution: Power andIdentity in the Tiwanaku to Pacajes Transition,’’ in K. J.Vaughn, D. Ogburn and C. A. Conlee, eds., Foundations ofPower in thePrehispanicAndes.ArchaeologicalPapers of theAmer-icanAnthropological Association, vol. 14, Arlington, VA, 175–210.
Janusek, J. W. 2008. Ancient Tiwanaku. Cambridge: CambridgeUniversity Press.
Janusek, J. W. 2013. ‘‘Social Diversity, Ritual Encounter, and theContingent Production of Tiwanaku,’’ in A. Vranich andC. Stanish, eds., Visions of Tiwanaku. Los Angeles: TheCotsen Institute of Archaeology, 197–209.
Janusek, J. W., and A. L. Kolata. 2003. ‘‘Prehispanic Rural Historyin the Katari Valley,’’ in A. L. Kolata, ed., Tiwanaku and itsHinterland: Archaeology and Paleoecology of an Andean Civiliza-tion. Washington D.C.: Smithsonian Institution Press, 129–172.
Kennett, D. J., A. J. Anderson, M. J. Cruz, G. R. Clark, andG. R. Summerhayes. 2004. ‘‘Geochemical Characterizationof Lapita Pottery via Inductively Coupled Plasma-MassSpectrometry (ICP-MS),’’ Archaeometry 46(1): 35–46.
Kennett, D. J., H. Neff, M. D. Glascock, and A. Z. Mason. 2001.‘‘A Geochemical Revolution: Inductively Coupled Plasma MassSpectrometry,’’ The SAA Archaeological Record 1(1): 22–26.
Knudson, K. J., T. D. Price, J. E. Buikstra, and D. E. Blom. 2004.‘‘The Use of Strontium Isotope Analysis to InvestigateTiwanaku Migration and Mortuary Ritual in Bolivia andPeru,’’ Archaeometry 46(1): 5–18.
Kolata, A. L. 1993a. The Tiwanaku: Portrait of an Andean Civiliza-tion. Oxford: Blackwell.
Kolata, A. L. 1993b. ‘‘Understanding Tiwananku: Conquest,Colonization and Clientage in the South Central Andes,’’ inD. S. Rice, ed., Latin American Horizons. Washington, D.C.:Dumbarton Oaks, 193–224.
Kolata, A. L. 2003. Tiwanaku and its Hinterland: Archaeology andPaleoecology of An Andean Civilization Vol. 2. Washington,D.C.: Smithsonian Institution Press.
Kolata, A. L., and C. R. Ortloff. 2003. ‘‘Agroecological Perspec-tives on the Decline of the Tiwanaku State,’’ in A. L.Kolata, ed., Tiwanaku and its Hinterland: Archaeology andPaleoecology of An Andean Civilization. Washington, D.C.:Smithsonian Institution Press, 181–202.
Korpisaari, A. 2006. Death in the Bolivian High Plateau: Burialsand Tiwanaku Society. B.A.R. International Series 1536.Oxford: Archaeopress.
Korpisaari, A. and M. Parssinen. 2011. Pariti: The CeremonialTiwanaku Pottery of An Island in Lake Titicaca. Helsinki: Aca-demia Scientiarum Fennica.
Mallory-Greenough, L. M., J. D. Greenough, and V. J. Owen.1998. ‘‘New Data for Old Pots: Trace-Element Characteriz-ation of Ancient Egyptian Pottery Using ICP-MS,’’ Journalof Archaeological Science 25: 85–97.
McAndrews, T. L., J. Albarracin-Jordan, and M. Bermann. 1997.‘‘Regional Settlement Patterns in the Tiwanaku Valley of Boli-via,’’ Journal of Field Archaeology 24(1): 67–83.
Mirti, P., M. Gulmini, M. Pace, and D. Elia. 2004. ‘‘The Prove-nance of Red Figure Vases from Locri Epizephiri (SouthernItaly): New Evidence by Chemical Analysis,’’ Archaeometry46(2): 183–200.
Moseley,M.E., P.R.Williams,D. J.Nash, S.DeFrance,A.Miranda,and M. Ruales. 2005. ‘‘Burning Down the Brewery: Establishingand Evacuating An Ancient Imperial Colony at Cerro Baul,Peru,’’ Proceedings of the National Academy of Sciences of theUnited States of America 102(48): 17264–17271.
Nakassis, D., W. A. Parkinson, and M. L. Galaty. 2011.‘‘Redistributive Economies from a Theoretical and Cross-Cultural Perspective,’’ American Journal of Archaeology 115:177–184.
