An immense orchard

Distribution of trees and geoglyphs underscores the impact of pre-Columbian human populations in the Amazon Forest

Imagem: Carolina Levis Trees used by humans predominate in an area of dark earth at the Humaitá National Farm, state of AmazonasImagem: Carolina Levis

Well before Europeans arrived in the Americas, the Amazon Region had abundant populations of indigenous peoples who were quite knowledgeable about using the forest for their own benefit. They selected and cultivated vegetation to the point of altering the plants’ properties; they excavated circular or square-shaped trenches that are visible from high in the air; and they built water reservoirs. Thousands of years of such alterations have resulted in a forest that retains few traces of virgin stands, according to botanists and archeologists.

“We found that, in areas surrounding archeological sites, there is a higher concentration and diversity of trees used by the Indians,” says biologist Carolina Levis, a doctoral candidate at the National Institute for Research on the Amazon (INPA) and Wageningen University in the Netherlands, first author of a paper published in the journal Science in February 2017. Levis was the first to examine the correlations between archeological and botanical data. The differences are so striking that she proposes using flora composition as a signature for locating ancient human settlements. Taking into account present-day trees, she found 85 species used and domesticated by Indians, such as açaí palm, the Brazil nut tree and the rubber tree.

The results were obtained thanks to the use of two extensive databases. One of them was compiled by archeologist Eduardo Tamanaha, a doctoral candidate at the University of São Paulo’s Museum of Archeology and Ethnology (MAE-USP), and ecologist André Junqueira, currently a postdoctoral researcher at Wageningen University, and it includes data from more than 3,000 archeological sites. The other database, which was created by botanist Hans ter Steege of the Naturalis Biodiversity Center in the Netherlands, draws on a network of researchers who took botanical inventories at 1,170 sample plots in Amazonia, and it contains over 4,000 tree species.

Imagem: Edison Caetano Horticultural geoglyph, in Acre, is considered complex due to its variety of elementsImagem: Edison Caetano

Anyone who has spent time in an orchard and learned which trees yield the best fruit can conceive of this type of selection as the first step in domestication, and it leads to a gradual process of differentiated plant management. The next step is to grow the plants outside the forested area, such as in home gardens and plantation plots. Over time, the morphology and genetics of the fruit become altered, creating populations and individuals with features very different from those of the original fruit.

Searching for clues to the domestication process in the genetic material of these plants is the focus of American biologist Charles Clement of INPA, a coauthor of the paper and coadvisor of Levis, along with biologist Flávia Costa. One example is the peach-palm (Bactris gasipaes), which he notes is appreciated for the fruit as well as for the hearts of palm it is known for in Southeastern Brazil. In the wild variety, each fruit weighs about one gram (g), but the domesticated version can grow to as large as 200 g in western Amazonia. The studies he has conducted since his master’s degree indicate that domestication of the peach-palm began in what is now the Bolivian Amazon and spread via two different routes. The abundant form in western Amazonia is the source of a pulp that is good for fermentation. “Peach-palm beer is a favorite of the Indians,” he says. In the area around the cities of Manaus and Belém, the oil-rich fruit does not ferment well. In this case, the fruit is cooked and enjoyed as an appetizer. Its genetic markers, however, provide no clues about when these events occurred.

Botanist Priscila Moreira, Clement’s doctoral student, obtained more detailed results with the calabash tree, or treegourd (Crescentia cujete), using sequencing methods carried out in partnership with geneticist Yves Vigouroux of the French Research Institute for Development. In a paper accepted for publication in the journal Evolutionary Applications, she refutes the idea that the treegourd was domesticated in Amazonia or Mexico. The plant arrived in Amazonia already domesticated, and it crossbred with the wild treegourd, producing hybrids that yield smaller fruits than those of the domesticated trees. The shape varies as well, and each type is used in its own specific way by present-day Indians.

