The Archaeology ofSettlement Abandonment in

Middle America

Edited by Takeshi Inomata and Ronald W. Webb

THE U N I V E R S I T Y OF UTAH PRESSSALT L A K E CITY

FOUNDATIONS OF A R C H A E O L O G I C A L I N Q U I R YJames M. Skibo, editor

Copyright © 2003 by the University of Utah PressAll rights reservedoo Printed on acid-free paper

The Defiance House Man colophon is a registered trademark of the University of Utah Press.It is based upon a four-foot-tall, Ancient Puebloan pictograph (late PHI) near Glen Canyon,Utah.

2003 04 05 06 07 08 095 4 3 2 1

L I B R A R Y O F C O N G R E S S C A T A L O G I N G - I N - P U B L I C A T I O N DATA

The archaeology of settlement abandonment in Middle America / edited byTakeshi Inomata and Ronald W. Webb.

p. cm. — (Foundations of archaeological inquiry)Includes bibliographical references and index.

ISBN 0-87480-752-2 (hardcover : alk. paper) —•ISBN 0-87480-753-0 (pbk. : alk. paper)i. Indians—Antiquities. 2. Indians—Urban residence—Latin America.

3. Land settlement patterns—Latin America. 4. Ethnoarchaeology—LatinAmerica. 5. Extinct cities—Latin America. 6. Latin America—Antiquities.I. Inomata, Takeshi, 1961- II. Webb, Ronald W,1960- III. Series.£65 .A67 2003

2—dc2i 2003001348

Contents

List of Figures vii

List of Tables ix

Foreword xiMICHAEL BRIAN SCHIFFER

Preface xv

1. Archaeological Studies of Abandonment in Middle America iTAKESHI INOMATA AND RONALD W. WEBB

I. Rapid Abandonment

2. From Episodic to Permanent Abandonment:Responses to Volcanic Hazards at Tetimpa, Puebla, Mexico 13PATRICIA PLUNKET AND GABRIELA U R U N U E L A

3. Xochicalco, Morelos:The Abandonment of Households at an Epiclassic Urban Center 2.9RONALD W. WEBB AND KENNETH G. HIRTH

4. War, Destruction, and Abandonment:The Fall of the Classic Maya Center of Aguateca, Guatemala 43TAKESHI INOMATA

5. Volcanic Activity and Abandonment Processes:Ceren and the Zapotitan Valley of El Salvador 61BRIAN R. McKEE AND PAYSON SHEETS

II. Gradual Abandonment

6. A Place of Continued Importance: The Abandonment of Epiclassic La Quemada 77BEN A. NELSON

7. The Abandonment of Teotihuacan 91LINDA M A N Z A N I L L A

8. Monte Alban and Late Classic Site Abandonment in Highland Oaxaca 103MARCUS WINTER

7

The Abandonment of Teotihuacan

LINDA MANZANILLA

Abandonment is a recurrent phenomenon insettlements. It is often related to hazards,stemming from either natural or socialcauses, which may have different effects insites and regions. Thus, the first step in aban-donment research is to determine the archae-ological correlates of different types of disas-ters—natural or social (Manzanilla 19973);then the type of abandonment process—gradual or abrupt—should be defined, fol-lowed by the characterization of the societyaffected by a particular type of hazard.

As for the first step, natural hazards in-clude floods, droughts, earthquakes, volcan-ism. Floods may be detected in mud and siltstrata covering sites, a phenomenon that isfrequent in the Lower Mesopotamian plain,particularly in Ur and Shurrupak (Buringh1957:37; Jacobsen and Adams 1958; Raikes1966:61-62) and the Nile Valley (Butzer1984) but also in coastal Peru because of ElNino-Southern Oscillation effects (Moore1991; Moseley 1987; Ortlieb and Machare1992; Thompson et al. 1984, 1985, 1986).Erosive sequences and vegetational changesmay reveal droughts, to which urban cen-ters—such as Teotihuacan in central Mexicoor Tiwanaku in Bolivia—were often very vul-nerable (Garcia 1974; Manzanilla 1992.,i993b, i997b; Metcalfe et al. 1989, 1991;Ortloff and Kolata 1993; Paulsen 1976). Along-lasting drought stimulated generalizedregional collapse of urban centers in the NearEast around 2000 B.C. (Bard 1993; Malek

1986; Rosen 1993; Weiss et al. 1993). Fis-sures, fractures, faults, and collapse of build-ings signal the effects of seismic movements(Vittori et al. 1991). Tephras, lahars, pyro-clastic flows, and lava are evidence of volcan-ism (Sheets and Grayson, eds. 1979).

