Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

4 Management of oligotrophic areas

There is much we can learn about the proper management of Amazonian ecosystems from the region's indigenous population, which coped with its limitations and in some cases rose above them by changing the composition of forest and soil. What we conclude from our examination of these two ecosystems is that native populations used both areas, but that they restricted their manipulations of the ecosystem to the "islands" of terra firme forest in the oligotrophic regions and focused on the eutrophic "islands" within the vaster nutrient-poor areas of terra firme.

Human populations have inhabited the upper Rio Negro for at least 6,000 years. Ceramics and anthropogenic soils have been identified in the Rio Negro, dated at 3750 BP (Clark and Uhl, 1987: 7; Saldarriaga and West, 1986; Sanford et al., 1985). The anthropogenic soils are found not in the caatinga areas but in upland tropical forests growing on oxisols, suggesting a very ancient preference for the patches of tropical forest, rather than caatinga, for occupation and agriculture.

Populations of the upper Rio Negro practise slash-and-burn agriculture, clearing areas of between 0.5 and 2.0 hectares between September and November each year. It has been observed that contemporary native populations generally avoid locating their swiddens in areas of caatinga, which through ethno-ecological taxonomies are identified as inappropriate for agriculture (Hill and Moran, 1983). Clark and Uhl (1987) estimated that in the region near San Carlos de Rio Negro, in Venezuela, only about 20 per cent of the soils were not spodosols. This extremely restricted availability of soils capable of supporting crops for even a couple of years runs counter to very old assumptions about the ease with which native populations could relocate in Amazonia and the lack of "environmental circumscription" in the basin (Carneiro, 1970). There have been no reports of nutrient-rich alfisols (terra roxa estruturada eutrófica) in the Rio Negro Basin. Galvão (1959: 24) noted another factor limiting the agricultural potential of the upper Rio Negro: the apparent avoidance of areas which required penetration into the forest and a preference for areas along river banks by horticultural populations. Chernela (1983) observed a similar preference among the Uanano, a Tukanoan population in the Vaupés, as have those who have studied the Bara Makú (Silverwood-Cope, 1972) and the Hupdu Makú (Reid, 1979). This tendency to limit territorial occupation to the river-banks may represent a compromise response to the poverty of the terrestrial ecosystem, the availability of restricted areas of fertile levees, and the importance of fisheries based on ancient territorial claims.

Similarly, flooded forests (igapós) are avoided for agriculture, given their importance for fisheries (Chernela, 1983; Clark and Uhl, 1987; Dufour, 1983). Many of the fish in Amazonian river channels enter the flooded forest during the rainy season to gain weight and to spawn (Goulding, 1980, 1981). When fishing gives poor results, the Wakuenai in Venezuela say that the fish are spawning and locate their swiddens distant from the flooded forests. Chernela (1982, 1986a) noted the same explicit avoidance of flooded forest for agriculture among the Uanano in Brazil.

The length of the cultivation period and the size of clearings is of special significance in these oligotrophic habitats. The smaller the area cleared, the easier it will be for seeds from the native vegetation to recolonize the area. The length of the cultivation period affects the levels of nutrients available to the incoming seeds and the growth rate of secondary vegetation. In black-water ecosystems, the return of the original vegetation may take over a hundred years (Uhf, 1983; Uhl et al., 1982). Uhl et al. (1982: 319) found 271 seeds per m² in an area studied at the end of the cultivation cycle, 90 per cent of them secondary successional species. It is quite likely that secondary successional species are better adapted to low levels of nutrients and can thrive where domesticated plants cannot. Succession is much slower in black-water ecosystems: above-ground biomass after three years was only 870 g/m² compared to 2,000 g/m² in areas of oxisols on upland forest. It appears that oligotrophy, as well as the flooding, are responsible for this lower level of above-ground biomass production (Uhf et al., 1982: 320). After sixty years, the above-ground biomass is only 40 per cent of that of the original vegetation (Clark and Uhl, 1987: 12; Jordan and Uhl, 1978), compared to 90 per cent of above-ground biomass in eight years on ultisols in the Peruvian Amazon (Sanchez, 1976: 351).

One of the ways in which native peoples help accelerate the recolonization of cleared areas is by planting fruit trees in the swiddens. Not only does this increase the utility of the land, but it serves to attract birds and bats, which are the principal agents of primary forest seed dispersal in the humid tropics. In a controlled experiment, areas planted in this manner had nine times the number of seeds of native trees than an area which was not planted with fruit trees at the end of the cultivation cycle. The shade of the fruit trees serves to provide the needed shade to primary tree species of slow growth, and reduces leaching of nutrients.

