Ballcourts and Ceramics: The Case for Hohokam Marketplaces in
Transcription
Ballcourts and Ceramics: The Case for Hohokam Marketplaces in
BALLCOURTS AND CERAMICS: THE CASE FOR HOHOKAM MARKETPLACES IN THE ARIZONA DESERT David R. Abbott,Alexa M. Smith, and Emiliano Gallaga During the middleSedentaryperiod (ca. A.D. 1000-1070) in the deserts of southernand centralArizona, crowdsfrom near andfar regularlygatheredat the centers of Hohokamvillages to participate in ritual ballcourtfestivities. These events were ideal venuesfor barter and exchange, leading some theorists to hypothesize that periodic marketplaceswere associated with the ritual ballgames. Recent ceramic provenance and vessel-form evidencefrom the Phoenix basin have shown that the productionof decorated and utilitarianpots was highly concentratedduring this time and large numbersof bowls and jars were evenly distributedtofar flung consumers. Thesefindings have supportedthe marketplacehypothesis, suggesting that an efficient and reliable mechanismwas available for moving large numbers of commodities across the region. The high volumeof ceramic transactions, however,seems to have placed the Hohokamcase beyond the capabilities of nascent marketplacesdocumentedfrom ethnohistoric and ethnographic evidence. In this paper, we support the idea that marketplace barter was a central componentof the Hohokameconomy by presenting new ceramic datafrom the lower Salt River valley, which temporallylinks the demise of the ballcourt ceremonialismwith a transformationin the organization of potteryproductionand distribution.Wethen examine some unusualcircumstancespertaining to the Hohokamregional system that may help to explain how consumers could have so heavily depended on a networkof horizontallyorganized,periodic marketplacesfor basic necessities like earthenwarecontainers. Duranteel periodo Sedentariomedio(1000-1070 d.C) en los desiertosdel centroy sur de Arizona,multitudesde cerca y lejos se juntaban regularmenteen los centros de las villas Hohokampara participar en las festividades rituales de los juegos de pelota. Estos eventoseran lugares ideales para regateare intercambiar,llevandoa hipotetizara algunos teoricos que la realizacionperiodica de mercadosestaba asociada con los rituales de los juegos de pelota. Recientes investigacionesde provenienciacerdmicay evidenciadeforma-de-vasijade la cuencade Phoenix,han mostradoque laproduccionde vasijasdecoradas y utilitariasestaba concentradaprincipalmenteduranteel periodo Sedentariomedio y un gran numerode cuencos y ollas fueron distribuidasequitativamentea consumidoreslejanos. Estas investigacionesdan soporte a la hipotesis del mercado, sugiriendoque un eficientey confiablemecanismoestaba en uso para mover un gran numerode comodidadesa travesde la region.Sin embargo,el alto volumende transaccionescerdmicas,parecen haberpuesto el caso Hohokammuypor encima de las capacidadesde los incipientesmercadosdocumentadosmediantela evidencia etnohistoriacay etnogrdfica.En este articulo, apoyamosla idea de que el regateo en los mercadosfue un componentecentral de la economia Hohokampresentando nueva evidencia e informacioncerdmica del Vallebajo del Rio Salado, que temporalmentese conecta con la debacle ceremonial de los juegos de pelota con la transformacionen la organizacionde la producciony distribucioncerdmica.Posteriormenteexaminamosalgunascircunstanciasinusualesconcernientesal sistemaregionalHohokamquepuedenayudara explicar como los consumidorespudierondependertantode una red organizadahorizontaly periodicamenteen mercadospara cubrir necesidadesbdsicas como recipientesde barro. conceptualizetheHohokam regionalsystemas a set of geographically Archaeologists separatebut interactingprehistoriccommunitiesspreadacrossa vastregionalterritoryand dependenton one anotherthroughthe exchangeof goods and services(CrownandJudge,eds. 1991; Wilcox 1979).DuringthemiddleSedentaryperiod (ca. A.D. 1000-1070), when the regionalsystem hadexpandedto its largestextent,it encompassed some 80,000 km2that includedmost of southern andcentralArizona(Doyel 1991a:247).One manifestationof the system's integrationwas a dense networkof ballcourts,whose limitslargelydemarcated the regionalsystem boundariesand whose David R. Abbott School of HumanEvolutionand Social Change,Arizona State University,Tempe,AZ 85287-2402 Alexa M. Smith 1233 N. 5th Avenue,Tucson,AZ 85705 Emiliano Gallaga Calzadade los HombresIlustress/n (Museo Regional de Chiapas),c.p.29000, Col. Centro,Tuxtla Gutierrez,Chiapas AmericanAntiquity,72(3), 2007, pp. 461-484 Copyright© 2007 by the Society for AmericanArchaeology 461 462 AMERICAN ANTIQUITY [Vol. 72, No. 3, 2007 of the largestandmostimpressiveirrigationworks of the prehistoricNew World.At severallocations alongboth sides of the lowerSalt andmiddleGila Rivers(knowncollectivelyas the Phoenixbasin), the desertfarmersmaintaineda centuries-oldpatternof constructingintakefacilitiesto drawwater fromthe riverchannelintohundredsof kilometers of main and distributioncanals thatdeliveredthe flow to tens of thousandsof hectaresplantedin corn, beans, squash,and cotton (e.g., J. Howard 1991, 1993; Hunt et al. 2005). As OmarTurney ( 1929) noted,thehydrauliclayoutin eachrivervalley was dividedinto canalsystems,each of which consistedof canalswith a commonheadgatelocation and the irrigatedfarmlandsand settlements spreadoutalongthecanalroutes.Clearly,thefarmers in each canal system cooperated with one anotherto operate and maintaintheir irrigation facilities.In the lowerSaltRivervalley,therewere four majorirrigationunits- CanalSystems 1 and 2, the ScottsdaleCanalSystem,andtheLehiCanal System- as well as severalsmallerandunnamed systems (Figure1). The denselyinhabitedandirrigatedlowlandsin the Phoenixbasinconstitutedthe heartlandor core of theHohokamregionalsystem.Probablynearthe beginningof the ninthcentury,the firstHohokam ballcourts were constructedthere. These large, oval-shapedfeatureswere flat-bottomedexcavations with preparedfloors and sidewallsthatconEmil tinuedupontothesurrounding embankments. Haury(1937a) was the firstto proposethatthese earthenconstructionswereballcourts,wherea versionof theMesoamericanballgamewasplayed.He cited similaritiesin various attributes,including high parallelembankmentsthatflankeda playing alley withinwhich threemarkers,one in the centerandone ateachendwereplaced(see alsoWilcox 1991b:103). Haury(1937b) also documentedthat rubberballs,probablymadefromguayule,a plant nativeto the ChihuahuanDesert, were presentin the Hohokamterritory. Some of the largestof the Hohokamballcourts were built along the middle Gila River at Snaketown (Haury1937a) and Grewe (Marshall2001) andhad embankmentsthatstood 2.5 m abovethe surroundingterrainand stretchedaroundan area The Hohokam Regional System more than 60 m long and 33 m wide (crest to The Hohokamare best known for theirhydraulic crest). Doyel (1991b:249) has estimatedthat as infrastructure of ditchesandcultivatedfields,some manyas 500 spectatorscould havebeen seatedon presenceimplied a sharedset of religiousbeliefs acrosstheregion,whichprovideda basisfor social and economic interaction (e.g., Wilcox 1991a; Wilcox and Steinberg 1983). At last count, 238 ballcourtsat 194 sites have been identified(Marshall 2001:120). Crowdsof spectatorsandparticipantsin the ritual ballgamespresumablygatheredfromnearand far, providing an ideal venue for barter and exchange. Indeed, theoristshave long suspected thatballcourtfestivitieswereconduitsforthemovement of various commodities, leading them to hypothesizethatthe largegatheringsaffordedthe opportunityfor tradefairs and incipientmarkets (e.g., Bayman 2002; Doyel 1979, 1985; Haury 1976:78;Heidke2000;Wilcox 1991a;Wilcox and Sternberg1983:213).Inthispaper,we explainhow recent ceramicevidence has supportedthis idea. The potterydatasuggestthattheremusthavebeen a reliable and efficient mechanism during the eleventhcenturyfor distributinglargenumbersof clay containers from concentrated production sources.We concurwith those theoristswho suggest thatperiodicmarketplacesassociatedwithritual ballgamesare good candidatesfor explaining thedistributionpattern.However,theceramicfindings also imply a level of dependenceon marketplacebarterfor basic necessities,like earthenware containers,thatis greaterthantypicallyassociated with nascent marketplaces. Concerned by the imperfectfit betweenthe Hohokamevidence and extantmodelsdrawnfromethnohistoricandethnographicevidence, we asked ourselvesif the marketplace hypothesis for the Hohokam is supportable.We have respondedwith additional ceramic information that temporally links the demise of the Hohokamballcourtswith a transformationin theorganizationof potteryproduction anddistribution. Finally,we considersomeunusual circumstancesthat characterizedthe Hohokam regionalsystem and offer some tentativepropositions, which, if correct,imply that the variability among emergent economies, like that of the Hohokam,exceedstheprevalentmodelsof nascent marketingnetworks. Abbott et al.] BALLCOURTS AND CERAMICS 463 Figure 1. Canal Systems and major Hohokam settlements in the lower Salt River valley (based on information in Howard and Huckleberry 1991:Figure 1.1). the berms in stadium-likefashion to watch the game below. During subsequentcenturies,the network of ballcourtsin southernand central Arizona was expandedgreatly,as mostlycourtsof a smallertype were built on the edges of the central plazas in nearlyeveryvillagein theHohokamterritory.DuringthemiddleSedentaryperiod(A.D. 1000-1070), whenthe networkhadgrownto its greatestextent, the densest distributionremainedin the Phoenix basin,butmorewidely spacedlines of courtsparalleledthe watercoursesthatextendedin all directions from the centralclusters (Figure 2). These linesreachedintovariousecologicalzones,including non-Hohokamlandsinhabitedby the Sinagua, Cohonina,andmiddleVerdeculturalgroupsin and aroundthe Flagstaff and Prescottareas (Wilcox 1999;Wilcox et al. 1996). The irrigatedvalleys were highly productive agriculturallybut were poor in naturalresources. In contrast,the surroundinguplandsto the north, east,andsouth,andthe desertlowlandsto the west werejust the opposite:marginalagriculturallybut environmentally diverse and rich in natural resources.Not surprisinglythen,rawmaterialsand finishedproductsmade from obsidian,turquoise, argillite,galena, steatite,serpentine,and shell, as well as manos, metates,tabularknives, wild vegetablesand fruits,large-bodiedgame, and probably hideswere importedfromthe outlyingareasto the villages lining the banks of the Salt and Gila rivers.Paintedred-on-buffceramicsandprobably surplusesincludingcottonmovedin the agricultural directions (e.g., Doelle 1980; Doyel opposite Kwiatkowski and Gasser 1991; Teague 1991a; 1998). Most Southwesternarchaeologistsagree thattheballcourtgatheringsfacilitatedthe transfer of the richdiversityof goods fromvariousecological districtsacross considerabledistances.This connection between ballcourts and regional exchangehas become a centraltopic of research and is consideredby many analyststo be of the utmostsignificancefor comprehendingHohokam prehistory(e.g., CrownandJudge,eds. 1991). Previouswork has establishedthat the assortmentof tradegoods was impressive,althoughwith a closer look we see thatthe mix of materialculturevariedfromone end of the regionalsystemto the other.The variationimplies that the network lacked an overarching,centralizedcontrol (e.g., redistributionfrom a centralplace) thatregulated the exchange transactionsand other associated activities (Bayman 2002:70; Crown 1991:402403). Instead,thenetwork'sdecentralizedstructure was a patchworkof communities,horizontallytied together for the benefits afforded by the farreaching interaction(e.g., Crown 1991: Wilcox 1979). The Hohokampresentresearcherswith a case studythatcan informon how largequantities of goods can be transferredin the absenceof overarchingpoliticalunity. therearevarBeyondthatgeneralconfiguration, ious butoverlappingideasabouthow the ballcourt network and the movement of goods throughit 464 AMERICAN ANTIQUITY [Vol. 72, No. 3, 2007 Figure 2. Distribution of Ballcourts in Arizona (based on information in Marshall 2001:Figure J.l.)