August 17, 2012

Rainforest Crunch: Origins of the Brazil nut in ancient Amazonia (Bertholletia excelsa)

If you still miss "Rainforest Crunch" ice cream, you can finally rest assured that Amazonian Indians really were behind all those Brazil nuts in the recipe, only not in quite the way that Ben & Jerry's had advertised.
The stately Brazil nut tree appears to have been managed
and perhaps cultivated by ancient Amazonian peoples.

Recent scientific studies show that Brazil nut groves have been managed, facilitated and probably spread throughout the Amazon by indigenous peoples since before European conquest. As highlighted in this month's issue of the Brazilian science news magazine, Revista FAPESP, "the human factor" has played an important role in shaping this emblematic rainforest landscape.

The Brazil nut comes from colossal trees up to 60 m tall (200 ft) and 16 m in circumference (53 ft) which are found in discontinuous patches throughout the Amazon basin. It is the most important non-timber commercial product of the South American rainforest, with an annual harvest in Brazil alone of over 25,000 tons employing some 200,000 people. The grove-like nature of Brazil nut stands led early researchers like Adolpho Ducke to guess that they might be plantations left behind by ancient indigenous peoples.[2] However more recent scientific research criticized this theory, claiming that the agouti, a large tropical rodent, was solely responsible for their dispersal.[3]

The Brazil nut harvest provides a livelihood for
tens of thousands of Amazon rainforest dwellers.

However new studies have reasserted an important role for indigenous peoples and their modern descendants in spreading and maintaining Brazil nut groves throughout the Amazon. "Made in Brazil," published last year in Economic Botany[4], presents results of a research project I initiated in 2000 with plant geneticists Rogerio Gribel and Maristerra Lemes, and continued to develop through later collaborations with archeologist Eduardo Neves and linguist Henri Ramirez. 

The softball-sized seed case within which the Brazil "nuts" are held is extremely hard and heavy, making it particularly inefficient for long-distance dispersal. While other trees in the same family (Lecythidaceae) have seed cases that open when ripe, the Brazil nut's seed case remains tightly sealed and nearly impenetrable. Agoutis are virtually the only animals with teeth sharp enough to open the seed case. Although agoutis do play a role in dispersing Brazil nut seeds a short distance within existing groves, it is hard to explain how they alone could have spread the seeds to form discrete, far-flung groves across much of the Amazon basin. If the Brazil nut was spread throughout its vast range by natural processes only, it would have taken a very long time.

Brazil nut seeds are encased in an unusually thick, hard outer shell that does not open naturally, usually requiring either agoutis or humans to break them open.

Yet our genetic studies comparing eight widely separated Brazil nut groves across the entire basin indicate that all may be descended from a single original seed bank that was dispersed quickly through a large area.[5] The observed genetic pattern -- low overall genetic variability, and greater variability within groves than between them -- is rare in wild trees but common in cultivated trees like Eucalyptus[6], indicating a strong role of human intervention.  

This result is consistent with Ducke's original suggestion that the Brazil nut was planted or managed by ancient indigenous populations. Indeed, the 11,000 year-old archeological site of Pedra Pintada, near the modern city of Santarém, contains Brazil nut remains.[7]

To further test this possibility, linguist Henri Ramirez collected and analyzed the vocabulary for Brazil nut in over 70 indigenous languages. The pattern of loan words and tentative reconstructions of proto-language terms suggests that Brazil nut was present and culturally important some 2000 to 4000 years ago in the northern part of the Amazon basin where the the Arawak and Carib language families originated, encompassing the region where Brazil nuts were first documented in the archeological record at Pedra Pintada. 
Linguistic analysis suggests a pattern of borrowing especially to the south and west
(Shepard & Ramirez 2011, p. 53).