The lack of a marked dry season in this region would lead us to expect that burns would be of poor quality. Generally, it is the quality of the burn that determines the yield of slash-and-burn cultivation in non-volcanic areas of the humid tropics (Moran, 1981). That is clearly the case in areas with high above-ground biomass. But in areas with lower biomass, high insolation, and high albedo resulting from reflection from the white sands, the biomass dries sufficiently to burn so well that, in fact, areas of xeromorphic vegetation tend to experience burns beyond the areas cleared. Clark and Uhl (1987) documented the problem of natural burns in this habitat, where hundreds of hectares can catch fire when twenty rainless days occur. The destructive impact of fire is a real threat in this ecosystem, in contrast to other areas of the humid tropics, where fire rarely extends beyond the area cleared (or in areas scarred by logging activities). Planting follows, and is dominated by bitter manioc.

Dufour (1988) found more than a hundred varieties of bitter manioc among a Tukanoan population in the Colombian Vaupés. Very few varieties of sweet manioc were known and cultivated. Chernela (1986b) found that repeated efforts by Uanano peoples in the Vaupés of Brazil to introduce sweet cultivars of manioc failed and only bitter cultivars persisted, when she did a study of their manioc varieties. Bitter manioc resembles the native vegetation of blackwater regions by its toxic quality, which serves to conserve nutrients for the plant through reduction of herbivory. Montagnini and Jordan (1983) found that insects consumed less than 3 per cent of the tissue of bitter manioc plants due to the cyanogenetic glucosides present.

Bitter manioc cultivation solves one of the great problems of Amazonian populations: how to cultivate soils that are extremely poor in nutrients, extremely acid, and have toxic levels of aluminium. Manioc, a plant that appears to have evolved in just such areas of South America, can produce impressive yields in areas where nothing else will grow (Moran, 1973). One of the few limitations to its cultivation is its inability to withstand waterlogging, which explains why it is cultivated on higher ground. Manioc is even adapted to drought, during which it loses its leaves and goes into dormancy, gaining its leaves again with the return of soil moisture. Beans and corn, by contrast, are unable to produce a predictable crop in these nutrient-poor areas, and unable to cope with even short-term droughts. Galvão (1959: 24) noted that corn had been abandoned by the populations of the Içana and that it probably never had much importance in the Rio Negro Basin.

Bitter manioc produces the bulk of the calories for black-water basin populations. Dufour (1983) showed that among the Tukancans she studied in Colombia, 70 per cent of the energy came in the form of manioc flour and manioc bread (casabe or beijú), tapioca, manioc beer, and other forms of prepared manioc. The energy efficiency of manioc is impressive: it yielded 15.2 calories for every calorie spent on its production. Seventy percent of the production costs occur during processing. Nevertheless, the oligotrophy of the black-water regions depresses the total yields. Yields vary between 3 and 8 tons per hectare, with a mean of 4.7 (Clark and Uhl, 1987). By comparison, the world mean is 8.4 tons per hectare, and reaches 12.7 tons in Brazil. The relatively low mean harvests confirm the nutrient-poor conditions of the environment.

Clark and Uhl (1987: 19) estimated the fish productivity of the Rio NegroCasiquiare-Guaima rivers to be between 6.6 and 13.2 kg/ha/ year - one of the lowest values for any tropical basin. In Africa and Asia the mean values are 40-60 kg/ha/year and Goulding (1979) estimated the productivity of the Madeira River in Brazil at 52 kg/ha/year.

The lower mean productivity results, in large part, from the absence of some particularly large species like pirarucú or paiche (Arapaima gigas), aruanã (Osteoglossum bicirrhosum), several large species of the genus Colossoma, and several large catfishes or pimeloids (Goulding, 1979: 15). The absence of aquatic grasses in black-water rivers, an important food source for many of the larger fishes, influences the species composition of these rivers and favours smaller species. This is not to say that black-water rivers are species-poor. Goulding et al. (1988) have shown that the lower Rio Negro is among the most species-rich rivers in the world, with approximately seven hundred species. Fish species in the Rio Negro, however, are dominated by smaller species with a mean length of only 40 mm with over one hundred species of less than 30 mm length (Goulding et al., 1988: 109).

Success in fishing in black-water ecosystems depends mostly on territorial control over cataracts and flooded forest, where the fish volume is greatest. The most successful method of fishing is through the use of large fixed traps which require considerable investment in their construction and maintenance. The importance of control over the best fishing spots is evident in such a situation, and reaches a considerable level of sophistication in the Rio Negro (Chernela, 1986a; Moran, 1990). Fish was found to be a part of meals in 78-88 per cent of cases sampled by Dufour (1987: 389).

5 Management of eutrophic areas

By contrast, management of the eutrophic areas of terra firme offers the challenge of diversity. This was handled by native populations by creation of "islands of resources." They concentrated species of economic interest within close proximity to soils favourable for horticulture and favourable in locational factors, such as ease of river transportation, well-drained areas for habitational sites, and cool breezes to provide comfort and freedom from insects.

The creation of anthropogenic forests and soils solves two of the commonly mentioned problems posed by Amazonia:poor soils and species dispersal. This process need not have been conscious. The behaviour of ancestors could very well have been certified in myths and descendants continued traditions found to be favourable to their lifestyle.

