Society for American Archaeology

Specialized Production of San Martín Orange Ware at Teotihuacan, Mexico

Author(s): Kristin S. Sullivan
Source: Latin American Antiquity, Vol. 17, No. 1 (Mar., 2006), pp. 23-53
Published by: Society for American Archaeology
Stable URL: http://www.jstor.org/stable/25063035 .
Accessed: 23/04/2013 23:36

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].

Society for American Archaeology is collaborating with JSTOR to digitize, preserve and extend access to Latin
American Antiquity.

http://www.jstor.org

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
 SPECIALIZED PRODUCTION OF SANMART?N ORANGE WARE
AT TEOTIHUACAN,MEXICO

Kristin S. Sullivan

Specialized production forms an important component of the socioeconomic organization of ancient complex societies. Exca
vation provides critical information on the internal organization of production within a particular workshop and permits
the recovery of features not preserved on the surface. In large settlements like Teotihuacan where modern occupation cov
ers much of the ancient city, it is not feasible tofully excavate every suspected craft production locale. At Teotihuacan and
in other complex societies, the use of surface indications to generate thematic maps permits discussion of the spatial rela
tionships between economic and social units within the settlement. I use the surface collections made by the Teotihuacan
Mapping Project and the partial excavation of one apartment compound and ceramic workshop in the Tlajinga district to
consider the organization of specialized production in this neighborhood. I rely on the co-occurrence of ceramic produc
tion indicators to define
likely workshops, using the surface collections from the excavated workshop as a reasonable indi
cator of production. the evidence for vessel form specialization
Considering and dimensional standardization and previous
ethnoarchaeological analyses of the excavated materials, Tlajinga district pottery production appears to have been maxi

mally organized at the community level, with individual apartment compounds forming the basic production units.

La producci?n especializada es un componente importante de la organizaci?n de las sociedades complejas. La excavaci?n

proporciona importante informaci?n sobre la organizaci?n interna de la producci?n artesanal y permite recuperar elementos
culturales raramente preservados en superficie. En ciudades tan grandes y complejas como Teotihuacan, donde la construc
ci?n moderna cubre la ciudad antigua, no es posible excavar cada lugar que sugiera una producci?n cer?mica. En Teotihua
can como en otras sociedades complejas, el uso de colecciones de superficie para crear mapas tem?ticos permite la discusi?n
de relaciones entre unidades productivas. Utilizo las colecciones del Teotihuacan Mapping Project y la excavaci?n parcial de
un conjunto y taller alfarero del distrito de Tlajinga para considerar la organizaci?n de su producci?n especializada. Me baso
en que un probable taller incluye por lo menos dos indicadores razonables de producci?n. A partir de la evidencia que dis
cuto acerca de la especializaci?n de las formas y la estandarizaci?n de los tama?os de vasijas, as? como en la de los restos
excavados en investigaciones etnoarqueol?gicas anteriores, la alfarer?a de Tlajinga parece haber sido extremadamente orga
nizada al nivel comunitario, en la que los conjuntos residenciales individuales formaban las unidades b?sicas de producci?n.

Archaeologists have linked the emergence Earle 1987; Charlton 1994; Charlton et al. 2000;
of craft specialization to the development Graves 1991; Hodge and Mine 1990; Pool 1992),
of political complexity, and in state-level as well as that between production and the politi
societies many of the goods in circulation are pro cal economy (Arnold et al. 1993; Hirth 1996;
duced by specialists. Studies of craft specialization Nichols et al. 2002; Pool and Santley 1992; Stark
have focused on identifying production areas 1992). InMesoamerica, researchers have investi
(Arnold 1990; Deal 1988; Pool 1992; Stark 1984, gated broad temporal and spatial changes in the
1985) and assessing the organization of production organization of ceramic production and exchange
(Arnold 1991b; Arnold 1993; Costin 1991, 2001; using intensive surface collections (Feinman 1985;
Feinman 1985; Reina and Hill 1978; Rice 1981, Kowalewski et al. 1989; Santley and Arnold 1996;
1996; Sinopoli 1991). Other researchers have Santley et al. 1989; Stark 1992; Stark and Garraty
focused on evaluating the degree of specialization 2004). Excavation provides the opportunity to con
(Blackman et al. 1993; Charlton et al. 1991 ;Curet firm activities inferred from surface indications and
1993) and investigating the interaction among pro permits detailed discussion of specialized struc
duction, distribution, and exchange (Brumfiel and tures, features, or other indications of production

Kristin S. Sullivan Department of Anthropology, Arizona State University, Tempe, AZ 85287-2402 ([email protected])

Latin American Antiquity, 17(1), 2006, pp. 23-53

Copyright? 2006 by the Society for American Archaeology

23

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
24 LATINAMERICAN
ANTIQUITY [Vol. 17, No. 1,2006

Ifeotihuacw Chronology
(ca.150B.C-AJD.630)

Metepec A.D. 550-650

Xolalpan AD. 350-550

Tlamimilolpa AD. 200-350

Miccaotii AD. 150-200
Tracualh AD. 1-150
Pad achique 150-1 B.C.

Figure 1. Map of Teotihuacan showing the Tlajinga district and the excavated apartment compound and San Martin
Orange ware workshop at Tlajinga 33. Teotihuacan map, copyright 1972 by Ren? Mill?n, reprinted with permission.
Teotihuacan chronology after Rattray 2001:435.

not preserved on the surface. Surface collections, pie is of particular interest because Teotihuacan
however, offer the opportunity to consider spatial was a
major economic center producing
an array

patterns and the possible relationships between of goods that circulated throughout the city and
individual productive units. beyond during the Early Classic period (Cowgill
The production of San Martin Orange (SMO) 1992a, 1992b, 1997;Mill?n 1973,1981). Evidence
ware pottery in theTlajinga district of Teotihuacan for the production of different craft goods has been
provides a unique opportunity to consider the orga found in several of the city's more than 2,000 apart
nization of production within a neighborhood of the ment compounds, and previous archaeological
ancient city utilizing surface collections as well as studies have demonstrated the utility of thematic
information from the partial excavation of one maps based on surface collections in examining
apartment compound and SMO workshop (Fig social and spatial relations at Teotihuacan (Altschul
ures 1-2). Collections from the full-coverage sur 1981,1987; Robertson 1999, 2001).
vey of Teotihuacan conducted by the Teotihuacan In this article I present the results of a system
Mapping Project (TMP) permit consideration of the atic analysis of the surface materials for evidence
organization of some of the most specialized pot of specialized SMO production in theTlajinga dis
tery produced locally in one neighborhood of the trict. I utilize the ceramics collected by the TMP to
ancient city (Mill?n 1973;Mill?n et al. 1973). The detect likely areas of specialized ceramic produc
partial excavation of the apartment compound and tion (Mill?n 1973;Mill?n et al. 1973). Artifacts col
SMO workshop at Tlajinga 33, or 33:S3W1, also lected from the surface of the excavated SMO
contributes to this discussion of the organization of workshop at Tlajinga 33 serve as a baseline against
SMO production in the Tlajinga district (Hopkins which surface collections from other Tlajinga apart
1995; Krotser 1987; Krotser and Rattray 1980; ment compounds
are compared. I focus on the sur

Sanders et al. 1982; Sheehy 1992, 1998;Widmer face collections in order to consider the spatial
1987,1991;Widmer and Storey 1993). This exam relationship between likely SMO workshops in the

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
Sullivan] SPECIALIZED OF SAN MARTIN
PRODUCTION ORANGE
WARE

Figure 2. Late Xolalpan and Metepec San Martin Orange amphora, basin, and crater forms: (a) late Xolalpan crater, (b)
late Xolalpan amphora, (c) Metepec amphora, (d) Metepec crater, (e) Metepec basin. Scale in whole cm. After Rattray
2001:Figures 165 and 192.

Tlajinga district.1 This study ismeant to form the duction changed over time at different workshops
basis for future research in the area, including exca in the neighborhood.
vation of other apartment compounds with strong
surface indications for SMO pottery production.
Models of Specialized Production
These excavated materials could be compared with
the existing excavated collections from Tlajinga 33 Researchers have presented several models for the
for a more detailed understanding of the variabil organization of specialized production (Charlton et
ity inherent in the organization of Tlajinga SMO al. 2000; Earle 1981; Peacock 1982; Santley et al.
production and of how the organization of pro 1989; Sinopoli 1988). I rely on Costin's (1991:

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
 26 LATIN AMERICAN ANTIQUITY [Vol. 17, No. 1, 2006

Table 1.1,2001:275-278) typology for the organi In community specialization, where the maxi
zation of production because it takes into consid mal production unit would be the lineage or com
eration social, economic, and political factors pound, as is argued for Tlajinga 33 (Sheehy 1992;
influencing the organization of production. The Widmer and Storey 1993), the ceramic assemblage
results of my study do not fall neatly into a single should exhibit low investment in labor and moder
category; rather, they reflect variability in the orga ate skill (Costin and Hagstrum 1995:629). Iwould
nization of production. expect little or no evidence for vessel form spe
One aim of my study is to define areas where cialization or dimensional standardization by indi
SMO pottery was likely produced and to discuss the vidual production loci, with each potting household
organization of that production within the Tlajinga independent of others in theTlajinga district. Alter
district. Four of the productive configurations dis natively, in labor-based nucleated workshops?
cussed by Costin (1991)?dispersed corv?e, indi where skill and availability rather than kinship
vidual retainer, nucleated corv?e, and retainer determine recruitment by the production unit?
workshop?are not considered here because they potters may maximize production output with one
are associated with attached contexts of production, workshop,
or a group of clustered workshops, spe
whereas the Tlajinga district potters were located cializing in a particular vessel form, size, or both
far from the civic-ceremonial core of the city and and/or sharing productive facilities such as firing
were not attached to elite or governmental sponsors. areas (Santley et al. 1989:119). In this case, Iwould
Two other productive units addressed by Costin expect specialists in certain vessel forms or sizes
(1991)?individual and dispersed workshops? to be clustered together, reflecting organization at
involve production locales dispersed throughout a a level beyond that of the lineage or compound,
settlement and thus do not reflect the concentration including the sharing of tools and the scheduling
of potters in the Tlajinga neighborhood. Because of firings and other tasks. I also expect SMO assem
community specialization and nucleated workshops blages from nucleated workshops to be signifi
aremost pertinent to the present investigation, these cantly more standardized than assemblages from
are the two organizational configurations discussed nonworkshop contexts, following the standardiza
in detail here. Community specialization involves tion hypothesis.
"autonomous individual or household-based pro
duction units, within a single commu
aggregated The Standardization Hypothesis
nity, producing for unrestricted regional
consumption" (Costin and Hagstrum 1995:621). I The standardization hypothesis maintains that spe
refer to units practicing this form of specialized pro cialists' products will be more standardized?in
duction as
"aggregated small autonomous produc dimensional, compositional, stylistic, or formal
tion" units because they represent independent attributes?than those of relatively less specialized
household producers clustered in close proximity household potters. Many studies have relied, at
to one another. Nucleated workshops
are
"larger least in part, on the standardization hypothesis to
workshops aggregated within a single community, examine specialization in archaeological assem
producing for unrestricted regional consumption" blages where evidence such as raw materials,

and are expected to produce more specialized goods wasters, specialized facilities, tools, and reliably
(Costin and Hagstrum 1995:621). identified production locations is lacking (Barnes
The other goal of the present study is to assess 1987; Blackman et al. 1993; Feinman et al. 1984;
the level of specialization of theTlajinga SMO pot Hagstrum 1985, 1986, 1988; Rice 1981, 1989,
ters based on the standardization hypothesis 1991 ;Sinopoli 1988). Many researchers, however,
(Arnold and Santley 1993; Blackman et al. 1993; have noted deviations from the expectation that
Sinopoli 1988; Stark 1995). According to this specialists produce more standardized goods than
model, more specialized producers are expected to nonspecialists (Arnold 1990; Arnold and Santley
make more specialized and/or standardized prod 1993; Longacre et al. 1988; Stark 1995).
ucts as the result of repetitive behavior, increased Stark (1995:232) discusses additional factors
skill, and the need or desire to increase production affecting standardization, including the use of
efficiency tomeet consumer demand. stacking adjustments in firing and transporting ves