Abandonment processes may also be re-lated to social and ecological causes. Saliniza-tion of soils due to overirrigation or shep-herding, deforestation (O'Hara et al. 1993),raids, warfare, and conquests are ecologicalor social phenomena that may cause differentscales of abandonment in sectors, sites, or re-gions.

With regard to the second step, two typesof abandonment processes, in their varioussubtypes, are mentioned in this book: grad-ual and rapid. Gradual abandonment may betriggered by deterioration or salinization ofcultivation areas, drought, changes in rivercourses or exchange routes, nonfulfillment oftributary tasks, messianic movements, or col-lapse of urban organizations. The amount ofobjects subtracted from activity areas is re-lated to the portability of the objects andvaluables, the distance to the new site, the ob-jects' personal or economic value, their inac-cessibility in the new settlement area, andtheir requirement for the fulfillment of cer-tain specialized tasks. Rapid abandonmentmay have different modalities: the most dras-tic may be due to the destruction of settle-ments by raids and conquest, volcanic laharsand lava, floods, mud flows, or earthquakes.

L I N D A M A N Z A N I L L A

There is another type that allows people totake some of their belongings, even thoughthe event may also be dangerous: volcanicash emissions, differential destruction of sitesby invaders, some types of earthquakes, andthe like.

The third step in abandonment research isthe characterization of the society that isaffected by the natural hazard or social con-flicts. A disaster is the result of the confluencebetween a dangerous natural hazard and avulnerable society (Garcia Acosta 1996:18).The vulnerability of a given society may beassessed by answering the following ques-tions: ( i ) What was the duration and the geo-graphical scale of impact? (2) What was thesettlement hierarchy of the affected system,and what type of settlements were disrupted?(3) What developmental stage was the givensociety in? Was it in a vulnerable stage withrespect to resource access, social integration,regional networking, or centralized author-ity? (4) Did the society have preventive tech-nological measures regarding the type of haz-ard or conflict that most commonly affectedthe region? and (5) What were the conse-quences of the hazard or conflict? Did it bringabout changes in, among other things, settle-ment patterns, site abandonment, bufferingmechanisms for cyclical disasters, massivestorage, or integration of symbolic interpre-tations of these events in myths?

The final effects of natural and socialdisruptions may be ( i ) migration, with socialeffects in the population that migrates aswell as in the societies that receive the mi-grants; (z) devastation of specific areas withrespect to their productivity, with conse-quent changes in productive strategies; (3)hydrological changes with the regional con-sequences they represent; (4) increase in thesedimentation of nearby areas with conse-quences for agriculture; and (5) disruption oftrade routes.

The scale may also vary. Macroregionalcollapses brought about by major climaticchanges have recently been proposed byscholars working throughout the Near East(Weiss et al. 1993). By the end of the thirdmillennium B.C. a catastrophic collapse of

Early Bronze Age societies from Turkey to In-dia coincided with a severe shift toward adrier climatic era. Volcanic eruptions havealso been detected in some of these areas, andlamentations of famine and drought charac-terize the epoch. Almost all the cities andtowns in western Palestine were deserted(Rosen 1993). The decreased frequency andamount of rainfall changed the hydrologicalregime, eliminating the buffering effect offloodwater farming. Primary data for thestudy of this climatic change come frompollen, paleolimnology, oxygen isotopes, andgeomorphology (Rosen 1993). Evaporationand increased salt deposition, decline of oakpollen, wadi incision, and flooding of valleybottoms are some of the phenomena relatedto this change.

Why did these societies fail to adapt to thenew situation? In the Near East there aregreat fluctuations in precipitation, streamflow, and crop yield from year to year. Rosen(1993) cites some of the predictable measuressocieties took: diversification of crops andherd animals, food storage, retention and dis-tribution of information on famine foods,transformation of surplus food into nonper-ishable items of value that could be traded forfood in times of stress, and extension of thesocial network to allow access to food re-sources from other regions. He notes severalfactors that may explain why the agriculturalsectors of Early Bronze III society failed to re-spond successfully to the climatic desiccationat the end of the period: ( i ) state control oversurplus production, (z) nondiversification ofsubsistence crops, (3) loss of floodwaterfarming as a buffer, and (4) a slow responsetime on the part of elite managers.

