Drained Fields at la Tigra, Venezuelan Llanos: A Regional Perspective

Society for American Archaeology

Drained Fields at la Tigra, Venezuelan Llanos: A Regional Perspective Author(s): Charles S. Spencer, Elsa M. Redmond, Milagro Rinaldi Source: Latin American Antiquity, Vol. 5, No. 2 (Jun., 1994), pp. 119-143 Published by: Society for American Archaeology Stable URL: http://www.jstor.org/stable/971559 Accessed: 09/08/2010 20:12 Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=sam. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Society for American Archaeology is collaborating with JSTOR to digitize, preserve and extend access to Latin American Antiquity.

http://www.jstor.org

DRAINED F IFl ns AT LA TIGRA, VENEZUELAN LLANOS: A REGIONAL PERSPECIIVE CharlesS. Spencer,Elsa M. Redmond, and MilagroRinaldi Thispaperdiscussesdrained-fieldstudiesin Venezuela beginningwiththefirst investigationstwo decadesago thatfocused on field systems themselvesand proceedingto recent researchby the authors that examined the drainedfields of La Tigraas part of a regional-scaleprojectin the state of Barinas. The La Tigrafields are dated to the Late Gavanphase (A.D. 550-1000) a time of extensivehabitationin the region.An analysis of excavated pollen samplesfrom the drainedAlds and a nearbyvillagesite has revealedthat whereasmaize was thepredominant plant there was notableintersitevariabilityin the secondarycultigens.The paper also considerswhether populationpressurecouldhavepromptedthe constructionof the La Tigra>Ids. A comparisonof archaeological populationestimatesto estimatesof potentialpopulationundervaryingassumptionsof productivecapacityyields no indicationof demographicpressures.We suggest that drained-ld constructionin this case was motivated primarilyby political-economicconsiderationspart of a strategywherebythe regionalelite sought to stimulate and mobilizethe productionof surplusby villagefarmers. Discutimosen este articulolos estudiossobre campos drenadosen Venezuela desde sus inicios hace veinte anos hasta llegara la investigacionrecientepor los autoresde esta obrasobreel estudiode los camposdrenados de La Tigra queforman parte de un proyectoa escala regionalen el estado Barinas.Los camposcorresponden a la fase GavanTardla(55S1 000 D.C.) una epocade ocupacionextensivaen la region.El analisispalinologico de las muestrasexcavadasen los camposdrenadosy en un yacimientocercanosugiereque el maizfue el cultivo dominante mientrasque los cultlgenossecundariosvarlanentrelos dos sitios. Tambiendiscutimossi los campos drenadosde La Tigrafueron construidosen respuestaa presiones demograMicas. Una comparacionentre la estimacionde la poblacionarqueologicay estimacionesde la poblacionpotencial bajosuposicionesvariables no prod2xce indicacionesde presiondemografica.Sugerimosquelos camposdrenadosde La Tigrafueronconstruidos por notivospolltico-economicosy queformaronparte de una estrategiade la elite regionalpara estimulary movilizarla produccionexcedente. It is by now well established that drained-field cultivation was practiced in many parts of the Precolumbian world (Darch, ed. 1983; Denevan 1982; Denevan et al. 1987; Farrington 1985). Truly massive canalization and field-construction projects were carried out along lakeshores in the highaltitude basins of Mexico and Lake Titicaca (Erickson 1987; Kolata 1991; Parsons 1976). No less impressive were the drained-field systems built in various lowland riverine settings, including the Maya Lowlands (Flannery 1982; Turner 1974; Turner and Harrison 1982), the Rio San Jorge Basin in Colombia (Parsons and Bowen 1966; Plazas and Falchetti de Saenz 1981, 1987), the Guayas Basin of Ecuador (Denevan and Mathewson 1983; Marcos 1987; Parsons 1969), and the Llanos de Mojos of Bolivia (Denevan 1966; Enckson 1980). Archaeological studies of drained fields have vaned somewhat in their goals and methodology, but we think it fair to say that the history of research in most regions has followed a broadly similar trajectory. Dunng the initial phases of research the remains of drained fields are located, often through a combination of aenal photography and ground reconnaissance. Mapping and excavation programs are then directed at the drained fields themselves, with the goal of ascertaining how the systems were constructed and utilized. This phase often draws on ethnographic analogy or expenmental work to support the archaeological reconstructions. Later phases of research, when they CHARLES S. SPENCER and ELSA M. REDMOND *Anthropology Department,Amencan Museum of NaturalHistory, CentralParkW. at 79th Street,New York, NY 10024-5192 MILAGRORINALDI * Laboratoriode Paleoecologia,Centrode Ecologiay Ciencias Ambientales,Instituto Venezolanode InvestigacionesCientificas,Apartado21827, Caracas1020-A, Venezuela Latin AmericanAntiquity, 5(2), 1994, pp. 119-143. Copyrightt 1994 by the Society for AmericanArchaeology

119

LATIN AMERICANANTIQUITY

120

Archaeological Sites O Modern Cities

[Vol. 5, No. 2, 1994

North

_

Okm

_

40km

)tof

rnap 7

W

>

i

Figure 1. Map of westernVenezuela,showingthe study regionof the 1983-1988 BarinasProjectand other relevantarchaeologicalsites. occur, usually involve analysis of the fields in the context of regional settlement patterns, with an eye toward elucidating the role of drained-field systems in the regional subsistence and/or political economy. In this report, we begin with a discussion of previous drained-field studies in Venezuela (Denevan and Bergman 1975; Denevan and Zucchi 1978; Garson 1980; Zucchi and Denevan 1979), and then present the results of our own recent research on the La Tigra drained fields in the Venezuelan state of Barinas. PREVIOUS DRAINED-FIELD RESEARCH IN VENEZUELA Extensive Prehispanic earthworks have been known for some time in the western llanos (humid savanna grasslands) of the Apure-Orinoco Basin (Cruxent and Rouse 1958; Rouse and Cruxent 1963;Zucchi 1972), and a quarter century ago Cruxent (1966) hypothesized that some ofthe features might have had agricultural functions. In 1968 James H. Terry spotted what he suspected was evidence of drained fields on an aerial photograph of the Cano Ventosidad, a cano (stream) some 20km north of the Rio Apure in the state of Barinas (Figure 1). Terry showed the photograph to JamesParsons, who sent it to William Denevan; Denevan agreed that the features were probably artificialconstructions. In 1972 Denevan and Alberta Zucchi carried out a mapping and excavation project at Cano Ventosidad (Denevan and Zucchi 1978; Zucchi and Denevan 1979). The Cano Ventosidad project recorded250 pairs of parallel ridges running perpendicular to the natural levees of the cano (Figure 2),distributed over an area of about 15.5 km2 (Denevan and Zucchi 1978:237-240). In a measured sampleof 16 ridges and 9 ditches (Denevan and Zucchi 1978:240), the width of individual ridges rangedfrom 6.7 to 25.3 m (mean = 15.5 m), the width of the ditch between ridges varied between

Spencer et al.]

DRAINED FIELDS AT LA TIGRA

121

Figure2. Map of the drained-fieldsystem at Cano Ventosidad,Hato Candelaria(afterDenevanand Zucchi

3.7 and 6.7 m (mean = 4.4 m), and the distance between pairs of ridgesrangedfrom 26.2 to 59.4 m (mean = 48.8 m). The ridgesvaried greatlyin length, with some reaching1,000 to 1,500 m. An excavatedcross section revealedan originalridgeheight of 1.59 m, with an interveningcanaldepth of 1.2 m (Denevan and Zucchi 1978:241). The ridged fields were viewed by Denevan and Zucchi as artificialextensionsofthe naturalcano levees. Sucha facility,they suggested,could have extended the naturalgrowingseason in this region of seasonal inundations,and permittedtwo harvestsper year (Denevan and Zucchi 1978:242-243; Zucchi and Denevan 1979:36-37). Whenwerethe CanoVentosidadfieldsbuiltand used?As is oftenthe case in drained-fieldresearch, the fact that very few artifacts(just three small, nondescriptsherds)were found in the excavation of the fields precludeddirect chronologicalassignment.The investigatorsalso did not locate any nearbyhabitationsites that might have been used to date the fields throughassociation, although they acknowledgedthat such sites probably exist (Denevan and Zucchi 1978:243). The nearest habitation sites they recordedwere some 15 to 20 km from the drainedfields. At one of these, E1 Choque(Figure1),fourtest pits producedpotterysimilarto the small samplefrom CanoVentosidad (Denevan and Zucchi 1978:243). El Choque is describedas occupyingthe easternedge of an oval mata (clump of forest) measuring200 x 600 m; a reasonableapproximationof site size would be less than 5 ha (Zucchiand Denevan 1979:39). Ceramic analysis (Zucchi and Denevan 1979:4149) groupedthe E1Choque material with the Arauquinoidseries ofthe latterpartof PeriodIII (A.D. 300-1000) and PeriodIV (A.D. 100>1500) in the Rouse and Cruxent(1963:22, 90-95) chronology.The investigatorsconcludedthat the ridged fields were constructedby people from E1Choquebetween about A.D. 1200 and 1400 (Zucchiand Denevan 1979:82), or perhapsas early as A.D. 1000 (Denevan and Zucchi 1978:244). A trek of 15-20 km from village to field seems exceedinglylong for preindustrialfarmers,far greaterthan the 2-5 km that most catchment-areaanalystsacceptas a reasonabledistance(Alden 1979;Brumfiel

