Add offer

Amazonian Dark Earth: Terra Preta and Indigenous Agricultural History

Explore Amazonian terra preta: archaeological dates, Indigenous soil management, settlement and crop histories, and the debate over natural origins.

Seventh century CE · Osvaldo example; origins much earlier · Amazonia: Brazil, Bolivia and other parts of the basin · Land & water

Beneath parts of the Amazonian forest survives an archive of agricultural history: enriched dark soils known as terra preta. Their development connects settlements, food preparation, plant management and prolonged attention to the ground. Research reveals multiple formation pathways and the active contributions of Indigenous communities. There is no single invention date, however, and no simple treatment that explains every archaeological site.

Dark earth within a varied Amazonian landscape

Terra preta, literally “black earth”, describes soils whose dark surface horizons preserve the effects of prolonged human activity. Researchers commonly use the term Amazonian Dark Earths, or ADE. Charcoal particles, accumulated organic matter and pottery fragments are characteristic indicators. Colour alone, however, cannot identify an archaeological soil: the profile and its surroundings must also be investigated. Many examples occupy non-flooding ground overlooking waterways, where inhabitants could combine settlement with access to rivers and wetland resources.

Amazonia contains many different soils. The contrast between dark earth and strongly weathered, nutrient-poor surroundings is striking in numerous locations, but it does not describe every river valley. Research at the Jabuti archaeological site in Bragança examined dark earth developed in a periodically waterlogged setting. Compared with neighbouring soil, the sampled horizons had improved chemical properties and greater water retention. This case shows that people modified different environments.

Multiple chronologies rather than one invention date

Dark earth did not appear simultaneously across the basin. At Teotônio on the Madeira River, researchers documented early plant use followed by deposits interpreted as anthropogenic soils. Dates associated with the latter phase extend to approximately 6,500–6,400 calibrated years before 1950. Useful plants occurred earlier, indicating that landscape management preceded the formation of recognisable dark horizons. Such early dates belong to particular archaeological contexts. They do not establish when every ADE developed or demonstrate a single agricultural tradition spreading unchanged across Amazonia.

Osvaldo, near the meeting of the Negro and Solimões rivers, illustrates a different timescale. Eduardo Neves and colleagues connected much of its dark-earth accumulation with seventh-century CE occupation. Their analysis suggested that a substantial layer could develop relatively rapidly, over approximately a century, although subsequent cultivation had disturbed the upper profile. A centimetre of dark soil therefore cannot be translated into a fixed number of years. Deposition depended on occupation intensity, movement of materials and later disturbance.

From food remains to the enduring layout of a settlement

Fertility had material sources. Preparing meals, processing plants, consuming fish and animals, and clearing hearths concentrated nutrients around dwellings. Bones could supply phosphorus and calcium, while ash and incompletely burned residues altered the receiving soil. Molecular-marker research by Jago Birk and colleagues also identified contributions from faecal material in the soils they examined. These findings reveal several pathways through which matter accumulated.

The position of enriched ground can help reconstruct settlement organisation. Research in the Upper Xingu, Central Amazon and lower Trombetas identified recurrent arrangements of raised refuse deposits and flatter activity surfaces. Nutrient concentrations and archaeological material varied with the use of each place. A swept plaza, a road, a household backyard and a disposal area could leave different signatures. Soil therefore preserves evidence of daily work after wooden buildings have disappeared and forest has covered the settlement.

Why some properties endured for centuries

Carbon produced by incomplete biomass combustion was an important component. Research led by Biqing Liang showed that the investigated charcoal-rich soils had greater potential cation-exchange capacity per unit of organic carbon than adjacent comparison soils. Oxidation of particle surfaces and the attachment of organic matter helped retain positively charged nutrient ions. This mechanism explains part of the persistence of fertility, but charcoal did not manufacture phosphorus or calcium. Those elements had to enter the soil through other materials.

Microbial life also mattered. Julie Grossman and colleagues compared anthropogenic soils and neighbouring substrates at four sites around Manaus. Bacterial and archaeal communities in the dark earths shared similarities despite differences in current land use, while remaining distinct from the comparison samples. The results connect management history with subsequent soil functioning. They do not identify one special organism responsible for fertility. Enduring properties emerged from interactions between chemistry, organic matter and biological processes.