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
Sullivan] SPECIALIZED OF SAN MART?N
PRODUCTION ORANGE
WARE 27

sels and consumer acceptance of variability. In the (Cowgill 1987; Hopkins 1995;Krotser and Rattray
Tuxtla Mountains of Veracruz, ethnographic stud 1980; Robertson 1999). Previous studies of ceramic
ies indicate that repetitive actions and the use of production using the TMP survey collections have
molds and measuring techniques (e.g., pieces of identified probable ceramic workshops and discuss
wood, finger and hand spans) by relatively unspe product uniformity and specialization in a number
cialized household potters may also contribute to of ceramic wares and at apartment compounds
the relative standardization of their goods (Arnold located throughout the city (Hopkins 1995; Krotser
1991a; Arnold and Santley 1993). Increased motor 1987; Krotser and Rattray 1980; Rattray 1988).
skill and experience in pottery making, the ratio of The spatial patterning of artifactual and architec
the number of producers to consumers or total prod tural remains argues for at least some degree of spa
ucts, the effects of market controls, and the choice tial clustering of social groups and economic
of fabrication technique all influence the variabil activities atTeotihuacan (Altschul 1981,1987; Mil
ity of vessel dimensions (Arnold and Nieves Ion 1973; Robertson 2001; Sload 1982; Spence
1992:112; Stark 1995:232). 1981). In Tlajinga, the clustering of probable
Other factors influencing the apparent stan ceramic workshops within the district suggests
dardization of particular assemblages include "the some social relationship among occupants of the
choice of variables, the nature of assemblages to apartment compounds (Altschul 1981, 1987;
be compared, technology, and methods of analy Sheehy 1992:778).
sis" (Stark 1995:234). The time depth of prehistoric The Tlajinga district is of particular interest
remains also poses a problem for comparing because the area has evidence for some of themost
archaeological phases with ethnographic examples specialized ceramic production at Teotihuacan
where the number of producers and the years of (Figure 3). Large quantities of the popular Xolal
production are known variables (Arnold 1991a; pan (A.D. 350-550) andMetepec (A.D. 550-650)
Stark 1995; Sullivan 2002:Table 5). Costin and phase SMO ware were used for cooking (craters)
Hagstrum also note that the "vagaries of quantify and serving (basins) food and for liquid storage
ing skill lead us to suggest that ceramic techno {amphorae, Krotser and Rattray 1980; Rattray
logical attributes related to skill and control, to 1988; Robertson 2001:263; Sheehy 1998:299). The
standardization, and to labor investment be con ceramics collected by the TMP from the surface of
sidered together since they combine in distinctive the partially excavated apartment compound and
ways, discriminating different kinds of specializa SMO workshop at Tlajinga 33 provide a useful
tion" (1995:623). baseline against which other TMP surface collec
tions can be compared (Sanders et al. 1982; Sheehy
1992, 1998; Storey 1987, 1992; Widmer 1987,
Archaeological Background
1991;Widmer and Storey 1993). Although SMO
Beginning in the 1960s, the TMP created topo ware was produced from the same clay as other

graphic and archaeological maps andmade surface locally made Early Classic period ceramics, skilled
collections of nearly 5,000 survey units (Mill?n firing at relatively high temperatures in an oxidiz
1973,1981,1993; Mill?n et al. 1973:3-18). These ing environment made it extremely durable and
collection tracts include multifamily apartment produced its distinctive orange color (Hopkins
compounds (some of which were also craft work 1995; Krotser and Rattray 1980:96; Sheehy 1992,
shops), temples, and other structures, as well as 1998; Sotomayor Castenada 1966).
nonstructural areas such as plazas. Collection tracts SMO pottery production at Tlajinga 33 began
were assigned alphanumeric site numbers in the in the Early Xolalpan phase, as a complement to
order they were surveyed within one of the arbi the lapidary craft specialty that dates to the pre
trary 500 x 500-mmap squares (e.g., 33:S3W1 was ceding Tlamimilolpa phase (Rattray 2001:385;
the 33rd tract surveyed in square S3W1). All fea Widmer 1987). The Tlamimilolpa phase also wit
ture artifacts, including diagnostic rim sherds, were nessed the construction of nearly 2,000 planned
collected from the collection tracts (Mill?n 1973). apartment compounds to house the city's residents
The TMP collections have been used to exam and an increase in the importance of craft produc
ine the distribution of SMO throughout the city tion and external ties with other regions of

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
 28 LATIN AMERICAN ANTIQUITY [Vol. 17, No. 1, 2006

UL!J--"TA
28a
o .9;
'-'I_I
^
mn3<]D
' *? D Q
Didvq_g q
I
/j
- e
?.dD EH
a r s o
DD

.. D
ann
d q ?*aa a
LEGEND
apartment compound
f J possiblecompound
El temple
Q platform
^collection tract,
V
no structure
"j*river,deepbarranca
D
^"'probable oldwatercourse
jfr likelySMQworkshop_
|S4W2 Jsffl]_Q.

3. Map of the Tlajinga district. The apartment compounds and collection tracts mentioned in the text are labeled.
Figure

Mesoamerica (Mill?n 1981:209-210). The pro and S4W1-W3 (Mill?n 1973; Mill?n et al. 1973).
duction and distribution of SMO during theXolal These collections include sherds set aside by Paula
pan phase coincided with the city's largest Krotser (1987) as evidence of ceramic production.2
population and the height of its interaction abroad Although SMO is prevalent during two distinct
(Mill?n 1993:28-30; Rattray 2001:385-390). Rat ceramic phases (Xolalpan and Metepec), I con
tray (2001:389) characterizes the Late Xolalpan sider these to be a single combined phase because
phase as "a period of stabilization" in the pottery of disagreements in the diagnostic features of
industry, followed by a decline in quality during Metepec-phase SMO (e.g., Rattray 2001:297;
theMetepec phase. SMO continued to be an impor Sheehy 1998:311).3 I recorded 348 ceramic pro
tant ware during the Metepec phase, although duction indicators, including lunates (specialized
researchers disagree on whether SMO production crescent-shaped ceramic smoothers), sherd scrap
at the Tlajinga 33 apartment compound slowed as ers, a possible rotational support, a possible mold,
parts of the residence were abandoned (Rattray lumps of fired clay, and possible wasters, which are
2001:399) or ceased prior to the Metepec phase discussed inmore detail below.
(Sheehy 1992:300). By the end of the Metepec One concern that arose during this study
phase, the civic-ceremonial core of Teotihuacan involved differences in interpretations of the archi
had been looted and burned, with later occupation tectural remains at 20, 21, 24, and 25:S3W2.
occurring away from the Street of the Dead (Mil Krotser (1987:421) and Rattray (1988:263) argue
Ion 1981:238). that these structures may have been temple and
platform structures involved in the administration
of ceramic production in the district. The TMP
Tlajinga SMO Production
database indicates that 24:S3W2 is interpreted as
To assess specialized SMO production at Tlajinga, a residence of uncertain status, and Mill?n notes
I examined ail SMO sherds (nearly 7,000) from 181 that "the structures involved did not appear to be
TMP surface collections inmap squares S3W1-W3 built on large platforms, except for those fronting

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
Sullivan] SPECIALIZED OF SAN MARTIN
PRODUCTION WARE 29
ORANGE

the 'Street of the Dead.' But the grouping of these order to eliminate problems associated with small
structures around a large or open area was sizes. Unusual concentrations may reflect
plaza sample
confirmed" (1973:18). Survey notes indicate that sherds from vessels that broke during firing or some
this large plaza, 35:S3W1, may have been an open other part of the production process prior to being
firing area.Alternatively, the areamay have served used (Balkansky et al. 1997; Redmond 1979). Stark
as a local market for the goods produced by Tla notes that although unusual concentrations of pot
jinga potters (Mill?n 1981:225). Other important tery alone may not signify production, "arguments
structures in the area include the Eastern Platform, based on unusual concentrations of particular wares
which delimits the eastern extent of the Tlajinga or forms aremore promising. Absent inmost cases
district and is located along the Street of theDead. has been use of 'typical' domestic inventories as a
Temples and platforms at 7A-7G:S3W1 are located benchmark against which to compare such atypi
on top of the Eastern Platform and may have been cal concentrations" (1985:167).
involved in controlling access to the district or in To establish such a reference point in the pre
the distribution of SMO outside of the potting sent study, I consider both the overall ratio of SMO
neighborhood.4 to total Xolalpan andMetepec sherds from collec
tions throughout the city and the ratio from vari
ous contexts (i.e., production, domestic, ritual)
Probable SMO Production Workshops
within the Tlajinga district. The ratio of SMO to
The detection of ceramic workshops has been a total Xolalpan and Metepec sherds for the entire
topic of interest to researchers using ethnographic TMP survey is 54,550 to 156,396 sherds, or 34.9
(Arnold 1985; Arnold 1990; Graves 1991; Lon percent. This is lower than both themean (42.0 per
gacre and Stark 1992) and archaeological data cent) and the median (41.9 percent) of SMO per
(Arnold et al. 1993; Charlton et al. 1993; Curet total Xolalpan andMetepec sherds collected from
1993), as well as a combination of the two (Arnold survey tracts in theTlajinga district. Within the dis
and Santley 1993; Deal 1988; Pool 1992; Santley trict, atypical domestic assemblage (e.g., 2:S3W1)
et al. 1989; Stark 1985). Archaeologically recov has 39.3 percent SMO (57 of 145 sherds). An
erable indicators used to define ceramic production assemblage associated with a typical ritual context
locales include unusual concentrations of ceramic (e.g., 7A.-S3W1) has 23.2 percent SMO, or 13 of
wares or forms and the presence of wasters, spe 56 sherds. Surface collections from the excavated
cialized structures or features, and production SMO workshop at 33:S3W1 include 272 of 337
tools.51 require the presence of multiple produc sherds or 80.7 percent SMO (Altschul 1981;
tion indicators to define a collection as likely to have Sanders et al. 1982; Sheehy 1992, 1998; Storey
produced SMO (Arnold and Santley 1993; Stark 1992;Widmer and Storey 1993).
1985,1992; Stark and Garraty 2004). I rank these I define high proportions of SMO sherds per
collections based on the quality and quantity of total Xolalpan- and Metepec-phase sherds as equal

surface evidence for SMO production, following to or greater than a threshold value of 60 percent,
the assumption that qualitatively and quantitatively or the third quartile of this ratio for all collections
stronger indicators of production are generally from the Tlajinga district, following research in
associated with more intensive production (Sant other parts of Mesoamerica (Pool and Santley
ley et al. 1989:113). 1992:212). This proportion is greater than those
from the nonproduction contexts discussed above.
Unusual Concentrations of Wares or Forms
Figure 4 shows the distribution of the proportion
The first criterion for defining a likely SMO work of SMO sherds per total Xolalpan- andMetepec
shop is that the collection should include an unusu phase sherds plotted against the proportion of pos
ally high concentration of SMO sherds relative to sible SMO wasters per total SMO sherds. Those
total Xolalpan- and Metepec-phase sherds. collection tracts to the right of the vertical line have
Although I examined all Tlajinga district SMO for unusual concentrations of SMO sherds at or above
evidence of production, only those collections with the third quartile. Excluding collections with a
at least 100 Xolalpan- and Metepec-phase sherds small quantity (i.e., less than 100 sherds) of Xolal
were considered possible production locales in pan and Metepec pottery, collection tracts with at

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
 30 LATIN AMERICAN ANTIQUITY [Vol. 17, No. 1,2006

60.0 % SMO sherds per total

Xolalpanand Metepec sherds

40
moderatd strong
04S4W2

30

20 h
20.S3W2Q

12:S4W2OO20:S4W2

25:S3W2#
9:S3W2<?
10 17:S3W1
14.S3W2
01S4W2 6.4 % SMO wasters
per totalSMO sherds
46-.S3W1 21:S3W2
28:S3WlO (mean forall Tlajinga)
16:S3W2C 03:S4W2 WEM
^aasawg
%3;S3W1
3.S4W? IT 2S?4W1
5?wmm
02*S3W2
JT% SMO wasters
#15:S3W f Q32;S3W1
per total SMO sherds
j 9:S3W1??*:S3W1* 7:S3W2 } 49:83W1#|
J_| 6S3^r
-%1 (proportionat Tlajinga 33)
10 20 30 40 50 6D 70 80 90
SMO Sherds per Total
Xolalpan-Metepee Sherds

Figure 4. Scatterplot of total possible San Martin Orange (SMO) wasters to total SMO sherds and total SMO sherds to
total Xolalpan and Metepec sherds. Solid circles indicate collection tracts with at least 100 Xolalpan and Metepec sherds.
Note: 17,21, and 18:S3W2 have 97,98, and 99 total Xolalpan and Metepec sherds, respectively.

least 60 percent SMO include 5, 6, 7, 18, 21, 24, workshop at 33:S3W1 (3.3 percent) as a reason
and 25:S3W2; and 33 and 49:S3W1. able lower threshold for the relative proportion of
SMO wasters at likely Tlajinga district SMO work
Possible Wasters
shops. Those collections above the solid horizon
To decrease the possibility that the high proportion tal line in Figure 4 have unusual proportions of
of SMO reflects specialized use rather than pro SMO wasters that are equal to or exceed the pro
duction, I also require that collections from prob portion of SMO wasters I identified in theTMP sur
able workshops include waster sherds from vessels face collection from Tlajinga 33. Six collection
that would have been unusable (Figure 5). These tractsmeet the criteria of an unusual proportion of
production errors include vitrified, bloated, warped, SMO to totalXolalpan andMetepec sherds, a large
and fused sherds and sherds with severe cracking number of total sherds, and a high proportion of
(Pool 1992:291; Pool and Santley 1992:212; Rye SMO wasters per total SMO sherds: 5, 6, 21, 24,
1981:108-115). Also considered wasters are sherds and 25-.S3W2; and 33:S3W1. With 3.1 percent
from vessels that have been severely overtired, SMO wasters per sherd, 18:S3W2 is just slightly
refired, underfired, heated too quickly, or cooled below the 3.3 percent cutoff and has other indica
very rapidly and sherds with thick gobs of clay tors of pottery production.
adhering to the surface (Pool 1992:291;Rice 1987;
Stark Specialized Structures or Features
Rye 1981; Shepard 1985; Sinopoli 1991;
1985). I use the proportion of possible wasters col Although it is difficult to reliably identify special
lected by the TMP from the excavated SMO crater ized structures or features associated with produc

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
Sullivan] SPECIALIZED
PRODUCTION ORANGEWARE
OF SAN MART?N 31

Figure 5. Pottery production evidence from the Teotihuacan Mapping Project Tlajinga collections: (a) lumps of worked
and fired clay from 25:S4W1, (b) lunates from 24:S3W2, (c) a possible mold from 25:S3W2, (d) a possible rotational sup
port from 18:S3W2, and (e) a sherd with thick gobs of clay adhering to the surface from 24:S3W2.

