Heckenberger et al 1999

VILLAGE SIZE AND PERMANENCE IN AMAZONIA: TWO ARCHAEOLOGICAL EXAMPLES FROM BRAZIL

Michael J. Heckenberger, James B. Petersen, and Eduardo GoCs Neves

~-

Recent archaeological investigations along the lower Negro and upper Xingu Rivers in the Brazilian Amazon provide important new evidence bearing on long-standing debates about the size and permanence of Amerindian settlements in the region. Preliminary regional surveys and more in-depth study of selected large (30-50 ha) sites, particularly analyses of the associations between structural features, anthropogenically altered soils, and artifact distributions, lead us to conclude that large, permanent settlements, likely associated with fairly dense regionalpopulations, existedprehisrorically in both areas. Thesefindings cast doubt on the view that environmental limitations prevented sedentism and demographic growth among Amerindian populations throughout much or all of the region. Specifically, we conclude that fully sedentary and relatively large populations emerged in a r.arietj of Amazonian settings prehistorically, nor necessarily correlafed with the distriblrfion of one or another narrowly defined ecological variable (e.g., high fertiliiy soils). Thus, a critical evaluation of core concepts in Amazonian anthropology, such as the varzealterrafirme dichotomy or tropical forest culture, is advised.

Investigaciones arqueoldgicas recientes a lo largo de 10s rios bajo Negro y alto Xingli, en la Amazonia brasiletia, proporcionan nuevas e importantes evidencias para 10s prolongados debates acerca del tamatio y la permanencia de 10s asentamientos amerindios en la regidn. Relevamientos preliminares realizados en la regidn y estudios mas profundos de sitios de 30 a 50 hectcireas, parri- cularmente un analisis de la asociacidn entre rasgos estructurales, suelos antropoge'nicamente alterados y distribuciones de artefatos, nos llevan a corrcluir que estos asentamientos de gmn tamatio y permanenres, probablemente asociados a poblaciones regionales bastante densas, existieron prehistdricamente en ambas areas. estos hallazgos plantean serias dudas a la visidn, sostenida por mucho tiempo, de que limitaciones medioambientales habrian impedido el sedenrarismo y el crecimiento demogrcifico entre /as poblaciones amerindias de la regidn. Concluimos, espec@camente, que poblaciones plenamente sedentarias y relativamente amplias emergieron prehistdricamente en una variedad de escenarios amazdnicos, no necesariamente correlacionadas con la dis- rribucidn de ninguna variable ecoldgica estrechamentc dejinida. Por consiguiente, se requiere una evaluacidn cri'tica de ciertos conceptos incorporados en la antropologia amazona tales como la dicotomia valleltierra firme o la cultura de la jloresra tropical.

Amerindian settlement patterns in Amazonia historic variability in Amerindian settlement patterns have been a source of considerable debate, in the region (e.g., Meggers 1996, 1997). This view particularly focusing on the general causes conforms to initial definitions of the "tropical forest

of increasing sedentism and regional population culture" (Lowie 1948; Meggers and Evans growth or the lack thereof. Diverse viewpoints have 1957: 17-32; Steward 1948, 1949), or what Viveiros been expressed, but a gross distinction can be made de Castro (1996:180) has aptly called the "standard between those arguments that suggest large, fully model." The alternative "revisionist" position, which sedentary populations rarely, if ever, developed has gained increasing currency over the past several indigenously in Amazonia and those that argue such decades, expands on Steward's (1948:886,1949:698) populations existed but were largely restricted to suggestion that ecological differences between flood- select riverine areas. Proponents of the former view plain (~~drzea) and upland (terrajrme) areas were believe that the generally small, impermanent, and the predominant factor in Amazonian cultural evo- relatively dispersed settlements characteristic of lution. Protagonists of the revisionist "~ldrzea model" ethnographic communities adequately represent pre- suggest that large, fully sedentary villages were not

Michael J. Heckenberger .Department of Anthropology, Turlington Hall, University of Florida, Gainesville, FL 3261 1 James B. Petersen .Department of Anthropology, Williams Hall, University of Vermont, Burlington, VT 05405 Eduardo Gobs Neves .Museu de Arqueologia e Etnologia, Universidade de Slo Paulo, Avenida Prof. Almeida Prado, No. 1466, S lo Paulo 05508-900, Brazil

Latin American Antiquity, 10(4), 1999, pp. 353-376 Copyright O 1999 by the Society for American Archaeology

354 LATIN AMERICAN ANTIQUITY [Vol. 10, No. 4, 1999

only present but densely clustered in some riverine areas by 1492, with a possible time depth of several millennia (e.g., Beckerman 1979, 1991; Brochado and Lathrap 1982; Carneiro 1970, 1995; DeBoer 1981; Denevan 1976, 1996; Gross 1983; Lathrap 1970a, 1977; Lathrapet al. 1985; Myers 1973,1992; Porro 1993, 1996; Roosevelt 1980, 1989, 1994; Whitehead 1994, 1996).

Among archaeologists working in the tropical lowlands, Betty Meggers was the first to propose and remains the main advocate of the standard model. Based on ethnographic examples, such as the Mekragnoti Kayap6, the Yanomamo, and groups from the Guiana highlands, among others, Meggers has repeatedly espoused a general model for Ama- zoniain which settlements are portrayed as uniformly small (averaging 50 to 150 people), impermanent, dispersed, and politically autonomous (e.g., Meggers 1954, 1994, 1995a, 1995b, 1995c, 1996; Meggers and Evans 1957). Using selected archaeological data from the Programa Nacional de Pesquisas Arque- ol6gicas na Bacia AmazBnica (PRONAPABA), she states that there is "no evidence that terrafirme com-munities were larger, more closely spaced or more sedentary in Precolumbian times than indigenous communities today" (Meggers 1992: 199). Moreover, this basic pattern has "existed for at least two millennia without significant alteration in village size or permanence," an equilibrium that she feels results from unique human adaptations to severe environmental constraints (Meggers et al. 1988:291; see also Meggers 1994:416, 1995c:69). The vcirzea habitats, narrowly defined as the floodplain and adjacent riverbluff areas of the Amazon and its major Andean- derived tributaries (i.e., the so-called "white-water" rivers), are viewed as slightly more propitious for sedentism and demographic growth than other Ama- zonian environments. Nevertheless, population density is considered to be comparable between the two areas (roughly 0.3/krn2) and, in general, settlement patterns are "not significantly different from [those] reconstructed along clear and black-water tributaries of the terra firme" (Meggers 1992:201-203).

Increased recognition of early ethnohistoric accounts of large, complex societies and identification of extensive anthropogenic "black earth" (terra preta do indio, hereafter TP)' sites along the Ama- zon River led revisionist scholars to focus on appar- ent differences between these relatively dense riverine societies, largely decimated within a century

of initial European contact (ca. A.D. 1550), and the typically small and autonomous villages known ethnographically from upland areas. Proponents of the vcirzea model suggest that higher levels of productivity, in terms of agriculture and/or faunal capture, provided the economic basis for sedentism and population growth in these riverine areas. In contrast, low soil fertility and scarce and dispersed game of the terra jirme are seen to limit productivity and, hence, population growth (e.g., Brochado and Lath- rap 1982; Brochado 1984; Cameiro 1986, 1995; Denevan 1984, 1996; Gross 1975; Lathrap 1968, 1970a; Lathrap et al. 1985; Meggers 1995b, 1996; Roosevelt 1980, 1991, 1994). (There is disagree- ment on the distribution of the riverine, or "vdrzea," pattern, as described in the historic chronicles, but most authors emphasize rich floodplains, thus restricting the term to Andean-derived rivers, since most black- and clear-water rivers lack nutrient-rich floodplains and are commonly included in the generic category terrafirme, along with interfluvial areas).2 Thus, most revisionists agree that sedentary populations emerged early in the Amazonian vcirzea and developed into regionally organized, hierarchical societies (chiefdoms) in many cases, but still view the standard model from ethnography as generally applicable to broadly defined terra jirme settings, before and after European contact.

Detailed archaeological field studies and research methodologies designed to identify regional-level settlement patterns in Amazonia are still largely lack- ing. Even basic aspects of chronology, regional settlement patterns, and the characteristics of individual sites and site components (e.g., size, duration, and internal variability) are poorly understood and minimally reported from most areas. Models of Pre- columbian occupations therefore typically are hypothetical, based largely on nonarchaeological data (ethnographic, ethnohistoric, and/or ecological) and untested theoretical assumptions. In other words, various models have been proposed for what prehistoric occupations should or may have been like, based on the expectations of general models, but few relevant archaeological studies directly bear on what they were, in fact, like, in terms of demonstrable archaeological patterning. DeBoer et al. (1 996) drew attention to a clear example of this problem, the dis- juncture between empirical evidence and hypothetical models, in an incisive critique of Meggers' reconstructions of prehistoric settlement behavior in

355 Heckenberger, Petersen, and Neves] VILLAGE SIZE IN AMAZONIA

Amazonia based on analysis of seriated ceramic sequences (e.g., Meggers 1986,1990,1991; see also Machado 1993: 107-1 12; Raymond 1995:225-226, for recent critiques of Meggers' seriation technique). Specifically, they concluded that the evidence provided by Meggers (1991) does not justify her interpretation that large TP sites in Amazonia represent palimpsests of numerous successive and only partially overlapping reoccupations, as opposed to extensive contemporaneous occupations and long- term settlement permanence.

Meggers' reconstructions are particularly significant since they represent one of the few attempts to infer prehistoric settlement patterns based on archaeological data from broadly defined terrafirme settings in Amazonia. Nonetheless, her conclusions seem to stem from two unwarranted assumptions: 1) TP is composed of relatively undifferentiated deposits that largely lack stratigraphy or other significant internal structure, regardless of their areal extent or depth; and 2) this uniformity is expectable for partially overlapping, sequential occupations by small communities. Assessment of prehistoric settlement patterns requires the development of appro- priate methods of identification, analysis, and interpretation of spatial patterning, such as the distribution of cultural remains across sites or settlement distributions within discrete regions (see Roosevelt 1991). Minimally, such methods must include careful examination of the actual archaeological contexts (i.e., the configuration of cultural deposits and their internal associations within natural depositional matrices), from which the raw materials for seriation and other analyses are derived (see Brochado and Lathrap 1982:20-21; DeBoer et al. 1996; Lathrap 1970b; Lathrap et al. 1985; Roo- sevelt 1991: 115-1 16). Given the very limited testing (one or a few small excavations and uncontrolled surface collection) and minimal attention paid to archaeological context in the studies employed by Meggers, we concur with DeBoer et al. (1996) that there is no sound empirical basis to support her con- tention that the standard model can be extended into prehistory (e.g., Meggers 1990, 1991; Meggers et al. 1988). Not only has Meggers failed to provide a compelling archaeological case for the standard model, but there is clear evidence that large and permanent settlements were present in both vcirzea and terra firme areas of Amazonia in the prehistoric period.

