Radiochim. Acta 97, 513–518 (2009) / DOI 10.1524/ract.2009.1643
© by Oldenbourg Wissenschaftsverlag, München
Characterization of pottery from Cerro de Las Ventanas,
Zacatecas, Me´xico
By H. Lo´pez-del-Rı´o1 ,∗, F. Mireles-Garcı´a1 , R. Y. Me´ndez-Cardona2, M. Nicola´s-Caretta3 ,4, R. J. Speakman5 and
M. D. Glascock6
1 Unidad Acade´mica de Estudios Nucleares, UAZ, Apdo. Postal 579C, 98068 Zacatecas, Me´xico
2 Unidad Acade´mica de Antropologı´a, UAZ, Apdo Postal 555 Suc C, Zacatecas, Me´xico
3 INAH Delegacio´n Zacatecas, Me´xico
4 Coordinacio´n de Ciencias Sociales y Humanidades, UASLP, Av. Industrias 101-A, Fracc. Talleres 78494 SLP, Me´xico
5 Museum Conservation Institute, Smithsonian Institution, Suitland, MD 20746, USA
6 Research Reactor Center, University of Missouri, Columbia, MO 65211, USA
(Received February 24, 2009; accepted March 4, 2009)
Cerro de las Ventanas / Mexico / Pottery /
Elemental analysis / INAA
Summary. With the aim of classifying prehispanic pot-
tery from Cerro de Las Ventanas site, Juchipila, Zacatecas,
Me´xico, instrumental neutron activation analysis (INAA) was
used to analyze ceramic samples at the University of Missouri
Research Reactor Center. Thirty-two chemical elements were
measured: Al, As, Ba, Ca, Ce, Co, Cr, Cs, Dy, Eu, Fe, Hf, K,
La, Lu, Mn, Na, Nd, Rb, Sb, Sc, Sm, Sr, Ta, Tb, Ti, Th, U, V,
Yb, Zn, and Zr. Two multivariate statistical methods, cluster
analysis and principal component analysis, were performed
on the dataset to examine similarities between samples and
to establish compositional groups. The statistical analyses of
the dataset suggest that the pottery samples form a unique
chemically homogeneous group, with the exception of one
pottery sample. The compositional data were compared to an
existing Mesoamerican ceramic database. It was found that
the newly generated data fit best with data from a previous
chemical analysis of pottery from the Malpaso Valley. How-
ever, despite the apparent similarity, pottery samples from the
site of Cerro de Las Ventanas represent a new and unique
chemical fingerprint in the region.
1. Introduction
Pottery is often the most common artifactual material re-
covered during archaeological field research. It can pro-
vide archaeologists with valuable information about the age
of a culture, length of site occupation, social organization,
technological phase, economic system, and cultural contact
across a region [1]. In pottery analysis, archeologists are
most commonly interested in stylistic variation through time
and space, in the technological level involved in ceramic
manufacture, and in the function and use of artifacts. With
these purposes in mind, a classification is made on the ba-
sis of shared stylistic and functional attributes (e.g., typolo-
gies) [2, 3]. However, visual examination of artifacts does
*Author for correspondence (E-mail: hlopez@uaz.edu.mx).
not always provide definitive information with respect to the
study of exchange networks or pottery provenance [4–6].
Chemical and mineralogical characterization of pottery
provides valuable quantitative analytical data useful for in-
ferring the location and techniques involved in its manufac-
ture. Such studies include mineralogical composition, chem-
ical composition, microstructure, and surface traits [1]. In
particular, chemical analyses of pottery may be used to trace
the source of raw material (provenance studies) by matching
elemental abundances of wares with those of clays. Com-
positional analyses identify the chemical elements of the
paste that represent a unique geochemical fingerprint [6].
According to the “provenance postulate” [7] the variation in
chemical composition of the raw materials must be greater
between two spatially separated regions than the variation
within a particular region. Additionally, chemical analyses
may be useful for investigating geographic displacements
among spatially separated regions by comparing ceramic
elemental patterns [8].
Compositional datasets typically are examined by apply-
ing pattern-recognition multivariate statistical methods in
order to group ceramic samples according to their similari-
ties. Thus, a new chemical-based classification of the pottery
is obtained, and homogeneous groups are formed [9]. Com-
positional groups act as a new pseudo-typology that can be
used to compare different ceramic types.