122

LATIN AMERICANANTIQUITY

[Vol. 5, No. 2, 1994

1976; Flannery1976). There is, however, no known habitationsite closer to the drainedfields than E1 Choque, a situation that only a regional settlement-patternsurvey (e.g., Blanton et al. 1982; Parsons 1971) can rectify. What crops were grown on the Cano Ventosidad fields? Zucchi and Denevan suggestedthat manioc was the principalcrop, in part becauseArauquinoidsites generallyyield budares (griddles), but not manos or metates (Denevan and Zucchi 1978:244;Zucchi 1973). Pollen samplesrecovered from test excavationsin the canal betweentwo ridgesand analyzedby Mana Salgado de Labounau (1979:97-99) show that savanna grasses (Gramineae)comprise 83.4 percent of all pollen grains counted. Squash (Cucurbitaceae)are also present;the cultivated genus Cucurbita constitutes 2.6 percentof the count, and wild squash (Melancio sp.) makes up only .1 percent.Manioc (Manihot esculenta)pollen was not found,but becausethis cultigen(whichis propagatedvegetatively) generally producesvery low quantitiesof pollen, it cannot be ruled out as a possible crop. Maize (Zea mays) was also not positively identified.Salgadode Labouriau,however, noted grains identifiedas Gramineae, which were too largeto be so classifiedwith confidence,but yet were not quite largeenough to be identifiedsecurelyas maize or anothercultigen.She suggests(Salgadode Labouriau1979:99) that early cultigens may have had somewhat smallerpollen grainsthan their present-daycounterparts, which leaves open the possibility that maize was among the cultigens. The role of the Cano Ventosidad fields in the economic organizationof the region can only be partiallyaddressed,given the lackof an intensive regionalsettlement-patternsurvey.It is particularly difiicultto determinewhetherthe fields should be seen as a responseto demographic pressures,a position not infrequentlytaken by drained-fieldresearchers(Darch 1983; Denevan 1982). Nevertheless, Denevan and Zucchi (1978:244) have assertedthat iithe presenceof enormous ndged fields at Cano Ventosidad is indicative of a large regionalpopulation which could not be sustained by shiftingcultivation in the available forest on high ground."The premise underlyingthis statement appearsin the precedingsentence:iiThe existence of intensive forms of agricultural productionand reclamationof marginalland by subsistencefarmersin the humid tropics is invariablythe result of dense populations." Propositions of this sort reveal the potential importance of drained-field researchto general anthropologicalconcerns such as the causes and consequencesof agricultural intensificationand population growth (Boserup 1965; Cowgill 1975; Sanderset al. 1979; Spencer 1979). If this potential is to be fully realized,however,drained-fieldsystems will have to be incorporatedinto regional-scaleresearchdesigns that entail settlement-patternsurvey as well as site mappingand excavation. Ethnographicwork was carriedout by Denevan and Bergmanin 1972 among the Karinyaof the easternllanos, a Carib-speakinggroupstudiedby KarlSchwerinin the 1960s(Denevan and Bergman 1975;Denevan and Schwerin 1978; Schwerin 1966). Along with other pursuits, includingfishing, swiddenmixed cropping,house gardens,and cash cropping,the Karinyawere practicing intensive drained-fieldcultivation in the swampy river-bottom and flood-plain zone called the morichal, namedfor the dense groves of morichepalm (Mauritiaflexuosa or Mauritia minor) that thrive there.A detailed map (Figure 3) was made of one drained-fieldsystem, called Bajo Hondo, that covered6.24 ha (Denevan and Schwerin 1978:3s31). Manioc was the principalcrop grownon the fields,which also producedbanana,ocumo, maize, beans,and squash(Denevanand Schwerin1978: 25-26).The investigatorsdid not collect productivitydata, but they noted that the Bajo Hondo fieldsystem provideda significantportion ofthe sustenancefor 16 households;the averagewas .39 haper family, a figureiiindicativeof highly productiveagriculture"(Denevan and Schwerin 1978: 31).The Karinyaalso cultivatedhouse gardensthat produceddietarysupplements and other useful products such as coconut, cashew, various fruits, and ornamentalflowers(Denevan and Schwerin 1978:34-35).In the rainy season, house gardens also contained basic subsistence crops such as maize,beans, squash, and chili peppers, whereas in swidden zones the major food crops were manioc,maize, squash,beans, banana,plantain,and watermelon(Denevan and Schwerin 1978:34, 18).Cashcroppingof cotton, maize,peanuts,and sesamewith mechanizedequipment on the savanna andOrinoco flood plain began in 1972 (Denevan and Schwerin 1978:38-40). The overall picture isone of a highly diversifiedsubsistenceeconomy, with individual households cultivatingfields in several differentlocations.

Spencer et al.]

DRAINED FIELDS AT LA TIGRA

123

Figure3. Map of the Bajo Hondo drainedmorichalfields (after Denevanand Schwerin1978:Figure5).

In the mid-1970s Garson (1980) conducteda regionalsurveyaugmentedby 19 test pits at 5 sites in the Hato La Calzadaarea. In his study regionof 120 km2he locateda total of 22 habitationsites and nine calzadas (causeways)that he identifiedas Osoid-seriesoccupations,and another 13 habitation sites of the Cano Caroni period (Garson 1980:98). Zucchi (1967, 1973, 1976) places the Osoid series between 230 B.C. and A.D. 1200 and the Cano Caroniperiod betweenA.D. 1200 and 1500. Garson (1980:130) also reports evidence of drained fields, in the form of parallel ridges orientedperpendicularto the levees of the Ticoporo River, coveringan area of about 10.3 ha. The fieldsarejust 1.3 km from Mound GroupA, an Osoid habitationsite (Figure4), but Garson(1980: 129-130) was reluctantto draw an association between the sites because he discovered the fields while examining aerial photographsof the region well after the fieldworkperiod. Hence although the La Calzada drained fields were found in the context of a regional-scaleresearchdesign, they await verificationthroughgroundreconnaissance. In the sixteenthcenturythe llanosof westernVenezuelawereinhabitedby a numberof independent polities, some of which are reportedto have used artificiallydrainedor raisedfields(Gumilla 1963: 429434 [1745]; Morey 1975:147). A number of Caquetio chiefdoms that occupied a territory stretchingfrom the llanos properinto the sub-Andeanvalley of the Rio Claro near Barquisimeto are said to have irrigatedsome 19 km of fields on both sides of the river (Morey 1975:51).Among the crops grown were two varieties of maize, squash, manioc, sweet potatoes, cotton, and tobacco (Morey 1975:51, 85). LA TIGRA: DRAINED FIELDSIN REGIONALCONTEXT Between 1983 and 1988 Spencerand Redmond carriedout a regional-scalearchaeologicalproject in a 450-km2portion of the Rio Canaguadrainageof Barinas(Redmondand Spencer1989; Spencer 1991; Spencerand Redmond 1992). Threeseasons of fieldworklocateda total of 103 archaeological sites througha systematic pedestriansurvey of the entire study region, which overlaps the llanos and the adjacentAndean piedmont (Figure 1). Duringthe fourth season 201 test excavationswere carriedout at 10 differentsites, whereasthe fifthseason was devoted to horizontalblockexcavations at the largest site found on survey. The work produceda four-phaseceramic sequence based on

LATIN AMERICAN 124

ANTIQUITY

[Vol. 5, No. 2, 1994

Figure 4. Map of sites, causeways,and (after Garson 1980:Map24; and based on drainedfields locatedby Garsonin his Hato La Calzadastudy Denevan l991:Figure91.3). region ceramic cross dating with published materials as well for both as radiocarbon and the piedmont and the llanos that stretches in the from A.D. 300 to thethermoluminescence dating sixteenthcentury (Table 1). time of European Although the discovery of ofwork, our we were contact drained fields was not the not surprised, given the aknown major focus place locally as earlier work just La described, to find such ofsettlement-pattern Tigra. To situate these fields in their evidence at regional context, a brief development is required. description

Regional SettlementHistory The first phase of occupation is called Early Gavan in (A.D. the llanos and Early 300-550).Only four sites, three Curbati in the piedmont in the llanos and one time (Figure 5). There was no in the piedmont, clear settlement were inhabited at this the area of the of each hierarchy; the piedmont site three llanos sites is 3 covers 2 to 3 ha, ha, although B12 may be slightly larger. whereas The total

Spencer et al.]

