Civic and Residential Settlement at a Late Center Author(s): Vernon L. Scarborough and Robin A. Robertson Source: Journal of Field Archaeology, Vol. 13, No. 2 (Summer, 1986), pp. 155-175 Published by: Boston University Stable URL: http://www.jstor.org/stable/530218 Accessed: 18/10/2010 15:48

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=boston.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

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

Boston University is collaborating with JSTOR to digitize, preserve and extend access to Journal of Field Archaeology.

http://www.jstor.org Civic andResidential Settlement at a Late PreclassicMaya Center

VernonL. Scarborough Universityof Texasat El Paso

Robin A. Robertson SouthernMethodist University Dallas, Texas

The Late Preclassic period in the Maya Lowlands (300 B.c.-150 A.c.) docu- ments the transition toward increased social and economic complexity culmi- nating in the Classic (250-900 A.C.). The Late Preclassic Maya community of in northern has revealed a settlement pat- tern of dispersed household clusters and scattered public architecture. More- over, the site manifests a clear, three-part concentric zonation, similar to later Classic period communities. The authors' analysis provides a definition through time of civic and residential architecture and of the division between elite and non-elite domiciles. The study draws heavily on a functional analy- sis of the excavated ceramic assemblage. The unique settlement pattern of the semitropical Maya is suggested to be an environmental adaptation with rural elites coordinating the dispersed sustaining population through public monuments and associated ritual.

Introduction munity did not change radicallyfrom that of its Late The Late Preclassicperiod in the Maya Lowlandshas Preclassic antecedents. Although public and private been viewed as a transitionalperiod in which levels of building styles were alteredsignificantly through time, social complexitydeveloped from chiefdomsinto state- the overall settlementpattern was always composedof like institutions.'This transitionin other areas of the dispersedhousehold clusters. At the sites of ,3Dzi- world has involved markedchanges in settlementform bilchaltun,4Altar de Sacrificios,5Seibal,6 Barton Ra- and density,giving rise to nucleatedcities.2 In the Maya Lowlands,however, the of the com- spatialorganization 3. W. R. Coe, Tikal: a Handbook of the Ancient Maya Ruins (Uni- versityMuseum, University of Pennsylvania:Philadelphia 1967); D. 1. R. E. W. The Civilization Adams, ed., Origins of Maya (Univer- E. Puleston, The Settlement Survey of Tikal. Tikal Report 13, Uni- of New Press: W. R. sity Albuquerque1977); Coe, "Tikal, versity Museum Monographs (University of Pennsylvania: Philadel- and Science 147 , EmergentMaya Civilization," (1965) phia 1983); R. F. Carrand J. E. Hazard,Map of the Ruinsof Tikal, D. A. "CultureAreas and Interaction 1401-1419; Freidel, Spheres: El Peten, Guatemala. Tikal Report 11, University Museum Mono- to the of Civilizationin the ContrastingApproaches Emergence Maya graphs (Universityof Pennsylvania:Philadelphia 1961). Lowlands,"AmAnt 44 (1979) 36-54. 4. E. B. Kurjack, Prehistoric Lowland Maya Community and Social 2. W. T. Sanders,J. R. Parsons, and R. S. Santley, The Basin of Organization. Middle American Research Institute Publication 38 Mexico: Ecological Processes in the Evolution of a Civilization (Ac- (TulaneUniversity: New Orleans1974); E. W. AndrewsV, "Dzibil- ademicPress: New York K. V. andJ. 1979); Flannery Marcus,eds., chaltun," in J. A. Sabloff, ed., Supplement to the Handbook of Middle The Cloud Evolution the People: Divergent of Zapotec and Mixtec AmericanIndians 1 (Universityof Texas Press: Austin 1981) 313- Civilizations(Academic Press: New York 1983); R. M. Adams, 344. Heartlandof Cities (Universityof Chicago Press: Chicago 1980); 5. G. R. and A. L. The Ruins Altar de Kwang-Chih Chang, The Archaeology of Ancient China (Yale Press: Willey Smith, of Sacrificios. No. New Haven 1977); B. J. Kemp, "TheEarly Development of Towns PapPeaMus62, 1 (HarvardUniversity: Cambridge 1969); A. L. Excavations at in Egypt,"Antiquity 51 (1977) 185-200; M. E. Moseley and K. C. Smith, . PapPeaMus 62, No. 2 Day, eds., Chan Chan: Andean Desert City (University of New (HarvardUniversity: Cambridge 1972). Mexico Press: Albuquerque1982); B. Allchin and R. Allchin, The 6. G. R. Willey, A. L. Smith, G. TourtellotIII, and I. Graham, Rise of Civilization in India and Pakistan (Cambridge University Excavations at . MemPM 14, No. 1 (Harvard University: Cam- Press:Cambridge 1982). bridge 1975); G. Tourtellot,"The Peripheriesof Seibal: an Interim 156 Late Preclassic Maya Center/Scarborough and Robertson

mie,7 ,8Colha,9 and the Lake Yaxhaarea,'0 there was little majorsettlement reorganization after the Late MEXCO Santa Elen MEXICO Preclassicdespite an increasein the size and density of these communitiesduring the Classic period. Nucleated \?, I rlSantaRita cities were the exception ratherthan the rule." Given A CERO Bay GUATEMALA?, AventurCCE the sociopolitical heights to which the Maya were to aspire, the absenceof a spatiallywell-controlled consti- SBenqueViejo tuency has mistakenlycontributed to the notion of the iNohmul "mysterious"character of Maya civilization. MEXICO hipstern The communityof Cerrosin present-daynorthern Be- SonLuis Sn Antonio lize (FIG.i) has recently revealed a settlementpattern datingto this transitionalperiod.12 The site is uniquein been a Late Preclassic with little having community ElPbsito Kic6on- subsequentClassic and Postclassic construction.Con- sequently, the data permit a closer inspection of the 1. of Belize. transitionand provide a baseline for assessing the sub- Figure Map northern tleties involved in the developmentof later centers. bulk of the architectureat this early date, and they are The dichotomy between civic monumentalarchitec- clusterednear the present-daybay. Divisions between ture and residentialarchitecture has providedthe best elite and non-elite seem not to have been very great. evidence for shifts in settlementwhich did occur during During the late facet of the Late Preclassic, the civic the LatePreclassic. The earlyfacet of the LatePreclassic architectureincreased in size, frequencyof appearance, at Cerrosmanifests a relativelysmall portionof the total and complexity;though there was a slight closure of public architecturefound at the site. The little monu- space within a well-definedcentral precinct, a consid- mentalarchitecture that does occur is associatedwith an erableamount of civic architecturewas constructedaway open centralprecinct. Residential structures make up the from this zone. Although the growing elite may have deliberatelybegun delimiting the central precinct and thereby made it less accessible, the remainderof the in R. and Report," W. Bullard, ed., Monographs Papers in Maya communityappears to have incorporateda more open Archaeology.PapPeaMus 61 (HarvardUniversity: Cambridge 1970) 405-421. display of civic architecture.Outside the central pre- cinct, access the to activitieswas 7. G. R. et Prehistoric Settlement in the Belize by people manypublic Willey al., Maya maintained.The constructionof a 1200-m Valley. PapPeaMus 54 (Harvard University: Cambridge 1965). curvilinear, canal and its apparentfunction as a territorialmarker 8. P. M. Thomas, Jr., Prehistoric Maya Settlement Patterns at Becan, a of controlledaccess for the Campeche, Mexico. Middle American Research Institute Publication provided degree greater 45 (TulaneUniversity: New Orleans1981); D. L. Webster,Defensive core zone, but the scatteredpattern of civic architecture Earthworks at Becan, Campeche, Mexico. Middle American Research withinthis residentialzone was well maintained.Despite InstitutePublication 41 (TulaneUniversity: New Orleans1976). the increasingdifferentiation of elites and non-elites,the 9. H. J. Shaferand T. R. Hester,"Ancient Maya ChertWorkshops two groups sharedthe core zone definedby the canal. in NorthernBelize, CentralAmerica," AmAnt 48 (1983) 519-543; J. Before the implicationsof this settlementdesign are in D. Eaton, "Colha:an Overviewof Architectureand Settlement," discussed,a descriptionof the settlementpattern and the T. R. Hester, H. J. Shafer, and J. D. Eaton, eds., Archaeologyat thatare not Colha, Belize: the 1981 Interim Report (Center for Archaeological data-gatheringmethodologies fully presented Research,University of Texas: San Antonio 1982) 11-20. elsewhere'3seems appropriate. 10. D. S. Rice and P. M. Rice, "The NortheastPeten Revisited," AmAnt 45 (1980) 432-454. Survey Design 11. Dense prehistoricpopulations comparable to those defined in Three kinds of survey were employed on Lowry's HighlandMexico are reportedfor the two late periodsites of Chun- Bight and the adjacentland. Beginningwith reconnais- chucmiland Mayapain.D. T. Vlcek, S. GarzaT., and E. B. Kurjack, sance, each level of surveyreduced the amountof terrain "ContemporaryFarming and Ancient Maya Settlement:Some Dis- examined but increasedthe and of D. Harrisonand B. L. Turner accuracy intensity concertingEvidence," in P. II, eds., control. Prehispanic Maya Agriculture (University of New Mexico Press: survey Albuquerque1978) 211-224; H. E. D. Pollock, R. L. Roys, T. Proskouriakoff, and A. L. Smith, Mayapdn, Yucatdn, Mexico. 13. Ibid.; Freidel,loc. cit. (in note 1); idem, "Civilizationas a State D.C. CarnlnstPub619 (Washington, 1962). of Mind: the CulturalEvolution of the LowlandMaya," in G. D. 12. V. L. Scarborough,"A PreclassicMaya Water System," AmAnt Jones and R. R. Kautz, eds., The Transition to Statehood in the New 48 (1983) 720-744. World(Cambridge University Press: Cambridge 1981) 188-227. Journal of Field Archaeology/Vol. 13, 1986 157

Four iMile Lagoon

Aventura COROZALBAY

New River

Cerros St •

Puerto Nuevo Saltilloo Sa tillo

Esperanza

Cudjoe

Lagoono \a4 P~Point Aleg re San Antonio al( Chunox ateRamonal/

Cocos Braot Lagoon B Pond

0 km 5 Km. 1 , a , l ..... I

Figure2. Mapof Lowry'sBight and adjacent archaeological sites. The size of theblack dot reflectsthe estimated size of thesesites relativeto Cerros.The SAR readout,provided by R. E. W. Adams,was subjectivelyfiltered for tightlattice and dendriticpatterns. Survey data were collected by reconnaissancemethods.