In the present study, prehistoric settlement size and permanence are discussed in relation to large TP sites in two separate areas of Arnazonia: 1) the "black- water" setting of the lower Negro River in the central Amazon and 2) the headwater region of the Xingu River, a "clear-water" southern tributary of the Ama- zon (Figure 1). The results of preliminary regional surveys and more detailed analysis of the large (30+ ha) Aqutuba site, lower Negro River, and three large sites (30-50 ha) in the upper Xingu, Nokugu, Kuhikugu, and Hialugihifi, including surface artifact and TPdistributions, site stratigraphy, subsurface sediment chemistry, visible structural features, and overall site plans, provide important new evidence incompatible with the standard model. Specifically, if multiple partially or non-overlapping occupations by small mobile or semi-sedentary groups (100 to 300 people) cumulatively produced extensive anthropogenic deposits, as Meggers suggests, we could expect relatively homogeneous (vertically and hori- zontally) or patchy distributions of TP, interspersed with areas of little or no sediment alteration. TP deposits are neither patchy nor homogeneous at AFU- tuba or the upper Xingu sites, but instead form broad contiguous strata with considerable variability in depth and composition within and between sites and within individual excavation units, due to differential deposition, intensity of use, functional variability, and, in some cases, significant landscape modifica- tion. Likewise, TP generally lacks intervening non- cultural sediments (paleosols) or other evidence of abandonment. In short, small-scale occupations, as inferred by the standard model, simply do not create the areally extensive alterations characteristic of many prehistoric occupations in both regions (see Woods and McCann 1998), an assertion amply supported in our studies by the lack of extensive TPdeposits related to small-scale occupations by contemporary neo- Brazilian (caboclo) hoe-farmers at Aqutuba and recent Amerindian villages in the upper Xingu.

Our findings also document that certain areas of these sites, namely central site areas, were used far more intensively and over a much longer span of time than other, more peripheral occupation areas. Stratig- raphy and associated radiocarbon dates indicate that TP, most notably in these central site areas, are accretional deposits formed by gradual deposition and sediment enrichment related to relatively continuous human occupation. The substantial structural elaboration at each of these sites, including the construc-

LATIN AMERICAN ANTIQUITY [Vol. 10, No. 4, 1999

Figure 1. Map of Amazonia showing location of the lower Negro and upper Xingu study areas (the parkland and savanna vegetation, or cerrado, of central Brazil is marked by cross-hatching).

tion of central plazas, earthworks, and specialized midden deposits, documents the type of landscape alteration and functional variability that would be expected for large, sedentary occupations, but highly unlikely for small, impermanent communities. Rather than suggesting broadly similar, repeated occupations across these sites, they indicate long-term, relatively continuous occupation of central site areas, continuity in overall site layout (with occupations gravitating toward a central plaza), and settlement growth outward from core areas. Based on these fmd- ings, we conclude that large sedentary communities, likely associated with relatively high regional population densities, were present prehistorically in both the lower Negro and upper Xingu regions.

The Agutuba Locality

Recent archaeological investigations along the lower Negro River provide a rather different picture of pre-

historic settlement patterns than that offered by the standard model. Preliminary regional survey in the lower Negro study area in 1995 and 1997, involving initial assessment of site size through visible surface distributions of artifacts and TP, sediment coring, and occasional test-pitting, has identified 19 prehistoric archaeological sites (Neves and Bartone 1998);~ an additional 12 sites were previously identified in this area by Hilbert (1968) and Simdes (1970,1974a, 1974b, 1983; Simdes and Kalkman 1987)~ (Table 1, Figure 2). These surveys document substantial settlement variability: small and medium-sized sites (roughly 2 to 10ha of TP) are located in diverse ecological settings, including major rivers, side chan- nels, small to medium streams (igarapks), and upland (terra$rme) lakes, but the largest sites (some with over 30 ha of TP) are apparently most common on river bluffs along the Solimdes and Negro Rivers.

A more detailed investigation was conducted at


Table 1. Uncalibrated Radiocarbon Dates from the A~utuba Site and Other Sites in the Middle-Lower Negro River Region.

Site/Provenience Lab. # I4cAge B.P. dI3c Material Dated

Acutuba IA Modeled-Incised: Unit 2, 30 to 40 cm Beta 97527 1030 f 100 -25.0% wood charcoal Unit 2, 70 to 80 cm Beta 90724 1800 f 80 -25.0% wood charcoal Aqutuba IIB Initial site occupation (?): Unit 1, 90 to 100 cm Beta 90721 6850 f 100 -25.0% wood charcoal

Modeled-Incised: S231W859 130 to 140 cm Unit 1, 30 to 40 cm Unit 1, 40 to 50 cm Unit 1, 50 to 60 cma Unit 1 ,60 to 70 cm

Beta 109182 Beta 90723 Beta 106437 Beta 106438 Beta 97528

1100 +-60 1270 * 60 1230 * 70 1590 * 60 23 10 * 140

-25.0% -25.0% -25.0% -25.0% -25.0%

wood charcoal wood charcoal wood charcoal wood charcoal wood charcoal

S250W500 130 to 140 cm Beta 109179 1830 f80 -25.0% wood charcoal

Guarita: Surface (Guarita polychrome) Beta 90009 Caraipk Temper Surface (Guarita "grooved") Beta 97529 caraipk temper S23 1 W859 40 to 50 cmb Beta 109180 wood charcoal Surface (finger-punctate) Beta 97530 cauixi temper S350W970 60 to 70 cm Beta 109 184 wood charcoal S250W500 40 to 50 cm Beta 109178 wood charcoal S23 1 W859 70 to 80 cmb Beta 109181 wood charcoal S230W889 40 to 45 cm Beta 109183 wood charcoal

Related Sites Santa Rosa (lower N e g r ~ ) ~ SI 275 1 Pared30 (lower ~ e g r o ) ~ GrN 4329 Pared30 (lower GrN 4330 Engheno Velho (middle Negro)' SI 4055 Nova Vida (middle Negro)' SI 4054 Unini River (middle Negro) Beta 106436 Vista Alegre (middle N e g r ~ ) ~ SI 4053 Terra Preta (lower N e ~ r o ) ~ - SI 2752

SI 4052 390 k 95 a Level 50-60 was redated since the first radiocarbon date (2910 f70 B.P., or 960 B.C., Beta 90722) was out of sequence and because the dated sample may have been mixed (or switched) with another in the lab

The weighted average of these two dates (Beta 109180 and 109181) is 615 * 57 B.P. (A.D. 1335). 'SimBes and Kalkman 1987.

Hilbert 1968.

one such large site, A~utuba, on the lower Negro, Excavation of 18test pits (.5 m2) and 5 larger units actually a series of related habitation areas (desig- ( .5 x 1m, 1m2 and 1x 1.5m) were conducted in cen-nated Aqutuba I, 11, and III) located on a long linear tral areas of Aqutuba I, 11, and numerous sedi- riverbluff terrace, variably incised by drainage gul- ment cores also were taken to help determine the extent lies (Figure 3). Field research at Aqutuba (conducted and depth of the TP (Figure 4). In composite, the site over some 180 person-days from 1995 to 1997) extends at least 3 km along the Negro River, and pri- focused on mapping, surface collection of select mary habitation areas minimally extend over an area areas, and subsurface excavations. The horizontal of 30 ha, based on the clearly observable distribution extent of TP deposits and variations in surface soil of TP. The extent of cultural activities is larger (50+ color and artifact distributions were readily appar- ha), however, as documented by an artificial ditch and ent in most areas due to broad exposure by mecha- Amerindian ceramics located well beyond the TP nized and hoe agriculture (disturbing the upper 20 deposits at A~utuba II. The overall occupation area to 40 cm of the cultural stratigraphy in most areas). may extend farther upstream and downstream to adja-


Figure 2. Location of sites identified by the authors in the lower Negro study area, Amazonas, Brazil; previously iden-tified sites are marked with small blackened circles (Note: Site sizes are approximated, dashed lines mark edge of high ground, and short horizontal lines mark seasonally inundated areas within the study area).

NEGRO R/VER .

LOW u

A C U T U B A III KEY:

aEXTENT OF TERRA PRETA

TEST PIT

0 EXCAVATION UNIT

0 .,"OOM

SCALE

Figure 3.Map of the Afutuba locality (AM-IR-02) showing approximate distribution of primary terra preta deposits and 1995 excavations at Afutuba I, 11, and 111 (Note: 35-m contour roughly defines riverbluff).

Heckenberger, Petersen, and Neves] VILLAGE SIZE IN AMAZONIA 359

TEST PIT

EXCAVATION U N I T- MODERN ROAD

-.-.- CONCENTRATED DISTRIBUTION OF FINE-WARE

100 METERS

-**- RECONSTRUCTED DITCH LOCATION

Figure 4. Map of Aqutuba I1 showing location of 1995 and 1997 excavation units and test pits, trash middens, proba-ble central plaza, and primary surface distribution of Guarita "fine-ware."

cent lobes of the terrace, but, if so, these were appar- 150 * 80 (70 to 80 cm) and A.D. 920 k 100 (30 to ently used much less intensively,for considerablyless 40 cm) minimally bracket the intact cultural deposits time, and/or for nondomestic activities (e.g., agricul- in this area (Table 1,Figure 5). Amerindian cultural turd fields), as indicated by light-or non-TP sediments remains, including TP, also were encountered across and sparse surface artifacts. A~utubaIIIA.

Stratigraphy and Distributions

The depth and color of the TP varies dramatically across the Aptuba site due to the variable intensity, duration, and nature of cultural activities, as well as to differential natural deposition, largely related to eolian transport from the extensive dry-season beaches of the Negro River. Three test pits (.5 m2) and one larger (1 m2) unit were excavated at ACU-tuba I. These investigations served to document 30 to 80 cm of TP deposits. Radiocarbon dates of A.D.

Eight test pits and four larger units (one .5 x 1m, two 1m2,and one 1 x 1.5 m units) were excavated at A~utubaI1 (subdivided into IIA,IIB, and IIC based on topography in relation to a low central area; see below) (see Figure 5). Cultural remains extend well beyond the area of primary TP at A~utuba11, as evi-denced by very light- or non-TP soils, artifact distri-butions,andthe presenceof artificialearthworks (e.g., excavated ditches) in one or more peripheral areas. Likewise, earlier,presumably cultural, deposits were encountered in the less altered (light TP) sediments

-#


ACUTUBA I U N l T 2

SOUTH WALL WEST WALL

m m

:

ACUTUBA It U N l T I

S O U T H W A L L W E S T W A L L

K E Y :

TZ7 C E R A M I C SHERD

Figure 5. Profile of south and west walls of Aqutuba I Excavation Unit 2 (top) and Afutuba I1 Excavation Unit 1 (bot-tom). Munsell color descriptions of strata include: (Afutuba I, Unit 2) I =10 YR 518 (yellowish brown); 1/11 = 10 YR 4- 514 (dark yellowish brown); I1 =10 YR 313 (dark brown); I11 =10YR 313; (Aqutuba 11, Unit 1) I =I0YR 5-618 (yellowish brown); 1/11 = 10 YR 518; I1 = 10 YR 3-413 (dark brown); III=lO YR 212 (very dark brown); IV =10 YR 212 (very dark brown). Note: Sediment texture is fine to medium sand with small quantities of silt; Stratum I in both units apparently predates cultural deposits, Stratum I11 at Afutuba I apparently postdates cultural deposits and is plow-disturbed, and Stratum IV at Afutuba I1 is plow-disturbed.


