How, why. and when did Olmec culture col- which to investigate the fate of the Olmecs.
lapse and what do we meon by the concept of Located on the western margin of thc Olmcc
a collapse in this context! heartland, the site contains a long archaeo-
Richard A. Diehl, 1989 logical sequence that includes Olmec and Epi-
Olmcc components in addition to later Classic
. nothing is 1<nown ohout the Olmec-posl- and Postclassic occupations. Although Tres
Olmec transition beyond the bore foct thol Zapotes has been studied longcr than any other
Son Lorenzo ond Lo Vento were abondoned at major Formative site in the Olmec heartland,
approximately this time. The limited infor· previous studies failed to ascertain the overall
mation we hove on Tres Zapotes suggests thol extent of the site or to produce an accurate site
research there will provide importont insights map, much less provide detailed information
into this tronsition. on the organization and history of settlement
Richard A. Diehl, 1989 of the site. In 1995 I initiated a new phase of
research at Tres Zapotes to address questions
concerning the evolution of political and eco-
1\7
an updated interpretation of site chronology
and apply it to a reconstruction of the occu-
pational history of Tres Zapotcs. This recon-
struction provides the basis for the subsequent
discussion of continuity from Olmec to Epi-
Olmcc culture and the evolution of political
organization at Tres Zapotes. I conclude with
a model of political evolution that takes into
account the ecological setting of Tres Zapotes,
the history of regional political and economic
systems, and the development of new forms
of political expression.
History of Research
Tres Zapotes first attracted scholarly attention
in 1869 when Jose Melgar reported the dis-
covery of a colosscll head by a compesino on
the Hacienda Hucyapan (fig. i). Seventy years
later, in 1939, Matthew Stirling initiated the
first modern exploration of an Olmec site at
Tres Zapotes. His discovery of Stela C, and
Marion Stirling's reconstruction of a Cycle 7
baktun coefficient for its inscribed Long Count
date, provided early support for a Formative
placement of Olmec culture [fig. 2) (Stirling
1940). Working with Stirling, Philip Drucker
[19431 conducted the first stratigraphic exca-
vations in an Olmec center and worked out a
general ceramic chronology, later revised by
Michael Coe in 1965 and refined by Ponciano
Ortiz in 1975. The stone monuments of Tres
Zapotes, which now numher more than forty,
have been the subiect of several studies (Porter
1989), including Howell Williams' and Robert
Heizer's (196 sllandmark petrographic c1l1alysis,
and the obsidian assemhlage of the site was one
of the first in Mesoamerica to he characterized
hy physicochemical means [Hester et al. Il)71).
Although Tres Zapotes figured prominently
in the early history of Olmec studies, it was soon
eclipsed hy the spectacular finds cit La Venta
Regional Setting
[Stirling 1943,1947; Drucker [9<;2; Drucker et
a1. 1959) and San Lorenzo (Stirling 1947; Cae The westernmost of the major Formative I. Tr~s Z"potes Monumcnt
A, the Caheza Colusal Je
1968; Coe and Diehl i9S0). As these eastern period centers in the O]mec heartland, Tres
HoeyapclI1, Formative periuJ,
sites became the paragons of Olmec culture, Zapotes occupies an area of rolling sedimen- basalt
ecological explanations of Olmec evolution tary uplands between the volcanic massif of
came to focus on the peculiarities of their low- the Sierra de los Tuxtlas on the east and the 2. Stela C. upper portiun
land riverine settings, and Olmec social com- alluvial plain of the Rio iJapaloapan and its showing Initial Series glyph
anJ baktun codncicnt of
plexity became the "Gift of the River" (Coe tributaries on the west (fig. 31. This ecologi- Lung Cuunt date, Formative
1\)81). As a result, scholars have underapprcci- cally diverse setting provided the people of Tres period, ~tooe
ated the significance of variation in the regional Zapotes with most of the resources they
settings of heartland Olmec sites. required for their basic livelihood. The lclkes
and swamps of the Papaloapan basin teemed
138 POOL
with aquatic resources, and the alluvial plain argued that the distribution of major Olmec
provided vast expanses of fertile agricultural centers and their association with specific sets
land. If, as Drucker (194J: 8) believed, the sedi- of natural resources reflect a system of coop-
mentary uplands were less intensively culti- erative exchange based on zonal complemen-
vated, they would have provided diverse forest tarity, which would have been under the
resources in addition to underlying deposits of control of chiefs who may have reinforced the
high-quality pottery clays. Most significantly, ties between centers through marriage alliances
the inhabitants exploited the nearby slopes of [Grove 1994: 228; see also Arnold, this volume).