For agriculture, the preferred areas were the vine forests, because of their association to alfisols. Near the villages, the population concentrated further resources by creating palm forests, bamboo forests, food/fruit-rich forests and the garbage from their settlements built up over time anthropogenic soils which would serve in the future as usable soils in previously inferior areas from an agronomic viewpoint. This is equivalent to the population creating ecotones, or transitional areas, to favour high productivity.

In eutrophic areas, cultivation of grains and other nutrient-demanding plants is possible, although this will shorten the time before fertilisation will be required. On the other hand, it makes no sense to build cities on top of these superb soils (as has been done in recent years, for example, at Tucumã, Pará) or to put them into the production of export and tree crops (as was the case in Rondônia, with cacao). Agricultural zoning of these areas will need to achieve particularly high levels of sophistication, matching the quality of the soils to the products developed. Unfortunately, much of this will have more to do with market forces than with the micro-ecological characteristics of the areas.

The eutrophic areas of terra firme can be put to intensive use. Doing so should begin with attention to the processes of environmental modification undertaken by indigenous populations, and build on these efforts. The better soils can withstand cultivation for prolonged periods of time because of their fine structure and high initial cation exchange capacity. However, fertilisation will be required for any form of continuous and intensive cultivation. The viability of continuous cultivation in these areas of terra firme is not an environmental issue per se but, rather, a political and economic one, of the government providing the conditions for access to fertilizers at competitive prices and of matching market demand to the products supplied by the areas put under cultivation. So far, for most of Amazonia, this has been poorly done. One of the few success stories has been the Japanese colony at the favourably located Tome-Açú, in which the inhabitants identified a low-weight/high-value crop early on, and which invested in technical and human resources from the outset.

6 Conclusions

Black-water ecosystems such as those found in the Rio Negro, Vaupés, and Içana represent the poorest and most limited areas of Amazonia. Aboveground biomass is lower, litterfall is poor in nutrients, decomposition is slow due to the acidity of the soils and the water, and the drainage limitations resulting from the podzolic soils contribute to making these areas a true challenge to human populations.

Selection of particularly toxic varieties of bitter manioc and hierarchical control over fishing spots further adjusted native peoples to these poor areas of Amazonia. The areas are rich in endemic species and in toxic plants with secondary compounds with great promise for advances in medicine and pharmacology. More than any others they should be protected from largescale development activities. These are important ethnic refuge areas, with low potential for intensification but high potential for specialized and more selective exploitation for genetic, pharmacological, and medical research. While the total biomass is less than elsewhere in Amazonia, the species diversity is as high, if not higher, than other phytogeographic areas of the Amazon Basin. The native populations of this region are particularly attuned to the limitations of this ecosystem and their expertise should guide land use and research in the years ahead. Their potential for medicine and pharmacology, in particular, should have highest priority, especially the caatinga and bane areas, which are too poor for most other economic uses. On the "islands" of terra firme forests, extensive forms of resource use and extraction may continue, but intensive use is out of the question except in very restricted areas.

By contrast, intensive agriculture in Amazonia is possible and should largely be restricted to the eutrophic areas, especially where there are anthropogenic forests and soils. These have demonstrated productivity and, under the right market conditions, are likely to be economical. In contrast to more virgin areas, use of these areas will have minimal significance in terms of species extinction and biodiversity, since these areas are already simplified versions of the original ecosystems, with a predominance of species of economic value. This strategy would, in turn, recommend that for the foreseeable future large areas of the Amazon be protected from further clearing until such time as, through research and conservation, their potential can be better known.


The author wishes to thank the John Simon Guggenheim Memorial Foundation for a Fellowship during 1989-1990, which permitted him to take the time to reflect upon some of the matters discussed in this paper. He also wishes to thank the Institute for Advanced Study at Indiana University for the space in which to work during the Fellowship year; and Prof. Toshie Nishizawa for the invitation to participate in the stimulating meeting at Tsukuba.


Anderson, A., and E. Anderson. 1985. "A 'tree of life' grows in Brazil." Natural History 94: 41-6.

Anderson, A., and D. Posey. 1985. "Manejo de cerrado pelos Indios Kayapó." Boletim do Museu Paraense Emílio Goeldi, Botánica 2 ¹: 77-98.

Andrade, E.B. 1983. Relatório da expedição para coleita de germoplasma de "caiaui," Elenis oleifera, na Amazonia Brasileira. Biblioteca da EMBRAPA/CPATU (Empresa Brasileira de Pesquisa Agropecuaria/Centro de Pesquisa Agropecuaria do Trópico Umido), Belém.

Araújo-Costa,, F. 1983. "Projeto Baixo Tocantins: Salvamento arqueológico na regiao de Tucuruí." Dissertação de Mestrado, USP.

