INSTITUTO NACIONAL DE PESQUISAS DA AMAZÔNIA – INPA PROGRAMA DE PÓS-GRADUAÇÃO EM ECOLOGIA

ECOLOGIA HISTÓRICA DE FLORESTAS DA BACIA DO RIO IÇANA, ALTO RIO NEGRO, AMAZONAS: UM LEGADO BANIWA

NAS PAISAGENS

JULIANO FRANCO DE MORAES

Manaus, Amazonas Junho, 2016

JULIANO FRANCO DE MORAES

ECOLOGIA HISTÓRICA DE FLORESTAS DA BACIA DO RIO

IÇANA, ALTO RIO NEGRO, AMAZONAS: UM LEGADO BANIWA

NAS PAISAGENS

ORIENTADOR: DR. GLENN HARVEY SHEPARD JR.

COORIENTADOR: DR. CHARLES ROLAND CLEMENT

Dissertação apresentada ao Instituto

Nacional de Pesquisas da Amazônia como

parte dos requisitos para obtenção do título

de Mestre em Biologia (Ecologia).

Manaus, Amazonas Junho, 2016

ii

Banca examinadora da defesa oral pública:

Dra. Joana Cabral de Oliveira (UNICAMP)

Dra. Helena Pinto Lima (MPEG)

Dra. Juliana Schietti de Almeida (INPA)

iii

Ficha catalográfica:

M827e Moraes, Juliano Franco de Ecologia histórica de florestas da bacia do rio Içana, alto rio Negro, Amazonas: um

legado Baniwa nas paisagens / Juliano Franco de Moraes. --- Manaus: [s.n.], 2016. 58f. : il., color. Dissertação (Mestrado) --- INPA, Manaus, 2016. Orientador: Glenn Harvey Shepard Jr.. Coorientador: Charles Roland Clement. Área de concentração: Biologia (Ecologia).

1.Ecologia histórica. 2. Ecologia histórica – território Baniwa. 3.Floresta – Rio Içana. I.Título

CDD 574.5

Sinopse:

Estudou-se a relação entre a composição florística atual de florestas de terra firme da bacia do rio Içana (alto rio Negro, AM, Brasil) e o manejo florestal de séculos atrás praticado pelos índios Baniwa. Foi avaliado se a composição florística é relacionada ao manejo ou se somente é relacionada a condições edáficas.

Palavras-Chave: domesticação, manejo, composição florística, florestas culturais, florestas de interflúvio

iv

Em memória de Eduardo Filhos do Sertão, grande apreciador e amante da natureza,

uma inspiração por toda a vida.

v

Agradecimentos: À minha família por todo o amor e apoio que me deram e por terem sempre me incentivado a enfrentar o novo para alcançar meus sonhos e planos de vida, como esse mestrado.

Ao Glenn Shepard (MPEG), pela orientação e oportunidade da realização do projeto, por ter me incentivado a vestir a pele da onça, e por ter confiado em mim e feito com que meu sonho de vivenciar a cultura indígena e desvendar os mistérios das florestas do alto rio Negro se realizasse.

Ao meu co(des)orientador, Charles Clement (INPA), por todo o embasamento teórico e ensinamento que me transmitiu, toda confiança e entusiasmo que me passou, pelas centenas de conversas que cada vez mais faziam-me apaixonar pela floresta Amazônica e pela história humana inerente a ela, e por estar sempre presente quando necessário.

Aos docentes do INPA, pelo conhecimento transmitido por meio das disciplinas

Aos avaliadores do plano de dissertação e aos membros da aula de qualificação: Ulisses Albuquerque (UFRPE), Henrique Pereira (UFAM), Eduardo Neves (USP) e Maria Piedade (INPA).

A todos os pesquisadores do INPA que, de alguma maneira, colaboraram com a realização deste projeto: Bruce Nelson; Charles Zartman; Fabrício Baccaro; Juliana Schietti; Mike Hopkins; Newton Falcão; Niro Higushi; e Paulo Bobrowiec.

À Flávia Costa, pelos ensinamentos estatísticos, pela colaboração nas análises dos dados, e pelas revisões do artigo.

Muito obrigado a todos do Laboratório Temático de Solos e Plantas (LTSP) do INPA: Jonas, Gabriela, Raimundo, Edivaldo, Roberta, Mozanei e, em especial, a Laura por terem me auxiliado durante as análises do solo.

Meu profundo agradecimento ao Sérgio Tadeu Meirelles (USP), por ter me ensinado a fazer ciência, me incentivado a desvendar a floresta Amazônica, por ter dado o pontapé inicial nas ideias desse projeto, e sem o qual nada disso teria acontecido.

Ao pessoal do Instituto Socioambiental: Beto Ricardo, Adeílson Lopes e Wilzer, pelo apoio logístico em São Gabriel da Cachoeira. Adeílson, em especial, por ter me ajudado em momentos decisivos desse mestrado.