Currently, many analytical techniques can be applied to
determine elemental composition of pottery: X-ray fluores-
cence (XRF) [10], instrumental neutron activation analysis
(INAA) [11], proton induced X-ray emission (PIXE) [12],
laser ablation-inductively coupled plasma-mass spectrom-
etry (LA-ICP-MS) [13], and electron probe microanalysis
(EPMA) [14]. However, INAA has been and continues to
be used by many researchers to study pottery. It allows
simultaneous determination of a large number of elements
with high accuracy, precision and sensitivity. It is sensitive
enough to measure concentrations in the µg g−1 range or be-
low and sample preparation is relatively easy and fast.
In this work, we report the results obtained from chem-
ical analysis of 15 ceramic potsherds from the Mexican ar-
514 H. Lo´pez-del-Rı´o et al.
Fig. 1. Map showing the location of Cerro de Las Ventanas site.
chaeological site of Cerro de Las Ventanas. The site was an
important Caxcan settlement, and archaeologists believe that
it was included in a great exchange network extending from
northern Zacatecas to Teotihuacan City in central Me´xico,
during the Classic period [15]. The research performed at
this site will allow archaeologists to attain a greater un-
derstanding of the interactions between prehispanic social
groups in northwestern Mexico.
1.1 Archaeological background
The Cerro de Las Ventanas site is located in the southern part
of the state of Zacatecas, Me´xico, in Juchipila town, cov-
ering an area of approximately 119 ha (Fig. 1). It is about
175 km away from the site of La Quemada, which is one of
the most important prehispanic settlements of northwestern
Mesoamerica. Cerro de Las Ventanas, along with La Que-
mada and other major settlements from Los Altos, Jalisco,
and northern Guanajuato, were all connected to Teotihuacan
through a large exchange network [15].
The most important archaeological structures are located
on Las Ventanas hill (because of this it was called Cerro de
Las Ventanas), but a portion of the site is encompassed by
Cerro Chihuahua, Cerro Pico de Pecho, and Cerro Pico de
Aguila. According to 14C dating, there was a long period
of occupation by the Caxcan culture from 20 ∼ 70 AD until
1400 AD [16]. The archaeological record consists of ceram-
ics, lithics, and milling materials. The main monumental
structures lie on Cerro de Las Ventanas, which encompass
a pyramid-altar-square complex (i.e., ceremonial center) and
several residential houses including a cliff-house. Terraces
and residential houses also are present on the piedmont and
in the valley.
Although descriptions of the archaeological site have ap-
peared in several documents from the early 20th century, the
first academic research was carried out in the late 1980s and
1990s. Unfortunately, research at the site was interrupted for
several years, but ultimately a more systematic INAH (Insti-
tuto Nacional de Antropologia e Historia) Project was initi-
ated at Las Ventanas under the direction of Nicola´s-Caretta
and Jime´nez in 2002. The materials analyzed for this paper
were recovered during the excavation phase of this project in
2003.
2. Experimental
2.1 Pottery samples
Fifteen ceramic sherds were analyzed. The samples reported
here were collected from Terrace No. 1 during Nicola´s-
Caretta and Jimenez-Bets’s excavation [17], and they rep-
resent a ceramic occupation ranging from 30 to 280 cm in
depth. One ceramic fragment (LVZ003) was collected on the
surface. The specimens were selected based on the diagnos-
tic quality they represented within the spectrum of material
collected.
2.2 Sample preparation
Standard MURR procedures for pottery analysis were ap-
plied to samples [11]. An area of ∼ 1 cm2 is removed
from potsherd, and all surface layers were removed using
Characterization of pottery from Cerro de Las Ventanas, Zacatecas, Me´xico 515
a tungsten-carbide burring tool. Next, the sample was
cleaned and washed with deionized water and allowed to dry
for several hours. A fine powder was obtained by crushing
the sample in an agate mortar. Finally, the powder was stored
in clean glass vial and dried at 100 ◦C for at least 24 h.