125

DRAINED FIELDS AT LA TIGRA

Table 1. CeramicChronologyfor Andean Piedmontand AdjacentLlanos in the Rio CanaguaDrainage,Barinas. Years A.D.

High Llanos

Piedmont

Chuponal Late Cano Seco

Late Cano Seco

Early Cano Seco

Early Cano Seco

Late Gavan

Late Curbati

Early Gavan

Early Curbati

1550

1350 1000 550

300

occupied area is estimated to be 2 to 3 ha in the piedmont and 9 to 11 ha in the llanos. There is no evidence of any earthworkconstructionduringthis early phase. Majorchanges occurredin the region in the succeedingphase, which we call Late Gavan in the llanos and Late Curbatiin the piedmont (A.D. 55s1000). A total of 42 sites can be dated to this time period, 10 in the piedmont and 32 in the llanos (Figure6), figuresthat imply a considerable increasein the population of both zones. We estimate the total occupied area for Late Curbatiin the piedmont at 16 ha, a five-fold increase from Early Curbatitimes. In the llanos, population growthwas even more dramatic;the total occupied area duringLate Gavan is estimatedto be 124 ha, an increaseof 1,140 percentover the EarlyGavan phase.

, *

ErJrly

*

Early Gava'n Sites

Curbutr

Sites,

V\

)S,>X\vC

\

>YX\\S,\

t

\

\

7

\

)

Figure5. RegionalsettlementpatternsduringEarly Gavanand Early Curbatiphases.

LATIN AMERICANANTIQUITY

126

*

Curbot Complex Sites

\7ft

O Petroglyphs

Gavon Complex Settlement Hierarchy \

lst-Order

g

<

XX7

* 3rd0rder Sites

sti7r

\

gs

\

+,505\,

0

'

(

\

\

\

\ V

t

* C a z ada

Figure6. RegionalsettlementpatternsduringLateGavanandLateCurbatiphases. The Late Gavan occupation of the llanos exhibits a three-level regional settlement hierarchy in site size and associated mounded architecture (Figure 6). The Gavan site, B12, is at the top of this hierarchy.With an area of at least 33 ha it was by far the largest site of its time, and it also contained an impressive assortment of earthworks that included two large mounds (12 m and 10 m tall, respectively) on opposite sides of a plaza that measures 500 m across (Figure 7). The taller of the mounds measures 90 m across at the base and has the remains of an earthen ramp that still extends 80 m into the plaza. A number of internal calzada-like features were also recorded, as well as 134 smallerearthen mounds, all of which appear to have been housemounds. Variations in housemound height, housefloor area, associated artifacts, and burial patterns undoubtedly reflect considerable socialdiffierentiation at this time (Spencer and Redmond 1992). The scale of all the housemounds, however, is commensurate with a nuclear family, as was the casefor the ethnohistorically documented Caquetio (Morey 1975:92; Spencer and Redmond 1992). Becausehabitation debris was found in parts of the site without visible housemounds, we suspect thatsome inhabitants built their homes on the ground surface or on mounds so low that they are notnow visible without excavation. We estimate the total population of B12 to range from a minimum of 134 households to a maximum of 200 households, or approximately 670 to 1,000 peopleat 5 persons/household (Baker and Sanders 1972). This would produce an overall density of20.3-30.3 persons/ha for B12, a figure not out of line with the densities presented by Parsons (1971:23) for the Basin of Mexico settlement types called Compact Low-Density Village (1>25 persons/ha)and High-Density Compact Villages (25-50 persons/ha). Circumscribing the occupation at B12 is a calzada-like earthwork that forms an enormous oval about950 by 470 m. A portion of the feature has been eroded by a seasonal stream (madre vieja) thatpasses by the northwest corner of the site. The earthwork measures some 2s25 m wide at the

127

DRAINED FIELDS AT LA TIGRA

Spencer et al.l

/2t,

"\

* f

/St

g S/

s

Figure 7. Map of the Gavan site (B12), showing the majormounds,smaller housemounds,causeways,the encirclingearthwork,the areas of block excavation(markedwith capital letters),and the numberedtest pits.

base, 6-8 m wide at the top, and still reaches 1 meter or more in height (Figure7). Other calzadas radiate out from the B12 site, linking it to a number of smaller Late Gavan sites (Figure6) in a network that is surely a manifestationof substantialregionalcentralizationand cohesion by this phase. Five of the sites (B97, B2 1, B25, B 17, and B30) can be interpretedas second-tieroccupations

128

LATIN AMERICANANTIQUITY

[Vol. 5, No. 2, 1994

Figure 8. The Gavan locality, showing the first-orderregional center (B12), a seconderder site (B97), a numberof third-ordervillage sites (B52, B96, B98, B26), and the drained-fieldfacility (B27).

in the Late Gavan regionalsettlementhierarchy:B97, B21, B25, and B30. These sites rangein size from6 to 10 ha and have two to four mounds that reach 2 to 6 m in height. There are also 26 habitationsites in the third level of the Late Gavan regional settlement hierarchy.No mounded architecturewas noted at any of these sites, which vary in size from .5 to 4.4 ha. In the transition from Early to Late Gavan we see the emergence of a three-level settlement hierarchyfocused on a large regionalcenter, the constructionof monumentalearthworks, considerablepopulationgrowth,and social-statusdifferentiation.We feel that "chiefdom"is an appropriate labelfor this development in view of the evidence of centralizedregional organizationand, at the sametime, the relativelylow degreeof diversityin ceremonialor public architectureat the regional center(Spencer 1987, 1990; Spencerand Redmond 1992).

TheLa TigraFields:Morphology, Chronology, and Use The calzadathat issues from B 12 towardthe southeastpasses alongsideB96 and B26, both thirdordervillage sites, before skirtingthe edge of a drained-fieldfacility, B27 (Figure8). B27 is situated inthe vega(flood plain) of the Rio Canagua,which has soil of a type that is more fertile and far easierto till than the surroundingsavanna clays. Two remnant river bends (now oxbow lagoons) anda cano(the Cano Colorado)were linked by a networkof canals to create approximately35 ha ofdrainedfields(Figure9). Sucha combinationof naturalfeaturesand artificial construction,Zucchi (1985)has argued,was highly characteristicof Prehispanicagroengineeringprojects in the Venezuelanllanos. The canals at La Tigra are variable in size; the widest is some 8 meters across, the majority have widths of 4 to 6 m, and the narrowestis just a meter wide. Canaldepths vary from 30cm to 2.0 m, with most in the .5-1.5 m range.We mapped a total length of 2,960 m of canals withalidade and plane table (Figure 10);an additional832 m had to be located less precisely(with compass, tape measure,and aerialphotography)becauseof extremelydense vegetation.In addition

Spencer et al.]

DRAINED FIELDS AT LA TIGRA

129

Figure9. La Tigra (B27). A small stream or cano, the Cafio Colorado,was linked to two oxbow lagoons creatinga 35-ha field systemthat drainedinto the Rio Canagua.Numberedtest pits are indicated,as are lettered

Figure 10. View of the B27 drained fields. A field assistant (Rafael Gasson) holds a stadia rod in the canal between Field A and Field D.

130

Figure 11.

LATIN AMERICANANTIQUITY

[Vol. 5, No. 2, 1994

Potrero Urpianero (B26), a third-order village site covering 3 ha. Numbered test pits are indicated.

to the clearlyartificialcanals, there were modified sections in the course of the Cano Colorado;we estimated the total length of modified sections to be 534 m. A grandtotal of 4,326 m is produced by combining the lengths of canals and modified cano sections. The two oxbow lagoons appear largelyunmodified,althoughthere is one possible boat landing (Figure9). When was the B27 facility created?The largesthuman occupationever seen in this locality was during the Late Gavan phase. At the present the area is sparselyinhabited, with only 5 families (about 25 people) scatteredin isolated households within 2 km of B27. We made an effiortto talk with all the farmersin the Gavan locality in orderto collect ethnographicdata on land-use history and currentpracticesin the area. Without exception, they claimed that the networkof canals at La Tigrawas extremelyold and not the work of currentfarmersor thefundadores (foundingfarmers), all of whom enteredthe regionwithin the past four decades.Our informantswere uniformlyof the opinion that the La Tigracanals were dug by the same ancient peoples (usuallyreferredto as "los Indios")who built the mounds (cerritos) and calzadasthat dot the llanos landscape.We wereinclined to agree with their judgment and date the fields by association with the abundant Late Gavan occupation of the area. We placed four test pits in the fields and recovered soil samples but no potterythat we could use for dating.Thoughdisappointing,the resultwas not unanticipatedbecause domestic artifactswould probablynot have been deposited with any frequencyunless people had lived on or alongside the field system. At 1 km away, B26 is the nearest habitation site and the most likely home village of those who cultivatedthe B27 fields (Figure11). Any provisionsbrought by workersto the site were probablypackedin gourdsor bags (as llanerosdo today) ratherthan in pottery vessels. In 1988 we were able to interviewthe currentowner of the small farm on which B27 is situated. He told us he was the firstfundadorin the locality, and had settled there more than 30 yearsbefore. He uncoveredthe canal network as he was first clearingthe monte(thick brush)in order to plant maize, beans, manioc, and other crops. After some cleaning,he found that he could use the canals to facilitate drainage of the area during the rainy season, particularlyfrom September through November when severe floodingcan damageplants.Moreover,the canalscould be used to marshall the waterthat falls duringthe sporadicshowersthat markthe end of the dry season (January-May). The owner informed us that the canal networknot only minimized crop loss throughdehydration

Spencer et al.]