Balée, William. 1984. "The ecology of ancient Tupi warfare." In R.B. Ferguson (ed.), Warfare, culture and environment (Academic Press, New York), 241-65.

. 1989. "The culture of Amazonian forests." Advances in Economic Botany 7: 1-21.

Balick, M.J. 1984. "Ethnobotany of palms in the neotropics." Advances in Economic Botany 1: 9-23.

Black, G.A., T. Dobzhansky, and C. Pavan. 1950. "Some attempts to estimate species diversity and population density of trees in Amazonian forests." Botanical Gazette 111: 413-25.

Boer, J.G. 1965. "Palmae." In J. Lanjouw (ed.), Flora of Suriname, vol. 5, part 1. Brill, Leiden.

Braga, P. 1979. "Subdivisao fitogeográfica, tipos de vegetação, conservação, e inventario florístico da floresta amazônica." Supplement to Acta amazônica 9 (4): 5380.

Carneiro, Robert. 1970. The transition from hunting to horticulture in the Amazon basin. Eighth Congress of Anthropological and Ethnological Sciences 3: 243-51.

Chernela, Janet. 1982. "Indigenous forest and fish management in the Vaupés basin of Brazil." Cultural Survival Quarterly 6 ²: 17-18.

. 1983. "Hierarchy and economy of the Uanano speaking peoples of the middle Vaupés basin." Ph.D. dissertation, Columbia University.

. 1986a. "Pesca e hierarquização tribal no alto Uaupés." In B. Ribeiro (ed.), Suma Etnológica Brasileira, vol. 1 (Editôra Vozes, Petrópolis), 235-50.

. 1986b. "Os cultivares de mandioca na area do Vaupés. In B. Ribeiro (ed.), Suma Etnológica Brasileira, vol. 1 (Editôra Vozes, Petrópolis), 151-8.

Clark, K., and C. Uhl. 1987. "Farming, fishing, and fire in the history of the upper Rio Negro region of Venezuela." Human Ecology 15: 1-26.

Dufour, Darna. 1983. "Nutrition in the northwest Amazon." In R. Hames and W. Vickers (eds.), Adaptive responses of native Amazonians (Academic Press, New York), 329-55.

. 1987. "Insects as food: A case study from the northwest Amazon." American Anthropologist 89: 383-97.

. 1988. "Cyanide content of cassava cultivars used by Tukanoan Indians of Northwest Amazonia." Economic Botany 42: 255-66.

Falesi, I.C. 1972. Os solos da Rodovia Transamazônica. IPEAN (Instituto de Pesquisa Agronomica do Norte), Belém.

Galvão, Eduardo. 1959. "Aculturação indigene no Rio Negro." Boletim do Museu Paraense Emílio Goeldi, N.S. Antropología 7: 1-60.

Geisler, R., and J. Schneider. 1976. "The element matrix of Amazon waters and its relationship with the mineral content of fishes." Amazoniana 6 ¹: 47-65.

Goulding, Michael. 1979. Ecologia da pesca do Rio Madeira. INPA (Instituto Nacional de Pesquisas da Amazonia), Manaus.

. 1980. The fishes and the forest. University of California Press, Berkeley.

. 1981. Man and fisheries on an Amazon frontier. W. Junk, The Hague.

Goulding, M., M.L. Carvalho, and E.G. Ferreira. 1988. Rio Negro: Rich life in poor water. SPB Academic Publishing, The Hague.

Heindsdijk, D. 1957. Forest inventory in the Amazon valley. FAD, Rome.

Heinen, H.D., and K. Ruddle. 1974. "Ecology, ritual and economic organization in the distribution of palm starch among the Warao." Journal of Anthropological Research 30 ²: 116-38.

Herrera, Rafael. 1979. "Nutrient distribution and cycling in an Amazonian caatinga forest on spodosols in S. Venezuela." Ph.D. dissertation, University of Reading.

Hill, J., and E. Moran. 1983. "Adaptive strategies of Wakuenai people of the Rio Negro basin." In R. Hames and W. Vickers (eds.), Adaptive responses of native Amazonians (Academic Press, New York), 113-35.

Jacobs, Maurius. 1988. The tropical rain forest: A first encounter. Springer-Verlag, New York.

Jordan, C.F. 1982. "Amazon rain forests." American Scientist 70: 394-401.

. 1985. Nutrient cycling in tropical forest ecosystems. Wiley, New York.

Jordan, C.F., and C. Uhl. 1978. "Biomass of a sierra firme forest of the Amazon basin." Oecologia Plantarum 13: 387-400.

Kitamura, P.C., and C.C. Muller. 1984. Castanhais nativos de Marabá:: Fatores de depredaçãoãe bases para sua preservação. Documentos no. 30. EMBRAPA/ CPATU, Belém.

Klinge, Hans. 1978. "Studies of the ecology of Amazon caatinga forest in S. Venezuela." Acta Científica Venezolana 29: 258-62.

. n.d. "Low Amazon caatinga or bane." Mimeo.