« toproducegoodsandservices shouldbe characterized.In one of the best devel[T]heobligations forceremonies,whenharnessedto a scheduled oped of these models, Wilcox conceptualizesthe roundof ceremonialevents among a set of ballcourtceremonialismand some of the associated exchanges as embedded in social arrangenearbycommunitieswouldhavecreatedperitocarry odicpoolsof goodsandthemotivations mentsandritualobligations.He proposedthatthe themto anotherplaceas giftsor presentations ballcourteventsfacilitatedthecirculationof exotics in a continuingseriesof reciprocalexchanges and other goods among emergent local leaders, who usedthe commoditiesto buildandmaintaina [WilcoxandSternberg1983:213]. constituency (Wilcox 1991b:123-124). He pre- Inaddition,he also positedtheexistenceof regional sentedevidencethatthe ballgameswere calendritrade fairs associated with ballgames to supplecally timed (Wilcox 1991b: 115-122), and ment exchangescontingenton kinshipconnections suggestedthat: (Wilcox 1991b:123-124). Abbott et al.] BALLCOURTS AND CERAMICS Doyel(1979, 1981:58,1985, 1991a,1991b:252) has offeredan economic perspective.He has also proposed that trade fairs or incipient markets accompaniedtheballgames,emphasizingthatpeople of all socialstatusesenjoyedfull anddirectparticipation in the exchange of goods. By this reckoning,the games themselveswere clearlyritualaffairs.Undera ceremonialumbrella,however, bargainingandbarterof a purelyeconomicnature tookplace.Periodicmeetings,wherelargegatherings formfor ritualperformances,inevitablyfacilitatesecularintercourseas well, which,as Renfrew (2001:19, 22) notes, can sometimesincludeperiodic markets.Recent advancesin the Hohokam ceramicchronology,pottery-provenance research, and subsequentfindingsconcerningthe organization of ceramicproductionanddistributionduring the middle part of the Sedentary period have stronglysupportedthe marketplaceidea. Wallace's Refinements of the Red-On-Buff Typology 465 MiddleSacaton2 Red-on-buff." Therecanbe many such combinationswith differentdegreesof temporalprecision. Pottery Provenance Also in recentyears,a significantadvancementhas been achievedin the analysisof Hohokampottery fromthe lowerSaltandmiddleGilaRivervalleys. Previousworkhadnotedmineralogicalvariability in the temperof Hohokampots and suggested a provenancerelationshipbetweentempervarieties andproductionsources(Abbott1983, 1988;Doyel and Elson 1985; Lane 1989;Weaver1973). New compositional and geological studies, begun at Pueblo Grande(Abbott 1994, 2000; Abbott and Schaller1994;Schaller1994) andcontinuedelsewhere(Abbott1993, 1995,2001a;AbbottandLove 2001; Miksa 1995, 2001; Miksaet al. 2004), have subsequentlyextendedand refinedthe documentation of tempervariation,codified it in termsof geological zones, and demonstratedthat the prehistoric artisansrelied on the raw materialsthat werenearestathand(Figure3). Consequently, tempertype has been shownto be an excellentindicator of productionsource. Because the tempertypes are mineralogically distinctive,theyaredistinguishablewithjusta standardbinocularmicroscope,makingit possible to determinethe provenanceof individualceramics quicklyandinexpensively.As a result,tensof thousandsof ceramicshavebeenlinkedto spatiallynarrow productionsources, and pottery exchanges have been tracedwith remarkableprecision.The movementof earthenwarevessels betweeninhabitantswho lived as little as 5 km apartis now recognized,demonstratingthatlargenumbersof clay containerschangedhands.By tracingtheseceramic exchanges,the organizationof potterymanufacture,thespatialandtemporalpatternsof potterydistribution,andthesocialandeconomicrelationships impliedby theceramicdataarebeingreconstructed withunprecedented detail(e.g.,Abbott1993, 2000, 2003; Abbott and Walsh-Anduze 1995; WalshAnduze 1996). workforbuildingtheHohokam Themostimportant chronologysince Haury's(1937c, 1976) eminent studies at Snaketownis Wallace's (2001, 2004) recenttypologicalrefinementsof the red-on-buff ceramicsmadealongthe middleGilaRiver.Using a rigorouslydesigned, fine-scale, temporalseriationof thepainteddesignsandvessel forms,Wallace modeled the rise and fall of decorative traditions, leading to temporal subdivisions of Haury'spottery types and objective criteriafor assigningindividualceramicsto the moreprecise forourpresentation temporalcategories.Important is thefour-partdivisionof SacatonRed-on-buff,the type that Haurydefinedto representthe Sacaton phase,whichis equivalentto the Sedentaryperiod. Wallacehas been ableto dividethe type intoEarly Sacaton,MiddleSacaton1, MiddleSacaton2, and LateSacatonRed-on-buff.Ineffect,whathadbeen a single potterycategoryrepresentinga 200-year block of prehistoryhas been split into four types, each associatedwith a relativelyshortintervalof time.In the applicationof Wallace'srefinedtypology, some decoratedsherdscanbe pinneddownto a specific type (e.g., Late SacatonRed-on-buff), Concentrated Production whereas others are sometimes assigned to categoriesthatencompassa smallandcontinuousrange Potterywas not madein most villages of the lower of new types. One example is "EarlySacatonto Salt River valley during the middle part of the 466 AMERICAN ANTIQUITY [Vol. 72, No. 3, 2007 Figure 3. Zones of sand and rock types in the lower Salt River valley: la = basalt, Phoenix Mountains; Ib = basalt, Fountain Hills; II = phyllite; III = Squaw Peak Schist; IV = Camelback Granite; V = quartzite; VI = quartzite and schist; VII = South Mountain Granodiorite;VIII = Estrella Gneiss; IX = Usery Mountains sand (based on information in Abbott 2000:Figure 5.1). Sedentaryperiod.Instead,morethan90 percentof the bowls andjars in the valley householdscould be accountedfor by the outputof artisansresiding in only five pottery source areas.Three of these sourcesexisted in the lower Salt Rivervalley and twootherswerepresentalongthemiddleGilaRiver to the south. The first production source, associated with phyllitetemper,has been pinpointedat Las Colinas, a sprawlingsettlementat the terminalend of CanalSystem2 (see Figure1). Onthewesternedge of the site, thick layers of well-sortedclay were repeatedlyaccumulatedin and mined from two huge settlingbasins(approximately10 m in diameterand5 m deep)intowhichcanalwaterwithsuspendedsedimentshad been divertedand allowed to dry (Nials and Fish 1988). The levigated clay was ideal for fabricatingthousandsof mediumandlarge-sizedplain-warejars (probablyused for cookingandstorage)andlesser numbersof plainware bowls, which were distributedthroughout CanalSystem 2 andthe ScottsdaleCanalSystem. The secondandthirdsupplyzones werelocated on the southside of the Salt River.The two halves of SouthMountainwerethebedrockoriginsof two distinctive sands (South MountainGranodiorite and EstrellaGneiss; see Figure3). Local artisans used the sandtypes to producelarge,wide-orifice, thick-walled ollas, which possibly functionedas water coolers. These vessels were widely traded acrossthe lowerSaltRivervalley,includingto Las Colinas,wherejars of this shapeandsize were not made.Becausetwo discretesourcesfor these pots can be discriminatedby their temper, we have inferredthatat least two groupsof potterslabored to makethem.Wehaveno knowledge,however,of precisely where and at how many places in the South Mountain vicinity these producers were located. Manufacturersof plain-warecontainers also workedin the middleGila Rivervalley,supplying nearlyall of the medium-and large-size jars and some bowls to the Phoenix-areahouseholdson the south side of the Salt River.Many of these Gila Abbott et al.] BALLCOURTS AND CERAMICS 467 Figure 4. Distribution of Red-on-buff Jars and Bowls during the Middle Sedentary Period. Riverimportswerelikewisetransported to the residentsof theScottsdaleCanalSystem.Ineffect,the middleGila plain-waremakerssuppliedthe communitieson the south side of the Salt River with the samevessel formsmadeat Las Colinas,which weredistributedto the settlementson thenorthside of the river.Only in the ScottsdaleCanalSystem do we finda significantoverlapof the plainwares from Las Colinas and the middle Gila River valley.The middleGilaceramicsweretemperedwith coarse-grainedmica schist,whichoutcropsat severalbedrockexposuresthatencirclethe valley and exist along Queen Creek, a short distanceto the north.Consequently,we do not knowthe extentto which plain-wareproductionwas concentratedin thatsouthernpartof the Phoenixbasin. Finally,thosepottersspecializingin red-on-buff bowls and smalljars were also situatedalong the middleGilaRiverat Snaketownandprobablyelsewhere in that vicinity (Figure4). Haury's(1976) excavationsat Snaketownunearthedthe facilities, tools, andrawmaterialsforpotterymanufacturein a workareanearthecenterof the site. Experiments with those raw materialsdemonstratedthat they were intendedfor buff-waremanufacture(Abbott and Love 2001). Analogous evidence has been reportedfromthe MaricopaRoad site, a few kilometers to the west of Snaketown(Lascaux and Ravesloot 1993), and extreme densities of buffwarepieces areknownfromthe surfaceof the Gila Butte site, suggestingbuff-wareproductionthere, about4 km east of Snaketown(Rafferty1982). Nearly all of the bowls in use in the lower Salt River valley duringthe middle Sedentaryperiod were decoratedforms, and the small red-on-buff jars complementedthe largerplain-ware containers in size and aperturediameter,with little overlap. The buff-warepots were also temperedwith coarse-grainedmica schist,butapproximatelyhalf the decoratedceramicsthatreachedthe lowerSalt River valley also contained a sand component. Extensivegeological analysis of the sands in the middleGila Rivervalley has yielded a petrofacies model that demarcatesseveral sand-composition zones (Miksa 2001; Miksa et al. 2004). Recent analyses of the sand fractionin severalhundred 468 AMERICAN ANTIQUITY [Vol. 72, No. 3, 2007 weremadeby specialistsforbroadscaleexchange. We define"specialistproducers"as those artisans who practicedtheircraft,at least in part,to supply theirwaresbeyondtheirown householdsto external consumers. Specialized production and exchange have been proposed for several commoditytypes, althoughthe supportingdataarenot as well-developedas theyarefortheceramics(e.g., shell jewelry [A. Howard1993], projectilepoints [Hoffman1997], manos and metates[Hoffmanet al. 1985],tabularknives[Bernard-Shaw 1983],axes [Doyel 1991a],andsome kindsof plantfoods and fibers [Fish et al. 1992; Gasserand Kwiatkowski 1991; Hutira1989; Teague 1998]).1The one-way flow of potteryfromthe middleGila Riverpotters to the communitiesalongthe lowerSaltRiveralso implied that large quantitiesof othergoods were moved in the oppositedirection.At Grewe,in the heartof the middleGila area(see Figure1), it was recentlyestimatedthata paltry.03 percentof the pottery(i.e., 3 sherdsper 10,000)originatedin the lower Salt River valley (Abbott 2001b:265). In contrast,probablythousandsof pots each year, roughlyaccountingfor half of all ceramicsalong the lower SaltRiver,were importedfromthe middle Gila sources.2 Second,whenconsideringthehugenumbersof consumersrelativeto the muchsmallernumberof potterysuppliers,we recognizethatthe clay pots musthavebeendistributed beyondthelimitsof kinship networksand othersocial ties. Artisansmust havemanufacturedsome of theirwaresfor people they did not know. For instance,the uniformdistributionof a standardset of red-on-buffceramics over a large territory(Abbottet al. 2001) clearly differedfromexpectationsfordown-the-linetransactions typically associatedwith reciprocalkinbased exchange. Presumably,the trade in other commoditiesalso transcendedthe connectionsin the social field. Third,the ceramicresultsimplythatthe Sedentaryperiodwas markedby an efficientandreliable mechanismfor commoditydistribution.One idea to considerwas a politicaleconomyarticulatedby Economic Implications dominatingelites(e.g.,Earle1997:70-75;Feinman Theceramicresultshaveatleastfourobviousimpli- et al. 1984:169;Johnsonand Earle2000), but the cationsforhow to conceptualizethemiddleSeden- available evidence offers little support.In such taryperiodeconomy.First,whereproductionwas economies, an obligatoryproductionof surplus concentratedfor one kind of commodity(i.e., the goods and labor was paid to a centralfigure for pots), it standsto reasonthatotherkindsof goods financingthe activitiesof governinginstitutions. Sedentaryperiodbuff-waresherdsrecoveredfrom varioussites along the lower Salt Riverhave tentativelyshown thatmost of these specimenscontainedsandfromthe Snaketownarea(Abbottet al. 2007). These findingssuggestedthatthose middle Gila potterswho suppliedpaintedpots to the Salt River communitieswere mostly concentratedat and aroundSnaketown,on the northbank of the Gila River. Despitethisconcentration,a standardset of redon-buffvessel forms was uniformlydisseminated fromthemiddleGilaRiverregionto domesticunits throughoutthe lower SaltRivervalley.Duringthe middleSedentaryperiod,buff-waresherdsconsistently accountedfor about20 percentof the total ceramic assemblagesin all parts of the Phoenix area. Moreover,the ratios of decoratedbowls to decoratedjars andthe relativenumbersof red-onbuff bowls of particularsizes were repeatedfrom each locationin the lower Salt Rivervalley to the next (Abbottetal. 