The earliest signs of change in the patterns of social and political organization in the Amazonian floodplains are visible from ca. 2500 BC, with the first evidence of anthropogenic dark earths (ADEs) or terras pretas on the Jamari River, in the upper Madeira basin (Miller et al. 1992). The inception of the Ananatuba phase about 1400 BC marks the first sign of large sites on Marajo Island sites (Meggers and Danon 1988; Meggers and Evans 1957; Schaan 2004; Simoes 1969). A little later, from 700 BC to AD 400, in the Upano basin of the Ecuadorian Amazon, at the Sangay site, a group of artificial mounds forming anthropomorphic figures when seen from the air was built and occupied (Rostain 1999).
From what is currently known, however, these seem to be isolated phenomena since no corresponding developments are seen at the same time elsewhere in the Amazon. It is rather later, from around the beginning of the Christian era on, that a widespread and visible pattern of population growth, site aggregation and noticeably anthropogenic landscape
Changes become visible throughout the area (Neves and Petersen 2006; Petersen et al. 2001). These changes are matched, in the archaeological record, by the sudden appearance, at different times and places, of large sites with deep stratified ceramic deposits associated with anthropogenic dark soils (Kern et al 2003; Neves et al. 2003, 2005; Petersen et al. 2001); artificial earthworks (Parssinnen, Ranzi, Saunaluoma and Siiriainen 2003) such as raised fields and causeways (Denevan 1966, Erickson 2000); large villages surrounded by moats and connected by road networks (Heckenberger 2005; Heckenberger et al 2003); artificial residential and funerary mounds associated with elaborate pottery (Meggers and Evans 1957; Roosevelt 1991a, 1996; Schaan 2001a, 2004); quasi-urban settlement systems also are associated with elaborate pottery, polished stone statuettes and long-ranging trade networks (Gomes 2002; Nimuendaju 2004; Roosevelt 1999); and the construction of circular megalithic structures (Nimuendaju 2004).
Most of these developments are organized into a cultural chronology first outlined almost fifty years ago. It divided Amazonian ceramics into four wide horizons, from older to younger, based on their decoration and paste: Zoned-Hachured, Incised Rim, Polychrome and Incised-Punctated (Meggers and Evans 1961, 1983). The cultural historical reconstitution embedded in this chronology is no longer valid, since it proposed that all of the horizons had their origin outside of the Amazon, either in the Andes or in northern South America, which is not the case. The chronology has also been revised, given that it originally proposed a shallow span for each of the horizons, which was later proven to be wrong, and hence the horizons are now classified as traditions (Meggers and Evans 1983; Roosevelt 1991; Schaan 2001b: 157). Finally, further work has shown that there are earlier complexes not identified in the early 1960s, such as the Mina phase ceramic-bearing shell mounds of the Atlantic shore (Roosevelt 1995; Simoes 1981), and even earlier ceramics from the lower Amazon, in the sites of Taperinha and Pedra Pintada cave (Roosevelt 1995; Roosevelt et al. 1991). Now these earlier components must be added to the chronology. These problems notwithstanding, the backbone of the sequence retains utility warranting continued use (Lima et al. 2006). Ideally it would be interesting if data on ceramic chronology could be matched by more information on things such as settlement patterns, site occupation chronology, intra-site spatial patterns and so forth, but this is not available in the vast majority of cases (Schaan 2005). That was the case, for instance, of the Manacapuru phase in the central Amazon, for which an earlier component was recently recognized, the Agutuba phase, based on chronological, contextual and formal patterns (Lima et al. 2006). Again, criticism of previous work needs to be tempered by the recognition that field work in the Amazon is complex and expensive: distances are vast and in many cases sites are hidden under the forest or bush.