With these issues in mind, let us addressthe evidence for the abandonment of a hugemetropolis of the Classic horizon in centralMexico, Teotihuacan. Of particular impor-tance is a multifactorial explanation of thefall of the city and the different steps involvedin its abandonment.

ABANDONMENT AT TEOTIHUACANTeotihuacan was one of the most extensiveurban developments of preindustrial times.

/Quetzalpapalotl Palace

Pyramid of the Moon

Xalla'alace of the Sun

Xolalpan^Tlamimllolpa.. . , _ .it K? Merchants Barrio

Yayahuala

Zacuala

Tetitla

Oaxaca B

yramld of the Sun

Pirul and Varillas Tunnels

Street of the Dead Complex

1D Compound

1E Compound

Ciudadela

Figure 7.1. The city of Teotihuacan. (Redrawn from Millon 1973)

L I N D A M A N Z A N I L L A

A city of order and cosmopolitan activities,a pilgrimage center, a political capital, and amanufacturing center, it housed one of thelargest populations of Mesoamerica. Itsabandonment was no ordinary fact, and itshook Classic Mesoamerica in many ways.

The importance of Teotihuacan (Figure7.1) in the Prehispanic history of centralMexico is such that human groups that cameafter its collapse referred to it as a mythicalplace. The Legend of the Fifth Sun is centeredin Teotihuacan, where the new sun, or era,began (Sahagun 1969, 2:258-259). Sahagun(1969, 3:209-210) mentions that the Mexicacame to Teotihuacan to make sacrifices andto invest a new lord. The relevance of the me-tropolis in the political scenario of the Classicperiod stimulates us to discriminate the dif-ferent elements that are linked to its collapse:fire, dismantling, ritual destruction, violence,abandonment rites, looting, and reoccupa-tion.

Fire in the Central Portion of the CityEvidence of fire has been detected by Millon(1988:149-152) in nearly all monumentalstructures on the Street of the Dead (the city'smain north-south axis) and in temples andassociated buildings in the rest of the city. Yetof the 965 apartment compounds examined,only 45 showed clear evidence of burning(5 percent). In general, burning was found infront of and on both sides of the staircasesand on top of the temple platforms (Millon1988:150).

In many cases carbonized wooden beamsand jambs have been found on top of floorsof the last occupational level, particularly inthe Quetzalpapalotl Palace, to the west of theMoon Plaza (Figure 7.2) (floors of the ante-chamber with stains of the carbonized beamsand pillars: Acosta 1964:25); the House ofthe Priests, near the Pyramid of the Sun (Ba-tres 1906:15); the Viking Group, in the cen-tral portion of the Street of the Dead (wherean ash layer was found before the structurecollapsed, as well as carbonized beams andpillars, alterations in the color of wall paint-ings, smoke stains, etc.; Armillas 1944); thecentral plaza of the northern palace (Struc-

ture iD) of the Ciudadela (Jarquin Pachecoand Martinez Vargas 1982:113); the Xallacompound; and Teopancazco, an apartmentcompound to the southeast of the Ciudadela(I unearthed carbonized wooden beams nearportico floors of the last occupation in our1997 field season; Manzanilla 1998), amongother structures.

Some of this carbonized wood has beensent for radiocarbon dating. Bernal (1965)wonders why the uncalibrated dates rangefrom A.D. 200 to 290 and speculates that thespan occurs as a result of the reuse of Tlami-milolpa beams by Xolalpan builders. I foundsimilar results in Teopancazco, where thewooden beams gave early dates.

Wolfman (1990) also sampled manyburned floors, pestholes, and column anddoorjamb molds for archaeomagnetic dat-ing. Some came from Structure iD in La Ciu-dadela, the Viking Group, and Teopancazco;because of the crossover point, the archaeo-magnetic dates concentrate roughly aroundtwo possible means: A.D. 310 and 475 (Wolf-man 1990:300). Our archaeomagnetic datesfrom Teopancazco seem to fall at the end ofthe sixth century A.D. (Hueda 2000) and thusare placed at the end of the Xolalpan phase.