2001). Wenow recognizethatceramicproductionduring the middle Sedentaryperiod was organized acrossanexpansiveterritoryso thatartisangroups, sometimesseparatedfrom one anotherby one or two days'travel,each fabricateda restrictedset of waresandvesselformsthatoftenfunctionallycomplementedthecontainersmanufactured by theother pottergroups.In this way, each Salt Riverhousehold was dependenton multipleand often distant producersfor the full complementof its domestic potteryinventory.This concentrationof production could not havebeen moredissimilarfromthe patterndocumentedfor the subsequentearlyClassic period(ca.A.D. 1100-1 300), whena full range of vessel forms was produced by local potters throughoutthe lower Salt River valley (Abbott 1993, 1995, 2000, 2003). This shift implieda radically reducedscale of productionduringthe later time periodwith a greateremphasison self sufficiencyandless dependenceon distantsuppliersfor basic domesticneeds. Abbott et al.] BALLCOURTS AND CERAMICS This mechanism of accumulationand reapportionmentprobablydoes notexplainthedistribution of Hohokampottery.Inotherculturalsettings,elites had little interest in a monopoly control over domestic pottery production and distribution andRowlands1978;Fry 1980;Rice (Frankenstein 1987),probablybecauseit wouldhavebeen especially difficultto attain.Cheaplymade utilitarian items thatcould be and often were manufactured across wide areaswere not the stuff from which political hierarchieswere built (Costin 1991:11; cf. Earle 1987). controlover FortheHohokam,anadministrative potterymakingprobablywouldhavebeen impossible becausethe clays andtemperfor fabricating pots were some of the few rawmaterialsfor crafts productionthatwerewidespreadin the valleylowlands (see Abbott 2000; Abbott and Love 2001; Miksa 2001). The knowledge and skill to utilize these materialswere probablyeasy to acquireas well, as witnessedduringthe Classicperiodwhen pots were made in all sections of the lower Salt Rivervalley (Abbott2000). Moreover,we findno evidence for stockpilingat redistributivecentral placesduringthe Sedentaryperiod.In fact, as discussed above, we know of productionduringthe Sedentary period outside the most politically prominentvillages (e.g., large headgatevillages thatmayhavecontrolledwaterdistribution through the canals), which contradictsexpectations for interestedin controllingmanufacadministrators tureby locatingit in administrativecenters. Fourth,we can rejectthe idea thatpotteryproductionwas regulatedas partof a tributeeconomy. Costin( 199 1:11) has notedthatthe outputof independentspecialistswas usually governedby the generalprinciplesof supply and demand,where costs servedto regfabricationandtransportation ulatedistributionfrom makersto consumers.She also noted, however, that politics and taxation sometimesinfluencedtheiractivitiesto fulfilltheir obligations.In the case of potteryproductionat LasColinas,we couldentertainthepossibilitythat thelocal artisanssuppliedcookingandstoragejars to the othervillages in theircanal system as tribute or in exchangefor assuredwaterdeliveriesto theirfields at the tail end of the irrigationworks. LasColinas,however,notonly suppliedCanalSystem2 withutilitarianvesselsbutalsopartiallyfilled the householdinventoriesin the ScottsdaleCanal 469 System.The Scottsdalevillagesimportedthe same vessel formsfromthe middleGilaplain-waresuppliers as well. In fact, the ScottsdaleSystem was theone portionof thelowerSaltRivervalleywhere a significantoverlapexistedfor the distributionof both the Las Colinas and middle Gila plain-ware ceramics,and it was the place within the supply areas of both producergroups that was furthest from the source.In the Scottsdalearea,both potter groups incurredtheir greatest transportation costs, andpresumablytheydirectlycompetedwith one anotherin thatvicinity (Abbott2001c). From theperspectiveof the ScottsdaleCanalSystem,the potteryproductionat Las Colinas and along the middle Gila appearedto have been an economic undertakinggovernedby the forces of supplyand demandratherthan the dictates of a controlling centralauthority.Moreover,productionas tribute poorly explains the massive importof clay containersfromthe middleGila throughoutthe lower Salt River valley. Political hegemony that influencedthedistributionof largenumbersof clay containersat an intervalleyscale duringthe Sedentary periodseems farfetched. Marketplace Hypothesis A likely candidatefor the mechanismthat facilitated commodity distributionwas periodic marketplacesassociatedwith the ritualballgames.As notedabove,this idea is not new. But whatis new concernsthe greatquantitiesof materialsthatpossibly changed hands via marketexchange. Marketplacebarterhasimportantadvantagesfromboth theperspectivesof supplyandof distribution.Markets offer ideal conditionsfor producersby minimizing their transportationcosts when multiple buyersassemblein one place (Alden 1982:86-88; Belshaw 1965; Stine 1962). We can imagine Hohokampottersin the middle Gila Rivervalley packinguploadsof theirwares,walkingone ortwo daysto ballcourteventsin thelowerSaltRivervalley,whileeagerbuyersanticipatedthesemerchants' arrival.Manykindsof artisanscanalsobe attracted to a single marketplacethereby concentratinga greatvarietyof goods at a particularlocation(Hassig 1982; Hirth1998). On the distributionside, a supplier'swarescan reach sparsely settled, low-demand areas when buyersare drawnfrom the countrysideto a mar- 470 AMERICAN ANTIQUITY ketplace (Hassig 1982). This process creates a homogenizingeffect on the distributionof goods producedin limitedareas(Hirth1998),exactlylike the uniformdistributionof Hohokamred-on-buff potteryin the lower Salt River valley duringthe middle Sedentary period (Abbott et al. 2001). Finally,participantsin marketexchangerequireno priorkinshiprelationshiporothersocialtie in order to transactwith one another(Carrasco1983;Hirth 1998). As noted above, barterbetween strangers almost certainlycharacterizedthe distributionof Hohokamceramics. At thisjuncture,we shouldclearlystatethatthe econHohokamdidnotinitiatea market-dominated omy,suchas thosein industrializednations,where all membersof the society acquiretheirlivelihood by selling something(includingtheirlabor)to the market.Instead,the Hohokampossibly operated whatBohannanandDalton(1962:7-9) havecalled "peripheralmarkets"within a multicentriceconomy.Distincttransactionalsphereswere probably extantamongtheHohokam,eachdistinguishedby differentmaterialitemsandservices,differentprinciplesof exchange,anddifferentmoralvalues(e.g., Abbott2000:133-140). One of thosetransactional spheresmay have been marketplaceexchangein whichtheforcesof supplyanddemanddominated. In otherspheres,"themovementof materialgoods in reciprocative[sic]andredistributive transactions cannot be understoodoutside the context of the social situationof which they [formed]an integral part"(BohannanandDalton 1962:4). Ceramicbowls andjars were basic necessities of Hohokamlife, yet householdsacrossthe lower Salt River valley did not make their own pots. Instead,each familydependedon multiplepottery suppliers, often situated considerable distances away.As such, the clay containerslead us to an interestingconundrum.The ceramicdata, on the one hand,imply thattheremust have been a reliable andefficientmechanismfor distribution,thus idea.Ontheotherhand, supportingthemarketplace if basicnecessitieswereindeeddistributedvia marketplace barter,there must have been a heavy dependenceon those kinds of transactions,which seems inconsistentwith expectationsassociated with extantmodels of nascentmarketplaces. Nascent marketingsystems, often called market rings, include uniformlydistributedmarketplaces wheretradeat each locationis periodicand [Vol. 72, No. 3, 2007 scheduledto offset with marketingdays at other nearbysites. Differentcommoditiesare available more or less equivalentlyat each location,where barteris horizontal,withoutoversightfrom some overarching,verticallystructuredinstitution.Marketringseverywherearecomposedof mostlylocal people, who are small-scaleandpart-timesuppliers, servingonly a single centeror a single market ring (Bromley et al. 1975; Hodder 1962; Smith 1974). Theirproductiondecisions are little influencedby the markets,whicharesimplyoutletsfor occasionalsurplusesof craftsand foodstuffs.For these reasons, consumers are unable to obtain necessitiesin marketrings with regularityor certainty(Smith 1976:39^4). Marketringsarerootedin a setof economicconditions,namelydirectparticipationof small-scale producers,horizontalorganization,and a considerabledegreeof local self-sufficiency.The conditionsimplya low dependenceon marketexchange and a poorly developeddivision of labor.In contrast, what we find in the Hohokam case is an advanceddivisionof laborevidentby transactions of productsthatcouldhavebeen,butwerenot,produced by self-sufficienthouseholds (cf. Blanton 1983:53; Formanand Riegelhaupt 1970; Smith 1974, 1976) and a scale and efficiencyof market distributionthat exceeded expectationsfor periodic, horizontallyorganizedmarketrings. The poor fit between the Hohokamcase and othermarketringscausesus concern,to whichwe respondwith new information.Ournew findings will not settle the issue, but they do link ballcourt of ceramicproceremonialismandtheorganization ductionanddistribution morecloselythanhadbeen accomplishedbefore,therebyaddinga new measureof supportto thepropositionthattheritualballgames were associatedwith marketplaces. Supporting The Marketplace Hypothesis Our results are based on Wallace's(2001, 2004) refinementsto the Hohokamceramicchronology and on precise controlsover potteryprovenance. Southwesternarchaeologistsgenerallyagree that mostif notall of theHohokamballcourtshadfallen into ruin by the Classic period, and the ceramic provenancedata from the Classic period in the lowerSaltRivervalleyalso implythatpotteryproductionhadrevertedto the local manufactureof a Abbott et al.] BALLCOURTS AND CERAMICS full rangeof vessel formsby thattime. The question thatwe posed was whetheror not those two developmentswere linkedsynchronously. Webeginwith a reviewof the evidencethatthe ballcourtnetworkcollapsedrathersuddenly,probably nearthe end of the middleSedentaryperiod. Wethenpresentnew evidencefromLasColinasto supportthatclaim. Then, using ceramicpercentages and provenancedata from Las Colinas and elsewhere,we show thatthe organizationof pottery productionand distributionchangedconsiderably,preciselyatthetimethattheballcourtswere abandoned. Collapse of the BallcourtNetwork Thefirstapplicationof Wallace'srefinementsto the red-on-bufftypologyimmediatelyyielded