In the southern and western part of the basin, by contrast, languages in the Tupi family and others refer to the Brazil nut either with apparent loan words, or words derived from the term for "peanut," a cultigen first domesticated in the southern Amazon. This raises the possibility that the Brazil nut was not originally present or culturally important in these regions. Archeologist Eurico Muller, who has conducted excavations in Rondônia in the southern Amazon, was surprised to find no evidence of Brazil nut consumption in ancient sites, despite how dominant Brazil nut is in the current landscape.

Though somewhat speculative, these results provide further evidence that the Brazil nut may have been spread by ancient indigenous peoples, extending its range as new groups were introduced to the seed crop by their neighbors.

Indeed, the distribution of the Brazil nut, found in central and eastern Amazonia but largely absent in the western part of the basin, closely matches the distribution of "terra preta do índio" ('indigenous black earth'), patches of highly fertile dark soil found in association with ancient Amazonian agricultural settlements.[8,9] Thus Brazil nuts may have been managed or cultivated as part of the intensive agricultural and agroforestry practices that permitted indigenous populations in some parts of the Amazon to grow and flourish during and after the first millennium A.D.[10,11]

The territory of the Wari' people of Rondônia contains vast Brazil nut groves that are important in subsistence, ritual and social life as well as the modern cash economy.

More recent work published in Human Ecology by Ricardo Scoles and Rogerio Gribel[12] contributes further evidence in favor of this interpretation by elucidating how management practices from ancient through more recent times are crucial to the maintenance, productivity and perhaps establishment of Brazil nut groves. 

Rainforest Crunch now lies defunct in Ben & Jerry's Flavor Graveyard, after journalist Jon Entine revealed that indigenous peoples were not in fact receiving the kind of economic benefits originally promised.

And yet the ancient legacy of Amazonian peoples lives on in the global Brazil nut trade: think about that the next time you dig into a bag of trail mix.


2. Ducke, A. 1946. Plantas de cultura précolombiana na Amazônia brasileira. Boletin Técnico do Instituto Agronomico do Norte 8: 2-24.
3. Peres, C. A. and C. Baider. 1997. Seed dispersal, spatial distribution and population structure of Brazil nut trees (Bertholettia excelsa) in southeastern Amazonia. Journal of Tropical Ecology 13: 595-616.
4. Shepard G.H. Jr. and H. Ramirez. 2011. Made in Brazil:  Human dispersal of the Brazil nut (Bertholletia excelsa, Lecythidaceae) in ancient AmazoniaEconomic Botany 65(1): 44-65.
5. Gribel, R., M. R. Lemes, L. G. Bernardes, A. E. Pinto and G. H. Shepard Jr. 2007. Phylogeography of Brazil-nut tree (Bertholletia excelsa, Lecythidaceae): evidence of human influence on the species distribution. Paper presented at the meetings of the Association for Tropical Biology and Conservation, Morelia, Mexico, July, 2007. Abstract at online conference proceedings, pg. 281.
6. Kanashiro, M., S. A. Harris and A. Simons. 1997. RAPD diversity in Brazil nut (Bertholletia excelsa Humb. & Bonpl.: Lecythidaceae). Silvae Genetica 46(4): 219-223.
7. Roosevelt, A., C. M. Lima da Costa, C. Lopes Machado et al. 1996. Paleoindian cave dwellers in the Amazon: The peopling of the Americas. Science 272: 373-384.
8. Kern, D., G. D’aquino, T. Rodrigues, F. Frazao, W. Sombroek, T. Myers and E. Neves. 2004. Distribution of Amazonian dark earths in the Brazilian Amazon. In J. Lehmann, D. Kern, B. Glaser, and W. Wodos [eds.], Amazonian Dark Earths, 51-75. Dordrecht: Kluwer Academic Publishers.
9. Arroyo-Kalin, M. 2010. The Amazonian formative: Crop domestication and anthropogenic soils. Diversity 2: 473-504.
10. Neves, E. G. and J. B. Petersen. 2006. The political economy of pre-Columbian Amerindians: Landscape transformations in central Amazonia. In W. Balée and C. Erickson [eds.], Time and Complexity in the Neotropical Lowlands: Explorations in Historical Ecology, 279-310. Columbia University Press, New York.
11. Heckenberger, M.J., J. C. Russell, C. Fausto et al. 2008. Pre-Columbian urbanism, anthropogenic landscapes, and the future of the Amazon. Science 29(321): 1214-1217.