Table 2. Sediment Data from 1995 Excavation Units at Aqutuba I and 11.

% Org./Total Mehlich Ca P K Mg Zn Mn Cu P f r a ~ t . ~ carbon ppm-P ppm ppm ppm ppm ppm ppm ppm sum mgkg

A ~ u t u b a IAIUnit 2 0 : depth (cm) 3 0 - 4 0 ~ ~ ,961.73 3 0 - 4 0 ~ ~ ,911.85 40-50 1.051.70 5 0 4 0 ,691.65 60-70 .61/.62 70-80 ,511.47 80-90 ,321.37

Aptuba IIBIUnit IA: depth (cm) 20-30 2.2311.81 619.2 3212 2281 301.3 227.8 58.2 294.4 16.4 3 W O 1.9511.61 646.9 3176 2035 182.3 208.4 48.8 226.7 15.7 4 0 - 5 0 ~ ~ 1.5411.35 674.3 2626 1819 160.1 177.1 44.7 177.8 13.8 1327.2 4 0 - 5 0 ~ ~ 1.2211.16 379.2 1890 1406 123.3 137.9 42.8 158.7 12.5 5 0 4 0 ,871.94 216.4 1078 948 73.4 93.7 30.2 105.3 9.2 60-70 ,671.94 165.3 767 89 1 56.1 68.5 30.9 81.2 8.1 601.3 70-80 ,651.80 123.6 645 817 75.1 59.1 21.6 59.0 5.7 80-90 S11.82 105.3 502 776 60.9 56.4 20.3 54.8 5.1 90-100 ,571.93 111.8 428 758 16.9 46.4 18.9 72.7 4.6 560.5 100-1 10 ,961.72 136.3 404 820 101.7 47.4 21.2 72.2 5.0 Note: Data in Table 2 are a summary of sediment analyses being conducted under the direction of William Woods (Southern Illinois University-Edwardsville) and they represent total values. The upper levels (0-20) were clearly disturbed by recent farm- ing activities and were excluded from the preliminary analysis. A~utuba IA pH values range from 4.44 to 4.88 (m=4.73) and Aqutuba IIB values range from 5.53 to 6.17 (mean=5.98) (with values measured for each of the samples described in Table 2). a Phosphate fractionation (P fract.) results presented above (sum of fractions) are summarized from analyses conducted by Soils and Physical Geography Laboratory, University of Wisconsin-Milwaukee b~ denotes dark terra preta and B represents lighter terra preta.

underlying the dark TP in several excavation units; 1996; Moraet al. 1991; Pabst 1991; Smith 1980556; in fact, possibly cultural non-TP deposits lying imme- Woods and McCann 1998; cf. Petersen 1996), par- diately abov2 basal stratum I (a compact layer of lat- ticularly associated with the upper dark TP of AFU- erite pebbles and cobbles) in Unit 1 at A~utubaII may tuba 11. The radiocarbon date (A.D. 150 k 80) from date as early as 4900 f 100 B.c .~ near basal cultural deposits (very light TP) in Unit 2

The limited amount of subsurface testing con- at Aptuba I (the deepest TP identified at Aqutuba I) ducted thus far precludes full assessment of the over- indicates that this portion of the site was likely occu- all variability of site deposits at A ~ u t u b a . pied later and, based on the lighter soils and lower Nonetheless, excavations, extensive surface collec- sediment chemical values and artifact densities, less tions, and sediment coring conducted to date do per- intensively than Aqutuba 11. mit several general observations. In almost all That such extensive anthropogenic alteration as excavation units, there is a clear vertical increase in documented at A~utuba is fairly typical of perma- artifact frequency and soil alteration (darkening and nent occupation but unlikely for short-term occupa- chemical alteration) toward the surface (Tables 2 and tions, as suggested elsewhere (Eidt 1984; Woods and 3).Dark TP (Munsell lOYR 2-311 -3) generally dates McCann 1998), is supported by visible sediment to after ca. A.D. 800 to 1000 in radiocarbon-dated alterations associated with small households of neo- units, but includes older deposits in Unit 1 at AFU- Brazilian (caboclo) hoe-farmers at A~utuba I. These tuba 11. Analysis of sediment chemistry from this unit permanent occupations have failed to produce sig- and Unit 2 at A~utuba I documents the high degree nificant, areally extensive deposits of TP even after of human-influenced alteration, relative to both nat- more than 50 years of continuous residence. They ural (latosol) sediments and other TP deposits in instead prompt localized sediment alteration through Amazonia (e.g., Eden et al. 1984:131- 133; Kern concentrated refuse disposal or domestic activities


W E S T W A L L N O R T H WALL

K E Y :

Lcl C E R A M I C S H E R D

a R O D E N T B U R R O W

e R O C K

Figure 6. Profile of north and west walls of Excavation Unit 9230-231W857.5-859. Strata descriptions include: I=10 YR 518 (yellow brown); II=10 YR 518; III= 10 YR 211-2 (very dark brown to black) grading to 10 YR 311-2 (very dark gray to black); IV=lO YR 211-2; V = 10 YR 211-2. Note: Sediment texture is fine to medium sand with limited amounts of silt.

(see below for similar distributions among Amerindian communities in the upper Xingu).

The distribution of ceramic remains, pertaining to two distinctive macro-ceramic complexes defined at the site: also is revealing insofar as it indicates the broad contemporaneity and significant functional variability of late prehistoric occupations across the site. Ceramics pertaining to the earlier complex, roughly correlated with the so-called "Modeled- Incised" (Barrancoid) or "Incised-Rim" traditions, and more locally to the Itacoatiara and Manacapuni phases (Hilbert 1968; Lathrap 1970a, 1970b; Meg- gers and Evans 1961, 1983)', are radiocarbon dated from ca. 360 B.C. to A.D. 850 at Agutuba 11, and ca. A.D. 150 to A.D. 920 at A~utuba I. These ceramics are commonly found in large quantities in subsurface contexts, but are rare on the site surface (see Table 3). On the other hand, diagnostic ceramics related to the later ceramic complex, associated with

the Guarita phase of the Amazonian Polychrome tra- dition (Hilbert 1968) and radiocarbon dated from ca. A.D. 980 t0A.D. 1440 at Agutuba 11, are widely distributed across the entire site surface (A~utuba I, 11, and III) .~ The most elaborately decorated styles, "Guarita fineware" (notably including red- and/or black-on-white painted, wide-line incised ["grooved"], flanged, and complicated modeled designs), are concentrated in central portions o f A ~ u - tuba I1 in close association with a probable central "plaza" and several prominent Guarita-age mounded middens. lo

Definition of a Central Plaza

The immediate impression of the A~utuba locality is one of occupations gravitating toward a fixed center prominently marked by a rectangular, low-lying area in the middle of A~utuba I1 (see Figure 4). We tentatively interpret this hollow as a central public


Table 3. Ceramic Remains from Select Excavation Units at Aptuba I and 11.

Acutuba I, Unit 2 Acutuba 11, Unit 1 Acutuba 11. S231W858-859a

Depth countweiRhtb (Total ~ims/MvC)' Countweight (Total Rims/MVC) Countweight (Total Rims) 0-10 501125 (414) 62511,641 (22122) 324018369 (168) 10-20 4011 62 (212) 40311,731 (18118) 3956113,807 (233) 20-30 1061463 (1 1/10) 1,12319,592 (73164) 3027/10,717 (178)

3 W O 89019,387 (76167) 3,654119,360 (84180) 242918443 (1 19) 40-50 1,04714,398 (35128) 1,957114,584 (73PO) 253211 1,898 (173) 5 M O 2881960 (18118) 60014,785 (39138) 163818754 (150) 6G70 12816 10 (17115) 1721983~ (13112) 220011 1,090 (156) 7G80 441199 (414) 16512.1 lod (13111) 144115450 (78) 80-90 17/36 (111) 46/560d (111) 85813094 (58) 90-100 10129 (414) 122315771 (77) 10G110 111 2132P541 (114) 11G120 186717171 (95) 120-130 55011951 (35) 130-140 7151568 1 (30) 140-150 2151420 (5) 15G160 1218 walle 12211,346 37114,370 297819623

Totals 2681117,391 9,126159,746 31,01311 19,787.2 a S23 1W858-859 denotes coordinates of 2-x-1-m unit in site grid established in 1997 (only 1 x 1.5 was excavated; other two units are 1 m2 units designated using 1995 system %eight is rounded to nearest gram

Denotes the total number of rim sherds over the minimal vessel count (MVC), which represents the subset of rims that clearly ertain to separate vessels (MVC for 1997 1-x-1.5 unit not available);

'Ceramic counts include clearly intrusive ceramics from a rodent burrow in Subunit 1C and are therefore artificially high (Subunit 1C constitutes 61 percent, 92 percent, and 72 percent of the total in the three levels, respectively); by averaging the number of sherds in the other three subunits we can calculate the following, likely more realistic estimates: 60 to 70 cm = 891381.6; 70 to 80 cm=15/47.6; 80 to 90 cm=26/148.4 eWall refers to sherds recovered when walls were cleaned (brought to true vertical) at various stages of excavation

(nondomestic) area or "plaza." It forms an obvious 8 to 10 m above the plaza floor, densely concentrated sunken amphitheater-like feature, characterized by ceramics (the highest frequencies of ceramics and, two roughly parallel, steep-sided (eastern and west- particularly, Guarita fineware, from across the site) em) banks and a more gradually sloping bank at the and extremely dark TP, associated with three large northern end. The symmetry and uniformity of this mounded middens, further indicate that this central central area indicates that it may represent an artifi- area was functionally distinctive. Although impre- cially reworked natural drainage feature. Shallow TP cisely defined, these oblong mounds each extend deposits, largely disturbed by plowing, were noted across areas of 1500 to 2000 m2 or more. Excava- in one test pit and numerous sediment cores. Surface tions in and around the easternmost mound document artifacts are likewise present, particularly in extreme that it is about 50 to 60 x 30 to 35 m in areal extent, northern areas of the plaza. It seems likely, however, about 1 m thick, and associated with late prehistoric that the plaza area was originally characterized by Guarita occupations, based on diagnostic ceramics, low artifact densities and limited TP development radiocarbon dates, and the recovery of an earlier bur- since current distributions at least partially reflect ial urn. The urn, which was encountered 110 to 150 erosion from the higher flanking areas, particularly cm below the surface, was associated with a radio- due to north to south drainage and recent plowing. carbon date of A.D. 850 rt 60. Excavations and/or An excavated ditch (1 10 m in length, 5 to 6 m across, surface collections around the other mounds suggest and 1 to 2 m in depth) presumably marks the south- that they also relate to Guarita occupations. No other em boundary of the plaza floor, which thus measures prominent mounded deposits have been encountered at least 450 x 100 m in extent." elsewhere at A~utuba.