Wolfman argued that if the extensiveburning at Teotihuacan occurred ca. A.D.475-500, the Maya hiatus (around A.D.534-593) could be seen as a partly relatedevent (Wolfman 1990:301, 303). If the fire isdated before the beginning of the exodus ofthe population, then it may be seen as an in-ternal revolt, a prelude to the disarticulationof central authority in the city. If it is dated tothe end of the city's history, then all the eventspossibly related to the collapse may havebeen intertwined. Some of the events thathave been cited include the following (Man-zanilla 1995).

1. Nomad incursions. Jimenez Moreno(1959:1066) proposed that Otomi nomadgroups inhabiting the northern sector of thevalley could have been involved in theburnout of the city's core.

2. Agricultural collapse and deforestation.Mooser (1968:31) thought that the growthof the city provoked the destruction of

T H E A B A N D O N M E N T O F T E O T I H U A C A N

Figure 7.2. The central courtyard of the Quetzalpapalotl Palace, with evidenceof destruction of the carved pillars. (From Acosta i964:Figure 37; reproducedby permission of the Instituto Nacional de Antropologfa e Historia)

nearby forests, and thus originally advanta-geous natural conditions turned into adverseones. Millon (1967:48) added that immoder-ate woodcutting for construction and limeprocessing produced erosion, causing a de-crease in agricultural soil humidity. Recently,Barba (1995) calculated that the city neededz million m3 of calcium carbonate for itsplastering, and thus z.2. million m3 of woodto burn the limestone into lime (four timesthe surface of the valley floor). If persistentdrought occurred around A.D. 650 in centralMexico (Garcia 1974; Lorenzo 1968; Man-zanilla 1992.; Metcalfe et al. 1989; O'Hara etal. 1993), many seminomadic or nomadicgroups may have migrated to more benignenvironments. At that time, the city may havesuffered from a lack of food and increasingsocial deterioration, making it vulnerable toexternal threats. Isotopic data support thepossibility of long-lasting droughts in centralMexico during the sixth and seventh cen-turies A.D., with consequences for the subsis-tence strategies of Epiclassic groups in thecity (Manzanilla et al. 2.000).

3. Powerful marginal groups. Palerm andWolf (1971:191-193) thought that in thenorthern frontier of Mesoamerica, increasingaridity would require the organization of

irrigation systems and the need for protectionagainst the incursions of nomads, factorsthat would reinforce military patterns. Thispower could then be used against ancient civ-ilization centers such as Teotihuacan.

4. Blocking of provisioning networks.Chadwick (1966:2.) noted that Mixtecs,Olmeca-Xicallanca, and Chocho-Popoloca(who were in contact with Teotihuacan dur-ing its last phases) would have taken advan-tage of the social unrest in the city to blockexchange and access routes.

Millon (1988:149) has mentioned thatsome of the causes of Teotihuacan's collapsewere bad administration of economy andpolitics, inflexibility toward change, an in-efficient bureaucracy, and the deteriorationof exchange networks. Many if not all ofthese factors would have coincided in the lastperiods of Teotihuacan history. If the govern-ment was collective and in the hands of sectorheads, often depicted as anonymous priests(whose main function was related to fertility-rites, as the mural paintings suggest), one canimagine how the population of the city mayhave reacted toward drought, soil erosion,and lack of food. The result was populationdispersion toward the west, east, and south,although during Coyotlatelco times (A.D.

L I N D A M A N Z A N J L L A

Figure 7.3. Dismantled staircase from the Quetzalpapalotl Palace. (FromAcosta i964:Figure 23; reproduced by permission of the Institute Nacional deAntropologia e Historia)

600-850) the site was still inhabited, particu-larly in its central and eastern portions.