Reconnaissance tions. The side-looking airbornereadout (SAR) is be- lieved to representancient canals of uncertaindate.14 Broad reconnaissanceof the 10 sq km defining the penisulawas conductedby land, sea, and air. Additional site-orientedsurvey was carriedout nearthe immediate SystematicSurvey Area of the these involved margins bight. Although operations The second of was conductedaround and little more than a camera,a and kind survey compass, graphpaper, within the site zone of Cerros. no formalex- a sharpmachete, they providedthe interpretivebackdrop Although for the more systematicprogram. Figure 2 illustratesthe numberand location of LatePreclassic sites nearCerros. 14. V. L. Scarborough,"Raised Field Detectionat Cerros,Northern for the Postclassic of on Belize," Drained Field Agriculture in Central and South America. Except community Esperanza 189 R. E. W. the side of the all of BARSupp (Oxford 1983); Adams, W. E. Brown, Jr., opposite bight, the sites have a Late andT. P. Culbert,"Radar Mapping, Archaeology, and AncientMaya Preclassic component, based on their ceramic collec- Land Use," Science 213 (1981) 1457-1463. 158 Late Preclassic Maya Center/Scarborough and Robertson

5

132

N -!*107 1140

S5ocbe V 77 w37 .2627 "77 w64 -9770UIV 37 v27

922m 672 4190 Ca34 96?9

131" 3 Au i ,129

S139 Aguad o

14869 roblemtic Mounds 93

guo

Figure3. Mapof Cerros. cavationswere done in this area, a systematictransit and community. Although the transects were minimally alidadesurvey was employedto accuratelymap structure 200 m wide, the presenceof old bush trails, the undu- density as well as environmentalrelationships. The 1.51 lating edge of the coastline, and our attemptsto reduce sq km areaenclosed by the surveytract involved clearing "boundaryeffects" in the sample,'"increased their width over 26 sq km of brecha, or bush trails, duringwhich by as much as 100 m. 181 moundedfeatures were uncovered. Transect1 extended800 m fromthe systematicsurvey The dense secondaryvegetation on the bight neces- and excavationarea along the coast. It was designedto sitated the use of an on-the-groundgrid system cut examine the specific functionof the coastline as a site throughthe foliage and across the site. The north-ori- boundary.Was it a civic or residentialboundary and ented, 100-m intervalgrid providedthe frameworkfor what economic implicationswould follow (see below)? structurelocation. The surveyteam withineach circum- Transect2 was designed to bisect the long axis of the scribedhectare surveyed the bush on foot, separatedby bight. It was hypothesizedthat if landwardintruders 10-15 m. When a structureor related featurewas lo- were a threat,a defensivefeature might be exposed near cated, an alidade map was renderedand tied into the the base of the peninsulasomewhere along the transect. grid system (FIG.3). Reconnaissance,however, suggested otherwise. After The boundaryof the systematicsurvey zone was ar- locating only four house moundsin the area systemati- bitrarilydefined. The area enclosed was comparableto cally surveyed,which revealed a cleardrop-off in mound that defined by the systematic survey and excavation density from the center of the communitytoward the zone aroundthe site. In additionto the contiguoussouth- edge, it was decided to terminatethe transect. ern block, three transectswere cut into the interiorof the bight, away from the site center. These transects 15. I. Hodder and C. Orton, Spatial Analysis in Archaeology (Cam- were chosen to help discern the outer boundaryof the bridge University Press: Cambridge 1976) 41. Journal of Field Archaeology/Vol. 13, 1986 159

m was directed toward the :en Transect 3, 900 long, large Precinnag? Aguada 2 (or pond), near the opposite side of the bight. Reconnaissance had indicated an early facet of Late Preclassic occupation at this location (Ixtabai, 300-200 B.C.). The transect was positioned to assess how discrete Periphery these two locations might have been and to determine the density of occupation in the vicinity of the aguada. 0 SystematicallySurveyedArea

Systematic Survey and Excavation Area The final kind of survey involved no change in the Transect2 Transect 3 actual research design but did add the dimension of controlled excavation. This area encompassed 75 ha, half the total area enclosed by the systematic survey and is the focus of the remainder of this area, paper. Figure 4. Map of Cerros zonation. Settlement Pattern in Excavated Area The three zones of civic and residential settlement at houses in this area during the later part of the C'oh phase Cerros have been revealed in the systematic survey and (100-50 B.C.). A density figure of 14.3 houses per ha excavation area (FIG. 4). Extending outward from the has been derived from these data. During the Tulix phase central precinct of monumental architecture, each zone (50 B.C.-150 A.C.), the village nucleus was buried by the represents a larger concentric area than does the zone construction of a two-meter-thick raised plaza, on which before it. Structure density and structure size decrease additional monumental architecture was constructed. correspondingly. Most of the grand architecture at Cerros is concen- A three-ring volumetric zonation is also apparent at trated in the central precinct, with over 133,000 cu m Cerros, suggesting the concentric zonation model (see of rubble fill used in the construction of these buildings. below). The amount of fill necessary for the construction Structure 4 is the largest single edifice, having basal of a mound is understood to be one index of energy dimensions of 58 m x 68 m and rising to a composite expenditure. Sixty-three percent of all architecturalvol- height of 22 m. The fill volume of this one structure is ume occurs in the 6 ha defining the central precinct. 32,038 cu m. Approximately 22,184 cu m of fill per ha was used in the construction of the central one-tenth precinct. Only Core Zone of this architectural volume per ha was manifest within the core zone, excluding the central precinct, and only Surrounding the buried village nucleus and the central one-hundredth of this architectural volume per ha was precinct of the site is a core zone of compact residential defined in the area outside the canal. and civic structures. The arrangementof public and pri- vate space is less ordered than the planned appearance of the central The orientation of these struc- Central Precinct precinct. tures, however, suggests that the buildings were con- The central precinct,16 or center zone, covers 6 ha of structed in a radial/radiating pattern focusing on the raised plaza and associated monumental architecture. It central. precinct. This pattern is particularly evident in was the most densely built-up area of the community, the western portion of this zone. In addition, the two encompassing both monumental architecture and a ballcourts, Structure Groups 50 and 61, are positioned nearby village nucleus. Excavations by Cliff"17 in the along a broad N-S medial axis that bisects the core zone village nucleus indicate there were 86 ground-level (FIG.3). They appear to have been deliberately located with respect to the isolated pyramid, Structure 29. The 16. An examination of the central precinct can be found in Freidel, westward medial axis of that structureintersects a cit. note "Maritime and the Rise of point loc. (in 1); idem, Adaptation from either and the Maya Civilization: the View from Cerros," in B. Stark and B. Voor- approximately equidistant ballcourt, hies, eds., Prehistoric Coastal Adaptations in (Aca- north and south ends of its summit structures appear to demic Press: New York 1978) 239-265. be oriented toward Groups 61 and 50, respectively."8 17. M. B. Cliff, "Lowland Maya Nucleation: a Case Study from 18. V. Scarborough et al., "Two Late Preclassic Ballcourts at the Northern Belize," unpublished Ph.D. dissertation, Southern Methodist Lowland Maya Center of Cerros, Northern Belize," JFA 9 (1982) University (Dallas 1982). 21-34. 160 Late Preclassic Maya Center/Scarborough and Robertson