significantresults.EvenbeforeWallace'sstudyappeared in print,his typologicalcriteriawereappliedto the Sedentaryperiodassemblageat Palo VerdeRuin, located in the uplandzone northof the Phoenix basinandapproximately35 kmfromtheSaltRiver (see Figure 2). The Palo Verde analysis demonstratedthatpartof a stringof single-familyfarmsteads,parallelingthe west bankof the New River, existedon the easternmarginof the site duringthe earlypartof the Sedentaryperiod.Then,quitesuddenly, a large Hohokam village, complete with courtyardgroupsof pithousesclusteredinto residence units with associatedcemeteries,was constructedarounda centralplazaandballcourtduring the middleportionof the Sacatonphase.The ballcourt village was short-lived,being abandoned priorto the inceptionof LateSacatonRed-on-buff (Abbott2002a). WithoutWallace'stypology, we wouldhaveneverdiscernedthis rapidtransformationof the site structurein themiddleof the Sedentaryperiod.In addition,an abundanceof tree-ring dated intrusive ceramics and a large suite of absolutedatestentativelyindicatedthatthe middle Sedentaryballcourtvillagewasestablishedaround A.D. 1010-1020 andsecurelydatedthesite'sabandonmentby A.D. 1070 (Abbott2002a). A chronologicalreassessment(Abbott2002b) of the ceramicdatafromvarioussmall sites previously investigatedalong the New Riverby Doyel and Elson (eds., 1985) implied that the ballcourt village at Palo VerdeRuin representedan aggreTheabsolutedatingsupgationof localinhabitants. Wilcox's (1999) hypothesis for the ported 471 expansionof the regionalballcourtnetwork,when the Palo Verdeballcourtbecame an intermediary node in a rapidly expanding ceremonial and exchangenetwork.It helpedto link the Hohokam core area with otherethnic groups in the middle Verde,Cohonina,and Sinaguacultureareasto the north(cf. Wilcoxet al. 1996). The abandonmentof PaloVerdeRuinby A.D. 1070 impliedthe loss of those northernconnections. Doyel (2000) has also describeda majordisruptionof theHohokamregionalsystembeginning at A.D. 1075, which correspondedwith the abandonmentof Snaketownbasedon the presenceand absenceof tree-ring-datedintrusiveceramictypes (Doyel 1993). In additionto the abandonmentof thetwoballcourtsat Snaketown,Doyel pointedout thattheballcourtsatGatlinandPuebloGrande(see Figure2) were also abandonedat aboutthe same time, suggestingto himthatthatballcourtnetwork haddisintegratedaroundA.D. 1075 andno longer functionedas a key element of horizontalsocial integration.Doelle and Wallace (1991:319-321) havealso suggesteda ballcourtcollapsein theTucson Basin wheretherewere no settlementswhose occupationswere spatiallyfocused on a ballcourt by the beginningof the Late Rinconphase,probably datingto aroundA.D. 1100 (HenryWallace, personal communication 2003) - an idea supportedby Heidke (1996:62). To these instances, we can now addthe demiseof the PaloVerdeRuin ballcourtat A.D. 1070. Admittedly,only a handful of Hohokamballcourtshave been excavated andevenfewercourtsarewell-dated,buttheavailable evidence does suggest thatthe ballcourtnetworkwas in rapiddecline,if not totallycollapsed, by the end of middle Sacatontimes, aroundA.D. 1070. Dating the Ballcourtat Las Colinas Duringthe excavationsat Las Colinasin the early 1980s,the fieldcrewencountereda ballcourtfilled with dense depositsof soil andculturaldebristhat includedan abundantquantityof ceramicpieces. Below the uppertrashfill, they removeda lower stratumin whichfew artifactswerepresent.Based on our applicationof Wallace'srefinedbuff-ware types, we suggestthatthe upperstratumaccumulated duringthe late partof the Sedentaryperiod, indicatingthatthe courtwas abandonedand partiallyfilledpriorto thattime (Table1).Amongthe 472 [Vol. 72, No. 3, 2007 AMERICAN ANTIQUITY Table 1. TemporallyDiagnostic Buff WareSherdsin the Las Colinas Ballcourtand Associated Pits. Feature Ballcourt(upperstratum) 4024 (intrusivepit) 4025 (possible intrusivepit) 4029 (possible intrusivepit) 4178 (pit in ballcourtfloor) E Sacaton- E Sacaton- Middle M Sac 1 M Sac 2 Sacaton 1 11 12 few buff-warespecimens saved from the lower deposits,nonecouldbe placedin a temporallydiagnostic category. In addition, a large trash pit (Feature4024) intrudedthe ballcourtfill andtwo otherpits (Features4025 and4028) probablyintrudedthe court's embankment,which was not preservedon the historicsurface(see Gregory1988a:273).All threepits containeddecoratedceramics that indicatedthe featureswere filled duringor after Late Sacaton times(see Table1).Feature4178 was a sherd-lined pit that in contrastprobablywas coeval with the ballcourt's use, possibly marking the division between the playing floor and the northernend court(see Gregory1988a:272-273).The sherdlining was composed, in part, of decorated bowl sherds,oneof whichwe identifiedas pre-LateSacaton.An abandonmentdateduringthemiddleSacaton2 intervalseemedlikelyfortheballcourt,which, interestingly, precisely agreed with Teague's (1988:56) assessment,publishedin the Las Colinas projectreport.Based on the presenceof redslipped sherds in the upper stratum and their absencein the lower stratumof the ballcourtfill, Teague suggested that the court was abandoned beforethe late end of the Sedentaryperiod. Dating the Reorganizationof PotteryProduction Previousevidence has shown that the supply of buff-warevessels fromthe middleGila Rivervalley declined duringthe Sedentaryperiod from a high of roughly 20 percent of the total ceramic assemblageto a low of about5 percentin the Classic periodassemblages.Moreover,thatdeclinehas been characterizedas a reductionin the manufactureof paintedbowls.The earlierworkat Las Colinasestablishedthatanincreaseinplain-warebowls replacedthe shortfallin buff-wareforms during the Sedentaryperiod(Abbott1988).We havenow recognizedthatthe shift occurredduringthe Mid- Middle Sacaton 1/2 4 3 1 pre-Late Sacaton 4 11 Late Late Sacaton/ Sacaton Casa Grande 12 1 4 1 dle Sacaton2 times and that the decline in buffwarepercentageswas not a gradualandprotracted of thebuff-ware declinebuta one-timeinterruption the timingof thatinterruption supply.Importantly, preciselycorrespondedto the demise of the ballcourt. We evaluatedthe sherdcountsfor each feature, which allowedus to assignsome featuresto a specific temporalcategory(e.g., MiddleSacaton2) or to a slightlylongertemporalinterval(e.g., Middle Sacaton2/Late Sacaton).In addition,there were numerousinstanceswheretheceramicset included specificallydatedsherdsthattogetherrepresented threeor moretemporalunits (e.g., an assemblage containinga Middle Sacaton 1 sherd, a Middle Sacaton2 sherd,and a Late Sacatonsherd).We interpretedthese cases as featureswithtemporally mixed deposits. Usingthedatapublishedin theLasColinasproject report(Gregory1988b),we thencalculatedthe percentageof buff-ware ceramicsfor each of the features in our sample (see Table 2). When we excluded the temporallymixed assemblagesand plottedthe featuresby buff-warepercent,we got an unexpectedresult(Figure5). Insteadof a continuousdecline over time, thereseemedto be two somewhatvariablegroups,with a distinctgap in the distributionof buff-warepercentages,between 12 and 17.5percent.Thisgapwas a resultto which we assignedconsiderablesignificance.Wesuggest thatit representeda disruptionin the supplyof redon-buffcontainersat Las Colinas. Small relative frequenciesof decoratedsherdsexistedin all of the MiddleSacaton2/LateSacatonfeaturesandall of the Late Sacatonassemblages.In contrast,all of the Middle Sacaton 1 and all but one of the Middle Sacaton1/2 featurescontainedmorethan 17.5 percentbuff ware.The one exceptionwas Feature 4029, fromwhichfivetemporallydiagnosticsherds were collected: four cases classified as pre-Late Sacaton and one piece that was either a Middle Abbott et al.] BALLCOURTS AND CERAMICS 473 Table 2. CeramicData from Contextsat Las Colinas. Plain WarePercent 4254 22.3 South Middle Canal Mt.b Gila R.c System 2d Early Sacaton/MiddleSacaton 1 46.9 38.8 2.0 2.0 4001 4122 4124 4125 4179 4262 5066 5103 7089 25.4 18.7 17.8 22.7 21.1 27.9 28.1 17.5 19.8 34.0 58.0 52.7 62.0 53.8 48.0 43.1 51.0 . Middle Sacaton 1 50.9 9.4 30.0 4.0 14.5 0.0 16.0 0.0 23.1 7.7 40.0 4.0 33.3 7.8 19.6 3.9 . . 4000 4029 4178 4253 5314 7021 7023 21.1 7.8 21.3 20.2 18.2 18.7 18.6 37.7 62.7 70.0 43.1 50.0 42.0 66.7 Middle Sacaton 1/2 43.4 7.5 13.7 0.0 20.0 5.0 35.3 7.8 38.2 5.9 34.0 4.0 18.5 3.7 0.0 5.9 0.0 2.0 2.9 6.0 0.0 0.0 3.9 0.0 0.0 0.0 0.0 0.0 11.3 13.7 5.0 11.8 2.9 14.0 11.1 53 51 20 51 34 50 54 4091 4215 5034 5073 5116 5126 7.5 10.7 19.9 9.2 11.7 18.5 83.3 53.1 56.9 26.4 66.0 63.5 Middle Sacaton 2 10.4 0.0 24.5 4.1 27.5 2.0 15.1 5.7 13.2 0.0 21.2 5.8 4.2 2.0 0.0 22.6 9.4 0.0 0.0 2.0 0.0 0.0 1.9 0.0 2.1 14.3 13.7 30.2 9.4 9.6 48 49 51 53 53 52 4025 4150 4280 8.1 9.5 7.6 60.8 64.0 35.3 27.5 18.0 13.7 9.8 6.0 7.8 0.0 0.0 5.9 0.0 12.0 31.4 51 50 51 4019 4060 4250 4273 5038 11.2 8.6 9.5 11.2 11.1 51.0 64.7 64.3 42.0 36.0 8.2 11.8 10.7 14.0 10.0 4.1 5.9 1.8 8.0 6.0 20.4 0.0 16.1 10.0 34.0 2.0 0.0 0.0 2.0 6.0 14.3 17.6 7.1 24.0 8.0 49 51 56 50 50 5074 1O5 23A 46,2 O0 1L5 7/7 11.5 42 Feature Buff % Local3 Scottsdalee Otherf Total 0.0 10.2 49 0.0 0.0 1.8 0.0 2.6 0.0 2.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 ... 5.7 8.0 30.9 22.0 12.8 8.0 13.7 3.9 53 50 55 50 39 50 51 51 Middle Sacaton 2/Late Sacaton 2.0 0.0 5.9 Late Sacaton aphyllite bSouthMountainGranodiorite,EstrellaGneiss ccoarse-grainedmica schist dCamelbackGranite,Squaw Peak Schist equartzite,quartziteand schist fotherschist, unidentified. Sacaton1 ora MiddleSacaton2 Red-on-buffsherd. Based on the low percentage of buff-ware, we stronglysuspectthatthis featureactuallydatedto the MiddleSacaton2 interval.Finally,the Middle Sacaton 2 features straddledthe gap, with four cases to the left and two cases to the right, suggesting that the disruptionin buff-ware supply occurredduringthattemporalspan. AMERICAN ANTIQUITY 474 [Vol. 72, No. 3, 2007 Figure 5. Buff Ware Percentages in Features of Different Time Intervals at Las Colinas. Consequently,the disruptionin buff-waresupplies fromthe middleGila Rivervalley transpired at Las Colinasexactly at the time the Las Colinas ballcourtwas abandonedand when the regional ballcourtnetworkis now thoughtto have been in rapiddecline.As we discussnext,thoseeventswere coeval with a reorganizationof plain-wareproductionas well. Las Colinas Plain Ware We beganwith the plain-ware sherdsat Las Colinasrecoveredfromthe3 1 Preclassicstructuresand pitsthatwereeithercoevalwiththe ballcourt's use or postdatedit.3 From each of these features,we randomlyselectedabout50 plain-waresherdsand carefully reviewed each sherd set to eliminate pieces fromthe samevessel (see Table2). Locally madecases containedcrushedphyllitefragments. The ceramicsmadein the centralandeasternportionsof CanalSystem2 weretemperedwithSquaw Peak Schist and CamelbackGranite,respectively. Thosespecimensmadein theScottsdaleCanalSystem containeda quartzite-richsand or a sand of mixed quartziteand schist. Othercases included those tempersrepresentingproductionin the two SouthMountainareas(SouthMountainGranodiorite,EstrellaGneiss),manufacturealongthemiddle GilaRivervalley (coarse-grainedmica schist), and a few cases with undifferentiatedtempering material. Forour analysis,the percentagesof plainware with eitherSquawPeak Schist, CamelbackGranite, quartzite,orquartziteandschistin eachfeature was of interestbecauseceramicswith these "eastern"tempertypes were not madein the lowerSalt Rivervalley duringthe middle Sedentaryperiod. As discussedabove, nearlyall of the ceramicsin usein thevalleyduringthemiddleSedentaryperiod were made either at Las Colinas, in the South Mountainarea,or along the middleGila River.In contrast,bowls andjars were producedin abundance throughoutthe valley during the Classic period,includingthe upperreachesof CanalSystem 2 and the ScottsdaleCanalSystem, and were commonly exchanged among the valley inhabitants at that time (see Abbott 1995, 2000). We variwantedto knowwhenthe"eastern-tempered" eties fromthecentralandeasternportionsof Canal