Imagem: Val Moraes / Central Amazon Project Ceramics over 1,000 years old at the Hatahara site, near the Solimões River…Imagem: Val Moraes / Central Amazon Project

According to Clement, the research suggests that agriculture is not intrinsic to large societies. For forest dwellers, domesticating the landscape can produce as much food as farming in other parts of the world. The abundance of plants used by the Indians suggests that many domestications occurred in southwestern Amazonia, where important linguistic families such as Tupi and Arawak also arose. “Those groups could have brought the plants over long distances,” Levis suggests. The correlation between hyperdominant trees and evidence of ancient human populations is stronger in southwestern Amazonia, such as in the state of Rondônia, and also in the area around the mouth of the Amazon River, but definitive conclusions are not possible because of the sheer amount of information that remains unknown, from both a floristic and an archeological standpoint. One of the difficulties lies in knowing whether the distribution of the trees was really altered by succeeding generations of Indians, or whether people settled in places where there were valuable resources. Levis supports the first of these explanations. “We found trees with different ecological preferences growing in the same sample plots, which is unlikely to have occurred naturally.”

Life in the forest
“The findings corroborate the view that there have always been people in the Amazon and that the present-day forest is therefore not so natural,” comments archeologist Eduardo Góes Neves, a professor at MAE and coauthor of the study. The next step is to refine the work further by using archeological data to investigate which species were domesticated in the past. “Archeology is a social science, but it requires a dialogue with the natural sciences.”

Imagem: William Farfan-Rios …and rock art in Hinkiori, PeruImagem: William Farfan-Rios

Neves says that one problem lies in the lengthiness of archeological studies. “Botanists are able to collect samples at several predefined points in a week’s time, but we spend years excavating a single site. Knowledge is produced at a very slow pace.” For this reason, the study published in Science considered only the presence or absence of archeological sites. There can be immense variation, however, both in the traits of the inhabitants and in time.

One important part of this ancient story can be found in Rondônia, near the city of Porto Velho, where Neves began excavating 10 years ago with archeologist Fernando Almeida, who was then his student and is now a professor at the Federal University of Sergipe. An abundance of dark earth, which resulted from successive burn events that break down organic matter, including excrement (see Pesquisa FAPESP Issue nº 183), suggests a large population over a long period. It is a more fertile soil that does not erode with the passage of time and water. Moreover, nearly neutral soil pH preserves plant remains, ceramics and other signs of human life—in this case from up to 6,500 years ago. “Domesticated plants already existed at that time,” says the MAE professor.

During migrations over the immense stretches of Amazonia, human groups probably encountered a barrier in the Teotônio rapids, on the Madeira River. Industrial quantities of fish were available until the area was flooded by the Santo Antônio hydroelectric dam in 2011. “Fishing promoted a sedentary lifestyle,” Almeida notes. But there are also indications that cassava cultivation began there—in large part, he says, in order to make beer. Thousands of years ago, the beverage was used not only in rituals, but also as food, as he describes in a 2015 paper in the journal Estudos Avançados.

Imagem: Jorge Contreras Domesticated fruits: wild sugar-apple (Anonna mucosa)…Imagem: Jorge Contreras

Fermentation leaves a crackled appearance typical of ceramics, but identifying the plants that were used (most often corn and cassava) requires analysis of plant remnants. English archeologist Jennifer Watling, a postdoctoral researcher at MAE who is supervised by Neves, has found ancient samples of plants at Teotônio, such as pumpkin, beans and corn, including inside ceramic fragments, indicating that they were used for food.

Some 400 kilometers away in the state of Acre, Watling also studied both ancient and present-day vegetation around trenches measuring up to 11 meters (m) wide and 4 m deep. These geoglyphs, as they are called, are characterized by geometric shapes up to 300 m in diameter. The remains of plants, called phytoliths, have led to a better understanding of the formation and use of these structures, according to a paper published in PNAS in February 2017. “Phytoliths bear testimony of a part of the forest in the past,” she says in summation. They show that 6,000 years ago the region was covered in bamboo, still a dominant plant today, and that the oldest geoglyphs were built between 3,000 and 3,500 years ago in a landscape previously altered by human activity. This vegetation, which is easier to cut down, was removed only at the excavation sites, so the structures remained hidden. Another sign of anthropic alteration is a large number of palm trees 2,000 years ago—a period when there were fires, and climatic conditions were not conducive to an increase in density of this type of plant.