Our dates from Late Coyotlatelco occupa-tions in the tunnels to the east of the Pyramidof the Sun (Manzanilla et al. 1996) are placedaround A.D. 550-700. I propose that thecity's fall began when its core was set on firearound A.D. 550 (the collapse of the elites),together with the first incursions of the Coy-otlatelco groups, perhaps from the Bajio re-gion or farther north (our strontium isotopedata suggest that some Epiclassic and EarlyPostclassic burials from the tunnels to theeast of the Pyramid of the Sun are migrants tothe valley of Teotihuacan [Price et al. zoooj).It is possible, however, that the Coyotlatelcolived in the city some time before the collapseand participated in the transition and loot-ing. Eventually, the apartment compoundswere deserted, and finally, the entire periph-ery was abandoned.

Dismantling of Decorative Elements andCut Stones from Staircases and Altars

Episodes of political unrest in Teotihuacan(particularly around A.D. 250) may have re-sulted in the desacralization of structures andlooting, as in the destruction and conceal-

ment of the Pyramid of the Feathered Ser-pent. When new construction levels werebuilt, walls were torn down, beams werereused, rooms were filled with adobes and de-bris, wall stones were dismantled and reused,burials were reopened, and so forth.

For the end of the city's life, however,many staircases lining the Street of the Deadwere dismantled (Bernal 1963:2.1), particu-larly those of the Plaza of the Moon. Acosta(1964:24, Figure 23) mentions that the cutstones of the staircase of the QuetzalpapalotlPalace were taken out, except for the firststep (Figure 7.3), which had on top of itcarved stones (serpent parts, seashells, coni-cal stones, etc.) from a dismantled altar orstructure. Similar stones, as well as others inthe shape of corncobs and chalchihuites(beads of precious green stone), were foundin the door between the main patio (the Pil-lars' Patio) and Antechamber i of the palace.

In Structure 2 of the Street of the DeadComplex (Matos Moctezuma 1980:86), justto the north of the San Juan River, fragmentsof stucco and carbonized wood were foundon top of the floor of the temple, bearingtraces of the burned roof. Morelos Garcia(1993:66-67) notes that the feline head

96

T H E A B A N D O N M E N T O F T E O T I H U A C A N

Figure 7.4. Shattered braziers and jars on top of Rooms C6 and Cj at Oztoy-ahualco i5B:N6W3. (FromManzanilla 19933; photograph by L. Manzanilla)

sculptures that decorated the main temple'salfarda (stairway limits) at the West PlazaCompound (part of the Street of the DeadComplex) were dismantled and thrown tothe plaza's floor, together with a carved stonefrieze, a feline's paw and arm, and other ob-jects. The same situation occurred in themain courtyard at Xalla, particularly at thebottom of Structure 2 (Manzanilla and LopezLujan 2.001), which also displayed disman-tled feline heads and paws, as well as carvedstones that belonged to a frieze.

In apartment compounds such as Xolal-pan (to the northeast of the Pyramid of theSun), Linne (1934:48) found traces of de-struction of the main central red altar. Theritual objects related to this structure (orna-mented cylindrical and circular plaques,greenstone plaques, theater-type censers, aHuehueteotPs censer, etc.) were found scat-tered between the altar and the staircase inthe eastern platform.

Shattering and Ritual Destruction ofVessels and Sculptures

Many ritual elements have been found on topof the last occupational floors in differentparts of the city. Millon (1988:151) mentions

the violent burning and destruction of thePuma Group (on the Street of the Dead), withthe consequent shattering and scattering of agreen onyx sculpture.

In Structure iD (the northern palace) ofthe Ciudadela many traces of violent destruc-tion were reported (Jarquin Pacheco andMartinez Vargas 1982:103), especially shat-tered Tlaloc vases, Tlaloc disks, Huehue-teotl's braziers, masks, theater-type censers,Mezcala figures, obsidian eccentrics, jadebeads, decorated slate, zoomorphic sculp-tures, candelero concentrations (one of thesehad 160 candeleros [small, two-chamberedceramic devices] in the central plaza of Struc-tures iD), and the like.

The frequent find of Huehueteotl's bra-ziers in the last occupational floors (see theCiudadela northern palace or Xolalpan) hascalled to our attention similar finds in periph-eral apartment compounds, such as the one Iexcavated extensively at Oztoyahualco 156:N6W3 (Manzanilla 19933:108), where simi-lar shattered braziers were found in Room 7.Pottery vessels were also broken on top offloors (Figure 7-4)(Manzanilla :i993a:iO9,Figures 61 and 62).