System 2 and the ScottsdaleCanal System were first importedto Las Colinas, which presumably was the time when plain-wareproductionbecame localizedacrossthe region. Not surprisingly,we found only tiny amounts of the eastern-temperedvarieties in the middle Sedentarycontexts coeval with the ballcourt.In contrast,the contextsthatpostdatedthe ballcourt typicallycontainedconsiderablygreaterpercentages of these ceramics (Figure6). Based on the importsreceivedby peoplelivingatthewesternend of CanalSystem2, it appearedthatjust at the time thattheballcourtatLasColinaswasabandonedand the supplyof red-on-buffpotteryfrom producers far to the southrapidlydeclined,local production of plain-warebowls andjarsbeganthroughoutthe lower Salt Rivervalley. Pueblo GrandePlain Ware A similarresultwas obtainedfromthe plain-ware potteryatPuebloGrande.Recentexcavationsthere by Soil Systems, Inc. (Abbott,ed. 2003), at the Abbott et al.] BALLCOURTS AND CERAMICS 475 Figure 6. Percentages of Plain Ware Ceramics with "Eastern"Tempers in Features at Las Colinas. easternend of CanalSystem 2, uncovered11 late Sedentaryperiod structures,all containing late Sacatonphaseceramicassemblages(Abbottet al. 1994). The locally made plain ware contained sands,whichwere abundantin the granite-derived Classicperiodcontextsat PuebloGrande(Abbott 2000).As indicatedin Table3, theywerealsoabundant at Pueblo Grandeduringthe late Sedentary period,indicatingthatlocalproductionbeganin the PuebloGrandevicinityjust afterthe ballcourtsystem collapsednearthe end of the middle Sedentary period. Also in the late Sedentary period structureswere large quantitiesof Squaw Peak Schist-temperedplain ware from the centralportion of CanalSystem2, whereplain-wareproduction was rareor absentpriorto the late Sedentary period(see Table3). Los Hornos Plain Ware Finally,we turnedto two late Sedentaryandeight middle Sedentary assemblages at Los Hornos, which we identifiedusing Wallace'srefinedbuffwaretypology.These featureswere excavatedby ArizonaStateUniversity(Wilcoxet al. 1990) and SWCA,Inc.,(Chenaultet al., ed. 1993).The number of plain-waresherdswas few in each feature, and all pieces clearlyfrom differentvessels were includedin our analysis. Los Hornoswas situatedneartheeasternendof South Mountainand in Canal System 1, on the southside of the Salt River.The readerwill recall thatthe villagerson the southside, includingthose residing at Los Hornos, received much of their plain-warepotteryfromthe middleGilaRivervalley duringthe middle Sedentaryperiodwhen the ballcourtnetworkwas at its height. The middle Gila wares were temperedwith copious amounts of the glitterycoarse-grainedmica schist. In the eight middleSedentaryperiodassemblagesin our sample,themicaschist-tempered potteryaccounted forhalformoreof theplain-warespecimens(Table 4). In the two later contexts, which presumably correspondedto the time after the ballcourtcollapse,the middleGila Riverimportsdroppedsubstantially.In orderto adjustfor the shortfall,local artisansrampedup theirown productionof plainwarepots (i.e., SouthMountainGranodioritetemper),just as theircounterpartshaddone elsewhere throughoutthe valley. Summary Ourresultshave indicatedthatthe collapse of the Hohokamballcourtnetworkwas synchronouswith the dissolutionof a regionallyorganizeddivision of labor for the manufactureand distributionof clay containersin the Phoenix basin. Priorto the collapse,duringthe middleSedentaryperiod,artisanswho madea narrowrangeof vessel formsand laboredto supplythe householdsof the lowerSalt Rivervalley were highly concentratedin only five productionzones. Thousandsof pots were probably tradedandexchangedeachyear.Theyincluded [Vol. 72, No. 3, 2007 AMERICAN ANTIQUITY 476 Table 3. CeramicData from Late SedentaryContextsat Pueblo Grande. Plain WarePercent CentralCanal System 2b 37.8 28.6 37.2 21.8 29.4 17.6 15.0 24.9 35.6 19.3 WesternCanal System 2C 19.7 22.6 30.9 21.2 23.5 17.6 49.6 33.7 18.0 20.7 424 Total Otherd 12.2 360 30.3 591 168 13.1 614 35.7 94 8.5 23.4 687 170 23.3 34.7 710 68 28.0 780 19.1 51 17.7 47.1 786 274 13.9 21.5 2027 193 9.2 33.2 2032 239 22.6 23.8 2099 14.8 135 45.2 2233 125 110 224 2786 23^2 6,0 aCamelbackGranite,arkosicsand wih minor Squaw Peak Schist bSquawPeak Schist cphyllite dSouthMountainGranodiorite,EstrellaGneiss, quartzite,quartziteand schist, coarse-grainedmica schist, unidentified. Feature Buff % 5.7 9.0 7.6 1.3 7.7 9.0 7.3 4.7 11.9 7.0 Local3 largequantitiesof red-paintedbowlsandsmalljars, distributedthroughoutthe valley fromproduction localities on the northside of the Gila River,and thousandsmore plain-warejars deliveredto Salt Rivercommunitieson the south side of the river fromproducersto the south.But as the ballcourts atLasColinas,PaloVerdeRuin,Snaketown,Gatlin, across the Tucson basin, and elsewhere fell into ruins,probablyaroundA.D. 1070, buff-ware supplies to the Salt River householdswere abruptly reduced.At Los Hornosand probablyelsewhere on the south side of the Salt River, the coarsegrainedmicaschist-tempered plainwaresfromthe middleGila regionwere also cut, as local potters filledthe local demandsby fabricatinga full range of vessel forms to replacethe formerlyimported decoratedandplain-warecontainers.InCanalSystem 2, across the riverfrom Los Hornos,red-onbuffsupplieswerealso severelydiminishedas local ceramic productionin all parts of that area was begunanew. We suggestthatthe underpinningsof the economy, repletewith marketplacesand concentrated production,werelost,including( 1) a moralauthority that promotedcooperationacross the region, guaranteedthe peace of the market,andfacilitated barterandtradeamongstrangers;(2) regional-scale integrationthat fostered the flow of information aboutsupplyanddemandat distantplaces;and(3) a dense andwell-organizednetworkof interaction that minimizedtransportation costs. As indicated by the analysis of Classic period ceramics(e.g., Abbott 2000; Abbott and Walsh-Anduze1995), pockets of specialized productionwere reestablishedduringthattime,butneverto the impressive scale associatedwith the ballcourtsof the middle Sedentaryperiod.4 Some Tentative Propositions Now thatwe have addednew evidenceto support the Hohokammarketplacehypothesis,we return to the problemof reconcilingthe specifics of the Hohokamcase with expectationsfor nascentmarket rings. Economic anthropologistsand cultural geographershavelong notedthatmarketexchange occurs sporadicallyin all kinds of societies.Various formsandscales of marketsystemscreatedifferentconditionsfor marketparticipationfor both suppliersand consumers(Bohannonand Bohannon 1968; Bohannonand Dalton 1962; Brumfiel 1980;Haggettetal.1977:157;Plattner1975, 1985, 1989; Smith 1976:51).In the Hohokamcase, theoristshavesuggestedthata systemof marketrings was in place, where marketdays were timed in accordancewitha calendricalscheduleof ballcourt events.As thenameimplies,marketringsarecharacterizedby relativelylocal, noncentralized,overlappingmarketsthataremoreorless equivalentfor meeting local needs. Commoditiesflow horizontally, ratherthan vertically througha settlement hierarchyandwithoutconstraintby politicalbound- Abbott et al.] BALLCOURTS AND CERAMICS 477 Table4. CeramicData from Contexts at Los Hornos. Plain WarePercent Feature 12 15 16 146 157 172 485 10 174 176 Excavator Middle Gila R.a ASU ASU ASU ASU ASU ASU ASU SWCA 70.5 55.5 47.8 50.0 48.2 87.5 68.7 50.0 ASU 32.4 ASU [5,6 acoarse-grainedmica schist bSouthMountainGranodiorite cEstrellaGneiss dphyllite,Squaw Peak Schist, unidentified. WesternSouth Localb Mountains0 Middle Sacaton 1/2 16.4 11.5 28.6 6.3 37.7 7.2 25.0 17.6 17.2 20.7 0.0 0.0 13.2 10.5 0.0 25.0 Late Sacaton 37.8 594 aries.Both producersand consumershave access to multiplemarkets,andtheirregularvisits aredue as muchto socialreasonsas to economicconcerns. Noncentralizedoverlappingmarketingsystems tend to be characterizedby weakly specialized economies in which marketplacebarterhas little influenceontheorganization of production.Dependence on the marketsis minimal.And withouta developedmarkethierarchy,economicinteraction tends to rapidlydecline with distance.Goods are overshortdistances,anddecisionsabout transported productionare little effectedby the conditionsof supplyanddemandin otherlocalities.Intheseways, theHohokamcircumstancesmatchpoorlywiththe expectationsfor marketrings.Specializedproductionaccountedfor nearlyall of thepotteryin use in thelowerSaltRivervalleyduringthemiddleSedentaryperiod,therewas a heavy dependenceon the marketforacquiringbasicnecessitieslikeclay conwas tainers,anda considerable degreeof articulation evident among distantly separatedartisanswho respondedto one another'sproductionanddistributiondecisions.Thesetraitsaretypicallyassociated withmorehierarchically organizedmarketarrangements(e.g.,Alden1982;Bromleyetal. 1975;Brumfiel 1980;Hirth1998;Smith 1974, 1976). Do Hohokammarketplacesmake sense? We havearguedthatthereis mountingevidenceto indicate thatthey did exist and played a largerpartin theHohokameconomythanpreviouslyrecognized. 2.7 3J Otherd Total 1.6 9.5 7.2 7.4 13.8 12.5 5.2 25.0 61 63 69 68 29 8 38 4 27.0 2L8 37 32 But to defend that proposition,we have had to diverge from currentthinking about the modest extentto whichmarketringscanintegratea regional economy.Wemakeno claimthatwe thoroughlyand completelyunderstandhow thatshouldbe so, but we can offer some observationsand suggestions. We posit that severalimportantcircumstances werecombinedduringthemiddleSedentaryperiod in theHohokamterritoryto fostera noncentralized overlappingmarketingsystemon whichthedependencefor a varietyof goods, includingsome basic necessities,was considerable.Thesecircumstances included(1) widely acceptedreligiousbeliefs that promotedcooperationacrossanexpansiveterritory, (2) large-scaleirrigationinfrastructure capableof surplusproduction,(3) benignenvironmentalconditions that fosteredregionalinteraction,and (4) ecological diversity that encouragedspecialized production. We startwith a simple fact: one or more ballcourts stood near the center of just about every moderateto large-sized Hohokamsettlementduring themiddleSedentaryperiod.Theballcourtceremoniesprobablydrewpeople togetherfromnear andfar andmusthavebeen commonlycelebrated throughoutthe Hohokamterritory.In a palpable sense, cooperationwas writlargeacrossthe desert landscapewith the constructionand use of ballcourtsat nearlyeveryplace of consequencein the Hohokamworld. 