Journal of Field Archaeology

ISSN: 0093-4690 (Print) 2042-4582 (Online) Journal homepage: http://www.tandfonline.com/loi/yjfa20

Ancient West Mexican Metallurgy: A Technological

Chronology

Dorothy Hosler

To cite this article: Dorothy Hosler (1988) Ancient West Mexican Metallurgy: A Technological
Chronology, Journal of Field Archaeology, 15:2, 191-217, DOI: 10.1179/009346988791974475

To link to this article: http://dx.doi.org/10.1179/009346988791974475

Published online: 18 Jul 2013.

Submit your article to this journal

Article views: 25

View related articles

Citing articles: 1 View citing articles

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=yjfa20

Download by: [Universite Laval] Date: 08 April 2016, At: 11:43

 191

Ancient West Mexican Metallurgy:

A Technological Chronology

Dorothy Hosler
Massachusetts Institute of Technology
Cambridge, Massachusetts

Data from laboratory studies of chemical composition and fabrication techniques of some
400 prehispanic metal artifacts from the M useo Regional de Guadalajara collections were
used to identify two primary artifact groups. One consists of objects made from coppe1;the
other of subtypes of the same artifact types made from copper-arsenic) copper-tin) and other
alloys. The studies showed that these artifact subtype/materials associations were often tech-
nically imperative. A two-phase chronology for the West Mexican metallurgical technology
Downloaded by [Universite Laval] at 11:43 08 April 2016

was constructed by brinjfinf{ these data to bear on datable artifacts. From approximately
A.C. 800 to between A.C. 1200 and 1300) copper was used to manufacture bells) tweezers)
and other objects by lost-wax casting) cold work) and other techniques. Between A. c. 1200
and 1300 to A.C. 1520 copper was used as before but morphologically distinct subtypes of
these artifact classesalso appeared) whose designs required the alloys. The work reported here
is the first to identify associations between designs and materials that characterize particular
artifact classesand to use them to construct a chronology. The presence of the artifact types
described here can be used to date) in approximate terms) otherwise undatable material and
employed as an independent source of temporal information for archaeological material for
which approximate dates may already exist.

mold casting, and cold work and annealing. The second

Introduction phase extended from between A.C. 1200 and 1300 to the
A technically complex metallurgy based on copper de- Spanish conquest in the 1,6th century. Copper continued
veloped in West Mexico in the prehispanic era. The tech- to be used, as before, for the same artifact types, but three
nology, which flourished from approximately A.C. 800 to binary alloys~opper-silver, and the two bronzes (i.e.,
the Spanish conquest, was closely related to the metallur- copper-tin and copper-arsenic)-were also employed with
gies of Central and South America from which important the alloying element sometimes present in high concen-
components had been introduced (Hosler 1986a, in press) . trations. A ternary copper-arsenic-tin alloy was also uti-
Analytical laboratory studies of prehispanic copper and lized. In most cases the artifacts made from the alloys were
copper-alloy artifacts in the large collection of the Museo morphologically distinct subtypes of the objects that were
Regional de Guadalajara (MRG), in Jalisco, Mexico, have still being made from copper. Information from technical
identified the major characteristics of the technology studies performed on these artifacts coupled with materials
(Hosler 1986a, 1988, n.d.). It is now possible to distin- engineering data demonstrate that the consistent use of
guish chronological phases in ancient West Mexican me- particular metals or alloys to fashion certain artifact sub-
tallurgy by bringing the results of these laboratory studies types was not fortuitous; rather, the alloys were necessary
to bear on information from West Mexican and Mesoam- to optimize the design of the subtype and sometimes were
erican metal artifacts found in datable contexts. technical imperatives for its success and for the proper
The evidence indicates that there were at least two function of the object. Hot work, a fabrication technique
chronological phases in the development of West Mexican required when tin is alloyed with copper in high concen-
metallurgy. The first extended from approximately A.C. trations, also appeared at this time.
800 or slightly earlier to between A.C. 1200 and A.C. 1300.
During this period copper was the primary metal used to The Approach
fashion a constellation of artifacts that included tweezers, Metal artifacts are not abundant in West Mexican and
bells, needles, open rings, axes,·awls, and ornaments by a Mesoamerican archaeological sites. For information con-
number of techniques including lost-wax casting, open cerning distribution of metal artifacts in Mesoamerica the
 192 Ancient West Mexican Metallurgy/Hosler

reader may consult Bray (1977), Castillo Tejero (1980), position is unknown, while others have been chemically
Flores de Aguirrezabal and Quijada Lopez (1980), Hosler analyzed but lack specific archaeological provenience. For
(1986a), and Pendergast (1962b). Moreover, prior to the a smaller proportion of artifacts, information exists about
technical study of the MRG collection only a relatively both archaeological context and artifact chemistry; not
small proportion of these artifacts had been analyzed surprisingly, the chemical analytic data available on these
chemically, and studies of fabrication techniques were rare. artifacts substantiate the relations found between partic-
Obviously these circumstances hampered efforts to define ular subtypes and compositional groups in the MRG col-
the major characteristics of the technology or to place lection. It can now be argued that artifacts from datable
them in temporal context. The work reported here ad- contexts that are morphologically identical to objects an-
dresses these very problems and has revealed systematic alyzed in the study are made from the same metal and
use of metals, alloys, and fabrication techniques to fashion alloy types, and produced using the same fabrication tech-
distinct functional artifact types and their subtypes. Those niques. The high level of confidence for this contention is
data provide the key to the chronology. based on materials engineering principles coupled with
The technical studies show that the corpus of MRG the MRG study results. The body of information from
metal artifacts comprises two groups on the basis of chem- the MRG collection and from other Mesoamerican copper
Downloaded by [Universite Laval] at 11:43 08 April 2016

ical conlposition. One consists of tweezers, needles, open and copper-alloy objects can now be organized into a two-
rings, axes, and bells made from copper. The other consists phase scheme identifying the time of appearance of par-
of morphologically distinct subtypes of these same artifact ticular artifact subtypes, as well as the fabrication tech-
types made from the copper-tip or copper-arsenic bronze niques and metals and alloys used to fashion them. Other
alloys, a copper-arsenic-tin ternary alloy, or, occasionally, metals used in Mesoamerica, particularly gold and its al-
from a copper-silver alloy. Materials engineering data can loys, were not examined in this study; their conformity to
explain such patterning on the basis of the markedly dif- this chronology has yet to be explored.
ferent physical and mechanical properties that characterize
copper and these alloys. Such inherent differences in ma-
terials profoundly influence the design possibilities and West Mexican Metallurgy: Technological
functional capabilities of objects made from them. By de- Characteristics
sign I refer to the arrangement and specific dimensions of
the structural elements of any object; among the objects Methods
treated here, elements such as height, degree of curvature, The data were derived from information gained through
length, width, thickness, and internal volume are the most the macroscopic examination of the approximately 3200
important. For example, a depilatory tweezer of a given copper and copper-alloy artifacts in the MRG collection
thickness, length, and width that is made from copper and from the detailed laboratory studies of some 400 of
may not function correctly, whereas an object of these these. The laboratory studies investigated fabrication tech-
same dimensions made from a copper-tin bronze alloy may niques, metal composition, microharclness, and al1oy-
be a fully serviceable implement. At the same time, because property relations in artifacts representing the eight most
manufacturing technique significantly affects the inherent common functional types in the collection as well as a few
properties and mechanical behavior of copper and these objects in the other categories. Table 1 shows the number
alloys, the use of particular manufacturing techniques may of artifacts comprising each type.
also be technically imperative for the success of a design. The MRG artifacts had been acquired over some years
Such interdependence among design, material composi- by Ingeniero Frederico Solorzano, Department of Pa-
tion, and manufacturing technique in metal objects has laeontology and Prehistory, Museo Regional de Guada-
been documented in materials science research and can be lajara, to serve as a study corpus. The artifacts are regis-
experimentally reproduced. tered with the Instituto Nacional de Antropologia e
The two-phase chronology for West Mexican metal- Historia in accordance with Mexico's antiquities laws and
lurgy developed here utilizes the systematic and predict- are government property. The collection was assembled
able associations found between metal composition and without explicit bias: .Solorzano obtained virtually every
artifact design in the MRG objects to assess and chrono- metal artifact or scrap made available (Ing. Frederico So-
logically order the data concerning other Mesoamerican lorzano, personal communication 1981) except for occa-
assemblages containing artifacts formally identical to those sional extremely-expensive items of gold or silver. The
in the study collection. Some artifacts have been found in collection thus includes many fragmentary, partial, and
datable archaeological context, but their chemical com- extremely corroded objects. The stated provenience of the
 Journal ofField Archaeology/Vol. 15) 1988 193

objects was from the West Mexican states of Michoacan, primary quantitative method used was atomic absorption
Jalisco, Colima, and Nayarit. spectrometry, which provided concentrations of major,
The authenticity of the MRG artifacts is supported by minor, and trace elements. Neutron activation studies
both laboratory and archaeological data. Layered external were performed on some objects when the concentration
corrosion products (a cuprite layer in direct contact with of key trace elements occurred at levels below the detection
the metal altered to green copper minerals at the surface) limits of the atomic absorption method.
and intergranular corrosion were microscopically visible.
These were present on all artifacts examined and charac- Fabrication Methods
terize copper and copper-alloy objects that have under- Fabrication methods were identified through interpre-
gone corrosion at slow rates over long periods of time tation of artifact microstructure using metallographic tech-
(centuries) and thus far cannot be faked. Further, most of niques. Metallography involves a series of laboratory pro-
the artifact types identified in the collection have also been cedures to prepare a metal cross-section so that its
found in archaeological context (Hosler 1986a; Pender- microstructure can be examined with a metallurgical mi-
gast 1962b). croscope. To examine a metal artifact using these tech-
niques one or more intact samples (as distinct from filings
Downloaded by [Universite Laval] at 11:43 08 April 2016

or drillings) has to be removed from the artifact. The

Chemical Composition sample is mounted, ground flat, and polished to produce
Investigation of chemical composition of the artifacts a plane, scratch-free surface; it is then etched with a chem-
selected for study consisted of qualitative and quantitative ical reagent to reveal the metallic microstructure.
examinations to determine the metals and alloys present. The microstructure of a metal artifact records and pre-
Alloys, mixtures of two or more metals, can occur natu- serves the history of the procedures used to fabricate the
rally. For example, deposits of gold in Colombia often object. Shaping methods, such as casting, hot and cold
contain as much as 20% silver, and such deposits are, working, and other common techniques produce highly
effectively, natural alloys of gold and silver. In other cases characteristic microstructures that have been experimen-
the metals that eventually form the alloy co-occur in the tally reproduced, studied, and catalogued (Mehl 1961).
metallic minerals of their ores. This is not infrequent in By examining and interpreting the microstructure of a
highland Andean ores of copper, for example, which con- metal artifact it is possible to describe, step by step, the
tain both copper and arsenic in mineral form. When such fabrication history of that object. Often, artifact micro-
ores are smelted, the product is an automatic alloy of structure can also indicate whether or not the object was
copper and arsenic known as arsenical copper or arsenic- ever used since use produces characteristic deformation of
bronze. Other alloys, however, require the intentional the metal crystals or grains.
mixing of two metals by melting them together after each
has been won from its ore.
Mechanical Properties and Artifact Design
The qualitative studies of chemical composition were All metals are characterized by specific intrinsic physical
carried out using emission spectrographic techniques. The and mechanical properties that include hardness, ductility,
toughness, elasticity, and color, among others. These
properties are conferred by the inherent qualities of the
Table 1. Artifact frequency by functional type in metal or alloy and are influenced by fabrication technique.
the MRG collection. To fashion an object successfully from particular metals
Type Total % of Total and to successfully use it, the object must meet certain
Bells 1936 60.5 criteria with respect to mechanical properties. The prop-
Open rings 685 21.4
Axe monies 185 5.8 erties of copper and the copper alloys used by West Mex-
Ornaments 136 4.2 ican smiths can be affected dramatically by the concentra-
Needles 87 2.7 tion of a particular element: the hardness of a cold-worked
Cutting tools 44 1.3
Tweezers 42 1.3 copper-tin alloy axe containing 5% tin is far greater than
Awls/Narrow chisels 23 <1 one containing 1% tin. Fabrication method also markedly
Bell ornaments 22 <1
Pins 17 <1
influences mechanical properties: a cold-worked 5% tin
Fishhooks 14 <1 bronze can be made very hard. A 5% tin bronze that is
Beads 9 <1 cast, or cold worked and left annealed, is nearly as soft as
Other 32* <1
unalloyed copper.
*Number could not be precisely determined.
Mechanical properties thus derive from metal compo-
 194 Ancient West Mexican Metallurgy/Hosler