À Natalia Camps, por ter financiado o campo piloto, pelo auxílio na organização dos campos, e pela companhia durante alguns momentos na Terra Indígena e em São Gabriel da Cachoeira.

vi

Aos índios Baniwa da bacia do rio Içana pelo acolhimento durante os 70 dias que passei morando em suas aldeias e vivenciado seus modos de vida. Com os Baniwa aprendi a respeitar a floresta, os animais, e todas as forças que ali vivem. Aprendi que o conhecimento indígena é tão rico em sabedoria quanto qualquer outro conhecimento. Aprendi que com conscientização, respeito e conhecimento, a floresta pode ser preservada e seus recursos utilizados. E aprendi que o conhecimento ecológico indígena é tão (ou mais) profundo quanto o científico.

Aos meus ajudantes de campo Baniwa: Gerunso, Paulo Filho e André (comunidade Mauá Cachoeira); Plínio, Pedro, Samuel e Wilson (comunidade Tucumã Rupitá); Francisco, Charles e Carlos (comunidade Bela Vista); Januário e seu filho, e Armando (comunidade Santa Marta); Gilberto e Názário (comunidade Santa Rosa).

Um especial e ilustre agradecimento ao Baniwa Armindo Brazão por ter me acompanhado durante todos os dias de meu trabalho de campo e me ensinado sobre botânica e ecologia tanto quanto outros professores acadêmicos me ensinaram.

Ao CNPQ, pela bolsa de mestrado concedida.

À National Geographic Society, pelo financiamento do projeto, por ter acreditado na realização do mesmo, e confiado em meu potencial.

Ao amigo Valdecy e sua família, pelo super apoio quando cheguei em Manaus.

A todo pessoal da república do Hospício (antigos e atuais) que me receberam desde o princípio em Manaus e fizeram meus dias mais alegres: Erickson, Shirley, Betão, Maíra (ai ai!), Tarcísio, Iñaki, Marco, Aroldo, Ramon, Renata, Cinthya, Rubens, Israel e Patrick. Essa casa doida, cujo os 6 gatos são os únicos sãos, com seus integrantes descontínuos lókis, me fez passar ótimos momentos que nunca esquecerei.

À toda galera da minha turma de Mestrado em Ecologia, pelos ótimos momentos durante as disciplinas (em aula e em campo), e pelas conversas e risadas. Um agradecimento em especial a querida Clarice Matos, pelo companheirismo nestes 2 anos, pelo carinho de todos os dias, e por estar comigo nos momentos alegres e tristes, sempre me apoiando.

À Maihyra (INPA) e ao Val Kinupp (IFAM) pela ajuda na identificação das plantas.

Ao Tijolo (UFAM) pelo empréstimo do GPS e pelo auxílio com o método da amostragem de solo.

À Val e todos os outros da secretaria da PG-ECO, pela ajuda com as burocracias do mestrado em momentos importantes.

vii

“ O aspecto que mais chamou a atenção naquela viagem foi o grande número de aldeias indígenas à beira do rio. Toda vez que a expedição tentava encontrar um remanso para descansar, deparava com mais aldeias e mais flotilhas de canoas hostis. No dia 25 de junho [de 1542] eles foram atacados por mais de duzentas

canoas, cada qual com uns vinte ou trinta índios, provavelmente tapajós. Carvajal ficou impressionado com as decorações plumárias dos índios, bem como com sua música: eles entravam em combate gritando e tocando trombetas, tambores, flautas e rabecas

de três cordas, enquanto seus companheiros, em terra firme, saltavam, gritavam e agitavam ramos de palmeira à passagem das embarcações. O território dessa tribo

era densamente povoado e se estendia por cerca de 240 quilômetros. Aqueles índios rejeitaram a tentativa da expedição de fazer escambo de alguns objetos e assediou-a

até ela sair de seu território. ”

John Hemming, Ouro Vermelho, pág. 290.