DRAINED FIELDS AT LA TIGRA

131

Table 2. Pollen Analysis Results from Sites B27 and B26: Family- and/or Species-Level Identificationsand PotentialUses.

Plant Maiz (Zea mays) Aji(Capsicumfrutescens,Solanaceae) Corozo(Acrocomiasclerocarpa,Palmae) Guapo(Marantaarundinacea,Marantaceae) Palomero(AWyricapubescens, Myricaceae) Quina (Chenopodiumquinoa) Tomate(Lycopersicumesculentum,Solanaceae) Anil (Indigoferatinctoriaor I. anil, Leguminaceae) AstroloCa (Aristolochiaringens,Aristolochiaceae) Guayabo(Psidiumguajava,Myrtaceae) Icaco or hicaco (Chrysobalanusicaco, Chrysobalanaceae) Ocumo(Xanthosomasagittifolium,Araceae) Yopo(Piptadeniaperegrina)

B26 B27 VilFields lage

PotentialUse

Reference

6 1 2 1 1 2 2 0 0 0

6 0 0 0 Q 0 0 4 2 1

maize chili peppers fermenteddrink arrowrootstarch treefruit quinoa fruit possible dye ornamentalplant tree fruit

Schnee 1973:692 Alcorn 1984:674 Schnee 1973:56 Schnee 1973:364

0 0 1

1 1 1

tree fruit tuber hallucinogen

Schnee 1973:392 Schnee 1973:505 Morey 1975:85

Morey 1975:46 Morey 1975:47 Schnee 1973:232 Morey 1975:78 Schnee1973:287

Note:Keyto relativefrequencies:0 = Absent;1 = verylow frequency;2 = low frequency;3 = low to intermediate frequency;4 = intermediatefrequency;5 = intermediateto high frequency;and 6 = high frequency.

or flooding but also extended the effective growing season. He recalledthat the land consistently yielded two successfulharvestsof maize per year owing to the longergrowingseason that the canal system permitted,whereasother farmersin the areacould only count on a single harvesteach year. He clearly rememberedannual yields of 3,600 kg of husked maize/ha/yearon the drained fields, approximatelydouble the amount that was generallyharvestedon other vega soils in the area. Was maize also grownon the B27 fields in Late Gavan times? Rinaldi (1990) conducteda pollen analysis of soil samples from B27 fields (Sample 442, Test Pit 186,40 cm DBS) and also the B26 village site (Sample 114, Test Pit 78,40 cm DBS). Her family- and/or species-levelidentifications are presentedin Table 2; identificationsto genus/subfamilyand/or family level are given in Table 3. She found that maize (Zea mays) pollen was presentin high frequenciesin both samples (Table 2, Figure 12). The predominanceof maize is noteworthyin view of the earlierCano Ventosidad research,which concluded on the basis of artifactualevidence that manioc was the primarycrop grown on drained fields (Denevan and Zucchi 1978:244). As at Cano Ventosidad, the B27 palynological analysis recordedno manioc pollen (Rinaldi 1990), an observationthat does not rule out the possibility of manioc cultivation given the low pollen productiontypical of this plant (Salgado de Labouriau1979:99). Nevertheless, the results suggestthat maize was a majorcrop at both B26 and B27, an interpretationconsistent with Morey's (1975:46) assessment of the ethnohistorical accounts: "Corn receives more emphasis than yuca [manioc] in the reports of food crops for all northernand western areas of the Llanos." Although the B27 and B26 samples both have high levels of maize pollen, they differ in their representationsof other plant species, one of which occurredin intermediatefrequencyand the others in very low or low frequencies(Table 2). This suggeststhat maize was extensively grown both on the B27 drainedfields and in conucos(swiddenplots) and/or house gardensnear or within the B26 village, whereasthe secondarycrops were variouslycultivatedin the two locations.At B27 the secondaxyplantsincludedajz(Capsicumfrutescens),corozopalm (Acrocomiasclerocarpa),guapo (Maranta arundinacea),palomero (Myricapubescens),quina (Chenopodiumquinoa), tomato (Lycopersicumesculentum),and yopo (Piptadeniaperegrina).At the village site the followingsecondaxy plants were in evidence: anil (Indigoferatinctoria),astroloVa(Aristolochiaringens),guava (Psidium guajava), icaco (Chrysobalanusicaco), ocumo (Xanthosomasagittifolium),and yopo (Piptadenia peregrina).Note that one root crop, ocumo, was evidently cultivated in or near the B26 village, whereasguapo (a root starchthat may not have been, strictlyspeaking,a domesticate[Morey 1975:

-

t

-

-

-

-

-

-

-

-

-

-

-

w -

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

s

-

-

t

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

v

-

-

-

-

r

-

-

-

x

-

-

w -

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

F -

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

z

s

-

-

-

w -

x

-

-

-

X -

-

-

-

-

-

-

-

-

-

LATIN AMERICANANTIQUITY

-

[Vol. 5, No. 2, 1994

132

Table 3. Pollen Analysis Results from Sites B27 and B26: Genus-/Subfamily- and/or Family-Level Identifications.

Plant

B27 Fields

B26 Village

5 0 0 0 0 2 0 1 0 0 0 0 0 2 1 1 1 1 2 1 1

3 4 4 2 3 5 2 5 2 3 1 1 1 0 0 0 0 0 0 0 0

Compositae Plantago Sida (Malvaceae) Trigonia (Trigoniaceae) Symplocos Dyckia (Bromeliaceae) Talinum (Portulacaceae) Alternantera (Amaranthaceae) Onograceae Pistacia (Anacardiaceae) Herpalyce (Leguminoceae) Alsophila Polypodium Cuphea (Lythraceae) Cecropia (Moraceae) Gomphrena (Amaranthaceae) Polycarpae (Caryophyllaceae) Cyperaceae Juglans (Juglandaceae) Ruellia-Distictella (Bignoniaceae-Acanthaceae) Ipomoea (Convolvulaceae)

Note: Key to relative frequencies: 0 = Absent; 1 = very low frequency; 2 = low frequency; 3 = low to intermediate frequency; 4 = intermediate frequency; 5 = intermediate to high frequency; and 6 = high frequency.

TOMATE QUINA PALOMERO GUAPO COROZO AJI YOPO

OCUMO ICACO GUAYABO ASTROLOJA ANIL

-

o

-

sssssS

Wsssssssssssl

MAIZ -M--N-i

i

i

r

2

3

4

l

1

W

l

r

6

Relative Frequencies N

B26VillageSite

B27 DrainedFields |

Figure12. Bar graphof plant species identifiedby pollen analysis of samplesfromB26 village site and B27 drainedfields.

Spencer et al.]

DRAINED FIELDS AT LA TIGRA

133

78]) was grown on the B27 drainedfields. The plant species identifiedin the drainedfields sample were largely those suitable for food (maize, aji, guapo, palomero, quina, and tomato) or drink (corozo) although traces of yopo (a hallucinogen)were also found. The village site, on the other hand, producedevidence for ornamentalhouse plants(astroloja), a plantdye (anil ), and yopo, along with food plants (maize, ocumo, guava). Such variabilityin the plants cultivatedby a single community is consistent with ethnohistoricand ethnographicdata on the diversifiedagriculturalstrategies of indigenousllanos groups (Denevan and Schwerin 1978; Morey 1975:45-86). At 35 ha, the scale ofthe B27 drainedfields is unimpressivecomparedto the 15.5 km2estimated for the Cano Ventosidad field system (Denevan and Zucchi 1978:240), although it is somewhat largerthan the 10.3-ha system spotted on aerial photographsof Hato La Calzada(Garson 1980: 130) and the 6.24-ha morichal fields used by the Karinya(Denevan and Schwerin1978). Why were no other drainedfields found in the regionalsurveyofthe 1983-1988 BarinasProject?One possible answeris that the La Tigraarea was unusuallyprone to inundations-the elevation of 160 m asl at B27 places it among the lower elevations in our study region. On the other hand, this elevation is hardlyunique, and the inhabitantsof many of the sites near the Rio Canaguaflood plain (including B25, B96, B97, B98, B12, B52, and B64) would have cultivated fields that were no less prone to flooding than the B26-B27 area. We consequentlysuspect that there were other drained fields in the Gavan locality duringLate Gavan times. Indeed, we were told by informantsthat canals like those at B27 were visible in the vega between B97 and the Rio Canagua,but were destroyedin the 1970s by a previouslandownerwho used heavy machineryto modify the terrain.We surveyedthis area and noted no evidence of drained fields. Because several other parts of our study region are also owned by wealthy individuals who have worked their farms over the years with large-scale equipment, we must recognize the very real possiblity that drained-fieldcultivation was more widespreadin this region than the single field system at B27 indicates. Demographic and Economic Implications