On the high terrace areas flanking the north- It seems extremely unlikely that it is only by coin- western margins of the presumed plaza, lying some cidence that the large trash middens, the highest arti-


fact concentrations and most pronounced TP, and the excavated earthworks occur in close association with the plaza. These features were obviously interrelated and together comprise a settlement layout organized around a definitive center, the plaza, which formed the futed and highly constructed core of the local built environment. We argue that the large mounded middens overlooking the plaza area were possibly associated with special activities linked to communal ritual and public affairs concentrated near the village center and plaza. This is indicated not only by the middens' central location, flanking the plaza, but also by the frequency of very large sherds, multiple sherds from single vessels (including nearly whole vessels), and ritually "killed vessels encountered in excavations (particularly S231W857-858), which distinguishes these features from the TP deposits in other parts of the site. This reconstruction also would explain the notable concentration of finely decorated non-utilitarian ceramics (Guarita fineware), likely manufactured for ceremonial or otherwise restricted use, around the plaza. It follows that social groups occupying areas adjacent to the plaza likely had greater control or involvement in activities conducted in these core areas than those living in more peripheral areas. In short, we propose that the central area represents a locus communis of public affairs, which was apolitico-religious center of community life and perhaps dominated by limited segments of society.

The structural elaboration of the site, both intentional and unintentional, and patterned variability in stratigraphy, sediment chemistry, and artifact distributions, particularly insofar as they relate to a fixed center or plaza, are incompatible with the view that such large sites represent homogeneous deposits created by multiple, randomly placed reoccupations of the same locale by small communities over time (e.g., Meggers 1990, 1991). The massive size, restricted distribution, and unique ceramic remains of the mounded middens indicate that, although composed of refuse, they are not the product of general- ized refuse disposal by small households over a short period of time, nor likely even the aggregate of such occupations. Although site abandonment and reoccupation undoubtedly occurred over the long sequence of occupation (perhaps spanning 7,000 years), we suggest that relatively fixed occupations were established in central portions of the site (central Agutuba 11) by A.D. 1, if not before. About this time, occupations may have expanded to adjacent

portions of the long terrace, at least Agutuba I,based on the radiocarbon date 0fA.D. 150 from near basal deposits on Unit 2 at Agutuba I. By late prehistoric times (after A.D. 1000), central areas of the settlement at Agutuba I1 became the focus of concentrated and perhaps functionally specialized residential activities, although contemporaneous, if less dense, occupations also occurred in more peripheral areas of A~utuba 11, as well as across A~utuba I and 111.

In sum, we suggest that Agutuba was occupied almost continuously, if not continuously, throughout the Christian era, if not before, and, although we cannot precisely estimate the overall extent of contemporaneous occupations, it was likely the home of a large, sedentary population in the centuries immediately preceding European contact. Stretch- ing over at least 3 krn along the Negro River, it appears to be more akin to the large, complicated settlements reported from the middle Amazon in the six- teenth and seventeenth centuries (e.g., Porro 1993, 1996) than the general semi-sedentary pattern suggested from ethnography and predicted by the standard model. Regional survey has documented several other extensive (30 ha or more of TP), generally contemporaneous late prehistoric sites (i.e., containing diagnostic Guarita ceramics) within a 30-km radius of Agutuba, providing further evidence of a high regional population density.

The Upper Xingu

In the headwater region of the Xingu River, a major southern (clear-water) tributary of the Amazon, an even clearer example of large village size and permanence has been documented. The upper Xingu has long been prominent in debates over village size and permanence in Amazonia due to Robert Carneiro's (e.g., 1957, 1960, 1983, 1985) observations regard- ing subsistence and settlement patterns in the region. He noted that the Kuikuru and other Xinguano villages, although relocating every so often (10 to 30 years), had moved no more than a few hundred meters in nearly 100 years (ca. 1860 to 1960) in some cases. Thus, village relocation was not caused by the exhaustion of garden soils or local ecological degradation. Furthermore, Xinguano agricultural technology, he argued, could support a large, permanent population, perhaps as many as 2,000 in a single settlement. Carneiro's conclusions have been challenged on the grounds of both agricultural productivity and the availability of fished, hunted, and

Heckenberger, Petersen, and Neves] VILLAGE SIZE IN AMAZONIA

@ Major Earthwork Site

@ Earthwork Site of Unknown Configuration

b Non-Earthwork Site

o Site of Unknown Type

Figure 7. Map of Kuikuru study area in the upper Xingu showing location of sites mentioned in text: 1) Nokugu (MT- FX-06); 2) Kuhikugu (MT-FX-11); 3) Hialugihi'ti (MT-FX-13); 4) Hatsikugi (MT-FX-17); 5) Itsagahiti' (MT-FX-18); 6) Secu (MT-FX-19); 7) Sehu (MT-FX-20); 8) Meijeinei (MT-FX-21): 9) Agikuangaku (MT-FX-22); 10) Intagu (MT-FX- 23); 11)Morenh (MT-AX-08); 12) Yakare (MT-FX-09); and 13) Tuatuari (MT-FX-07). (Note: Darker stippling marks seasonally inundated areas and lighter stippling marks seasonally waterloggd areas; small blackened circles in the study area represent a distinctive, late prehistoric to early protohistoric cultural group).

foraged resources in the region (e.g., Descola FX-06), Kuhikugu (MT-FX-1 I), and Hialugihi'ti' 1996: 183-186; Gross 1975; Johnson 1982; Ross (MT-FX- 13)demonstrate that these ditches are 1978). Nevertheless, vestiges of ancient fortified vil- constructed earthworks and that prehistoric villages lages, notably including peripheral ditches or moats, were dramatically larger and more structurally elab- provided evidence that Xinguano villages were much orated than those of contemporary Xinguano vil- larger in the past (Carneiro 1957:212; Dole lages, although sharing the same circular central 196111 962; see also Agostinho 1993). The ditches plaza configuration (Figure 8; see Heckenberger were misinterpreted as natural features by several 1996 for a more detailed discussion). Demonstrable archaeologists who briefly worked in the region, and cultural continuity throughout the chronological have therefore received little attention in discussions sequence (ca. A.D. 1000-2000), documented by of Precolumbian settlement patterns in Amazonia marked conservatism not only in village spatial orga- (Becquelin 1978,1993; Kneip 1969; Simdes 1967). nization (plaza and spoke-like road configuration)

Recent archaeological investigations at three late but also in ceramic technology, subsistence eco- prehistoric sites in the upper Xingu-Nokugu (MT- nomics, and village location (Heckenberger 1996,


Table 4. Uncalibrated Radiocarbon Dates from the Upper Xingu River.

Siteprovenience Lab. # I4cAge B.P d13c Material Dated Nokugu Occupation surface A Beta 72261 1000 f70 -25.0% wood charcoal Occupation surface B Beta 78979 700 f70 -25.0% wood charcoal Ditch 1 fill (middle) Beta 81301 360 & 70 -25.0% wood charcoal Ditch 1 fill (top) Beta 72260 180 It 60 -25.0% wood charcoal

Kuhikugu Occupation surface A Beta 72263 900 It 60 -25.0% wood charcoal Base of road fill Beta 72262 440 It 70 -25.0% wood charcoal

Hialugihiii' Base of terra preta Beta 88363 910 & 80 -25.0% wood charcoal Top of terra prera Beta 88362 690 & 60 -25.0% wood charcoal

MorenBa Gif 3308 920 It 90 wood charcoal Tuatuaria Gif 5365 680 f 60 wood charcoal a Bequelin 1993.

1998a, 1999a), permits fairly detailed direct histor- to the plaza and two sectioning the interior berm of ical comparisons. Fieldwork relevant to the present a peripheral ditch), one at Hialugih'ifi (located over discussion included: 1) systematic mapping of earth- the plaza-marginal curb), and an excavation trench works, including single and paired ditches around at Nokugu (11 contiguous 1 m2 units bisecting a settlements and contiguous linear ridges (curbs) ditch); 3) systematic surface collection of 1,336 and demarcating plaza areas and causeways within vil- 1,949 2 m2 units at Nokugu and Kuhikugu, respec- lages; 2) excavations into and beneath the artificial tively (surface soil color and vegetation also were earthworks, including three 1 m2 excavation units at recorded); and 4) sediment coring conducted in var-Kuhikugu (one positioned over a road curb adjacent ious portions of the three sites (e.g., plaza areas,

Figure 8. Site map of prehistoric site of Nokugu (Mt-Fx-06). Note: Closed named circles (Ati', Ahanitahagu and Waura) represent Xinguano occupations between ca. 1850 to 1971."


Mt-Fx-06

EXCAVATION TRENCH PROFILE West Wall

Key Ceram~c Sherd - Charcoal Lens

Figure 9. Profile of Excavation Trench 1 bisecting Ditch 1 at Nokugu (Mt-Fx-06).

roads, areas between roads, and ditches located near and distant from the plaza) to evaluate subsurface sediment characteristics.

Stratigraphy and Chronology

In almost all excavations, intact anthropogenic TP soils, designated Stratum I1 at all sites, were encountered beneath reworked overburden deposits related to earthwork construction (stratum III) (Figure 9). The intact TP ranged from about 40 to 50 cm in depth, although artifacts extend into a reddish brown transitional zone between the TP and sterile basal sediments (red-colored Stratum I).Shallower, lighter, and non-TP deposits also were documented in more peripheral areas of each site, as documented in surface soil characteristics, sediment cores, and excavations. Radiocarbon dates obtained from slightly above the stratum I/II interface at Nokugu (A.D. 950 f70), Hialugih'iti' (A.D. 1040 f 80), and Kuhikugu (A.D. 1050 f 60) roughly date initial site occupations to the period immediately prior to A.D. 900-1000. No earlier occupations are known from the upper Xingu basin.

The dates from all three sites provide internally

consistent cultural sequences for the intact TP (stratum 11), indicating accretional formation. Correlation of dates for intact Stratum 11between the three sites indicates roughly contemporaneous initial occupations (base of Stratum 11) by ca. A.D. 800 to 900 and earthwork construction (top of Stratum 11) ca. A.D. 1400 to 1450. The maintenance of earthworks apparently declined within a century or so of European contact, based on a date of A.D. 1590 f 70 from the middle of Stratum N (cultural fill deposited in ditch). The sites were completely abandoned by the mid- 1700s as the result of depopulation from introduced diseases and territorial consolidation (Heckenberger 1996).

Village Size and Configuration

Like contemporary Xinguano communities, late prehistoric villages were constructed around a large, central plaza (Figure 10). This is an area of low artifact density and minimal soil alteration (i.e., little TP formation). In both contemporary and prehistoric villages, areas directly outside of the plaza are characterized by high artifact densities and pronounced soil alteration related to domestic activities (Table


Figure 10. Site map of prehistoric site of Kuhikugu (Mt-Fx-11). Note: Closed named circles (Atika, Kuhikugu, Lahatua, and Lamakuka) represent Xinguano occupations between ca. 1860 to 1961.