478 AMERICAN ANTIQUITY [Vol. 72, No. 3, 2007 Anothercrucial factor to consideris the ecoIf true,the Hohokamregionalsystem,by virtue of thedensityandfrequencyof theballcourt-related logical diversityin the Hohokamregion and its events,may have been characterizedby consider- potential relationshipto specialized production. caseshaveshownthatecoableconnectivitythatpromoteda broad-scale eco- Numerousethnographic nomic integrationbasedon moralauthorityrather nomic specializationcan thrivein politicallysimthanon a hierarchicallyorganizedpoliticalecon- ple societies (Cobb 1993). In the absence of a omy of powerrelations.We postulatethatthe hin- politicalhierarchy,butin the presenceof a reliable drancesto regionalexchange and otherforms of mechanismfor exchange,increasesin productivinteractionweremetby extending ity via specialized productionfor regional scale intercommunity the moraleconomy of kinshipwith a sharedcon- distributioncanbe stimulatedin at leasttwo ways. sciousnessof religiousidentity(Rappaport1971, Both of them are dependent on pronounced 1984; Yoffee 1994). As suggested by Yoffee inequalitiesin the availabilityof naturaland agri(1994:353), similarpatternsmay have been com- culturalresourcesassociatedwith ecological hetmon in the Southwest where centralization erogeneity. remainedweak even though social and religious The firstis surplusproductionfor the purpose institutionsintegratedrelativelylarge geographic of bufferingagainst agriculturalshortages(food regions (see also Renfrew 2001:21). Among the andcotton).It is interestingto notethatthe known Hohokam,commoditiesandtheinformationabout suppliersof decoratedceramics,the craftspersons demand and prices may have efficiently flowed at andnearSnaketown,were situatedat the downhorizontally,therebyminimizingthe risk of sell- streamend of the middleGila Rivervalley,in one ing at distantmarketsand rewardingthe efficien- of the last canalsystemsin line to takewaterfrom cies of concentratedproduction. the river channel. Similarly, Las Colinas was The couplingof favorableenvironmentalcon- locatedatthe tailendof CanalSystem2. Whenthe ditions and large-scale irrigationmay have also firstLasColinasinhabitants establishedthemselves been a crucialpartof the middleSedenatryperiod there,they immediatelybegan supplyingthe rest case. Graybill's (1989) tree-ring-based recon- of theirirrigationcooperativeandthepeoplein the structionof the streamflowhistoryin the SaltRiver ScottsdaleCanalSystem with cooking pots, stordemonstratedthatthe waterwaywas a dependable age jars, and plain bowls. Perhaps,as is common benefactorof the Hohokameconomy,providinga cross-culturallyamong intensivecultivatorswho steady and predictablesource of irrigationwater. haveexcesslaborandinsufficientarablelandand/or Inturn,theirrigationagriculturesustainedlargeand water (e.g., Arnold 1985; Netting 1990:43;Rice geographicallystablepopulations,allowingthem 1987:195;Stark 1991:72), agriculturaldisadvanto enter into long-term relationships with one tages due to relativelocationin the Phoenixbasin anotherand with surroundingpopulations.The stimulatedpotteryproductionforexchangeatthose Sedentaryperiod, in particular,was a salubrious locations. intervalfor the Hohokam,duringwhich both the Second, one need only look at the maps of andthe ballcourtnetwork Hohokamcanalsystems,andespeciallythe irrigairrigationinfrastructure were expanded.And it seems to have been partof tion infrastructureon the broadvalley floor sura wider pattern. As noted by LeBlanc roundingthe lower Salt River (see Figure 1), to (1999:195-196), the A.D. 900-1 150 intervalwas appreciatethe prehistoricpotentialfor agricultural an exceptionaltime of peace and culturalflores- surpluses.As notedabove,thatpotentialprobably cence across the Southwest that coincided with variedwith locality,but,at the regionalscale, agriwarmingclimaticconditions.At thattime, a wide culturalcapacitywas considerablygreaterin the rangeof habitatswas farmed- leadingto a period irrigatedrivervalleys thanelsewhere.One idea is of raisedcarryingcapacity,littlestresson resources, that raw cotton fiber,thread,or cloth, functioned and little incentivefor warfare.It is certainlynot as a medium of exchange to facilitatethe barter happenstancethat the expansionof the ballcourt among suppliersof differentcommodities (e.g., networkand probablyits associatedmarketplace Schele and Freidel 1990:92-93; Wilcox 1987). exchangecoincidedwith a time of favorableenvi- Withoutsometypeof generalcurrency,onlytraders ronmentalconditionsand dampenedhostilities. who were interestedin each others'wares would BALLCOURTS AND CERAMICS Abbott et al.] successfullyexecute a transaction.If cotton productionwas literallya cash crop and agricultural productswere key componentsof the interaction network,thenperhapswe shouldconceptualizethe of thelowerSaltRivervalhydraulicinfrastructure ley as a centrally located economic engine that poweredthe regionalsystem.In thepresenceof an efficient mechanismfor trade,the regularavailabilityof agriculturalsurplusesmay havewhetted the appetite of regional demand, whereby the widely desirableagriculturalproductsof the irrigated valleys were exchanged for the relatively abundantnaturalresourcespresentin the desert lowlands and the upland zones surroundingthe Phoenix basin. Geographicalvariabilityand the relatedresourceinequities,whenmatchedwiththe high connectivityengenderedby the widespread acceptanceof religiousbeliefs, the regularparticipationin the ballcourtceremonialism,andfavorableenvironmentalconditionsmayhelpto explain theassociationbetweena networkof periodicmarketplacesanda specializedeconomyintegratedat a regionalscale. Conclusion Evidencehasbeengrowingin recentyearsthatthe manufactureof plain-wareand red-on-buffpots duringthe middle Sedentaryperiod was highly concentratedandthatlargenumbersof those conto consumersacross tainerswerewidelydistributed the Phoenix basin. Interestingly,those arrangementswere in effect at a time when the Hohokam ballcourtnetworkhad reachedits greatestextent, in terms of the numberof settlementswith ballcourtsandthe territoryover whichthose ballcourt Archaeologistshavelong villagesweredistributed. speculatedthat because the ballgames probably drewcrowdsfromnearandfar,theeventswerealso ideal venues for the exchange of goods. The ceramicdatacertainlysupportthatidea,and,in fact, the heavy volume of potterytransactionsimplies thatperiodicmarketplacesserveda moresubstantialrole in the Hohokameconomythanpreviously thought.Insupportof thathypothesis,we havepresentednewceramicdatafromLasColinasandelsewhere,whichdemonstrateda temporalconnection among the Las Colinas ballcourt, possibly the demise of the ballcourtnetworkin general,and a radicalreorganizationof the productionand dis- 479 tributionof clay containersin the Phoenixbasin. The new andpreviousevidence,however,have confrontedus with a conundrum.The Hohokam ballgamesarethoughtto havebeenassociatedwith horizontally organized market rings, which, as observedelsewhere,supplyconsumerswith basic necessitiesonly unreliablyandwithoutregularity. However,Hohokammarketplacesmay have been heavily relied upon by specializedproducersand consumers alike for the distribution of goods, including some basic necessities. We have suggestedthatseveralunusualcircumstancesmaycollectively clarify how the Hohokam economy exceeded the dependenceon marketplacebarter typically associatedwith nascent marketingsystems. Those circumstancesincluded (1) widely acceptedreligiousbeliefs thatpromotedcooperation, consensus, social stability,and the peace of the marketplace;(2) a large-scaleirrigationinfrastructure capableof surplusagricultural production; (3) a benignenvironmentalregimethatreinforced the cooperativespiritof the ballcourtceremonialism andfosteredthe geographicallyexpansiveand horizontallyintegratedsocial, religious, and economic network across which commodities and informationfreely flowed; and (4) a diverseecological landscapein which the availabilityof various and important resources were distributedand, when coupled disproportionately withsocialandenvironmentalstability,led to "reliable dependencies"among populations.Clearly, ourtentativemodelhas not explainedthe developmentof a region-sizeeconomythatincludedmarket rings. Less ambitiously, we have simply attemptedto enumeratesome,butnotall,of thenecessaryconditionsfor the unusualcoexistenceof a periodic and horizontallystructuredmarketplace networkwith a high dependenceon markettransactions for the supply of commodities,including some basic necessitieslike pottery. Acknowledgments.A portion of this paper was first presented at the Spring 2003 meeting of the Arizona Archaeological Council, organizedby Sarah Herr.Funding and support for the ceramic analyses were provided by grants from the National Science Foundation (DCB9201161, BCS-0204330), Southwest Parksand Monuments Association, the Arizona Archaeological and Historical Society, and a Robert H. Lister Fellowship from Crow Canyon Archaeological Center. The work at Las Colinas, Pueblo Grande,and Los Hornosresultedfrom contractswith the Arizona Departmentof Transportationto the Arizona 480 AMERICAN ANTIQUITY [Vol. 72, No. 3, 2007 pp. 37-49. ProjectReportNo. 2000:057.Archaeological ResearchServices,Tempe,Arizona. 2002a Chronologyat PaloVerdeRuin.InPhoenixBasinto edited PerryMesa:Rethinkingthe "NorthernPeriphery," by Mark R. Hackbarth,Kelley Hays-Gilpin,and Lynn Neal, pp. 41-46. The ArizonaArchaeologistNo. 34. ArizonaArchaeologicalSociety,Phoenix. 2002b Ceramicsand the Organizationof Regional Ceremonial and ExchangeNetworksat Palo VerdeRuin. In Archaeological Investigations at Palo VerdeRuin, AZ T:8:68(ASM):The TerramarProject, Volume2:Artifact AnalysisandSynthesis,editedby DouglasB. Craig,David R. Abbott, and Mark R. Hackbarth,pp. 13.1-13.16. AnthropologicalPapersNo. 02-02. NorthlandResearch, Tempe,Arizona(draftreport). 2003 Ceramics,Communities,andIrrigationManagement. References Cited In Centuries of Decline during the Hohokam Classic Periodat Pueblo Grande,editedby DavidR. Abbott,pp. David R. Abbott, 148-165. Universityof ArizonaPress,Tucson. 1983 A TechnologicalAssessmentof CeramicVariationin 2007 A ProvenanceAnalysis of Buff WarePotteryFrom the Salt-GilaAqueductArea: Towarda Comprehensive FourVillages in the LowerSalt RiverValley.In Crismon Documentation of Hohokam Ceramics. In Hohokam Ruin: A HohokamSettlementat theHeadof theLehiCanal the Central AriSalt-Gila Aqueduct, ArchaeologyAlong System,editedby T. KathleenHenderson.Anthropologizona Project,Vol. 8, editedby LynnS. Teagueand Patrical PapersNo. 44. Centerfor DesertArchaeology,Tuccia L. Crown,pp. 1-118. ArchaeologicalSeries,No. 150. son, in press. ArizonaStateMuseum,Universityof Arizona,Tucson. 1988 Form,Function,Technology,and Style in Hohokam Abbott,David R. (editor) 2003 Centuriesof Decline during the HohokamClassic Ceramics.In The1982-1984 Excavationat Las Colinas: Period at Pueblo Grande.Universityof ArizonaPress, MaterialCulture,by DavidR. Abbott,Kim E. Beckwith, Tucson. PatriciaL. Crown,R. ThomasEuler,DavidA. Gregory,J. RonaldLondon,MarylinB. Saul,LarryA. Schwalbe,and Abbott,David R., SusanL. Stinson,and ScottVanKeuren 200 1 The EconomicImplicationsof HohokamBuff Ware Mary Bernard-Shaw,pp. 73-198. ArchaeologicalSeries Exchange During the Early Sedentary Period. Kiva No. 162(4).ArizonaStateMuseum,Universityof Arizona, 67:7-29. Tucson. 1993 ThePuebloViejoCeramics.InPuebloViejo:Archae- Abbott,David R., andTinaK. Love 200 1 SpecializedStudiesandObservationson PotteryProological Investigationsat a ClassicPeriodCemeteryin El duction.In TheGreweArchaeologicalResearchProject, Roposo Park,Phoenix,Arizona,edited by M. Zyniecki, Volume2, Part 1: CeramicStudies,edited by David R. pp. 69-107. SWCA ArchaeologicalReportNo. 92-75. Abbott, pp. 141-154. AnthropologicalPapersNo. 99-1. Flagstaff. NorthlandResearch,Tempe,Arizona. 1994 ChemicalAnalysis of Clay Fractionsin Hohokam Pottery.In ThePuebloGrandeProject,Volume3: Ceram- Abbott,David,Douglas R. Mitchell,andJ. A. Merewether 1994 Chronology.In ThePueblo GrandeProject,Volume ics andtheProductionandExchangeof Potteryin the Cen2: FeatureDescriptions,Chronology,and Site Structure, tralPhoenixBasin,editedby DavidR.Abbott,pp.91-148. editedby DouglasR. Mitchell,pp. 157-254. Publications Publications in Archaeology No. 20. Soil Systems, in ArchaeologyNo. 20. Soil Systems,Phoenix. Phoenix. 1995 ChemicalAnalyses of PlainwareClay Fractions.In Abbott,David R., and DavidM. Schaller 1994 Ceramics among the Hohokam:Modeling Social Archaeologyat the Head of the ScottsdaleCanalSystem, Organizationand Exchange. In Archaeometryof PreVolume2: Studies of Artifactsand Biological Remains, ColumbianSites and Artifacts,editedby DavidA. Scott editedby T. KathleenHendersonandDavidR. Abbott,pp. andPieterMeyers,pp. 85-109. GettyConservationInsti15-56. Anthropological Papers No. 95-1. Northland tute,Los Angeles. Research,Tempe,Arizona. 2000 Ceramicsand CommunityOrganizationAmong the Abbott,David R., andMary-EllenWalsh-Anduze 1995 TemporalPatternsWithoutTemporalVariation:The Hohokam.Universityof ArizonaPress,Tucson. Paradoxof HohokamRed WareCeramics.In Ceramic 200 1a ElectronMicroprobeAnalysesof the GARPPottery Productionin theAmericanSouthwest,editedby Barbara and Raw Clays. In The GreweArchaeologicalResearch J. Mills andPatriciaL. Crown,pp. 88-1 14. Universityof Project, Volume2, Part 1: Ceramic Studies, edited by ArizonaPress,Tucson. David R. Abbott,pp. 47-66. AnthropologicalPapersNo. Abbott,David R., JoshuaWatts,andAndrewLack 99- 1. NorthlandResearch,Tempe,Arizona. 