tion based on surface artifacts alone (although, con 1999:72; Krotser and Rattray 1980:96; Redmond
cerning surface indications of lime kilns inPuebla, 1979; Santley et al. 1989; Stone andTYirnbuU 1941 ;
see Castanzo and Anderson 2004), the TMP col Whalen 1981:88-105). I expect surface collections
lections permit suggestions for future investiga from possible open firing areas to include unusual
tion. The partial excavation of Tlajinga 33 revealed proportions of deformed wasters, large overfired or
an open firing area on a patio outside the northwest retired sherds thatmay have been used as kiln fur
corner of the compound (Sheehy 1992; Storey niture, and concretions of baked or burnt clay. Other
1992;Widmerand Storey 1993). At Tlajinga 33 and indicators such as reddened soil or fire-cracked
in other parts of Mesoamerica, open firing areas or rock with angular fracture patterns from repeated,
open pit kilns have been identified in excavation high-temperature firings might be identified on the
based on the presence of a pit dug into the subsoil surface. However, because open firings do not
and associated with a thick layer of charcoal and maintain high temperatures for an extended period
ash, high densities of wasters and broken pottery, of time, they often are not preserved (Arnold
concentrations of prepared and of burnt clay, burnt 1978:342; Pool 1992:289;Rye 1981; Shepard
adobe and bedrock, and soil rich in carbon Abas( 1985; Stark 1985:165).
calMacias 1975,1976;Balkanskyetal. 1997:146; Iwas unable to confidently identify sherds fired
Bordaz 1964; Stark 1985;Winter and Payne 1976). to vitrification as a result of having been used to
More formal, updraft kilns have been excavated support vessels during firing. This may be related
elsewhere inMesoamerica, although evidence from to the relatively low firing temperatures (below
Terminal Formative and Early Classic period Teoti 1,000?C) of the open firing areas likely used at
huacan indicates that firing occurred in open areas Teotihuacan. Such firings did not sustain high
similar topit kilns inHidalgo, Nay arit,Oaxaca, and enough temperatures to induce the vitrification and
the Puebla-Tlaxcala area (Cabrera Castro 1988; bubbling often associated with kiln-fired ceramics
Flannery and Marcus 1983:299; Hernandez et al. (Hopkins 1995:580; Rye 1981:25). I was able to

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
 32LATIN AMERICAN ANTIQUITY [Vol. 17, No. 1, 2006

identify lumps of fired clay or daub thatmay reflect and in some cases old pots may be inverted to serve
stockpiles of raw material that burned in postoc as rotational aids (Arnold 1978:336,339,347; Hop
cupation structural fires ormay be thematerial used kins 1995:152; Rattray 1988:259-260; Reina and
in lining firing areas (Figure 5). These may be sim Hill 1978:22-23, plate 191; Sheehy 1992:506;
ilar to the fired adobes found in excavation of Tla Stark 1985:166; Widmer 1987:361-362). Other
jinga 33 and the clay concretions found in researchers argue thatTlajinga bowls were used as
abundance at other pottery production areas in molds for forming the base of SMO vessels (Rat
Mesoamerica (Balkansky et al. 1997:148; Red tray 1988:Figure 4). However, I do not consider
mond 1979; Sheehy 1992; Stark 1985). Lumps of them specialized tools used exclusively in pottery
fired clay or daub occur in four Tlajinga collections: production, as
they
were sometimes also "used as

17,24, and 34:S3W2 and 25:S4W1, of which only serving bowls and ultimately as a container for bur
24:S3W2 has other strong surface indications for ial of young infants" (Sheehy 1998:311).
the production of SMO. In the TMP Tlajinga collections I identified 24
Although the largecleared areas at35:S3Wl and lunates from 24:S3W2, three from 25:S4W1, and
7:S3W2 would provide considerable space for fir one from 48:S3W1.1 identified a possible mold at
ing pottery, excavation of Tlajinga 33 indicates that 25:S3W2 and a possible rotational support at
SMO potters fired their wares on residential patios 18:S3W2, with both collection tracts adjacent to
adjacent to apartment compounds rather than in 24:S3W2.?Molds, powdered with sand or dry clay
the cleared areas between residences. I also did not to prevent thewet vessels from sticking, were likely
find unusual proportions of SMO wasters and used to form the pocked or rough bases character
sherds or specialized production tools in the col istic of SMO craters (Hopkins 1995:152; Rattray
lection from 35:S3W1. Collection tract 7:S3W2 1988:259-260; Sheehy 1992:Chapter 9;Widmer
does have an unusual concentration of SMO sherds; 1987:361-362). Although equivocal as production
however, I only identified one SMO waster in this evidence because of their utility in a number of
collection. Excavation of areas with strong surface household and other specialized tasks, I also iden
indications for production like 24:S3W2 will pro tified six sherd scrapers at 25:S4W1 and one each
vide larger sample sizes and additional evidence at 34:S3W1 and at 9 and 24:S3W2.
that cannot typically be recovered from the surface,
including unfired vessels, lumps or stockpiles of Summary of Production Evidence
unfired clay, small stone or bone tools, and pigment Based on the evidence presented above, I distinguish
(e.g., used in painting some of the SMO basins). between strong and moderate evidence for the pro
duction of SMO. Collections that I consider to have
Production Tools
strong evidence of SMO production have at least
The presence of production tools like lunates, 100 Xolalpan andMetepec sherds and an unusual
molds, or possible rotational supports further proportion of total SMO sherds and of SMO waster
strengthens the ranking of these likely Tlajinga sherds. I define an unusual concentration of total
SMO workshops, although excavation is necessary SMO sherds as greater than or equal to the third
to confirm or refute the suggestions presented here. quartile of the distribution of SMO sherds per total
Production tools identified in the Tlajinga collec Xolalpan and Metepec sherds (Pool and Santley
tions include several lunates, a possible mold, a pos 1992:212). Unusual proportions of wasters are
sible support or rotational device, and reworked greater than or equal to the proportion of wasters
sherd scrapers (Figure 5). Other studies of pottery collected from the excavated SMO crater workshop
production at Teotihuacan and elsewhere in at Tlajinga 33. Several of the collections that fit the
Mesoamerica have identified lunates or crescent criteria for strong SMO production evidence also
shaped ceramic smoothers as being associated with contain specialized production tools such as lunates,
pottery production (Hernandez et al. 1999:76; Hop a possible mold, and a possible rotational support.
kins 1995:152; Krotser and Rattray 1980:96). Seven Tlajinga district apartment compounds have
Molds are used to help form the base of a vessel strong surface indications for SMO production: col
and to rotate the vessel in coiling and shaping. lection tracts 5,6,18,21,24, and 25:S3W2, as well
These tools vary greatly in shape, size, and form; as the excavated SMO crater workshop at 33:S3W1

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
Sullivan] SPECIALIZED PRODUCTION OF SAN MARTIN ORANGE WARE 33

*^" B.
o^^^^w

D D

Figure 6. Spatial distribution of San Martin Orange (SMO) production evidence in the Tlajinga district showing likely
SMO workshops with (a) >60 percent SMO per total Xolalpan- and Metepec-phase sherds, (b) >3.3 percent SMO wasters
per total SMO, (c) lumps of fired clay, and (d) specialized production tools (solid black indicates lunates, diagonal black
lines indicate a possible mold, horizontal black lines indicate a possible rotational support).

(Figure 6; Table 1). Although they have less than criteria of moderate SMO production evidence:
100 Xolalpan andMetepec sherds, I have included 48:S3W1; 14, 17, and 33:S3W2; and 25:S4W1.
18 and 21:S3W2 because their totals are close to Although the proportion of SMO wasters falls
this cutoff (99 and 98 sherds, respectively) and both below the cutoff of 3.3 percent, I also included
have other indicators of SMO production. Collec 25:S4W1 in this group because the surface collec
tion tract 21 :S3W2 has a high proportion of possi tion includes lunates and lumps of fired clay. Res
ble SMO wasters per total SMO sherds, exceeding idents at the apartment compounds with moderate
proportions and counts of SMO wasters from the evidence for SMO production may have produced
surface of 33:S3W1. The proportion of possible low quantities of SMO and/or some non-SMO
SMO wasters (3.1 percent) from 18:S3W2 is very ware. The collection from 25:S4W1 includes sev
close to the 3.3 percent collected from 33:S3W1, eral griddle sherds, and Krotser (1987:421) notes
and a possible rotational support was recovered that three apartment compounds in S3W1 and
from the compound. S3W2 made both SMO and griddles and that one
Moderate evidence of SMO production includes in S4W1 also made Matte Ware covers (for a
collections with greater than 100 Xolalpan and description of non-SMO vessel forms, see Rattray
Metepec sherds and unusual concentrations of 2001). The possibility of production of some other
SMO wasters, although proportions of total SMO ware at 24:S4W1?griddles or Matte Ware cov
fall below the third quartile. Five collections fit the ers?should be investigated further.

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
34 LATINAMERICAN
ANTIQUITY [Vol. 17, No. 1,2006

Table 1. Summary of Pottery Production Evidence from Tlajinga District Collections.

Possible SMO workshops.

Collection SMO SMO wasters Tools and other indicators

18:S3W2a 65/99 (65.66) 2/65 (3.08) 1 possible rotational support

21:S3W2 79/98 (80.61) 6/79 (7.59)
24:S3W2 496/578 (86.81) 61/496 (12.30) 24 lunates, 1 reworked sherd, 4 fired clay lumps
25:S3W2a 229/287 (79.79) 28/229 (12.23) 1 possible mold
33:S3Wla 272/337 (80.71) 9/272 (3.31)
5:S3W2a 190/248 (76.61) 7/190 (3.68)
6:S3W2a 295/360 (81.94) 12/295 (4.07)

Possible small-scale SMO production, or production of SMO and some non-SMO ware.

Collection SMO SMO wasters Tools and other indicators

14:S3W2a 77/155 (49.68) 6/77 (7.79)

17:S3W2b 52/97 (53.61) 5/52 (9.62) 1 fired clay lump
25:S4W1C 370/845 (43.79) 10/370 (2.70) 3 lunates, 6 reworked sherds, 9 fired clay lumps
33:S3W2b 58/136 (42.65) 2/58 (3.42)
48:S3Wla 26/222 (11.71) 2/26 (7.69) 1 lunate

Possible smaller-scale or household pottery production.

Collection SMO SMO wasters Tools and other indicators
20:S4W2d 14/29 (48.28) 2/14 (14.29)
26:S3W2e 31/53 (58.49) 2/31 (6.45)
29:S3W2b 55/69 (79.71) 1/55 (1.82)
32:S3Wlb 41/61 (67.21) 1/61 (1.64)
34:S3Wla 117/223 (52.47) 1/117 (.85) 1 reworked sherd
4:S3W2a 163/347 (46.97) 4/163 (2.45)
49:S3Wla 173/219 (79.00) 1/173 (.58)
9:S3Wla 165/328 (50.31) 1/165 (.61)
9:S3W2b 36/60 (60.00) 4/36 (11.11) 1 reworked sherd

Other collections with some indication of possible pottery production.

Collection SMO SMO wasters Tools and other indicators
1-.S4W2 22/31 (70.97) 2/22 (9.09)
12:S4W2 7/15 (46.67) 1/7 (14.29)
13:S4W1 41/108 (37.96) 1/41 (2.44)
15:S3W2 62/115 (53.91) 1/62 (1.61)
16:S3W2 49/82 (59.76) 2/49 (4.08)
20:S3W2 12/20 (60.00) 2/12 (16.67)
28:S3W1 14/31 (45.16) 1/14 (7.14)
3:S4W1 48/155 (30.97) 1/48 (2.08)
3:S4W2 44/67 (65.67) 2/44 (4.55)
34-.S3W2 13/37 (35.14) 0/13 (.00) 2 fired clay lumps
4:S4W2 11/31 (35.48) 4/11 (36.36)
44:S3Wlf 35/57 (61.40) 1/35 (2.86)
6:S3W1 55/148 (37.16) 1/55 (1.82)
7:S3W2f 71/113 (62.83) 1/62 (1.61)
Note: Columns indicate counts and proportions (in percent) of SMO per total Xolalpan and Metepec sherds, and of SMO
wasters per total SMO sherds, and production tools from Tlajinga collections.
aPossible ceramic workshop based on survey field notes.
bSpecial ceramic concentration based on survey field notes.
cPossible ceramic workshop (griddles, Matte Ware handled covers) based on survey field notes.
Possible firing area (Krotser 1987:421).
eSpecial ceramic concentration or place where pottery was fired based on survey field notes.
fPlazas with no evidence of architecture according to field notes.

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
Sullivan] SPECIALIZED OF SAN MART?N
PRODUCTION WARE
ORANGE 35

15.3% SMO amphora sherds

per totalSMO sherds
(meanof all Tlajingacollections)
25 mod?r?t?
ir-r
strong
CD00

20
o o
JIJE
o.a.
15
\ OO
10

2.1 % SMO amphorai

wasters per total
SMO amphora sherds!
(meanof allTlajinga
collections)_
\ ?Li 4m ??XU?
?48.-S3W1-'
2feS3W2*
10 20 30 40 50 60
SMO Amphora Sherds
per Total SMO Sherds

Figure 7. Scatterplot of the total possible San Martin Orange (SMO) amphora wasters to total SMO amphora sherds
and the total SMO amphora sherds to total SMO sherds. Solid circles indicate likely SMO workshops.

Nine Tlajinga collections also include some evi majority of the metamorphic inclusions and the
dence for pottery production and were noted in sur only example of polycrystalline quartz inclusions
vey as having evidence for possible pottery derive from the northwest of the city inmap squares
production: 9, 32, 34, and 49:S3W1; 4, 9, 26, and N6W2-W3 and N5W2-W3.
29:S3W2; and 20:S4W2.6 Fourteen other collec
tion tracts (6,28, and44:S3Wl; and Vessel Form Specialization
7,15,16,20,
34.-S3W2; 3 and 13:S4W1; and 1, 3, 4\ and Elsewhere inMesoamerica, the scale and intensity
12:S4W2) have some indications of pottery pro of pottery production have been assessed by con
duction but were not noted in survey as possibly sidering specialization in particular wares and
having been involved in pottery production. forms with larger, more specialized production
units tending to specialize in fewer wares and forms

Assessing Specialized Production (Santley et al. 1989:126). Vessel form specializa

tion is related to the scale and intensity of produc
Specialized production may be manifest in ware tion, in that the reduced number of actions required
and form specialization and dimensional stan formaking a few forms, rather than thewhole range
dardization (Arnold 1991a:364; Blackman et al. of forms, may permit more efficient production of
1993). Standardization may also be evident in clay massive quantities.
and temper composition as well as in stylistic and I consider amphorae, basins, and craters as these

formal attributes (Arnold and Santley 1993:231; were the three most common SMO forms distrib
Rice 1987:203; Sinopoli 1988). For the purposes uted throughout the city during the Xolalpan and
of this article, I focus on vessel form specialization Metepec phases (Hopkins 1995:132,152). Figure
and dimensional standardization because I have 7 shows the proportion of possible amphora wasters
not yet identified compositional subgroups within as a ratio of the total SMO amphora sherds plotted
theTlajinga district SMO. Hopkins ( 1995:464) has against the total SMO amphora sherds as a ratio of
documented distinctions in the composition of the total SMO sherds. Figures 8 and 9 plot these
SMO pastes throughout the city, observing that the proportions for basins and craters, respectively.