  1. Thank you for the note! We can even go further to propose that human management (harvesting plus cleaning around the crown projection) helps to increase the annual crop. I have seen the crops in Tambopata and compared to the non harvested grooves in Manu basin are incredible higher.

    1. Dear Cesar, Thanks for your comment, you certainly know the situation of Peruvian Brazil nut harvest as well as anyone. This is precisely the point made in the 2011 Scoles & Gribel article in Human Ecology, though over a time scale of centuries: heavily managed Brazil nut groves are much more productive than ones that are abandoned by humans. In fact, "natural" Brazil nut groves with no human intervention go slowly extinct with time, since regeneration falls to nil. Thanks again, Glenn

  2. we have two questions here. one is how much did ancient peoples in amazonia depend on the brazil nut. Considering how much fish and game a well organized tribe can complement a diet based on cassava, I think it is hardly necessary. Other agricultural societies were much larger based on less productive crops, like wheat, maize or rice. It´s much more a problem of social organization than of agricultural technology. The other question is how much they helped spread this tree over long distances.
    There is already a large scientific literature that proves that traditional slash and burn agriculture, hunting and gathering of nuts improves bertholletia`s regeneration. I indeed saw it in the field. On the shores of igarapé pupunha I saw many brazil nut trees growing in the igapó near abandoned vilages. This is obviously impossible to occur without the help of man, and although it does not thrive as much as in the terra firme, it shows how much people can help spread the tree.

    I´d like you to comment on some pictures I took from a tree in São Gabriel da Cachoeira, it is one of the most productive trees I´ve ever seen, in a region with a very short dry season (I think there is hardly a month without rain in a decade there), which goes against general observations of most brazil nut trees.

    good blog, congratulations!

  3. Dear Marcos,

    Thanks for your knowledgeable observations and questions. Hunter-gatherer populations 11,000 years ago already seemed to be eating Brazil nuts. No one "depends" on Brazil nut, but considering how nutritious and delicious it is, and how adaptable it is to anthropogenic environments, it seems natural that people would use and manage it. But other than that one mention of Brazil nut from Pedra Pintada, and lots of contemporary evidence, we don't really know how extensive Brazil nut use was in ancient times.

    Nor do we have clear evidence how, when or if it was cultivated or managed. The genetic and linguistic evidence suggests a pattern of expansion that would have to be confirmed by additional work especially by archeologists.

    I'd be glad to look at your photos. I've seen Brazil nuts planted in native gardens all the way up on the Içana River, it's easy enough for people to carry and plant it, which makes me suspect they have done so for quite a long time. Send me your email and we can correspond further.

    Thanks again for your comments,


  4. Glenn: I'm glad to be reminded of Ducke's prescience about the anthropic Brazil nut groves and to hear about your and the others recent studies confirming it. And thank you for remembering Caverna de Pedra Pintada's Terminal Pleistocene Brazil nut specimen, for which Scott Mori confirmed the botanical identification.

  5. Also, Glenn and Marcos: While working at lithic sites in the middle Xingu along the Curua river, traveling in riverboats, we saw many Brazil nut groves at abandoned villages. They furnished a great supplement to our monotonous diet of rice, beans, and manioc.

    1. Yes, it took a while to dig into the nested Matrioshka-doll references and work my way back to Ducke's original hypothesis. Always interesting to hear more evidence, circumstantial as it may be, about the association of Brazil nut groves with ancient sites of human occupation. The "smoking gun" is going to have to come from archeology, not ethnobotany, so I urge you all to go out there and please! Prove me wrong! Or right! Or some other explanation altogether different! Thanks for the feedback, Glenn