Nash, D. J. 2002. The Archaeology of Space: Places of Power in theWari Empire. Ph.D. dissertation, University of Florida,Gainesville. Ann Arbor: University Microfilms.
Nash, D. J, and P. R. Williams. 2009. ‘‘Wari Political Organization;The Southern Periphery,’’ in J. Marcus and P. R. Williams, eds.,Andean Civilization: A Tribute to Michael Moseley. Los Angeles:Cotsen Institute of Archaeology Press, 157–175.
Neff, H. 1994. ‘‘R-Q Mode Principal Components Analysisof Ceramic Compositional Data,’’ Archaeometry 36(1): 115–129.
Neff, H. 2002. ‘‘Quantitative Techniques for AnalyzingCeramic Compositional Data,’’ in D. M. Glowacki andH. Neff, eds., Ceramic Production and Circulation in theGreater Southwest: Source Determination by INAA and Comp-lementary Mineralogical Investigations Monograph 44. LosAngeles: Cotsen Institute of Archaeology. University ofCalifornia, Los Angeles, 15–36.
Neff, H., F. J. Bove, B. Lou, and M. F. Piechowski. 1992. ‘‘CeramicRaw Materials Survey in Pacific Coastal Guatemala,’’ inH. Neff, ed., Chemical Characterization of Ceramic Pastes inArchaeology. Madison, WI: Prehistory Press, 59–84.
Niedershlag, E., E. Pernicka, Th. Seifert, and M. Barthelheim.2003. ‘‘The Determination of Lead Isotope Ratios by MultipleCollector ICP-MS: A Case Study of Early Bronze AgeArtefacts and Their Possible Relation with Ore Deposits ofthe Erzgebirge,’’ Archaeometry 45(1): 61–100.
Niziolek, L. 2013. ‘‘Earthenware Production and Distribution inthe Prehispanic Philippine Polity of Tanjay: Results fromLaser Ablation-Inductively Coupled Plasma-Mass Spec-trometry (LA-ICP-MS),’’ Journal of Archaeological Science40: 2824–2839.
Oka, R., and C. M. Kusimba. 2008. ‘‘The Archaeology of TradingSystems, Part 1: Towards a New Trade Synthesis,’’ Journal ofArchaeological Research 16: 339–395.
Ortloff, C. R., and A. L. Kolata. 1993. ‘‘Climate and Collapse:Agro-ecological Perspectives on the Decline of the TiwanakuState,’’ Journal of Archaeological Science 20: 195–221.
Sharratt et al. Pottery production, regional exchange, and state collapse during the Middle Horizon (A.D. 500– 1000)
Journal of Field Archaeology 2015 VOL. 00 NO. 00 15
Owen, B. 1997. ‘‘Informe de Excavaciones en los SectoresMortuorios de Chen Chen,’’ in Proyecto Rescate de ChenChen de 1995. Unpublished report submitted to the InsitutoNacional de Cultura, Lima, Peru.
Owen, B. 2005. ‘‘Distant Colonies and Explosive Collapse: TheTwo Stages of the Tiwanaku Diaspora in the OsmoreDrainage,’’ Latin American Antiquity 16(1): 45–80.
Pari Flores, R., E. Elias, R. Ochoa P. and N. Ramos Rosales.2002. Proyecto de Rescate Arqueologico Chen Chen 2002;Informe Labores del Campo. Report submitted to the InsitutoNacional de Cultura, Lima, Peru.
Parkinson, W. A. 2010. ‘‘Beyond the Peer: Social Interaction andPolitical Evolution in the Bronze Age Aegean,’’ in D. J.Pullen, ed., Political Economies of the Aegean Bronze Age.Oxford: Oxbow Books, 11–34.
Parkinson, W. A., and M. L. Galaty. 2009. Archaic StateInteraction: The Eastern Mediterranean in the Bronze Age.Santa Fe: School for Advanced Research.
Perez-Arantegui, J., M. I. Urunuela, and J. R. Castillo. 1996.‘‘Roman Glazed Ceramics in the Western Mediterranean:Chemical Characterization by Inductively Coupled PlasmaAtomic Emission Spectrometry of Ceramic Bodies,’’ Journalof Archaeological Science 23: 903–914.
Phillips, S. C., and M. Morgenstein. 2002. ‘‘A Plains Ceramic ClaySource Characterization by Comparative Geochemical andPetrographic Analyses: Results from the Calhan PaintMines, Colorado, U.S.A.,’’ Geoarchaeology 17(6): 579–599.
Pollard, M., C. Batt, B. Stern, and S. M. M. Young. 2007.Analytical Chemistry in Archaeology. Cambridge Manuals inArchaeology. Cambridge: Cambridge University Press.