Cerro El Vigfa and the ravines descending from I argue here that the location of Tres Zapotes
them for the distinctive porphyritic basalt from vis-a-vis other Gulf Coast centers and natural
which they fashioned stone monuments and resource zones is important for understanding
grinding implements. The only commonly used the history of its growth and sociopolitical
material that was not available nearby was ob- organization. First, however, I update the pic-
sidian; it does not occur naturally in the Sierra ture of the site's geography as it has been re-
de los Tuxtlas. Chemical analyses indicate that vealed through recent archaeological fieldwork.
the people of Tres Zapotes looked westward
for sources of obsidian, the bulk of which they
Site Layout
obtained from the Pico de Orizaba, Guadalupe
Victoria, ZanIgoza, and Oyameles sources in The archaeologiccll site of Tres Zapotes covers
central Veracruz and Puebla (Hester et al. 1971). about 450 hectares on either side of a large
As David Grove [1994: 227-228) has empha- bend in the Arroyo Hueyapan (fig. 4). Alluvial
sized, the upland environment of Tres Zapotes terraces hound the floodplain of the arroyo to
differs significantly from the riverine and estu- the east and west. Cerro Rabon and Cerro
arine settings of the more intensively studied Nestepe, two hills formed by resistant volcanic
eastern heartland sites of San Lorenzo and La ash deposits, or la/a, rise above the phlin on
Venta. Taking note of the environmental diver- the east bank of the arroyo. A broad ravine
sity of the Olmec heartland, Grove has recently delimits the northern edge of the site.
GROUP 2
E
0 r-1-~
, c.•
A • B
low mounds on center lines within plazas, and GROUP 1 • E F A
o 0
I
~
/
~
b 00 I
I
0° ~ 0 :
ODJj
(It- \
N
&.. 0 1
kilometers
in Ortiz' type collection arc large and well pre-
served, suggesting minimal f1uvial transport.
Furthermore, auger tests conducted in 1996
encountered the daub-rich remains of a house-
mound below the volcanic ash on the east side
of the arroyo, confirming Formative period resi-
dential occup8tion on the alluvial plain.
Occupational History
The distribution of diagnostic rim shercls in
our systematic transect surface collections
reveals significant differences in the organiza-
tion of Olmec and Epi-Olmec occupation at
Tres Zapotes.
Early to Middle Formative diagnostics at
Tres Zapotes include white-rimmed black
wares and white wares. Although tecomate
rims are also diagnostic of Early to Middle For-
mative occupation, I have not included them
in this analysis hecause their functional equiv-
7. Tres Z,Jpotes J\!\olllJlllcnt
alents in the Late Formative period are non- Q, Fonndtivc period, sron,;
diagnostic striateu coarse ware ollas, which
continue in large frequencies in the Classic
8. Tres Z,lpotes Monulllent
period. I have also not separated Early from H, Forllldtive period, stone
Middle Formative phases. The most diagnos-
tic Middle Formative wares are white wares,
which are quite rare and occur in association
with Black and White ceramics and tecomates
in Ortiz' collections; separating them creates
a probably erroneous impression of population
decline in the Middle Formative. Furthermore,
discriminating between Late Formative and
Protoclassic occupation is difficult due to the
erosion of the diagnostic Polished Orange
shercls of the Hueyapan phase in surface collec-
tions. For these reasons the following analysis
only distinguishes between Olmec (Early to
Middle Formative) and Epi-Olmec (Late For-
mative to Protoclassiel occupations.
Surface materials of the Otmec occupation
are concentrated on the elevated terrace to the
west of the arroyo and on Cerro Rabon to the
cast of the arroyo (fig. (,). The 1996 survey also
encountered Olmec ceramics on the lower
slopes of terrace remnants farther to the east.