The outer boundary of the core zone was clearly de- Table 1. Structuretypes in the systematicallysurveyed and limited by the 1200-m curvilinear canal, reflecting the excavatedsettlement area. radial of the from the central dispersion community away Type Description Frequency the Tulix phase, the core zone extended precinct. During 4 or more 2 over 34 ha and contained a of 1.21 1 buildingplatforms population density on a sharedsubstructure individual house mounds ha. per 2 3 buildingplatforms on a 4 sharedsubstructure 3 2 on a 4 Periphery buildingplatforms sharedsubstructure The peripheral zone is defined by the area outside the 4 Substructuregreater than 150 18 main canal and within the arbitrarylimits of the survey sq m and more than I m tract. The locations of structures in this zone do not high appear to have been the result of any coordinated plan, 5 Substructuregreater than 150 28 m and few civic structures were observed. A marked drop- sq and less thanor equal off in the of structures occurs in the interior of to 1 m high density 6 Substructureless than 150 34 the to the south Transects 2 and bight, especially along m and less thanor 3. The distribution of structures the coast- sq equal along present to 1 m high line, however, indicates a denser occupation than in the interior. Such a settlement pattern may have been pro- duced by a service population engaged in petty exchange or maritime resource extraction outside the immediate Table2. Structuretype frequencyand exposurein systematicallysurveyed and excavatedsettlement area. control of the central precinct.19 The area examined within the peripheral zone encom- Total passes 35 ha and is comparable to that in the core zone. Number Percentage Exposure/ The density of house mounds in this tract is 0.40 indi- Type Frequency Tested Tested Stratum vidual structures per ha during the Tulix phase. 1 2 2 100 451 sq m 2 4 4 100 269 sq m 3 4 3 75 101 sq m Structure Typology 4 18 13 72 80 sq m* 5 28 6 21 74 sq m From a of view, the focus of the methodological point 6 34 7 21 45 sq m survey has been on spatial variability within and between *Excludes extensive on Structure 29B. mound groups. To deal with this variability, a working exposure morphological typology of mounded features using form, size, and groupings was devised for the settlement complicated mound grouping may be seen as a reflection zone. Tables 1 and 2 present the classification that was of his rank in the community. In addition, intrasite dis- developed, due in part to the absence of standing ma- tance relationships were also expected to reveal elements sonry in the settlement. The settlement typology and of social organization related to the clustering of mound excavation program included only the core and periph- groups.20 eral zones. Underlying this typology was the implicit assumption The typology was used to stratify the sample of that the majority of mounds served a residential function. mounds around the center and to provide the basis for Recent information, however, suggests that this assump- selecting particular mounds for excavation in order to tion must be reevaluated (TABLES3-5). Most of the large investigate social organization. It was hypothesized that and formally complex mounds within the core zone are the size and form of these mounds and mound groups now considered to be monumental civic architecture. would correlate roughly with the status of their occu- Even though the functional justification for stratifying pants. Regardless of whether another family resident, the mound excavation sample must be reassessed, it ancestor, or relative of non-resident status was respon- should be noted that the typology still stands on. mor- sible for the initial construction of the mound (as opposed phological grounds. Consequently it may prove appro- to the mound's occupant being responsible for its con- struction), the individual occupant's ability to elevate 20. After Kurjack, loc. cit. (in note 4); E. W. Andrews V et al., himself above others by residing on a larger or more ": an Early Maya Community in Northwest Yucatan," paper presented at the Sociedad M6xicana de Antropologia, in San Cristo- 19. Freidel, loc. cit. (in note 16). bal, , 1981. Journal of Field Archaeology/Vol. 13, 1986 161

Table 3. Civic and residentialconstruction and occupationfrequencies through time (excavatedsample). Early Late MoundCount* C'oh Tulix Classic Postclassic TotalMound Construction (52) 11 21% 33 62% 6 12% 2 4% TotalMound Occupationt (52) 11 21% 13 25% 19 37% 8 15% CombinedTotals (52) 22 42% 46 88% 25 48% 10 19% Civic MoundConstruction (20) 4 20% 13 65% 3 15% 0 0% Civic MoundModification (20) 0 0% 4 20% 0 0% 0 0% Civic ConstructionInside Canal(17) 2 12% 12 17% 3 18% 0 0% Civic ConstructionOutside Canal (3) 2 67% 1 33% 0 0% 0 0% ResidenceConstruction Inside 7 35% 12 60% 0 0% 1 5% Canal(20) ResidenceConstruction Outside 0 0% 8 67% 3 25% 1 8% Canal(12) TotalResidence Occupation Inside 11 30% 19 51% 12 32% 8 22% Canalt (37/71%) TotalResidence Occupation Outside 7 47% 8 53% 10 67% 2 13% Canal$(15/29%) Area Count* Total MoundConstruction (35) 9 26% 21 63% 3 9% 2 6% TotalMound Occupationt (35) 6 17% 8 23% 9 26% 6 17% CombinedTotals (35) 15 43% 30 86% 12 34% 8 23% Civic MoundConstruction (11) 2 18% 8 73% 1 9% 0 0% Civic MoundModification (11) 0 0% 3 27% 0 0% 0 0% Civic ConstructionInside Canal(10) 1 10% 8 80% 1 10% 0 0% Civic ConstructionOutside Canal (1) 1 100% 0 0% 0 0% 0 0% ResidenceConstuction Inside 7 39% 10 56% 0 0% 1 6% Canal(18) ResidenceConstruction Outside 0 0% 3 50% 2 33% 1 17% Canal(6) TotalResidence Occupation Inside 11 39% 15 54% 6 21% 6 21% Canalt (28/80%) TotalResidence Occupation Outside 2 29% 3 43% 5 71% 2 29% Canalt (7/20%) *Occupationpercentages need not add up to 100%due to thereoccupation of moundsthrough time. tExcludescoeval construction loci. tExcludesutilized civic structures. Parenthesesindicate total possible within each category. Percentages refer to totalwithin the systematicsurvey and excavation area.

priate for other sites where the principal objective of a a stratified judgmental sample.21 At minimum, a 20% typology is the stratificationof the basic unit of analysis: sample was taken from each of the typological divisions mounded features. in the mound typology (TABLE2). The typological divi- sions with fewer mounds, however, were sampled more Excavated Structures extensively. Specific mound selection within each type This partof the programinvolved a series of 4 sq m was determined by the desire to maintain a dispersed test excavations dispersed throughout the 75 ha imme- geographical representation of mounds and by fortuitous diately surrounding the center. Moreover, in order to trash exposures revealed by natural agents (tree falls, place the formal typology on a strong interpretive foot- erosional shoreline profiles, etc.). The judgmental sam- ing, extensive lateral exposure was carried out on at least one structure from each of the six structure in the 21. C. L. Redman,"Multivariate Artifact Analysis: a Basis for Mul- types tidimensional in C. L. settlement. Excavation and included Interpretations," Redman,M. J. Berman,E. recovery techniques V. Curtin,W. T. Langhorne,Jr., N. M. and J. C. Wanser, the of natural and architectural strata. Versaggi, screening eds., Social Archeology: Beyond Subsistence and Dating (Academic Mound groups were selected for test excavation from Press:New York 1978) 167. 162 Late Preclassic Maya Center/Scarborough and Robertson

Table4. Civic and residentialdensity figuresthrough time (extrapolatedfrom excavationsample to visible moundpopulation). Early Late MoundCount* C'oh Tulix Classic Postclassic

No. Density No. Density No. Density No. Density TotalMound Construction (108) 23 0.33 68 0.99 13 0.19 4 0.06 TotalMound Occupationt (108) 23 0.33 27 0.39 40 0.58 16 0.23 CombinedTotals (108) 45 0.65 95 1.38 52 0.75 21 0.30 Civic Mounds& Plaza Construction(25) 5 16 4 0 Civic Mounds& Plaza Modification(25) 0 5 0 0 Civic ConstructionInside Canal(22) 3 16 4 0 Civic ConstructionOutside Canal (3) 2 1 0 0 ResidenceConstruction Inside 21 0.62 35 1.03 0 3 0.09 Canal(59) ResidenceConstruction Outside 0 16 0.46 6 0.17 2 0.06 Canal(24) TotalResidence Occupation Inside 24 0.71 41 1.21 26 0.76 18 0.53 Canalt (81) TotalResidence Occupation Outside 13 0.37 14 0.40 18 0.51 4 0.11 Canalt (27) Area Count* TotalMound Construction (90) 23 0.33 57 0.83 8 0.12 5 0.07 TotalMound Occupationt (90) 15 0.22 21 0.30 23 0.33 15 0.22 CombinedTotals (90) 39 0.57 77 1.12 31 0.45 21 0.30 Civic Mounds& Plaza Construction(16) 3 12 1 0 Civic Mounds& Plaza Modification(16) 0 4 0 0 Civic ConstructionInside Canal(15) 1.5 12 1.5 0 Civic ConstructionOutside Canal (1) 1 0 0 0 ResidenceConstruction Inside 22 0.65 31 0.91 0 3 0.09 Canal(56) ResidenceConstruction Outside 0 9 0.26 6 0.17 3 0.09 Canal(18) TotalResidence Occupation Inside 28 0.82 38 1.12 15 0.44 15 0.44 Canalt (71) TotalResidence Occupation Outside 6 0.17 8 0.23 13 0.37 6 0.17 Canalt (19) *Density= count/ha. tExcludescoeval construction loci. tExcludesutilized civic structures. Parenthesesindicate total possible within each category.

pie was developed to take advantageof known surface Testing Program achieve maximumdata retrieval indicatorsin order to The of the test excavation program in the not purpose for minimumlabor costs. Althoughthe samplewas settlement was threefold. First, it was to establish chron- controlled environmental rigidly geographically,every ological control. This was accomplished primarily by and sector of the was tested.22 spatial community penetrating and examining the contents of the mound. A sealed dating context was defined as primary material 22. Scarborough,loc. cit. (in note 12) discussesmicroenvironments capped by an impenetrable layer of flooring or thick at Cerros.It shouldbe notedthat the decisionto test a largepercentage plaster melt. Unfortunately, few burial or cache offerings of 4 structureswas a one. Because was Type pragmatic preservation were found. Primary trash was identified as such by the in this mound and in 5 and 6, a generallygood type poor Types of axis than 10 cm), uner- greaternumber of Type 4 moundswere excavatedin an attemptto presence large (long greater furtherrefine our architecturaltypology. oded sherds from a single time period. Journal of Field Archaeology/Vol. 13, 1986 163