Why werethe drainedfieldsconstructed?Both the pollen analysisand our informantdata indicate that successfulcultivation can occur without drained-fieldtechniquesin the rich vega soils of the B26-B27 area. Eachyear several farmersharvesta healthy crop of maize here, along with a variety of other plants including beans, chili peppers, squashes, manioc, papaya, and plantains. Swidden methods are widely used. The slashing takes place early in the dry season, usually in Januaryor early February,with the burningfollowingin March.Sowinggenerallyoccursin May afterthe rains begin in earnest, and harvestingis usually concluded by September,just before the heaviest rains and consequentflooding. Without techniquesof water management,only a single crop per year is feasible. Fallowingis generallynot essential to restorethe fertilityof swidden plots, because fresh silts are widely deposited by the annual flooding. Nevertheless, after two to five years of steady cultivation, fields tend to become weed infested, and farmers sometimes consider it less laborintensive to slash and burn a plot of secondaryforestgrowththan to pull weeds. Fallow terms vary widely, from as little as two years to 10 years or even longer;the span has more to do with the needs, aspirations,and abilities of a household than with the fertilityof the soil. Given the considerable labor entailed in digging and maintainingcanals, it is reasonableto ask what would have promptedthe Late Gavan inhabitantsof the area to build drainedfields. We have noted that an essential advantage offered by the drained fields is their potential for regulardouble cropping.Whatmight have motivated such an increasein productiveintensityduring Late Gavan times? Some scholarshave arguedthat Prehispanicdrainedor raised fields were constructedin response to demographicpressures.For example, in a wide-rangingreview of the literature,Denevan (1982:193) proposedthat "populationssupportedby raised-fieldagriculturewere substantial"and that "even a small number of raised fields in a region is probablyindicative of demographicpressureon the land." In a similarvein, Darch(1983:7) concludedthat "drainedfields are clear evidence of the control of wet season inundation and dry season droughtswhich when combined with activities of good managementhave produceda very lucrativeform of agriculture, certainlyan answerto the food requirementsof large populations."

LATIN AMERICANANTIQUITY

134

[Vol. 5, No. 2, 1994

Table 4. Site Sizes and PopulationEstimatesfor the Gavan Locality.

Site

Area (ha)

Population

B 12 B97 B96 B98 B52 B26 Total

33 6 1.25 1 1.25 3 45.5

67s1,000 122-182 25-38 20-30 25-38 61-91 92F1,379

Households 13F200 2o36 5-7 4-6 5-7 12-18 185-275

Note: Population- and household-size estimates are based on a density of 20.3-30.3 persons per hectare of occupation, derived from the housemounddensity estimate for the B 12 site, with an assumption of 5 persons/household (Baker and Sanders 1972; Morey 1975:92).

Are there indications of demographic pressure on the land in Late Gavan times? We have noted that there was considerable population increase in the general study area during the period (cf. Figures 5 and 6). If we extrapolate from the population density inferred for site B12 (20.3-30.3 persons/ha), we derive a population estimate for the entire Late Gavan regional system (124 ha) of 2,517-3,757 people, or 503-751 households at 5 persons/household. Examination of the overall settlement pattern makes it clear, however, that the area of densest Late Gavan occupation is the Gavan locality (Figure 8). Hence if there was a condition of population pressure in Late Gavan times, it is in the Gavan locality that we should find the evidence for it. We calculate that the six habitation sites in the Gavan locality covered a total of 45.5 ha of occupation during Late Gavan times (Table 4). Use of the same density estimate of 20.3-30.3 persons/ha produces a total of 924-1,379 people or 185-276 households of 5 persons each (Table 4). In the Gavan locality there are 3,125 ha of land overall, of which 1,264 ha qualify as prime vega alluvium (Figure 8). In order to investigate whether demographic pressures on an extensive production system prompted the construction of drained fields, we must ask if the population levels of Late Gavan could have placed excessive demands on the vega under a single-cropping regime. Because maize was the predominant plant identified in the pollen analysis at both B26 and B27, we base the following estimates on maize production and consumption. This simplifying assumption admittedly introduces an element of error into the analysis. To the extent that farmland was planted in crops other than maize, the exercise will overestimate maize production and hence potential population. On the other hand, the magnitude of the potential error is likely to be small, because the traditional multicrop conuco agriculture in this area calls for cultivation of a variety of plants (including beans, chili peppers, squashes, tree fruits, and even manioc) along with maize in the same fields. An additional potential source of error is incorrect estimates of the cropping and fallow terms. Drawing upon our own ethnographic inquiries, we assume that cropping terms range from two to five years and fallow terms from two to 10 years. These figures yield cropping fractions of .2-.5, which in turn reduce the total arable vega in the Gavan locality to 253-632 ha available for cultivation at any given point ln tlme. But were there enough potential farmers at the six Gavan locality sites to keep 253-632 ha under cultivation? Kirkby's (1973:Table 10) data from the Valley of Oaxaca indicate that a traditional farmer with a digging stick can keep 2 ha under cultivation. If each of the estimated 185-275 households in the Gavan locality had a single full-time farmer, then some 37O550 ha could have been kept under continuous cultivation, figures that fall comfortably within the estimated range based on the cropping-fallow cycle. This suggests that a productive strategy of extensive cultivation could have been maintained in the Gavan locality vega without exhausting the soil or placing undue labor demands on individual households. Moreover, if each household had fielded just two farming .

.

Spencer et al.]

DRAINED FIELDS AT LA TIGRA

135

members (Kowalewski 1980:156; Spencer 1979:50), a total of 740-1,100 ha could have been cultivated, an estimate that does not just exceed the range based on the ethnographic fallow cycle but amounts to 59-87 percent of the total 1,264 ha of vegain the Gavan locality. Farming of this much land in Late Gavan times would, of course, have called for a shorter fallow term than the one currently followed. On the other hand, there is no reason to rule out a shorter cycle, in view of the fact that our ethnographic informants cited weed invasion and not declining soil fertility as the major reason for letting a field return to fallow. Because fertility is apparently sustained by the seasonal flooding, the fallow cycle could presumably be shortened with no significant decline in crop yields-though admittedly with rather more effort expended in weeding. How much maize could have been produced by such an extensive single-cropping regime on the vega?In her pollen analysis Rinaldi (1990) notes that the maize pollen in the B26 and B27 samples is somewhat smaller in diameter than the pollen of contemporary maize, and suggests that an indigenous race of maize, smaller than modern races, was grown in Late Gavan times. If this was the case, then we would be ill-advised to base our estimates of Late Gavan maize productivity solely on the 1,800 kg/ha per harvest figure obtained through our ethnographic inquiry in 1988. Kirkby (1973: 126) has shown how mesoamerican maize cobs steadily increased in length over time, with a concomitant increase in overall maize productivity. Although no measurable maize cobs were recovered in our excavations, the productivity of maize in Late Gavan times would be about 63 percent of present-day yields, according to Kirkby's graph (1973:126). For single cropping, this would be about 1,134 kg of kernels/ha/year. A different view is taken by Roosevelt (1980:239), who proposes that the La Betania (A.D. 6501200) maize excavated by Zucchi was no smaller than modern races of Venezuelan maize, in particular the variety known as Chandelle. If the maize grown in the Gavan locality during Late Gavan (A.D. 550-1000) was similarly productive, then the Kirkby correction would seem conservative. We have generated estimates of maize productivity using both assumptions, as well as varying assumptions of the available labor force that range from one to two farmers per household (Table 5). The estimates were computed for all six habitation sites of the Gavan locality separately and together, assuming one harvest per year. We also present estimates for labor requirements and maize productivity at the drained fields (B27), in this one case assuming two crops per year, in line with our ethnographic information. We combined the data with dietary estimates of maize requirements in order to produce estimates of the population that could have been supported by the Gavan locality vega(Table 6). We assumed an annual maize requirement of 1,100 kg/household, the midpoint ofthe 1,000-1,200 kg/household range usually cited in the literature (Coe and Diehl 1980:78; Flannery 1976; Parsons 1976). We then generated estimates for both the Kirkby correction for cob size and productivity (called Estimate 1 in the table) and the Roosevelt assessment (Estimate 2). Following this we varied the estimates according to assumptions of one or two farmers per household and the range of the estimated available labor supply. Separate estimates of potential population were produced for site B27, the drained fields (Table 6). A summary line graph of the Table 6 results is presented as Figure 13. To construct the graph, we used the midpoints of the actual population estimates, along with the midpoints of the range of potential population estimates under Estimate 1 (the Kirkby correction), assuming a labor supply that varied from 1 to 2 farmers/household (Figure 13). The outcome of this exercise is unambiguous. Whether the Gavan locality habitation sites are considered separately or together, the potential population that could have been sustained with a single-cropping strategy on the vega was considerably higher than the archaeologically estimated population. As Table 5 indicates, the B27 fields could have been kept under cultivation quite easily by just the people who lived at adjacent B26. The potential yields would, however, have greatly exceeded the nutritional needs of the 12 to 18 households at B26; the drained fields could have sustained an estimated 72 to 114 households (Table 6). The estimates are not consistent with the idea that drained-field agriculture was a direct response to demographic pressures. How, then, should the B27 fields be interpreted? We draw attention to the potential of the fields for surplusproduction.By using drained fields to produce two successive