sition and fabrication technique. The specific artifact di- Table 2. Artifact types in the MRG collection examined
mensions, however, and the instrumental task to be per- for laboratory studies.
formed in turn determine the metals and alloys that can Type Number sampled
be used. For example, to perform the same task, an axe Bells 128
made from copper must be thicker than an axe made from Open rings 81
Cutting Tools (axes, splitting tools) 44
copper-tin bronze, because the bronze is intrinsically Tweezers 39
stronger than copper alone. Because of its strength, the Axe monies 33
Needles 30
bronze axe can be made thinner; hence a bronze alloy is
Sheet ornaments 24
more appropriate for fine, sharp tools. These properties AwlslN arrow chisels 16
are, in turn, affected by manufacturing technique: if a Other (buttons, bell ornaments, fishhooks) 13
bronze axe were to be cold worked beyond a certain point
it would become brittle and subject to fracture.
The factors underlying the choice of particular metals the extremely high cost of the chemical analyses. At the
and alloys to fashion specific MRG bell, tweezer, needle, same time, since fabrication technique~asting by the
and other types were investigated by evaluating the me- lost-wax method-for all bells proved virtually the same,
Downloaded by [Universite Laval] at 11:43 08 April 2016

chanical properties theoretically required to execute the it was not necessary to verify that process through frequent
design in question, and, when appropriate, to actually use and repetitive metallographic examination. In most cases
the object. Whenever possible the design and mechanical the sample goal of 100/0was far exceeded (e.g., axe monies,
properties were evaluated against a set of objective criteria needles, awls); in other cases where technical questions
for the usability or success of artifacts with the same ap- posed by design, composition, or fabrication technique
parent instrumental function. The criteria were developed were especially complex, all or virtually all artifacts within
from materials engineering data and, in the case of the the type were examined.
tweezers, through a computer simulation study of the Seven of the eight most common types shown in Table
artifact design. Once the mechanical properties normally 2 have been reported from datable contexts and will be
required to execute the particular design and for its func- discussed here. 1 A few objects sampled from the "other"
tional success were determined, they were measured for category in Table 2, i.e., fishhooks, buttons, and bell or-
the MRG objects in various ways. Some properties were naments also appear in excavated contexts and have been
measured directly: hardness, for example, by microhard- incorporated into the chronology.
ness tests made on the sampled cross-sections. In other
cases (such as a measure of springiness for the tweezers)
West Mexican Metallurgy: Design and Materials
properties were derived through standard calculations that The results of the technical studies show that within
use experimentally-determined data for the properties of each MRG functional artifact class some objects tend to
specific metals and alloys as a function of their fabrication be made from copper, and others from bronze alloYS.2The
techniques. In these cases, physical data obtained from the sole exception are the ornaments, made nearly exclusively
sampled cross-sections were incorporated into the engi- from gold and gold alloys,3 from silver and from alloys of
neering formulas to arrive at the appropriate value for the silver and copper. What is more, in most cases the designs
property being measured (e.g., elasticity, yield strength, of objects within a single functional category (e.g., bells)
etc.). The basic data about both fabrication techniques made from the bronze alloys differ systematically from
and composition were obtained in the laboratory studies. those made from copper with respect to key parameters,
The technical studies thus provided fundamental, repro- such as thickness and length. As the discussion of each
ducible data that identified fabrication techniques and artifact type will show, when the materials engineering
chemical composition. These in turn provided bases for data are used to evaluate the designs, the systematic as-
understanding the interdependence of design, composi- sociations among particular design attributes, the use of
tion, and fabrication techniques. specific materials, and manufacturing techniques are expli-
cable and predictable. The quantitative chemical analytic
The Sample
Table 2 shows the artifact types selected for the study 1. The subtype of axe money, thin metal objects in the shape of axes,
that is characteristic of West Mexico, has not yet been reported from
sample and the number of artifacts chosen from each. As archaeological excavations.
Tables 1 and 2 indicate, the general goal of a 10% sample 2. The object is identified as an alloy when the alloying elements
was achieved in all cases but one. Bells, the exceptional appear in concentrations sufficiently high to alter the mechanical prop-
erties of the metal.
type, contained far too many artifacts (1,936 items of 3. The gold and gold-alloy objects were not made available for anal-
which a 6.6% sample was made) to meet that goal given ysis.
 Journal ofField A rchaeologyl Vol. 15) 1988 195

Table 3. Bells: compositional groups and dimensions of subtypes sampled.

Ratio of
Metal Average Average height Number
or thickness height* to Number in
Subtype alloy Wall (em) (em) thickness sampled subtype
la Cu, Cu-Sn smooth .05 .61 12.2 8 894
2a Cu smooth .06 .9 15 9 136
4a Cu,t Cu-As wirework .05 1.3 26 5 10
5b Cu smooth .09 1.6 17.8 11 33
6a Cu wirework .06 1.2 20 9 316
6b Cu wirework .07 1.4 20 9 79
7a Cu, Cu-As,4: Cu-Sn wirework .06 1.9 31.7 9 30
8a-cl Cu-As, Cu-Sn wirework .05 3.2 64 12 41
9a Cu, Cu-As, Cu-Sn smooth .08 2.0 25 11 51
9c Cu wirework .10 2.5 25 8 95
lOb Cu-As, Cu-Sn wirework .07 3.6 51.4 18 27
lIb Cu smooth .13 3.0 23.1 6 10
* Height: distance from base of bell to top of resonator.
Downloaded by [Universite Laval] at 11:43 08 April 2016

t Subtype 4a contains one unalloyed copper bell.

4:Two thirds of the bells in 7a are alloys.

data (by atomic absorption spectrometry) for artifacts be- easily-and is known for its resonant properties. Copper-
longing to each type treated in the following discussion tin (Cu-Sn) and copper-arsenic (Cu-As) bronzes are also
appears in Table 12.4 appropriate materials for bells, and for some purposes are
superior to copper. These alloys are, for example, easier
to cast than copper because the presence of the alloying
BELLS
element lowers the melting point of the metal. They also
Among the bells in the collection, 11 types were con- increase fluidity by altering the solidification process itself.
stituted on the basis of formal criteria. Each contains one Alloys solidify over a range of temperatures, a circum-
or more subtypes (Hosler 1986a) as shown in Figure l. stance that permits the molten metal to fill in the detail
The 12 subtypes sectioned are indicated in Table 3. (such as wirework walls) of a mold as the metal solidifies.
All of the bells examined metallographically were cast Alloys are also stronger than copper, permitting castings
using the lost-wax method. Models for the smooth -walled of the former to have thinner walls. In terms of resonance,
bells were apparently fashioned from a single piece of wax. however, bells of the West Mexican design made from the
By contrast, the bells exhibiting "wirework" surfaces were bronze alloys do not differ appreciably from those made
built with individual strands of wax in the same way that from copper (Hosler 1986a).
coiled pottery is built up. The cross-hatch designs in sub- These relations between materials and properties were
type 11b were incised. clearly appreciated by the West Mexican metalsmiths. As
The results of the studies of MRG bell chemical com- Table 3 and Figure 1 show, the alloys were generally used
position (TABLE 3) indicate that some subtypes were made for casting thin-walled, larger wirework bells (subtypes
exclusively from copper. Others contained bells made ei- 4a, 7a, 8a-<i, and 1Ob). The walls of the two subtypes
ther from copper-tin or copper-arsenic alloys, tin concen- (8a-<i and 1Ob) for which alloys were used exclusively are
tration varying between 1.62 and 12.3% by weight and extremely thin in relation to their height, and the bells are
arsenic concentration varying between 0.49 and 23.47% highly symmetrical. The original wax model of the bell
by weight (see TABLE 12). Some of the quantitative data was, presumably, symmetrical; that it was successfully re-
cited in the text of this article do not appear in Table 12; produced reflects the strength and castability the alloys
this is because analyses of certain artifacts were not per- afford.
formed by atomic absorption. Still other subtypes con- Materials engineering data indicate that the strength
tained some bells made from alloys and others from cop- and fluidity that characterize the copper-tin and copper-
per. arsenic alloys were necessary for these larger thin-walled
Copper is a good material for a cast bell because it is designs. As the overall size of the bell decreases, however,
relatively fluid-it flows into and takes the shape of a mold the technical requirements of the design become less strin-
gent. The strength and fluidity provided by copper can
4. Those artifacts analyzed quantitatively using other methods, e.g., suffice for the smaller bells, as exemplified by subtypes 4a
neutron activation, are not included in Table 12. and 7a that contain bells made from copper and from
 196 Ancient West Mexican Metallurgy/Hosler

1a 1b 1el11 1el21 2a 3a

•• 4a Sa 5b
6a 6b
Downloaded by [Universite Laval] at 11:43 08 April 2016

7a 7b 7e

8a 8b 8e 8d

9b ge 9d
9a

ge 9f 10b
10a

10e
o 3 em.
11a
W 11b

Figure 1. Bells from the MRG collection and their subtypes, so designated. Note that the design on
the upper resonator of 11b is incised.
 Journal of Field A rchaeologyl Vol. 15) 1988 197

Table 4. Compositional groups and dimensions of selected bell subtypes.

Ratio of
Average Average thickness
thickness height to Number Number in
Subtype Composition Wall (em) (em) height sampled subtype

leI Cu smooth .05 1.4 28 1 6

Ic2 Cu smooth fragment .9 1 2
7b Cu-Sn wirework .05 1.9 38 2 7
7c Cu-Sn wirework .06 2.5 41.7 2 30
IOc Cu-Sn wirework .05 3.6 72 2 5

bronze. In the cases of smaller bells, the alloys facilitated but are the "wirework" type. They are nearly consistently
the casting but were not essential to it. The bells belonging misshapen, however. Such asymmetry reflects a design
to subtype la, which are extremely small, were likewise executed with non-optimal materials, which in this in-
made using both copper and bronze with only minor stance lacked both necessary strength and fluidity.
differences in wall thickness. The bells in subtypes that contain both copper and
Downloaded by [Universite Laval] at 11:43 08 April 2016

In addition to mechanical advantages, copper-tin and bronze objects (subtypes la and 4a) are small (less than
copper-arsensic alloys were also used for their brilliant 1.3 cm high) or, as in the case of 9a, thick in relation to
colors. Copper-arsenic alloys become silvery as arsenic height. In these examples, dimensional parameters are
concentration increases, copper-tin alloys become golden such that the strength provided by either material suffices
as tin increases. The use of these bronze alloys in high to execute the design.
concentrations for their golden or silvery colors marks one A few objects from other subtypes were analyzed for
of the major, original, technical achievements of West chemical composition. The results, which substantiate the
Mexican metallurgy. These colors were culturally appre- pattern identified in the basic sample, are shown in Table
ciated because they were associated with divinities-silver 4. Larger bells with thin, wirework walls require the
with the moon, gold with the sun. The cultural factors strength provided by the copper-tin and copper-arsenic
that shaped this technology is an issue treated in detail alloys. Copper is successfully used for smaller bells.
elsewhere (Hosler 1986b, 1988, n.d.).
Copper, like the alloys, was also used to cast large bells, 1WEEZERS
but they were smooth-walled and relatively thick in rela- The interdependence among artifact subtypes, manufac-
tion to their height, such as subtypes 5b and lIb. Their turing techniques, and the "metals and alloys used is par-
size requires a strong material but strength is provided by ticularly dramatic in the case of the tweezers in the study
the thickness of the bell walls rather than by the use of collection. Two designs were identified, as shown in pro-
the alloys. Subtype 9c appears to constitute an exception file in Figure 2 (left). The blades of one, called here "beam
to these patterns. These large bells are made from copper tweezers," lack three-dimensional curvature. All beam

2c 2d
Beam Shell 2a 2b

Beam Shell
a 3 em.

o 3 em.