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
 36 LATIN AMERICAN ANTIQUITY [Vol. 17, No. 1, 2006

2. Summary of Likely SMO Workshops

Table and the
25:S3W2). The easternmost compounds in this
Vessel Forms Emphasized. Indicates Strong Evidence for
cluster have strong evidence for basin (5 and
the Production of a Particular Form. Indicates Moderate
Evidence for the Production of a Given SMO Form. 6:S3W2) and crater production (6:S3W2 and
33:S3W1), with moderate evidence for amphora
Probable Workshop Amphoras_Basins Craters production at 33:S3W1. Only one possible work
18:S3W2 shop shows a strong indication of having empha
2LS3W2 sized just one form (basins at 5:S3W2). As
24:S3W2 discussed below, an analysis of variance (ANOVA)
25.-S3W2
to test for the significance of observed differences
33:S3W1
inmeans suggests that the rim diameters of small
5:S3W2
6:S3W2 andmedium-large basins from 5:S3W2 are not sig
nificantly different from those from other Tlajinga
SMO workshops. No collection shows strong evi
Table 2 summarizes the vessel forms emphasized dence for having produced only craters or
at each of the seven possible SMO workshops that amphorae.
meet the criteria detailed above. Amphora Specialization. Although eclipsed by
In assessing the evidence for vessel form spe the crater form during the Xolalpan phase,
cialization in the Tlajinga district, I distinguish amphorae become the major SMO form in the
between strong and moderate evidence for the pro Metepec phase (Rattray 2001:265, 299). Strong
duction of a particular vessel form. These desig evidence for amphora production occurs only at 24
nations are based on the expectation thatworkshops and 25:S3W2, with moderate evidence at 33:S3W1
specializing in a particular form should have (Figures 7 and 10). Surface indications from
unusual proportions of waster sherds exhibiting fir 33:S3W1 are supported by excavation, which
ing errors (Stark 1985). Collections with propor revealed the production of SMO amphorae in small
tions greater than themean of both waster and total quantities at the apartment compound (Rattray
sherds of a particular form are considered to have 1988:263; Sheehy 1992). The distribution of
strong evidence for specializing in that form (e.g., lunates, tools especially useful for smoothing the
24:S3W2 in Figure 7).Workshops with moderate restricted necks of amphorae, reflects this limited
evidence for specializing in a particular SMO form production (Hopkins 1995; Krotser and Rattray
have greater than themean proportion of possible 1980:96). The vast majority (85.7 percent) of the
SMO wasters of that form but less than the mean lunates in the TMP Tlajinga district collections
proportion of total sherds of that form (e.g., derive from 24:S3W2. The collection from
33:S3W1 in Figure 7). In all cases, I also require 25:S4W1 includes 10.7 percent, and that from
that a given workshop assemblage contain greater 48:S3W1 has 3.6 percent of the total lunates.
than themean number of sherds ofthat vessel form Although 25:S4W1 and 48:S3W1 do not have
for all collection tracts in the Tlajinga district (i.e., strong evidence for the production of SMO ware,
so that unusual proportions are based on relatively residents may have produced ollas, jars, or other
large samples). The mean number of amphora similar forms with restricted necks or other con
sherds from Tlajinga collections is four, for basins stricted areas more easily smoothed with the
themean is one, and for craters themean is 30. crescent-shaped tools.

Most of the probable SMO workshops show Basin Specialization. SMO basins seem to have
moderate to strong evidence for producing two or been produced in much smaller quantities than
three of themajor SMO forms. Figure 10 summa either craters or amphorae, and Sheehy (1998:313)
rizes the evidence for strong and moderate pro has argued that they may have been imitations of
duction of the three major SMO forms from the another form. Basins are the only consistently
likely workshops identified above. The western painted form of SMO and may have been used for
portion of the cluster of SMO workshops in theTla serving particular kinds of food or for presenting
jinga district has strong evidence of amphora and food at specific functions like feasts or ceremonies
basin production (24 and 25:S3W2) and moderate (Rattray 2001:301). Strong evidence for basin pro
evidence of crater production (21, 24, and duction occurs at 5, 6, 24, and 25:S3W2 (Figures

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
Sullivan] OF SAN MART?N
PRODUCTION
SPECIALIZED WARE
ORANGE 37

2.5 % SMO basin sherds

per total SMO sherds
(mean of all Tlajinga collections)
100 "2B:S3W2
T
(moderate) strong
90
80

70

60

24:S3W2
50

40 h l6:S3W2
23,0 % SMO basin
wasters per total
SMO basin sherds ?
30 k #25:S3W2 #5:S3W2 {mean of all Tlajinga
collections)_
20

21;S3W2
09:S3W1

i.
^
28;S3Wr

10 15 20
SMO Basin Sherds
per Total SMO Sherds
Figure 8. Scatterplot of total possible San Martin Orange (SMO) basin wasters to total SMO basin sherds and total SMO
basin sherds to total SMO sherds. Solid circles indicate likely SMO workshops.

8 and 10). tracts 21, 24, and 25:S3W2 show moderate evi
Crater Specialization. During the Late Xolal dence for crater production. The collection from
pan phase SMO craters were distributed to most 18:S3W2 falls below the cut-off for all SMO forms;
parts of the city, despite their waning popularity in however, the collection from this apartment com

later Metepec times (Rattray 2001:265, 299). pound includes a possible rotational support that
Craters are by far themost abundant Tlajinga SMO may have been used to form SMO vessels.
form in terms of both possible wasters and total
Dimensional Standardization
sherds, and evidence for the production of this form
ismost widespread in theTlajinga district. The col As with specialization by ware and form, dimen
lection from 6:S3W2 shows strong evidence for sional standardization may also be related to the
SMO crater production (Figures 9-10). The exca scale and intensity of production as producers opt
vated SMO crater workshop at 33:S3W1 has a for the increased efficiency of making a single size
slightly lower proportion of SMO crater wasters for vessels produced in large quantities. Standard
than the 2.6 percent mean for all Tlajinga collec ized sizes facilitate stacking of vessels in firing,
tions, with 2.4 percent crater wasters. Collection storage, and transport to markets (Arnold 1991a;

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
38 LATINAMERICAN
ANTIQUITY [Vol. 17, No. 1,2006

82.0 % SMO crater sherds

per totalSMO sherds
(mean of all Tlajinga collections)
15 T
moderate strong

04&S3W1

Q9S3W2
bfe 10r
25:S3W2 24.S3W2
Q28.-S3W1

026:S3W2

01:S4W2
2.6 % SMO crater
01T:S3W2
wasters per total
#21:S3W2 SMO crater sherds
8:S3W2? 044:S3W1
3.S4W20 or'4:S3W2
(mean of all Tlajinga
fipJl?Ct?ons)
"
15:S3W20J 06:S3W1
14:S3W20 25:S4Wld
32:S3W1;3.541??

|5:S3W2 ? ?l 49:S3W I
Iq29S3W? &34:S3W2
0

40 50 60 . SOi 70 13:S4W1
100
tprr20:S4W2
12-.S4W2;
SMO Crater Sherds
per Total SMO Sherds
Figure 9. Scatterplot of total possible San Martin Orange (SMO) crater wasters to total SMO crater sherds and total
SMO crater sherds to total SMO sherds. Solid circles indicate likely SMO workshops.

Blackman et al. 1993:71-73; Sinopoli 1988). A In the analysis presented here, I use 1,891 of the
review of the ethnographic and archaeological lit 2,055 total rim diameters that I recorded from Tla
erature reflects the complexity inherent in linking jinga SMO sherds with an arc length greater than
product standardization to specialized ceramic pro or equal to 5 cm. These include 281 amphora (13.7
duction (Arnold 1991a; Arnold and Santley 1993; percent of the 2,055 total rims), 180 basin (8.8 per
Blackman et al. 1993; Costin and Hagstrum 1995; cent), and 1,430 crater (69.6 percent) sherds.
Longacre et al. 1988; Sinopoli 1988). Other Another 164 sherds (7.9 percent of the total rim
researchers have used various methods to examine diameters recorded) are from other vessel forms
dimensional standardization (Arnold 1991a:366; thatwere not typically made in SMO ware for dis
Arnold and Santley 1993:232; Blackman et al. tribution outside of Tlajinga but which may have
1993:71; Costin and Hagstrum 1995:631; Lon been produced for local consumption by Tlajinga
gacre et al. 1988:103-105; Plog 1985; Pool 1990; residents (Rattray 2001:263, 299).
Stark 1995:238-239). I consider histograms, I examined histograms of rim diameters from
descriptive statistics, box plots, andANOVA, which the three different vessel forms for evidence of mul
is based on the F distribution and permits simulta tiple modes indicating possible size groups. I have
neous
comparison of two or more samples to eval divided basins into small (14-35 cm), medium
uate whether there is a significant difference in the (36-60 cm), and medium-large (36-86 cm) size
means (Kirby 1993; Sheskin 2000:511; Stevens classes based on modalities I observed in the basin
1986). rim diameter distribution (Figure ll).71 have com

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
Sullivan] SPECIALIZED
PRODUCTION ORANGEWARE
OF SAN MARTIN 39

A. B.
X

C.

Figure 10. Summary of vessel form specialization evidence. Strong evidence indicated in solid black; moderate evidence
indicated by diagonal black lines: (a) amphorae, (b) basins, and (c) craters.

bined the medium and large categories together hypothesis, a comparison of coefficients of varia
because I only recorded 14 examples of large tion for amphora and small basin sherds indicates
(61-86 cm) basin rim sherds in the district, half of that the combined workshop assemblages are
which were from workshops. I did not divide slightly more standardized than those from the
amphora and crater rim diameters into separate size combined nonworkshop assemblages. Unexpected
classes because the distributions of rim diameters based on the standardization hypothesis, coeffi
for these vessel forms are unimodal and do not sug cients of variation from assemblages of medium
gest multiple size classes. large basin and of crater sherds demonstrate that
I also inspected various summary statistics for combined workshop assemblages are slightly less
assemblages of each of the threemajor vessel forms standardized than the combined nonworkshop
in Tlajinga collection tracts including median, assemblages. In all cases differences are slight, as
mean, standard deviation (s), variance (s2), and is supported by a comparison of the interquartile
coefficient of variation (CV) as presented in Table ranges of the box plots of rim diameters from the
3 (Sheskin 2000; Sullivan 2002:Tables 7-8). The three vessel forms from both workshop and non
coefficient of variation is the standard deviation workshop contexts (Figures 12-15).
divided by the mean and is particularly useful in The distribution of amphora rim diameters from
studies of dimensional standardization as it pro the likely SMO workshops follows the expectation
vides a standardized measure of dispersion (Shen of the standardization hypothesis (Figure 12).
nan 1992:44). As expected by the standardization Likely SMO workshops with strong indications for

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
40 LATIN AMERICAN ANTIQUITY [Vol. 17, No. 1,2006

A B

0.2 ^

10 20 10 20 30 40 50 60 70 80 90
AmphoraRimDiameter Basin RimDiameter

0 10 20 30 40 50 60 70 80 90
CraterRimDiameter

Figure 11. Distribution of San Martin Orange (a) amphora, (b) basin, and (c) crater rim diameters from Tlajinga dis
trict collections.

amphora production (24 and 25:S3W2) have expectation of the standardization hypothesis.
smaller interquartile ranges (i.e., the dispersion I also used ANOVA to test for significant dif
around themean, or typical observation) than the ferences in the collections (Table 4).9 Other
excavated workshop at 33:S3W1 where some researchers have argued that SMO amphorae are
amphorae
were
produced. For small basins, the uniform enough that they may be assigned to spe
expectation of the standardization hypothesis is not cific workshops (Rattray 1988:261, 2001:263).
met, in that the only likely SMO workshop thatmay ANOVA does not support this argument in terms
have focused on producing basins (5:S3W2) has a of dimensional standardization of rim diameters;
very large interquartile range, reflecting greater dis however, other attributes such as rim form and sur
persion of scores around themean (Figure 13). For face finish should also be considered. Although
medium-large basins, the distribution of rim diam workshop assemblages of small andmedium-large
eters from 5:S3W2 does conform to the expecta basin rim diameters tend to be somewhat smaller
tions of the standardization hypothesis (Figure 14).8 than those for nonworkshops, the only significant
Craters do not conform to the expectation of the difference between workshop and nonworkshop
= =
standardization hypothesis, in that the excavated assemblages occurs with craters (F 43.55; df
SMO crater workshop has a relatively less stan =
1, 1428; p .00).10 ANOVA results comparing
dardized assemblage (larger interquartile range) crater assemblages from the individual workshops
than the likely SMO workshops (24 and 25:S3W2) also show a significant difference (F = 5.31; df=
that show only moderate surface indications for 6, 523; p = .00). Collections from 5 and 6:S3W2
producing the crater form (Figure 15). Crater rim and 3 3 :S 3W1 have relatively larger craters than the
diameters from 6:S3W2, with strong surface indi other likely SMO crater workshops to the west.
cations for crater production, do conform to the Four different post hoc tests indicate significant

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
 Sullivan] SPECIALIZED PRODUCTION OF SAN MARTIN ORANGE WARE 41

Table 3. Sample Size (AO,Median, Mean, Standard Deviation (s), Variance (s2), and Coefficient of Variation (CV) for
Amphora, Small Basin, Medium-Large Basin, and Crater Rim Diameters.