Klinge, H., and R. Herrera. 1978. "Biomass studies in Amazon caatinga forest in S. Venezuela." Tropical Ecology 19 ¹: 93-110.

Laraia, R., and R. da Mata. 1968. Indios e castanheiros. Zahar, São Paulo.

May, P.H., A. Anderson, J. Frazão, and M. Balick. 1985. "Subsistence benefits from the babassu palm." Economic Botany 39 ²: 113-29.

McKey, D., P. Waterman, C. Mbi, J. Gartlan, and T. Struhsaker. 1978. "Phenolic content of vegetation in two African rain forests: Ecological implications." Science 202: 61-3.

Medina, Ernesto, M. Sobrado, and R. Herrera. 1978. "Significance of leaf orientation for leaf temperature in an Amazonian sclerophyll vegetation." Radiation and Environmental Biophysics 15:131-40.

Montagnini, F., and C.F. Jordan. 1983. "The role of insects in productivity decline of cassava on a slash and burn site in the Amazon territory of Venezuela." Agriculture, Ecosystems, and Environment 9 (3): 293-301.

Moran, Emilio F. 1973. "Energy flow analysis and Manihot esculenta crantz." Acta amazônica 3 (3): 28-39.

. 1977. "Estrategias de sobrevivencia: o uso de recurses ao Longo da Rodovía Transamazônica." Acta amazônica 7 (3): 363-79.

. 1981. Developing the Amazon. Indiana University Press, Bloomington.

-. 1990. A ecologia humana das populações da Amazonia. Editôra Vozes, Petrópolis),.

. 1993. Through Amazonian eyes: The human ecology of Amazonian populations (revised English edition). University of lowa Press, Iowa City.

Pesce, C. 1985. "Oil palms and other oilseeds of the Amazon." Cited in Balée, 1989.

Pires, J.M. 1973. Tipos de vegetação da Amazonia. Publicacães Avulsas do Museu Paraense Emílio Goeldi, Belém.

Pires, J.M., and G. Prance. 1985. "The vegetation types of the Brazilian Amazon. In G. Prance and T. Lovejoy (eds.), Key environments: Amazonia (Pergamon Press, Oxford), 10945.

Posey, Darrell. 1985. "Indigenous management of tropical forest ecosystems: The case of the Kayapó Indians of the Brazilian Amazon." Agroforestry Systems 3: 139-58.

Prance, G.T. 1978. "The origin and evolution of the Amazon Flora." Interciencia 3(4): 20722.

Prance, G.T., W. Rodrigues, and M.F. da Silva. 1976. "Inventário florestal de um hectare de mate de terra firme km 30 da estrada Manaus-Itacoatiara." Acta amazônica 6: 9-35.

Reid, Howard. 1979. "Some aspects of movement, growth, and change among the Hupdu Makú Indians of Brazil." Ph.D. dissertation, University of Cambridge.

Richards, Paul. 1952. The tropical rain forest. Cambridge University Press, Cambridge.

Salati, Eneas (1985). The climatology and hydrology of Amazonia. In G. Prance and T. Lovejoy (eds.), Key environments: Amazonia (Pergamon Press, Oxford), 18-48.

Saldarriaga, J., and D. West. 1986. "Holocene fires in the Northern Amazon basin." Quaternary Research 26: 358-66.

Sanchez, Pedro. 1976. Properties and management of soils in the tropics. Wiley-Interscience, New York.

Sanford, R.L., J. Saldarriaga, K. Clark, C. Uhl, and R. Herrera. 1985. "Amazon rainforest fires." Science 227: 53-5.

Silverwood-Cope, P. 1972. "A contribution to the ethnography of the Colombian Makú Ph.D. dissertation, University of Cambridge.

Simões, M., and F. Araújo-Costa, 1987. "Pesquisa arqueológicas no baixo Rio Tocantins." Revista de Arqueologia 4 (1): 11-28.

Simões, M., C. Correa, and A. Machado. 1973. "Achados arqueológicos no baixo Rio Fresco." O Museu Goeldi no Sesquicentenario. Publicações Avulsas do Museu Paraense Emílio Goeldi, Belém

Sioli, Harald. 1951. "Zum Alterungsprozess von Flussen und Flusstypen in Amazonas Gebiet." Archiv fur Hydrobiologie 45 (3): 267-84.

Smith, Nigel. 1979. A pesca no Rio Amazonas. INPA, Manaus.

. 1980. "Anthrosols and human carrying capacity in Amazonia." Annals of the Association of American Geographers 70: 553-66.

Sombroek, W. 1966. Amazon soils. Centre for Agriculture Publications and Documents, Wageningen.

Stark, N., and C.F. Jordan. 1978. "Nutrient retention by the root mat of an Amazonian rain forest." Ecology 59 (3): 434-7.

Takeuchi, M. 1961. "The structure of the Amazonian vegetation, II: Tropical rain forest." Journal of the Faculty of Science (University of Tokyo), 8: 1-26.

