The focus on the floodplain as a differentiated area in the cultural history of the Amazon builds on a long and illustrious tradition in anthropology that divides the Amazon basin in two general ecological areas: the varzeas, or floodplains, and the terras firmes, or uplands (Lathrap 1968a; Meggers 1996; Moran 1993: 26; Roosevelt 1980). That tradition goes back at least to Erland Nordenskiold’s early synthesis of Amazonian archaeology (Nordenskiold 1930). In it, Nordenskiold called attention to the fact that, in lowland South America, the three major river basins-from north to south: Orinoco, Amazon and Parana (Figure 1.2)-are directly or indirectly connected, creating a network that would have allowed the movement of people throughout the continent. Nordenskiold’s insight holds some culture-historical truth (Heckenberger 2002), but most importantly, it laid a methodological foundation for differentiating between fluvial-oriented versus hinterland settlements in Amazonia, in terms of size, density and length of occupation. Fluvial in this case means those occupations related to first, second or even third order streams. Nordenskiold’s idea was subsequently used by Lowie (1948) in his definition of the tropical forest culture type. For Lowie, one of the key features of socio-political organization characteristic of the type was emphasis on fluvial occupation (Lowie 1948: 1). Finally, based on Nordenskiold and Lowie, as well as
On the work of cultural geographer Carl Sauer (1969: 21-23), Donald Lathrap was the first archaeologist to explicitly propose that human occupation of the Amazon would be directly conditioned by access to aquatic resources, mostly as regards demographic growth and the emergence of social inequality (Lathrap 1968a: 62). For Lathrap, drastic ecological differences between river or lake-oriented versus hinterland-oriented settlements, especially in access to animal protein, would have a direct influence on the duration, density and size of settlements in pre-colonial Amazonia. In this model, the abundance of fish, mammals, reptiles and waterfowl in fluvial settings would provide a stable and predictable source of food resources, providing conditions for long-term population growth and sedentary settlements. Hinterland occupations, on the other hand, would be limited to small size and short time span due to the relatively low availability and unpredictable nature of resources in those settings (Lathrap 1968b).
Following Lathrap’s work it became accepted among archaeologists and cultural anthropologists that in the Amazon, fluvial-oriented settlements would be consistently larger and occupied for much longer periods than hinterland settlements. The problem, however, is that not much research has been done on hinterland settings, due to the usual logistical problems inherent to working in the Amazon. Every once in a while, though, evidence from hinterland settings shows that, although valid at a general level, Lathrap’s hypothesis probably simplifies a more complex pattern. This is the case, for instance, in the geometric-shaped earthworks recently found in the upper Acre basin (Parssinnen, Ranzi, Saunaluoma and Siiriainen 2003) or in the upper Xingu basin in Brazil (Heckenberger 2005; Heckenberger et al. 2003). These are areas located far from the major channel of the Amazon, but it is clear that similar research in comparable places will show that important floodplain environments exist, even though far removed.
The term “Amazonian floodplain” is employed here to designate those areas placed immediately adjacent to first, second and third order rivers, always downstream from the rapids that mark the transition from the greater Amazonian plain to the Andes in the west, and to the Brazilian and Guiana plateaus in the south and north. They are settings that normally lie less than 100 masl, even when very far from the Atlantic Ocean. For instance, the city of Iquitos, in the Peruvian Amazon, which lies roughly 3,400 km from the mouth of the river, is only 100 masl. The Amazon itself forms nearby, where its two major upper tributaries-the Maranon and Ucayali rivers-meet after flowing northward from the Central Andes. From there the Amazon flows eastward until it reaches the Atlantic Ocean close to the city of Macapa, in Brazil.
It is common that rivers in the Amazon basin change their names when crossing national borders or even after joining other rivers. The Amazon River itself is a good example: it has this name when it flows through Peru and Colombia, after the Maranon and Ucayali rivers meet. Then, after it enters Brazil, the Amazon is known as Solimoes until it joins the Rio Negro, in central Amazonia. After that, it is known again as Amazon until its mouth. The same happens with the Putumayo-Iga, the Caqueta-Japura and the Gauinia-Negro. In the text, I will use these regional names.
The Amazon River and all of its large tributaries that have headwaters in or near the Andes of Bolivia, Peru, Ecuador and Colombia are white water rivers. These rivers contain sediment eroded during rainy seasons and sediment from snow-melt in mountainous catchment areas. The Andes are a recent geological formation. As a result, the floodplains of white-water rivers are typically fertile, although their agricultural cycle is limited by the length of the dry seasons and the unpredictability of the flood regimes (Meggers 1996). More important for human occupation, these rivers and floodplains support a rich wild life,
Including fish, mammals, fowl and reptiles that were a reliable source of protein for the people settled along and around them.
White water rivers are active, and as a result their channels are constantly being rebuilt by both erosion and deposition. One visible manifestation of that process is the development of ox-bow lakes whose different stages are in fact ancient stretches of river cut off from the main channel, to eventually become lakes and swamps. Lathrap (1968a) has shown how along the Ucayali River this pattern of erosional secession had important consequence for dating archaeological sites. Older sites are normally located away from the present-day course of the river, next to swamps or ox-bow lakes. The intense erosive action of rivers can also lead to the destruction of archaeological sites. One good example is the site of Miracanguera, located in the central Amazon, close to the city of Itacoatiara, which was identified and partially excavated by the botanist Barbosa Rodrigues in the 1880s (Barbosa Rodrigues 1892: 2). Forty years later, when visited by Curt Nimuendaju in the 1920s, the site had disappeared, having been carried away by the Amazon River (Nimuendaju 2004: 160-161).
Besides white water rivers, there are two other broad categories of rivers in the Amazon basin: clear water and black water rivers. Clear water rivers run from geologically old catchment areas, such as the Guiana Plateau in the north, and the central Brazilian Plateau in the south. Their sediment load is much smaller than white water rivers, and consequently their floodplains are comparatively less productive. Among them are the Uatuma, Nhamunda-Trombetas, Tapajos, Xingu and Tocantins rivers. Black water rivers are probably the most common in the Amazon basin. However, with the exception of the Rio Negro, black water rivers are normally small to medium in size, at least in Amazonian terms, being tributaries of larger, white or clear water rivers. Black water rivers have their catchment areas in geologically old settings, swamps or in areas of poor, sandy and well-drained soils. Their black color results from the dissolution of tannic acids from the decomposing litter on the ground. As a consequence, black water rivers are considerably less productive for humans than clear or white water rivers.
Discussion of the occupation of Amazonian floodplains presented here includes the alluvial plains of white, clear and black water rivers, although black and clear water rivers do not form classic varzeas, since the term applies only to the nutrient-rich floodplains of white water rivers (Moran 1993: 24-31). It also includes data from areas where the flood regime depends more on the annual pluvial cycle than on changes in river level, such as Marajo Island, located at the mouth of the Amazon, and the Llanos de Mojos in the Bolivian Amazon. The archaeology of the Orinoco River and the Guianas are not discussed at length here (see Chapters 13, 16, 17 and 23 in this volume).
World History