Concentrations of Olmec ceramics on the val-
ley plain are associated with mounds amI un-
doubtedly represent old deposits incorporated
in later mound fill. We do not at present know
the extent of Olmec occupation beneath the
alluvium of the valley plain. Nevertheless, the
distribution of Olmec sherds derived from the
shallower deposits of the alluvial terrace re-
veals a pattern of small, discrete communities Though scholars disagree about the tempo-
covering I to 40 hectares separated by zones ral placement of several monuments at Tres
with little or no occupation. Zapotes, most accept as Olmec the two colos-
Mound construction does not appear to have sal heads (Mons. A and QI (figs. 1,7), two seated
been typical of the Olmec occupation. Of the figures (Mons. I and 11, and the head of a were-
fourteen mounds sectioned by Stirling's proj- jaguar statue IMon. Hllfig. 8] and assign most
ect, none produced assemblages assignable of the remaining monuments to the Late For-
exclusively to the Olmec occupation [Drucker mative period [Lowe 1989: 43i Milbrath I979i
I943i Weiant 1943). The only possible excep- Porter 1989: 97-100). A basalt column cham-
tion is represented by Mound Ein Group I [fig. ber, excavated in 1978 in Group 2, is similar
4). The initial construction stage consisted of to Tomb A at La Venta (Lowe 1989: 60). The
ared clay mound about [ to 1.5 meters tall chamber contained a rectangular stone slab
with sandstone steps [Weiant 194): 6-71. Un- pierced by a circular hole in which was placed
fortunately, Stirling only excavated a corner an upright serpentine "plug" [Mons. 33 and
of this basal mound, and it was apparently 34), a damaged piece of dressed stone (Mon.
sterile. A single incised Black ware sherd found 32), and a basalt column with a crude petro-
just above the surface of the red mound prob- glyph face (Mon. )I). I On the hasis of their con-
ably dates to the Late Formative period. Rather text, these may also be counted among the later
than constructing mounds, the Tres Zapotes Olmec monuments of Tres Zapotes. The spa-
Olmecs appear to have taken advantage of nat- tial distribution of the known Olmcc sculp-
ural eminences, perhaps filling ami leveling ture reinforces the impression of small, discrete
them, as may be the case on Cerro Rabon and communities but does not correspond closely
on the projecting ridges of the Ranchito Group. to the ceramic distributions (fig. 6). The colos·
This method of construction parallels that sal heads, for example, were found in plazas
recently reported from San Lorenzo by Ann that do not exhibit high frequencies of diag-
Cyphers (1996: 69-70). nostic Olmec sherds. The most likely expla-
o 1
kilometers
10. Tres Z'lpmes Monument
fl), Lite Fort1l<lrtvl: pel'lml.
slOne
depicts a leftward-facing head amid curved, its greatest elaboratlOn on the Gulf Coast in
upward-radiating lines above the cleft brow of the inscription on La Mojarra Stela I (fig. 191.
an abstract were-jaguar mask (fig. lSI (see also Joyce Marcus (1992) has recently argued that
Porter [9Rl): pl. sa and my fig. 21 The Olmec early writing and calendrical systems in Meso-
affinity of the mask has been defended by Cae america developed in response to competition
(I965b: 756) and Porter [I9R9: 49-50). The upper among chiefly elites who legitimized their sta-
portion of the design, however, was found later tus through propagcmda directed at peers and
and has been discussed less frequently. The subordinates. In this context, the historical
leftward-facing head in this part of the carving accuracy of an inscription would have been
calls to mind figures on celts from Rio Pes- less important than the relation of elite activ-
quem, and elsewhere, which Reilly (I995: V~ ities to the mythical past and the prophetic
39) identifies as representcltions of the ruler as future. The Terminal Olmec stelae of La Vema
the axis mundi or world tree, thus reinforcing and the Epi-Olmec stelae of Tres Zapotes and
the Olmec conception of this celtifonn stela. La Moiarra appear to document the evolution
In contrast to the Early Formative colossal of this practice from its nonlitenlte roots to its
heads, the Late Formcltive stelae of Tres Zapotes literate climax as rulers sought new modes of
and its environs present a pronounced change legitimation in an increasIngly competitive
in sculptural themes related to rulership, from poli tical landscape. Indeed, at Tres Zapotes,
static representations of rulers to depictions competitors for rulership may have been as
of legitimizing acts. This shift docs not repre- near as the next mound group.