The earliest ceramics of the New World appear to have been produced in the Amazon (Roosevelt 1995; Roosevelt et al. 1991). The evidence comes from the Pedra Pintada cave, located near the city of Monte Alegre in the lower Amazon, where 8000-year-old ceramics were recovered (Roosevelt 1995). From the Taperinha shell mound, located across the Amazon River from Pedra Pintada, 7000-year-old ceramics were also found (Roosevelt 1995). Beyond the Taperinha shell mound, dates for early ceramics were also obtained from other shell mounds in the lower Amazon, the estuary zone and the Guianese shore: on the Xingu river (Perota and Botelho 1992), the Jauari River (Hilbert 1968), along the Atlantic shore east of the mouth of the Amazon (Simoes 1981) and in Western Guiana (Williams 1997: 344). This pattern indicates an association between the early pottery and fluvial and maritime shell mounds of Amazonia (Roosevelt 1995). The problem is that these early ceramic complexes are different from each other: at Pedra Pintada and Taperinha one sees
Sand temper and plastic decoration, whereas in the early occupations at the Jauari site (Castalia phase), the lower Xingu and the Atlantic shore (Mina phase), and Western Guiana (Hosororo creek - Mina phase) plastic decoration is absent and shell is the tempering material. If the dates and dated contexts are correct, these differences might indicate two distinct early pottery-making traditions in the Amazon: one, more localized and without further developments, was characterized by the Mina phase ceramics found in both riverine and maritime shell mounds. The other, with an emphasis on plastic decoration, shares common features with other known early complexes in northern South America (Meggers 1997).
The presence of two distinct unrelated early ceramic complexes in the lower Amazon could explain the large formal and technological differences between Mina and Zone-Hachured ceramics. At the Jauari site, a shell mound located on the lower Amazon, Hilbert (1968) found two distinct ceramic assemblages: one of them, belonging to the Castalia phase and composed by shell tempered bowls, can be associated with the Mina phase, whereas the other, belonging to the Jauari phase, has decorative patterns and cauixi (a fresh water sponge) temper, which is characteristic of the Zone-Hachured Tradition. Even if there was a connection between earlier Taperinha and later Mina and Castalia phase ceramics, one is dealing here with at least a 1,500-year-long gap between them. Jauari phase sites have not been dated, but other Zoned-Hachured sites belonging to the Ananatuba phase have. These are sites located on Marajo island, dating back to 1400 BC (Meggers and Danon 1988; Schaan 2001a; Simoes 1969). Of course, the gap may result from poor archaeological visibility and lack of fieldwork, but the sudden appearance of Zone-Hachured sites after a long chronological hiatus seems not to be restricted to the lower Amazon and Marajo Island alone.
About the first millennium AD, ceramic-bearing sites seem to appear rather suddenly after long hiatuses in regional sequences (Herrera, Bray and McEwan 1980-81; Hilbert and Hilbert 1980; Lima et al. 2006). The changes visible in the archaeological record from the beginning of the Christian era onwards cannot be associated with any of the ceramic traditions discussed here. Indeed, during most of the first millennium AD one sees throughout the Amazon, at least from the point of view of ceramics, a picture of cultural diversity marked by the simultaneous development of distinct phases or traditions in different places. At Marajo Island for instance, mound building is clearly associated with the Marajoara Phase (Polychrome Tradition), which lasted from the fourth to the thirteenth centuries AD (Meggers and Evans 1957; Roosevelt 1991a; Schaan 2004). Around the same time, in the central Amazon, one sees, from the fifth century AD on, the formation of large sites from the Manacapuru Phase (Incised Rim Tradition) associated with anthropogenic dark earths (ADEs) (Neves et al. 2003; Petersen et al. 2001). These are very fertile soils, highly sought after by current Amazonian farmers, that result from both unintentional and intentional management in the pre-Hispanic past (Glaser et al. 2004; Lehmann et al. 2003; Smith 1980; Woods and McCann 1999). In the central Amazon, ADEs are interpreted as the result of input into the soil of slow burn charcoal and organic waste from household activities in a context of sedentary occupations. They can, therefore, be seen as an indicator of social and demographic change, marking the inception of long-term occupation or of large settlements in the area (Neves and Petersen 2006; Petersen et al. 2001). A similar pattern seems to have occurred with other Incised Rim Tradition sites of the Japura-Caqueta river, associated with the Japura phase of Brazil and Colombia, dated from AD 600 to 800 (Herrera, Bray and McEwan 1980-81; Hilbert 1968).