viii

Resumo

A floresta Amazônica é a maior floresta tropical do mundo e abriga uma enorme proporção da biodiversidade do planeta. Povos indígenas convivem com essa biodiversidade há milhares de anos e dela adquirem seu sustento por meio de um preciso conhecimento acerca das relações bióticas e abióticas. Durante muito tempo acreditou-se que em períodos pré-Colombianos esses povos não teriam se adaptado às condições da floresta - como solos pobres e ácidos, e baixa quantidade de proteína disponível - e que esse ambiente, considerado hostil pelo homem moderno, teria imposto grandes dificuldades aos povos, tornando-os meros animais passivos em meio ao “inferno verde” e não permitindo que grandes populações se firmassem. Entretanto, esse cenário tem sido modificado uma vez que pesquisas tem demonstrado que grandes populações indígenas pré-Colombianas viveram na floresta Amazônica e a alteraram por meio de mudanças em seu solo, sua composição de espécies e sua paisagem para que o ambiente florestal se tornasse adequado para atender suas demandas. Informações a respeito da maneira como esses povos modificaram as paisagens sem alterar a resiliência da floresta e preservando sua biodiversidade, assim como da extensão com que essas modificações ocorreram, permanecem pouco conhecidas, principalmente em florestas de interflúvio. Nosso objetivo nesse estudo foi avaliar o impacto antrópico, de séculos atrás, nas florestas de interflúvio da bacia do rio Içana, alto rio Negro, Brasil, e verificar se ele possui relação com a composição florística atual. Inventários florísticos foram realizados em 12 parcelas localizadas em florestas de áreas de antigo manejo Baniwa e 4 parcelas localizadas em florestas imemoriais para os Baniwa quanto ao manejo. Amostras de solo foram coletadas para análises de areia, ECEC, fósforo, carvão e pH. Entrevistas foram feitas com os indígenas para obtermos informações orais sobre o manejo das paisagens. As florestas de terra firme do território histórico Baniwa, localizado na bacia do rio Içana, que estão a até 750 m de quaisquer cursos d’água, contêm cerca de 14,7% de abundância de espécies arbóreas manejadas e são formadas por um mosaico de paisagens culturais (áreas habitadas abandonadas pelos Baniwa a mais de 200 anos) e paisagens antigas imemoriais (áreas que os Baniwa se referem como intactas há centenas de anos, não sabendo se foram ou não manejadas). Em paisagens culturais a abundância de espécies manejadas chegou a até 57% e a área basal total média foi de 41,4 m² ha-1, sendo 40% dessa biomassa pertencente às espécies manejadas. Paisagens antigas imemoriais contêm em média 7,7% de abundância de espécies manejadas e área basal total média de 35,1 m² ha-1 - com somente 8% pertencente a espécies manejadas. O manejo histórico Baniwa nas paisagens tornou os solos da região menos ácidos por meio de queima controlada que deixou grandes quantidades de carvão nos solos. Os dados de carvão, associados a dados biolinguísticos e históricos, nos permitem sugerir que o manejo ocorre na região há mais de 4000 anos, época em que os ancestrais Aruak dos Baniwa chegaram à região. Antigas malocas localizadas em meio as matas e antigas áreas de manejo em meio a floresta continham, no passado, grandes quantidades de espécies manejadas que estavam sob cuidados dos Baniwa. Uma vez abandonadas, essas áreas se transformaram em florestas maduras com abundância de espécies manejadas - um legado Baniwa nas florestas deixado por meio da domesticação da paisagem. Assim como a domesticação das florestas de interflúvio da bacia do

ix

rio Içana, grandes porções de florestas de interflúvio da Amazônia podem ter sido domesticadas em períodos pré-Colombianos. Atualmente, muitas destas florestas estão distribuídas em muitos territórios indígenas, e mantê-las protegidas é fundamental para se evitar que milhares de espécies sejam extintas. Frente à expansão agrícola e à exploração madeireira, a proteção dessas florestas é um dos maiores desafios ambientais e deve ocorrer com auxílio dos povos que as manejam há séculos.

x

Abstract The Amazon rainforest is the largest rainforest of the world and is home to

a large proportion of the planet's biodiversity. Indigenous peoples live with this biodiversity for thousands of years and it acquire their livelihood through a precise knowledge about the biotic and abiotic relations of the forest. For a long time it was believed that in pre-Columbian times these people would not have adapted to forest conditions - such as poor and acid soils, and low amount of protein available - and that this environment, considered hostile by modern man, would have imposed major difficulties to people, making them mere passive animals amid the "green hell" and not allowing large populations establish themselves. However, this scenario has been modified since research has shown that large pre-Columbian indigenous population lived in the Amazon rainforest and altered it through changes in soil, species composition and landscape to the forest environment became suitable to meet their demands. Information about the way these people changed the landscapes without changing the forest resilience and preserving biodiversity, as well as the extent to which these changes occurred, remain poorly known, particularly in interfluvial forests. Our goal in this study was to assess the human impact of centuries ago in interfluvial forests of Içana river basin, upper Negro river, Brazil, and to verify if it has relation with the current floristic composition. Floristic inventories were carried out in 12 plots located in old management forests and 4 plots located in immemorial forests to the Baniwa as regards the management. Soil samples were collected for analysis of sand, ECEC, phosphorus, carbon and pH. Interviews were performed with indigenous to obtain oral information about the management of landscapes. Interfluvial forests from the historical territory Baniwa, located in the basin of Içana river, which are up to 750 m of any water courses, contain about 14.7% of abundance of managed tree species and are formed by a mosaic of cultural landscapes (old inhabited areas abandoned by Baniwa over 200 years) and imemorial ancient landscapes (areas that Baniwa refer to as intact for hundreds of years, not knowing whether or not managed). In cultural landscapes the abundance of managed species reached up to 57% and the average total basal area was 41.4 m² ha-1, 40% of biomass belonging to the species managed. Immemorial ancient landscapes contain average 7.7% of abundance of managed species and avarage total basal area of 35.1 m² ha-1 - with only 8% belonging to managed species. The historical management Baniwa in landscapes became soil less acidic through controlled burns that left large amounts of charcoal. Charcoal data associated with biolinguísticos and historical data, allow us to suggest that the management occurs in the region for over 4000 years, a time when the ancestors of the Aruwak Baniwa arrived in the region. Old indigenous roundhouses located amid the forest, and old management areas amid forest contained in the past large amounts of managed species that were under care of the Baniwa. Once abandoned, these areas became mature forests with abundance of managed species - one Baniwa legacy left in the forests through the landscape domestication. As the domestication of forests interfluvial the basin Içana river, large portions of interfluvial Amazon forests may have been domesticated in pre-Columbian times. Currently, many of these forests are distributed in many indigenous territories and keep them protected is crucial to prevent thousands of species become extinct. Front to agricultural expansion and logging, protection

xi

of these forests is one of the greatest environmental challenges and must take place with the help of people who managed them for centuries

xii

Sumário Banca examinadora da defesa oral pública ……………………………………………... iii

Ficha catalográfica ..................................................................................................... iv