5, Figure 11). One obvious contrast between prehistoric and contemporary villages is that the plaza location remained fixed during much or all of the long occupational history of the prehistoric villages, as documented by the substantial TP (k 50 cm) deposits surrounding the "clean" plaza areas (within a ring of approximately 100 to 200 m around the plazas). In other words, whereas contemporary settlement patterns result in a patchy distribution of TP, related to the doughnut-shaped distribution of TP around multiple non-overlapping plaza villages, the extensive and relatively deep TP deposits that formed around plazas in prehistoric sites were undoubtedly the result of long-term, continuous occupation of central site areas.

The earliest occupations of these sites were apparently smaller than those of terminal prehistoric times, at least at Nokugu and Kuhikugu where distributions of artifacts and TP are better understood. They were likely intermediate in size between historically known

villages and late prehistoric villages (compare Fig- ures 10 and 11). This hypothesis is based on the lack of notable anthropogenic soil alteration (i.e., TP formation), as would be expected with long-term occupation, in areas more distant from the plaza at Nokugu and Kuhikugu. Domestic ceramics (i.e., residential areas) were distributed across the surface of the sites (i.e., throughout the area confined within the ditches) except in road and plaza areas. We surmise that areas farthest from the central plaza(s) were occupied late (and apparently briefly) in the history of each site due, in part, to population nucleation concomitant with the construction of village fortifications.

The prominent village earthworks, such as ditches and curbed plaza/causeway systems, enable accurate reconstruction of village size and configuration, at least in terminal prehistoric times. Village earthworks were articulated in an integrated and, therefore, contemporaneous architectural plan. Even if one or another village had been temporarily abandoned

Heckenberger, Petersen, and Neves] VILLAGE SIZE IN AMAZONIA 369

Table 5. Sediment Data from Upper Xingu.

Ex. Mg Ex. Ca Ex. K Ex. Na P fract." SiteLocation % Or . Mat. m0m H~ sum m /k Kuikuru I Midden 13.5 68.7 509.2 38.1 7.2 7.5 479 House 5.1 7.9 19.5 15.5 5.2 5.6 216 plazab 5.4 5.4 14.9 10.8 5.1 5.2 76 Plaza 5.4 .3 4.1 2.1 4.1 4.6 72 Plaza 5.7 1 2.3 1.2 3.5 4.1 77 Midden 8.2 20.1 142.8 54.5 5.5 5.9 550

Nokugu Plaza 5.8 9.2 52.9 4.1 6.8 5.6 299 Domestic 7.5 18.9 143.8 3.8 5.8 6.4 556 Note: Sediment analyses were conducted by the Soils and Physical Geography Laboratory, University of Wisconsin-Madison (UWM) and represent total values. a Phosphate fractionation (P fract.) results presented above represent sum of fractions

Denotes sample from approximately 5 meters in front of house

prior to A.D. 1400 to 1500, it is clear from the scale of earthmoving at each site that villagers did not intend to abandon their villages once they had built these features. Structural elaboration of a village to this degree is exactly what would be expected of large, fully sedentary populations, but seems less typical of small, semi-sedentary groups. Moreover,

it is hard to imagine that communities occupying the large and structurally elaborate prehistoric villages practiced a more mobile settlement pattern than their contemporary descendants, who have an essentially sedentary lifestyle.

As is the case in contemporary villages, the central plaza was the gravitational center of prehistoric

Figure 11. Map of Kuikuru village of Ipatse (1973-1983). Note: phosphate fractionation sum values plotted along sediment coring transect bisecting the village (see Table 5).

370 LATIN AMERICAN ANTIQUITY [Voi. 10, No. 4, 1999

villages. Domestic areas surrounding the plaza were many times larger than today, although plaza dimen- sions are comparable to those of contemporary villages (up to 250 m in diameter). The partitioning of each site, conditioned by the placement of artificial earthworks, created discrete intravillage precincts, or "neighborhoods," situated between roads and also delimited by the plaza and excavated ditches. These partitions may well correspond to social divisions (e.g., kin groups or factions), an elaboration on the clustering of related households in contemporary villages. Likewise, households were more than likely positioned with respect not only to aftiliation with one or another rival kin group but also internal rank within that group. What is clear in contemporary villages is that the most powerful individuals (hereditary chiefs) live in houses positioned at key points around the plaza (e.g., cardinal directions or in opposition to one another), and other households position themselves with relation to these in either adjacent areas around the plaza or behind the house of a chief. We can assume that the plaza, in prehistory as today, was not only a gravitational center of the local built environment but also the center of political and ritual life.

In the upper Xingu, as suggested for Aqutuba, the plaza was controlled by a specific segment of society, namely adult men and particularly chiefs. In present-day villages, both the village and central plaza are symbolically "owned" by the two primary named chiefs (in Kuikuru they are called the etk-Bto, or "owner of the village," and hugogd Bto, or "owner of the centerlplaza," respectively). As the locus of diverse intercommunity rituals, the plaza also is a symbolic point of articulation between communities in the integrated regional system. In short, prehistoric cultural patterns fail to meet the expectations of standard model, i.e., communities were neither small and impermanent nor likely egalitarian and politically autonomous.

Without entering into a detailed discussion of population estimates, it is reasonable to conclude that populations in prehistoric villages were considerably higher than those from historically known villages (50 to 350 people), if for no other reason than their physical size (.3 to .5 km2 compared to .06 km2). The largest recent Xinguano village on record is the Kuikuru village, recorded by Heckenberger in 1995, at about .06 km2, or 6 ha, in size with a village population of about 330.It is important to recognize, however, that plaza space constitutes over half of the

entire village area in the contemporary Kuikuru village, but considerably less in prehistoric villages (i.e., there is far more residential space in prehistoric villages). Late prehistoric villages were not only up to 10 times larger than contemporary ones but also substantially more elaborated spatially and structurally, likely the product of a workforce considerably larger than available under current conditions. Therefore, it seems highly likely that village populations ranged into the low thousands, at least 1,000 to 1,500, as known historically from much smaller circular villages among central Brazilian peoples (Nimuendaj6 1967; Posey 1994; Turner 199 1 ;Wiist and Barreto 1999), although their populations may have ranged significantly higher, as suggested elsewhere (Agostinho 1993:275).

Regardless of exact village sizes, regional population must have been dramatically higher than historically documented (some 3,000 people in 20 or more villages during the 1880s), due to the much higher incidence of contemporaneous sites, as suggested by the density and regular placement of fortified villages constructed ca. A.D. 1400 to 1500. Specifically, excavations at Nokugu, Kuhikugu, and Hialugih'it'i not only have revealed similar develop- mental histories for these sites, but indicate that the ditches and the linear mounds placed at the edges of plazas and roads were constructed more or less simul- taneously across the region. Thus, the earthworks reflect a relatively sudden shift in regional settlement patterns that, by extension, applies to other undated fortified sites (i.e., Mt-Fx- 17,18 and 22 also were more or less contemporaneous).

Like contemporary villages, prehistoric settlements were undoubtedly interconnected by a well- developed system of roads and paths (i.e., extensions of the intravillage causeways), but there is no evidence that one or another of these villages held sway over any other. Given available site distribution data and dating in the upper Xingu, some contemporaneous settlements were larger and probably more powerful than others. Research to date suggests that numerous relatively large and closely spaced settlements with extensive earthworks were present in the area during late prehistoric and early historic times, but their precise interrelationship remains unknown. In the Kuikuru study area, where today there is only one plaza village, there may have been five contemporary plaza villages (Mt-Fx-06, 1 1, 13,17 and 18) and perhaps even more (Mt-Fx-22 and 23), with few


or no intervening second-order settlements. At the regional level, it appears that large villages existed as more or less peer communities, with perhaps one or a few smaller satellites (e.g., Mt-Fx-19, 20, and 21). Similar dense, sedentary social formations apparently characterized the Pareci groups, closely related Arawak-speaking peoples who lived west of the upper Xingu. In the 1720s, Pires de Campos (1 862:443-444; authors' translation) noted that they existed "in such numbers, that their villages could not be counted and often in a day's march, 10 to 12 villages were passed, each with 10 to 30 houses of 30-40 paces [meters] width. . . even their roads they make very straight and wide and conserved so clean and in good repair that not even a leaf is to be found."

Discussion

The findings from the lower Negro and upper Xingu underscore the substantial variability of Amazonian settlement patterns, including apparently large and fully sedentary communities and fairly dense regional population aggregates across a more diverse range of ecological settings than commonly accepted. Specif- ically, our findings neither support the standard model, insofar as it suggests that, due to environmental constraints, Amerindian settlements in Amazonia are necessarily small, impermanent, and dispersed (e.g., Meggers 1992,1995b, 1996,1997), nor do they agree with the more limited application of the standard model to non-vcirzea settings or those areas that lack nutrient-rich soils (e.g., Roosevelt 1980, 1991, 1994; cf. Cameiro 1985,1995). A further implication of our studies is that the "carrying capacity" of diverse Ama- zonian settings is substantially higher than commonly accepted and, therefore, cannot be easily correlated with any narrowly defined ecological variable (e.g., high fertility soils), as proposed by some revisionists (e.g., Lathrap 1 9 7 0 a : 3 W ; Roosevelt 1980:79-92, 112-1 19). In other words, although variability in set-tlement patterns across the region obviously corre- lates, in part, to ecological patterning, it cannot be explained through recourse to general models derived from broadly defined ecological variables (e.g., vhrzea vs. terra$rme, white-water vs. black- or clear-water rivers). The vcirzealterrajirme dichotomy (or vcirzea model) and the standard model from ethnography (i.e., the tropical forest culture) thus perpetuate images of homogeneity where it has not been demonstrated and undoubtedly grossly oversimplify Precolumbian patterns.

A particularly dubious assumption that arises from ignoring regional variability is that sociocul- tural formations were necessarily similar across vast areas, corresponding to one or a few narrowly defined cultural "types" (e.g., vcirzea chiefdoms, tropical forest tribes). The natural landscape of Amazonia was extremely varied, and the cultures that lived in and transformed it were no less diverse. Minimally, we can conclude that the "complex end" of cultural diversity in the terra jirme (i.e., black-water and clear-water rivers, small tributaries, and interfluves), in terms of sedentism, local and regional population size, economic production, and sociopolitical com- plexity (both hierarchical and heterarchical), has been significantly underestimated due, in large part, to assumptions derived from ethnography and recent human ecology. Ethnographic patterns of small village size and frequent relocation, such as predicted by the standard model, are well known from Ama- zonia, but many groups described as definitive exarn- ples of such a pattern have been demonstrably displaced as a result of contact (e.g., Kayapd and TarumP), or live in areas that seem to have supported relatively large, sedentary social formations prehistorically (e.g ., Jivaro and Siriono). As the upper Xingu example clearly demonstrates, we must exer- cise extreme caution in retrodicting ethnographic patterns, generally derived from observations made during the demographic nadir of Amazonian peoples (ca. 1925 to 1975), into the past (Beckerman 1978, 1979, 1991; DeBoer et al. 1996; Roosevelt 1989; Whitehead 1994).