Since the early 1960s there has been a strong debate on the chronology, origins, spread and ethnic meaning of sites with Incised-Rim ceramics. In the initial formulation
These were supposed to derive from older complexes in northern South America (Meggers and Evans 1961), reinforced by the identification, in the lower Magdalena River of Colombia, of sites with ceramics decorated by incision and modeling with dates going back to ca. 4000 BC (Meggers 1997; Oyuela-Caycedo 1995; Oyuela-Caycedo and Bonzani 2005). This hypothesis was criticized by scholars working in different parts of lowland South America, for whom Incised Rim ceramics represented an ancient and local development in the central Amazon (Lathrap and Oliver 1987; Rouse 1985, 1992). The proponents argued that populations expanded from the central Amazon due to demographic growth based on an effective economy combining agriculture and the exploitation of aquatic fauna. Pottery-making groups moved up the Rio Negro and down the Orinoco River towards the Caribbean mainland. The resulting archaeological manifestation of this process is seen in the Barrancoid series sites along the lower and middle Orinoco and western Guiana (Lathrap 1964, 1970a, b; Lathrap and Oliver 1987). In this formulation, the Barrancoid series would be a derivation from earlier Incised Rim complexes from the central Amazon. However, the available data do not support this claim. Barrancoid sites in the lower Orinoco are consistently older than Incised Rim sites along the Amazonian floodplain (Barse 2000; Boomert 2000; Gasson 2002; Hilbert 1968; Lima et al. 2006). On the other hand, the similarity among some Barrancoid and Incised Rim ceramics are too strong to be overlooked (Boomert 2000; Evans and Meggers 1968; Hilbert 1968). These include single and double line parallel incisions and modeled decoration around the rims or on labial flanges on the vessels. Maybe the best way to account for this is to accept Lathrap’s identification of a connection between Barrancoid and Incised Rim ceramics, but reject his historical hypothesis about its central Amazonian origin, and instead accept Meggers’ (1997) hypothesis that early Amazonian and lower Orinocan complexes derive from an initial center of ceramic production in northern Colombia.
Archaeological research in the central Amazon is at last revealing information about historical changes that go beyond ceramic chronology. For example, from the seventh to the eleventh centuries AD, ring-shaped sites associated with ceramics belonging to the Paredao Phase appear (Donatti 2003; Moraes 2006; Neves 2003). Ring villages in lowland South America are ethnographically associated with the Ge speaking peoples from the central Brazilian Plateau (Wust and Barreto 1998) or with the first Arawak speakers in the Caribbean (Heckenberger 2004; Petersen 1996), but, with the exception of Paredao Phase sites, they are unknown along the Amazonian floodplain (Myers 1973). In the second half of the first millennium AD one also sees the establishment of long lasting regional systems in the central Amazon, based on the presence of Manacapuru ware in contemporary Paredao sites and vice-versa (Donatti 2003; Moraes 2006), and interaction that could have been based on trade and marriage, as described ethnographically among today’s Tukanoan Indians of the northwestern Amazonia (Jackson 1983).
If, in the central Amazon, the first millennium AD was characterized, on one hand, by the flourishing of distinct local cultural traditions, and on the other, by a concomitant pattern of population growth and increase in sedentary lifestyles throughout the Amazon, the advent of the second millennium AD marked profound changes that are clearly visible in the archaeological record. Those changes relate to a process that can be interpreted as the emergence of a wide cultural patterning characterized by a replacement of sites of the Incised Rim and other local traditions by sites of the Polychrome Tradition, and also by the gradual expansion of the Polychrome Tradition over a vast area, from the lower Amazon almost to the Andean piedmont in Colombia, Ecuador and Peru. The ethnic and political processes behind the Polychrome expansion are not clear. Brochado
(1984) interpreted the archaeological record as the correlate of agricultural and demographic expansion of Tupi speaking populations over the floodplains of the Amazon and its major tributaries, from a central Amazonian origin center. Meggers (Meggers et al. 1988), on the other hand, has noticed how most of the sites of the Polychrome Tradition tend to be found downstream of the rapids that mark the geological transition from the Amazonian floodplain and the surrounding Guiana and central Brazilian plateaus to the south and north.