Inside Outside Table5. Numberof areasand moundsexcavated and surveyedin Civic Residential Civic Residential Total the settlement,inside and outsidethe Total Number 10 18 1 6 35 core area. of Areas Excavated TotalNumber 17 20 3 12 52 of MoundsExcavated Total Number 15* 56t 1* 18t 90 of Areas Visible Total Number 22* 59t 3* 24t 108 of MoundsVisible Total SurfaceArea 34 35 69 in Hectares (excludingcenter) *Presumedcivic tPresumedresidential

Although such material would generally be classified cavation exposure. Unfortunately, the quality and quan- as midden or habitation debris at other sites, this as- tity of architecture revealed in the limited exposures did sumption was not made at Cerros because bone, char- not produce the correlations with the morphology of the coal, and other domestic debris such as manos, metates, mounds prior to excavation necessary to modify the and other stone tools or debitage were sometimes absent typology itself. from the deposits. Moreover, in some instances, the The third focus of the testing program was to deter- ceramics were deposited as part of a termination ritual mine the function of structures. It was anticipated that marking the abandonment of civic structures,23and ac- certain kinds of information would be culled from our tually represented offerings ratherthan habitation debris. limited exposures which could be compared with the Because Late Preclassic pottery is easily broken into quality and quantity of information collected from our small bits (long axis of sherd less than 2 cm) and eroded lateral stripping operations. In meeting this goal, the when exposed, trampled, or transportedand redeposited, ceramics recovered proved to be most useful. sherd size and erosion could be used to detect these processes. In these situations chronological homogeneity Functional Analysis of Ceramics of the excavated sherds, while always utilized, became a critical factor in determining whether or not the deposit The functional analysis of the pottery from Cerros24 was primary. Although deposits subjected to extended indicated that the distribution of Late Preclassic ceramic exposure prior to burial were not used to directly date types within a given context reflects definable social and mound construction, some deposits containing "freshly" behavioral phenomena. As a part of that study, seven redeposited sherds were. In all cases the redeposited excavated contexts containing well-defined primary hab- material postdated the C'oh Phase, the terminus post itation debris and trash were assigned a function. This quem of interest to this study. was based on location within the community and degree The second intent of the testing programwas to expose of elaboration in the associated features and artifacts. any architectural features that would aid in describing The types were then plotted against these contexts. Table the structuresfrom the perspective of a refined settlement 6 illustrates the results for the two complexes of interest typology. Exposure of well-preserved architecturein the in this study. It should be noted that the two very rare settlement was also designed to aid in the identification types (Bribri Black Incised and Unslipped: Bribri Vari- of those mounded types deserving significant lateral ex- ety; and Iguana Creek White: Variety Unspecified) were eliminated from consideration because no more than 20 sherds of these types have been recovered at Cerros to 23. R. A. "Functional Robertson, Analysis and Social Process in date. Table 6 shows that the occurrence and relative Ceramics:the Potteryfrom Cerros,Belize," in R. M. Leventhaland A. L. Kolata, eds., Civilization in the Ancient Americas: Essays in Honor of Gordon R. Willey (University of New Mexico Press: Al- buquerque1983) 105-142; J. F. Garber,"Patterns of JadeConsump- 24. R. A. Robertson-Freidel,"The Ceramicsfrom Cerros:a Late tion and Disposalat Cerros,Northern Belize," AmAnt 48 (1983) 800- PreclassicSite in NorthernBelize," unpublishedPh.D. dissertation, 807. HarvardUniversity (Cambridge 1980). 164 Late Preclassic Maya Center/Scarborough and Robertson

Table6. Functionalassessment of ceramics.

C'oh Deposits TulixDeposits Interpretation of Type Operation Number 1 33 34e 107 34w 1 33 34 35 111 34w Function Assessed D/NE D/NE D/E D/E R D/NE D/E D/E R R R by Excavation Evaluation Chiculte Slipped Rim c c c c Domestic Striated Bobche Smudged c c c c Domestic Crabboe Washed c c c c Domestic Sierra Red: Society Hall c c c c Domestic Canop Red-on-Red Trickle c c c Elite Cockscomb Buff c c Elite Teabox Unslipped c c c c c Domestic Canxun Red c c c c c Domestic Lanillo Groove Incised c c c c c Domestic Xaxnik Through the r r r r c Spec. Occ. Slip Incised ZorraBlack-on-Red r r r r m Spec. Occ. MatamoreDichrome: r r r r m Spec. Occ. Matamore TintaUsulutan r r r r m Spec. Occ. HokabImpressed c Ritual Hole Dull: Hole c Ritual Cassada Red-over-Black c Ritual Paila Unslipped: Non-elite Unspecified c c c Domestic Sapote Striated: Chacah c c c c m c c c m m m Domestic Chamah Washed c c c c c c c Domestic Sierra Red: Xaibe c Non-elite Other Forms SierraRed: Xaibe Form c c c Domestic SapoteStriated: Chichem Variety c c c m m m Domestic CabroRed c c c m m m Domestic LiscanalGroove Incised c c c m m m Domestic PahotePunctated c c c m m m Domestic Tuk Red-on-Red Trickle c c m m m Elite Nictaa Buff c c m m m Elite Savannah Bank Usulutan: Unspecified r r r m m m Spec. Occ. Pixoy Usulutan r r r m m m Spec. Occ. Sangre Red r r r m m m Spec. Occ. MufiequitaAppliqued m m m Ritual TacisteWashed m m m Ritual KuxcheOrange m m m Ritual RemaxPunctated m m m Ritual MatamoreDichrome: m m m Ritual Shipyard Chatoc Dichrome m m m Ritual Ritual Zapatista Tickle-on-Cream m m m Brown: Zoon Hole Dull: Hukup m m m Ritual

c = commonoccurrence r = rareoccurrence E = elite R = ritual m = presentin moderatefrequencies D = domestic NE = non-elite Journal of Field Archaeology/Vol. 13, 1986 165 frequencyof a type withina given contextappears to be Table 7 providesthe initial constructiondata and the relatedto the natureof the contextrather than to chron- two functionalassessments of the structuresexcavated ological factors. Subsequentanalysis of other contexts at Cerros.The formalfield identificationis basedon the has repeatedlyconfirmed this observation. criteriadiscussed below. In 23 of the 36 cases, the formal Using the resultingpatterns, each type was assigned field identificationand the ceramicassessment were the to one of five functionalclasses-general domestic,non- same. In 14 of these 23 cases, however, the ceramic elite domestic, elite domestic, ritual, and special occa- data permitteda more precise functionalinterpretation sion. It was notedthat general domestic pottery is found of the structure.On the otherhand, in 13 of the 36 cases in all domestic contexts. Non-elite domestic potteryis the ceramic data were of no help either because the restrictedto non-elite contexts. Elite domestic pottery, samplesize was too smallor becausethe initialconstruc- however, is found not only in elite domestic contexts tion of the mounddated to the EarlyClassic or the Late but in ritual ones as well. The same is true of the Postclassic. Unfortunately,given the condition, sample dominantred wares and some of the utility wares that size, and stage of analysisof the ceramicmaterial from have been classifiedas generaldomestic pottery. these later time periods, similarinterpretive reconstruc- In the initial study, the analyzedritual contexts were tion is not yet possible. restrictedto deposits associatedwith the abandonment of monumentalarchitecture. Later analyses, which in- cluded caches, burials, and civic structures,indicated ResidentialStructure Identification: Architecture and that such ceramictypes also occurredin those contexts, necessitatinga broadeningof the class to ritual/civic SpatialRelationships pottery.In contrastto the elite and the generaldomestic The Cerrossettlement data have been used to treatthe types, potteryof this kind is found only in ritualand/or most elusive andrecurrent problem in settlementpattern civic contexts. analysis:the identificationof a house. The problemhere Special occasion types could have been interpretedas is perhapsmore extreme because of the poorpreservation raregeneral domestic types, but withinthe ethnographic of ruins and the absence of dependableethnohistoric literaturethere is evidence that such raredomestic types continuities. These factors are a consequenceof the are in fact the special occasioncounterparts of everyday amountof time thathas elapsedsince majorconstruction domestic vessels, used in weddings, householdsancti- and occupationof the site. Most settlementstudies that fications,and otherindividual or family-orientedprivate have accuratelyapproached the problemof household rituals. It is distinguishedfrom the general domestic identificationand population density have dealt with Late pottery by surface treatment,size, and/orsource.25 At Classicand Postclassic populations.27 The Cerrosmound Cerros, some of the types match the patternsof distri- populationhas been affected by the passage of at least butionand the frequencyof occurrenceof theircounter- double that amountof time. partsdescribed in the ethnographicrecord. When these House mounds have been defined at Cerrosthrough observationsare coupled with the fact that the ancient various independentchecks. The size, complexity,and Maya had such privaterituals,26 it seems only prudent abundanceof moundsgenerally have been the basis for to differentiatethese types from domesticpottery. identification,given the low frequencyof densedomestic Havingestablished that these classes of potteryexisted trash deposits and household features such as hearths and were consistentlyidentifiable in primarycontexts, it and burials.28A high percentageof the domesticobjects was then possible to suggest functionalinterpretations (fishnetweights, spindlewhorls, etc.) andceramic types for other excavated contexts using the ceramic data. (TABLE6) from the settlementoccurred in association When these ceramically-basedinterpretations were used with structuresdating to the Late C'oh and Tulix phases in conjunctionwith the informationderived from the (100 B.C.-100 A.C.). These kinds of pottery and other architecture,settlement patterns, and excavations, it was artifactsare generallytaken to indicatethe locationof a possible to interpretthe function of a given Late Pre- household. In addition, there is reason to believe that classic structurewith a higherdegree of confidencethan ritualobjects (such as jade fragmentsand painted stucco) is usuallypossible in the Maya Lowlands. and certain ceramic types (TABLE6) are expectably less frequentin house mound locations than in those with