Amphoras_

_N_Median_Mean_s_s?_CV__
281 All 12.00 12.63 2.52 6.33 .20
120 Workshops 12.50 12.68 4.81.17
2.19

Non-workshops 161 12.00 12.60 7.51

.222.74

Small Basins_

_N_Median_Mean_s_s?_CV__
91 All 28.00 27.15 5.00 .19
24.98

Workshops 49 27.00 26.98 4.70 22.08 .18

Non-workshops 42 28.00 27.43 5.37

28.84 .20

Med-Lg Basins_

_N_Median_Mean_s_s?_CV__
89 All 46.00 49.19 10.82 .22
117.04
40 Workshops 45.00 49.00 12.00 .25
143.95

Non-workshops 49 47.00 49.35 9.8897.56 .20

Craters_

_N_Median_Mean_s_s^_CV__
1430 All 44.00 44.81 12.39 .28
153.62

Workshops 530 42.00 42.03 12.07 .29

145.60

Non-workshops_900_45.00 46.44_12,30_151.28_127_

differences between 33:S3W1 and 24 and 25:S3W2 cialized SMO workshop at 6:S3W2, with strong
(for probabilities, see Figure 16).11With both the evidence for both basin and crater production, has
Bonferroni adjustment and the Tukey test, signifi a relatively more standardized assemblage of
cant differences are also documented between craters than does 33:S3W1, with strong evidence
33:S3W1 and 6:S3W2 and between 25:S3W2 and for crater production. The threewestern workshops
5:S3W2. The Fisher test, which allows the small with only moderate indications for crater produc
est difference between means, indicates tion also have relatively more standardized assem
significant
differences between 33:S3W1 and all other possi blages than 33:S3W1, with strong indication for
ble workshops except 5:S3W2, between 5:S3W2 crater production.
and 24 and 25:S3W2, and between 6.S3W2 and
25:S3W2. The Organization of Tlajinga
Considering the spatial patterns, those work District SMO Production
shops with the strongest evidence for SMO crater
production are located to the east and generally To interpret the vessel form specialization and
produced larger craters (6:S3W2 and 33:S3W1); dimensional standardization data and discuss the
meanwhile, those compounds with moderate evi organization of SMO production in the Tlajinga
dence for crater production, and which were gen district, I return to Costin's (1991, 2001) four
erally producing smaller craters, are located to the parameters of production?context, concentra

west (21,24, and 25:S3W2). However, not all east tion, scale, and intensity. Context involves either
ern and western SMO crater workshops show sig independent or attached specialists. Excavation
nificant differences, as themore conservative post of Tlajinga 33 indicates that in the Tlajinga dis
hoc tests show that the collection from 6:S3W2, in trict, SMO potters were independent, with pro
the eastern cluster, is not significantly different duction facilities located within individual
from any of those from the western cluster work apartment compounds. I did not find evidence for
shops. The assemblage from the relatively less spe SMO production in the open area at 35:S3W1

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
 42 LATINAMERICAN
ANTIQUITY [Vol. 17, No. 1,2006

18:S3W2 21:S3W2 24:S3W2

n= 4 n= 5 n = 68

i....
-n>
i.... i
?QH ?[D?
10 15 10 15 10 15 20
RimDiameter Rim RimDiameter

25:S3W2 33:S3W1 5:S3W2

n = 19 n=12 n= 3

KD-<
Imilnil.?iin* IIIJ iiiniin
HXh HD>
tiiniiiil
20 10 15 20 10 15
RimDiameter RimDiameter

6:S3W2 NONWKSPS WKSPS

n=9 n = 161 n = 120

-Oh
.i.... i.... i. I i i i t lit i i Ii
15 10 15 20
nDiameter Rim

Figure 12. Amphora rim diameters from individual San Martin Orange workshops, all workshop collections combined,
and all non-workshop collections combined.

near the Eastern Platform. None of the possible the Street of the Dead and could have restricted
SMO workshops identified in this study is in access from the city's main north-south artery into,
direct association with the Eastern Platform, and and out of, theTlajinga district. Access to localmar
the only other platform or temple structure in the kets at Tlajinga may have been regulated by struc
district has been variously interpreted by differ tures like the four-temple complex (7A-7D) on this
ent researchers (Krotser 1987:421; Mill?n platform, and the complex's proximity to the Street
1973:18; Rattray 1988:263). of the Dead may have facilitated themovement of
The second parameter, the concentration of pro finished products to city markets. Alternatively,
duction, involves the spatial arrangement of pro these concentrations may be the result of cooking
duction units as either more nucleated or more and serving food and fermenting and storing pulque
dispersed. The Tlajinga district itself is isolated used in rituals or feasts. Further consideration of
from themain civic-ceremonial core of the city and the distribution of particular SMO forms and ves
is spatially delimited by the R?o San Lorenzo and sel sizes from surface collections and excavations
the Barranca Patlachique to the north and south throughout the city may clarify some of themech
west, respectively. The Eastern Platform creates a anisms by which SMO was distributed. Distribu
formal boundary between the Tlajinga district and tional studies will also permit further investigation

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
Sullivan] SPECIALIZED
PRODUCTION ORANGEWARE
OF SANMART?N 43

18:S3W2 21.S3W2 24.S3W2

n=0 n=2 ns20

m ?CE}-4
I t t A li I I I I I tul ili I I I I I I I I ln.iliiliili?I I I i I I I I I
1,1 I .1 lam?n?
0 10 20 30 40 0 10 20 30 40
RlmC

25:S3W2 33:S3W1 5.S3W2

n=8 n*8 n=5

?O HH] -C
11111111111.>.i
10 20 30 10 20 40 10 20 30
RitnDwTWtef Rim RbnDiameter

6:S3W2 NONWKSPS WKSPS

n=6 n = 42 n = 49

-TTh ??rxH
I i i i * I i i i i I ii i i L?LiiIi i >il 11111111111111111 I..!, t.i il i t i i I i i i i I i i i i I
0 10 20 30 40 0 10 20 30 0 10 20 30 40
RknDi&rnster RimDiameter Rim G

Figure 13. Small basin rim diameters from individual San Martin Orange workshops, all workshop collections com
bined, and all non-workshop collections combined.

of Manzanilla's (1992:328) suggestion that SMO production (e.g., children can help with certain
may have been used for state-sponsored storage as tasks like collecting some rawmaterials). I assessed
unusual concentrations have been found about 300 the scale and intensity of SMO production in the
m west of the northern portion of the Street of the Tlajinga district by considering vessel form spe
Dead (Cowgill 1987). cialization and dimensional standardization within
The third and fourth parameters of specialized the threemajor SMO forms. I expected greater spe
production are concerned with the nature of labor cialization in forms (e.g., producing only one form)
relations and the level of skill of the potters. The and more highly standardized products in nucle
labor-oriented production units of nucleated work ated workshops because the "products of large
shops are based on recruiting potters with high lev workshops are likely to be more standardized than
els of skill and efficiency who are able to meet those of individual forms of production because
considerable consumer demand for their goods. workshop artisans generally work in close prox
Skill also plays a role in doling out tasks to family imity under supervision, share technology, and
members in community specialization; however, draw from a common store of tools and raw mate

kinship may be more important than skill and effi rials" (Costin and Hagstrum 1995:624).
ciency in terms of participation in some aspects of Evidence for vessel form specialization in the

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
44 LATIN AMERICAN ANTIQUITY [Vol. 17, No. 1,2006

18.S3W2 21:S3W2 24:S3W2

n= 0 n= 2 n = 17

4J]->?
30 60 90 0 30 60
RimDiameter RimDiameter

25-.S3W2 33:S3W1 5:S3W2

n=7 n=4 n= 7

HD HZD ? ? ?
{EH
0 30 60 30 60 90 0 30 60
RimDiameter RimDiameter RimDiameter

6:S3W2 NONWKSPS WKSPS

n= 3 n = 49 n = 40

0 HID? CD?
.ii iii i i i i A. I iliiti I ft i I I I I tiltil
0 30 60 0 30 60 30 60
Rim Diameter RimDiameter RimDiameter

Figure 14. Medium-large basin rim diameters from individual San Martin Orange workshops, all workshop collections
combined, and all non-workshop collections combined.

Tlajinga district indicates thatmost of the probable dardization hypothesis with workshops that were
SMO workshops produced two of the three major more specialized in terms of vessel form produc
forms, as confirmed in the excavation of Tlajinga tion (e.g., craters at 33:S3W1) producing less stan
33. In some cases, possible SMO workshops with dardized assemblages than workshops with strong
strong evidence for the production of two of the indications for the production of two SMO forms
three major forms and basins at 24 craters at 6:S3W2, where there is also a strong
(e.g., amphorae (e.g.,
and 25:S3W2) also have some more moderate evi indication for basin production). Significant differ
dence for producing the third major form (e.g., ences in crater rim diameters from adjacent work

craters at 24 and 25:S3W2). Only one compound shops with the strongest evidence for crater
has strong evidence for emphasizing a single form production (6:S3W2 and 33:S3W1) argue for rela
(basins at 5:S3W2). However, ANOVA results show tively less specialized production by independent
workshop and nonworkshop assemblages to be sig producers in a single community rather than the
nificantly different only in the case of craters, with aggregation of workshops producing similar forms
workshops generally having assemblages of smaller and sizes expected of nucleated workshops.
craters than nonworkshops. Further analysis
com Evidence for vessel form specialization and
paring the assemblages of individual workshops, dimensional standardization within the three ves
however, contradicts the expectations of the stan sel forms indicates that SMO production in theTla

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
Sullivan] SPECIALIZED
PRODUCTION ORANGEWARE
OF SAN MART?N 45

18.S3W2 21:S3W2 24:S3W2

n = 18 n = 12 n = 200

-o hid?
'* * '
30 60 30 60 30 60
RimDiameter RimDiameter RimDiameter

25:S3W2 33:S3W1 5:S3W2

n = 34 n = 67 n = 60

?CD?" ?CO
i i i i i i
30 60 30 60 30 60
RimDiameter RimDiameter RimDiameter

6:S3W2 NON WKSPS WKSPS

n = 139 n = 900 n = 530

-CO ?rn

30 60 30 30 60
RimDiameter Rim RimDiameter

Figure 15. Crater rim diameters from individual San Martin Orange workshops, all workshop collections combined, and
all non-workshop collections combined.

jinga district was relatively less specialized than obtain high-quality fuel for firing) and the consid
expected for "themass production and distribution erable time pressures they may have faced in pro
of San Martin Orange wares" (Krotser 1987:421). ducing goods distributed and used throughout the
Based on the standardization hypothesis, I expected city (Hopkins 1995:626; Sheehy 1992).
more
specialized producers to make more stan The present analysis indicates that probable Tla
dardized goods because of the increased efficiency jinga SMO workshops were residential units that
and skill related to repeated pottery production. likely engaged in pottery production on a full- or
The results from this study indicate that SMO pot part-time basis. Variability in the quality and quan
tery produced in the Tlajinga district was neither tity of production evidence at probable Tlajinga
very specialized in form nor very standardized in SMO workshops may reflect differences in pro
size. The results of this study do support other duction intensity, as other researchers have found
research that has documented irregularity in the that the incidence of production evidence tends to
apparent firing temperature and in the forming and increase with the intensity of production (Santley
finishing of SMO craters (Hopkins 1995; Sheehy et al. 1989:113). The qualitative and quantitative
1992). These characteristics may reflect both the differences in the types of SMO production evi
relatively poor socioeconomic status of the Tla dence recovered from 24 and 25:S3W2 (unusual
jinga potters (e.g., making it difficult for them to proportions of SMO wasters plus specialized pro

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
 46 LATINAMERICAN
ANTIQUITY [Vol. 17, No. 1,2006

Table 4. Summary of ANOVA Results Testing for a Significant Difference in Rim Diameters of Different SMO Vessel
Forms from Workshop and Non-Workshop Apartment Compounds.