Polychrome ceramics share a series of common features that render them distinct. Among them there is the use of painted decoration with motifs in red and/or black over a white slip. Plastic decorative techniques, such as incision, excision, grooving and modeling are also found. In the central and upper Amazon, polychrome vessels tend to be tempered with crushed and burned tree barks (caraipe) whereas on Marajo Island grog is the preferred antiplastic. Study of the timing and direction of the polychrome expansion has been a focus of research since the 1950s. Initially it was proposed that it had an Andean or circum-Caribbean origin (Evans and Meggers 1968; Meggers and Evans 1957). As better chronologies became available the hypothesis of an external origin was abandoned and a central Amazonian origin was proposed (Brochado 1984; Lathrap 1970a; Lathrap and Oliver 1987; Oliver 1989). However, the hypothesis of a central Amazonian origin is not supported by the available chronologies for that area (Heckenberger et al. 1998; Hilbert 1968; Neves 2003). Along the main channel of the Amazon, the earlier polychrome sites are related to the Marajora phase, found on Marajo Island and the adjacent mainland, with dates going back to the fifth century AD (Meggers and Danon 1988: 248; Roosevelt 1991a: 313-314; Schaan 2001b: 157), but it is only after AD 750 that dates are more frequent and display a smaller standard deviation (Boomert 2004: 259). This provisional picture may change if earlier dates are confirmed for the upper Madeira basin where polychrome ware, related to the so-called Jamari or Jatuarana Phase, has been (briefly) reported with dates clustering around the beginning of the Christian era (Miller et al. 1992). If these dates are confirmed, we will have to infer a southern Amazonian connection between the upper Madeira and the Marajo Island sites.
By the twelfth and thirteenth centuries AD, most of the floodplains of the Amazon/ Solimoes were occupied by villages of different size where polychrome ware was produced. The available data show a clear pattern in the dates: older in the upper Madeira, old in Marajo Island and consistently more recent as one moves upstream from the lower to the upper Amazon (Brochado and Lathrap 1982; Evans and Meggers 1968; Heckenberger et al. 1998; Herrera, Bray and McEwan 1980-81; Hilbert 1968; Meggers and Evans 1983; Neves 2003; Neves and Petersen 2006; Schaan 2001b, 2004; Simoes 1974; Simoes and Kalkmann 1987; Simoes and Lopes 1987). Can the polychrome expansion be correlated with a single ethnic or linguistic component as proposed by Lathrap, Brochado and Oliver? There is no single answer, but it appears that by the late 1400s Amazonian social formations were multi-ethnic (Hornborg 2005; Whitehead 1994), similar to what one sees today in such different areas as the upper Xingu (Heckenberger 2005) and the northwest Amazon (Chernela 1983; C. Hugh-Jones 1979; Jackson 1983). In that sense, one would not find a simple correlation between the polychrome expansion and a single linguistic group in late precolonial times. On the other hand, there is mounting evidence across the world that correlates demographic expansion - and the corresponding expansion of material culture and genes - with the advent of agriculture (Bellwood 2001). From this comparative perspective, the match among variables is stronger in cases of expansions into areas previously unoccupied, or in situations where the expanding population brings with it a different
Technology enabling the exploitation of new niches as, for instance, agriculture in areas previously occupied by hunter-gatherers (Renfrew 2000).
If the above hypothesis is valid, then a correlation between early polychrome expansion and the expansion of agriculture can be postulated; the earliest dates for polychrome sites come from the upper Madeira basin, the same area that genetic evidence suggests as the center for manioc domestication, Manihot esculenta (Olsen and Schaal 1999) as well as the peach palm, Bactris gasipaes (Clement 1999). Indeed those initial polychrome sites of the upper Madeira have the earliest ADE sites currently known in the Amazon. If ADEs, as proposed above, are formed in contexts of sedentary occupation, being therefore markers of social and economic change, and also since, together with polychrome ware, the earliest ADEs are found in the upper Madeira, it can be argued that early polychrome expansion is correlated with the expansion of manioc and peach palm farming, by Tupi speaking populations from the upper Madeira basin beginning about 2,500 years ago.