25. R. E. Reina and R. M. Hill, Traditional Pottery of Guatemala 27. W. A. Haviland, Excavation of Small Structures in the Northeast (University of Texas Press: Austin 1978); R. H. Thompson, Modern Quadrant of Tikal, Guatemala (University Microfilms: Ann Arbor Yucatecan Mayan Pottery Making. SAAMem 15 (1958). 1963); Pollock, loc. cit. (in note 11). 26. Alfred M. Tozzer, ed., Landa's Relaci6n de las Cosas de Yuca- 28. After W. A. Haviland, "Tikal, Guatemala and Mesoamerican tdn. PapPeaMus 18 (Harvard University: Cambridge 1941). Urbanism," WA 2 (1970) 186-198. 166 Late Preclassic Maya Center/Scarborough and Robertson

Table 7. Functional of interpretation Structureor Formal Initial structuresexcavated in the Structure Field Ceramic Construction settlement. Group Number Identification Assessment Date 9 civic facility civic/ritual Tulix 10 civic facility indeterminate EC 11 elite residence elite residence Tulix 13 residence non-eliteresidence Tulix 14 civic facility indeterminate* Tulix 15 residence elite residence L. C'oh 16 residence elite residence L. C'oh 18 residence elite residence Tulix 19 civic storage civic/ritual Tulix platformt 21 civic facility indeterminate* Tulix 22 residence non-eliteresidence Tulix 24 outbuilding non-eliteresidence L. C'oh 26 residence non-eliteresidence Tulix 29 civic facility civic/ritual Tulix 33 groundlevel elite residence L. C'oh residence 34 residence elite residence L. Tulix 38 residence non-eliteresidence L. C'oh 46 elite residence indeterminate EC 50 ballcourt civic/ritual Tulix 53-1st residence elite residence L. C'oh 53-2nd civic facility civic/ritual Tulix 54 civic facility indeterminate* Tulix 57 outbuilding non-eliteresidence Tulix 61 ballcourt civic/ritual L. C'oh 65 residence indeterminate* L. C'oh 66 residence elite residence Tulix 76 civic facility civic/ritual Tulix 77 residence indeterminate LPC 84 residence indeterminate EC 94 residence indeterminate LPC 98 residence elite residence Tulix 102 residence indeterminate* L. Tulix 112 dockingfacility civic/ritual Tulix 115 residence elite residence Tulix 116 residence indeterminate* Tulix 165 groundlevel indeterminate* Tulix residence *Prohibitivesample size. tStructuresummit area is 1,280sq m. civic architecture.29 Domestic objects and other small form and comprise 74% of the mounds in the systematic finds from the settlement are currently under study.30 survey and excavation area. Of the 21 excavated structures that were assigned a More than half (7) of the excavated Type 4 mounds residence function, approximately 80% (17 mounds) were residences as indicated by both the ceramic and were Type 4, 5, or 6. These structures are simple in field criteria. The ceramic inventory also revealed that only two of these Type 4 mounds were civic facilities. 29. J. F. Garber,personal communication. Excavations on Structure 53 suggested that during the Late C'oh this functioned as an elite residence 30. J. F. Garber, Material Culture and Patterns of Artifact Con- phase gee the Tulix took on a civic function. sumption and Disposal at the Maya Site of Cerros in Northern Belize but, during phase, (University Microfilms: Ann Arbor 1981). This was confirmed by both the formal field data and Journal of Field Archaeology/Vol. 13, 1986 167 the ceramicevaluations. Although clear architecturalin- dicationswere lacking, the locationof Structures53 and 65, equidistantfrom the westward-facingTulix phase Structure29 and the two ballcourts,suggests an inten- tional civic plan.31 The excavatedType 5 moundswere assigneda resi- dence function on the basis of the field and ceramic criteriaoutlined above, though small lots precludeda clear ceramic identificationfor approximately50% of these mounds. Such criteria,however, did indicatethat the other half of the mounds sampled were elite resi- dences. The horizontalexposure of Structure34 also revealed a small Tulix midden deposit off the eastern corner of the structure(FIG. 5), confirminga residential functionfor the structure. o The excavated Type 6 structuresincluded five resi- dences and two Except for the ground- outbuildings. 5. Houseplatform Structure 34: a Type 5 residentialstructure. level elite residence Structure33, all of the structures Figure assessed as residencesproduced non-elite domestic pot- tery. Structures165 and 22 were horizontallystripped to stucco facadingwithin an exteriorwall niche, coupled test their residentialfunction, but unfortunatelyneither with the overall architecturaldesign of the structure, produceddistinctive household features. Structure165 suggests that it was an elite dwelling. was a ground-levelresidential structure near the main The ceramic assessmentspermitted a furthertest of canal. The outbuildings(Structures 24 and 57) were the functional significanceof the mound typology. A identifiedas such by theirsmall size andclose proximity comparisonof moundvolumes with ceramically-identi- to othermounds or featuresin the settlement.The pottery fled elite and non-elite residencesdid not prove to be recoveredin both cases was that typical of a non-elite useful, but a clear correlationwas evident when the residence. summit surface-areaor presumedfloor space of a resi- More elaboratestructures belonging to Types 1, 2, dence (derivedfrom 1:50 or largerscale contourmaps)34 and 3 were also assigned a residentialfunction. These was comparedwith the functionalceramic inventory. raised plaza groups have from two to four additional structureson top of their plazas. In all cases within the systematicallysurveyed area, one of the summitstruc- 34. V. L. Scarborough, The Settlement System in a Late Preclassic tures is larger and more prominentthan the others, an Maya Community,unpublished Ph.D. dissertation,Southern Meth- arrangementalso present at ,32Tikal,33 and odist University(University Microfilms: Ann Arbor1981). other sites in the Lowlands.Their form may be related to the elevatedstatus of a householdhead and his nuclear family. Presumably,members of the associatedextended family would have occupied some of the remaining structures.The ceramic inventories from these more elaboratestructures are those characteristicof elite res- idences. The exception, Structure116, did not produce enough sherdsfor an evaluation. As an example of this plazuela association, the Struc- ture 11 group (FIG.6) was horizontally stripped. Although little midden debris was located, the presence of painted facading 31. Scarborough,op. cit. (in note 18) 22. meters 32. A. L. Smith, op. cit. (in note 11) 218. 33. WilliamA. Haviland,"Maya Settlement Patterns: a CriticalRe- view," in E. W. AndrewsIV et al., eds., ArchaeologicalStudies in Figure6. Elite houseplatform Structure 11B (thelargest of three Middle America. Middle American Research Institute Publication 26 mounds).This is a Type 2 residentialstructure. The backthird of the (TulaneUniversity: New Orleans1966) 31. moundwas not preservedand no doorwaycould be defined. 168 Late Preclassic Maya Center/Scarborough and Robertson