Figure 2. Left: profiles of beam design and shell-design tweezers. Right: front view of the typical blade shapes of
beam and shell tweezers; the designations 2a-<i for the shell tweezers refer to subtypes.
 198 Ancient West Mexican Metallur;gy/Hosler

tweezers but one are made from copper, some left in the tion of their ability to actually function as tools-was
cold-worked condition, others left annealed. The second prompted by data from ethnohistoric sources (Hosler
group, "shell tweezers," possess three-dimensional, shell- 1986a). Those data revealed that at least some tweezers
like curvature below the tweezer hinge, giving the blade were important ceremonial objects, worn by priests and
a dome-like appearance. All shell tweezers but one are elites on state occasions. The study assessed the capacity
made from alloys of copper-tin, copper-arsenic, copper- of such tweezers to also function as utilitarian artifacts.
silver or a low-arsenic ternary copper-arsenic-tin alloy. The The dimensions of particular specimens were replicatec
copper-tin alloys contain tin in concentrations from 2.54 mathematically and the design subjected to the stresse~
to 12.38% by weight, the copper-arsenic alloys from 2.7 experienced in use (FIG. 3). The simulation demonstratec
to 4.43%, and the two copper-silver alloys contain 24.3 that when model implements of the particular dimensions:
and 55.9% silver respectively (TABLE 12). Most shell tweez- compositions, and fabrication techniques were repeated!)
ers made from the bronze alloys are hot worked to shape. closed (and a hair plucked) that they could withstand the:
These two geometries, of beam and shell, represent stresses incurred; all tweezers were functional depilatory
fundamentally different designs for an object required to tools. By changing key dimensional parameters, however
operate as a depilatory implement. In brief, both designs the models also showed that the success of the bean:
Downloaded by [Universite Laval] at 11:43 08 April 2016

distribute the stress at the tweezer hinge. The beam tweez- tweezer design, when made from copper, required tha1
ers focus the load or stress in one small area of the imple- the tweezers be relatively thick, the hinges relatively nar·
ment blade. In the case of the shell tweezers, however, the row, and the blades relatively long. When beam tweezet
load or stress is also distributed across the domed surface thicknesses were decreased to approximate the much thin·
of the blade. The two designs and the subtypes based on ner wall thicknesses of the shell tweezer design, the stresse~
blade shape are illustrated in Figure 2. exhibited were sufficient to induce failure. Tweezers made:
Apart from the differences in design and composition, from copper lack the strength of identical tweezers made:
the beam and shell tweezers differed systematically with from bronze alloys; by increasing dimensions such a~
respect to length, thickness, and hinge width as well (TA- thickness however, West Mexican smiths achieved a work·
BLE 5). The beam design is relatively longer, the metal able depilatory implement. 5
thicker, and the hinges narrower than the shell tweezers. The simulation studies showed that the shell tweezet
A computer simulation study examined the relation, on design demanded the strength provided by the copper·
the one hand, between the specific dimensions of artifacts tin, copper-arsenic, and other alloys. The thinness of the:
representing these two designs, their fabrication tech- metal and the width of the hinge require the strengtb
niques, and their metal compositions, on the other, to the afforded by the alloys to tolerate the stresses generated
respective ability of each tweezer design to function as a when depressing and releasing the implement. The avail·
depilatory implement (Hosler 1986a, n.d.). One objective ability of the alloys with their inherent strength permitted
was to assess the extent to which the use of copper for a design as thin and efficient as the shell type.
the beam tweezers and alloys for the shell tweezers was a 5. The single shell tweezer made from copper was approximately twic(
technical requirement of each design. Another-the ques- the thickness of its bronze counterparts.

Table 5. Comparison of average values of beam and shell tweezers.

Ratio of
Ratio of average
average Average length to
Metal Average Average length to hinge average
or length thickness average width hinge
alloy N (em)* (em)* thicknesi'* (em)* width*
Beam Cu 9 6.5 .13 50 .50 13
Cu-Sn I
Shell Cu I 5.0 .05 100 1.4 3.6
Cu-Sn II
Cu-As 3
Cu-Ag 2
Cu-As-Sn 12
* The averages are calculated for all 10 beam tweezers and for all 29 shell tweezers.
 Journal of Field Archaeology/Vol. 15) 1988 199

UISE9 F.QVIV. STREII

SECTION POINT 5
I.D. VALUE
1 +O.OOEv·OO
2 .•.1.00[+07
3 +1.10£+01
.• +2.40E:+05
I +:J.20E·U'
• +..•..
00£+0.
7 +.••• 0£+0.
8 +5.IOE+0 •
• +1.40£+0'
'0 +7.201£+0.
n +1.001:+01
Downloaded by [Universite Laval] at 11:43 08 April 2016

'1'-~~~NMISES PLOT
, ITEP 1 INCRtUEUT t AISAQUS VERSION "-5-·151

Figure 3. Stress contours generated in simulation study of one MRG

shell design tweezer during use. Note that stress concentrations occur
at tweezer hinge and across the domed blade.

CUTTING TOOLS Table 6. Composition vs. averages of length and

Patterns similar to those of bells and tweezers in the thickness of axes. A total of 17 copper axes and 18
use of copper and the alloys were found among these copper-alloy axes are listed in this table.
tools, the majority of which resemble axes. A few tools Ratio of
average
were too small to have been hafted and are not considered length
here. Two distinct groups are apparent: one consists of Composition Average Average to
MaRl or (percent length thickness average
tools made from copper; the other consists of tools made al/Qy by weight) (em) (em) thickness
from copper-tin bronze, copper-arsenic bronze, and the Cu 10.8 1.0 10.8
copper-arsenic-tin ternary alloy. Objects belonging to both Cu-Sn 0.77-8.06 11.5* .7* 16.1*
Cu-As 0.71-4.80
groups were made by casting to shape in an open mold. Cu-As-Sn As 0.64-1.83
All but a few were then selectively work-hardened at the Sn 1.33-5.3
blade edges (Hosler 1986a). Axes made from copper tend * The averages for the axes made of alloys are for all 18 speci-
to be thicker with respect to length than axes made from mens.
the alloys (TABLE 6).
Both copper and its alloys can be cast thin when the lacks strength and will deform upon impact when used
metal is poured into an open mold but copper is not an for cutting. Although a copper axe blade can be hardened
optimal material for an axe. When made thin, the metal through cold work, the hardnesses achieved are insuffi-
 200 Ancient West Mexican Metallu"lJy/Hosler

Table 7. Rings sampled from MRG collection. Of the 81 rings

sampled, this table lists the 68 that were analyzed for chemical
composition.
Average Average Ratiooj'
Metal or Number oj' thickness diameter diameter to
alluy specimens (em) (em) thickness
Rectangular Cu 6 .16 3.1 19.3
cross-section Cu-Sn 17 .09 2.7 30
Cu-As -
Round Cu 16 .17 3.01 17.7
cross-section Cu-Sn 23 .13 2.72 20.9
Cu-As 6 .13 3.16 24.3

cient to produce a blade with a very sharp cutting edge. a nearly circular form. Bronze permitted artifact designs
West Mexican copper axes were most likely used for wood with bands considerably thinner than those made from
splitting; the blade does not require a keen edge, but the copper (TABLE 7). The concentration of tin in these rings
activity does require a thick, tough implement able to is sufficiently high that the rings required hot-working
Downloaded by [Universite Laval] at 11:43 08 April 2016

absorb impact during use. Many of the copper axe-like (FIG. 5) to avoid embrittlement of the metal.
implements were both sufficiently hard and sufficiently The rectangular-cross-section rings made from copper
thick to have served as wood-splitting tools. are, predictably, substantially thicker than those made
The alloys, by contrast, wt?re used to make thin and from tin bronze; their average diameter with respect to
very hard axes that could be employed for tasks requiring band width is also far lower than that among those made
careful and precise control. Microhardness tests show that of the alloys.
their working edges are sometimes more than twice as The rings with round cross-section, are also most com-
hard as those of the copper axes. Maximum blade hardness monly of bronze. Again, the rings made from bronze tend
values for the cold-worked copper axes range from 128- to be thinner than their copper counterparts (TABLE 7).
141VHN (average 132VHN)6 and for the bronzes 131- As might be expected, rings with smaller diameters (2-
297VHN (average 197VHN). The alloys were used to 2.5cm) were made from either material since both copper
optimize the designs of axes; their thinness virtually re- and bronze provided the strength required for the suc-
quired these materials. Bronze could easily have been used cessful execution of such designs.
to execute the designs of the thicker wood-splitting tools,
NEEDLES
but its superior strength and hardenability was unneces-
sary for such implements. Needles, in two general designs, appear in the West
Mexican corpus (FIG. 6). The eye of one is perforated.
OPEN RINGS The eye of the other is formed by hammering a tab of
The open rings exhibit patterns in the use of copper metal around into a loop, and either tucking it into the
and the bronzes similar to those apparent in the other needle shaft (FIG. 6: 2a) or securing it with two flaps of
artifact types. Two subtypes were identified (FIG. 4): one metal (FIG. 6: 2b).
with a rectangular cross-section, the other with a round
cross-section. The rings that were sampled range in di-
Figure 4. Left) Ring with a rectangular cross-section;
ameter from 2 to ca. 4.4 cm. Table 7 shows the frequency
right) ring with a round cross-section.
of each subtype by compositional group. Many rings may
have been hair ornaments (Maldonado Cardenas 1980).

0-' 0--
Some are found encircling a thick braid of fiber; the ring
and the fiber forming a composite object that perhaps was

.
a hair band (Hosler 1986a: 396).
Most rings with rectangular cross-section were made
from copper-tin bronze in which tin concentration ranged
from 6.29 to 16.1 % by weight. They were fashioned by
- I I

hammering out a thin band of metal, then shaping it into

6. VHN (Vickers hardness number) is a widely used method for o!!!!!!!!!!!Iiiiiiiiiiil!!!!!!!!!!!I3cm.
measuring the hardness of a metal by indenting its surface with a conical
indenter; hardness is determined by the load applied and size of the
indentation.
 Journal ofField A rchaeologyl Vol. 15) 1988 201

the bronze alloys. The loop design required a material that

could be used to fashion a thin shaft from which a narrow
tab of metal, circa half the thickness of the needle shaft,
could be shaped into a loop, then doubled over and tucked
back into the still hollow shaft (FIG. 7). Needles in this
design made from copper often failed, presumably in use
(TABLE 8). The copper-arsenic alloys conferred those prop-
erties required to optimize the design, particularly for the
thin, narrow tab of metal in the loop-eye design. The
doubled over loop concentrated stress during fabrication
and during use.
Copper was used for longer needles of both designs,
however. While not an optimal material it was probably
preferable to copper-tin or copper-arsenic for these objects
because it is both relatively strong and less brittle than the
Downloaded by [Universite Laval] at 11:43 08 April 2016

cold-worked bronze alloys. Brittleness would b deleteri-

ous to needles of this length especially since they must
bend slightly when pulled through fabric or other mate-
rials.

AWLS/NARROW CHISELS.

The 16 objects analyzed in the category of Awls/N arrow

Figure 5. Longitudinal section of a copper-tin bronze ring (185x). Chisels were made both from copper and from the alloys.
Note the elongated pools of eutectoid reflecting hot work. Etchant: Three basic forms were distinguished by the shape of the
potassium dichromate and ferric chloride.
working end (or ends) of the implement: some taper to a
point (FIG. 8: 1); another group is bipointed (FIG. 8: 2);
The design and material composition of a needle must others taper to a narrow blade (FIG. 8: 3); variation occurs
result in an implement that achieves a balance between in the shape of the shank. Lengths range from 5 to 13.6
toughness and strength. The needle must be tough enough cm. In most cases the working tip or blade was severely
so that the stresses to which it is subjected when used- cold-worked to increase hardness (FIG. 9) but the body of
for example, when drawn through heavy cloth-do not the objects was left annealed. Microhardness values range
induce fracture, especially at the eye. At the same time the
needle must be strong enough so that it does not bend or
permanently deform when used. The perforated-eye nee-
Figure 6. Needles: subtype 1 with its perfo-
dles were hammered to shape from copper and have been
rated eye; subtypes 2a and 2b with loop eyes.
left in the cold-worked condition. The loop-eye needles
are made from copper and from copper-arsenic alloys (1.6
to 2.3% by weight) and in one case a copper-tin alloy.
Some were left in the cold-worked condition, others were
left annealed.
Table 8 records the condition of the needles, their
length, and midpoint thickness. Copper was a suitable
material for the perforated -eye needles not more than 10
cm long. Longer needles became plastically deformed with
use. The bending is not smooth and gradual but is irreg-
ular and nonuniform along the length of the needle. The
two exceptions to this pattern were two needles (5 and 6)
that were substantially thicker with respect to length than
those that were deformed, an attribute that provided the
2a 2b
requisite strength for the design~
The loop-eye needles are made from both copper and
 202 Ancient West Mexican Metallurgy/Hosler
Downloaded by [Universite Laval] at 11:43 08 April 2016

Figure 7. Transverse section of loop-eye needle showing tab tucked Figure 9. Tip of awl severely cold-worked to shape and deformed
into needle shaft; (48 X); etchant: potassium dichromate. through use (90 X); etchant: potassium dichromate.

from 106 to 114VHN for the shaft, to approximately during working and annealing (Lechtman 1979, 1980).
140VHN for the tip or blade. These artifacts exhibit no The properties of these alloys that motivated their use and
systematic association between types and chemical com- satisfied the requirements of design were their toughness,
position. Three were made from a copper-arsenic alloy and, above all, their ability to develop a silver color.
with the alloy concentration at about 2 %. There is nothing
inherent in the design of the bronze awls that requires the Summary
physical and mechanical properties of that alloy in pref- The two compositional groups that crosscut the entire
erence to those of copper, although their greater hardness corpus contain on the one hand, objects made from cop-
allowed a-wider range of uses than for those made from per-.either native copper or copper smelted from ores-
copper.
Figure 8. Forms of awls (1-2); form of narrow chisel (3).
ORNAMENTS

Most MRG ornaments were made from sheet metal and

were generally for personal adornment; these include cres- Q, '" '. . '.':.'.'.:.:. :.:".::..:: .::: :'.:: :::.'.:' : 00-

cent pendants, large discs, breast plates, semicircular neck I

pieces, and others. They were hammered exceedingly thin,
with sheet thicknesses averaging 0.04 cm. Those that
•
could be sampled were fashioned from silver and copper- C/'·.:·.:·.·:·.:.: :.'.;.: :: :.::;·:..i L:·:;·:~.'.:~.:~.:·;·!~~
:';::.\ ..;':~;: ;..:::.~'::. ::'.; :':.".:::: ::;;>
silver alloys, although a few, in the shape of discs, were I
found to be made from copper. The properties of the 2 •
copper-silver system make this alloy especially appropriate
for thin sheet metal. The lamellar microstructure of cop-
per-silver alloys (FIG. 10) permits the metal to be reduced
to a very thin sheet without embrittlement, resulting in a
material that is strong and extremely tough. What is more, o ~!!!!!5iiiiiiiiiiil!!!!!!!!!!!5iiiiiiiiiiil!!!!!!!!!1 5 em.
the copper-silver alloys produced material that looked like
silver because of the enriched silver surfaces that developed
 Journal ofField Archaeologyl Vol. 15) 1988 203
Downloaded by [Universite Laval] at 11:43 08 April 2016

Figure 10. Longitudinal section of cold-worked ornament of copper-silver sheet

metal (180 x ). The lamellar microstructure is readily apparent.