The problem is that, unlike sub-saharan Africa, Europe or Polynesia, there was not one but at least two major waves of linguistic expansion in the Amazon: one of Arawak and the other of Tupi-Guarani families. If the Bellwood-Renfrew hypothesis is correct, one can expect that these different expansions result from the expansion of two different sets of crop complexes. In the same way, the apparently late establishment of sedentary agricultural life across the Amazon indicates that, although food production systems in tropical northern South America area may date from the early Holocene (Gnecco and Aceituno 2004), the transition to a full dependence on agriculture was much later, dating to the beginning of the Christian era (Piperno and Pearsall 1996: 8).
The acceptance of a southwestern origin for the Polychrome Tradition may also help resolve another puzzle of Amazonian archaeology: the fact that there are no signs of polychrome sites in the lower Tapajos, Nhamunda or Trombetas rivers, an area where most of the known sites have ceramics of the Incised-Punctated Tradition. They are highly elaborated ceramics decorated by painting and, most notably, by modeling with zoomorphic and anthropomorphic motifs as well as incision (Gomes 2002, 2005; Nimuendaju 1949, 2004; Roosevelt 1999). The preferred temper is cauixi. These sites are associated with a small but significant corpus of anthropozoomorphic stone statuettes (McEwan, Barreto and Neves 2001; Nimuendaju 2004; Nordenskiold 1930) and little anthropomorphic or zoomorphic lithic amulets known as muiraquitas that have a wide distribution through the Amazon and northern South America (Boomert 1987). Incised-Punctated sites can be quite large, as it is the case of the site now under the modern city of Santarem (Roosevelt 1999). In the lower Tapajos river, Woods and McCann (1999) found archaeological sites with ADEs surrounded by fairly large areas of anthropogenic soils not associated with cultural remains. These areas, called “terras mulatas” are interpreted as records of former agricultural intensification and soil management in a context of large, sedentary occupations (Woods and McCann 1999).
The area of largest density of sites with incised-puntacted ceramics lies between the modern cities of Parintins and Santarem. Within that area, at least two distinct complexes can be identified: to the east are sites with so-called Tapajos ceramics, clustered around Santarem; to the west there are sites with Konduri ceramics, clustered around the Trombetas river. In the chronology of Meggers and Evans (1961), the Incised-Punctated Tradition was classified as recent, appearing after the Polychrome Tradition. Of course, following the pioneering research by Curt Nimuendaju in the 1920s (Nimuendaju 2004), a surprisingly small amount of fieldwork has been done in the area (Gomes 2002, 2005; Hilbert 1968; Hilbert and Hilbert 1980; Roosevelt 1999; Roosevelt et al. 1991, 1996). However,
The available chronological data indicate that Meggers and Evans’ scheme is basically valid on this matter, since the inception of the Incised-Punctated Tradition dates from the end of the first millennium AD, that is, later than the earliest polychrome sites elsewhere in the Amazon (Gomes 2005; Roosevelt 1999). Polychrome sites, as already mentioned, have a wide distribution throughout the Amazonian floodplain, from Marajo island to the foothills of the Andes. The only place along the floodplains where sites of the tradition are not found is precisely the Santarem area. There, incised punctuated sites overlap older occupations from the Poco phase, dating from 100 BC to AD 200 (Hilbert and Hilbert 1980), and not polychrome sites as it would be expected.
As with many other matters in Amazonian archaeology, there are no clear data on the early history of the Incised-Punctated Tradition. These ceramics share a basic pattern of incised and modeled decoration found in the earliest complexes throughout northern South America, going back to the pottery of San Jacinto 1, Puerto Hormiga and Valdivia (Meggers 1997; Oyuela-Caycedo 1995; Oyuela-Caycedo and Bonzani 2005), the Malambo Tradition of northern Colombia (Angulo 1981), the Incised Rim Tradition of the Amazon and the Barrancoid series of the Orinoco River (Gasson 2002). They do not represent a rupture with these earlier complexes. The closest similarities with the Incised-Punctated Tradition are found with the ceramics of the Arauquinoid series of the middle Orinoco (Zucchi 1985), coastal Surinam and coastal French Guiana (Rostain and Versteeg 2004: 239). In the middle Orinoco, cauixi-tempered ceramics of the Araquinoid series date from AD 400 to 1400 (Zucchi 1985). In the Guianese coastal plain, the dates are a little later, starting around AD 600 but also continuing until the arrival of the Europeans (Rostain and Versteeg 2004).