Uhl, Christopher. 1983. "You can keep a good forest down." Narural History 92: 6979.

Uhl, Christopher, C. Jordan, K. Clark, H. Clark, and R. Herrera. 1982. "Ecosystem recovery in Amazon caatinga forest after cutting, cutting and burning and bulldozer treatments." Oikos 38: 313-20.

Uhl, Christopher, and P. Murphy. 1981. "Composition, structure and regeneration of a sierra firme forest in the Amazon basin of Venezuela." Tropical Ecology 22 ²: 219-37.

Archaeological perspectives on the potential of Amazonia for intensive exploitation

1 Introduction
2 Evolutionary principles
3 Past and present settlement behaviour
4 Relations between várzea, and terra firme groups
5 The impact of climatic fluctuation
6 Overcoming environmental constraints
7 Conclusion


Betty J. Meggers

1 Introduction

The first European visitors were astonished by Amazonia. The lush vegetation, immense rivers, healthy climate, strange animals, and attractive people were truly a "new world," unlike anything known at home. Five hundred years later, the region continues to elicit wonder. Its biological diversity has raised the estimated number of species on the planet from three million to 30 million, and even that may be conservative. Interactions among species are intricate, life cycles are peculiar, and morphology is often exotic, especially among insects. These traits reflect the remarkably complex interaction among biota, soil, and climate, achieved during tens of millennia of biological evolution.

This successful natural configuration is now in conflict with the most successful product of cultural evolution, namely, modern industrial civilisation. We desire to replace it with the kind of landscapes we understand, to make it produce the foods we prefer and support the social and economic institutions we consider most suitable. To this end, we are converting forest to pasture, constructing roads, damming rivers, and encouraging immigration on a larger scale. But things are not working out as expected. Urban sprawl is affecting local weather. Cleared land rapidly declines in productivity and costs often exceed gains. Malaria has become a serious health hazard. The biologists warn of global disaster; the politicians say there is nothing to worry about. Who is right? How fragile are the tropics of Latin America?

We cannot answer these questions unless we recognize and rid ourselves of biases inculcated by our immersion in temperate ecosystems. We continue to judge the carrying capacity of tropical environments using temperate-zone criteria and to design projects for maximizing productivity using temperatezone models, although scientific evidence and practical experience make it increasingly evident that temperate and tropical ecosystems behave very differently. We assess the complexity of tropical forest cultures on the scale of hierarchical organization and urbanism evolved in the riverine cradles of Old World civilization, although archaeological and ethnographic evidence attests to alternative forms of social integration and population distribution. In this context, the numerous and often extensive archaeological habitation sites on the terra firme as well as along the várzea, are equated with large sedentary populations organized into stratified political entities. Similarly, the low density and high mobility of surviving indigenous groups are attributed to decimation and deculturation since European contact (Roosevelt, 1989: 31, 45; Smith, 1980: 553, 564, 566).

The temperate-zone perspective also pervades assessments of the disappointing results of development programmes. Poor planning, insufficient credit, corrupt administration, inadequate transportation, shortage of labour, defective infrastructure, and other sociopolitical and economic factors implicated in the failure of temperate-based projects are judged responsible, with the consequence that attention is focused away from potentially significant environmental and ecological factors (e.g. Collins, 1986; Schumann and Partridge, 1989).

I shall attempt to demonstrate that surviving unacculturated groups accurately depict pre-Columbian settlement behaviour and that their way of life reflects an ancient and remarkable accommodation to complex and variable conditions. In so doing, I will make two assumptions: ¹ the fundamental physical-chemical processes and evolutionary principles that have guided the transformation of life during the past five billion years remain unchanged and immune to our control; and ² our survival depends on exploiting necessary resources in sustainable ways. All organisms affect and are affected by their environments; humans differ mainly in the range of resources they use and in the elaboration of cultural mechanisms for obtaining them. Our global dominance is testimony to the superiority of cultural behaviour over nonsymbolic forms of learning, but does not justify assuming that the evolutionary processes responsible for this situation have been neutralized.By shedding another of our biases - namely, that culture is exempt from evolutionary constraints - we can take advantage of a large body of theory to discover clues for reconstructing prehistoric cultural adaptation in Amazonia, where the predominant physical remains are small undecorated fragments of pottery.

2 Evolutionary principles

Although natural selection is the best known evolutionary process, drift is also a significant mechanism of change. Characters relevant to survival and reproduction are subject to selection, whereas those that are adaptively neutral change directionally by drift. As long as the members of a population interact socially and reproductively, behavioural and genetic variations among individuals are homogenized. If interaction is impeded, however, the resulting segments begin to diverge, in part because of different adaptive constraints and in part because of drift. The best-known cultural manifestation occurs in language, where drift accounts for gradual alterations in pronunciation, meaning, and grammar, but it also operates in other aspects of culture.