At Teopancazco (SzEi), an apartment

97

Figure 7.5. Huehueteoti sculpture facedown inRoom Ciy at Teopancazco. (From Manzanilla2002:49; photograph by L. Manzanilla)

Figure 7.7. Restored Tlaloc vase found in Room€25 at Teopancazco. (From Manzanilla 2002:48;photograph by J. Saldana and L. Manzanilla)

compound to the southeast of the CiudadelaI also found a complete Huehueteoti sculp,ture, facedown nearly on top of the floor ofRoom 17, as if thrown down from a smallpedestal (Figure 7.5); its frontal carving hadbeen effaced. A Tlaloc vase was also foundshattered on top of Room 25, with the face ofthe god toward the floor (Figures 7.6 and 7.7)(Manzanilla 1998). It is interesting to notethe same type of behavior in different func-tional sectors of the city.

Similarly, in the northern room of the Pil-lars' Patio of the Quetzalpapalotl Palace atecali jaguar was found on top of the floor onits leftside (Acosta 1964:34).

At Tlamimilolpa, Linne (1942:115) found"a systematic tearing down and breaking upof large and complicated 'incense burners1 toform part of the filling on top of which thenew floors were laid," which may give us ahint of the renewal behavior in each con-struction level.

ViolenceThere are a few hints of violent behavior to-ward the inhabitants of the city. Batres(1906:15) mentioned that in the House of thePriests, near the Pyramid of the Sun, femaleand male skeletons were found in differentpositions, with their stone beads near theirheads. In Structure iD of the Ciudadela, dis-membered bodies were dispersed in the west-ern room and the plaza of Group A (includ-ing a man who died from a blow to the head),as well as articulated limbs (Jarquin Pachecoand Martinez Vargas 1982:103). These areexceptional examples.

Termination and Abandonment RitesIn studying activity areas at Teotihuacan for27 years, I have preferred the interpretationof distributional patterns of archaeological,chemical, paleobotanical, palynological, andpaleozoological data in specific structures(Manzanilla 1985, 1987, 1988-89, 19933,1996; Manzanilla and Barba 1990). Artifactsare mobile, and their location may be dis-rupted in the diverse circumstances related tothe abandonment processes (Cameron 1993),as well as by natural and cultural transforma-

T H E A B A N D O N M E N T O F T E O T I H U A C A N

Figure 7.6. "Killed" Tlaloc vase in Room Cz$ at Teopancazco. (Manzanilla1998; photograph by L. Manzanilla)

Figure 7.8. A "killed" stuccoed tripod found at Teopancazco. (From Man-zanilla 2000; photograph by L. Manzanilla)

tions. Yet chemical compounds (which are aproduct of continuous repetitive activities) instucco floors have been capped in a matrixthat does not allow much modification(Barba 1986). Thus, chemical analyses oftengive us information about past activities,data that are normally lacking in gradualabandonment situations.

Particularly in our anatomy of two Teo-tihuacan apartment compounds (Oztoy-

ahualco i5B:N6W3 and Teopancazco iNW:SiEz), I have detected what seem to be ter-mination and abandonment ceremonies.Termination rituals involve the "killing" ofdifferent types of vessels, mica, figurines,candeleros, lithics, slate, and so on, to endone construction phase before beginning thenext; these rituals were clearly documentedat Teopancazco, where Late Tlamimilolpastructures were replaced and covered by the

L I N D A M A N Z A N I L L A

Figure 7.9. One of the stuccoed tripods fromTeopancazco after restoration. (From Manzanillazooo; photograph by R. Reyes and L. Man-zanilla)

main courtyard (C6) around A.D. 350 (Man-zanilla 2000). Large stuccoed tripods were"killed" and delicately placed in the north-western corner of the main ritual courtyardof the compound (Figures 7.8 and 7.9).

Abandonment rites involve concentra-tions of candeleros, "killed" vessels, ceramicthree-prong braziers, theater-type censers(Figure 7.10), and so on at Teopancazco(Manzanilla 1998), or human mandibles de-posited with ceramic pots and seashells inlayers (each layer separated by a pile of earth)(Manzanilla 1993a: 101, Figure 101). In bothTeopancazco and Oztoyahualco I also found

"killed" stone vessels (i.e., Manzanilla 19933.447, Figure 331).