2.3 Neutron activation analysis
Approximately 150 mg of ceramic powder was weighed into
a clean polyethylene vial used for short irradiation. A second
portion of about 200 mg powder was weighed into a high-
purity quartz vial used for long irradiation. Both aliquots
were weighed to the nearest 0.01 mg. Quality control sam-
ples and standards of Ohio Red Clay, SRM-1633a (coal fly
ash), SRM-278 (obsidian rock), and SRM-688 (basalt rock)
were similarly prepared.
Samples and standards in polyethylene vials were se-
quentially subjected to a short irradiation (5 s) in the pneu-
matic tube system with a neutron flux of 8×1013 n/(s cm2).
The samples were allowed to decay for 25 min and counted
for 720 s using a high resolution, high-purity germanium
detector (HPGe). The short irradiation yields data for nine
short-lived elements: Al, Ba, Ca, Dy, K, Mn, Na, Ti, and V.
A long irradiation of samples, standards, and quality con-
trol samples in quartz vials was carried out in the reactor
pool using a neutron flux of 5×1013 n/(s cm2) for a period
of 24 h. The radioactive samples and standards were counted
twice: a middle count of 1800 s after a 7-d decay and a fi-
nal count for a period of 10 000 s after an additional 3- or
4-week decay. The first count measures seven medium-lived
elements: As, La, Lu, Nd, Sm, U and Yb. The final count
yields sixteen long-lived elements: Ce, Co, Cr, Cs, Eu, Fe,
Hf, Rb, Sb, Sc, Sr, Ta, Tb, Th, Zn and Zr.
3. Results and discussion
Elemental concentration values for 32 elements were ob-
tained from the analysis. A statistical summary of data is
listed in Table 1. Missing values were replaced with values
calculated from a Mahalanobis distance minimization rou-
tine to enable subsequent analysis. A series of GAUSS®-
language based routines [18] were used to calculate miss-
ing values, to transform data, and to perform multivariate
statistical analysis. Given that chemical characterization of
pottery produces a dataset with high-concentration values
(major elements) and low-concentration values (trace elem-
ents), elemental concentrations were converted to base-10
logarithms to compensate for differences in magnitude be-
tween the major elements and trace elements.
As a first approach to identifying compositionally ho-
mogenous groups within the matrix of data, a hierarchical
cluster analysis was performed. Squared-mean Euclidean
distances and average linkage were used for the clustering
algorithm. From the resulting dendogram, shown in Fig. 2,
partitioning of pottery samples into subgroups is possible,
with two individual samples (3 and 7). Clearly, sample 3
is related to the archaeological context from which it was
recovered making it the poorest match to any of the other
samples.
Principal component analysis (PCA) was applied to con-
firm these assumptions. PCA, a pattern-recognition tech-
Table 1. Statistical summary of compositional data for the Cerro de
Las Ventanas pottery (in µg g−1).
Element Mean S.D. Minimum Maximum
As 59.1 76.3 13.4 304.4
La 29.0 10.9 14.8 54.2
Lu 0.587 0.217 0.236 1.169
Nd 41.3 21.5 12.8 83.7
Sm 5.99 2.32 2.36 11.3
U 3.95 1.93 1.48 9.52
Yb 4.28 1.67 1.78 8.79
Ce 71.6 25.5 29.4 110.4
Co 5.69 3.27 1.18 12.1
Cr 26.0 24.0 10.5 100.3
Cs 12.92 9.78 2.31 31.0
Eu 0.755 0.410 0.235 1.728
Fe 27 256 6505 14 935 37 019
Hf 10.82 3.45 6.32 17.0
Rb 114 34 77 180
Sb 3.45 6.89 0.51 28.0
Sc 8.41 3.79 3.93 19.73
Sr 279 117 121 652
Ta 2.60 1.17 0.79 5.17
Tb 1.07 0.45 0.35 2.23
Th 14.8 3.30 7.78 20.9
Zn 90 21 50 120
Zr 241 73 124 367
Al 89 324 12 584 71 209 115 989
Ba 1209 1356 266 5843
Ca 9301 3312 5777 18 695
Dy 6.05 2.52 1.94 12.57
K 35 758 7056 26 505 51 599
Mn 346 182 96 589
Na 7887 1884 3703 10 707
Ti 2866 1114 1243 5944
V 54.9 30.9 17.9 151.8
Table 2. Eingenvalues and percentage of variance of Principal Compo-
nents based on a variance-covariance matrix.