The recent history of the Panare in south-eastern Venezuela provides an excellent illustration of the process. Less than a century ago, they began expanding from their traditional territory in the Guyana highlands toward the north-west, into land vacated by the previous occupants (fig. 4.1). Dispersal along several tributaries segregated the original community into three semiisolated populations, which have begun to diverge in dialect, kinship terminology, dress, and ornament, as well as in ceremonies associated with male initiation, female puberty, and funerals (Henley, 1982: 11-14). Similar incipient linguistic and cultural differentiation has been observed among other recently separated Amazonian groups (e.g. Yanomami: Spielman, Migliazza and Neel, 1974; Campa and Machiguenga: D'Ans, 1982: 252; northern Ge: Seeger, 1981: 229-30).

Temper and firing are among the characteristics of pottery susceptible to unconscious directional change. Classifying unselected samples of fragments using uniform criteria permits recognizing distinct trends in the relative frequencies of undecorated types. All samples that can be interdigitated into a single sequence represent the same prehistoric community, termed a phase (fig. 4.2). Phases that share a specific set of decorative characteristics are assigned to the same tradition. After these communities have been identified, their settlement behaviours can be reconstructed and their population densities can be inferred. Comparing the resulting patterns with those of surviving indigenous groups makes it possible to assess whether or not the latter are representative of the pre-Columbian demographic situation.

Figure 4.1 Panare expansion during the past hundred years from their traditional territory south-east of the Orinoco into lands vacated by the previous indigenous population. Groups occupying different rivers have begun to drift apart linguistically and culturally as a consequence of reduced interaction (After Henley, 1982.)

Figure 4.2 Seriated sequence of the Tucuruí Phase, Rio Tocantins. The pottery was classified into four undecorded types distinguished by differences in tempering material and four decorated types defined by technique of decoration. Vertical bars at the left connect successive levels of stratigraphic excavations. The direction of change is indicated by stratigraphic evidence for an increase in Jatobal Plain and a decrease in Tucuruí Plain in each excavation. The erratic occurrence of Turucuí Painted implies the presence of two matrilocal exagamous residential groups (subphases).

3 Past and present settlement behaviour

Survey along the principal tributaries of the Amazon during the past fifteen years by participants in the Programa Nacional de Pesquisas Arqueológicas na Bacia amazônica (PRONAPABA) has located and sampled hundreds of archaeological sites (fig. 4.3). The resulting data permit dozens of phases to be distinguished and their settlement behaviour to be described.

Figure 4.3 Regions surveyed during the Programa Nacional de Pesquisas Arqueologicas na Bacia Amazônica between 1977 and 1983. Areas 2, 5a, 5b, 8, and 21 by Mário Simões, Areas 3, 6, 7, and 22 by Celso Perota; Areas 9-12 and 23 by Enrico Th. Miller; Areas 13-17 by Ondemar Dias, and Areas 24-26 by Bernardo Dougherty.

Territorial configurations

The spatial distributions of habitation sites belonging to contemporary phases generally suggest contiguous territories along rivers. Along the lower Tocantins, for example, the sharp boundaries between the phases coincide with changes in the character of the river. Each zone provides different aquatic resources and requires different methods of exploitation (fig. 4.4). Similar contiguous territories have been reported among surviving indigenous communities in various parts of the lowlands, among them Akawaio River groups in westcentral Guyana (fig. 4.5).

Figure 4.4 Territorial distributions four contemporary phases (prehistoric communities) on the lower Tocantins, inferred from the locations of habitation sites. The boundaries correlate with physical changes in the character of the river and associated differences in subsistence resources.

Figure 4.5 Territories of Akawaio River groups in west-central Guyana. The riverine focus and contiguous along the river are shared with the prehistoric territories reconstructed in figure 44, suggesting that the archaeological phases represent similar endogamous communities. (After Colson. 1983/4, map 2.)

Figure 4.6 Territorial distributions four archaeological phases in eastern Acre, suggesting the existence of isolated territories.

Where steep slopes, annual inundation, and other geographical features make parts of a region unsuitable for settlement, territories are likely to be isolated by unoccupied land. Identifying this pattern archaeologically depends on more complete survey data than are available from most parts of Amazonia, but the known distributions of phases in eastern Acre (fig. 4.6) are generally comparable to the non-contiguous territories of Yukpa-Yuko communities in western Venezuela (fig. 4.7).

Village movement

Since a seriated sequence is a relative chronology, comparing the locations of sites with their chronological positions permits tracing village movement. The schematic settlement history of the Tucuruí Phase on the lower Tocantins is typical. Differences in the relative frequencies of painted decoration suggest the community was divided among two matrilocal villages, each of which moved independently and occasionally fissioned and refused. Although the sites are seldom shared, both patterns are centripetal within the territory and combine short and long moves, some to pristine and some to previously occupied locations (fig. 4.8).