Sinopse ...................................................................................................................... iv

Dedicatória .................................................................................................................. v

Agradecimentos ......................................................................................................... vi

Epígrafe .................................................................................................................... viii

Resumo ...................................................................................................................... ix

Abstract ...................................................................................................................... xi

Lista de tabelas …………………………………………………………………..………… 2

Lista de figuras ………………………………………………………………………….….. 3

Introdução ……………………………………………………………………………….….. 4

Objetivos ……………………………………………………………………...………......... 7

Capítulo I ……………………………………………………………………………….…… 8

Conclusões ………………………………………………………………………………... 40

Apêndice ………………………………………………………………………………..…. 41

Lista de tabelas Capítulo I.

Dados Estendidos Tabela 1. Resumo dos dados estruturais florísticos coletados nas parcelas amostradas. _______________________________________________ 36 Dados Estendidos Tabela 2. Resumo da localização das parcelas e dos valores das análises edáficas. __________________________________________________ 37 Dados Estendidos Tabela 3. Abundância total das 46 espécies mais abundantes nas parcelas amostradas. _______________________________________________ 38

2

Lista de figuras Capítulo I.

Figura 1. Gradientes de composição florística ordenados pelos eixos do NMDS. __ 17 Figura 2. Concentração de carvão no solo à diferentes profundidades em Cultural e Immemorial Ancient Forests. __________________________________________ 18 Figura 3. Comparação entre Cultural e Immemorial Ancient Forests quanto à abundância relativa de espécies manejadas. ______________________________ 19 Figura 4. Mapa da projeção da abundância de espécies manejadas no território Baniwa histórico. ___________________________________________________ 20 Dados Estendidos Figura 1. Mapa com as localizações de antigas áreas de manejo e habitação da região do médio rio Içana, território Baniwa histórico, identificadas pelos informantes das cinco aldeias que participaram do estudo. ___________________ 28 Dados Estendidos Figura 2. 10 espécies com maiores valores de importância ecológica (VIE). ____________________________________________________ 29 Dados Estendidos Figura 3. Comparação entre Cultural Forests (CF) e Immemorial Ancient Forests (IAF) quanto às variáveis edáficas. _________________________ 30 Dados Estendidos Figura 4. pH do solo em relação a concentração de carvão no solo (camada 0-20 cm). __________________________________________________ 31 Dados Estendidos Figura 5. Concentração de carvão no solo em relação a distância ao rio Içana. _______________________________________________________ 32 Dados Estendidos Figura 6. Comparação entre Cultural Forests (CF) e Immemorial Ancient Forests (IAF) quanto às variáveis estruturais. _______________________ 33 Dados Estendidos Figura 7. Riqueza, abundância e área basal relativas de espécies manejadas em relação às distâncias ao rio Içana e ao curso d’água mais próximo. 34

3

Introdução

Há um atual debate científico a respeito do grau de alteração antrópica nas

paisagens pré-Colombianas da Amazônia e de qual seria a proporção de florestas

primárias virgens e florestas antrópicas hoje nessas paisagens. Durante a segunda

metade do século XX, muitos pesquisadores sugeriram muitas hipóteses para

entender o passado Amazônico. Steward (1948) propôs que sociedades complexas

na Amazônia pré-Colombiana seriam resquícios de migrações sub-andinas que ao

se depararem com os limites ambientais impostos pela floresta teriam retraído sua

cultura. Tais culturas, portanto, não teriam se originado por processos endógenos e

sim seriam vestígios de culturas complexas de outros locais. Anos depois, Betty

Meggers (Meggers, 1954) sugeriu que populações humanas na Amazônia pré-

Colombiana foram pequenas e de curta duração devido ao ambiente ser hostil e

dotado de solos pobres, hipótese denominada de “Determinismo Ecológico”. Tal

hipótese foi contraposta por Carneiro (1961) que sugeriu que uma economia

baseada em agricultura de corte e queima, associada a caça e pesca, poderia

suportar densas populações sedentárias na Amazônia pré-Colombiana. Segundo o

autor, o cultivo de mandioca tornou possível a existência de habitações em áreas de

terra firme. Lathrap (1970) advogou que a região central da Amazônia seria o mais

antigo centro de origem da agricultura em toda a América, hipótese que ficou

conhecida como modelo “cardíaco”, o que contrariou as ideias de Steward e

Meggers de que culturas Amazônicas seriam vestígios de culturas andinas.

O debate continua vigente até os dias de hoje, principalmente em relação às

florestas de interflúvio. Alguns arqueólogos e ecólogos tradicionais defendem, assim

como Meggers, que o ser humano teria pouco modificado a floresta amazônica e as

áreas de interflúvio em períodos pré-Colombianos (Bush & Silman 2007; Peres et al.