Testing for a significant Significant

Vessel Form difference between difference?
JL
Amphoras Workshop and Non-Workshop .50 1,279 NO .82
All Workshops 1.63 7,111.15NO
Small Basins Workshop and Non-Workshop .23 .63 NO
1,89
All Workshops .95 .46
NO5,43

Medium-Large Basins Workshop and Non-Workshop 1,.02

87 NO
All Workshops .07 5,34 .67 NO

Craters Workshop and Non-Workshop 43.55 1,1428 .00 YES

_AllWorkshops_531_6,523_.00_YES

duction tools) compared with the surface indica zation of SMO crater production at Tlajinga 33,
tions for SMO production at Tlajinga 33 (no spe Sheehy (1992,1998) has found production similar
cialized production tools) may distinguish between to that observed ethnographically in highland
full- and part-time specialization in the craft.At Tla Guatemala, where themaximal unit of production
jinga 33, for example, residents also engaged in lap is the extended family?what Widmer and Storey
idary production (Widmer 1991; Widmer and (1993) identify as the lineage or compound. The
Storey 1993). multicraft focus at Tlajinga 33 lends further sup
Architectural changes at Tlajinga 33 during the port to the idea that SMO production was only a
transition from the Late Xolalpan to Metepec part-time pursuit, although it is possible that at cer
phases may be related to the barrio-wide adoption tain times more residents participated in pottery
of SMO production (Storey 1992:92), although production depending on production requirements
lapidary production continued at 33:S3W1 even and the scheduling of other tasks. This best fits the
after SMO production began (Widmer 1991 Wid
; model for community specialists rather than that
mer and Storey 1993:96). Examining the organi for nucleated workshops, where the high-volume

Fisher's Test
Least-Significant-Diffoeuce Bonferroni
Adjustment
Matrixofpairwisecomparison
probabilities: Matrixofpairwise
comparison
probabilities:

18:S3W221:S3W2 24:S3W2 25:S3W2 33:S3W1 5:S3W2 6:S3W2 18:S3W221:S3W2 24:S3W2 25:S3W2 33:S3W1 5:S3W2 6:S3W2

18:S3W2 1.00 18:S3W2 1.00

21:S3W2 0.76 1.00 21:S3W2 1.00 1.00
24:S3W2 0.86 0.59 24:S3W2 1.00 1.00 1.00
25:S3W2 0.32 0.61 25:S3W2 1.00 1.00 1.00 1.00
33:S3W1 r??n f??ll 1.00 33:S3W1 0.27 0.28 [cT5?] lO.OOl 1.00
5:S3W2 0.12 0.10 0.16 1.00 5:S3W2 1.00 1.00 026 {(?M 100 1.00
6:S3W2 0.58 0.40 0.08 6:S3W2 1.00 1.00 1.00 0.54 f?!?H 1.00 1.00
HU

TukeyHSDMultipleComparisons. Scheff?Test
Matrixofpairwisecomparison
probabilities: Matrixofpairwise probabilities:
comparison

18:S3W221:S3W2 24:S3W2 25:S3W2 33:S3W1 5:S3W2 6:S3W2 1&S3W221:S3W2 24:S3W2 25:S3W2 33:S3W1 5:S3W2 6:S3W2

18:S3W2 1.00 1.00

21:S3W2 1.00 1.00 1.00 1.00
24:S3W2 1.00 1.00 1.00 1.00 1.00 1.00
25:S3W2 0.96 1.00 0.55 1.00 0.99 1.00 0.78 1.00
33:S3W1 0.16 0.16 [???j 1.00 0.40 0.40 [???| [a??] 1.00
0.16 [???| 080 10? 0.88 0.83 0.40 0.10 0.92 1.00
5:S3W2 0.72 0.63 {wE]
6:S3W2 0.10 0.98 0.98 0.28 0.56 1.00 1.00 0.99 0.99 0.55 0.05 0.79 1.00
\OM?

Figure 16. Results of post hoc tests comparing results from analysis of variance of crater rim diameters.

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
 Sullivan] SPECIALIZED OF SAN MARTIN
PRODUCTION WARE 47
ORANGE

production of goods conflicts or interferes with evidence of ceramic production, but no kiln has yet
normal household tasks, requiring production facil been excavated atClassic period Teotihuacan (Hop
ities separate from domestic structures (Santley et kins 1995:78). Within the apartment compound
al. 1989:110). and SMO crater workshop at 33:S3W1, the only
Craft wards, like the community of aggregated possible firing area identified in excavation is an
small autonomous production unit-based SMO open area on a patio in the northwest corner of the
potters at Tlajinga, have also been documented in compound, adjacent to a possible area for refuse
several regions ofMesoamerica including Cotaxtla disposal (Widmer and Storey 1993). It is difficult
(Ohnersorgen 2001), Coyotepec (van de Velde and to document firing and other specialized areas used
van de Velde 1939), Matacapan (Santley et al. for ceramic production outside of residential struc
1989:120, 127), Otumba (Charlton et al. 1991; tures because excavations often focus on architec
Nichols 1994), Tula (Healan 1986), and Xochi turally delimited areas; it is nearly impossible to
calco (Hirth 2000). InTlaxcala, kilns were centrally confidently do so from surface collections alone.
located and surrounded by residences, possibly The presence of wasters also represents strong
indicating cooperation among potters of different evidence for ceramic production. However, various

households (Abascal Macias 1976). Outside of ethnographic examples have shown that the inci
Mesoamerica, village or community specialization dence of wasters varies greatly (Rye 1981; Santley
has been recognized archaeologically in theYana et al. 1989:111; Stark 1985), and in some cases
marca Valley of Peru (Costin and Hagstrum 1995) wasters have been found where there is no other
and in southeastern Utah (Hegmon et al. 1995) and evidence for ceramic production (Santley et al.
ethnographically in the Philippines (Longacre and 1989:111). Although proximity to raw materials
Stark 1992), southern Peru (Mohr Chavez 1992), may be important to production, this variable alone
and Egypt (Lacovara 1985). Internal differentiation is not sufficient to document pottery specialization
of ceramic production industries and variability in (Stark 1985:164). In the case of the Tlajinga SMO
the quality and quantity of production evidence potters this variable was not essential to defining
have been noted in other regions of Mesoamerica SMO production locales because all apartment
(Arnold and Santley 1993; Arnold et al. 1993;Ohn compounds are very close to the R?o San Lorenzo
ersorgen 2001; Pool and Santley 1992). and the Barranca Patlachique, both of which may
have supplied water, clay, and temper. Unusual
concentrations of wares and forms may indicate
Conclusions and Suggestions
for Future Research specialized production, although they may also
relate to specialized consumption (Santley et al.
This study suggests avenues for refining archaeo 1989:111). The use of multiple criteria for defin
logical studies of specialized production at Teoti ing areas of likely pottery production helps ensure
huacan and elsewhere. Various authors have argued that unusual concentrations alone are not sufficient
for the use of multiple lines of evidence in defin for recognizing pottery production (Stark and Gar
ing likely production locales (Feinman 1982; Rice raty 2004). Based on the Tlajinga SMO data, there
1981; Rye 1981; Santley et al. 1989; Stark 1985). may be little patterning when one type of produc
I have considered both the quantity and the qual tion evidence (tools) is plotted against another (pos
ity of several indicators of ceramic production in sible wasters), except where production may have
documenting probable SMO workshops in theTla been relatively more intensive (e.g., 24:S3W2 has
jinga district. Although lunates, ceramic scrapers, the greatest number and proportion of SMO wasters
and various types of ground- and chipped-stone and of specialized production tools).
implements may have been used in producing SMO Less intensive production that possibly also
vessels, lunates aremore strongly associated with occurred less frequently may be more difficult to
ceramic production at Teotihuacan than are items document using surface indications alone. The
like manos and metates or ceramic or obsidian excavated SMO crater and amphora workshop at
scrapers that could have been used in a variety of 33:S3W1, for example, has no ceramic production
household tasks or in producing other nonceramic tools and fewer than 10 possible SMO wasters (3.3
goods. Features such as firing areas represent strong percent of the 272 total SMO sherds) collected

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
 48LATIN AMERICAN ANTIQUITY [Vol. 17, No. 1, 2006

from the surface. Wasters, kiln furniture, and spe ethnographie examples because of differences in
cialized production tools were recovered in exca the time depth recorded in each study and the num
vation (Rattray 1988). Excavation at 24:S3W2 is ber of producers contributing to these samples
advisable considering the quality and quantity of (Costin and Hagstrum 1995:632). Problems also
SMO production evidence in the surface collections arise because researchers see attributes such as effi
and the contrast with indicators found on the sur ciency developing in distinct economic circum
face of the excavated SMO workshop at 33:S3W1. stances (Feinman 1982; Hagstrum 1988; Hopkins
The results of this study also provide a basis for 1995:51), illustrating the complex nature of spe
the theoretical treatment of specialized production, cialized production.
reinforcing the argument made by Costin (1991:4) Iwould like to thank Dr. George Cowgill,
Acknowledgments.
and others that production should be seen as a con
my M.A. committee chair, for providing his guidance and
tinuum of "types and degrees" of specialization.
insights throughout all phases of this research. I also thank
The arbitrary thresholds selected in this study were Drs. Arleyn Simon and Barbara Stark, members of my M.A.
chosen to recognize those Tlajinga collection tracts committee, as well as Drs. Ben Nelson, Thomas Charlton,
Charles Kolb, Linda Manzanilla, and Ian
with strong surface indications for SMO produc Mary Hopkins,
Robertson and an anonymous reviewer whose suggestions
tion and may have resulted in not documenting
have greatly enriched this article. Don Pedro Ba?os and Don
areas involved in less intensive production that Zeferino Ortega were extremely helpful in locating collec

yielded little to no surface indications for such pro tions at the lab and have made considerable contributions to
duction. As discussed above, even within those archaeological research at Teotihuacan. Oralia Cabrera

kindly offered helpful editorial comments on the Spanish

seven aggregated small autonomous production
abstract. Research for this article was made possible by
units with strong surface indications for SMO pro
funding from the Friends of the Center for Latin American
duction, variability in the quality and quantity of Studies and the Department of Anthropology at Arizona
the evidence may reflect distinctions in the inten State University (ASU), both of which supported my

sity with which SMO was produced by the resi research at the ASU-managed archaeological research center
in San Juan Teotihuacan, Mexico. Surface collections used
dents of different apartment compounds. Five other
in this study were collected by the Teotihuacan Mapping
apartment compounds in the Tlajinga district show
Project, directed by Ren? Mill?n with principal associates
indications of involvement in some pottery pro Bruce Drewitt and George Cowgill. I gratefully acknowl
duction. This may have included less intensive pro edge the support I have received in researching and writing
duction of SMO and/or the production of some this article; however, any errors or omissions are my own

other non-SMO ware, for they do not meet the cri responsibility.

teria of the probable SMO workshops suggested

here but do show other indications for pottery pro References Cited
duction. Nine other Tlajinga collection tractsmay
Abascal Macias, Rafael
have engaged in nonspecialized household pro 1975 Hornos prehisp?nicos en la regi?n de Tlaxcala. InXIII
duction; they do not fit the criteria required for Mesa Redonda: Balance y perspectiva de la antropolog?a
de Mesoam?rica y del norte deM?xico, Arqueolog?a I,No.
identifying likely SMO workshops but do show de Antropolog?a,
13:1, pp. 189-198. Sociedad Mexicana
some other indications of possible pottery produc
M?xico, D.F.
tion. These results echo variability in the organi 1976 Los primeros pueblos alfareros prehisp?nicos. In El
Proyecto Arqueol?gico Puebla-Tlaxcala, coordinated by
zation of craft production that has been identified
Angel Garc?a Cook, pp. 49-52. Suplemento, Comunica
elsewhere inMesoamerica (e.g., Charlton et al. ciones, Proyecto Puebla-Tlaxcala, III. Fundaci?n Alemana
1991; Ohnersorgen 2001; Santley et al. 1989). para la Investigaci?n Cient?fica, Puebla.
Costin and Hagstrum (1995:623) have noted the Altschul, Jeffrey H.
1981 Spatial and Statistical Evidence for Social Groupings
complexity of isolating skill, control, standardiza at Teotihuacan, Mexico. Ph.D. dissertation, Brandeis Uni
tion, and labor investment in the archaeological versity. University Microfilms, Ann Arbor.
1987 Social Districts of Teotihuacan. In Teotihuacan,
record; and more ethnographic and archaeological nuevos datos, nuevas s?ntesis, nuevos problemas, edited
studies are needed to clarify the influence of these de Tapia and Evelyn C. Rattray, pp.
by Emily McClung
factors. Pressures on potters to meet con 191-218. Universidad Nacional Aut?noma de M?xico,
placed
sumer demand also played a part in the apparent M?xico, D.F.
Arnold, Dean E.
standardization of goods (Sheehy 1992). There is 1978 The Ethnography of Pottery Making in the Valley of
some difficulty in comparing archaeological and Guatemala. In The Ceramics ofKaminaljuyu, Guatemala,

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
Sullivan] SPECIALIZED PRODUCTION OF SAN MART?N ORANGE WARE 49