In both the lower Negro and upper Xingu, settlement patterns show a strong tendency toward occupation of non-inundated forest areas that are directly adjacent to major waterways. Thus, tracts of arable terra$rme land and rich aquatic resources undoubtedly provided the economic base for these large settled populations (see Cameiro 1986; Denevan 1996). This conclusion conforms to a general pattern i n h a - zonia whereby fishing takes precedence over terres- trial hunting in areas where aquatic resources are abundant (Beckerman 1979, 1994; Carneiro 1970, 1986, 1995; Lathrap et al. 1985; Riviere 1984:ll).

There could have been little or no emphasis on local floodplain agriculture in either the upper Xingu or lower Negro areas. In the upper Xingu, the seasonally saturated, compact (gley) soils of the floodplain are unsuitable for cultivation, as are the sand levees (galeria forests) directly adjacent to major

372 LATIN AMERICAN ANTIQUITY [Vol. 10, NO. 4, 1999

rivers. As more fully explained elsewhere (Hecken- berger 1998a), manioc agriculture in terra$rme areas of the upper Xingu must have been the prehistoric economic base, as it is today (see Carneiro 1985; Dole 1961/1962). In the lower Negro case, there are no such floodplains at all. High terra$rme riverbluffs rise above seasonal sand beaches and inundated forests (igapds)along the Negro. Thus, the hypothesis that higher economic productivity of restricted floodplain settings provided the necessary basis for sedentism and population growth, as has commonly been proposed (e.g., Carneiro 1986, 1995; Denevan 1996; Lathrap 1970a; Meggers 1996; Moran 1991, 1993, 1995; Roosevelt 1980,1994), is untenable for both the lower Negro and upper Xingu. The large, sedentary villages in both areas were apparently supported by intensive terra firme agriculture (see Denevan 1992a), rather than extensive shifting agriculture as typical of contemporary Amerindian groups in Amazonia, or the intensive floodplain agriculture suggested for some areas in the prehistoric period (Roosevelt 1980, 1994).

An inescapable implication of our research is that Amerindian villages can no longer be generally portrayed as small, impermanent, or autonomous- isolated "clearings" in an otherwise vast, homoge- neously forested landscape. Our findings support a growing body of research that documents that Amerindian populations are not only highly selec- tive in their use of the natural landscape, but that over time choice settlement locations are not necessarily degraded but can become even more desirable through intentional modifications and unintentional alterations resulting from human occupation (e.g., BalCe 1989, 1994, 1995; Denevan 1992a, 1992b, 1996; Denevan and Padoch 1988; Posey and BalCe 1989). Long-term settlement continuity and outward growth from central or "core" areas at Aqutuba and in the upper Xingu resulted in increased internal dif- ferentiation and alteration (construction) of the local environment. As village populations increased and village spatial organization became more enduring and structurally elaborated, including maintenance of fixed central plazas and substantial earthmoving, specific places were radically transformed and communities became more committed to them. Agutuba and particularly the upper Xingu exemplify this kind of organization, but we strongly doubt they are unique. Some populations, such as in the Northwest Amazon, were strongly attached to specific places,

but remained small in terms of community and regional population (e.g., Mora et al. 1991; Neves 1998). Still other communities reached large sizes, but were unstable in their location and composition, such as some G2villages, for example (e.g., Nimuen- daju 1967; Posey 1994; Turner 1991; Verswijver 1992; Wiist and Barreto 1999). Furthermore, there was considerably greater variation within discrete regions than is commonly accepted, as indicated by both the lower Negro and upper Xingu studies.

In conclusion, we believe that prehistoric Ama- zonian societies, as a broadly comparable cultural universe, were neither ecologically prevented from settled village life nor narrowly constrained by an egalitarian ideology (Heckenberger 1998b, 199913). Nonetheless, the presence of large, sedentary population aggregates does not lead us to conclude that these societies were politically integrated into cen- tralized, hierarchical sociopolitical systems at local or regional levels (see Clastres 1977; Fausto 1992). Rather, it demonstrates that there is no a priori rea-son to assume that they were not.

Acknowledgments. The authors thank the Ideta family, particularly Kunitaka Ideta and his family, and Dona Chdida Jesus de Silva, her family, and the other families of the Agutuba community, co-owners of the Aptuba locality, who graciously provided permission to study the site, as well as accommodations and assistance throughout the project. Research in the lower Negro area is being conducted under a formal agreement between the Museu de Arqueologia e Emologia, Universidade de SBo Paulo (MAENSP) and the Camegie Museum of Natural History (CMNH). Heckenberger especially thanks the Kuikuru (Xinguano Amerindian) community for their gracious support of the upper Xingu research; in particular, Chief Afukaka Kuikuru, his brothers and family, and the numerous young men who aided in the conduct of the fieldwork. Various individuals at the MAENSP, the CMNH, and the University of Maine at Farmington Archaeology Research Center (UMF ARC) aided the project in a variety of ways, particularly at the MAE/USP where the authors and assistants conducted the laboratory work. We gratefully acknowledge the William Talbot Hillman Foundation (Pittsburgh), the Wenner-Gren Foundation (New York), the UMF ARC, the CMNH, and the Universidade de SZo Paulo for financial support of the Agutuba research. Heckenbeger thanks the Museu Nacional/UFederal do Rio de Janeiro (UFRJ) and the Museu Antropologico, UFederal de Goiis (UFG), institutional co-sponsors of the upper Xingu research, and, in particular, Bmna Franchetto (UFRJ) and Irmhild Wiist (UFG) for their support. Financial support for the upper Xingu research was provided by the Social Science Research Council (New York), the National Science Foundation (Grant No. DBS-9214806), Victor and Mary Heckenberger, the UMF ARC, and the W. T. Hillman Foundation. Denise Dal Pino de Souza (MAENSP) and Jeff Williams (UMF ARC) drafted Figure 2, and Williams con-


tributed to many of the other figures as well. Marina Valesco kindly provided the Spanish abstract. Cindy Longwell (University of Vermont) helped in final manuscript preparation. Discussions with Robert Carneiro, William Denevan, Anna Roosevelt, and William Woods provided helpful insights for the paper, as did the comments of six anonymous reviewers and the editorial staff of Latin American Antiquity. As customary, all errors or omissions are the responsibility of the authors.

References Cited Agostinho da Silva, P.

1993 Testemunhos da ocupaqgo pr6-xinguana na bacia dos for- madores do Xingu. In Karl von den Steinen: um sr'culo de antropologia noXingu, edited by V. P. Coelho, pp. 233-287. Editora da Universidade de Slo Paulo, Slo Paulo.

Balie, W. 1989 The Culture of Amazonian Forests. In Resource Man-

agement in Amazonia: Folk and Indigenous Strategies, edited by D. A. Posey and W. Balee, pp. 1-21. Advances in Eco- nomic Botany No. 7. New York Botanical Garden, New York.

1994 Footprints of the Forest: Ka'apor Ethnobotany. Colum-bia University Press, New York.

1995 Historical Ecology in Amazonia. In Indigenous Peoples and the Future of Amazonia: An Ecological Anthropology ofanEndangeredWorld, edited by L. E. Sponsel, pp. 97-1 10. University of Arizona Press, Tucson.

Beckerman, S. 1978 Comment on "FoodTaboos, Diet, and Hunting Strategy:

The Adaptation to Animals in Amazon Cultural Ecology" (Eric Ross). Current Anthropology 19: 17-19.

1979 The Abundance of Protein in Amazonia: A Reply to Gross. American Anthropologist 81533-560,

1991 A AmazBnia estava repleta de gente em 1492? In Ori-gens, adaptacdes e diversidade bioldgica de homem nativo da AmazGnia, edited by W. A. Neves, pp. 143-159. Museu Paraense Emilio Goeldi, Belim.

1994 Hunting and Fishing in Amazonia: Hold the Answers, What are the Questions? In Amazonian Indiansfrom Pre- histoiy to the Present: Anthropological Perspectives, edited by A. C. Roosevelt, pp. 177-200. University of Arizona Press, Tucson.

Becquelin, P. 1978 The Archaeology of the Upper Xingu, Mato Grosso.

Paper presented at the 77th Annual Meeting of the Ameri- can Anthropological Association, Los Angeles.

1993 Arqueologia Xinguana. In Karl von den Steinen: um siculo de antropologia noXingu, edited by V. P. Coelho, pp. 224-232. Editora da Universidade de Sgo Paulo, Slo Paulo.

Brochado, J. P. 1984 An Ecological Model of the Spread of Pottery andAgri-

culture into Eastern South America. Ph.D. dissertation, Uni- versity of Illinois at Urbana-Champaign. University Microfilms, Ann Arbor.

Brochado, J. P., and D. W. Lathrap 1982 Amazonia. Manuscript on file, Department of Anthro-

pology, University of Illinois, Urbana-Champaign, Carneiro, R. L.

1957 Subsistence and Social Structure: An Ecological Study of the Kuikuru Indians. Ph.D, dissertation, University of Michigan, Ann Arbor. University Microfilms, Ann Arbor.

1960 Slash-and-Bum Agriculture: A Closer Look at its Impli- cations for Settlement Patterns. In Men and Culture, edited by A. F.C. Wallace, pp. 229-234. University of Pennsylva- nia Press, Philadelphia.

1970 The Transition from Hunting to Horticulture in t h e h a -

zon Basin. Proceedings of the Eighth International Congress of Anthropological and Ethnological Sciences 3 : 1 U 8 . Tokyo.

1983 The Cultivation of Manioc among the Kuikuru of the Upper Xingu. In Adaptive Responses of Native Ama:onians, edited by R. B. Hames and W. T. Vickers, pp. 65-1 11. Aca- demic Press, New York.

1985 Slash-and-Bum Cultivation among the Kuikuru and Its Implications for Cultural Development in the Amazon Basin. In Native South Americans; Ethnology of the Least Known Continent, edited by P. Lyon, pp. 73-91. Waveland Press, Prospect Heights, Illinois.

1986 The Ecological Basis of Amazonian Chiefdoms. Manuscript on file, Department of Anthropology, American Museum of Natural History, New York.

1995 The History of Ecological Interpretations ofhazonia : Does Roosevelt Have It Right? In Indigenous Peoples and the Future ofAmazonia; An Ecological Anthropology of an Endangered World, edited by L. E. Sponsel, pp. 45-65. Uni- versity of Arizona Press, Tucson.

Clastres, P. 1977 Society against the State: Essays in Political Anthro-

pology. Zone Books, New York. DeBoer, W. R.

1981 Buffer Zones in the Cultural Ecology ofAboriginalha- zonia: An Ethnohistorical Approach. American Antiquity 46:36&377.

DeBoer, W.R., K. Kintigh, and A. Rostoker 1996 Ceramic Seriation and Settlement Reoccupation in Low-

land South America. Latin American Antiquity 7:263-278. Denevan, W. M.

1976 The Aboriginal Population of Amazonia. In The Native Population ofthe Americas in 1492,edited by W. Denevan, pp. 205-234. University of Wisconsin Press, Madison.

1984 Ecological Heterogeneity and Horizontal Zonation of Agriculture in the Amazon Floodplain. In Frontier Expan- sion in Amazonia, edited by M. Schmink and C. Wood, pp. 311-336. University of Florida Press, Gainesville.