Principal Eigenvalue % Total % Cumulative
component variance variance
1 0.5696 38.28 38.28
2 0.4283 28.78 67.07
3 0.1498 10.06 77.14
4 0.1114 7.48 84.63
5 0.0760 5.12 93.15
nique, is a multivariate statistical procedure used to identify
patterns in data of high dimension through data compres-
sion and dimension reduction without a significant loss of
information [19]. The data are transformed into a small num-
ber of linear combinations (principal components or scores)
of the original variables and the eigenvectors based on
a variance-covariance matrix. The first few principal com-
ponent scores account for a maximal amount of variance in
the data set. It was found that the first five principal compo-
nents account for more than 90% of the total variance in the
dataset (Table 2). A bivariate plot of principal components
1 and 2 (Fig. 3) with a 90% confidence ellipse used to es-
tablish group membership, suggests that the pottery samples
form a unique chemically homogeneous group, with the ex-
ception of pottery sample 3. This is in agreement with the
cluster analysis results.
516 H. Lo´pez-del-Rı´o et al.
Fig. 2. Dendogram for 15 pottery
samples from the site of Cerro de
Las Ventanas.
Fig. 3. Bivariate plot of princi-
pal components 1 and 2 for the
Cerro de Las Ventanas pottery.
The ellipse indicates the 90%
confidence level for group mem-
bership.
Subsequently, the compositional data generated for sam-
ples from the Cerro de Las Ventanas site were compared
to the database of Mesoamerican pottery at the MURR
Archaeometric Laboratory consisting of more than 10,000
analyzed pottery samples, in order to match pottery sam-
ples with well-characterized regional pottery. We found that
the compositional data for the Cerro de Las Ventanas pot-
tery samples best matched the data generated for a previ-
ous chemical characterization research of pottery from the
Malpaso Valley carried out by Strazicich [20]. The Mal-
paso Valley regional system, a geographical area located in
the southern end of the State of Zacatecas, is located about
175 km from the site of Cerro de Las Ventanas site and is one
of the northernmost and largest of the regional systems that
form the Mesoamerican frontier [21]. The Malpaso Valley is
dominated by the extensive center of La Quemada that was
founded around 500 AD and abandoned by 900 AD [22].
Strazicich identified three chemical compositional groups
Characterization of pottery from Cerro de Las Ventanas, Zacatecas, Me´xico 517
Fig. 4. Biplot of principal compo-
nents 1 and 2 based on PCA of
the Malpaso Valley, La Quemada,
and Cerro de Las Ventanas data
sets. The ellipses indicate the 90%
confidence level for group mem-
bership.
for the region: La Quemada A, La Quemada B, and Mal-
paso Valley. The first two are assumed to represent ceramic
production at the La Quemada site; the third includes pot-
tery from several minor sites: Los Pilarillos, Puerto Nuevo,
and Presa de Ambosco. Differences between the two data
sets are illustrated in Fig. 4. Despite the apparent similari-
ties, none of the Malpaso Valley samples chemically match
the Cerro de Las Ventanas samples, which is supported by
the calculated Mahalanobis distance probabilities.
4. Conclusions
Elemental compositions of fifteen ceramic fragments from
the site of Cerro de Las Ventanas were obtained using INAA.
At the 90% confidence level for group membership, one
compositionally homogeneous group was identified by mul-
tivariate statistical analysis, and one individual sample. Ana-
lytical results were compared with previous data for regional
pottery. We found that the samples from Cerro de Las Ven-
tanas matched best with data generated for pottery from
the Malpaso Valley located in the same general region of
Me´xico. However, the pottery samples from Cerro de Las
Ventanas represents a unique new chemical fingerprint for
this region.
Acknowledgment. This work was supported in part by US Department
of Energy grant DE-FG07-02ID14380, and the University of Missouri
Research Reactor (MURR).
References
1. Rice, P. M.: Pottery Analysis – a Sourcebook. The University of
Chicago Press, Chicago (1987).
2. Rice, P. M.: Recent ceramic analysis: 1. Function, style, and ori-
gins. J. Archaeol. Res. 4, 133 (1996).
3. Hegmon, M.: Archaeological research on style. Ann. Rev. Anthro-
pol. 21, 517 (1992).