2010; Barlow et al. 2012; McMichael et al. 2012; Piperno et al. 2015). Eles defendem

que a Amazônia é uma formação ecológica que existe a milhares de anos, porém

com presença humana pré-Colombiana pequena, relativamente recente e com

mínimo grau de impacto sobre as paisagens e processos ecológicos de grande

escala, sugerindo que a maioria das paisagens da Amazônia não possuem legados

humanos na sua composição e não foram domesticadas. Eles sugerem que a maior

parte da Amazônia, principalmente os interflúvios, é virgem e contêm e sempre

conteve baixo impacto e influência antrópica indígena sobre a biodiversidade. Outro

4

ponto de vista está em estudos com abordagem em ecologia histórica e é

compartilhado por muitos antropólogos, arqueólogos e alguns ecólogos (Balée 1989;

Heckenberger et al. 2003; Denevan 2011; Levis et al. 2012; Clement et al. 2015).

Estudos com abordagem ecológica histórica sugerem que a Amazônia é uma vasta

paisagem antrópica domesticada, onde grupos indígenas, desde caçadores-

coletores pré-históricos até os grandes cacicados da época da invasão europeia e

os atuais povos, exerceram e exercem um efeito na formação e na biodiversidade da

maioria das paisagens, incluindo as florestas de interflúvio.

Entretanto, a carência de dados coletados em florestas de interflúvio - devido

principalmente ao seu difícil acesso -, que representam a maioria das florestas da

Amazônia, acarreta em hipóteses não testadas sobre este tipo de ambiente e em

inferências sem embasamento empírico. Em nosso estudo foi avaliado o impacto

indígena de séculos atrás na formação de paisagens de florestais de interflúvio da

bacia do rio Içana, alto rio Negro, AM, Brasil. A pesquisa foi realizada em território

pertencente à etnia Baniwa e elucidou a seguinte questão: Qual a relação entre o

manejo de séculos atrás, praticado pelos Baniwa, e a composição florística atual de

florestas de interflúvio da bacia do rio Içana, alto rio Negro, AM, Brasil?

Contamos com uma interdisciplinaridade pouco recorrente na maioria das

pesquisas ecológicas realizadas no Brasil. Métodos da ecologia, arqueologia, e

antropologia foram utilizados para que existisse uma interação entre vários

conhecimentos, tanto tradicionais quanto científicos. A biodiversidade que queremos

preservar/conservar hoje não se originou somente através de milhões de anos de

evolução e seleção natural, mas também se originou através da relação do ser

humano com a natureza, por meio do conhecimento tradicional, ao longo de

milhares de anos. É necessário conhecer o passado para entender o presente e,

desse modo, compreender que povos indígenas tiveram e têm um papel

fundamental na formação e na proteção das paisagens Amazônicas.

5

Referências bibliográficas:

Balée, W. (1989). The culture of Amazonian forests. In: Posey, D.A.; Balée, W. (Eds.). Resource management in Amazonia: Indigenous and folk strategies, Advances in Economic Botany, 7. Bronx, NY: New York Botanical Garden. p.1-21. Barlow, J., Gardner, T. A., Lees, A. C., Parry, L., & Peres, C. A. (2012). How pristine are tropical forests? An ecological perspective on the pre-Columbian human footprint in Amazonia and implications for contemporary conservation.Biological Conservation, 151(1), 45-49. Bush, M. B., & Silman, M. R. (2007). Amazonian exploitation revisited: ecological asymmetry and the policy pendulum. Frontiers in Ecology and the Environment, 5(9), 457-465. Carneiro, R. L. (1961). Slash-and-Burn cultivation among the Kuikuro and its implications for cultural development in the Amazon basin. In The Evolution of Horticultural Systems in Native South America: Causes and Consequences, a Symposium, edited by J. Wilbert, pp. 47-67. Sociedade de Ciencias Naturales La Salla, Caracas. Clement, C. R. et al. (2015). The domestication of Amazonia before European conquest. Proc. R. Soc. B. 202. Denevan, W. M. (2011). The “pristine myth” revisited. Geographical Review,101(4), 576-591. Heckenberger MJ, Kuiruro A, Kuikuro UT, Russel JC, Schmidt M, et al. (2003) Amazonia 1492: Pristine forest or cultural parkland? Science 301: 1110–1114. Lathrap, D. W. (1970). The Upper Amazon. Praeger, New York. Levis, C., de Souza, P. F., Schietti, J., Emilio, T., da Veiga Pinto, J. L. P., Clement, C. R., & Costa, F. R. (2012). Historical human footprint on modern tree species composition in the Purus-Madeira interfluve, central Amazonia. PloS one, 7(11), e48559. McMichael, C. H., Piperno, D. R., Bush, M. B., Silman, M. R., Zimmerman, A. R., Raczka, M. F., & Lobato, L. C. (2012). Sparse pre-Columbian human habitation in western Amazonia. Science, 336(6087), 1429-1431. Meggers, B. J. (1954). Environmental Limitation on the Development of Culture1. American anthropologist, 56(5), 801-824. Peres CA, Gardner TA, Barlow J, Zuanon J, Michalski F, et al. (2010) Biodiversity conservation in human-modified Amazonian forest landscapes. Biol Conserv 143(10): 2314–2627. Piperno, D. R., McMichael, C. H. & Bush, M. B. (2015). Amazonia and the Anthropocene: what was the spatial extent and intensity of human landscape modification in the Amazon Basin at the end of prehistory? The Holocene 25, 1588-1597. Steward, J. H. (1948). Handbook of South American Indians, edited by J.H. Steward. Bureau of American Ethnology Bulletin 143 Vol. 4. Smithsonian Institution, Washington D.C.