Table 5. Labor,Land, and Maize-ProductivityEstimatesfor the Gavan Locality Kirkby Productivity

Labor Supply (Households)

Cultivable Land

B12, B97, B96, B98, B52, and B26

185-275

370-550 ha (1 farmer/household) 740-1,100 ha (2 farmers/household)

419,580-625,968

B12

13v200

268-400 ha (1 farmer/household) 536-800 ha (2 farmers/household)

303,912-453,600

kg/yr

607,824-907,200

kg/yr

Sites

B97

B96

B98

B52

B26

B27

Note:

2v36

5-7

4 6

5-7

12-18

(drained fields)

48-72 ha (1 farmer/household) 96-144 ha (2 farmers/household) 10-14 ha (1 farmer/household) 20-28 ha (2 farmers/household) 8-12 ha (1 farmer/household) 16-24 ha (2 farmers/household) 10-14 ha (1 farmer/household) 20-28 ha (2 farmers/household)

kg/yr

839,160-1,251,936

54,432-81,648

kg/yr

kg/yr

108,864-163,296

kg/yr

11,340-15,876 kg/yr 22,680-31,752

kg/yr

9,072-13,608 kg/yr 18,144-27,216 kg/yr 11,340-15,876 kg/yr 22,680-31,752

kg/yr

24-36 ha (1 farmer/household) 48-72 ha (2 farmers/household)

27,216-40,824

kg/yr

35 ha (18 farmers required; 9-18 households)

79,380 kg/yr

54,432-81,648 kg/yr

Singlecroppingis assumed for all calculationsexcept for those pertainingto B27, wheredouble croppingis assum

137

DRAINED FIELDS AT LA TIGRA

Spencer et al.]

Table 6. Estimatesof Potentialand Actual Populationfor Gavan LocalityHouseholdsDuringthe Late Gavan Phase.

Sites

Population Estimate

Potential Population Estimate la

Potential Population Estimate 2b

B12, B97, B96, B98, B52, and B26

185-275

381-569 (1 farmer/household) 762-1,138 (2 farmers/household)

605-903 (1 farmer/household) 1,21>1,806 (2 farmers/household)

B 12

134-200

276-412 (1 farmer/household) 552-824 (2 farmers/household)

438-654 (1 farmer/household) 876-1,308 (2 farmers/household)

B97

24-36

49-74 (1 farmer/household) 98-148 (2 farmers/household)

78-117 (1 farmer/household) 156-234 (2 farmers/household)

B96

5-7

1>14 (1 farmer/household) 2s28 (2 farmers/household)

16-22 (1 farmer/household) 32-44 (2 farmers/household)

B98

66

8-12 (1 farmer/household) 16-24 (2 farmers/household)

13-19 (1 farmer/household) 26-38 (2 farmers/household)

B52

5-7

1>14 (1 farmer/household) 2>28 (2 farmers/household)

16-22 (1 farmer/household) 32-44 (2 farmers/household)

B26

12-18

2F37 (1 farmer/household) 48-74 (2 farmers/household)

39-58 (1 farmer/household) 78-116 (2 farmers/household)

72

114

B27 (drained fields)

Note: Single cropping is assumed for all calculations except for those pertaining to B27, where double cropping is assumed. Annual maize requirement of 1,100 kg/household is assumed (Coe and Diehl 1980:78; Flannery 1976: Parsons 1976). a Estimate 1 is based on the Kirkby (1973: 126) correction. b Estimate 2 is based on Rooseveltss (1980:239) assessment of prehistoric maize productivity.

harvests, a given agriculturalvillage could have doubled its yearly output without doubling the amountof land kept underextensivecultivation,a strategythat wouldalmost certainlyhave required a much largerlabor force. Althoughthe initial diggingof the canals surelyentailed significantlabor costs, the effort probably could have been carried out during the nonagriculturaldry season (or perhapsover a series of dry seasons)usingthe existinglabor force. As RobertD. Drennan(personal communication 1993) has suggestedto us, the drainedfieldsmay have been a way to increaseoutput in a context where available labor was a major factor limiting surplusproduction. We have seen that even a relatively small community such as B26 could have generated a considerablesurplus,which may have been of no small importancein the regionalpoliticaleconomy. Becauseboth B27 and B26 are connected to B12 by a calzada, it seems reasonableto propose that much if not all of such surpluswas sent to the regionalcenter.The regularmovement of goods and people into the center would have reinforcedthe dominant position of the leadership,and put the mobilized surplus itself at the elite's disposal. Future excavations at B12 could aim to recover

Actual

/

Pop.

138

Est.

+ Potential

Pop.

Est.

[Vol. 5, No. 2, 1994

LATIN AMERICANANTIQUITY

c

500

/

O O

/

400

-

/ /

I

300

/

200

/

100--

1 B98

B96

B52

B26

B97

<

B12

All

Sites

Figure 13. Graphof potentialand actual populationestimatesfor six sites in the Gavanlocality, basedon a single-croppingregime in the vega alluvium.Midpoint of the estimate range was used for numberof actual households.For potential population,midpointof the Estimate I (Kirkbycorrection)range was used for 1-2 farmers/household .

evidence of largestoragevessels and/or facilities associatedwith the elite. For example, one might find evidence of a featurelike the maize storagebin that explorerNicolas Federmannobserved in the house of a sixteenth-centuryCaquetiochief. The bin he reported,was as tall as three men and had a storagecapacity large enough to hold 12 men; it sat on a platformsupportedby four large posts (Federmann1958:112-113 [1555]). To what purposes would such surplus have been put by the regionalleadershipin Late Gavan times? B12 is situated alongside a large expanse of prime vega, which had potential yields (even with single cropping)that would have greatlyexceeded the nutritionalneeds of the center'sinhabitants (Figure 13). It seems unlikely that the surplusgeneratedby B26/B27, not to mention that producedby othervillagesin the region,was destinedfor the routinesustenanceof B 12'spopulation. We propose that mobilized surplusplayed an importantlogistical role in the regionalelite's management of external affairs,particularlylong-distanceexchange and warfare,both of which were characteristicof sixteenth-centuryllanos chiefdoms and were also in evidence in our excavations at B12 (Morey 1975; Spencer 1991; Spencerand Redmond 1992; Whitehead 1988). In peacefultimes a good partof the surpluscould have been used by the Gavan elite for exchange purposes.Morey (1975:252, 255, 257-269) has recordedthat sixteenth-centuryllanos groupsoften exchangedlowland agriculturalproducts for workablestone and other highland items. In several high-statusLate Gavan contexts we excavated beads and pendants made of exotic polished stone includingserpentinite,as well as lithics made of a nonlocal red jasper. A numberof distant source areasfor these materialshave been identified,includingthe high VenezuelanAndes, the Caribbean mountains near Caracas,the SierraNevada de Santa Martaof Colombia, and the high Colombian Andes (Wagner1973; Wagnerand Schubert1972; CarlosSchubert,personalcommunication 1986; Ramon Sifontes, personalcommunication 1988). When war threatened,on the other hand, mobilized surplus could have been used to create a stockpile of foodstuffsat the regional center, to provision warriorsand others during battles and sieges. Morey (1975:280) has also noted how one sixteenth-centuryllanos society used a "palisade of tree trunks and earth" to fortify the "main village," which was kept well supplied in case of

Spencer et al.]