Table 8. Condition and dimensions of needles from the MRG collection.

Thickness
Metal or Length (midpoint)
Subtype Artifact no. alloy Condition* (em) (em)

Perforated eye 1 Cu Undeformed 7.4 .20

2 Cu Undeformed 8.0 .10
3 Cu Undeformed 8.4 .15
4 Cu Undeformed 8.9 .20
5 Cu Undeformed 12.9 .30
6 Cu Undeformed 14.7 .20
7 Cu Deformed 10.8 .10
8 Cu Deformed 1l.1 .20
9 Cu Deformed 12.2 .15
10 Cu Deformed 14.0 .15
11 Cu Deformed 14.0 .13
12 Cu Deformed 14.4 .12
13 Cu Deformed 14.8 .15
Loop eye 1 Cu Deformed 13.5 .18
2 Cu Damaged 10.0 .10
3 Cu Intact 22.5 .20
4 Cu Damaged/deformed 8.2 .14
5 Cu Damaged 9.0 .10
6 Cu Damaged/deformed 12.8 .15
7 eu Deformed 14.0 .15
8 Cu-As Deformed fragment .15
9 Cu-As Intact 10.7 .10
10 Cu-As Intact 8.7 .10
11 Cu-As Intact 8.5 .10
12 Cu-As Intact 8.5 .10
13 Cu-As Intact 8.4 .10
14 Cu-As Intact 8.2 .10
15 Cu-As Intact 8.6 .10
16 Cu-Sn Damaged fragment .10·

* "Damaged" refers to the eye; "deformed" refers to the shaft of the needle.
 204 Ancient West Mexican Metallurgy/Hosler

and on the other objects made from copper-tin and cop-

per-arsenic bronze, the ternary copper-arsenic-tin alloy,
and sometimes also from copper-silver alloys. The artifact
GULF OF MEXICO
classes discussed here contained objects made both from
copper and from copper alloys, but in most cases the
designs of such artifacts differed. West Mexican smiths
consistently used the alloys for designs that required the
TOMATLAN- T1Z:PAN - - COJUMATU(N

properties of the alloys: to achieve the design itself, so TUxeeSCO ~ZlNTZUNTZAN- _ • MEXICO CITY

EL CHANAl -APATZINGAN CALIXTLAHUACA

that the object could be used, and, in high concentrations, -INFIERNILLO

for color. The inherent strength of the alloys made possible

designs such as the shell tweezers, the thin rectangular- PACIFIC OCEAN

cross-section open rings and the thin, fine-edged cutting

implements capable of extensive work hardening. The
strength and fluidity of the alloys also permitted casting Figure 11. Map of Mesoamerica showing the principal sites men-
of the bells possessing intricate, thin walls. Copper, by tioned in text.
Downloaded by [Universite Laval] at 11:43 08 April 2016

contrast, was employed to achieve designs whose dimen-

sions and function were compatible with the lesser from collections whose chemical composition has been
strength and fluidity provided by that material: for ex- determined. A review was made of published chemical
ample, the casting of relatively. thick-walled, undecorated analyses of Mesoamerican metal artifacts from both exca-
bells. Copper was also utilized to fashion the thick wood- vated and unexcavated contexts. These data (TABLE 9)
splitting tools, which were cast and then cold worked, and demonstrate the same associations between subtypes and
to fashion beam tweezers with narrow, thick hinges and chemical compositions as characterizes their counterparts
blades. Any of these designs made from copper could have in the MRG collection.
been made from the copper alloys. The converse, however, A temporal framework (TABLE 10) for the appearance
was not true: the artifact designs executed using the al- of these attributes of the technology was constructed by
loys--copper-tin, copper-arsenic, copper-arsenic-tin, and utilizing archaeological parallels of MRG artifact subtypes
copper-silver-frequently could not be successfully accom- but which were found in datable contexts and that had
plished using copper. been chemically analyzed. The dates attributed to the West
Mexican sites or phases prior to about 1970 are from
West Mexican Metallurgy: Temporal Phases Meighan's summary (1974: 1255) of West Mexican pre-
The systematic associations between design subtypes history. For more recent excavations and for sites outside
and the metals, alloys, and fabrication techniques estab- of West Mexico, dates are those provided by the archae-
lished among the artifacts analyzed in the MRG collection ologist, unless otherwise noted. 7
can now be placed in temporal context. That body of data This body of data was enlarged (TABLE 11) by including
is sufficiently large and the associations of artifact designs archaeological counterparts of MRG artifact types from
and artifact chemistries sufficiently well demonstrated to datable contexts which lacked analyses of chemical com-
constitute a reliable foundation from which to infer the position. In these cases, chemical composition was in-
technical characteristics of metal artifacts from excavated, ferred from design.
datable assemblages. These associations constitute strong The resulting chronology is .bound to be refined as
evidence that artifacts from datable contexts whose designs archaeological excavations in West Mexico yield metal ar-
are identical with those of objects analyzed in the study tifacts that are chemically analyzed. The large corpus of
are made from the same metal or alloy types. The question data now available from the technical studies of the MRG
of a temporal sequence for the important attributes of the collection, however, when assessed in conjunction with
technology can thus be addressed by examining the dates
for the sites at which such metal artifacts were excavated. 7. Not all artifacts reported from sites cited in this text appear in the
Sites mentioned in the discussion appear on the map (FIG. tables. At Tomatlcin some artifacts are dated between A.C. 1000 and
1525, others more precisely, to between A.C. 1010 and 1250. Objects
ll). with dates falling within the very broad range, although present in the
First an initial assumption was examined: that patterns MRG collection, were of little use in refining the chronology and were
identified in the principal MRG artifact classes between excluded. At Bernard some artifacts were not from excavated context
and were excluded from Table 10. Objects from Infiernillo burials (TABLE
specific subtypes and their chemical composition also 11) dated solely on the basis of the presence of metal artifacts were also
holds for those same subtypes from excavated contexts or excluded.
 Journal of Field A rchaeologyl Vol. 15) 1988 205

Table 9. West Mexico and other regions of Mesoamerica: artifact subtypes and metal and alloy
types. The numbers in parentheses refer to the list of sites.
MRG West MexiaJ Other Regions of Mesoamerica
Bells
Subtype
la Cu, Cu-Sn Cu (1), Cu-Sn (3) Cu (13, 15*), Cu-Sn (3, 9, 15*, 16*)
2a Cu Cu (1, 4) Cu (7*, 13, 15*)
5b Cu Cu (1, 4)
6a Cu Cu (17*)
6b Cu Cu (1)
9a Cu, Cu-As, Cu-Sn Cu-Sn (3)
9c Cu Cu (5*)
lOb Cu-Sn, Cu-As Cu-As (3) Cu-Sn (8*, 9, 15* 17*t)
lIb Cu Cu (1, 5*) Cu (10, 12, 14, 15*)
Rings
Rectangular ds+ Cu, Cu-Sn Cu (2), Cu-Sn (3)
Round ds Cu, Cu-As, Cu-Sn Cu (1, 2), Cu-Sn (3)
Tweezers
Beam Cu Cu (1) Cu (13)
Downloaded by [Universite Laval] at 11:43 08 April 2016

Shell
2a Cu-Sn Cu-Sn (14, 18*)
2b Cu-Sn Cu-Sn (14)
Axes Cu, Cu-As, Cu-Sn Cu-Sn (3) Cu (14), Cu-Sn (9, 16*)
Needles
Perforated Cu Cu (1)
Loop Cu, Cu-As, Cu-Sn Cu-Sn (17*)
AwlslNarrow chisels
Unipointed Cu, Cu-As Cu-Sn (3) Cu-Sn (7*)
Blade Cu, Cu-As Cu-Sn (9)
Button Cu Cu (4) Cu (14)
Fishhook Cu Cu (1)
Ornaments
Discs Ag, Ag-Cu Ag-Cu (11)
Sheet Ag-Cu, Cu Ag-Cu (3, 6*)

* Artifacts cannot be dated.

t This artifact is a variant of lOb
+ ds = cross-section.

Sites or regions of origin of artifacts:

1. Amapa, Nayarit, Mexico (Hosler notes on Amapa collection 1981; Meighan 1976; Pendergast 1962a).
2. Tomatlan, Jalisco, Mexico (Hosler notes on Tomatlan collection 1981; Mountjoy and Torres 1985).
3. Bernard, Guerrero, Mexico (Brush 1962, 1969).
4. Guasave, Sinaloa, Mexico (Ekholm 1942; Hosler notes on American Museum of Natural History collections
1983; Lothrop 1952).
5. Guerrero (Hosler notes on American Museum of Natural History collections 1983; Lothrop 1952).
6. El Chanal, Colima, Mexico (Kelly 1985).
7. Calixtlahuaca, State of Mexico (Hosler notes, American Museum of Natural History collections 1983; Lothrop
1952).
8. Nochistlan, Central Mexico (Hultgren 1925).
9. Chiapa de Corzo, Chiapas, Mexico (Lee 1969; Root in Lee 1969).
10. Tajumulco Guatemala (Dutton and Hobbs 1943).
11. Zaculeu, Guatemala (Woodbury and Trik 1953; Root in Woodbury and Trik 1953).
12. Toninci, Chiapas, Mexico (Becquelin et Baudez 1982).
13. Mayapan, Yucatan (Pollock, Roys, Proskouriakoff and Smith 1962, Root 1962).
14. Lamanai, Belize (Hosler notes and chemical analyses: Lamanai collection).
15. Cenote de Sacrificios, Yucatan (Lothrop 1952).
16. Valley of Mexico (Arsandaux and Rivet 1921).
17. "Mexico" (Arsandaux and Rivet 1921, 1923).
18. Oaxaca (Arsandaux and Rivet 1923).
 206 Ancient West Mexican Metallur;gy/Hosler

Table 10. Temporal sequence for development of West Mexican

metallurgy based on chemical composition of specific artifact types. Phase
1 is dated to A.C. 800 to 1200/1300; Phase 2 dates from A.C. 1200/1300
to 1520.
Phase 1 Phase 2
Bells
subtype
la Cu (Amapa) Cu-Sn (Chiapa de Corro)
Cu (Mayapan)
2a Cu (Amapa) Cu (Guasave, Mayapan)
Sb Cu (Amapa) Cu (Guasave)
6b Cu (Amapa)
9a Cu-Sn (Bernard)
lOb Cu-As (Bernard)
Cu-Sn (Chiapa de Corro)
lIb Cu (Amapa, Tonini) Cu (Lamanai, Tajumulco)
Rings
Rectangular ds* Cu (Tomatlan) Cu-Sn (Bernard)
Round ds Cu (Tomatlan) Cu-Sn (Bernard)
Downloaded by [Universite Laval] at 11:43 08 April 2016

Needles
Perforated Cu (Amapa)
Tweezers
Beam Cu (Amapa) Cu (Mayapan)
Shell
2a Cu-Sn (Lamanai)
2b Cu-Sn (Lamanai)
Awl Cu (Amapa) Cu-Sn (Bernard, Chiapa de Corro)
Axe Cu (Lamanai), Cu-Sn (Chiapa de Corro)
Button Cu (Guasave, Lamanai)
Fishhook Cu (Amapa)
Ornaments Ag-Cu (Bernard, Zaculeu)
(Sheet)
*ds = cross-section

Table 11. Temporal and spatial placement of MRG collection

artifacts. Phase 1 is dated to A.C. 800 to 1200/1300; Phase 2 dates
from A.C. 1~00/1300 to 1520.
Metal or
Type Alloy Phase 1 Phase 2
Bells
la Cu, Cu-Sn Amapa (Cu) Apatzingan*
Cojumatlan t Culiacan+
Tomatlan Guasave (Cu)
Infiernillo§ Infiernillo
lb Tuxcacuesco**
lcl Cu Cojumatlan
2a Cu Amapa (Cu) Apatzingan
4a Cu-As Tzintzuntzan tt
Sb Cu Amapa (Cu) Apatzingan
Guasave (Cu)
Tuxcacuesco
La Villita++
6a Apatzingan
6b Cu Amapa (Cu)
7a Cu, Cu-Sn Tzintzuntzan
7b Cu-Sn Tzintzuntzan
7c Cu-Sn Tuxcacuesco
8a Cu-As, Cu-Sn Tomatlan§§ Tzintzuntzan
9a Cu, Cu-Sn Bernard (Cu-Sn)
Tzintzuntzan
lOb Cu-Sn, Cu-As Tzintzuntzan
 Journal of Field Archaeology/Vol. 15) 1988 207

Table II. (cont.)