Renfrew, C. 1979. ‘‘Systems Collapse as Social Transformation:Catastophe and Anastrophe in Early State Societies,’’ inC. Renfrew and K. L. Cooke, eds., Transformations: Math-ematical Approaches to Culture Change. New York: AcademicPress, 481–506.
Rivera Casanovas, C. 2003. ‘‘Ch’iji Jawira: A Case of CeramicSpecialization in the Tiwanaku Urban Periphery,’’ in A. L.Kolata, ed., Tiwanaku and Its Hinterland: Archaeology andPaleoecology of an Andean Civilization. Washington, D.C.:Smithsonian Institution Press, 296–315.
Schwartz, G.M., and J. J. Nichols. 2006.After Collapse: The Regener-ation of Complex Societies. Tucson: University of Arizona Press.
Schur, C. 2011. ‘‘Tiwanaku Ceramics: Production and Distributionin the Moquegua Valley.’’ Unpublished B.A. thesis, Northwes-tern University, Evanston.
Sharratt, N. 2011a. Social Identities and State Collapse:A Diachronic Study of Tiwanaku Burials in the MoqueguaValley, Peru. Ph.D. dissertation, University of Illinois atChicago. Ann Arbor: University Microfilms.
Sharratt, N. 2011b. ‘‘Identity Negotiation During Tiwanaku StateCollapse,’’ in L. Amundsen-Meyer, N. Engel and S. Pickering,eds., Identity Crisis: Archaeological Perspectives on SocialIdentity. Proceedings of the 42nd (2010) Annual ChacmoolConference. Calgary: University of Calgary, 167–177.
Sharratt, N., M. Golitko, P. R. Williams, and L. Dussubieux.2009. ‘‘Ceramic Production During the Middle Horizon;Wari and Tiwanaku Clay Procurement in the MoqueguaValley, Peru,’’ Geoarchaeology 24(6): 792–820.
Sharratt, N., P. R. Williams, M. C. Lozada Cerna, and J. Starbird.2012. ‘‘Late Tiwanaku Mortuary Patterns in the MoqueguaDrainage, Peru: Excavations at the Tumilaca la ChimbaCemetery,’’ in A. Vranich, E. Klarich and C. Stanish, eds.,Advances in Titicaca Basin Archaeology III. Ann Arbor:Museum of Anthropology Publications, 193–203.
Shennan, S. 1997. Quantifying Archaeology, 2nd edn. Iowa City:University of Iowa Press.
Sherriff, B. L., P. Court, S. Johnston, and L. Stirling. 2002. ‘‘TheSource of Raw Materials for Roman Pottery from Leptiminus,Tunisia,’’ Geoarchaeology 17(8): 835–861.
Shortland, A. J. 2002. ‘‘The Use and Origin of Antimonate Color-ants in Early Egyptian Glass,’’ Archaeometry 44(4): 517–530.
Sillar, B. 1996. ‘‘The Dead and The Drying: Techniques for Trans-forming People and Things in the Andes,’’ Journal of SocialArchaeology 1(3): 259–289.
Sims, K. 2006. ‘‘After State Collapse: How Tumilaca CommunitiesDeveloped in the Upper Moquegua Valley, Peru,’’ in G. M.Schwartz and J. J. Nichols, eds., After Collapse: The RegenerationofComplexSocieties.Tucson:UniversityofArizonaPress, 114–136.
Squier, E. G. 1877. Peru: Incidents of Travel and Exploration in theLand of the Incas. London: Macmillan and Co.
Stanish, C. 1989. ‘‘Household Archaeology: Testing Models ofZonal Complementarity in the South Central Andes,’’American Anthropologist 91(1): 7–24.
Stanish, C. 2002. ‘‘Tiwanaku Political Economy,’’ inW.H. Isbell andH. Silverman, eds., Andean Archaeology I: Variations in Sociopo-litical Organization. NewYork: Kluwer Academic, 169–198.
Stanish, C. 2003. Ancient Titicaca: The Evolution of ComplexSociety in Southern Peru and Northern Bolivia. Berkeley:University of California Press.
Stanish, C. 2013. ‘‘What Was Tiwanaku?’’ in A. Vranich andC. Stanish, eds., Visions of Tiwanaku. Los Angeles: CotsenInstitute of Archaeology, 151–166.
Stanish, C., K. L. Frye, E. de la Vega, and M. T. Seddon. 2005.‘‘Tiwanaku Expansion into the Western Titicaca Basin,Peru,’’ in C. Stanish, A. B. Cohen and M. S. Aldenderfer,eds., Advances in Titicaca Basin Archaeology 1. Los Angeles:Cotsen Institute of Archaeology, 103–114.