DRAINED FIELDS AT LA TIGRA

139

surpriseattack. Although standingarmies were not maintained,regionalchiefs could muster temporaryfightingforces of considerablesize from the villages of their domain. The sustenanceof such forces would have been facilitatedby stored surpluses(Morey 1975:277). Evidence of hostilites duringLate Gavan times was recoveredin our excavationsat B 12. In the lowest constructionlevel of the site's largestmound we excavatedpieces of at least one disarticulated human, which we tentatively interpretas a sacrificedvictim, possibly a war captive. Two other nondomestic constructionsalso yielded disarticulatedskeletons;whole skeletons were only found as burialsassociatedwith residences(Spencerand Redmond 1992). We also suspect that the oval earthworkat B 12, evidentlybuilt at the interfacebetweenEarlyand LateGavan, could have served, in part, as a defensive feature(Figure7). In our excavation Area B, we recoveredan alignmentof carbonizedpost molds along the centerline of the earthwork,which suggeststhat it once bore a palisade (Spencerand Redmond 1992:144). Signs of warfareare prevalentat the end of the Late Gavan occupationas well. Not only the palisadebut also the latest level of the Area A housemound showed signs of burning,as did many of the uppermostlevels of our excavationsat B 12. It appears that this elite center was set afireby attackersat the end of its occupation. Evidence dating from the beginningto the very end of Late Gavan suggeststhat violence and warfarewere regularoccurrences.Moreover,becausewe found indicationsof fortifications,possible human sacrifice,and conflagrationin our excavations at B12 and not in our excavations at any other Late Gavan sites, we infer that the regional leadershipat B12 was directingboth offensive and defensive forms of warfare.The drained fields and calzadanetwork would have provided importantlogisticalsupportfor these activitiesby ensuringthe expeditiousflow of mobilizedsurplus and musteredwarriorsinto the regionalcenter.

CONCLUSIONS The La Tigra (B27) drained-fieldsystem was discovered duringthe 1983-1988 BarinasProject, which collected household, community, regional,and interregionaldata that document the emergence of a chiefdom polity in the Late Gavan phase, A.D. 550-1000. We assigna Late Gavan date to the drained fields based on association with surroundinghabitation sites. Our mapping and excavationprogramat La Tigradocumentedthe morphologyof the field system, and also revealed (throughpollen analysis)that maize was probablythe predominantcultigen,althougha numberof secondaryfood plantswere also grown.Excavationsat a nearbyvillage site (B26) also showed maize to have been the main crop;the secondaryplantsincluded some species that probablyhad nonfood uses. We have addressedthe demographicand economic implications of the drainedfields in part by consideringwhethertheir constructioncould cogentlybe seen as a responseto populationpressure in Late Gavan times. We have focused on the Gavan locality, the part of our study region that contained not only the drained fields but also the greatestconcentrationof settlement. Use of a combination of archaeologicaland ethnographicinformation permitted us to compare the actual estimatedpopulationwith estimates of potentialpopulationbased on an extensive, single-cropping regimein the vegaalluvium. We found no evidence of demographicpressure,which suggestedthat other factorsmust have motivated the constructionof the drainedfields. Noting their capacityfor doublecropping,we proposedthat the drainedfieldsfunctionedimportantlyin the regionalpolitical economy by facilitatingthe productionof surplus,which could have been transportedexpeditiously to the regionalcenter by way of the calzadasystem. This mobilized surpluswould have provided important logistical support for the regional elite in their managementof interregionalexchange and warfareactivities, both of which were evidenced in our excavationsat the regionalcenter. We do not contend that all drained-fieldsystems had the political-economicsignificancewe have proposedfor the La Tigra fields, and we do not deny that populationpressuremay have prompted drained-fieldconstructionin other cases. We suggest,however, that such questions can be feasibly addressedwhen drained fields are studied through suitable researchdesigns, not in isolation, but ratherwithin their regionalcontexts.

140

LATIN AMERICANANTIQUITY

[Vol. 5, No. 2, 1994

Acknowledgments The 1983-1988 Barinas Projectreceived generousfunding from the National Science Foundation (BNS 85-06192), the ConnecticutResearch Foundation (No. 00220-35-220), the Wenner-Gren Foundationfor AnthropologicalResearch(Grant#4798), and the Universityof ConnecticutFacultyFellowship Program.Spencerand Redmond have been colaboradoresvisitantessince 1983 at the Departamentode Antropologia of the InstitutoVenezolanode InvestigacionesCientificas(IVIC),and Rinaldi is a researchassistantat the Laboratoriode Paleoecologia,Centrode Ecologiay CienciasAmbientalesof IVIC. For makingour yearsof work at IVIC so stimulatingand enjoyable,we thank Erika Wagner,AlbertaZucchi, Carlos Schubert,Jesus EduardoVaz, RafaelGasson,Ines Frias,LilliamArvelo,LuisMolina,and RamonSifontes.Duringthe fieldwork in Barinaswe wereassistedby MariaAndueza,RafaelGasson, Ines Frias,JavierFernandez,TheodoraMeijers, Pablo Novoa, Alejo Novoa, Lucio Laviano,and Raiza Ron. Figures1 and 6 are reproducedwith the permission of the Universityof MichiganMuseumof Anthropology.Figures5, 7, and 11 arereproducedwith the permission of WorldArchaeology.Figures2 and 8-9 were drawnby Diana Salles of the AmericanMuseum of Natural History.Partof this paperwas presentedat the 1992 AnnualMeetingof the Society for AmericanArchaeology in Pittsburgh.

REFERENCESCITED Alcorn, J. B. 1984 HuastecMayan Ethnobotany.University of Texas Press,Austin. Alden, J. R. 1979 A Reconstructionof Toltec PeriodPoliticalUnits in the Valleyof Mexico. In Transformations:MathematicalApproachesto CulturalChange,edited by C. Renfrewand K. Cooke, pp. 169-200. Academic Press, New York. Baker,P., and W. Sanders 1972 DemographicStudies in Anthropology.AnnualReviewof Anthropology1:151-178. Blanton,R., S. Kowalewski,G. Feinman,and J. Appel 1982 Monte Alban'sHinterland,Part I: The PrehispanicSettlementPatternsof the Centraland Southern Parts of the Valleyof Oaxaca, Mexico. Memoirs No. 15. Museum of Anthropology,University of Michigan,Ann Arbor. Boserup,E. 1965 The Conditionsof AgriculturalGrowth.Aldine, Chicago. Brumfiel,E. 1976 RegionalGrowth in the EasternValley of Mexico: A Test of the 66Population Pressure"Hypothesis. In The Early A>esoamericanVillage,edited by K. V. Flannery,pp. 234-249. Academic Press, New York. Coe, M. D., and R. A. Diehl 1980 The People of the River.In the Land of the Olmec, vol. 2. University of Texas Press, Austin. Cowgill,G. 1975 On the Causesand Consequencesof Ancientand ModernPopulationChanges.AmericanAnthropologist 75:505-525. Cruxent,J. M. 1966 Apuntes sobre las calzadas de Barinas,Venezuela.Boletln Informativodel Instituto Venezolanode InvestigacionesCientlyicas4: 10-24. Cruxent,J. M., and I. B. Rouse 1958 An ArchaeologicalChronologyof Venezuela,vol. 1. Social ScienceMonographsNo. 6. Pan American Union, Washington,D.C. Darch, J. P. 1983 DrainedFields in the Americas:An Introduction.In DrainedField Agriculturein Centraland South America,editedby J. P. Darch,pp. 1-10. BAR InternationalSeries 189. BritishArchaeologicalReports, Oxford. Darch, J. P. (editor) 1983 DrainedField Agriculturein Centraland South America.BAR InternationalSeries 189. British ArchaeologicalReports,Oxford. Denevan, W. M. 1966 TheAboriginalCulturalGeographyof the Llanos de Mojos of Bolivia.Ibero-AmericanaNo. 48. University of CaliforniaPress, Berkeley. 1982 HydraulicAgriculturein the American Tropics:Forms, Measures,and Recent Research.In Maya Subsistence:Studiesin Memoryof DennisE. Puleston,editedby K. V. Flannery,pp. 181-203. Academic Press, New York. 1991 PrehistoricRoads and Causewaysof Lowland Tropical America. In Ancient Road Networksand SettlementHierarchiesin the New World,edited by C. D. Trombold,pp. 23s242. CambridgeUniversity Press, Cambridge.

Spencer et al.]