1992a Stone vs. Metal Axes: The Ambiguity of Shifting Cul- tivation in Prehistoric Amazonia. Journal of the Steward Anthropological Society 20:153-165.

1992b The Pristine Myth: The Landscape of the Americas in 1492.Annals ofthe Association of American Geographers 82:369-385.

1996 A Bluff Model of Riverine Settlement in Prehistoric Amazonia. Annals of the Association ofAmerican Geogra- phers 86:65448 1.

Denevan, W. M., and C. Padoch (editors) 1988 Swidden-Fallow Agroforestry in the Peruvian Amazon.

Advances in Economic Botany No. 5. New York Botanical Garden, New York.

Descola, P. 1996 In the Society of Nature: A Native Ecology ofAmazo-

nia. Cambridge University Press, Cambridge. Dole, G. E.

196111962 A Preliminary Consideration of the Prehistory of the Upper Xingu Basin. Revista do Museu Paulista n.s. XIII:399-423.

Eden, I. M., W. Bray, L. Herrera, and C. McEwan 1984 Terra Preta Soils and Their Archaeological Context in

the Caqueti Basin of Southeast Columbia. American Antiq- uity 49:125-140.

Eidt, R. 1984 Advances in Abandoned Settlement Analysis: Applica-

tions to Prehistoric Anthrosols in Columbia, South Amer- ica.The Center for Latin America, University of Wisconsin, Milwaukee.


Fausto, C. 1992 Fragmentos de historia e cultura Tupinambl: dd etnolo-

gia como instrumento critico de conhecimento etno-his- torico. Histdria dos indios no Brasil, edited by M. Carneiro da Cunha, pp. 381-396. Companhia das Letras/FAPESP/SMC, Slo Paulo.

Gross, D. 1975 Protein Capture and Cultural Development in the Ama-

zon Basin. American Anthropologist 77526549. 1983 Village Movement in Relation to Resources in Amazo-

nia. In Adaptive Responses ofNative Amazonians, cdited by R. B. Hames and W. T. Vickers, pp. 429-449. Academic Press, New York.

Heckenberger, M. J. 1996 War and Peace in the Shadow of Empire: Soci~poiitical

Change in the Upper Xingu of Southeastern Amazonia, ca. A.D. 1400 - 2000. Ph.D. dissertation, University of Pitts- burgh, Pittsburgh. University Microfilms, Ann Arklor.

1997 Investiga~Bes arqueol6gicas no baixo Rio Jali (medio Negro), Amazonas: relatorio de pesquisas prelin~inar em 1996. Manuscript on file, Funda~io Vitoria AmazBnica, Manaus.

1998a Manioc Agriculture and Sedentism in Amazonia: The Upper Xingu Example. Antiquity 72:633-648.

1998b Hierarquia e economia politica em Amazonia: a estrutura de desigualidade na sociedade Xinguana. Pi~per presented at the 20th Annual Meeting of the Associa$lo Brasileira de Antropologia, Vitoria.

1999a Estrutura, historia e transforma$50: a cultura Xinguana no longue dure'e (1000 a 2000 d.C.). In Histdria 5 clrltura indigena no altoxingu: visdes antropoldgicas, edited by B. Franchetto e M. J. Heckenberger. Editora da Universidade Federal do Rio de Janeiro, Rio de Janeiro, in press.

1999b 0 enigma das grande cidades: o corpo e o estado Amerindia. In A outra margem do Occidente. B,easil 500 anos: experitncia e destino, vol. 2, edited by Adauto Novaes. Companhia das Letras, Slo Paulo, in press.

Heckenberger, M. J., E. G. Neves, and J. B. Petersen 1998 A onde surgem os modelos?: considera$Ber, sobre a

origem e expans50 dos Tupi. Revista de Antropologia 41 :69-96.

Hilbert, P. P. 1968 Archaeologische Untersuchungen am Mittler.en Ama-

zonas. Dietrich Reimer Verlag, Berlin. Johnson, A.

1982 Reductionism in Cultural Ecology: The Ama2.on Case. Current Anthropology 23:413428.

Kem, D. 1996 Geoquimica e pedogeoquimica em sitios arqut~oldgicos

com terra preta na Floresta Nacional de Cawiuanii (Portel- PA).Unpublished Ph.D, dissertation, Centro de Gec~cihcias, Universidade Federal do Parl, Par&, Brazil.

Kneip, L. M. 1969 Relatdrio sobre as "valetas" do Parque National do

Xingu, estado de Mato Grosso. Manuscript on fill:, Depar- tamento de Antropologia, Museu Nacional, Rio d<: Janeiro.

Lathrap, D. W. 1968 The "Hunting" Economies of the Tropical Forest Zone

of South America: An Attempt at Historical Persp~ctive. In Man the Hunter, edited by R. B. Lee and I. Devore, pp.23-29. Aldine, Chicago.

1970a The Upper Amazon. Praeger, New York. 1970b Review of Archaeologische Untersuchunger, am Mit-

tleren Amazonas, by P. P. Hilbert. American Antiquity 35:499-501.

1977 Our Father the Cayman, Our Mother the Gourd: Spin- den Revisited, or a Unitary Model for the Emergence of

Agriculture in the New World. In Origins of Agriculture, edited by C. A. Reed, pp. 713-751. Mouton, The Hague.

Lathrap, D. W., A. Gebhart-Sayer, and A. M. Mester 1985 The Roots of the Shipibo Art Style: Three Waves on

Imariacocha or There Were "Incas" before the Incas. Jour-nal of Latin American Lore 11:3 1-1 19.

Lowie, R. H. 1948 The Tropical Forests: An Introduction. In The Tropical

Forest Cultures, edited by J. H. Steward, pp. 1-56. Hand- book of South American Indians, vol. 3. Bulletin No. 143, Bureau of American Ethnology, Smithsonian Institution, Washington, D.C.

Machado, A. M. 1993 As tradigdes ceramistas da bacia AmazBnica: uma

andlise critica baseada nus evidtncias arqueoldgicas do me'dio Rio Urubu (AM). Unpublished M.A. thesis, Univer- sidade Federal de Pemambuco, Recife.

Meggers, B. J. 1954 Environmental Limitation on the Development of Cul-

ture. American Anthropologist 565301-24. 1986 El uso de secuencias ceramicas seriadas para inferir con-

ducts social. In Actas del primer simposio de la fundacion de argueologia del Caribe, edited by 0.M. FonsecaZamora, pp. 11- 32. Caracas.

1990 Reconstru$io do comportamento locacional prt- hist6rico na AmazBnia. Boletim do Museu Paraense Emilio Goeldi, Sene Antropol6gico 6: 183-203.

1991 Cultural Evolution in Amazonia. In Projiles in Cultural Evolution, edited by A. T. Rambo and K. Gillogly, pp. 191-216. Anthropological Papers 85. Museum of Anthro- pology, University of Michigan, Ann Arbor.

1992 Prehistoric Population Density in the Amazon Basin. In Disease and Demography in the Americas, edited by J. W. Verano and D. H. Ubelaker, pp. 197-205. Smithsonian Insti- tution Press, Washington, DC.

1994 Pre-Columbian Amazonia. National Geographic Research and Explorer 10:398-42 1.

1995a Amazonia on the Eve of European Contact: Ethnohis- torical, Ecological and Anthropological Perspectives. Revista de Arqueologia Americana 8:91-115.

1995b Judging the Future by the Past: The Impact of Envi- ronmental Instability on Prehistoric Amazonian Populations. In Indigenous Peoples and the Future ofiimazonia: An Eco- logical Anthropology of an Endangered World, edited by L. E. Sponsel, pp. 15-43. University of Arizona Press, Tucson.

1995c Archaeological Perspectives on the Potential of Ama- zonia for Intensive Exploitation. In The Fragile Tropics of Latin America: Sustainable Management of Changing Envi- ronments,edited by T. Nishizawa and J. I. Uitto, pp. 68-93. United Nations University Press, Tokyo.

1996 Amazonia: Man and Culture in a Counterfeit Paradise. Rev. ed. Smithsonian Institution Press, Washington, D.C.

1997 Review ofAmazonian Indiansfrom Prehistory to thePre- sent: Anthropological Perspectives, edited by A. C. Roo- sevelt. Journal of the Royal Anthropological Institute 2: 194-196.

Meggers, B. J., 0 . F. Dias, E. T. Miller, and C. Perota 1988 Implications of Archaeological Distributions in Amazo-

nia. In Proceedings of a Workshop on Neotropical Distribu- tion Patterns, edited by W. R. Heyer and P. E. Vanzolii, pp. 275-294. Academia Brasileira de Ciencias, Rio de Janeiro.

Meggers, B. J., and C. Evans 1957 Archeological Investigations at the Mouth of the Ama-

zon. Bulletin No. 167. Bureau of American Ethnology, Smithsonian Institution, Washington, D.C.

1961 An Experimental Formulation of Horizon Styles in the Tropical Forest Area of South America. In Essays in Pre-


Columbian Art and Archaeology, edited by S. K. Lothrop, pp. 372-388. Harvard University Press, Cambridge.

1983 Lowland South America and the Antilles. In Ancient South Americans, edited by J. D. Jennings, pp. 287-335. W. H. Freeman, San Francisco.

Miller, E. T. 1992 Archaeology in the Hydroelectric Projects of Elec-

tronorte: Preliminary Results. Electronorte, Brasilia. Mora C. S., L. F. Herrera, I. Cavalier F., and C. Rodriguez

1991 Cultivars, Anthropic Soils and Stability: A Preliminary Report of Archaeological Research in Araracuara, Columbian Amazon. Latin American Archaeology Reports 2. University of Pittsburgh, Pittsburgh.

Moran, E. F, 1991 0 estudo da adapta~iio humana em ecosistemas

Amazbnicos. In Origens, adaptagbes e diversidade bioldgica do homem nativo da AmazBnia, edited by W. A. Neves, pp. 161-178. CNPq-Museu Paraense Emilio Goeldi, Belem.

1993 Through Amazon Eyes: The Human Ecology of Ama- zonian Populations. University of Iowa Press, Iowa City.

1995 Disaggregating Amazonia: A Strategy for Understand- ing Biological and Cultural Diversity. In Indigenous Peoples and the Future of Amazonia: An Ecological Anthropology of an Endangered World, edited by L. E. Sponsel, pp. 72-95. University of Arizona Press, Tucson.

Myers, T. P. 1973 Toward a Reconstruction of Prehistoric Community Pat-

terns in the Amazon Basin. In Variation in Anthropology: Essays in Honor ofJohn C.McGregor,edited by D. W. Lath- rap and J. Douglas, pp. 233-252. Illinois Archaeological Survey, Urbana.

1992 The Expansion and Collapse of the Omagua. Journal of the Steward Anthropological Society 20: 129-152.

Neves, E. G. 1998 Paths through Darkwaters: Archaeology as Indigenous

Histoy in the Upper Rio Negro, Northwest Amazon. Unpub-lished Ph.D. dissertation, Department ofAnthropology, Indi- ana University, Bloomington.

Neves, E. G., and R. N. Bartone 1998 Preliminary Results of Archaeological Survey in the

Central Amazon. Paper presented at the 65th Annual Meet- ing of the Society for American Archaeology, Seattle.