Metal/or
Type Alloy Phase 1 Phase 2

lIb Cu Amapa (Cu) Guasave

Infiernillo
Rings
Round ds*** Cu, Cu-Sn, Cu-As Amapa Apatzingin
Infiernillo Bernard (Cu-Sn)
Tomatlin (Cu) Culiacin
Infiernillo
Tuxcacuesco
Rectangular ds Cu, Cu-Sn Infiernillo Infiernillo
Tomatlin (Cu) Bernard (Cu-Sn)
Needles
Perforated Cu Amapa (Cu) Amapa
Cojumatlin Apatzingin
Infiernillo Infiernillo
Tizapinttt La Villita
Loop Cu-As, Cu-Sn Infiernillo
La Villita
Downloaded by [Universite Laval] at 11:43 08 April 2016

Tzintzuntzin
Tweezers
Beam Cu Amapa (Cu) Infiemillo
Tomatlin Apatzingin
Shell
2a Cu-Sn Apatzingin
La Villita
2c Cu-Sn, Cu-As, Cu-Ag Infiernillo
Tzintzuntzin
2d Cu-Sn, Cu-As Apatzingin
Tzintzuntzin
Awl
Unipointed Cu, Cu-As Infiernillo Apatzingin
Bernard (Cu-Sn)
Infiernillo
Tzintzuntzin
Bipointed Cu, Cu-As Amapa (Cu) Apatzingin
Tomatlin Infiernillo
Tzintzuntzin
Blade La Villita
Axe Cu, Cu-As, Cu-Sn Tzintzuntzin
Button Cu Tomatlin Apatzingin
Culiacin (Cu)
Guasave (Cu)
Fishhook Cu Infiernillo Amapa (Cu)
Tizapin Apatzingin
Tomatlin La Villita
Tzintzuntzin
Bell-Ornament Cu-Sn Tzintzuntzin

*Kelly 1947 ttRubin de la Borbolla 1944

t Lister 1949 :j::j:CabreraCastro 1976
:j:Kelly1945 §§Dates to A.C. 1140
§Maldenado Cardenas 1980 ***ds = cross-section
**Kelly 1949 tttMeighan and Foote 1968

information from excavated collections and existing chem- passes the period from the initial appearance of metal
ical analyses from excavated and unexcavated material, artifacts in West Mexico, at about A.C. 800, or perhaps
more than suffices to make a strong case for the scheme earlier (Cerritos phase at Amapa, AzatIcin complex at To-
set forth here. mathin, Jalisco), to between A.C. 1200 and 1300. The
The chronology for the West Mexican metallurgy con- second phase extends from the latter date until the Spanish
tains two technological phases, or periods, defined by conquest. The evidence at hand indicates that during the
artifact composition and by artifact design, which the tech- first phase, metal objects were made from copper, and in
nical studies show are interdependent. The first encom- several objects at Tomatlan, possibly from a low-arsenic
 208 Ancient West Mexican Metallurgy/Hosler

copper-arsenic alloy.8 The metallurgy of this early period the alloying element. For the copper-arsenic and copper-
was defined by the production of those subtypes of nee- tin bronzes that concentration level was set at approxi-
dles, rings, tweezers, and bells found consistently to be mately 1% and above for the metal to be considered an
made from copper. The objects were made by lost-wax alloy. In the few cases where only qualitative analyses were
casting and open-mold casting, the latter often followed performed, artifacts were classified as alloys based on the
by sequences of cold work and annealing. analyst's estimate of the approximate quantitative value
The second phase is defined by the use of copper-tin corresponding to the qualitative result. Artifacts were ex-
bronze, low and high-arsenic copper-arsenic alloys, alloys cluded where such estimates gave ranges that were too
of copper-silver, and the copper-arsenic-tin ternary alloy. imprecise (e.g., 0.2 to 2%) to determine whether or not
Concomitant changes occurred in artifact design and in the artifact was an alloy.
fabrication techniques. These changes took place primarily The sites from which the artifacts were excavated appear
in the SE sector of West Mexico, which encompasses high- in Table 9. An entry for a site means that at least one
land Michoacan, western Guerrero, and the southern part artifact of the particular subtype is found there; in some
of the state of Mexico. The changes were not accompanied cases many artifacts of the same subtype occur at a single
by the disappearance of those artifact types made of copper site. Analyzed objects that lack datable context are indi-
Downloaded by [Universite Laval] at 11:43 08 April 2016

but, rather, they amplified the earlier technical repertoire. cated with an asterisk.
Information summarizing the chemical compositions of A striking similarity exists between the metal and alloy
artifacts found either in archaeological context or in col- types used to fabricate objects belonging to MRG sub-
lections with designs identical ..to MRG artifacts is re- types and the metal and alloy types used to fabricate ex-
viewed in Table 9. As mentioned earlier, those data amples of those same subtypes appearing in the archaeo-
strongly support the assumption that the metals and alloys logical sites listed, as well as non-datable objects.
used to fabricate such objects were the same metals and Examining the bell data, we see that the MRG subtypes
alloys used to fabricate their counterparts in the MRG consistently made from copper are subtypes 2a, Sb, 6a,
collection. The table is organized by artifact type with 6b, 9c, and 11b. As the information in the second and
compositional data appearing for artifacts that are coun- third columns demonstrates, examples from those same
terparts of specific subtypes of seven major artifact classes subtypes that· have been chemically analyzed prove also to
in the MRG collection. These seven include bells, open be made from copper. As Table 9 also indicates, lOb is
rings, axes, tweezers, needles, awls, and ornaments. The the only MRG bell subtype always made from the bronzes
first column shows the composition of the artifact sub- and that also has been found in datable contexts or other
types analyzed in this study. The second (West Mexico) collections and chemically analyzed. Excavated examples
and the third (Other Regions of Mesoamerica) show the of subtype lOb are likewise made from bronze, as are
composition of such artifacts when found in excavated analyzed but undatable examples. Some MRG bells (sub-
contexts. Because the total number of analyses is not great, types la and 9a) were made either from copper or from
analyses were included of objects from regions outside the alloys. As the table shows, excavated examples as well as
west Mexican area. The examples cited consist of objects examples in museum collections of these bell subtypes are,
I was able to examine personally, or for which a photo- similarly, made from the unalloyed metal as well as from
graph·or illustration exists.9 Artifacts included in the table the alloys.
exhibit those design traits used to classify the type when Turning to the objects that were worked to shape,
found in the MRG collection, unless otherwise noted. chemically-analyzed open rings are made from copper and
The compositional type (Cu, Cu-Sn, etc.) shown in the from the bronzes; the open rings in the MRG collection
table was designated on the basis of the concentration of are made from these same materials.
The site of Amapa, in the state of Nayarit, has yielded
a large number of perforated-eye needles, all made from
8. All Tomatlan artifacts analyzed contain arsenic in trace concentra-
tions but several were as high as 1% and 2%, enough to affect the copper. To my knowledge, no loop-eye needles found in
mechanical properties of the metal. (Very high values cited for two excavated contexts have as yet been analyzed. Arsandaux
artifacts resulted from an error in calculations [Luis Torres, personal and Rivet (1921) analyzed a group of loop-eye needles
communication 1983]). In describing analytic procedures Mountjoy and
Torres (1985) advise caution in evaluating the results for arsenic, so that from the Valley of Mexico that are copper-tin bronze
while the use of copper-arsenic ores at Tomatlan is a good possibility alloys (TABLE 9). One example from the MRG collection
the question requires additional analyses. is made from copper-tin bronze also.
9. I was able to examine the artifacts from Amapa, Tomatlan, La-
manai, and some currently in the collections of the American Museum Beam-design tweezers that were analyzed, like beam
of Natural History analyzed by William Root (Lothrop 1952). tweezers in the MRG collection, are made from copper.
 Journal of Field ArchaeologylVol. 15, 1988 209

Tweezers of the shell design are made from bronze as are and 1300 and is defined by the appearance of the other
all but one of their MRG counterparts. alloys--copper-tin, a high arsenic copper-arsenic alloy,
MRG axes were made from copper, the bronzes, and copper-arsenic-tin, and copper-silver-as well as by the
the ternary copper-arsenic-tin alloy; their counterparts artifact subtypes made from them such as shell tweezers,
were made from copper and the bronzes. Analyzed awls thin-walled wirework bells, loop-eye needles, and others.
and narrow chisels were made from copper-tin bronze: Table 10 presents the outlines of the chronology using
those in the MRG collection were made from copper and data only from excavated and chemically-analyzed objects
the two bronze alloys. with counterparts in the MRG collection. The first column
The MRG collection sheet metal ornaments are made lists artifact subtypes in the MRG collection that were
from silver and copper-silver alloys; several fragments of found in datable contexts and that have undergone chem-
copper sheet (discs) were also analyzed. Sheet metal or- ical analysis. The second column lists the phase 1 sites or
naments made from a copper-silver alloy appear in exca- archaeological cultures where such artifacts have been
vated contexts in two regions. found, together with the artifact composition. The third
Buttons and fishhooks that underwent chemical analysis column lists phase 2 sites where MRG collection subtypes
are made from copper as are their counterparts in the have been found and the chemical compositions of the
Downloaded by [Universite Laval] at 11:43 08 April 2016

MRG collection. artifacts.

The data presented in Table 9 thus establish that the The data presented in Table 11 strengthen the argument
metals and alloys typically used to fabricate particular for the development of West Mexican metallurgy through
kinds of metal artifacts in the MRG collection are char- two phases. They demonstrate that the MRG artifact types
acteristic of those same artifact subtypes when found in manufactured from copper first appear in the earlier phase.
excavated contexts and in other collections. These associ- Those MRG collection types that were frequently or al-
ations are quite predictable. The technical studies have ways made from an alloy of copper-arsenic, copper-tin, or
demonstrated the interdependence of design and chemical copper-silver never appear in the earlier phase.
composition, findings which are in keeping with theory Although chemical analyses have not been performed
and experimental data on design-property relations gen- on all such artifacts excavated at the sites shown in this
erated in materials science and engineering studies. As the table, the results of the technical studies of the MRG
technical studies have also indicated, however, fairly small collection artifacts and the chemical analyses of their coun-
changes in certain critical artifact dimensions (such as terparts summarized in Table 9 and presented in detail in
thickness) can alter the mechanical requirements of the Table 12, indicate that s9me artifact subtypes were con-
object sufficiently so that the use of other materials (such sistently made from copper and others were consistently
as copper in place of bronze, for example) become tech- made from the bronze alloys. Given this consistent (and
nically feasible. As a result, the overall patterns observed often technically necessary) association between compo-
in design and composition in these data will vary in cases sition and design, and in view of the data presented in
where small changes in sensitive parameters-thickness, Table 10 that assign chronological placement to excavated
height, alloy concentration, and so forth-allow them. and analyzed artifacts, the probability is high that those-
as yet-unanalyzed objects that appear at the earlier period
sites are made from copper or possibly from a low-arsenic
A Chronology For West Mexican Metallurgical copper-arsenic alloy. It is also predicted that unanalyzed
Technologies artifacts belonging to the later sites or archaeological cul-
The proposed chronology for West Mexican metallurgy tures and which represent subtypes shown consistently to
can now be outlined. Metallurgy, elaborated over some have been made from bronze (e.g., subtype lOb wirework
seven centuries, was marked by two major technological bells) because of design requirements will likewise be
phases in West Mexico. They are defined by both the made from those alloys.
materials used and the kinds of objects made from those
materials. The first phase began at approximately A.C. 800. Discussion
Lost-wax cast bells made of copper first appeared during The earlier phase or period of West Mexican metallurgy
this period. Also present during this period were various involved the production of a constellation of artifacts
classes of objects shaped by cold work and annealing- made primarily from copper. The metal and ore types
including open rings, needles, tweezers, and awls-and exploited likely consisted of native copper, copper sulfides
made from copper, or from a very low-arsenic copper- such as chalcopyrite, and copper oxides and carbonates.
arsenic alloy. The second phase begins between A.C. 1200 The possible presence of arsenic in low concentrations at
 210 Ancient West Mexican Metallurgy/Hosler