Watling explains that different groups likely used the geoglyphs for long periods of up to 3,000 years for something akin to a meeting place. But there are no signs of large settlements nearby. “The people were nomads and camped outside the geoglyphs,” says archeologist Denise Schaan, a professor at the Federal University of Pará. Using satellite images as well, she has been studying these structures since 2005 in collaboration with Finnish colleagues, covering the entire region that includes the borders between the states of Acre, Rondônia and Amazonas, and Bolivia (see Pesquisa FAPESP Issue nº 186). Since late 2016, they have begun to use a technique known as LiDAR (Light Detection and Ranging), employed from aboard drones, to map the topography of the terrain in forested areas. “We’re finding structures that we used to have trouble identifying even in deforested areas.”

Imagem: Carolina Levis …and açaí (Euterpe precatoria)Imagem: Carolina Levis

A long distance away, Shaan is also studying signs of settlements in the Santarém-Belterra region, in central Pará State. There are more than 100 archeological sites in a region controlled by Tapajós Indians that has deteriorated as a result of mechanized soybean farming. It was a surprise to find signs of ancient peoples on plateaus far from major rivers. “They used depressions to dig wells ranging from 8 m to 100 m in diameter, where they collected rainwater and practiced farming.” She says that until recently, local populations used these wells, but the practice was abolished due to pollution from pesticides.

In the aggregate, these studies reinforce the notion that there were many people in Amazonia before the Europeans arrived, perhaps 8 million to 10 million Indians. Some interpretations, however, are viewed with caution by other researchers, such as paleoecologist Crystal McMichael, a professor at the University of Amsterdam. In a paper published in PNAS in January 2017, she and colleagues point out that a large percentage of botanical studies are done in locations where humans lived, so that alterations left in the forest receive disproportionate weight in our ecological understanding of the region. “We’ve studied such a small fraction of Amazonia, in both archeological and ecological terms, that we should be careful when extrapolating to unexplored areas in either discipline,” she cautions.

Imagem: Salman Kahn Laser technique (LiDAR) employed from aboard drones enables researchers to see structures beneath the forest coverImagem: Salman Kahn

McMichael is not convinced of a causal relationship between the ancient populations and the forest alterations, noting that proof is needed as to the timing of the domestications. In Neves’ opinion, these disagreements are positive and have led the researchers to rework their data to find answers. “I think much of the disagreement comes from the fact that archeologists, ecologists and paleoecologists do not have a common language, although we’re beginning to share ideas,” McMichael suggests.

1. Hydrology-vegetation structural relationships in the Purus-Madeira interfluve (nº 09/53369-6); Grant Mechanism Regular Research Grant; FAPEAM Agreement; Principal Investigator Javier Tomasella (INPE); Investment R$138,310.07 (FAPESP).
2. Pre-Columbian human environment interactions in the Upper Madeira basin, southwest Amazônia (nº 14/21207-5); Grant Mechanism Postdoctoral research grant; Principal Investigator Eduardo Góes Neves (USP); Grant Recipient Jennifer Watling; Investment R$280,149.35.

Scientific articles
LEVIS, C. et al. Persistent effects of pre-Columbian plant domestication on Amazonian forest composition. Science. V. 355, No. 6328, p. 925-31. March 3, 2017.
WATLING, J. et al. Impact of pre-Columbian “geoglyph” builders on Amazonian forests. PNAS. V. 114, No. 8, p. 1868-73. February 6, 2017.
MCMICHAEL, C. N. H. et al. Ancient human disturbances may be skewing our understanding of Amazonian forests. PNAS. V. 114, No. 3, p. 522-7. January 17, 2017.
CLEMENT, C. et al. The domestication of Amazonia before European conquest. Proceedings of the Royal Society B. V. 282, No. 1812.  August 7, 2015.
ALMEIDA, F. O. de. A arqueologia dos fermentados: A etílica história dos Tupi-Guarani. Estudos Avançados. V. 29, No. 83, p. 87-118. January-April 2015.