Fields, household cultivation and forest resources

At Triunfo and Versalles in the Bolivian Amazon, spatial sampling revealed a mosaic of dark earth and lighter anthropogenic brown earth. Geochemistry, isotopes and plant remains allowed researchers to interpret these soils as parts of a food-producing landscape. Settlement, crop cultivation, the use of trees and different forms of fire management operated together. Occupation at Triunfo began around 500 BCE. Differences between the soils show why agricultural history cannot be reconstructed solely from the blackest and most conspicuous patches.

Later cultivation also tests assumptions about suitable crops. James Fraser studied Middle Madeira communities where bitter manioc was grown on relatively poor soils, fertile dark earth and alluvial ground. Farmers adjusted varieties, cropping periods and fallows to local conditions. Their practices challenge the simple assumption that a crop adapted to infertile soil cannot perform well on fertile land. However, present-day observations are evidence about possibilities and adaptations, not direct measurements of precolonial yields.

Intentional enrichment and Indigenous knowledge

Research published in 2023 with Kuikuro and Wauja coauthors is especially significant. Upper Xingu practices included spreading organic residues and hearth materials in ways that residents explicitly associated with improving soil. These observations were compared with spatial and chemical patterns in archaeological settlements. The findings strongly support intentional enrichment and the possibility of comparable practices in the past. They do not show that every ancient horizon resulted from a planned project. Deliberate soil management could coexist with the cumulative, unintended consequences of living, cooking and disposing of materials in a settlement.

Earlier ethnoarchaeological work in this region demonstrated the value of following an entire settlement cycle: the house, its surroundings, an abandoned location and subsequent reuse. This approach connects particular activities with the soil traces they create. Indigenous communities contribute knowledge of plants and landscapes as interpreters and researchers, not simply as unnamed sources of “tradition”. Innovation need not have been a single discovered formula; it could consist of knowledge about materials, space and labour developed and transmitted across generations within living communities.

The Caldeirão debate and the natural-origin hypothesis

In 2021, Lucas Silva and colleagues proposed a different explanation for soils at Caldeirão near Iranduba. They interpreted elemental patterns, isotopic signatures and dated charcoal particles as evidence for ancient river-derived inputs. In their hypothesis, some exceptional fertility preceded recognised agricultural management, and inhabitants benefited from an already enriched location. This was a proposal based on one intensively investigated area. The authors made its wider significance conditional on corroboration elsewhere.

A response published by Umberto Lombardo and numerous colleagues in 2022 emphasised the extensive archaeological evidence for anthropogenic formation. The critics argued that chemistry and isotopes alone cannot determine whether nutrients arrived in river sediments or materials transported by people. They also challenged the adequacy of the proposed geomorphological mechanism. The disagreement concerns how specific evidence should be interpreted and how different processes contributed. It provides no basis for replacing well-documented human activity with a general account in which all dark earth formed naturally.

Soil as an archive of a larger historical landscape

The distribution of known sites also reflects where researchers have looked. A model published by Crystal McMichael and colleagues in 2014 related recorded dark earth to terrain, hydrology and soil conditions to identify promising survey areas. Its predictions were not a complete inventory of surviving ADE. Work on the Napo highlighted a related problem: poorly surveyed regions can appear sparsely inhabited in the past because evidence has not been recorded.

Agricultural legacies extend beyond soil profiles. Carolina Levis and colleagues found relationships between the distribution of domesticated woody species in present-day forests and pre-Columbian archaeological sites. People shaped landscapes by planting, protecting and dispersing useful plants as well as modifying soil. Terra preta forms one part of this history; it does not demonstrate equally intensive transformation across the entire basin. Removing soil together with its artefacts destroys spatial relationships that cannot be recovered simply by measuring the remaining material’s fertility.

Questions and answers

Is terra preta simply soil mixed with charcoal?

No. Terra preta records multiple material inputs and prolonged settlement-related transformations. Charcoal alone does not recreate that history.

Are all Amazonian dark earths equally old?

No. Dates describe particular sites and deposits, not the entire history of Amazonian dark earth.

Is the exact area of Amazonian dark earth known?

There is no complete inventory. Models identify likely distributions and survey targets; their results are not direct measurements of total area.

Sources

History of agriculture: all topics