Table 12. Quantitative analyses of artifact composition by atomic absorption spectrometry. Number in first column
refers to catalogue and acquisition number, Museo Regional de Guadalajara. Analyses were carried out at the Central
Analytic Laboratories, Massachusetts Institute of Technology.
Ag As Au Bi Cu Fe In Mg Ni Pb Sb Sn
Bells
Subtype la
F124b 0.080 0.000 * 0.000 * * * * * 0.000 0.000 0.420
F1220 0.010 0.000 * 0.000 * * 0.010 * * 0.000 * 3.240
F1228 0.220 0.090 0.000 0.003 * 0.041 * 0.001 0.024 0.003 0.000 0.000
F1253 0.150 0.020 * * * * 0.030 * * 0.010 * 4.810
Fall 0.093 0.000 * 0.031 * * * * 0.020 0.000 0.000 0.170
Fa12 0.140 0.000 * 0.000 * * * * * 0.000 0.000 2.230
Subtype 2a
F1546 0.019 0.000 * 0.000 * * * * 0.000 0.000 0.000 0.000
Subtype 4a
F1595 0.050 1.230 * 0.000 * * * * * 0.000 0.000 0.020
F2589 0.040 0.490 * 0.000 * * * * * 0.000 0.000 0.170
Downloaded by [Universite Laval] at 11:43 08 April 2016

F2791 0.030 0.000 * 0.000 * * * * * 0.000 0.000 0.270

Subtype 5b
F209 0.060 * * * * 0.020 * 0.001 0.001 0.004 * *
F723 0.163 0.000 0.000 0.000 * 0.043 * 0.000 0.000 0.011 0.010 0.000
F1538 * 0.130 * * * * * * * * * 0.000
F1539 0.110 0.000 "0.000 0.001 * 0.290 * 0.001 0.001 0.007 0.016 0.000
F1567 * 0.270 * * * * * * * * * 0.000
F2411 0.080 * * 0.030 * * * * 0.040 * * *
F2413 0.070 * * * * 0.020 * * * * * *
F2724 0.060 1.000 * 0.000 * * * * * 0.000 0.000 0.040
Subtype 6a
F2080 0.053 0.000 0.000 0.000 * 0.160 * 0.001 0.009 0.000 0.010 0.000
F2126 0.089 0.000 * 0.000 * * * * 0.063 0.000 0.090 0.000
Subtype 6b
F219b 0.054 0.000 * 0.000 * * * * 0.014 0.000 0.090 0.000
F1608 0.148 0.300 * 0.000 * * * * 0.000 0.000 0.160 0.022
Subtype 7a
F891 0.130 * * * * * * * * 0.009 * 10.430
F893 0.050 0.170 * * * * 0.040 * * * 0.030 12.300
F895 0.050 0.230 * * * * 0.020 * * 0.020 * 10.550
F897 0.160 0.000 * 0.000 * * 0.010 * * 0.000 * 7.270
F978 * 0.340 * * * * * * * * * 0.000
F1630 * 1.790 * * * * * * * * * 0.000
F1825 0.032 0.110 * 0.000 * * * * 0.024 0.000 0.000 *
F1830 * 1.140 * * * * * * * * * 0.000
Subtype 8
F197 0.025 1.350 * 0.000 * * * * 0.000 0.000 0.010 0.000
F1783 * 0.650 * * * * * * * * * 0.000
F1784 * 0.980 * * * * * * * * * 0.000
F1795 0.030 1.940 * 0.000 * * * * * 0.000 0.000 0.000
F1798 0.060 1.110 * 0.002 * * * * 0.002 0.002 0.000 0.000
F2393 * 1.010 * * * * * * * * * 0.000
F2649 * 0.340 * * * * * * * * * 1.050
Subtype 9a
Fl446 0.030 0.350 * 0.000 * * * * * 0.000 0.000 0.090
F2183 0.100 0.110 * 0.000 * * 0.060 * * 0.000 0.000 6.380
F2538 0.020 0.380 * 0.000 * * * * * 0.000 0.000 0.160
F2539 0.310 0.100 * 0.000 * * 0.000 * * 0.010 0.000 5.670
F2550 0.127 0.190 * 0.000 * * * * 0.000 0.000 0.009 0.160
Note: copper is the major element when indicated by an asterisk. An asterisk in all other columns indicates that the element was not analyzed
by the atomic absorption method. .
 Journal of Field Archaeology/Vol. 15) 1988 211

Table 12. (cont.)

Ag As Au Bi Cu Fe In Mg Ni Ph Sh Sn
Subtype 9c
F130a 0.190 0.000 0.000 0.001 * 0.340 * 0.000 0.004 0.008 0.010 0.000
F838 0.047 0.000 * 0.000 * * * * 0.000 0.000 0.008 0.000
Subtype lOb
F128a 0.030 0.450 * * * * * * * 0.010 0.020 6.930
F201 0.110 22.120 0.015 0.144 * 0.150 * 0.100 0.015 0.015 1.060 0.000
F204 0.070 12.820 * 0.019 * * * * 0.010 0.020 0.220 0.000
F207 0.110 23.470 * 0.080 * * * * 0.020 0.010 l.080 0.030
F1473 0.020 0.260 * 0.000 * * 0.020 * * 0.000 0.000 4.000
F1475 0.100 12.900 * 0.080 * * * * * * 0.320 *
F1484 0.100 0.100 * * * * 0.010 * * 0.030 * 3.490
F1485 0.050 13.800 * 0.008 * * * * 0.014 0.010 0.210 0.000
F1499 0.020 0.190 * * * * * * * 0.008 * 7.120
F2495 0.030 * * * * * * * * * * 4.780
F2571 0.030 0.210 * 0.000 * * 0.000 * * 0.000 0.000 2.810
Downloaded by [Universite Laval] at 11:43 08 April 2016

Subtype lIb
F1437 0.011 0.000 * 0.023 * * * * 0.023 0.000 0.000- 0.000
F2720 0.104 0.250 * 0.000 * * * * 0.000 0.019 0.070 0.000
F2786 0.070 0.150 0.000 0.000 * 0.015 * 0.000 0.000 0.000 0.020 0.000
Open Rings
Rectangular
cross section
F36a 0.750 * * * * 0.010 * 0.001 * 0.090 * *
F620 0.010 0.190 * 0.000 * * 0.060 * * 0.000 0.000 8.500
F682a 0.190 0.070 * 0.000 * * 0.030 * * 0.000 * 13.380
F682b 0.040 0.490 * 0.000 * 0.020 * * * * * 9.260
F1718 0.050 0.130 * * * * * * * 0.020 * 10.940
F1747 0.040 * * * * * 0.020 * * 0.030 * 12.850
F2313 0.050 0.020 * 0.008 * * 0.000 * * 0.020 * 10.430
Fa8 0.060 0.200 * 0.000 * * 0.000 * * 0.000 * 7.950
Round cross
section
F40a 0.010 0.040 * * * * 0.020 * * 0.030 * 8.890
F873 0.170 0.050 * * * * * * fr 0.009 * 15.160
F997 * * * * * 0.010 * * * * * *
F1664 0.180 0.000 0.000 0.000 * 0.015 * 0.000 0.000 0.068 0.000 0.015
Tweezers
Beam
F5 0.130 * * 0.001 * * * * * 0.001 0.011 0.008
F12 0.090 * * * * * * * * 0.009 * 4.340
F224 0.630 0.000 0.000 0.000 * 0.029 * 0.008 0.000 0.000 0.000 0.000
F225 0.130 * * 0.001 * * * * 0.002 0.003 0.016 0.000
F2344 0.120 2.030 * 0.010 * * * * * 0.060 * 8.840
F2345 0.060 * * * * 0.020 * * * * * *
F2346 0.150 * * 0.001 * 0.010 * 0.001 * 0.090 * *
F2678 0.110 * * * * 0.010 * * * * * *
F2679 0.098 0.000 0.000 0.000 * 0.102 * 0.000 0.000 0.000 0.000 0.000
F2686 0.140 0.130 * 0.002 * 0.042 * 0.011 0.000 0.007 0.020 0.000
Shell
F2 0.080 0.000 * 0.000 * * * * * 0.000 0.000 2.540
F3 0.030 2.700 * 0.000 * * * * * 0.000 0.000 0.000
F4 0.030 3.020 * 0.002 * * * * 0.017 0.000 0.040 0.000
F7 0.030 0.060 * * * * 0.010 * * 0.020 * 10.680
F8 0.110 2.250 * * * * 0.040 * * 0.020 0.040 10.060
F9 0.100 0.390 * * * * 0.130 * * * * 9.930
Fll 0.140 1.190 * * * * * * * 0.010 0.160 9.090
F13 0.170 1.140 * * * * * * * * 0.160 8.960
Note: copper is the major element when indicated by an asterisk. An asterisk in all other columns indicates that the element was not analyzed
by the atomic absorption method.
 212 Ancient West Mexican Metallurgy/Hosler

Table 12. (cont.)

Ag As Au Bi Cu Fe In Mg Ni Ph Sb Sn
F32a 0.050 0.780 * * * * * * 0.030 0.090 * 5.870
F339/76 0.140 0.790 * * * * * * * 0.010 * 9.550
F939 24.300 * 0.009 * 75.690 * * * * * * *
F2017 0.090 * * * * * * * * 0.010 * 9.600
F2343 0.100 1.210 * 0.020 * * 0.000 * * 0.230 0.170 11.760
F2513 0.070 0.280 * 0.020 * * * * * * 0.010 10.340
F2515 0.170 0.350 * 0.002 * * * * * 0.002 0.012 0.010
F2516 0.120 0.190 * * * * * * * 0.020 0.040 8.180
F2517 0.080 0.120 * * * * * * * 0.500 * 9.730
F2518 55.900 * * * 44.100 * * * * * * *
F2528a 0.030 0.150 * * * * * * * * * 11.020
F2528b 0.030 0.140 * * * * * * * 0.010 * 10.700
F2617 0.060 4.430 * 0.000 * * * * * 0.000 0.028 0.090
F2647 0.130 0.180 * 0.000 * * 0.000 * * 0.010 0.000 10.110
F2680 0.100 0.760 * * * * * * * 0.060 0.100 6.360
F2682 0.110 0.750 * * * * * * * 0.010 0.130 6.460
Downloaded by [Universite Laval] at 11:43 08 April 2016

F2683 0.130 2.430 * * * * * * * 0.010 0.100 6.020

F2684 0.060 0.620 * 0.000 * * 0.010 * * 0.000 * 7.930
F2687 0.070 0.810 * 0.000 * * * * * 0.500 0.120 5.690
Axes
F28 0.060 1.220 .' * 0.049 * * * * * 0.000 0.010 0.010
F351 0.070 0.710 * 0.000 * * * * * 0.100 0.040 1.330
F354 0.050 * * * * 0.020 * 0.001 * 0.007 * *
F357 0.100 * * * * 0.010 * 0.001 * * * *
F359 0.080 * * * * 0.010 * 0.000 * 0.006 * *
F367 0.120 0.090 * * * * 0.060 * * * * 2.480
F369 0.110 0.640 * * * * 0.020 * * * 0.180 5.310
F370 0.610 4.840 0.000 0.000 * 0.042 * 0.000 0.021 0.116 0.420 0.000
F371 0.230 * * * * 0.010 * 0.001 0.002 0.007 * *
F372 0.198 0.140 0.000 0.000 * 0.040 * 0.000 0.000 0.000 0.000 0.000
F373 0.220 0.000 0.000 0.000 * 0.250 * 0.000 0.030 0.000 0.000 0.000
F374 0.040 0.080 * * * * * * * * * 7.920
F378 0.030 0.960 * 0.000 * * * * 0.000 0.015 0.000 0.120
F379 0.030 0.060 * 0.000 * * 0.000 * * 0.010 0.000 0.770
F380 0.090 * * * * 0.030 * 0.004 * 0.020 * *
F385 0.060 0.710 * 0.000 * * * * * 0.000 0.280 0.000
F386 0.270 1.310 * 0.000 * * * * * 0.110 0.100 3.100
F387 0.150 0.000 * 0.000 * * * * * 0.000 0.000 0.000
F388 0.220 0.000 0.000 0.000 * 0.000 * 0.000 0.000 0.000 0.020 0.000
F393 0.040 * * * * 0.040 * * * * * *
F396 0.280 0.000 0.000 0.000 * 0.051 * 0.000 0.000 0.000 0.000 0.010
F397 0.050 * * * * 0.010 * 0.001 * * * *
F403b 0.120 0.070 * 0.000 * 0.050 * * * 0.000 0.030 8.060
F853 0.060 * * * * 0.010 * * * * * *
F2249 0.600 0.320 * * * * 0.290 * * 0.008 * 1.940
F2311 0.160 1.830 * 0.000 * * * * * 0.060 * 0.150
F2396 0.050 * * * * 0.040 * 0.001 * 0.010 * *
Needles
Subtype 1
F80 0.040 * * 0.002 * * * * 0.002 0.002 0.002 0.000
F2459 0.110 0.050 0.000 0.000 * 0.018 * 0.000 0.000 0.000 0.030 0.000
F2689 0.100 0.040 0.000 0.000 * 0.000 * 0.000 0.000 0.000 0.009 0.000
Fx14 0.060 * * * * 0.030 * 0.002 * 0.012 * *
Fxl5 0.039 0.000 0.000 0.000 * 0.000 * 0.000 0.000 0.010 0.010 0.000
Note: copper is the major element when indicated by an asterisk. An asterisk in all other columns indicates that the element was not analyzed
by the atomic absorption method.
 Journal ofField ArchaeologylVol. 15) 1988 213

Table 12. (cont.)