Nimuendajd, C. U. 1967 The ApinayP. Anthropological Publications, Oosterhout

N. B., The Netherlands. Pabst, E.

1991 CritCrios de distin~iio entre terra preta e latossolo na regiiio de Belterra e os seus significados para a discuss20 pedogenitica. Boletim do Museu Paraense Emilio Goeldi, Strie Antropoldgico,7:5-19.

Petersen, J. B. 1996 The Archaeology of Trants, Montserrat, Pt. 3. Chrono-

logical and Settlement Data. Annals of Carnegie Museum 65:323-361.

Pires de Campos, A. 1862 Breve noticia que dB o capit50 Antonio Pires de Cam-

pos. Revista Trimestral do Institute Histdrico, Geografico e Etnograjco do Bras i l25:43749 .

Porro, A. 1993 As crbnicas do rio Amazonas: notas etno-histdricas

sobre as antigas populaqdes indigenas da AmazBnia. Edi-tora Vozes, Petropolis.

1996 0povo das aguas: ensaios de etno-histdria AmazBnica. Editora Vozes, Petropolis.

Posey, D. A. 1994 Environmental and Social Implications of Pre- and Post-

contact Situations on Brazilian Indians: The Kayap6 and a

New Amazonian Synthesis. InAmazonian Indiansffom Pre- history to the Present: Anthropological Perspectives, edited by A. C. Roosevelt, pp. 271- 286. University of Arizona Press, Tuscon.

Posey, D. A,, and W. BalCe (editors) 1989 Resource Management in Amazonia: Folk and Indige-

nous Strategies. Advances in Economic Botany No. 7. New York Botanical Garden, New York.

Raymond, J. S. 1995 From Potsherds to Pots: A First Step in Constructing Cul-

tural Context from Tropical Forest Archaeology. In Archae-ology in the LowlandAmerican Tropics: Current Analytical Methods and Applications, edited by P. Stahl, pp. 224-242. Cambridge University Press, Cambridge.

Rivisre, P. 1984 Individual and Society in Guiana: A Comparative Study

of Amerindian Social Organization. Cambridge University Press, Cambridge.

Roosevelt, A. C. 1980 Parmana: Prehistoric Maize and Manioc Subsistence

along theAmazon and 0rinoco.Academic Press, NewYork. 1989 Resource Management in Amazonia before the Con- -

quest: Beyond Ethnographic Projection. In Resource Man- agement inAmazonia: Folk andlndigenous Strategies, edited by D. A. Posey and W. Balee, pp. 36-62. ~dvances in Eco- nomic Botany No. 7. New York Botanical Garden, NewYork.

1991 Moundbuilders of the Amazon: Geophysical Archaeol- ogy on Marajo Island, Brazil. Academic Press, New York.

1994 Amazonian Anthropology: Strategy for a New Synthe- sis. In Amazonian Indiansfiom Prehistory to the Present: Anthropological Perspectives, edited by A. C. Roosevelt, pp. 1-29. University of Arizona Press, Tucson.

Ross, E. B. 1978 Food Taboos, Diet, and Hunting Strategy: The Adapta-

tion to Animals in the Amazon. Current Anthropology 19:l-36.

Simdes, M. F. 1967 ConsideraQ3es preliminares sobre a arqueologia do alto

Xingu. In Programa Nacional de Pesquisas Arqueoldgicas: resultados preliminares do primeiro ano. 1965-1966. vol. 6, pp. 129-144. Museu Paraense Emilio Goeldi, Beltm.

1970 Pesquisas arqueoldgicas no baixoRio Negro (Projeto Rio Negro). Manuscript on file, Museu Paraense Emilio Goeldi, BelCm.

1974a Contribui~BoA arqueologia dos arredores do baixo Rio Negro, Amazonas. Publicaqbes Avulsas 26: 165-188. Museu Paraense Emflio Goeldi, Beltm.

1974b Pesquisas arqueoldgicas no baixo Rio Negro (AM): Projeto Rio Negro. Manuscript on file, Museu Paraense Emflio Goeldi, BelCm.

1983 Pesquisas arqueoldgicas no alto Rio Negro (AM): relatdrio sucinto. Manuscript on file, Museu Paraense Emilio Goeldi, BelCm.

Simdes, M. F., and A. Kalkman 1987 Pesquisas arqueoldgicas no mtdio Rio Negro. Revista

de Arqueologia 4:83-115. Smith, N. J. H.

1980 Anthrosols and Human Canying Capacity inhazonia . Annals of the Association of American Geography 7033-566.

Steward, J. H. 1948 Culture Areas of the Tropical Forests. In The Tropical For-

est Tribes, edited by J. H. Steward, pp. 883-899. Handbook of South American Indians, vol. 3. Bulletin No. 143, Bureau of American Ethnology, Smithsonian Institution, Washing- ton, D.C.

1949 South American Cultures: An Interpretive Summary. In


The Comparative Study of South American Indians, edited by J. H. Steward, pp. 669-772. Handbook of South Ameri- can Indians, vol. 5. Bulletin No. 143, Bureau of American Ethnology, Smithsonian Institution, Washington, D.C.

Turner, T. 199 1 The Mebengokre Kayapo: History, Social Consciouness

and Social Change from Autonomous Communities to Inter- Ethnic System. Manuscript on file, Instituto Socioambien- tal, Sio Paulo.

Verswijver, G. 1992 The Club-Fighters of the Amazon: Warfare among the

Kaipo Indians of Central Brazil. Rijksuniversiteit, Ghent. Viveiros de Castro, E. B.

1996 Images of Nature and Society in Amazonian Ethnology. Annual Review ofAnthropology 25: 179-200.

Whitehead, N. L. 1994 The Ancient Amerindian Polities of the Amazon, the

Orinoco, and the Atlantic Coast: A Preliminary Analysis of Their Passage from Antiquity to Extinction. In Amazonian Indiansfrom P r e h i s t o ~ to the Present: Anthropological Per- spectives, edited by A. C. Roosevelt, pp. 33-53. University of Arizona Press, Tucson.

1996 Amazonian Archaeology: Searching for Paradise? A Review of Recent Literature and Fieldwork. Journal of Archaeological Research 4:241-264.

Woods, W. I. 1995 Comments on the Black Earths ofhazonia . Papers and

Proceedings of the Applied Geography Conferences 18:159-165. Arlington, Virginia.

Woods, W. I., and J. M. McCann 1998 The Living Soils of Amazonia. Manuscript on file,

Department of Geography, Southern Illinois University, Edwardsville.

Wiist, I., and C. Barreto 1999 The Ring Villages of Central Brazil: A Challenge for

ArnazonianArchaeology. LadnAmerican Antiquity 10:3-23.

Notes 1. Terra preta do indio refers to anomalously dark (brown

to black) and nutrient-rich Amazonian soils of anthropogenic origin and is typically found in association with archaeological remains (e.g., Smith 1980; Woods 1995).

2. In Portuguese, vdrzea generally refers to any river floodplain or inundated bottomlands.

3. There are at least three distinct sites overlooking Lago Grande (designated AM-IR-12 on Figure 2) and three or more possibly related sites around Lago do Limiio (AM-IR- 9 on Figure 2).

4. The four sites identified by Hilbert ("ParedZo," "Refiniria," "Ponta de FAB," and "Manaus"), all lying on riverbluff areas within central portions of Manaus, are not plotted on Figure 2. Smith (1980:560) mentioned an additional six sites in the Manaus area; three of these may correspond to sites encountered in our survey (Praia Dourado, MT-MA-35), or those of Simdes (1974a; Ponta Negra and Siderama just upstream from Manaus). Neither these nor the remaining three small sites (.5 to 4 ha), located some distance from the Negro River on small perennial streams, can be precisely plotted on Figure 2.

5. All excavations were conducted in 10-cm levels within natural stratigraphy and all sediment from test pits and excavation units was screened (6.4 mm); samples of recovered sediment (5 to 15 liters) were waterscreened using a 1.0 mm

sieve, or floated using a 1.5 mm heavy fraction sieve. 6. Artifacts encountered in this dated level may have been

redeposited from higher levels and carbonized botanical remains may reflect natural burning. Nonetheless, the chemical values from these lower levels (80 to 110 cm) generally correlate well with those from other cultural deposits at the site and from other Amazonian sites (Eden et al. 1984; Kern 1996; Smith 1980), indicating that this may well be an early cultural horizon.

7. Apparently earlier ceramics were encountered in levels radiocarbon dated to between 6850 f 100 B.P. (90 to 100 cm) and 2310 f 140 B.P. (60-70 cm) in Unit 1, Aqutuba IIB.

8. Cauixi (sponge-spicule) is the dominant temper, although caraipk and caraipklcauixi- tempered ceramics are present throughout the sequence, as are other temper materials in lesser quantities, including grog, charcoal, and perhaps sand. Primary decoration includes incision, punctation, mod- eling, and a unique "cut" rim technique. Possibly associated with this complex are co-occuring variants of polychrome painted ceramics (red, orange, and other paints on a white slip clearly distinctive from later Guarita polychrome tech- niques and styles) and distinctive fine-line incised varieties, both from deposits securely dated to ca. A.D. 150 (see Heckenberger et al. 1998). Ceramics broadly comparable to these A~utuba ceramics are known from radiocarbon-dated deposits in the central Amazon at the Itacoatiara (ca. 95 B.C. andA.D. 86), Manacapurd (A.D. 425), Coari (ca. A.D. 763), and Caiambe sites (ca. A.D. 640 and 730) (Hilbert 1968:253).

9. The majority of the diagnostic Guarita ceramics are tempered with caraipe', with lesser amounts of cauixi- and caraipilcauixi-tempered sherds present. Similar ceramic materials have been recovered from sites on the Uatumi River, dated to ca. A.D. 890 to 1520 (Miller 1992:17-20), from sites on the Apuad River (middle Negro), dated to ca. A.D. 825 to 1560 (SimBes 1974a; SimBes and Kalkman 1987), and to A.D. 1150 at Coari on the middle SolimBes River (Hilbert 1968:253). A medially-flanged vessel, identi- cal to Guarita forms at A~utuba, from the middle Negro has recently been dated to A.D. 1340 (Heckenberger 1997).

10. These high concentrations of ceramics do not apparently relate to specialized production areas due to the general lack of ceramic manufacture refuse or sherds from unused vessels

11. The ditch bisecting the plaza may have connected with another ditch identified in northwestern portions of Aqutuba 11, but the exact configuration and position of this ditch is uncertain since it was filled in by the landowner and has since been plowed over numerous times. Our placement of this ditch in Figure 4 is approximate and based on landowner testimony.

12.North arrow in Figure 8 is revised from Heckenberger 1996, 1998.

13. The ph values from the central plaza samples (more distant from the houses and middens) correspond well with the values obtained by Carneiro (1983:Table 3.3) for chemi- cally unaltered soils from Kuikuru gardens.

Received February 7, 1998; accepted October 26 , 1998; revised May 19, 1999.

Heckenberger et al 1999

Documents

Transcript of Heckenberger et al 1999