Stanish, C., E. de la Vega, M. Moseley, P. R. Williams, C. Chavez,B. Vining, and K. LaFavre. 2010. ‘‘Tiwanaku Trade Patternsin Southern Peru,’’ Journal of Anthropological Archaeology29(4): 524–532.
Stovel, Emily. 2001. ‘‘Patrones funerarios de San Pedro deAtacama y el problema de la presencia de los contextosTiwanaku,’’ Boletin de Arqueologia PUCP 5: 375–396.
Strazicich, N. M. 1998. ‘‘Clay Sources, Pottery Production, andRegional Economy in Chalchihuites, Mexico, A. D. 200–900,’’ Latin American Antiquity 9(3): 259–274.
Sutter, R. C. 2000. ‘‘Prehistoric Genetic and Cultural Change:A Bioarchaeological Search for Pre-Inka Altiplano Coloniesin the Coastal Valleys of Moquegua, Peru and Azapa,Chile,’’ Latin American Antiquity 11(1): 43–70.
Sutter, R. C., and N. Sharratt. 2010. ‘‘Continuity and Transform-ation during the Terminal Middle Horizon (A.D. 950–1150):A Bioarchaeological Assessment of Tumilaca Origins withinthe Middle Moquegua Valley, Peru,’’ Latin American Antiquity21(1): 67–86.
Tainter, J. 1988. The Collapse of Complex Societies. Cambridge:Cambridge University Press.
Torres-Rouff, C. 2008. ‘‘The Influence of Tiwanaku on Life inthe Chilean Atacama: Mortuary and Bodily Perspectives,’’American Anthropologist 110(3): 325–337.
Torres, C.M., andW. J. Conklin. 1995. ‘‘Exploring the San Pedro deAtacma/Tiwanaku Relationship,’’ in P. Dransart, ed., AndeanArt: Visual Expression and Its Relation to Andean Beliefs andValues. Hampshire, U.K.: Avebury, 78–108.
Tykot, R. H. 1997. ‘‘Characterization of the Monte Arci (Sardinia)ObsidianSources,’’Journal ofArchaeologicalScience24: 467–479.
Vaughn, K. J., L. Dussubieux, and P. R. Williams. 2011. ‘‘A PilotCompositional Analysis of Nasca Ceramics from the KroeberCollection,’’ Journal of Archaeological Science 38: 3560–3567.
Vaughn, K. J., and H. Neff. 2004. ‘‘Tracing the Clay Source of NascaPolychrome Pottery: Results from a Preliminary Raw MaterialSurvey,’’ Journal of Archaeological Science 31(11): 1577–1586.
Vranich,A.1999. Interpreting theMeaningofRitualSpaces:TheTempleComplex of Pumapunku, Tiwanaku, Bolivia. Ph.D. dissertation,University of Pennsylvania, Philadelphia. Ann Arbor: UniversityMicrofilms.
Vranich, A. 2001. ‘‘La Piramide de Akapana: Reconsiderano elCentro Monumental de Tiwanaku,’’ Boletin de ArqueologiaPUCP 5: 295–308.
Vranich, A. 2006. ‘‘The Construction and Reconstruction of RitualSpace at Tiwanaku, Bolivia (A.D. 500–1000),’’ Journal of FieldArchaeology 31(2): 121–136.
Williams, P. R. 1997. The Role of Disaster in the Development ofAgriculture and the Evolution of Social Complexity in theSouth-Central Andes. Ph.D. dissertation, University of Florida,Gainesville. Ann Arbor: University Microfilms.
Williams, P. R. 2001. ‘‘Cerro Baul: A Wari Center on theTiwanaku Frontier.’’ Latin American Antiquity 12(1): 67–83.
Williams, P. R. 2002. ‘‘Rethinking Disaster-Induced Collapse inthe Demise of the Andean Highland States: Wari andTiwanaku,’’ World Archaeology 33(3): 361–374.
Williams, P. R. 2008. Informe de Campo e Informe Final: ProyectoArqueolgico Cerro Baul 2006–7. Unpublished report submittedto the Insituto Nacional de Cultura, Lima, Peru.
Williams, P. R. 2013. ‘‘Tiwanaku: A Cult of the Masses,’’ inA. Vranich and C. Stanish, eds., Visions of Tiwanaku.Los Angeles: Cotsen Institute of Archaeology, 27–40.
Yoffee,N., andG.L.Cowgill. 1988.TheCollapseofAncient States andCivilizations. Tucson: University of Arizona Press.
Sharratt et al. Pottery production, regional exchange, and state collapse during the Middle Horizon (A.D. 500– 1000)
16 Journal of Field Archaeology 2015 VOL. 00 NO. 00