Ag As Au Bi Cu Fe In Mg Ni Ph Sh Sn
Subtype 2
F74 0.040 1.610 * 0.003 * * * * * 0.003 0.007 0.000
F83 0.009 0.160 * 0.000 * * * * 0.000 0.000 0.009 0.035
F98 0.030 2.300 * 0.002 * * * * 0.002 0.002 0.003 0.030
F2352 0.040 0.260 * * * * * * * 0.000 * 11.810
F2575 0.070 0.000 * 0.035 * * * * 0.058 0.000 0.000 0.000
F2576 0.140 0.000 0.000 0.000 * 0.060 * 0.001 0.0024 0.002 0.000 0.000
F2577 0.030 * * 0.030 * * * * 0.030 * * *
F2677 0.070 0.000 0.000 0.000 * 0.067 * 0.000 0.070 0.007 0.000 0.000
F804a 0.030 2.170 * 0.000 * * * * * 0.000 0.010 0.000
F804b 0.030 2.120 * 0.000 * * * * * 0.000 0.019 0.000
F804d 0.030 2.130 * 0.002 * * * * * 0.003 0.000 *
F804e 0.030 2.130 * 0.000 * * * * * 0.000 0.098 0.070
F804f 0.030 2.130 * * * * * * * * * *
F806 0.180 0.000 0.000 0.000 * 0.000 * 0.000 0.000 0.000 0.020 0.000
Downloaded by [Universite Laval] at 11:43 08 April 2016

Awls/Narrow
Chisels
Fl12 0.035 2.010 * 0.000 * * * * 0.000 0.000 0.010 0.190
Fl14 0.030 * * * * 0.010 * 0.010 * 0.001 * *
Fl15 0.050 * * * * 0.010 * 0.001 * * * *
Fl17 0.050 * * * * 0.060 * * * * * *
F796 0.040 * * * * 0.040 * * * * * *
F799 0.170 * * * * 0.070 * 0.002 * * * *
F872 0.630 0.220 0.000 0.000 * 0.008 * 0.000 0.000 0.000 0.440 0.000
Sheet Metal
Ornaments
F19 96.330 * * * 3.670 * * * * * * *
F20 99.120 * 0.130 * 0.750 * * * * 0.010 * *
F21 85.220 * 0.072 * 14.700 * * * * * * *
F23b 26.100 * 0.024 * 73.880 * * * * * * *
F229 14.200 0.000 0.015 0.000 * 0.000 * 0.000 0.000 0.030 0.000 0.000
F231 99.460 * 0.011 * 0.530 * * * * * * *
F234 33.900 * * * 66.100 * * * * * * *
F235 98.370 * * * 1.630 * * * * * * *
F816 96.920 * * * 3.080 * * * * * * *
F874 93.430 * 0.290 * 6.280 * * * * * * *
F2586 95.220 * * * 4.780 * * * * * * *
F2627 99.990 * * * 0.009 * * * * * * *
F2628 99.980 * * * 0.019 * * * * * * *
Fspl 95.900 * 0.003 * 4.100 * * * * * * *
Fx 93.330 * 0.034 * 6.670 * * * * * * *
Fxl 93.300 * 0.280 * 6.430 * * * * * * *
Fx2a 96.330 * 0.011 * 5.500 * * * * * * 0.080
Fx4 88.840 * 0.059 * 11.100 * * * * * * *
Note: copper is the major element when indicated by an asterisk. An asterisk in all other columns indicates that the element was not analyzed
by the atomic absorption method.

Tomathin would suggest that the smelting regimes re- the same materials. But a new constellation of designs,
quired to produce such an alloy, which entailed co-smelt- subtypes of these functional types, appear as well, and
ing of arsenic-bearing ores with chalcopyrite, were in use they are made from tin bronze, arsenic bronze, a copper-
at least toward the end of this period. Dates for the To- arsenic-tin alloy, and from copper-silver alloys. The tech-
matlan objects in question are after A.C. 1025. nical evidence has demonstrated that these alloys and the
During the second phase, which began between A.C. new designs were inextricably linked; alloys were necessary
1200 and 1300, the same artifact types were made from for the fabrication and successful function of the objects,
 214 Ancient West Mexican Metallurgy/Hosler

given the particular characteristics of their designs. Or, to logically unlike the MRG artifacts considered here. More-
look at it another way, the alloys enabled or optimized over, chronological placement of the Morett chisel was
the new designs. Hot working was added to the technical problematic; the object could be intrusive (Meighan 1972:
repertoire, a fabrication process that was technically nec- 78). If the dates are correct, I suspect they are imports
essary for forging artifacts that contained tin in high con- from the Andean region.
centrations. The early use of bronze in West Mexico also has been
At the same time that the repertoire of designs and argued on the basis of traces of tin found in copper arti-
fabrication techniques expanded, the range of ores and facts at Amapa and Tomatlan. Meighan (1976: 115) and
extractive and smelting technologies greatly enlarged as Mountjoy and Torres (1985: 152) argue that such tin
well. Cassiterite deposits were first exploited during this could have resulted from the melting down of bronze
later period; the ore was smelted to produce metallic tin artifacts and reuse of the metal. Tin occurs in disseminated
destined for the copper-tin bronze alloys (Hosler 1986a). form, however, in chalcopyrite ores,10 and the smelting of
The extractive and smelting technologies required to pro- such ores could easily result in the incorporation of tin in
duce the high-arsenic bronze bells, and the copper-silver the copper metal in trace concentrations (Hosler 1986a).
alloys described in Hosler (1986a) were also developed Moreover tin concentrations in the MRG artifacts are
Downloaded by [Universite Laval] at 11:43 08 April 2016

during this period. generally either very low (less than 0.2%) which is con-
This temporal study has shown that in general, metal sistent with amounts occuring in disseminated form, or
was used to make the same kinds of artifacts in West high enough (> 0.5 %) to constitute an intentional alloy.11
Mexico from its first appearance there until the conquest; The trace amounts of tin in the Tomathin artifacts occur
the perceptions of the appropriate use of the material did in concentrations slightly higher than in the MRG objects,
not change. That issue is addressed explicitly elsewhere but such differences easily result from the differing ana-
(Hosler 1986a, n.d.). A major change did occur in the lytical regimes and standards used.
technology, sometime between A.C. 1200 and 1300, The spatial extent of these two technological phases
which greatly expanded the technical range of the West needs to be treated in at least a preliminary fashion, despite
Mexican smiths. The most dramatic aspect of the new the lack of systematic excavations in these areas. The two
diversity was the use of new materials: in particular cop- phases clearly do not appear at all West Mexican sites
per-tin bronze, alloys of copper-arsenic-tin and copper- where metal is found. The earlier phase is manifest on the
silver, and high-arsenic copper-arsenic alloys. The low- Michoacan-Guerrero border (the Infiernillo sites), in Mi-
arsenic copper-arsenic alloy became more widely used. The choacan (Cojumatlan, Tizapan el Alto), Jalisco (Tomatlan,
availability of these new materials was totally dependent Ayalan phase, and some material from the Nahualpa
upon the development of appropriate extractive and smelt- phase) and in the Cerritos and Ixcuintla phases at Amapa,
ing technologies, and, by the same token, the new alloys Nayarit.
made possible the new designs. Alloys were required to Artifacts and techniques characteristic of the later phase
produce thin-walled, intricate, lost-wax cast wirework appear in Michoacan at Apatzingan; Tzintzuntzan; on the
bells, and the impressively symmetrical, thin, shell-design Michoacan-Guerrero border at Infiernillo and at La Vil-
tweezers that were hot worked to shape. The new alloys lita; and in Guerrero at Bernard. Although the material is
also were used to fashion thin rings with rectangular cross- not yet analyzed the design and color of some artifacts
sections, hard cutting tools for precise use, and sewing excavated at Cuexcomate, Morelos, by Michael Smith in-
needles, with loop-eye designs requiring the strength and dicate that they are made from bronze alloys (author's
toughness of bronze. The toughness but flexibility of the observations). By contrast, the assemblages in Sinaloa at
copper-silver alloys was exploited to produce thin silver- Culiacan and Guasave, in Nayarit at Amapa, and in Jalisco
copper sheet metal ornaments used for status display. at Tuxcacuesco, lack artifacts that typify the second phase.
Although the evidence points to the use of copper-tin The evidence thus indicates that the first copper objects
bronze only in the late phase, several researchers have
suggested that tin bronze might have been used early in
10. Ulrich Petersen, Department of Geological Sciences, Harvard
the region. Meighan (1969: 14) states that two copper- University, personal communication 1985.
tin bronze objects-a chisel from the Morett site in Colima 11. This figure was suggested to me by Ulrich Petersen (see note 10).
and a tweezer from Barra de N avidad, Jalisc<r-date earlier The figure does not necessarily imply that tin concentration is high
enough to influence mechanical properties but that 0.5% is the highest
than A.C. 1275. They constitute the only published cop- level at which one might expect to find tin that is not the product of
per-tin alloys from the early period and both are typo- intentional alloying.
 Journal ofField Archaeology/Vol. 15) 1988 215

appeared in West Mexico sometime around A.C. 800 or phase, when alloys and new designs appeared, those re-
earlier at Amapa in Nayarit, and Tomathin in ]alisco, and lations were maintained but the system apparently ex-
the Infiernillo sites in the Balsas drainage. Metal in the panded to include goods managed by maritime traders
inland sites-Cojumatlan, Michoacan, and Tizapan el based at Chincha on the central coast of Peru (Hosler
Alto, ]alisco-seems to occur slightly later. Between A.C. 1986a, in press). Chincha traders are believed to have
1200 and 1300, a variety of new metals and alloys appear exchanged highland Peruvian copper for the spondylus
associated with new designs. They seem to be most prev- shell that flourishes in the warm water from Guayaquil
alent in the Tarascan region of Michoacan and its imme- north (Rostworowski 1970). It was likely through this
diate surroundings. To the NW, thus far only the wirework link that copper-tin bronze alloys, pervasive throughout
bell at Tomatlan is unquestionably typical of the second the southern Andes, and certain metal artifact types were
phase. introduced to western Mexico (Hosler in press). The or-
Despite the evidence that the technological changes ganization, spatial extent, items traded, and changes
took place primarily in the Tarascan wne and western and through time in this system have not yet been systemati-
southern Guerrero, it is likely that they also occurred in cally investigated. Marcos (1978: 46), for example, thinks
Downloaded by [Universite Laval] at 11:43 08 April 2016