Consequently,summit surfacearea must be considered Table8. LatePreclassic residential function assessed. a betterindex of "eliteness"than are gross earth-moving Structure SummitArea Ceramic of Table8 indicatesthat a structure expenditures energy. Number in sq m Assessment summitarea greater than 30 sq m is associatedwith elite residence. The occurs on Structure38. 98 160 elite residence only exception I 1B* 130 eliteresidence The of nonceramicritual from this lo- recovery objects 116B* 120 indeterminate cus,35 however, may provide an explanationfor this 115B* 100 eliteresidence reversalof the norm. 38 100 non-eliteresidence 15 96 eliteresidence Civic Structure Architectureand 18 80 elite residence Identification: Spatial 66 80 elite residence Relationships 53 72 eliteresidence Civic architectureis the other class of structurerec- 16 48 eliteresidence ognized at Cerros,defined as special-functionbuildings 33 8t elite residence 34 whichdo not incorporatethe rangeof domesticactivities 36 indeterminate carriedout on house mounds. were identified 102 36 indeterminate They by 13 25 non-eliteresidence their unusual elaborate imposing size, plan, masonry, 26 24 non-elite residence and/orfacades (the latterseldom preservedin situ) and, 57 16 non-eliteresidence in the Tulix phase, by theirspatial relationships to Struc- 24 16 non-eliteresidence ture 29.36 Similarly,the ceramicinventory is distinctive 165 16 indeterminate and domestic objects are presentin significantlylower 22 9 non-elite residence than in percentages domiciles.37 *Largeststructure in plazuelagroup. Civic architecturehas a in formthan greatervariability tLimited flat test exposure. do the house mounds. Given the activity-specificfunc- tion of many civic structuresand the numberof social at a and economic tasks performed Mayacenter, greater Within the four tested Type 2 structures,Groups 76 This in form variabilitywould be anticipated. variability and 10 were assigneda civic function.They date to the lateral on has been clarifiedby testing and by exposures Tulix phaseof the Late Preclassicand to the EarlyClas- the four mound that representativemounds from types sic, respectively. arguablyrepresent civic architecture. Both of the Type 1 structureswere excavated.Group In the Tulix six excavated 4 structures phase, Type 50 is clearly a Tulix phase ballcourt(FIG. 7),40 whereas The amount have been interpretedas civic architecture. Group116 has a residentialfunction (see above). Group of labor requiredto constructthese simple but more 116 is outside the core zone and is located within the imposingmounds would have been considerablygreater bajo, or internally-drainedswamp, suggestingits resi- structuresin than it would have been for the largest dents may have played a managerialrole in the raised m of fill in Structure29 Types 5 and 6. The 21,460 cu field agriculturepracticed at the site. If so, the residential the andits supportingplaza, for example,represent larg- functionof this groupwould have had civic components and exten- est constructioninvestment in the core zone, that may accountfor its greatercomplexity. sive lateral exposure has revealed architecturaldetails associatedwith a civic functionin this case.38 Problemsin Of the threeType 3 structuregroups that were tested, Interpretation the two Late Preclassic groups were assigned a civic Changes in the residential-and civic-structurepopu- function. Group 61 was unequivocallyidentified as a lations throughtime are importantindices for compara- ballcourt39and Group 19 appearsto be a storageplat- tive purposes.In orderto extractthese data, two units form. The Early Classic Group46, on the other hand, of occupationand constructionanalysis were designed. appearsto be a ruralelite residence. Single-moundpopulations were estimatedby counting all the moundedfeatures individually. A plazuelagroup 35. Garber,loc. cit. (in note 30). and all moundsin it were assigneda date on the basis indicatedthat 36. Scarborough,loc. cit. (in note 34). of a single test unit, unless additionaltests not all the mounds were Although 37. J. F. Garber,personal communication. contemporaneous. such inferentialdating is commonpractice in the Maya 38. Freidel,loc. cit. (in note 13). 39. Scarborough,loc. cit. (in note 18). 40. Ibid. Journalof Field Archaeology/Vol.13, 1986 169

64K

Figure7. BallcourtStructure 50. This is a Type 1 civic structure.

Lowlands,41it may give an inflatedidea of the actual Tikal are most evident duringthe EarlyClassic period. numberof mounds constructedor occupied at any one Althoughthey may continuethrough the entireoccupa- time. For this reasonan area-countassessment has been tional historyof the site, Cerrosground-level dwellings calculated(TABLES 3, 4). are most significantduring the Late Preclassicphases. The area count simply evaluatesdiscrete occupation Cliff,43 for example, has demonstratedthe presence loci by weightingan entireplazuela as if it were a single of ground-levelstructures along the coastline that date mound. This figure is perhapsmost meaningfulwhen to all phases of the Late Preclassic period at Cerros. discussing reoccupationof a mound group. In these Moreover,because mound constructionwithin the pe- cases, residenceupon each moundwas probablyunnec- rimetercreated by the canal apparentlywas restrictedto essary to create the observabledistribution of ceramics the Late C'oh and Tulix phases, it seems thatduring the and relatedlitter. In both methods, plaza-typefeatures Early and Middle C'oh phase much, if not all, of the have been excluded from the counts. Although these occupationof the dispersed settlementwas at ground must be considered major constructionexpenditures, level. Actualhuman population estimates during the Late they only directly affect percentagesof civic construc- C'oh and Tulix phases for the site must be considered tion. In postulatingthe numberand densityof civic and conservativeuntil Cliff has analyzedthe data. The struc- residentialstructures in the systematicsurvey and ex- turaldensity figures, however, are comparableto those cavationarea (TABLES4, 5), the functionalidentification at otherMaya sites.44 of unexcavatedstructures was necessary.This was ef- fected by comparingthe individualstructures with the A Settlement Reconstruction of identified structuresand was sample only possible Our excavationwithin the systematicsurvey and ex- afterthe excavationdata had been collected. cavation area permit the discussion of developmental A with the Cerros data base is the major problem trendsat Cerrosfollowing the initial occupationof the of structures set- presence ground-level throughoutthe site duringthe Ixtabiphase. tlement. Although the issue has been an elusive one, many researchershave devoteda greatdeal of attention C'oh Phase (200-50 B.C.) to it.42 Puleston indicates that "hidden"dwellings at In the Early and Middle C'oh phase most of the oc- cupationof the dispersedsettlement appears to havebeen 41. D. E. Puleston, Ancient Maya Settlement Patterns and Environ- mentat Tikdl,Guatemala (University Microfilms: Ann Arbor1973); Rice and Rice, loc. cit. (in note 10). 43. Cliff, loc. cit. (in note 17). 42. William T. Sanders, Prehistoric Ceramics and Settlement Pat- 44. Scarborough,loc. cit. (in note 34); idem, "LatePreclassic North- terns in , Mexico. CarnlnstPub 606 (Washington, D.C. ern Belize," Status, Structure, and Stratification: Current Archaeo- 1960); Haviland,loc. cit. (in note 27); Willey, loc. cit. (in note 7); logical Reconstructions. Proceedings of the 16th Chacmool Con- Puleston,loc. cit. (in note 41). ference (Calgary1985). 170 Late Preclassic Maya Center/Scarborough and Robertson

Tulixphase. Only two excavatedmonumental structures, Structure76 and StructureGroup 61, can be reasonably associatedwith the Late C'oh phase. Structure76 lies outside the core zone at a locationmidway between the main 2 and the main canal. its distance rm aguada Despite from the core area of the site, the constructionfill of ?. ? Structure76B containedC'oh pottery. Thus it would seem that there was C'oh occupationin the vicinity of the mound. Its location may indicatethat the structure k~ playeda role in linkingthe EarlyC'oh andperhaps even the Ixtabaioccupants of these two areas. Even though StructureGroup 61, one of the two ballcourts,was most heavilyutilized during the Tulixphase, constructionwas initiatedduring the Late C'oh phase. The constructionof the main canal was initiateddur- SMXJUND CONSTRUCTION ing the latterportion of the C'oh phase. It probablywas o oCCUPATION LOCI built for rainwatercatchment, with the bulk of the re- moved fill used in the constructionof the civic Figure 8. Distribution of C'oh Phase structures. being and residentialspace inside it, althougha defensiverole cannotbe dismissedat this early date.49 at groundlevel. During the Late C'oh, however, there The site appearsto have been heavilyoccupied during was a substantialamount of residentialconstruction in the C'oh phase. Occupationalloci for the C'oh phase this area of the site. Withinthe core zone, eight of the have been definedas secondarydeposits located in con- excavated structures(area count) were built duringthe structionfill or primarytrash sealed below mounded Late C'oh phase. Of these, six are house mounds, one features. In the former instance, the condition of the is an outbuilding(Structure 24), and one is a ballcourt sherdsindicates that they were not transportedany great (Structure61). Within the six house mounds, four are distance.Therefore, they reflectthe existenceof a nearby elite residences,one is indeterminate,and only one is a ground-leveloccupation. Based on these criteria,it ap- non-elite residence. It appearsthat with the successful pears that 39% of the known occupation"area" inside developmentof intensiveagriculture throughout northern the canal perimeterwas utilized duringthe C'oh phase, Belize,45the energy investmentin the constructionof while 29% of the known occupationspace outside the the main canal at Cerrosin the Late C'oh phase,46and canal was employed. An infield/outfieldagricultural ad- the growthof riverineexchange networks,47 the emerg- aptationis posited,50although riverine exchange directed ing elites began to express their authority.They either throughthe centralprecinct may accountfor the settle- moved away from the village nucleusor elevatedthem- ment attractiontoward that area and the village nucleus. selves above the non-elite. These elites, however, ap- withinthe boundariesof the canal, since parentlystayed Tulix Phase B.c.-150 no residentialconstruction can be identifiedoutside the (50 A.C.) core area (TABLES3, 4; FIG. 8). The Tulix phase occupationof the site representsthe Most of the civic architecturedating to this phase period of majorcivic construction.Nearly all the mon- underliesthe presentcentral precinct.48 In the settlement, umentalarchitecture visible at the site can be shown to it was considerablyless developedthan that of the later have been utilized duringthis phase, and at least 80% of all civic constructioninside the canal dates to this 45. P. R. Bloom et al., "Prehistoric Maya Wetland Agriculture and time. Outsidethe canal the only civic constructionwas the Alluvial Soils near San Antonio Rio Hondo, Belize," Nature 301 Structure 76D, which was added to the 76 group, sug- (1983) 417-419; A. H. Siemens, "Prehistoric Agricultural Use of the Wetlands of Northern Belize," in K. V. Flannery, ed., Maya Subsis- gesting the overall reuse of this group during this period tence: Studies in Memory of Dennis E. Puleston (Academic Press: (TABLES 3, 4; FIG. 9). New York 1982) 205-222; B. L. Turner and P. D. Harrison, "Pre- All of this grand construction corresponds with the historic Raised-Field Agriculture in the Maya Lowlands," Science 213 49. D. A. Freidel and V. "Subsistence, Trade and (1981) 399-405; Scarborough, loc. cit. (in note 14). Scarborough, Development of the Coastal Maya," in Flannery, ed., op. cit. (in note cit. note 46. Scarborough, loc. (in 12). 45) 131-151; Scarborough, loc. cit. (in note 12). cit. note cit. note 47. Freidel, loc. (in 16); Scarborough, loc. (in 44). 50. After R. M. Netting, "Maya Subsistence: Mythologies, Analo- 48. D. A. Freidel, personal communication. gies, Possibilities," in Adams, ed., op. cit. (in note 1) 299-334. Journal of Field Archaeology/Vol. 13, 1986 171

of an arbitraryN-s trending line from the Structure 146 Group to the canal, is 1.23 mounds/ha (the nearest mound east of the Structure 146 Group lies 240 m away, suggesting a less arbitrary division of this area). If 53 of these mounds were occupied during the Late Preclas- 4 " . sic (as has been indicated by the sample outside the *s canal in the systematic survey and excavation area), then CID-" " cT an occupation density of 0.65 mounds/ha is hypotheti- cally projected for this area. Thus the density drop-off outside the main canal is not significant along the coast- Do~3 line and in proximity to the New River until one ap- proaches the first river terrace. In contrast, the two transect lines to the south, into the interior of the bight, indicate a density figure much lower than that provided by the southern-block periphery zone. A total density of 0 MOUND CONSTRUCTKON 0.26 mounds/ha converts to 0.14 the Tulix oc- O oCCUPTK~NLOCI during