2007 The Provenanceand ConcentratedProductionof 200 1b ConclusionsfortheGARPCeramicAnalysis.In The HohokamRed-on-BuffPotteryfrom the ArizonaDesert. GreweArchaeologicalResearchProject, Volume2, Part Journalof AnthropologicalResearch,in press. 1: Ceramic Studies, edited by David R. Abbott, pp. 263-273. AnthropologicalPapers No. 99-1. Northland Alden, JohnR. 1982 MarketplaceExchangeas IndirectDistribution:An Research,Tempe,Arizona. IranianExample.In Contextsfor PrehistoricExchange, 200 1c TestingPhaseCeramicsFromEl Caliche.InResults editedby JonathanE. EricsonandTimothyK. Earle,pp. of Archaeological Testingof the Proposed Earll Drive 83-102. AcademicPress,New York. Right-of-wayAdjacentto El Caliche(AZU:9:110[ASM]), Dean E. Arnold, Salt RiverPima-MaricopaIndianCommunity, Maricopa 1985 CeramicTheoryand CulturalProcess. Cambridge County,Arizona,by DavidE. PurcellandDavidR.Abbott, State Museum;Soil Systems, Inc.; and SWCA, Inc., respectively. Thanks to Cory Breternitz, Mark Chenault, David Gregory,ChristineRobinson, and Lynn Teaguefor the sharing of informationrelatedto those projects.ArthurVokes at the Arizona State Museum is gratefully acknowledged for assistance with curated collections. Several colleagues, includingCharlesCobb, David Greenwald,RobertHunt,A. Trinkle Jones, Henry Wallace, and three anonymous peer reviewers graciously contributedtheir time and thought to reading and commenting on various versions of this manuscript. ChristopherWatkinsprovidedthe figures.Any errors in fact or judgmentremainours alone. Abbott et al.] BALLCOURTS AND CERAMICS UniversityPress,Cambridge. Bayman,JamesM. 2002 HohokamCraftEconomies and the Materialization of Power.Journalof ArchaeologicalMethodand Theory 9:69-95. Belshaw,CyrilS. 1965 TraditionalExchangeandModernMarkets.PrenticeHall, EnglewoodCliffs, New Jersey. Bernard-Shaw, Mary 1983 The Stone Tool Assemblageof the Salt-GilaAqueduct ProjectSites. In HohokamArchaeologyalong the Salt-Gila Aqueduct,CentralArizona Project, Vol. VIII (2-5): Material Culture,edited by Lynn S. Teague and PatriciaL. Crown, pp. 373-444. ArchaeologicalSeries No. 150. ArizonaState Museum,Universityof Arizona, Tucson. Blanton,RichardE. 1983 FactorsUnderlyingthe OriginandEvolutionof Markets Systems.EconomicAnthropology:Topicsand Theories, edited by Sutti Ortiz, pp. 51-66. Monographsin EconomicAnthropologyNo. 1. UniversityPressof America, Lanham,Maryland. Bohannan,Paul,andLauraBohannan 1968 Tiv Economy. Northwestern University Press, Evanston. Bohannan,Paul,andGeorgeDalton 1962 Introduction.In Marketsin Africa, edited by Paul Bohannanand George Dalton, pp. 1-28. Northwestern UniversityPress,Evanston,Illinois. Bromley,R. J., RichardSymansky,andCharlesM. Good 1975 The Rationaleof Periodic Markets.Annals of the Associationof AmericanGeographers65:530-537. Brumfiel,ElizabethM. 1980 Specialization,MarketExchangeandtheAztec State: 2 1:459^78. A ViewfromHuexotla.Current Anthropology Carrasco,Pedro 1983 Some TheoreticalConsiderationsAboutthe Role of the Marketin Ancient Mexico. In EconomicAnthropology,Topicsand Theories,editedby SuttiOrtiz,pp. 67-82. Monographsin EconomicAnthropologyNo 1. University Pressof America,Lanham,Maryland. Chenault, Mark L., RichardV. N. Ahlstrom, and Todd N. Motsinger(editors) 1993 In the Shadowof SouthMountain:The Pre-Classic Hohokamof La Ciudadde los Hornos,1991-1992 Excavations,2 vols.ArchaeologicalReportNo. 93-30. SWCA, Tucson. Cobb,CharlesR. 1993 ArchaeologicalApproachesto thePoliticalEconomy of NonstratifiedSocieties.InArchaeologicalMethodand Theory,Vol.5, editedby MichaelB. Schiffer,pp. 43-100. Universityof ArizonaPress,Tucson. Costin,CathyL. 1991 CraftSpecialization:Issues in Defining,Documenting, and Explainingthe Organizationof Production.In Archaeological Method and Theory,Vol. 3, edited by MichaelB. Schiffer,pp. 1-56. Universityof ArizonaPress, Tucson. Craig,DouglasB. 2001 The Demographic Implications of Architectural Changeat Grewe.In TheGreweArchaeologicalResearch Project,Volume3: Synthesis,editedby DouglasB. Craig, pp. 37-50. AnthropologicalPapersNo. 99-1. Northland Research,Tempe,Arizona. Crown,PatriciaL. 1991 The Role of Exchangeand Interactionin Salt-Gila Basin HohokamPrehistory.In Exploringthe Hohokam: 481 PrehistoricDesert Peoples of the AmericanSouthwest, edited by George J. Gumerman,pp. 383^16. Amerind FoundationNew WorldStudies Series No. 1. University of New Mexico Press,Albuquerque. Crown,PatriciaL., and SuzanneK. Fish 1996 Gender and Status in the HohokamPre-Classicto ClassicTransition.AmericanAnthropologist98:803-8 17. Crown,PatriciaL., andWilliamJ. Judge(editors) 199 1 Chaco and Hohokam:PrehistoricRegionalSystems in theAmericanSouthwest.School of AmericanResearch Press,SantaFe, New Mexico. Doelle, WilliamH. 1980 PastAdaptivePatternsin theWesternPapagueria: An Archaeological Study of Nonriverine Resource Use. UnpublishedPh.D. Dissertation,Universityof Arizona, Tucson. 1995 TontoBasin Demographyin a RegionalPerspective. In The Roosevelt CommunityDevelopmentStudy,New Perspectiveson TontoBasin Prehistory,edited by Mark D. Elson, MiriamT. Stark, and David A. Gregory,pp. 201-226. Anthropological PapersNo. 15. CenterforDesert Archaeology,Tucson. Doelle, WilliamH., and HenryD. Wallace 1991 The Changing Role of the Tucson Basin in the HohokamRegionalSystem. In Exploringthe Hohokam: PrehistoricDesert Peoples of the AmericanSouthwest, edited by George J. Gumerman,pp. 279-346. Amerind FoundationNew WorldStudies Series No. 1. University of New Mexico Press,Albuquerque. Doyel, David E. 1979 The PrehistoricHohokam of the Arizona Desert. AmericanScientist67:544-554. 1981 LateHohokamPrehistoryin SouthernArizona.ContributionstoArchaeologyNo. 2. GilaPress,Scottsdale,Arizona. 1985 Exchange and Interaction.In HohokamSettlement andEconomicSystemsin theCentralNewRiverDrainage, Arizona,editedby DavidE. Doyel andMarkD. Elson,pp. 715-726. PublicationsinArchaeologyNo. 4. Soil Systems, Phoenix. 1991a HohokamExchangeand Interaction.In Chacoand Hohokam:PrehistoricRegionalSystemsin theAmerican Southwest,edited by PatriciaL. Crown and William J. Judge,pp. 225-252. School of AmericanResearch,Santa Fe. 199 1b HohokamCulturalEvolutionin the PhoenixBasin. In ExploringtheHohokam:PrehistoricDesertPeoplesof theAmericanSouthwest,editedby GeorgeJ. Gumerman, pp. 231-278. AmerindFoundationNew World Studies Series No. 1. University of New Mexico Press, Albuquerque. 1993 On RiversandBoundariesin the PhoenixBasin,Arizona. Kiva 58:455^74. 2000 The Santan Phase in the Phoenix Basin. In The HohokamVillageRevisited,edited by David E. Doyel, SuzanneK. Fish, and Paul R. Fish, pp. 221-244. Southwesternand Rocky MountainDivision of the American AssociationfortheAdvancementof Science,FortCollins, Colorado. Doyel, David E., andMarkD. Elson 1985 CeramicAnalysis.In HohokamSettlementand Economic Systemsin the CentralNew RiverDrainage,Arizona,Vol.2, editedby DavidE. Doyel andMarkD. Elson, pp.436-520. PublicationsinArchaeologyNo. 4. Soil Systems, Phoenix. Doyel, David E., andMarkD. Elson (editors) 1985 HohokamSettlementand EconomicSystemsin the 482 AMERICAN ANTIQUITY CentralNew River Drainage, Arizona. Publicationsin ArchaeologyNo. 4. Soil Systems,Phoenix. Earle,Timothy 1987 Specializationandthe Productionof Wealth:Hawaiian Chiefdom and the Inka Empire. In Specialization, Exchange, and ComplexSocieties, edited by Elizabeth Brumfiel.CambridgeUniversityPress,Cambridge. 1997 How Chiefs Come to Power,ThePolitical Economy in Prehistory.StanfordUniversity,Stanford. Feinman,GaryM. 1999 RethinkingOurAssumptions:EconomicSpecialization at the Household Scale in Ancient Ejutla,Oaxaca, Mexico. In Potteryand People:A DynamicInteraction, edited by James M. Skibo and Gary M. Feinman, pp. 81-98. Universityof UtahPress,Salt LakeCity. Feinman,GaryM., andLindaM. Nicholas 2000 High-Intensity Household-Scale Production in AncientMesoamerica:A PerspectivefromEjutla,Oaxaca. In CulturalEvolution:ContemporaryViewpoints,edited by GaryM. FeinmanandLindaManzanilla,pp. 119-142. KluwerAcademic/PlenumPublishers,New York. Feinman,GaryM, RichardBlanton,andStephenKowalewski 1984 MarketSystemsDevelopmentin thePrehispanicValley of Oaxaca,Mexico. In Tradeand Exchangein Early Mesoamerica,editedby KennethG. Hirth,pp. 157-178. Universityof New Mexico Press,Albuquerque. Fish, SuzanneK., PaulR. Fish, andJohnH. Madsen 1992 Evidence for Large-ScaleAgave Cultivationin the MaranaCommunity.In The Marana Communityin the HohokamWorld,editedby SuzanneK. Fish,PaulR. Fish, and John H. Madsen,pp. 73-87. Universityof Arizona Press,Tucson. Forman,Shepard,andJoy F. Riegelhaupt 1970 MarketPlace andMarketingSystem:Towarda Theory of PeasantEconomicIntegration.ComparativeStudies in Societyand History 12:188-212. Frankenstein,Susan,andMichaelJ. Rowlands 1978 The InternalStructureandRegionalContextof Early IronAge Society in South-WesternGermany.Bulletinof the Instituteof Archaeology15:73-112. Fry,RobertE. 1980 Modelsof ExchangeforMajorShapeClassesof Lowland Maya Pottery.In Models and Methodsin Regional Exchange,editedby RobertE. Fry,pp. 3-18. SAA Papers No. 1. Society for AmericanArchaeology,Washington, D.C. Gasser,RobertE., and Scott M. Kwiatkowski 199 1 FoodforThought:RecognizingPatternsin Hohokam Subsistence. In Exploring the Hohokam: Prehistoric Desert Peoples of the American Southwest, edited by GeorgeJ. Gumerman,pp. 417-^-60.AmerindFoundation New WorldStudiesSeries No. 1. Graybill,DonaldA. 1989 The Reconstructionof PrehistoricSaltRiverStreamflow. In The1982-1984 Excavationsat Las Colinas,Volume 5: Environmentand Subsistence, by Donald A. Graybill,David A. Gregory,Fred L. Nials, Suzanne K. Fish, RobertE. Gasser,CharlesH. Miksicek,and Christine R. Szuter,pp. 25-38. ArchaeologicalSeriesNo. 162. ArizonaStateMuseum,Universityof Arizona,Tucson. Gregory,DavidA. 1988a The Ballcourt.In The 1982-1984 Excavationsat Las Colinas:TheSite and Its Features,by DavidA. Gregory,WilliamL Deaver,SuzanneK. Fish,RobertGardiner, RobertW. Layhe, Fred L. Nials, and Lynn Teague,pp. 269-273. ArchaeologicalSeries No. 162. Arizona State Museum,Universityof Arizona,Tucson. [Vol. 72, No. 3, 2007 1988b SherdCounts.In The1982-1 984 Excavationsat Las Colinas:SpecializedStudiesand Data Tables,by David A. Gregory,B. A. Murphy,WilliamL Deaver,RichardC. Lange, Mary Sullenberger,Christine R. Szuter, Mary Bernard-Shaw, Arthur W. Vokes, R. Thomas Euler, SuzanneK. Fish, CharlesH. Miksicek, and MarilynB. Saul,pp. 93-134. ArchaeologicalSeriesNo. 162.Arizona StateMuseum,Tucson. Haggett,Peter,A. D. Cliff, andAllan Frey 1977 LocationalModels. Halsted,New York. Hassig, Ross 1982 PeriodicMarketsin PrecolumbianMexico. American Antiquity47:346-355. Haury,EmilW. 1937a Ball Courts.In Excavationsat Snaketown:Material Culture,by HaroldS. Gladwin,EmilW. Haury,HaroldB. Sayles, and Nora Gladwin,pp. 36-49. MedallionPapers No. 25. Gila Pueblo,Globe,Arizona. 1937b A Pre-SpanishRubberBall fromArizona.American Antiquity2:282-288. 1937c PotteryTypesatSnaketown.InExcavationsat Snaketown:MaterialCulture,by HaroldS. Gladwin,Emil W. Haury,HaroldB. Sayles,andNoraGladwin,pp. 169-229. MedallionPapersNo. 25. Gila Pueblo,Globe,Arizona. 1976 TheHohokam,DesertFarmersandCraftsmen:Excavationsat Snaketown,1964-1965. Universityof Arizona Press,Tucson. Heidke,JamesM. 1996 ProductionandDistributionof RinconPhasePottery: Evidencefrom the JulianWashSite. In A RinconPhase Occupationat JulianWash,AZBB:13:17(ASM),editedby JonathanB. Mabry,pp. 47-71. TechnicalSeries No. 967. Centerfor DesertArchaeology,Tucson. 2000 MiddleRinconPhaseCeramicArtifactsfromSunset Mesa. In Excavations at Sunset Mesa Ruin, edited by MichaelW. Lindeman,pp. 69-1 18.TechnicalReportNo. 2000-02. DesertArchaeology,Tucson. Hirth,KennethG. 1998 The Distributional Approach,A New Wayto Identify Marketplace Exchangein theArchaeologicalRecord.CurrentAnthropology39:451-476 Hodder,BramwellW. 