6

Objetivos Objetivo geral:

Testar a hipótese de que a composição florística atual de florestas de interflúvio

da bacia do rio Içana, noroeste da Amazônia, possui relação com o manejo de

paisagens praticado séculos atrás por índios da etnia Baniwa.

Objetivos específicos:

1 - Identificar quais são as espécies arbóreas comestíveis manejadas pelos Baniwa

e quais as práticas de manejo e uso das paisagens;

2 - Verificar se a estrutura florística e as condições edáficas de florestas

historicamente manejadas são diferentes das de florestas imemoriais;

3 - Contextualizar o histórico de ocupação da região.

7

Capítulo I.

______________________________________________________ Moraes, J. F.; Shepard, G. H.; Costa, F. R. C; & Clement, C. R. Landscape domestication in interfluvial forests of northwestern Amazonia. Manuscrito submetido para Nature (England).

8

Landscape domestication in interfluvial forests of northwestern Amazonia Juliano Franco de Moraes 1*

Glenn H. Shepard 1,2

Flávia Regina C. Costa 1

Charles Roland Clement 3

1. Programa de Pós-Graduação em Ecologia, Instituto Nacional de Pesquisas da Amazônia, Avenida Rodrigo Otávio, 35 - Coroado, 69080-971 Manaus, Brazil.

2. Cordenação das Ciências Humanas, Departamento de Antropologia, Museu Paraense Emílio Goeldi, Avenida Perimetral, 1901 - Terra Firme, 66077-830 Belém, Brazil. 3. Coordenação de Tecnologia e Inovação, Instituto Nacional de Pesquisas da Amazônia, Avenida André Araújo, 2936 - Aleixo, 69060-001 Manaus, Brazil. * E-mail do autor: demoraes.franco@gmail.com

9

Landscape domestication in interfluvial forests of northwestern Amazonia Juliano F. Moraes1, Glenn H. Shepard1,2, Flávia R. C. Costa1, Armindo F. B. Baniwa3 & Charles R. Clement4

Recent research has identified significant long-term anthropogenic alterations in

tropical forests1,2. In Amazonia, such “cultural forests” 3 result from indigenous

management that has maintained forest resilience4. However, while anthropogenic

effects along major rivers are accepted5, little is know about interfluvial areas, assumed

to contain pristine forests6,7. Here we show the imprint of past human management on

the floristic composition of mature interfluvial forests in northwestern Amazonia.

Ancestral Baniwa village sites, abandoned centuries ago, have been transformed into

mature cultural forests with 40% of basal area consisting of managed species, compared

with 8% in nearby old-growth forests. These ancestral forests show a signature of

Baniwa occupation with as much as 57% of relative abundance of managed species,

although they are located as far as 13 km from the major river in regions predicted to

have minimal human impact on floristic composition8. High concentrations of charcoal

were found in the top 1 m of soil, providing a key sign of human occupation that is

thought to date back 4,000 years9. We extrapolated these results to predict an overall

15% relative abundance of managed species in interfluvial forests throughout historic

Baniwa territory, reflecting a significant degree of landscape domestication. Other

indigenous peoples may have likewise modified floristic composition throughout

significant areas of Amazonian interfluvial forests10. Protecting such areas helps avert

biodiversity extinctions11 while safeguarding the lifeways and food security of indigenous

and other traditional peoples.

Recent studies have suggested that some regions of pre-Columbian Amazonia were

occupied by large polities that altered their landscape through earth works12, soil

modification13, and agroforestry management resulting in lasting changes to floristic

composition10. These processes increased productivity for human needs and thus represent

1Programa de Pós-Graduação em Ecologia, Instituto Nacional de Pesquisas da Amazônia, Avenida André Araújo, 2936 - Petrópolis, 69067-375 Manaus, Brazil. 2Cordenação das Ciências Humanas, Departamento de Antropologia, Museu Paraense Emílio Goeldi, Avenida Perimetral, 1901 - Terra Firme, 66077-830 Belém, Brazil. 3Escola Indígena Baniwa-Coripaco / Federação das Organizações Indígenas do Rio Negro, Avenida Alvaro Maia, 79 - Centro, 69750-000 São Gabriel da Cachoeira, Brazil. 4Coordenação de Tecnologia e Inovação, Instituto Nacional de Pesquisas da Amazônia, Avenida André Araújo, 2936 - Petrópolis, 69067-375 Manaus, Brazil.

10

landscape domestication14. However the extent of such landscape-level alterations in

Amazonia remains little known and hotly debated5. Most authors assume that human impacts

are restricted to vicinity of large rivers, with minimal impacts in interfluvial areas8.

Recent studies conclude that the floristic composition of interfluvial forests in

Amazonia was unaffected by human management, since no evidence of intensive swidden

agriculture has yet been found in these areas6,7. Although they appear pristine, these forests

could nonetheless bear the legacy of human disturbance from centuries ago15. Pre-Colombian

indigenous populations had a hybrid lifestyle that included hunting, fishing, gathering and

horticulture without necessarily developing a complete dependence on agriculture16. By

clearing spaces for house construction and horticulture, ancient peoples created agroforestry

systems, enriching forests with useful species while suppressing undesired species10,17.