FIELDS a in the interior. P RAISED cupation, indicating density drop-off Consequently, it seems the western concentration rep- Figure 9. Distribution of Tulix Phase structures. resents an adaptation to riverine resources and exchange, with the service population engaged in petty trade along quarrying activities associated with the water manage- the shoreline. ment system.51The main canal was widened and dredged Generally, this density drop-off outside the canal cor- in at least one location, and raised field platforms were responds best to our understanding of the centralizing constructed in the core zone. forces at work in a Maya center. Local exchange systems In addition to the civic monuments, residential con- would have been coordinated and found to converge at struction was undertaken inside the canal. Approxi- the center. Additional support for this hypothesis comes mately 60% of the house mounds inside the canal were from the construction of Structures 112 and 19 to the constructed during the Tulix phase, indicating that the east and south of this area, respectively. Both are within core was a residential zone. The density of house mound the confines of the canal. Structure 112 has been inter- construction and occupation in this zone was 1.21 preted as a port facility based on its immediate proximity mounds/ha, demonstrating that the site was certainly not to the shoreline, the presence of a ramp-like gradient a vacant ceremonial center. In addition, more than a rising from the shore to the mound's summit, and the third of the house mounds in the core area were non- generous projected summit surface area of the struc- elite residences. In contrast, the house-mound density ture.53If this is the case, it contrasts with the earlier, but outside the main canal was 0.40 mounds/ha, suggesting still functioning, dock previously associated with the a "dispersed compact" model52for the settlement during village nucleus of the site,54 in that it has little additional the Late Preclassic florescence. Even if the entire sys- room for the storage of goods. Structure 19, by the same tematic survey area outside the canal is figured into the token, with its summit area of 1,280 sq m and slightly total (having an overall density of 0.75 mounds/ha), a greater distance from the center and coast, seems to have hypothetical figure of only 0.40 mounds/ha can be de- been reserved for the storage of goods.5 Perhaps the rived from a Tulix context, given the 53% occupation elites were exerting control over riverine exchange by total for this phase. It should be noted that the bajo area controlling not only the reception of goods at the site has not been subtracted from the density figures since but their distribution as well. Less supervised and more occupation did occur in these areas with frequency. localized exchange could have taken place along the There are, however, variations in the distribution of western shoreline and outside the canal, making settle- mounds outside the canal. Although the systematic sur- ment at this location advantageous to a subordinate ser- vey and excavation area was defined during the 1978 vice class. The production of agricultural products season, more recent survey and reconnaissance has re- vealed a sizeable concentration of mounded features sw 53. Scarborough, op. cit. (in note 34) 162-167. of the core zone. The mound density of this area, west 54. Cliff, loc. cit. (in note 17). 55. See D. A. Freidel and J. A. Sabloff, : Late Maya Set- 51. Scarborough, loc. cit. (in note 12). tlement Patterns (Academic Press: New York 1984) for similar struc- 52. Puleston, loc. cit. (in note 41). tures on Late Postclassic Cozumel Island. 172 Late Preclassic Maya Center/Scarborough and Robertson

small reservoirs of water. Sections of the canal are breachedby causeways or dams, suggesting that the originalcanal was an impedimentto foot traffic(TABLES 3, 4; FIG. 10). These data indicatethat the site was not abandoned following the Tulix phase occupation;instead, it was basedon a differentland-use pattern because of an over- all populationdecrease. The Early Classic adaptation probablywas similar to the infield/outfieldadaptation 4 o made duringthe C'oh phase, even thoughintensive ag- riculturein the form of raisedfields may have continued in the immediatevicinity of the core zone. The more elaboratehydrology of the Tulix phase, however, was certainlybeyond the organizationalinterests or capabil- ities of the Early Classic inhabitants.Although local 0 MOUND CONSTRUCTION o OCCUPATIONLOCI exchangemay have continuedto supportthe residential ) PROBABLEAGRICULTURAL ZONE populationinside the canal, agriculturemust have been &CATCHMENT BASINS the majorsubsistence mode. The communitymay have been attractedto the Tulix phase ruins because of the 10. Distributionof EarlyClassic structures. Figure functionaladvantage of elevatedground near the shore- line, to say nothingof the formerbeauty and glory of outsidethe canal for internalor externalconsumption is the site. It would seem that the managerialelite had suggested, but clear empiricalevidence is lacking. abandonedthe site by the end of the Tulix phase. The Althoughthe purviewof this paperhas been the Late service populationand their descendants,on the other Preclassic period, some discussion of the Classic and hand, adopteda new, less structuredorder. Postclassic periods is warranted.This brief treatment Cerroswas virtuallyabandoned during the Late Clas- permitsa clearerpicture of laterperiod densities at Cer- sic period. A cist containingTerminal Classic trashwas ros andunderscores the relativelypure nature of the Late exposed on Structure50E of the ballcourtgroup, but no Preclassiccomponents. additionalevidence for occupationhas been found. The abandonmentof the site duringthe EarlyClassic period is not well it has been Classic Period understood,although suggested that exchange networks and political associationscir- The EarlyClassic periodat Cerrosis identifiedby the cumventedLowry's Bight.56Alternatively, three sites in constructionof only one civic monument,Structure 10 northernBelize indicate that a rise in sea level or a Group, the bulk of which appearsto have been built complementarylandmass subsidence occurred immedi- during this period. Overall mound constructionat the ately following the Late Preclassicperiod.57 Given its site droppedto 12% of the total. Although a sizeable proximityto the bay, Cerroswould have been severely residentialreoccupation was assumed initially, further affectedif the relativewater level rose even a meter. analysishas shown thatthese laterdeposits were ephem- On the basis of presentinformation, it may be said eral and not indicative of permanent,long-term settle- thatthe EarlyClassic periodoccupation of the core area ment. The evidence indicates that once a mound was was ratherbrief and was initiatedby the formerservice constructed,it was used by all subsequentgroups. Later populationto revitalizethe fallen Tulix center.(Ceramic utilization,however, was usually slight and resemblesa analysisfor this periodhas not progressedfar enoughto sheet midden deposit covering the entire site. Much of clearlyassess the actuallength of time the EarlyClassic this occupationrepresents use of former civic monu- ments that had been constructedby the end of the Tulix 56. Scarborough,loc. cit. (in note 44). phase. 57. Scarborough,loc. cit. (in note 12); Bloom, loc. cit. (in note 45); Although the zone outside the canal representsa P. D. Harrisonand B. L. TurnerII, eds., PulltrouserSwamp: Ancient slightly lower density of occupation(0.51 mounds/ha) Maya Habitat, Agriculture, and Settlement in Northern Belize (Uni- two new versityof Texas Press:Austin 1983). The latterargue for raisedfield than does the area inside (0.76 mounds/ha), which would have been inundated the Late Pre- on the platforms following house mound areas were constructed periphery. classic period.W. J. Folan et al., "PaleoclimatologicalPatterning in The main canal probablydid not function as a major SouthernMesoamerica," JFA 10 (1983) 453-468, also suggesta wet water control device but ratherwas modifiedto collect periodduring the Late Preclassicin the Maya Lowlands. Journal of Field Archaeology/Vol. 13, 1986 173