sent an abandonment of Olmec themes, how-
ever, but a shift in emphasis already presaged
Conclusion
in La Venta Stelae 2, 3, cmd 5, for example. The
recording and display of such events suggest a Our continuing archaeological survey has
greater concern with historicity, a develop- helped clarify the l1dture of the Olmec OCCll-
ment that is expressed most explicitly in the piltion at Tres Zapotes and has documented
Long Count date of Stela C and that reaches the Epi-Olmec growth of the site. As has long
been suspected, Tres Zapotes no longer can be
considered a major Olmec center on a scale
equivalent to La Venta or San Lorenzo. Rather,
Olmec occupation at Tres Zapotes was dis-
tributed among several small communities.
Nevertheless, at least two chiefs in the Tres
Zapotes zone were able to commission colos-
sal head portraits in stone, emulating the rulers
of the eastern centers. These chIefs probably
extended their control over nearby vi]]ages,
and they may have exerted broader influence
on their contemporaries in the western periph-
ery of the Olmec heartland.
Although further analyses and investigation
will be required to isolate the Middle Forma-
tive component at Tres Zapotes, at present the
evidence from ceramic complexes and stratig-
raphy provide little support for a significant
disjunction in occupation at the end of the
Middle Formative. Olmcc villages appear to
have expanded and coalesced to form a site
extending over more than )00 hectares in the
Late FormCltive period. The Epi-Olmec growth
of Tres Zapotes coincided with the abandon-
18. Tres Zaputes Stela C,
LIpper fragment, front ment of La Venta, the growth of centers beyond
Author rhotUj.~r;.lph the western margin of the Olmec heartland,
and a pronounced change in obsidian technol-
19. La Mojarra Stela I ogy and resource utilization both at Tres Zapotes
and in the nearby Sierra de los Tuxtlas. I have
suggested in this essay that the underdevelop-
ment of political hierarchy in the Olmec period
and the expansion of the site in the Epi-Olmec
period are consistent with a hypothesis of zonal
complementarity in regional exchange systems
of the Formative period.
Reinterpretation of earlier mound excaVcl-
tions at Tres Zapotes suggests that the con-
struction of formal mound groups began in the
Late Formative period and continued into the
Classic period. The principal mound groups
are widely disperseo and of similar scale, sug-
gesting a weakly developed political hierarchy.
If true, this raises the possibility that rulership
may have been negotiated among elites with
competing claims to authority. Under the
model proposed above, that authority would
have extended to control over resource zones,
exchange networks, and productivc labor.
A prominent feature of mound groups at
Tres Zapotes is their association with Late For-
mative stelae that appear to record events,
either visually, as in Stelae A ilnd 0, or tcxtu-
"lily, as in Stelel C. Following Marcus' (1992)
arguments, these monuments are Interpretable
as propagandistic declarations to subordinates B1BLIOGRAPHY
and competing elites, which drew their legiti-
Barrett, Thomas P.
lTIacy from references to myth, legend, and !l)<)(; Formative Ohsidi'ln on the Culf Coast
prophecy. Moreover, they form part of a devel- of Mexicu: Industry l)evelnpment in the
opmental sequence of increasingly explicit Tuxtlas Regiun. Paper presented ,\I the
lTIythicohistoricaJ references beginning in the 61st Annual Meeting of the Society for
Terminal Olmec phase of La Venta and culmi- American Archaeolngy, New Orlc'lIls.
nating in the Protoclassic La Mojarra stela. Ikrn,t1, Ign'leio
In conclusion, the rumors of an Olmec col- 19(;<) Thc Olmet: World. Ikrkeley.
lapse have been greatly exaggerated. Instead, Chase, Jall1es E.
the Olmec to Epi-Olmec transition marks a 1<)81 The Sky Is Falling: The San Martin Tuxt]a
time when the inhabitants of the western 01- VOIc'"1ic Eruption and Its Effects on the
Olmec at Tres Zapores, Veracruz.
mec heartland slIccessfully adapted their
Vinculo.\' 7: 53 -(,9.
Olmec traditions to the political and economic
Clcwlow, C. William, II'.
landscape of the Late Formative Mesoamerican
1974 A Stylistic uni/ CIJlol1o/ogicul Stlld)' of
world. a/mee Monllmentlll SellljllUre. Berkeley.