edited by R. E. Wetherington, pp. 327^00. Pennsylvania 195-220. Institute for Mesoamerican Studies, Albany.
State University, University Park. Charlton, Cynthia Otis, Thomas H. Charlton, and Deborah L.
1985 Ceramic Theory and Cultural Process. Cambridge Nichols
University Press, Cambridge. 1993 Aztec Household-Based Craft Production: Archaeo
1993 Ecology and Ceramic Production in an Andean Com logical Evidence from the City-State of Otumba, Mexico.
munity. Cambridge University Press, Cambridge. InHousehold, Compound, and Residence: Studies ofPre
Arnold, Dean E., and Alvaro L. Nieves hispanic Domestic Units inWestern Mesoamerica, edited
1992 Factors Affecting Ceramic Standardization. In by Robert S. Santley and Kenneth G. Hirth, pp. 147-171.
Ceramic Production and Distribution: An Integrated CRC Press, Boca Raton.
Approach, edited by George J. Bey III and Christopher A. Charlton, Thomas H., Deborah L. Nichols, and Cynthia Otis
Pool, pp. 93-113. Westview Press, Boulder. Charlton
Arnold, Philip J., Ill 1991 Aztec Craft Production and Specialization: Archaeo
1990 The Organization of Refuse Disposal and Ceramic logical Evidence from the City-State of Otumba, Mexico.
Production within Contemporary Mexican Houselots. World Archaeology 23:98-114.
American Anthropologist 92:915-932. 2000 Otumba and Its Neighbors: Ex oriente lux. Ancient
1991a Dimensional Standardization and Production Scale Mesoamerica 11:247-265.
in Mesoamerican Ceramics. Latin American Antiquity Costin, Cathy L.
2:363-370. 1991 Craft Specialization: Issues in Defining, Document
1991 b Domestic Ceramic Production and Spatial Organi ing, and Explaining the Organization of Production. In
zation: A Mexican Case Study inEthnoarchaeology. Cam Archaeological Method and Theory, Vol. 3, edited by
bridge University Press, Cambridge. Michael B. Schiffer, pp. 1-56. University of Arizona Press,
Arnold, Philip J., Ill, Christopher A. Pool, Ronald R. Kneebone, Tucson.
and Robert S. Santley 2001 Craft Production Systems. InArchaeology at theMil
1993 Intensive Ceramic Production and Classic-Period lennium: A Sourcebook, edited by Gary M. Feinman and
Political Economy in the Sierra de los Tuxtlas, Veracruz, T. Douglas Price, pp. 273-327. Academic Press, New York.
Mexico. Ancient Mesoamerica 4:175-191. Costin, Cathy L., and Melissa B. Hagstrum
Arnold, Philip J., Ill, and Robert S. Santley 1995 Standardization, Labor Investment, Skill, and the
1993 Household Ceramics Production at Middle Classic Organization of Ceramic Production in Late Prehispanic
Period Matacapan. InPrehispanic Domestic Units inWest Highland Peru. American Antiquity 60:619-639.
ern Mesoamerica, edited by Robert S. Santley and Ken Cowgill, George L.
neth G. Hirth, pp. 227-248. CRC Press, Boca Raton. 1987 M?todos para el estudio de relaciones espaciales en
Balkansky, Andrew K., Gary M. Feinman, and Linda M. los datos de la superficie de Teotihuac?n. In Teotihuac?n,
Nicholas nuevos datos, nuevas s?ntesis, nuevos problemas, edited
1997 Pottery Kilns of Ancient Ejutla, Oaxaca, Mexico. by Emily McClung de Tapia and Evelyn C. Rattray, pp.
Journal of Field Archaeology 24:139-160. 161-190. Universidad Nacional Aut?noma de M?xico,
Barnes, Gina M?xico, D.F.
1987 The Role of the Be in the Formation of the Yamato 1992a Social Differentiation at Teotihuac?n. InMesoamer
State. In Specialization, Exchange and Complex Societies, ican Elites: An Archaeological Assessment, edited by
edited by Elizabeth M. Brumfiel and Timothy K. Earle, Diane Z. Chase and Arien F. Chase, pp. 206-220. Uni
pp. 86-101. Cambridge University Press, Cambridge. versity of Oklahoma Press, Norman.
Blackman, M. James, Gil J. Stein, and Pamela B. Vandiver 1992b Toward a Political History of Teotihuac?n. In Ideol
1993 The Standardization Hypothesis and Ceramic Mass ogy and Pre-Columbian Civilizations, edited by Arthur A.
Production: Technological, Compositional, and Metric Demarest and Geoffrey W. Conrad, pp. 87-114. School of
Indexes of Craft Specialization at Tell Leilan, Syria. Amer American Research Press, Santa Fe.
ican Antiquity 58:60-80. 1997 State and Society at Teotihuac?n, Mexico. Annual
Bordaz, Jacques Review of Anthropology 26:129-161.
1964 Pre-Columbian Ceramic Kilns at Pe?itas, a Post Curet, L. Antonio
Classic Site in Coastal Nayarit, Mexico. Ph.D. disserta 1993 Regional Studies and Ceramic Production Areas: An
tion, Columbia University. University Microfilms, Ann Example from La Mixtequilla, Veracruz, Mexico. Journal
Arbor. of Field Archaeology 20:427^140.
Brumfiel, Elizabeth M., and Timothy K. Earle Deal, Michael
1987 Specialization, Exchange, and Complex Societies: 1988 An Ethnoarchaeological Approach to the Identifica
An Introduction. In Specialization, Exchange, and Com tion of Maya Domestic Pottery Production. In Ceramic
plex Societies, edited by Elizabeth M. Brumfiel and Tim Ecology Revisited, 1987: The Technology and Socioeco
othy K. Earle, pp. 1-9. Cambridge University Press, nomics of Pottery, edited by Charles C. Kolb, pp. 111-142.
Cambridge. BAR International Series 436(ii). Oxford.
Cabrera Castro, Rub?n Earle, Timothy K.
1988 Horno cer?mico posteotihuacano en el palacio de 1981 Comment on P. Rice, Evolution of Specialized Pot
Atetelco. Arqueolog?a 4:41-15. tery Production: A Trial Model. Current Anthropology
Castanzo, Ronald A., and J. Heath Anderson 22:230-231.
2004 Formative Period Lime Kilns in Puebla, Mexico. Mex Feinman, Gary M.
icon 26(4):86-90. 1982 Patterns of Ceramic Production and Distribution, Peri
Charlton, Cynthia Otis ods Early I through V InMonte Alban's Hinterland, Part
1994 Plebians and Patricians: Contrasting Patterns of Pro 1: The Prehispanic Settlement Patterns of the Central and
duction and Distribution in the Aztec Figurine and Lap Southern Parts of the Valley of Oaxaca, Mexico, edited by
idary Industries. In Economies and Polities in the Aztec Richard E. Blanton, Stephen A. Kowalewski, Gary M.
Realm, edited by Michael Smith and Mary Hodge, pp. Feinman, and Jill Appel, pp. 181-206. Memoirs of the

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
 50LATIN AMERICAN ANTIQUITY [Vol. 17, No. 1, 2006

Museum of Anthropology. University of Michigan, Ann Hopkins, Mary R.

Arbor. 1995 Teotihuacan Cooking Pots: Scale of Production and
1985 Changes in the Organization of Ceramic Production Product Variability. Ph.D. dissertation, Brandeis Univer
in Pre-Hispanic Oaxaca, Mexico. InDecoding Prehistoric sity. University Microfilms, Ann Arbor.
Ceramics, edited by Ben A. Nelson, pp. 195-223. South Kirby,KrisN.
ern Illinois University Press, Carbondale. 1993 Advanced Data Analysis with SYSTAT. Van Nostrand
Feinman, Gary M., Stephen A. Kowalewski, and Richard Blan Reinhold, New York.
ton Kowalewski, Stephen A., Gary M. Feinman, Laura Finsten,
1984 Modeling Ceramic Production and Organizational Richard E. Blanton, and Linda M. Nicholas
1989 Monte Alb?n's Hinterland, Part 2: Prehispanic Set
Change in the Pre-Hispanic Valley of Oaxaca, Mexico. In
tlement Patterns in Tlacolula, Etla, and Ocotlan, the Val
The Many Dimensions of Pottery, edited by Sander van
der Leeuw and Alison C. Pritchard, pp. 297-337. Univer ley ofOaxaca, Mexico, Vol. 1.Museum of Anthropology,
voor Memoirs No. 23. University of Michigan, Ann Arbor.
sity of Amsterdam, Albert Egges van Giffen Instituut
Prae-en Protohistorie, Cingula VII, Amsterdam. Krotser, Paula H.
Flannery, Kent V, and Joyce Marcus 1987 Levels of Specialization among Potters of Teotihua
1983 Urban Mitla and Its Rural Hinterland. In The Cloud can. In Teotihuacan, nuevos datos, nuevas s?ntesis, nuevos
edited by Emily McClung de Tapia and Eve
People: Divergent Evolution of the Zapotee and Mixtee problemas,
Civilizations, edited by Kent V. Flannery and Joyce Mar lyn C. Rattray, pp. 417-427. Universidad Nacional
cus, pp. 295-300. Academic Press, New York. Aut?noma de M?xico, M?xico, D.F.
Graves, Michael W. Krotser, Paula H., and Evelyn C. Rattray
1991 Pottery Production and Distribution among the 1980 Manufactura y distribuci?n de tres grupos cer?micos
and Regional Organiza de Teotihuacan. Anales de Antropolog?a
Kalinga: A Study of Household 17(1):91-104.
tion and Differentiation. In Ceramic Ethnoarchaeology, Lacovara, Peter
edited by William A. Longacre, pp. 112-143. University 1985 The Ethnoarchaeology of Pottery Production in an
of Arizona Press, Tucson. In Ancient Technology toMod
Upper Egyptian Village.
ern Science, edited by William D. Kingery, pp. 51-60.
Hagstrum, Melissa B.
1985 Measuring Prehistoric Ceramic Craft Specialization: American Ceramic Society, Columbus, Ohio.
A Test Case in the American Southwest. Journal of Field Longacre, William A., Kenneth L. Kvamme, and Masashi

Archaeology 12:65-76. Kobayashi

1986 The Technology of Ceramic Production ofWanka and 1988 Southwestern Pottery Standardization: An Ethnoar
InkaWares from the Yanamarca Valley, Peru. In Ceramic chaeological View from the Philippines. Kiva 53:101-112.
Notes, edited by Prudence M. Rice, pp. 1-29. Occasional Longacre, William A., and Miriam T. Stark
Publications of the Ceramic Technology Laboratory, No. 1992 Ceramics, Kinship, and Space: A Kalinga Example.
3. Florida State Museum, Gainesville. Journal of Anthropological Archaeology 11:125-136.
1988 Ceramic Production in the Central Andes, Peru: An Manzanilla, Linda

Archaeological and Ethnographic Comparison. In A Pot 1992 The Economic Organization of the Teotihuacan Priest
hood: Hypotheses and Considerations. In Art, Ideology,
for All Reasons, edited by Charles C. Kolb and Louana M.
Tem and the City of Teotihuacan, edited by Janet C. Berlo, pp.
Lackey, pp. 127-145. Laboratory of Anthropology,
321-338. Dumbarton Oaks, Washington, D.C.
ple University, Philadelphia.
Healan, Dan M. Mill?n, Ren?
1986 Technological and Nontechnological of an 1973 Urbanization at Teotihuacan, Mexico, Vol. 1, Part 1:
Aspects
Obsidian Workshop Excavated at Tula, Hidalgo. In Text. University of Texas Press, Austin.
1981 Teotihuacan: City, State, and Civilization. In The
Research inEconomic Anthropology Supplement 2: Eco
nomic Aspects of Prehistoric Highland Mexico, edited by Handbook of Middle American Indians Supplement 1:
edited by Jeremy A. Sabloff, pp. 198-243.
Barry L. Isaac, pp. 133-152. JAI Press, Greenwich, Con Archaeology,
necticut. University of Texas Press, Austin.
R. Allison 1993 The Place Where Time Began: An Archaeologist's
Hegmon, Michelle, Winston Hurst, and James
1995 Production for Local Consumption and Exchange. In Interpretation of What Happened in Teotihuacan History.
Ceramic Production in theAmerican Southwest, edited by In Teotihuacan: Art from the City of the Gods, edited by
Barbara J.Mills and Patricia L. Crown, pp. 30-62. Uni Kathleen Benin and Esther Pasztory, pp. 16-43. Thames
and Hudson, New York.
versity of Arizona Press, Tucson.
Hernandez, Carlos, Robert H. Cobean, Alba Guadalupe Mas Mill?n, Ren?, R. Bruce Drewitt, and George L. Cowgill
1973 Urbanization at Teotihuacan, Mexico, Vol. 1, Part II:
tache, and Mar?a E. Su?rez
1999 Un taller de alfareros en la antigua ciudad de Tula. Maps. University of Texas Press, Austin.

Arqueolog?a 22:69-88. Mohr Chavez, Karen L.

1992 The Organization of Production and Distribution of
Hirth, Kenneth G.
1996 Political Economy and Archaeology: Perspectives on Traditional Pottery in South Highland Peru. In Ceramic
and Production. Journal of Archaeological Production and Distribution: An Integrated Approach,
Exchange
Research 4:203-239. edited by George J. Bey III and Christopher A. Pool, pp.
2000 Archaeological Research at Xochicalco, Vol. I: 49-92. Westview Press, Boulder.
Ancient Urbanism at Xochicalco: The Evolution and Orga Nichols, Deborah L.
nization of a Pre-Hispanic Society. University of Utah 1994 The Organization of Provincial Craft Production and
Press, Salt Lake City. the Aztec City-State of Otumba. In Economies and Poli
ties in the Aztec Realm, edited by Mary G. Hodge and
Hodge, Mary G., and Leah D. Mine
1990 The Spatial Patterning of Aztec Ceramics: Implica Michael E. Smith, pp. 175-193. Institute for Mesoameri
can Studies, State University of New York, Albany.
tions for Prehispanic Exchange Systems in the Valley of
17:415-437. Deborah L., Elizabeth M. Brumfiel, Hector Neff, Mary
Mexico. Journal of Field Archaeology Nichols,

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
Sullivan] SPECIALIZED PRODUCTION OF SAN MART?N ORANGE WARE 51

Hodge, Thomas H. Charlton, and Michael D. Glascock 257-279. University of Colorado Press, Niwot.
2002 Neutrons, Markets, Cities, and Empires: A1,000-Year 1996 The Archaeology of Wine: The Wine and Brandy
Perspective on Ceramic Production and Distribution in the Haciendas of Moquegua, Peru. Journal of Field Archae
Postclassic Basin of Mexico. Journal of Anthropological ology 23:187-204.
Archaeology 21:25-82. Robertson, Ian G.
Ohnersorgen, Michael A. 1999 Spatial andMultivariate Analysis, Random Sampling
2001 Social and Economic Organization of Cotaxtla in the Error, and Analytical Noise: Empirical Bayesian Methods
Post Classic Gulf Lowlands. Ph.D. dissertation, Depart at Teotihuacan, Mexico. American Antiquity 64:137-152.
ment of Anthropology, Arizona State University, Tempe. 2001 Mapping the Social Landscape of an Early Urban
Peacock, David P. S. Center: Socio-Spatial Variation in Teotihuacan. Ph.D. dis
1982 Pottery in the Roman World: An Ethnoarchaeologi sertation, Department of Anthropology, Arizona State Uni
cal Approach. Longman, New York. versity, Tempe.
Plog, Stephen Rye, Owen S.
1985 Estimating Vessel Orifice Diameters: Measurement 1981 Pottery Technology. Manual inArchaeology 4. Tarax
Methods and Measurement Error. In Decoding Prehis acum Press, Washington, D.C.
toric Ceramics, edited by Ben A. Nelson, pp. 243-253. Sanders, William T., Rebecca Storey, and Randolph J.Widmer
Southern Illinois University Press, Carbondale. 1982 Tlajinga 33, an Apartment Compound of the Pre
Pool, Christopher A. Columbian City of Teotihuacan. Technical report to the
1990 Ceramic Production, Distribution, and Resource Pro National Science Foundation, Washington, D.C.
curement at Matacapan, Veracruz, Mexico. Ph.D. disser Santley, Robert S., and Philip J. Arnold III
tation, Tulane University. University Microfilms, Ann 1996 Prehispanic Settlement Patterns in the Tuxtla Moun
Arbor. tains, Southern Veracruz, Mexico. Journal of Field Archae
1992 Integrating Ceramic Production and Distribution. In ology 23:225-259.
Ceramic Production and Distribution: An Integrated Santley, Robert S., Philip J.Arnold III, and Christopher A. Pool
Approach, edited by George J. Bey III and Christopher A. 1989 The Ceramics Production System atMatacapan, Ver
Pool, pp. 275-313. Westview Press, Boulder. acruz, Mexico. Journal of Field Archaeology 16:107-132.
Pool, Christopher A., and Robert S. Santley Sheehy, James J.
1992 Middle Classic Pottery Economics in the Tuxtlas 1992 Ceramic Production inAncient Teotihuacan, Mexico:
Mountains, Southern Veracruz, Mexico. In Ceramic Pro A Case Study of Tlajinga 33. Ph.D. dissertation, Pennsyl
duction and Distribution: An Integrated Approach, edited vania State University, University Park. University Micro
by George J. Bey III and Christopher A. Pool, pp. 205-234. films, Ann Arbor.
Westview Press, Boulder. 1998 Chronological Trends in the Ceramics of Tlaj inga 3 3.
Rattray, Evelyn C. In Los ritmos de cambio en Teotihuacan: Reflexiones y dis
1988 Un taller de cer?mica anaranjado San Mart?n en Teoti cusiones de su cronolog?a, edited by Rosa Brambila and
huac?n. In Ensayos de alfarer?a prehisp?nica e hist?rica Rub?n Cabrera Castro, pp. 299-316. Serie Arqueolog?a.
de Mesoamerica: Homenaje a Eduardo Noguera, edited Instituto Nacional de Antropolog?a e Historia, M?xico,

by Mari Carmen Serra Puche and Carlos Navarrete, pp. D.F.