Here we evaluate whether the modern floristic composition of mature interfluvial

forests of northwestern Amazonia is related to ancient human management, or whether it is

related solely to the edaphic conditions. The study was carried out along the Içana River

basin, territory of the Baniwa people of the Arawakan language family. Oral histories place

numerous ancestral Baniwa longhouses in interfluvial areas18, distant from the major river.

Soils are sandy and acidic, so the Baniwa focus their agroforestry activities in more fertile

interfluvial areas which are re-used cyclically, creating a dynamic patchwork of forests with

different degrees of landscape transformation.

Drawing on participatory mapping and oral histories of past habitation sites, we

selected 12 plots in ancestral forests and 4 in old-growth forests (Extended Data Fig. 1).

Ancestral forests are mature forests in the vicinity of ancient Baniwa longhouses where

modern villagers occasionally return to harvest resources. As such, they represent cultural

forests3 that bear the legacy of past human management. Old-growth forests are likewise

mature forests with similar canopy structure, but where the modern Baniwa identify no

ancestral settlements. Given Baniwa oral histories, the size of the trees, and the condition of

ancient houseposts and potsherds, we estimate that ancestral forests emerge around Baniwa

settlements that have been abandoned for at least 200 years.

We set out to test the hypothesis that ancient human management enriched these

ancestral forests with species valued by the Baniwa. We focused on forest trees with edible

fruits that are managed near current communities. Among the 14 focal species are Poraqueiba

sericea, which the Baniwa transplant to gardens and house patios, Humiria balsamifera,

which responds well to fire management, Pouteria ucuqui, which is protected from burning in

11

new swidden sites, and the wild cacao relative Theobroma subincanum, which grows

conspicuously along forest trails, apparently human-dispersed. The abundance of these

species10, the occurrence of soil charcoal19 and the content of Baniwa oral histories18 all point

towards human-induced alteration of floristic composition in apparently “virgin” forests1.

For each plot, edaphic conditions were measured: sand content, soil pH, phosphorus

and charcoal concentrations, and effective cation exchange capacity (ECEC). The forest

community structure was evaluated in terms of: richness, abundance, and relative basal area

of managed species; overall density, abundance, and basal area; and ecological importance

value (EIV). Relative abundance of managed species was then projected to visualize the

domestication of interfluvial forests across Baniwa territory.

Floristic composition was reduced to two dimensions by a NMDS that captured 83%

of floristic variation and showed relation with sand content (P = 0.004) and presence/absence

of past human habitation (P = 0.007). Other edaphic factors (ECEC, phosphorus, pH) showed

no relation with floristic composition (P = 0.704, P = 0.997, P = 0.066, respectively).

Ancestral and old-growth forests showed different floristic composition (Fib. 1b), with

ancestral forests having as much as 57% (mean 21.5% ± 15.3 s.d.) of relative abundance of

managed species, compared with a maximum of 10.5% (mean 7.7% ± 3.8) for old-growth

forests (P = 0.030, Fig. 2; Extended Data Tab. 1). Four managed species, Poraqueiba sericea,

Dacryodes sp., Oenocarpus bacaba and Humiria balsamifera, accounted for 19% of the EIV

in ancestral forest and were among the 10 most important species for these forests. Only one

managed species, Dacryodes sp., was found among the 10 most important species for old-

growth forests, accounting for only 2.3% of EIV (Extended Data Fig. 2). Managed species

accounted for 39.8% of basal area across ancestral forests, compared with only 8% for old-

growth forests (Extended Data Tab.1).

Floristic composition varies along a gradient from more to less sandy soils, but

ancestral and old-growth forests occur along the whole gradient (Fig. 1a). Thus the

differences in composition between the two forest types owe apparently not to soils per se, but

rather to past human management. Soil pH and charcoal concentration in the top 20 cm were

related (R2 = 0.67, P < 0.001, Extended Data Fig. 4) and both were higher in ancestral forests

(P = 0.052, Extended Data Fig. 3; P = 0.009, Fig. 3). Past Baniwa management reduced soil

acidity through periodic small burning, releasing ash containing calcium and magnesium into

the soil, as is the case for anthropogenic dark earths13.

12

Although we found differences in charcoal concentration for the top 20 cm soil layer,

below that depth (20-60 cm) charcoal concentrations were similar for ancestral and old-

growth forests (Fig. 3). Charcoal found in the top 20 cm in Amazonia is attributed to human

activities in the post-colonial period, while 95% of the charcoal found below 20 cm dates to

the pre-Colombian period, from 500 to as much as 8000 years ago6,19,20. Our data show a

greater intensity of burning in ancestral forests when compared with old-growth forests over

the past few centuries, but similar intensity in earlier times. This suggests that the observed

differences in floristic composition between ancestral vs. old-growth forests are due to

management by ancestral Baniwa populations during the post-colonial period18.

Charcoal concentration in the 0-20 cm and 20-60 cm soil layers does not diminish with

distance from the Içana River (P = 0.305 and P = 0.559, respectively; Extended Data Fig. 5),

as some authors would predict8. Charcoal was found as deep as 80-100 cm in some localities.