Palka (this volume) raises the interestingissue of the different reactions of diverse so-cioeconomic groups to the factor that stimu-lates abandonment. At Dos Pilas in the Petenpeople of higher socioeconomic rank mayhave abandoned their houses more quicklybecause they were affected first by warfareand were forced to leave. Lower-rank farm-ers were affected by environmental degrada-tion. At Teotihuacan both factors may havebeen related in the sixth century A.D., wheninternal revolts were occurring against theruling elite at the same time as large-scaleenvironmental degradation due to deforesta-tion and a long-lasting drought. Thus thecore of the city was burned, with much aban-donment-rite debris and de facto refuse inplace, whereas the periphery was abandonedwith some haste. In apartment compounds ofthe northwestern portion of the city I havedetected, for example, stucco mixtures—with basalt and scoria polishers—ready to beused in repairing graves, which were left ontop of the floors when the site was aban-doned (Manzanilla 19933:168).

I would expect that internal revolts andwarfare with other ethnic groups would leave

Figure 7.10. Dismantled theater-type censer found at Teopancazco's maincourtyard, C6. (Photograph by L. Manzanilla)

THE A B A N D O N M E N T O F T E O T I H U A C A N

different patterns. In the case of internal re-volts, even if large-scale rapid abandonmentensued, there might have been some chanceto carry belongings away. That may be thervpe of case that Inomata exposed with re-spect to Aguateca. In ethnic warfare there is aareater possibility for sudden abandonmentand the sacking of remaining valuables. Thispattern is very frequent in Near Eastern ar-chaeology during the Chalcolithic and BronzeAges (Manzanilla 1986).

LootingOffering pits are consistently sacked through-out Teotihuacan. Armillas (1944) remarkedthat in the Viking Group all the offerings hadbeen looted, and in one case this was doneduring Teotihuacan times, because new floorshad been constructed on top. In general,owing to the fixed position of the offerings infront of the staircases or in the corners, it isclear that they were sacked in Prehispanictimes. Some looters' tunnels were also dug inthose centuries (Bernal 1963:31). In theQuetzalpapalotl Palace large looting pitswere detected, many of which bore frequentexamples of carved stones from the Pillars'Patio (Acosta 1964:32).

Reoccupation or Continued OccupationAcosta (1972:149) believed that the violentdestruction of the ceremonial core ended theurban history of the site but that some groupscontinued living in the apartment com-pounds in the periphery for two or threemore generations. In many places one findslarge amounts of Coyotlatelco pottery in thedebris of the last Teotihuacan constructions,such as the Street of the Dead, Tetitla, andAtetelco (Armillas 1950; Bernal 1963); whenthe elite apartment compounds were totallyruined, the Mazapa people buried their dead

on top of the debris (Armillas 1950; Linne1934, 1942).

From our explorations in quarry tunnelsimmediately to the east of the platform of thePyramid of the Sun, one may conclude thatthe Coyotlatelco people were living in Teoti-huacan in the sixth century A.D. They wereburying their dead in these tunnels, and theywere certainly looting Teotihuacan ritualcontexts (Manzanilla et al. 1996). Perhaps,as with the Akkadian people living in the out-skirts of the Early Dynastic cities of Sumer,they participated in the collapse of the Teoti-huacan way of life and continued living in thepartially deserted city. I suspect that therewas not a continuous Coyotlatelco settle-ment in the city but perhaps clusters of peo-ple looting and dismantling former Teotihua-can buildings. There is scant evidence ofCoyotlatelco structures and activity areas(some of which have been dug by my team inthe tunnels to the east of the Pyramid of theSun). There is also some evidence that smallCoyotlatelco groups were living on top of thelast Teotihuacan floors in some of the apart-ment compounds. But there are no datawhatsoever of large Coyotlatelco settlementsand constructions, as have been found inTlalpizahuac, for example (Tovalin Ahumada1992).

CONCLUSIONSTeotihuacan, with its 20 km2, is still a vastfield of inquiry with respect to its last years ofexistence. In this chapter I have tried to givean idea of the recurrence of behavior in ritualand domestic contexts in the city. Yet muchwork remains to be done to elucidate whenand why the vast Classic metropolis col-lapsed, and who was involved in this extraor-dinary event.