Figure 4.7 Isolated territories of Yukpa-Yuko communities on the border between Colombia and Venezuela, resembling the disjunct pattern reconstructed for the archaeological phases in figure 4.6. (After Ruddle, 1971, map 1.)

Figure 4.8 Schematic reconstruction movement during the Tucuruí Phase, by subphase B derived from sequential positions of sites represented in the seriated sequence (fig. 42). Although more intensive investigation would undoubtedly identify additional episodes of occupation at many of the sites, the general pattern would not be affected.

The actual history of village movement during the past seventy-five years by the Mekragnoti of the upper Xingú reconstructed from interviews with informants, also incorporates long and short movements, reoccupation of several locations, and episodes of fission and fusion (fig. 4.9).


Changes in the relative frequencies of the undecorated pottery types in successive levels of a stratigraphic excavation provide the basis for estimating continuity and discontinuity of occupation of a site by the same phase and for recognizing reoccupation by a different phase or phases of the same or different traditions. The procedure is exemplified by the seriated sequences of the Kiusiu and Ibare phases on the Llanos de Moxos (fig. 4.10).

Two kinds of discontinuities are evident in both seriations. First, successive levels of each excavation (connected by vertical bars at the left) are separated by one or more levels from the other excavation. This interdigitation implies that the portion of the site represented by the excavation was abandoned and subsequently reoccupied. Second, the seriated positions of some levels are contiguous, but the magnitude of the differences in relative frequencies of the principal undecorated type is equal to or greater than that between levels with disjunct positions (e.g. BE-MO-8, Levels 60-80 and 80100 cm). This situation also implies discontinuous occupation of the location sampled.

This seriation provides evidence not only for discontinuous occupation at both sites during the Kiusiu Phase, but also for reoccupation of both by the Ibare Phase. The boundary is marked by reversal of the trend in the principal undecorated pottery type and by abrupt changes in occurrence of most of the decorated types. Carbon-14 dates confirm the stratigraphic evidence for priority of the Kiusiu Phase. This pattern of reoccupation of sites by the same phase and by successive phases of the same or different traditions is characteristic throughout the lowlands.

Figure 4.9 History of village movement by the Mekragnoti in the upper Xingú region during 75 years, reconstructed from interviews with informants. The centripetal pattern, long and short moves, fissioning and fusing, and reoccupation of several locations are characteristics shared with the Tucuruí Phase reconstruction in figure 4.8. (After Verewijver, 1978.)

Multiple reoccupation of the same locations is relatively common among surviving groups. For example, the Mekragnoti have moved their village thirty-six times since AD 1900. Ten of the sites had been occupied previously. Five were reoccupied once, several twice, and one three times (Gross, 1983: 439). Reoccupation of former locations is also characteristic among the Siona-Secoya and Tatuyo of eastern Colombia (Dufour, 1983: 330; Vickers, 1983: 471) and the Yanomami of southern Venezuela (Hames, 1983: 412-14). Archaeological evidence that the same location was reoccupied by phases belonging to different traditions has a modern counterpart in the reoccupation of prehistoric sites by contemporary groups (e.g. Balée 1987).

Figure 4.10 Seriated sequences derived by interdigitation of levels of straffgraphic excavations d two sites on the central Llanos de Moxos, identified by the vertical bars at the left. Replacement of the Kiusiu Tradition by the Ibare Tradition, marked by discontinuities in the trends of the undecorated pottery types and in the techniques of decoration, coincides with a brief, intense episode of aridity c. 700 HP (fig. 4.12).

Village area

Sites of several phases in Acre vary from 3 m diameter to 8 x 15 m, comparable to the ranges of house dimensions among the Mojo, Panare, Camayurá, Kayapó, and Witoto. Along the Xingú and Tapajós, by contrast, the minimum dimensions of sites are close to the upper limit of dwellings reported ethnographically (65 x 30 m and 50 m diameter). Differences in house size are associated today with differences in the size of the domestic group, rules of marriage and residence, and other social behaviour, implying the existence of similar variation prior to European contact.

The maximum dimensions of contemporary single-house villages suggest that archaeological sites with surface extensions exceeding about 50 m in diameter or 65 m in length are the product of multiple occupation. When a site is represented by more than one stratigraphic excavation, it is often possible to correlate the levels and estimate the maximum area attributable to each episode (Meggers, 1992, fig. 5). A more detailed reconstruction of settlement history can be made from a larger number of stratigraphic excavations distributed across the site (Miller et al., 1993, figs. 69-71). Although two or more contemporary dwellings are sometimes implied, none of the episodes identified thus far covers more than a small fraction of the surface of the site.

In summary, existing archaeological evidence does not confirm the existence of large permanent settlements prior to European contact along any of the tributaries of the Amazon. Rather, it indicates that the settlement behaviour exhibited by surviving unacculturated indigenous communities, including territorial configuration, village movement, village duration, and village size, was established by the beginning of the Christian era.


Contents - Previous - Next