]alisco, Colima, and Nayarit. Objects reportedly from Lo that the spondylus shell was being sought as far north as
Arado, a site in southern] alisco, and EI Chanal, in Colima, West Mexico and he may be right. Regardless- of the kinds
appear in the MRG collection and were analyzed in the of goods exchanged, there is no question that some com-
course of the study. Most are made from copper-tin and ponents of West Mexican metallurgy were introduced to
copper-arsenic alloys and include subtype lOb wirework the region through such a route and that it played an
bells and shell-design tweezers. Although they have not important role in New World cultural developments.
yet been found in the region in controlled excavations, To investigate these issues we need increasingly accurate
such objects are known to be present in the area (Isabel technological chronologies for the metallurgies of West
Kelly, personal communication 1983) and they were prob- Mexico, other regions of Mesoamerica, Central America,
ably manufactured locally. It is also likely that San Luis and the Andes. Regional sequences must be established in
Potosi and the Huastec area, where cassiterite ores exist, some cases and refined in others. The task requires that
will yield artifacts characterizing the later phase of the analyses of chemical composition, identification of fabri-
technology. An important recent report (Grinberg, Rub- cation techniques, and evaluation of alloy properties and
inovich, and Gasca 1986) describes the analysis of three design of excavated metal artifacts become a routine com-
samples of smelted material, from a Huastec site burial in ponent of archaeological investigations and that analytic
southern Tamaulipas, consisting of copper-tin, copper- regimes be comparable. In West Mexico and elsewhere
arsenic, and copper-arsenic-tin alloys containing arsenic in the presence of artifacts made from copper alloys may
low concentrations and tin in one sample, possibly in high become reliable chronological markers; analogous tem-
concentrations. poral indicators certainly exist in other areas. These data
Both phases of West Mexican metallurgy were strongly are important not only for regional culture histories, and
influenced by the metallurgies of Central and South Amer- for evaluating the cultural use of materials, but also to
ica (Hosler in press). The most plausible explanation at examine the broader cultural relations the contacts among
present for the shared technological attributes is that cer- the metallurgical centers of the New World clearly imply.
tain aspects of the metallurgy of both phases were intro-
duced to West Mexico through maritime trade managed Acknowledgments
by the Manteno in coastal Ecuador. Maritime contact be- The sampling and technical studies were carried out
tween West Mexico and other regions of Central and under permit 401-7/533 granted by the Consejo de Ar-
South America has long been offered as an explanation queologia, Instituto N acional de Antropologia e Historia,
for a variety of cultural similarities among these regions Mexico City in September 1981 and with the generous
(e.g., Kelly 1980; Marcos 1978; Meighan and Foote collaboration of the INAH staff of the Museo Regional
1968; Meighan 1969; Mountjoy 1969) although we de Guadalajara, Guadalajara, ]alisco, Mexico. I would par-
know little about factors that precipitated and sustained ticularly like to thank Arque610go Otto Sch6ndube and
such contact. The technical attributes characterizing the Ing. Frederico Solorzano for their help. The laboratory
earlier period of West Mexican metallurgy indicate that studies were carried out at the Massachusetts Institute of
relations during that time were most direct with coastal Technology at the Center for Materials Research in Ar-
Ecuador and Colombia/Central America. In the later chaeology and Ethnology. Support for the research was
 216 Ancient West Mexican Metallurgy/Hosler

provided by NSF doctoral dissertation improvement grant Bray, Warwick

BSN-25022, Wenner-Gren Foundation for Anthropolog- 1977 "Maya Metalwork and its External Connections," in Nor-
man Hammond, ed., Social Process in Maya Prehistory:
ical Research grant 4149, and a pre-doctoral fellowship at
Studies in Honour of Sir Eric Thompson. New York: Aca-
the Massachusetts Institute of Technology awarded by demic Press, 365-403.
Industrial Minera Mexico S.A. I would like to thank the
Brush, Charles F.
curatorial staff at the American Museum of Natural His- 1962 "Pre-Columbian Alloy Objects from Guerrero, Mexico,"
tory for permission to examine their Mesoamerican metal Science 138: 1336-1337.
collections. Drs. Joseph Mountjoy, David Pendergast, Mi- 1969 A Contribution to the Archaeology of Coastal Gum-ero, Mex-
chael Smith, and Clement Meighan permitted me to ex- ico. Ph.D. dissertation, Columbia University. Ann Arbor:
amine and cite information regarding artifacts from To- University Microfilms.
mathin, Lamanai, Cuexcomate, and Amapa respectively; Cabrera Castro, Ruben
artifacts from Amapa were subsequently returned to the 1976 Arqueologia en el Bajo Balsas Gum-ero y Michoacan: Presa
La Villita. Tesis Profesional, Escuela N acional de Antro-
Mexican government through the good offices of Dr.
pologfa e Historia, Mexico, D. F.
Jaime Litvak, then director of the Instituto of Investiga-
Castillo Tejero, Noemi
ciones Antropol6gicas, Universidad Nacional Aut6noma
Downloaded by [Universite Laval] at 11:43 08 April 2016

1980 "Distribuci6n Espacial de Objetos del Metal: Base de un

de Mexico. David Parks, Professor of Mechanical Engi- Intento TipoI6gico," Rutas de Intercambio en Mesoamerica
neering, Massachusetts Institute of Technology, and Matt y el Norte de Mexico. Saltillo, Mexico: Sociedad Mexicana
Lewis and Melissa Krawizcki, undergraduates in Mechan- de Antropologia, XVI Reuni6n de Mesa Redonda, 53-
62.
ical Engineering and Materials Science, respectively, pro-
vided invaluable help with the tweezer simulation study. Dutton, Bertha P., and Hulda R. Hobbs
1943 Excavations at Tajumulco, Guatemala. Monographs of the
I would like to thank Professors George L. Cowgill (Bran- School ofAmerican Research 9. Santa Fe, NM: The School
deis University), Heather Lechtman (Anthropology/Ar- of American Research.
chaeology and Materials Science, Massachusetts Institute
Ekholm, G. F.
of Technology), David Parks (Mechanical Engineering, 1942 Excavations at Guasave) Sinaloa. Anthropological Papers of
Massachusetts Institute of Technology), Robert Rose the American Museum of Natural History 38, no. 2. New
(Materials Science, Massachusetts Institute of Technol- York.
ogy), and Barbara Voorhies (University of California at Flores de Aguirrezabal, Ma., Dolores, and Cesar Armando Quijada
Santa Barbara) for reading and commenting upon the first Lopez
1980 "Distribuci6n de Objetos de Metal en el Occidente de
draft of the manuscript. Mexico" Rutas de Intercambio en Mesoamerica y el Norte
de Mexico. Saltillo, Mexico: Sociedad Mexicana de Antro-
pologia, XVI Reuni6n de Mesa Redonda, 83-92.
Dorothy Hosler received her Ph.D. in Anthropology in 1986
Grinberg, Dora, R. E. Rubinovich, and A. A. Gasca
with a specialization in Mesoamerican archaeologyfrom the 1986 "Elaboraci6n Intencional del Bronce en Mesoamerica,"
University of California at Santa Barbara. She has dnne field in Precolombian American Metallurgy. Bogota: Banco de
work in West Mexico and Ecuador and isprincipally interested la Republica, 37-65.
in ancient technology and the cultural uses of materials. She Hosler, Dorothy
has been appointed (effective 1989) Assistant Professor ofAr- 1986a The Origins) Technology and Social Construction of Ancient
chaeology at the Massa£husetts Institute of Technology. Mail- West Mexican Metallurgy. Ph.D. dissertation, University
ing address: Building 20D-11 0) M.l. T.) Cambridge) MA. of California, Santa Barbara. Ann Arbor: University Mi-
crofilms.
02139.
1986b "The Cultural Organization of Technology: Copper Al-
loys in Ancient West Mexico," in Precolombian American
Metallurgy. Bogota: Banco de La Republica, 65-86.
Arsandaux, H., and P. Rivet 1988 "The Metallurgy of Ancient West Mexico," in Robert
1921 "Contribution a l'Etude de la Metallurgie Mexicaine," Maddin and Ke Jun, eds., The Beginning of the Use of
Journal de la Societe des Amtricanistes de Paris 13: 261- Metals and Alloys. Cambridge: MIT Press, 328-343.
280.
in press "Ancient West Mexican Metallurgy: South and Central
1923 "Nouvelle Note sur La Metallurgie Mexicaine," L'Antro- American Origins and West Mexican Transformations,"
pologie 33: 63-85. article accepted for publication in American Anthropolo-
gist.
Becquelin, Pierre, and Claude Baudez
1982 Tonina Une Cite Maya du Chiapas Mexique. Mission Ar- n.d. "Style in Materials: The Metallurgy of Ancient West
cheologique et Etnologique Franfaise au Mexique. Paris: Mexico," article accepted by editor for publication in
Editions Recherche sur les Civilizations. volume on "style" edited by Margaret Conkey.
 Journal ofField Archaeology/Vol. 15) 1988 217

Hultgren, Axel soamerican Studies 4. Carbondale: University Museum,

1925 "Microscopical Investigations of a Bell from Mexico," Southern Illinois University, 11-25.
Journal de la Societe des Amiricanistes de Paris 17: 206-
1972 Archaeology of the Morett Site) Colima. University of Cali-
210.
fornia Publications in Anthropology 7. Berkeley and Los
Kelly, Isabel Angeles: University of California Press.
1945 Excavations at Culiacan) Sinaloa. Ibero-Americana 25 1974 "Prehistory of West Mexico," Science 184: 1254-61.
Berkeley and Los Angeles: University of California Press.
1976 "The Archaeology of Amapa, Nayarit," Monumenta Ar-
1947 Excavations at Apatzingan) Michoacan. Viking Fund Pub- queologica Vol. 2. University of California at Los Angeles:
lications in Anthropology 7. New York. Institute of Archaeology.
1949 The Archaeology of the Autlan- Tuxcacuesco Area of Jalisco) Meighan, C., and L. Foote
II: The Tuxcacuesco-Zapotitlan Zone. Ibero~mericana 27. 1968 Excavations ofTizapan el Alto) Jalisco. Los Angeles: Uni-
Berkeley and Los Angeles: University of California Press. versity of California at Los Angeles, Latin American Cen-
1980 Ceramic Sequence in Colima: Capacha) an Early Phase. ter.
Anthropological Papers of the University ofArizona 37. Tuc- Mountjoy, Joseph B.
son: The University of Arizona Press. 1969 "On the Origin of West Mexican Metallurgy," in J.
Charles Kelley and Carroll Riley, eds., Pre-Colombian
Downloaded by [Universite Laval] at 11:43 08 April 2016

1985 "Some Gold and Silver Artifacts from Colima," in Mi-

chael Foster and Phil Weigand, eds., The Archaeology of Contact Within Nudear America: Mesoamerican Studies 4.
West and Northwest Mesoamerica. Boulder, CO: Westview Carbondale: University Museum, Southern Illinois Uni-
Press, 153-179. versity,26-42.
Mountjoy, Joseph B., and Luis Torres M.
Lechtman, Heather
1985 "The Production and Use of Prehispanic Metal Artifacts
1979 "Issues in Andean Metallurgy," in Elizabeth P. Benson,
in the Central Coastal Area of Jalisco, Mexico," in Mi-
ed., Pre-Columbian Metallurgy of South America. Wash-
chael S. Foster and Phil C. Weigand, eds., The Archae-
ington, D. C.: Dumbarton Oaks Research Library and
ology of West and Northwest Mesoamerica. Boulder, CO:
Collections, 1-40
Westview Press, 133-152.
1980 "The Central Andes: Metallurgy Without Iron," in Theo-
Pendergast, David M.
dore Wertime and James D. Mulhy, eds., The Coming of
1962a "Metal Artifacts from Amapa," American Antiquity 2 7:
the Age of Iron. New Haven, CT: Yale University Press,
370-379.
267-334.
1962b "Metal Artifacts in Prehispanic Mesoamerica," American
Lee, Thomas A.
Antiquity 27: 520-545.
1969 The Artifacts ofChiapa de Corzo. Papers of the New World
Archaeological Foundation No. 26. Provo, UT. Pollock, H. E. D., R. L. Roys, Tatiana Proskouriakoff, and A. L.
Smith
Lister, R. H.
1962 Mayapan) Yucatan) Mexico. Carnegie Institution of Wash-
1949 Excavations at Cojumatlan) Michoacan) Mexico. University
ington 619. Washington, D.C.
of New Mexico Publications in Anthropology No.5, Albu-
querque: University of New Mexico. Root, William C.
1953 "Report on Some Metal Objects from Zaculeu," in R.
Lothrop, Samuel K.
B. Woodbury and A. S. Trik, eds., The Ruins of Zaculeu)
1952 Metals from the Cenote of Sacrijice) Chichen Itza) Yucatan.
Guatemala. Richmond: United Fruit Company, 262-
Memoirs of the Peabody Museum of Archaeology and Eth-
268.
nology Vol. 10, NO.2. Cambridge, MA: Harvard Uni-
versity. 1962 "Report on the Metal Objects from Mayapan," in H. E.
D. Pollack, ed., Mayapan) Yucatan) Mexico. Carnegie In-
Maldonado Cardenas, Ruben
stitution of Washington 619: 391-399. Washington, D.C.
1980 Ofrendas Asociadas a Entierros del Injiernillo en el Balsas.
Mexico, D.F.: Instituto Nacional de Antropologia e His- 1969 "Metal Artifacts from Chiapa de Corro, Chiapas, Mex-
toria. ico," in Thomas Lee, ed., TheArtifacts ofChiapa de Corzo.
Papers of the New World Archaeological Foundation 26.
Marcos, Jorge G.
Provo, UT, 203-207.
1978 "Cruising to Acapulco and Back with the Thorny Oyster
Set: A Model for a Lineal Exchange System," Journal of Rostworowski de Diez Canseco, Maria
the Steward Anthropological Society 9(1-2): 99-132. 1970 "Mercaderes del Valle de Chincha en la Epoca Prehis-
paruca: Un Documento y Unos Comentarios," Revista
Mehl, Robert F.
Espanola de AntropologiaAmericana 5: 135-177.
1961 Atlas of Microstructures of Industrial Alloys. Metals Hand-
book 7, 8th eOO., Metals Park, OH: American Society for Rubin de la Borbolla, Daniel
Metals. 1944 "Orfebreria Tarasca," Cuadernos Americanos 3(4): 127-
138.
Meighan, Clement W.
1969 "Cultural Similarities Between Western Mexico and An- Woodbury, Richard B., and Aubrey S. Trik, eds.
dean Regions," in J. Charles Kelley and Carroll Riley, 1953 The Ruins of Zaculeu) Guatemala. Richmond: United
eds., Pre~Colombian Contact within Nudear America. Me- Fruit Company.