The Santarem area, the middle Orinoco and the coastal plain of Surinam and French Guiana lie roughly equidistant from the Guiana plateau, a region predominantly occupied by Carib-speaking groups today. This has led authors to suggest that both the Arauquinoid series and the Incised-Punctated Tradition are local manifestations, from the late first millennium AD onwards, of a radiation of Caribs towards the Guyanese coast, the middle Orinoco and the lower Amazon (Brochado and Lathrap 1982; Zucchi 1985). Along the Amazonian floodplain, Incised-Punctated ceramics are restricted to a radius of roughly 300 km with its center at Santarem, and being surrounded both upstream and downstream by occupations related to the Polychrome Tradition. To the south of Santarem, sites of the Incised-Punctated Tradition are not known. This, together with the alleged similarities with the Arauquinoid series of the middle Orinoco River and Guianas, matches the hypothesis of a northern origin in the Guyana plateau correlated with the Carib expansion. As with other parts of the Amazonian floodplain, the Santarem area was densely occupied at the onset of European colonization. The relative wealth of reports about this area and the Tapajo Indians who occupied it indicate some measure of political centralization, including roads, settlement hierarchy, labor specialization and the presence of a nobility (Nimuendaju 1949, 2004; Roosevelt 1999). Signs of political centralization are, however, absent in the Arauquinoid sites of the Guianese coastal plain, where archaeological sites are small and monumental architecture is lacking (Rostain and Versteeg 2004: 239). In the middle Orinoco, on the other hand, between AD 600 and 800 there was an increase in the size and density of archaeological sites of the Araquinoid series probably related to the adoption of maize cultivation (Zucchi 1985: 33). Maize agriculture has also been described by early European chroniclers in the Santarem area. Although there are no direct paleobotanical data to confirm it, research done with terras pretas in the Santarem area has uncovered large extensions of anthropogenic soils, in one case with 120 ha, which could
Have been used for agriculture (Woods and McCann 1999: 12). As with other areas, more research is needed with Incised-Punctated sites in order to determine the extent to which the archaeological record matches early colonial reports.
The Tapajo Indians, settled along the major Amazonian floodplain, were easily accessible to Europeans traveling up and down the river, but in more remote areas the impact of European colonization was strong as well. In the upper Xingu basin, more than 1,000 km south of the Amazon floodplain, the archaeological record shows signs of population decrease and settlement abandonment starting in the sixteenth century AD, and most likely related to the indirect spread of infectious diseases (Heckenberger 2005: 74). These declines were preceded by several centuries of population growth and settlement accretion whose signs became visible after ca. AD 800 (Heckenberger 2005: 87-88, 103). Between AD 1250 and 1350 large earthworks were built, including ditches surrounding villages, sometimes paired settlements, and linear mounds placed at the margins of plazas or alongside causeways (Heckenberger 2005: 78).
From these features it can be inferred that there was a degree of social complexity within the communities.
At the mouth of the Amazon, the establishment of European outposts such as Belem and Cayenne also had a strong impact in what seems to have been one of the most culturally diverse areas of pre-colonial Amazonia. This diversity is attested by the flourishing of several distinctive, although correlated, ceramic complexes in an area ranging from Marajo Island towards what is currently eastern Surinam beginning in the fifth century AD (Guap-indaia 2001). Associated with these complexes one finds cave burials in zoomorphic and anthropomorphic urns in the Maraca area (Guapindaia 2001), artificial burial chambers, stone alignments and also cave burials related to Ariste pottery (Nimuendaju 2004; Meggers and Evans 1957; Chapter 16 in this volume). The people who made these ceramics were probably the ancestors of the Arawak-speaking Palikur and Lokono Indians, who currently live at the edge of the coastal plain in Surinam, French Guiana and the northern Amapa state in Brazil.
World History