IlJS2 Lo Vento, Tobliscu: A Siudy of Olma Pool, Christopher, Patty Wright, and Georgia
Cellimics ond Art. Bureau of American Mudd Britt
Ethnology Bulletin 1~1. Washington. 199) Formative House/ot Structure in Ilczua-
pan, Veracruz, Mexico. Report submitted
1')81 On the Nature of Olmec Polity. In The
to the H. ). Heinz, III Charitable Trust,
Olmec lind Their Neighbors, cd. EJizabeth
Pittshu rgh.
Benson, 2')-48. Washington.
Porter, James 13.
Drucker, Philip, Robert F. Heizer, 'Ind Robert J.
1989 The Monuments and Hieroglyphs of Tres
Squier
Z'lpOtes, Veracruz, Mexico. PhD. disser-
[lJ)l) [xLi/volium ot Lo Vento. Tohoscu, 19\5
tation, University of California, Berkeley.
Bureau of American Ethnology Bulletin
170. Washington. Reilly, F. I(ent, 111
19l)1 Art, Ritual, and Rulership in the Olmec
Grove, David C.
World. In The Olmec World: 1~/I[/(I1I//ld
IlJl)4 La Isla, Veracruz, IlJl)l: A Preliminary
Hulet'shil" 27-46Iexh. cat., The Art Mu-
Report with Comments on the Olmec
seum, Princeton UniversitYI. Princeton.
Uplands. Ancient Mesoomerico ):
227-22.8. Smith, Virgini'l G.
1984 IZI/I'I/ "elid Cllrvin,~: Form COlJlen/.
Hester, Thomas R., Robert N. j'lck, ~nd Robert F.
I! Illes lor Design and Role in Mesoomeri-
Heizer
can Art Hlstury a/lcl ArcJ1IIeolog)'.
1971 The Obsidi~n of Tres Zapotes, Veracruz,
Washington.
Mexico. In Poper, on Olmec and Moyo
Archoeo]ogy, <::d. John Graham, 6~ - " I. Stark, Barbara L., Lynette Heller, Mlch'lel D.
Berkeley. Glascock,]. Michael EI,II11, ~nd Hector Neff
1l)l)2 OhsiJia/l-Artif,lct Source Analysis for the
Justeson, John S., and T<::rrence Kaufman
Mixtequilla Region, S()uth-Centr~1 Ver~
llJlJ3 A Decipherment of Epi-Olmec Hienl-
cruz, Mexico. Lali/l Amencan A/lI/l/Ii/l)'
glyphic Writing. Science 2~l): 170.,-1711.
" 221-2'l).
Lowe, CMeth W.
Stirling, Matthew W.
\lJR9 The Heartland Olmec: Evolution of
1940 An l/liliol Serie" hom Tres Z0I'0leS,
M'lteri,,1 Culturc. [n !ies/onol PerSflec-
VeroCl"uz. Mexico. Washington.
lives 017 Ihe O]mec. cd. Robert Sharer
and D'lvid Grove, J ,-('7- Cambridge. 1941 Stone Monument.' of SoUl hem MexICO.
Bureau of Americ'ln Ethnology Bulletin
Marcus, Joyce
IJK. Washington.
[l)l)2 Mesoomericon Wrilins SyslelJJs. I'mflo-
glindo, Myth. ond I-listory in Four Ancienl 1l)47 On the Tr,JiI of La Venta Man. Nillionill
CiviliZe/lions. Princeton. Geogrophlc l)1(21: 1'17-172.
Melg,1I', Jose M. Weiant, Clarence W.
[R69 Antigiiedades mexlcanos. RoletllJ de 194, An Intmduc/ion 10 the Cerllmics of
]0 Sociedod Mexicono ric Ceogllifw y Tres Zilpoies. Vert/cruz. Mexico. Bureau
Esteufisticli, 2d series, I: 2lJ2-297. of American Ethnology 13ullctin I'l).
W'lsh ington.
Milbr~th, Susan
1971) A Study of Olmec Sculpll/Ilil Chronolosy Williams, Howell, and Robert F. Heizer
Washington. 1l)65 Source~ of Roch.\ Usecl in O]mec
Monumenl.'. Berkeley.
Miller, Mary E.
19x6 The Art of MesolllJJenco from Olmec
to Aztec. New York.
rOOL IS)