1962 The YorubaRound Market.In Markets in Africa, editedby PaulBohannanandGeorgeDalton,pp. 103- 117. NorthwesternUniversityPress,Evanston,Illinois. Hoffman,C. M. 1997 AllianceFormationandSocial Interactionduringthe SedentaryPeriod:A StylisticAnalysisof HohokamArrowpoints. UnpublishedPh.D. Dissertation,Departmentof Anthropology,ArizonaStateUniversity,Tempe. Hoffman,TeresaL., DavidE. Doyel, andMarkD. Elson 1985 GroundStoneToolProductionin theNew RiverBasin. In Proceedingsof the 1983 HohokamSymposium,PartII, editedby A. E. Dittert,Jr.and D. E. Dove, pp. 655-686. OccasionalPaperNo. 2. ArizonaArchaeologicalSociety, Phoenix. Howard,Ann V. 1993 MarineShell Artifactsand ProductionProcessesat Shelltownand the Hind Site. In Shelltownand the Hind Site:A Studyof TwoHohokamCraftsmanCommunitiesin SouthwesternArizona,editedby WilliamS. Marmaduke and Richard J. Martynec, pp. 321-423. Northland Research,Flagstaff. Howard,JerryB. 1991 System Reconstruction:The Evolutionof an Irrigation System. In The Operationand Evolutionof an Irrigation System:TheEast PapagoCanalStudy,by JerryB. Abbott et al.] BALLCOURTS AND CERAMICS HowardandGaryHuckleberry, pp.5. 1-5.33. Publications in ArchaeologyNo. 18. Soil Systems,Phoenix. 1993 A Paleohydraulic Approachto ExaminingAgricultural Intensification in Hohokam Irrigation Systems. In ResearchinEconomicAnthropology, Supplement7, edited andBarryL. Isaac,pp.263-324. byVernonL. Scarborough JAIPress,Greenwich. Hunt,RobertC, DavidGuillet,DavidR.Abbott,JamesM. Bayman,PaulR. Fish, SuzanneK. Fish, KeithW. Kintighand JamesA. Neely 2005 PlausibleEthnographic AnalogiesfortheSocialOrganizationof HohokamCanalIrrigation.AmericanAntiquity70:433^56. Hutira,Johna 1989 FlotationAnalysis. In Settlement,Subsistence,and Specializationin the NorthernPeriphery:The Waddell Project,editedby MargerieGreen,pp. 85 1-872. Cultural ResourcesReportNo. 65. ArchaeologicalConsultingServices, Tempe. Johnson,Allen W., andTimothyEarle 2000 TheEvolutionof HumanSocieties,2ndedition.StanfordUniversityPress,Stanford. Judge,WilliamJ. 1989 Chaco Canyon - San Juan Basin. In Dynamics of SouthwesternPrehistory,editedby LindaS. Cordelland GeorgeJ. Gumerman,pp. 209-261. SmithsonianInstitution Press,WashingtonD. C. Lane,Anne Marie 1989 The GrandCanal CeramicAssemblage: Ceramic TypeDescriptions.InArchaeologicalInvestigationsat the GrandCanalRuins:A ClassicPeriodSitein Phoenix,Arizona,Vol. 1, editedby DouglasR. Mitchell,pp. 249-293. Publications in Archaeology No. 12. Soil Systems, Phoenix. Lascaux,Annick,andJohnC. Ravesloot 1993 The MaricopaRoad Site Viewed in its Local and Regional Context. In The Maricopa Road Site: A PreClassic HohokamVillage,edited by John C. Ravesloot and Annick Lascaux, pp. 39^8. AnthropologicalField StudiesNo. 28. ArizonaStateUniversity,Tempe. LeBlanc,StephenA. 1999 PrehistoricWarfarein theAmericanSouthwest.Universityof UtahPress,Salt LakeCity. Marshall,JohnT. 2001 Ballcourt.In The GreweArchaeological Research Project, Volume 1: Project Background and Feature Descriptions,edited by Douglas B. Craig,pp. 109-124. AnthropologicalPapersNo. 99-1. NorthlandResearch, Tempe,Arizona. Miksa,ElizabethJ. 1995 PetrographicAnalysis of Ceramicsfor the McDowell to Shea Blvd. Portionof the Beeline HighwayExcavationProject.InArchaeologyat theHeadof theScottsdale CanalSystem,Volume2: Studiesof Artifactsand Biological Remains,editedby T. KathleenHendersonandDavid R. Abbott,pp. 15-38. AnthropologicalPapersNo. 95-1. NorthlandResearch,Tempe,Arizona. 2001 TemperProvenanceStudies.In TheGreweArchaeologicalResearchProject,Volume2, Part1: CeramicStudies, editedby DavidR. Abbott,pp. 7-46. Anthropological PapersNo. 99-1. NorthlandResearch,Tempe,Arizona. Miksa,ElizabethJ.,SergioF.Castro-Reino,andCarlosLavayen 2004 A CombinedPetrofaciesModel for the MiddleGila andPhoenixBasins,WithApplicationto PotteryFromthe DutchCanalRuin.InHohokamFarmingon theSaltRiver Floodplain: RefiningModels and Analytical Methods, editedby T. KathleenHenderson,pp. 7^4. Anthropolog- 483 ical PapersNo. 43, Centerfor DesertArchaeology,Tucson. Mills, BarbaraJ. 2000 Gender,CraftProduction,and Inequality.In Women & Men in the PrehispanicSouthwest:Labor,Power,and Prestige,editedby PatriciaL. Crown,pp.301-343. School of AmericanResearchPress,SantaFe. Netting,RobertMcC. 1990 Population, PermanentAgriculture, and Polities: Unpackingthe EvolutionaryPortmanteau.In TheEvolution of PoliticalSystems,editedby SteadmanUpham,pp. 21-61. CambridgeUniversityPress,Cambridge. Nials, FredL., and SuzanneK. Fish 1988 CanalsandRelatedFeatures.In The1982-1984 Excavationsat Las Colinas:TheSiteandItsFeatures,by David A. Gregory,WilliamL. Deaver,SuzanneK. Fish, R. Gardiner,RobertW. Layhe,FredL. Nials, andLynnTeague, pp.275-305. ArchaeologicalSeriesNo. 162.ArizonaState Museum,Universityof Arizona,Tucson. Plattner,Stuart 1975 Rural Market Networks. Scientific American 232:66-79. 1985 Marketsand Marketing.Monographsin Economic Anthropology4. UniversityPress of America,Lanham, Maryland. 1989 MarketsandMarketplaces.In EconomicAnthropology, editedby StuartPlattner,pp. 171-208. StanfordUniversityPress,Stanford. Rafferty,Kevin 1982 Hohokam Micaceous Schist Mining and Ceramic CraftSpecialization:An Example from Gila Butte,Arizona.Anthropology6:199-222. Rappaport,Roy A. 1971 The Sacredin HumanEvolution.AnnualReviewof Ecology and Systematics2:23^4. 1984 Pigsfor theAncestors:Ritualin theEcologyof a New GuineaPeople.2nded.YaleUniversityPress,New Haven. Renfrew,Colin 2001 ProductionandConsumptionin a SacredEconomy: TheMaterialCorrelatesof HighDevotionalExpressionat ChacoCanyon.AmericanAntiquity66:14-25. Rice, PrudenceM. 1987 PotteryAnalysis:A Sourcebook.Universityof Chicago Press,Chicago. Schaller,David M. 1994 GeographicSourcesof TemperIn CentralPhoenix Basin CeramicsBased on PetrographicAnalysis. In The Pueblo GrandeProject,Volume3: Ceramicsand theProductionand Exchangeof Potteryin the CentralPhoenix Basin,editedby DavidR. Abbott,pp. 17-90. Publications in ArchaeologyNo. 20. Soil Systems,Phoenix. Schele, Linda,andDavid Freidel 1990 A Forestof Kings: The UntoldStory of the Ancient Maya.W Morrow,New York. Smith,CarolA. 1974 Economicsof MarketingSystems:ModelsfromEconomic Geography. Annual Review of Anthropology 3:167-201. 1976 RegionalEconomicsSystems:LinkingGeographical ModelsandSocioeconomicProblems.InRegionalAnalysis, Volume1: EconomicSystems,editedby CarolA. Smith, pp. 3-68. AcademicPress,New York. Stark,MiriamT. 199 1 CeramicProductionandCommunitySpecialization: A CeramicEthnoarchaeologicalStudy.WorldArchaeology 23:64-78. Stine,J. H. 484 [Vol. 72, No. 3, 2007] AMERICAN ANTIQUITY 1962 TemporalAspectsof TertiaryProductionElementsin KoreaUrbanSystemsandEconomicDevelopment,edited by ForrestR. Pitts,pp.68-88. Universityof Oregon,School of BusinessAdministration,Eugene. Teague,LynnS. 1988 StructureandPopulation:GrowthandChangein the StudyArea. In The 1982-1984 Excavationsat Las Colinas, Synthesis and Conclusions, by Lynn Teague and WilliamL. Deaver,pp. 43-76. ArchaeologicalSeries 162. ArizonaStateMuseum,Universityof Arizona,Tucson. 1998 Textilesin SouthwesternPrehistory.University of New Mexico Press. Turney,Omar 1929 PrehistoricIrrigation. ArizonaHistoricalReview2(5). Phoenix. Wallace,HenryD. 2001 Time Sedation and TypologicalRefinementof the Middle Gila Buffware Sequence: SnaketownThrough Soho Phases.In TheGreweArchaeologicalResearchProject, Volume2, Part 1. CeramicStudies,editedby David R. Abbott,pp. 177-262. AnthropologicalPapersNo. 991. NorthlandResearch,Tempe. 2004 Update to the Middle Gila Buff Ware Ceramic Sequence.In HohokamFarmingon the Salt RiverFloodplain: RefiningModelsandAnalyticalMethods,editedby T. Kathleen Henderson, pp. 45-124. Anthropological PapersNo. 43. Centerfor DesertArchaeology,Tucson. Walsh-Anduze,Mary-Ellen 1996 Changesin the Structureof HohokamCeramicProduction:TheResultsof the PuebloSaladoCeramicStudy. InLifeon theFloodplain:FurtherInvestigationsat Pueblo Saladofor PhoenixSkyHarborInternationalAirport,Volume2: Data RecoveryandRe-evaluation,editedby David H. Greenwald,JeanH. Ballagh,DouglasR. Mitchell,and RichardA. Anduze,pp. 97-140. AnthropologicalPapers No. 4. PuebloGrandeMuseum,Phoenix. Weaver,DonaldE., Jr. 1973 The Site Characterization Program.In Definitionand PreliminaryStudyof the Midvale Site, edited by James Schoenwetter,SylviaW.Gaines,andDavidE. Weaver,Jr., pp. 92-154. AnthropologicalResearchPapersNo. 6. Arizona StateUniversity,Tempe. Wilcox, David R. 1979 The HohokamRegionalSystem.InAnArchaeological Testof Sites in the Gila Butte-SantanRegion,SouthCentralArizona,by GlennE. Rice,DavidR.Wilcox,Kevin Rafferty,and James Schoenwetter,pp. 77-116. AnthropologicalResearchPapersNo. 18. ArizonaStateUniversity,Tempe. 1987 FrankMidvale's Investigationof the Site of La Ciudad.AnthropologicalField StudiesNo. 19.ArizonaState University,Tempe. 1991a Hohokam Social Complexity. In Chaco and Hohokam:PrehistoricRegionalSystemsin theAmerican Southwest,edited by PatriciaL. Crown and William J. Judge,pp. 253-276. School of AmericanResearch,Santa Fe. 199 1b TheMesoamericanBallgamein theAmericanSouthwest. In TheMesoamericanBallgame,editedby Vernon L. Scarboroughand David R. Wilcox, pp. 101-125. Universityof ArizonaPress,Tucson. 1999 A PeregrineView of MacroregionalSystems in the North American Southwest, A.D. 750-1250. In Great Townsand RegionalPolities in the PrehistoricAmerican Southwestand Southeast,edited by Jill E. Neitzel, pp. 115-142. Universityof New Mexico Press,Albuquerque. Wilcox, David R., JerryB. Howard,and RuebenH. Nelson 1990 One HundredYearsof Archaeologyat La Ciudadde los Homos. PublicationsinArchaeologyNo. 16.Soil Systems, Phoenix. Wilcox, David R., TerrySamples, Donald Keller,and Maria Laughner 1996 TheWagnerBallcourtCommunityandOtherCohonina Sites. Kiva61:433^56. Wilcox, David R., andCharlesSternberg Archae1983 HohokamBallcourtsandtheirInterpretation. ological Series No. 160. ArizonaStateMuseum,University of Arizona,Tucson. Yoffee,Norman The 1994 Memorandum to MurrayGell-MannConcerning: Complicationsof Complexityin the PrehistoricSouthwest. In UnderstandingComplexityin the Prehistoric Southwest,editedby GeorgeJ. Gumermanand M. GellMann, pp. 341-358. Addison-Wesley,Reading,Massachusetts. Notes 1. Most scholars who study the Hohokam would agree that the Sedentaryperiod specialistswere part-timeartisans, and,following Feinman(1999; FeinmanandNicholas 2000), would concur that the craft activities were generallylocated in householdcontexts. 2. Obviously,this estimate will depend, in part,on how many people lived in the lower Salt River valley at any one time duringthe Sedentaryperiod, which is a topic of much debate.Nevertheless,a roughidea of the productionand distributionscales can be achieved with a few assumptions.If we accept the conservativeestimate that only 4,000 people were present in the lower Salt River valley during the Sedentaryperiod (Doelle 1995), and if we assume that each residentrequiredone new pot per year, then 4,000 new pots were producedannuallyfor the Salt Riverhouseholds.If half of them were made at middle Gila locations,then 2,000 vessels were transportedfrom the Gila to the Salt Rivercommunities each year.In contrast,if we acceptCraig's(2001) more liberal population estimates, then, perhaps, 8,000 pots or more were importedeach year to the lower Salt River valley from the middle Gila producers.Assuming higherconsumption rates would, of course, yield still higher estimates.For instance, Judge (1989:231) estimates a rate of potteryconsumption in domestic contexts in Chaco Canyon that was threetimes higher. 3. The plain-ware ceramicsfrom Feature7089 could not be located for analysis. 4. It is also interestingthat during the Sedentaryperiod most of the domestic laborfor crafts was devotedin particular householdsto a predominantproductratherthanto multicrafting (Mills 2000:314). These productiondecisions may have relied on the regional-scale integrationand reduced transportationcosts of the markets.In contrast,during the Classic period,when the economy was characterizedby more localized exchangenetworks,the craftactivitiesof individual households were often more diverse (Crownand Fish 1996; Mills 2000). Received February27, 2006; RevisedFebruary9, 2007; AcceptedJanuary18, 2007.