4. Rice, P. M.: Recent ceramic analysis: 2. Composition, production,
and theory. J. Archaeol. Res. 4, 165 (1996).
5. Bishop, R. L., Rands, R. L., Holley, G. R.: Ceramic compositional
analysis in archaeological perspective. In: Advances in Archaeo-
logical Method and Theory. (Schiffer, M. B., ed.) Academic Press,
New York (1982).
6. Neff, H., Bishop, R. L., Arnold, D. E.: Reconstructing ceramic
production from ceramic compositional data: An example from
Guatemala. J. Field Archaeol. 15, 339 (1988).
7. Weigand, P. C., Harbottle, G., Sayre, E. V.: Turquoise sources and
source analysis: Mesoamerica and the southwestern USA. In: Ex-
change Systems in Prehistory. (Earle, T. K., Ericson, J. E., eds.)
Academic Press, New York (1977).
8. Bishop, R. L., Blackman, M. J.: Instrumental neutron activation
analysis of archaeological ceramics: Scale and interpretation. Acc.
Chem. Res. 35, 603 (2002).
9. Glascock, M. D., Neff, H., Vaughn, K. J.: Instrumental neutron ac-
tivation analysis and multivariate statistics for pottery provenance.
Hyperf. Interact. 154, 95 (2004).
10. Pillay, E.: Analysis of archaeological artifacts: PIXE, XRF or
ICP-MS. J. Radioanal. Nucl. Chem. 247, 593 (2001).
11. Glascock, M. D.: Characterization of archaeological ceramics at
MURR by neutron activation analysis and multivariate statistics.
In: Chemical Characterization of Ceramic Pastes in Archaeology.
(Neff, H., ed.) Prehistory Press, Madison (1992).
12. Ontalba-Salamanca, M. ´A., Ruvalcaba-Sil, J. L., Bucio, L., Man-
zanilla, L., Miranda, J.: Ion beam analysis of pottery from Teoti-
huacan, Me´xico. Nucl. Instrum. Methods Phys. Res. B 161–163,
762 (2000).
13. Kennett, D. J., Sakai, S., Neff, H., Gossett, R., Larson, D. O.:
Compositional characterization of prehistoric ceramics: a new
approach. J. Archaeol. Sci. 29, 443 (2002).
14. Freestone, I. C.: Applications and potential of electron probe
micro-analysis in technological and provenance investigations of
ancient ceramics. Archaeometry 24, 99 (1982).
15. Jime´nez-Betts, P. F., Darling, J. A.: Archaeology of Southern Za-
catecas: The Malpaso, Juchipila, and Valparaiso-Bolaños Valleys.
In: Greater Mesoamerica. The Archaeology of West and North-
west Mexico. (Foster, M., Gorenstein, S., eds.) The University of
Utah Press, Utah (2000).
518 H. Lo´pez-del-Rı´o et al.
16. Mozillo, E.: Proyecto Las Ventanas. Boletı´n del Consejo de Ar-
queologı´a. Instituto Nacional de Antropologı´a e Historia, Me´xico
(1992).
17. Nicolas-Caretta, M., Jime´nez, P.: Zona arqueolo´gica Cerro de las
Ventanas-Informe te´cnico de actividades 2003. Instituto Nacional
de Antropologı´a e Historia, Me´xico (2003).
18. Neff, H.: RQ-mode principal components analysis of ceramic
compositional data. Archaeometry 36, 115 (1994).
19. Bilodeau, M., Brenner, D.: Theory of Multivariate Statistics.
Springer-Verlag, New York (1999).
20. Strazicich, N.: Prehispanic Pottery Production in the Chalchi-
huites and La Quemada Regions of Zacatecas, Me´xico.
Ph.D. Dissertation, State University of New York, Buffalo
(1995).
21. Trombold, C. D., 1985. A summary of the archaeology of the
La Quemada region. In: The Archaeology of West and North-
west Me´xico. (Foster, M. S., Weigand, P. C., eds.) Westview Press,
Boulder (1985).
22. Nelson, B. A.: Chronology and stratigraphy at La Quemada, Za-
catecas, Me´xico. J. Field Archaeol. 24, 85 (1997).
View publication stats