DRAINED FIELDS AT LA TIGRA

141

Denevan, W. M., and R. W. Bergman 1975 KarinyaIndian Swamp Cultivationin the VenezuelanLlanos. Yearbookof the Associationof Pacific Coast Geographers37:23-36. Denevan, W. M., and K. Mathewson 1983 PreliminaryResults of the SamborondonRaised Field Project,Guayas Basin, Ecuador.In Drained Field Agriculturein Centraland SouthAmerica,edited by J. P. Darch. BAR InternationalSeries 189. BritishArchaeologicalReports,Oxford. Denevan, W., and K. H. Schwerin 1978 Adaptive Strategiesin KarinyaSubsistence,VenezuelanLlanos.Antropologica50:3-91. Denevan, W., and A. Zucchi 1978 Ridged Field Excavationsin the CentralOrinoco Llanos, Venezuela.In Advancesin AndeanArchaeology, edited by D. L. Browman,pp. 235-245. Mouton, The Hague. Denevan, W. M., K. Mathewson,and G. Knapp(editors) 1987 Pre-HispanicAgriculturalFields in theAndeanRegion. BAR InternationalSeries359(i and ii). British ArchaeologicalReports,Oxford. Erickson,C. L. 1980 Sistemasagricolasprehispanicosen los Llanos de Mojos. AmericaIndigena40:731-755. 1987 The Dating of Raised-FieldAgriculturein the LakeTiticacaBasin,Peru.In Pre-HispanicAgricultural Fields in the Andean Region, edited by W. Denevan, K. Mathewson,and G. Knapp, pp. 373-384. BAR InternationalSeries 359(i and ii). BritishArchaeologicalReports,Oxford. Farrington,I. S. (editor) 1985 PrehistoricIntensiveAgriculturein the Tropics.BAR InternationalSeries 232. BritishArchaeological Reports,Oxford. Federmann,N. 1958 [1555] Historia indiana o primer viaje de Nicolas Federmann.Editado por Juan Friede. Academia Colombianade la Historia.Artes Graficas,Madrid. Flannery,K. V. 1976 EmpiricalDeterminationof Site Catchmentsin Oaxacaand Tehuacan.In The Early Mesoamerican Village,edited by K. V. Flannery,pp. 103-117. AcademicPress, New York. Flannery,K. V. (editor) 1982 PrehispanicMaya Subsistence:Studiesin Memoryof Dennis E. Puleston.AcademicPress,New York. Garson,A. G. 1980 Prehistory,Settlementand Food Productionin the SavannaRegion of La Calzadade Paez, Venezuela. UnpublishedPh.D. dissertation,Departmentof Anthropology,Yale University, New Haven, Connecticut. Gumilla, P. J. 1963 [1745] El Orinocoilustradoy definido.Bibliotecade la Academia Nacional de la Historia No. 69. Fuentespara la HistoriaColonialde Venezuela.Italgrafica,Caracas. Kirkby,A. 1973 The Use of Land and WaterResourcesin the Past and Present Valleyof Oaxaca, A>exico.Memoirs No. 5. Museumof Anthropology,University of Michigan,Ann Arbor. Kolata,A. L. 1991 The Technologyand Organizationof AgriculturalProductionin the TiwanakuState.Latin American Antiquity2:99-125. Kowalewski,S. 1980 Population-ResourceBalancesin Period I of Oaxaca,Mexico. AmericanAntiquity45:151-165. Marcos,J. G. 1987 Los campos elevados de la Cuencadel Guayas,Ecuador:E1ProyectoPenon del Rio. In Pre-Hispanic AgriculturalFields in the AndeanRegion, edited by W. M. Denevan, K. Mathewson,and G. Knapp, pp. 217-224. BAR InternationalSeries 359(ii). BritishArchaeologicalReports,Oxford. Morey, N. 1975 Ethnohistoryof the Colombianand VenezuelanLlanos. UnpublishedPh.D. dissertation,Department of Anthropology,University of Utah, Salt Lake City. Parsons,J. J. 1969 Ridged Fields in the Rio Guayas Valley, Ecuador.AmericanAntiquity34:76-80. Parsons,J. J., and W. A. Bowen 1966 Ancient Ridged Fields of the San Jorge River Floodplain, Colombia. The GeographicalReview 56: 317-343. Parsons,J. R. 1971 PrehistoricSettlementPatternsin the TexcocoRegion, Atexico.Memoirs No. 3. Museumof Anthropology, University of Michigan,Ann Arbor. 1976 The Role of ChinampaAgriculturein the Food Supplyof Aztec Tenochtitlan.In CultureChangeand edited by C. Cleland,pp. 233-257. Academic Continuity:Essays in Honor of James Bennett Gri.fgin, Press,New York.

142

LATIN AMERICANANTIQUITY

[Vol. 5, No. 2, 1994

Plazas,C., and A. M. Falchettide Saenz 1981 Asentamientosprehispanicosen el bajoRio San Jorge.Fundacionde InvestigacionesNacionales,Banco de la Republica,Bogota. 1987 Poblamientoy adecuacionhidraulicaen el bajo Rio San Jorge, Costa Atlantica,Colombia. In PreHispanicAgriculturalFields in the AndeanRegion, edited by W. M. Denevan, K. Mathewson,and G. Knapp,pp. 483-503. BAR InternationalSeries 359(ii). BritishArchaeologicalReports,Oxford. Redmond, E. M., and C. S. Spencer 1989 Investigacionesen el piedemonte andino y los llanos altos de Barinas.Boletin de la Asociacion Venezolanade Arqueologfa5:4-24. Rinaldi, M. 1990 Informede las Muestras54, 114, 442. Laboratoriode Paleoecologia,Centrode Ecologiay Ciencias Ambientales,Instituto Venezolano de InvestigacionesCientificas,Caracas.Manuscripton file, Departmentof Anthropology,AmericanMuseum of NaturalHistory, New York. Roosevelt, A. C. 1980 Parmana:PrehistoricMaize and Manioc SubsistenceAlongtheAmazonand Orinoco.AcademicPress, New York. Rouse, I. B., and J. M. Cruxent 1963 VenezuelanArchaeology.Yale University Press, New Haven. Salgadode Labouriau,M. 1979 Apendice 2: Resultadosdel analisis palinologico.In Camposelevadose historiaculturalprehispanica en los llanos occidentalesde Venezuela,by A. Zucchi and W. M. Denevan, pp. 97-99. Universidad CatolicaAndresBello, Caracas. Sanders,W. T., J. R. Parsons,and R. S. Santley 1979 The Basin of Mexico: Ecological Processesin the Evolutionof a Civilization.Academic Press, New York. Schnee, L. 1973 Plantas comunesde Venezuela.UniversidadCentralde Venezuela,Facultadde Agronomia,Institito de BotanicaAgricola,Maracay. Schwerin,K. 1966 Oil and Steel: Processesof Karinya CultureChange in Response to IndustrialDevelopment.Latin AmericanStudiesNo. 4. University of California,Los Angeles. Spencer,C. S. 1979 Irrigation,Administration,and Society in Formative Tehuacan.In PrehistoricSocial, Political, and Economic Developmentin the Area of the Tehuacan Valley,edited by R. D. Drennan,pp. 13-109. TechnicalReportsNo. 11. University of MichiganMuseum of Anthropology,Ann Arbor. 1987 Rethinkingthe Chiefdom.In Chiefdomsin the Americas,edited by R. D. Drennanand C. A. Uribe, pp. 369-390. University Press of America,Lanham,Maryland. 1990 On the Tempo and Mode of State Formation.Journal of AnthropologicalArchaeology9:1-30. 1991 Coevolutionand the Developmentof VenezuelanChiefdoms.In Profilesin CulturalEvolution:Papers from a Conferencein Honor of Elman R. Service,edited by A. T. Rambo and K. Gillogly, pp. 137165. AnthropologicalPapersNo. 85. Museumof Anthropology,University of Michigan,Ann Arbor. Spencer,C. S., and E. M. Redmond 1992 PrehispanicChiefdomsof the WesternVenezuelanLlanos. WorldArchaeology24:13v157. Turner,B. L. II 1974 PrehistoricIntensive Agriculturein the Maya Lowlands.Science 185:118-124. Turner,B. L. II, and P. D. Harrison 1982 PulltrouserSwamp:A PreliminaryStudyof Habitat,Land Use,and Settlementof the PrehistoricMaya. University of Texas Press, Austin. Wagner,E. 1973 The MucuchiesPhase:An Extensionofthe AndeanCulturalPatterninto WesternVenezuela.American Anthropologist75:195-213. Wagner,E., and C. Schubert 1972 Pre-HispanicWorkshopof SerpentiniteArtifacts,VenezuelanAndes,and PossibleRawMaterialSource. Science 175:888-890. Whitehead,N. 1988 Lords of the Tiger Spirit:A History of the Caribs in Colonial Venezuelaand Guyana, 1498-1820. Dordrecht,Netherlands. Zucchi,A. 1967 La Betania: Un yacimiento ar4ueologicodel occidentede Venezuela.2 tomos. Unpublished Ph.D. dissertation,Departamentode Antropologia,UniversidadCentralde Venezuela,Caracas. 1972 AboriginalEarthStructuresof the WesternVenezuelanLlanos. CaribbeanJournalof Science 12(1-2): 95-106. 1973 PrehistoricHuman Occupationsof the WesternVenezuelanLlanos.AmericanAnti4uity38:182-190. 1976 Cano Caronl: Un grupoprehispanicode la selva de los llanos de Barinas. Universidad Centralde Venezuela,Caracas.

Spencer et al.]

DRAINED FIELDS AT LA TIGRA

143

1985 Recent Evidencefor Pre-ColumbianWaterManagementSystemsin the WesternLlanosof Venezuela. In PrehistoricIntensiveAgriculturein the Tropics,edited by I. S. Farrington,pp. 167-180. BAR InternationalSeries 232. BritishArchaeologicalReports,Oxford. Zucchi,A., and W. Denevan 1979 Camposelevadose historiaculturalprehispanicaen los llanos occidentalesde Venezuela.Universidad CatolicaAndres Bello, Institutode InvestigacionesHistoricas,Caracas.

ReceivedMarch 11, 1993; acceptedNovember20, 1993

View publication stats