less meaningful.The overall density of reoccupationin the systematicsurvey and excavationarea was approxi- mately 0.30 mounds/ha.The adaptationmade by these occupantsis difficultto determine,but a variationof the "merchantpragmatism" model58 may be suggested by Z3 o the limitedarchitectural investment,59 and the ratherrich LatePostclassic caches whichhave been recoveredfrom the largermonumental structures. The Late Postclassic populationconfiguration at Cerrosis thoughtto be re- lated to the growth and dominanceof Santa Rita just threekm across the bay.60 A growthmodel throughtime hardlyseems appropri- ate for Cerros.The C'oh into the Tulix So phasedeveloped phase, while the Early Classic saw the reoccupationof the collapsed Tulix center. Sparse constructionduring 0 MOUND CONSTRUCTION the with the ab- O OCCUPATIONLOCI EarlyClassic, coupled nearlycomplete sence of Late Classic debrisat the site, stronglysuggests Classic was not a Figure 11. Distribution of Late Postclassic structures. that the Early adaptation successful one. An attemptto maintainLate Preclassictraditions, given the new social order,probably spelled the eventual occupationmay have continuedat Cerros.) This effort demise of the population.The absence of later Classic emphasizedthe maintenanceof the subsistencesystem periodoccupation was a consequenceof environmental to supportthe local population.With the abandonment factors to which the silting-in of the water catchment of the center by the elites and the loss of centralized system contributed.Although the geographicallycom- authority,however, the civic architectureeroded into mandingposition of Cerrosnear the mouthsof the New disrepair.The canal networkunderwent sedimentation, andHondo Rivers was not altered,the energyinvestment and the general water catchmentscheme failed. The necessaryto revitalizethe watercatchment system would environmentalsetting probablyreverted to a condition have been extremely costly. It was not until the Late not unlike that found at the site today, makingthe site Postclassicperiod and the rise of the materialistictrading nearly uninhabitable.Major reoccupationof the site colonies of the Yucatancoast that the geographicposi- would have requireda considerableenergy investment, tion and the potentialof an immediateeconomic return an investmentthat was never realizedagain. outweighed the physical unpleasantnessof the Cerros environment. Late Postclassic Period Conclusions The Late Postclassicoccupation at Cerrosmay spana The settlementand ceramicdata from Cerrospermit of time than the other described. greaterperiod phases a specific treatmentof one kind of pre-statecommunity this have inflatedthe estimates Although may density development.After the initial colonizationof Lowry's these to indicate the same somewhat, figures appear Bight, the communitydeveloped from a principallyres- trendsrevealed the Classic general by Early occupation identiallocus to a well-planned"central place" of civic The settlement this (TABLES3, 4). configurationduring monumentsand residential space. Even though civic is less reflectiveof deliberatePostclassic periodprobably constructionseverely modified the landscapeand pro- than it is of selectionfor available spatialdesign simple duced a differentresidential adaptation to the site envi- high ground (FIG.11). The setting probablywas very much like the present, althoughbush trails and cleared plaza space were likely the rule. No civic construction was carriedout at location in the and 58. J. A. Sabloff and W. L. Rathje, "The Rise of a Maya Merchant any settlement, Class," SAm 233 (1975) 72-82. only two house moundswere built. Most of the monu- mental was An es- 59. See P. D. Harrison, "The Lobil Postclassic Phase in the Southern architecture,however, reoccupied. Interior of the Yucatan in N. Hammond and G. R. dense domestic Peninsula," Willey, pecially trash deposit, for example, eds., Maya Archaeology and Ethnohistory (University of Texas Press: accumulatedat the foot of Structure9. Becausethe main Austin 1979) 189-207. canal probablywas not utilized in any significantman- 60. D. Z. Chase, "The Maya Postclassic at Santa Rita Corozal," ner, density figuresinside and outsidethe core zone are Archaeology 33 (1981) 25-33. 174 Late Preclassic Maya Center/Scarborough and Robertson ronment,the Late Preclassicpopulation remained dense Lowlands.Numerous authors67 have noted that a "rural and increasedthrough time. By the Tulix phase, the site elite"appears in the Mayaarea at varyingdistances from reflects"synchorism,"61 manifesting both civic attraction the central precincts and core areas of administrative and residentialaggregation. The growthof the commu- centers. The appearanceof relativelylarge civic struc- nity suggests the transitionfrom an economydependent tures away from centers also occurs with some fre- upon local resourcesduring its initial occupationto one quency. From the speed at which the rain forest focused on regional exchanges of goods and services regeneratesfollowing initialbush-clearing operations, it duringits final Late Preclassicoccupation.62 is clear that, in the absence of continualcropping, the The adaptationof a communityto a compactedsettle- spatiallimits of a communitycould be severelyaltered.68 ment design with substantialcivic constructionby the Withouta representativefrom the centralauthorities in Late Preclassic period has evolutionaryimplications. the ruralareas, a supportconstituency might easily be Previous researchin the Maya Lowlandssuggests that physicallylost. the "dispersedcompact" settlement adaptationwas a A recentset of articlestreating the concentriczonation recurrentadaptation made by later Classic period cen- model in the Maya area suggests that Late Classic ters. In light of the Cerrosdata, the questionbecomes: vaultedstructures (commonly associated with elite resi- why did the Maya continue this settlementdesign for dence) reveal little decreasein densityas one leaves the some 1500 years, given the complexityof their institu- epicenterof large centers.69Elsewhere it is arguedthat tions? In so doing they contradictmuch of what is elite house mounddensities do decline in proportionto known about state developmentand social controlfrom distance from these.70 Althoughthe Cerrosdata might other areas.63Although Maya centershad sizeable pop- be construedas somewhatequivocal, the need to define ulationaggregates and performedvarious "urban" func- and monitorthe hinterlands,even those in close prox- tions, they seldom developed into cities in the manner imity to the central precinct of a small center, is a of nucleatedresidential and civic centersfound in high- necessity with respect to any socially-complexpopula- land Mexico. This is not to say that the Maya were tion. However, the "lost cities" of the Maya represent incapableof founding and maintainingnucleated com- an exaggeratedcondition in which the dispersedpopu- munities, as evidenced by Mayapan64and Chunchuc- lation seems spatiallyintractable or at least difficultto mil.65 Agriculturalconstraints were no more severe in marshal. The engagementof the populationin public the Lowlandsthan in the Highlands,as indicatedby the rituals and the presence of elite monitorswould have extensive raised field systems identifiedthroughout the providedgreater cohesion and preventedany tendency Lowlands.66Although the Mayawere awareof nucleated towardcorporate village autonomy. urbanorganization and the advantagesof populationcen- Althougha formidableliterature has been compiledin tralizationfor social control, they opted to maintaina supportof the "swiddenhypothesis," recent studies have more dispersedsettlement design. challengedit. The corollary-that a dispersedsettlement The Maya developeda dispersedpattern of civic and patternresulted from a slash-and-burnagricultural ad- householdclusters which allowed the controland regu- aptation-has been shaken, and evidence of intensive lationof state institutions.Although the Mayaappear to agriculturehas been identifiedthroughout the Yucatan have changed their ritual orientationand their use of Peninsula.Even so, the populationremained more dis- public vs. private space at various points in time, the persedthan was the case in otherprimary civilizations. dispersed-clustersettlement design was constantlypres- Extendedfallow periodscan no longer be automatically ent. This dispersedpattern, when comparedwith many acceptedas a means of explainingthe Maya settlement othergreat civilizations, may have arisenas a condition system. Clearly the Maya, with their energy and crea- of the unique semitropicalenvironment of the Maya 67. W. R. Bullard, Jr., "Maya Settlement Pattern in Northeastern 61. C. L. Crumley, "Toward a Locational Definition of State Systems Peten, Guatemala," AmAnt 25 (1960) 355-372; Puleston, loc. cit. (in of Settlement," AmAnth 78 (1976) 59-73. note 41); J. E. Arnold and A. Forde, "A Statistical Examination of Settlement Patterns at Tikal, Guatemala," AmAnt 45 (1980) 713-726. 62. Freidel, loc. cit. (in note 1). 68. V. L. Scarborough, "Resourceful Landscaping: a Maya Lesson," 63. See L. Krader, Formation the State (Prentice Hall: Englewood of Archaeology 38 (1985) 58-59,72. Cliffs 1968); E. R. Service, Origins of the State and Civilization: the 69. cit. note Process of Cultural Evolution (W. W. Norton and Co.: New York Arnold, loc. (in 67). 1975); Sanders, loc. cit. (in note 2). 70. W. A. Haviland, "Where Rich Folks Lived: Deranging Factors 64. Pollock, loc. cit. (in note 11). in the Statistical Analysis of Tikal Settlement," AmAnt47 (1982) 427- 429; W. J. Folan et al., "An Examination of Settlement Patterns at 65. loc. cit. note 11). Vlcek, (in , Quintana Roo, Mexico, and Tikal, Guatemala: a Reply to 66. Scarborough, op. cit. (in note 12) 721. Arnold and Ford," AmAnt 47 (1982) 430-436. Journal of Field Archaeology/Vol. 13, 1986 175

tivity, were not limitedby any one ecologicalfactor, but perhaps one impinging environmentalelement-the rapidregeneration of the vegetationalcover-repeatedly affectedthe course of settlementevolution.

Acknowledgments We thankDavid Freidel,principal investigator for the Cerros Project, for the opportunityto work at Cerros. Beverly Mitchumand Susan Lewensteinprovided valu- able field assistance.Maynard Cliff, JamesGarber, Sor- raya Carr,and R. Kimball Smith providedinterpretive aid, both in the field and otherwise. Special thanksare extended to the editors, Peter Harrison,Howard Dau- distal, and the two anonymousreviewers. Karim Sadr was responsiblefor much of the drafting.The project was fundedby majorNSF grantsto David Freideland S.M.U. as well as a dissertationimprovement grant to Vernon Scarboroughand S.M.U. We are indebtedto HarriotTopsey, ArchaeologicalCommissioner of Belize and the kind and diligent people of Chunox and Cop- perbankvillages.

Vernon L. Scarborough was Director of the Meyer Pithouse Village Project at Fort Bliss, Texas. He has worked throughout the western United States and the Sudan. He was surveyor for the Cerros Project and is currently a Fulbright scholar in the Department of Archaeology at the University of Peshawar, Peshawar, Pakistan. Robin A. Robertson is Associate Dean for General Education and a Visiting Assistant Professor in the Department of Anthropology at Southern Methodist University. She was ceramicist for the Cerros Project.