249-266. Instituto de Investigaciones Antropol?gicas, Shennan, Stephen
Serie Antropol?gica 82. Universidad Nacional Aut?noma 1992 Quantifying Archaeology. 2nd ed. Edinburgh Uni
de M?xico, M?xico, D.F. versity Press, Edinburgh.
2001 Teotihuac?n: Cer?mica, cronolog?a, y tendencias cul Shepard, Anna O.
turales/Ceramics, Chronology, and Cultural Trends. Serie 1985 [1956] Ceramics for theArchaeologist. Carnegie Insti
Arqueolog?a de M?xico. Instituto Nacional de tution of Washington, Washington, D.C. (originally pub
Antropolog?a, M?xico, D.F.; and University of Pittsburgh, lished as Publication 609).
Pittsburgh. Sheskin, David
Redmond, Elsa M. 2000 Handbook of Parametric and Nonparametric Statis
1979 A Terminal Formative Ceramic Workshop in the tical Procedures. 2nd ed. Chapman and Hall/CRC, Boca
Tehuac?n Valley. In Prehistoric Social, Political, and Eco Raton.
nomic Development in the Area of the Tehuac?n Valley: Sinopoli, Carla M.
Some Results of the Palo Blanco Project, edited by Robert 1988 The Organization of Craft Production at Vijayana
D. Drennan, pp. 111-125. Technical Reports, Museum of gara, South India. American Anthropologist 90:580-597.
Anthropology. University of Michigan, Ann Arbor. 1991 Approaches to Archaeological Ceramics. Plenum
Reina, Rub?n E., and Robert M. Hill Press, New York.
1978 The Traditional Pottery of Guatemala. University of Sload, Rebecca S.
Texas Press, Austin. 1982 A Study of Status and Function in the Xolalpan
Rice, Prudence M. Metepec Community in Teotihuacan, Mexico. Ph.D. dis
1981 Evolution of Specialized Pottery Production: A Trial sertation, Brandeis University. University Microfilms, Ann
Model. Current Anthropology 22:219-240. Arbor.
1987 Pottery Analysis: A Sourcebook. University of Chicago Sotomayor Caste?ada, Alfredo
Press, Chicago. 1966 Estudio petrogr?fico y mineral?gico de cer?mica. In
1989 Ceramic Diversity, Production, and Use. In Quanti La cer?mica del centro ceremonial de Teotihuacan, appen
fying Diversity in Archaeology, edited by Robert D. dix 1. Instituto Nacional de Antropolog?a eHistoria, Depar
Leonard and George T. Jones, pp. 109-117. Cambridge tamento de Prehistoria, M?xico, D.F.
University Press, Cambridge. Spence, Michael W
1991 Specialization, Standardization, and Diversity: A Ret 1981 Obsidian Production and the State in Teotihuacan.
rospective. In The Ceramic Legacy of Anna O. Shepard, American Antiquity 46:769-788.
edited by Ronald Bishop and Frederick W. Lange, pp. Stark, Barbara L.

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
 52LATIN AMERICAN ANTIQUITY [Vol. 17, No. 1, 2006

1984 An Ethnoarchaeological Study of aMexican Pottery Bolet?n del Instituto Nacional de Antropolog?a eHistoria,
Industry. Journal of New World Archaeology 6:4-14. pp. 37^0. Instituto Nacional de Antropolog?a e Historia,
1985 Archaeological Identification of Pottery Production M?xico, D.F.
Locations: Ethnoarchaeological and Archaeological Data
inMesoamerica. InDecoding Prehistoric Ceramics, edited
by Ben A. Nelson, pp. 158-194. Southern Illinois Uni Notes
versity Press, Carbondale.
1992 Ceramic Production in Prehistoric La Mixtequilla, 1. Except in cases where artifacts are sealed under plas
South-Central Veracruz, Mexico. In Ceramic Production tered floors, admixture of sherds from earlier phases occurs in
and Distribution: An Integrated Approach, edited by excavated collections because sherds, figurine fragments, and
George J. Bey III and Christopher A. Pool, pp. 175-204. other artifacts were used as fill in the construction of more
Westview Press, Boulder.
recent architecture at Teotihuacan (Sheehy 1998:308).
1995 Problems in Analysis of Standardization and Spe
2. Krotser
and Rattray (1980:96) previously examined 32
cialization in Pottery. In Ceramic Production in theAmer
ican Southwest, edited by Barbara J.Mills and Patricia L. of the 181 collections from the Tlajinga district for evidence
Crown, pp. 231-267. University of Arizona Press, Tucson. of ceramic production, although other collections may have
Stark, Barbara L., and Christopher P. Garraty been examined at a later date as Krotser (1987:420) mentions
2004 Evaluation of Systematic Surface Evidence for Pot
having examined 40 collections from S3W1, S3W2, S4W1,
tery Production in Veracruz, Mexico. Latin American and S4W2. Evidence for ceramic production in these studies
Antiquity 15:123-143. included defective and burnt sherds as well as the presence of
Stevens, James
ceramic tools
(Rattray 1988:Figures 3-5). They examined
1986 Applied Multivariate Statistics for the Social Sciences.
one areain the Tlajinga neighborhood with 20.6 percent
Lawrence Erlbaum Associates, Hillsdale, N.J.
wasters per SMO sherds (106 SMO wasters of 515 total SMO
Stone, Doris, and Conchita Turnbull
1941 A Sulu-Ul?a Pottery Kiln. American Antiquity sherds), although the exact collection tract is not identified
1:39-47. and I have been unable to match the total SMO sherds with

Storey, Rebecca totals recorded in the TMP database for any of the collection
1987 A First Look at the Paleodemography of the Ancient tracts in the Tlajinga (Krotser and Rattray 1980:95).
district
City of Teotihuacan. In Teotihuacan, nuevos datos, nuevas 3. Rattray (2001:299) argues that SMO basins are diag
s?ntesis, nuevos problemas, edited by Emily McClung de nostic of the Metepec phase; however, 51 sherds collected by
Tapia and Evelyn C. Rattray, pp. 91-114. Universidad the TMP are classified as Xolalpan-phase SMO basins (50
Nacional Aut?noma de M?xico, M?xico, D.F.
percent of which derive from Tlajinga district collection
1992 Life and Death in the Ancient City of Teotihuacan: A
tracts). Basedon excavated materials from Tlajinga 33,
Modern Paleodemographic Synthesis. University of
Alabama Press, Tuscaloosa. Sheehy argues that basins occur at the workshop by the Early
Sullivan, Kristin S. Tlamimilolpa phase, with frequencies increasing slightly in
2002 Specialized Production of San Martin Orange Ceram the Late Tlamimilolpa and Xolalpan phases before "they vir
ics in the Tlajinga District of Classic Period Teotihuacan, tually disappear the Metepec (1998:311).
during phase"
Mexico. Master's publishable paper, Department of also notes typological and chronological
Hopkins (1995:646)
Anthropology, Arizona State University, Tempe. in Metepec-phase Burnished and SMO craters. I
ambiguity
van de Velde, Paul, and Henriette R. van de Velde
consider Xolalpan and Metepec as a single chronological
1939 The Black Pottery of Coyotepec, Oaxaca, Mexico.
phase for which sherds can be more securely assigned than
Southwest Museum Papers 13:7?43.
Whalen, Michael E. they can to either phase individually.
at Santo Domingo 4. These collection tracts are located on the Eastern
1981 Excavations Tomaltepec: Evolu
tion of a Formative Community in the Valley of Oaxaca, Platform the Tlajinga district from the Street of
that delimits
Mexico. Museum of Anthropology Memoirs 12. Univer the Dead. 7B:S3W1 is a small two-stage temple; 7A, 7C, and
sity of Michigan, Ann Arbor. 7D are small single-stage platforms; 7E and 7G are possible
Widmer, Randolph J. small room complexes with unclear structural limits; and 7F
1987 The Evolution of Form and Function in a Teotihua is an open area (e.g., no clear structural indications visible on
can Apartment Compound: The Case of Tlajinga 33. In
the surface) just north of 7A and 7B (Mill?n et al. 1973:114).
Teotihuacan, nuevos datos, nuevas s?ntesis, nuevos prob
5. Access to raw materials is important to pottery produc
lemas, edited by Emily McClung de Tapia and Evelyn C.
tion and may be used in conjunction with other indicators to
Rattray, pp. 317-368. Universidad Nacional Aut?noma de
support the
identification of likely production areas.
M?xico, M?xico, D.F.
1991 Lapidary Craft Specialization at Teotihuacan: However, in this case I do not compare this variable among
Impli
cations for Community Structure at 33:S3W1 and Eco apartment compounds because potters living and working in
nomic Organization in the City. Ancient Mesoamerica the apartment compounds in the Tlajinga district would have
2:131-147. had access to raw materials like water, clay, and temper avail
Widmer, Randolph J., and Rebecca Storey able from the nearby R?o San Lorenzo and Barranca
1993 Social Organization and Household Structure of a
Patlachique.
Teotihuacan Apartment Compound: S3WL33 of the Tla
6. Krotser also identifies 24:S3W2 as a possible SMO
jinga Barrio. In Prehispanic Domestic Units inWestern
workshop and argues that "seven sites in Tlajinga had in their
Mesoamerica, edited by Robert S. Santley and Kenneth
G. Hirth, pp. 87-104. CRC Press, Ann Arbor. collections many burned sherds indicative of firing floors, the
Winter, Marcus C, andWilliam O. Payne most outstanding in 20:S4W2 in a large open space"
1976 Hornos para cer?mica hallados en Monte Alb?n. In (1987:421). Although I have identified an unusual proportion

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions
Sullivan] SPECIALIZED PRODUCTION OF SAN MARTIN ORANGE WARE 53

of SMO wasters (14.29 percent or two of 14 SMO sherds) in shows a relatively low p value (p = .15), I consider this only
from 20:S4W2,1
the collection do not consider this a possi weak evidence of the differences being sought here.
ble SMO workshop because the collection includes only 14 Excluding craters, all other p values are large enough that
SMO sherds of 29 total Xolalpan and Metepec sherds. they may be the result of random chance rather than signifi
7. Sheehy (1992:548, 557, 563-565) infers eight vessel cant differences inmean rim diameter.
size classes for craters, three size classes of amphorae, and 11. Post hoc, or a posteriori, tests involve comparisons that
five size classes for basins. In her study of TMP collections are determined after data collection (i.e., I could not have
from throughout the city, Hopkins (1995:Figures 8.55 and determined a priori which Tlajinga collections might be
8.71) found that Xolalpan SMO craters have a multimodal defined as likely workshops, so the decision as to the particu
distribution, whereas Metepec SMO craters have a bimodal lar collections compared had to occur after the data had been
distribution with peaks at 30 and 40 cm (concerning distinc collected). Post hoc tests include Fisher's least-significant
tions in the size classes she and Sheehy identify in SMO difference test, the Bonferroni adjustment, the Tukey highly
craters, see Hopkins 1995:643). significant-difference multiple comparisons, and the Scheff?
8. Medium-large basin collections from 6 and 2LS3W2 test (Sheskin 2000:528-535, 537-540). Fisher's least
also have small interquartile ranges but only include three and significant-difference test conducts multiple t tests between all
two rims, respectively. of the pairs of groups (i.e., all possible workshops) but does
9. Amphorae have been assessed with ANOVA, although not include an adjustment for conducting multiple compar
the nature of these vessels likely constrained the acceptable isons. The Bonferroni adjustment is Fisher's test with a cor
size of rim diameters because large vessels would be difficult rection of the observed significance level. The Tukey test
to use and extremely small vessels would not be very useful. allows for all possible pair-wise comparisons in the data set
Three extreme outliers, two from workshops and one from a and is preferred over Bonferroni for comparison of a large

nonworkshop, were removed from the amphora assemblages number of pairs of means. The Scheff? test is the most con
in examining the distribution of amphora rim diameters. The servative of the four and accommodates unequal sample sizes.
outliers include one with a rim diameter of 30 cm from
48:S3W1 and two with rim diameters of 47 and 53 cm from
24:S3W2. Submitted August 23, 2004; Accepted May 5, 2005; Revised
10. Although ANOVA of amphorae from all workshops October 3, 2005.

This content downloaded from 129.81.226.149 on Tue, 23 Apr 2013 23:36:55 PM

All use subject to JSTOR Terms and Conditions