Besides indirect evidence of human habitation from charcoal in ancestral forests, we also

found direct surface evidence, including potsherds and durable wooden houseposts of

Minquartia guianensis. Based on charcoal, archeological studies9, and Arawakan pre-history21

we estimate that the forests of the Içana River basin have been occupied by humans for as

much as 4,000 years, with intensive management occurring over the past few centuries.

Ancestral and old-growth forests show similar values for overall density (mean ±

s.e.m. = 56.8 ± 4.0 and 63.5 ± 7.2 species per 0.086 ha, respectively), abundance (2,929 ± 115

and 3,156 ± 221 individuals per ha) and total basal area (41.4 ± 3.1 and 35.1 ± 1.5 m2 per ha)

(P > 0.05; Extended Data Fig. 6 Tab. 1). Basal area is a good predictor of tree biomass22 and

forest maturity23 in the upper Negro basin. Biomass resilience in neotropical forests varies

widely24; for the upper Negro was estimated that after 190 years, cleared forests attain the

basal area of mature forests, calculated at 34.8 m2 ha-1 (ref. 23). We measured an almost

identical value for old-growth forests (35.1), and a higher value for ancestral forests (41.4).

The higher basal area in ancestral forests may owe to improvements in soil fertility brought by

burning, or to the selection of high-biomass managed species like Poraqueiba sericea and

Pouteria ucuqui.

The richness, abundance and relative basal area of managed species did not decrease

up to 12.7 km from the Içana River, nor as far as 750 m from medium and small watercourses

(P > 0.05; Extended Data Fig. 7). The most distant ancestral forest surveyed, 12.7 km in a

straight line from the Içana, is considered by Baniwa to represent the site of an ancient

longhouse whose descendants currently reside in the village of Bobope (Bela Vista), 19 km

13

away (Extended Data Tab. 2). Combining this cartographic information with our data on

abundance, we project that managed species represent 14.7% (calculation in Fig. 4) of the

species abundance of all forests in the region located as far as 19 km from any currently

existing village, and as far as 750 m from any watercourse (Fig. 4), indicating a high degree

of landscape domestication. Satellite imagery of Baniwa territory beyond the sampled region

likewise shows intensive ongoing management near watercourses (Extended Data Fig. 1,

colored boxes), reinforcing the results of this projection. We considered that all 110 modern

Baniwa villages must have the same management pattern of the villages included in this

study.

This result contradicts prior studies that question any enduring legacy of ancient

human management on floristic composition for Amazonia interfluvial forests6,7. Recent

studies drawing on climatic data and lack archeological evidence failed to identify the Içana

basin as a likely area for cultural forests, instead predicting a pristine vegetation8,25. Multiple

approaches26 including ecological, remote sensing, archeological and ethnographic methods

are required to detect and document anthropogenic landscapes. This is the first study with

robust empirical data comparing ancestral and old growth forests, and reveals that interfluvial

forests can bear lasting legacies of ancient human management and landscape domestication,

manifest in the modification of soils and the concentration of useful species in areas of past

occupation10,17.

This result has important implications for contemporary Amazonia conservation.

Projecting current trends of agricultural expansion, 40% of Amazonian forests will no longer

be standing by 205027. Selective logging28 and overhunting of terrestrial frugivores29 may

reduce biodiversity and ecosystem functions long-term. Though their role in biodiversity

conservation has been questioned30, indigenous and other traditional peoples are increasingly

seen as key allies in defending forests, managing resources and preventing species extinctions

into the future11. Understanding how these people have domesticated forest landscapes in the

past provides a paradigm-shifting scientific background for contemporary management and

long-term socio-environmental resilience.

1. Willis, K. J. et al. How “virgin” is virgin rainforest? Science 304, 402-403 (2004). 2. Clement, C. R. et al. The domestication of Amazonia before European conquest. Proc. R. Soc. B. 202, http://dx.doi.org/10.1098/rspb.2015.0813 (2015). 3. Balée, W. (ed.) Cultural Forests of the Amazon: A Historical Ecology of People and Their Landscapes. (University of Alabama Press, Tuscaloosa, USA, 2013). 4. Clement, C. R. & Junqueira, A. B. Between a pristine myth and an impoverished future. Biotropica 42, 534-536 (2010).

14

5. Tollefson, J. Footprints in the forest. Nature 502, 160–162 (2013). 6. McMichael, C. H. et al. Sparse pre-Columbian human habitation in western Amazonia. Science 336, 1429-1431 (2012). 7. Piperno, D. R., McMichael, C. H. & Bush, M. B. Amazonia and the Anthropocene: what was the spatial extent and intensity of human landscape modification in the Amazon Basin at the end of prehistory? The Holocene 25, 1588-1597 (2015). 8. Bush, M. B. et al. Anthropogenic influence on Amazonian forests in pre‐history: An ecological perspective. Journal of Biogeography 42, 2277-2288 (2015). 9. Zucchi, A. A new model of the northern Arawakan expansion. In Comparative Arawakan Histories: Rethinking Language Family and Culture Area in Amazonia (eds. Hill, J. D. & Santos-Granero, F.) 199-222 (University of Illinois Press, Champaign, USA, 2002). 10. Levis, C. et al. Historical human footprint on modern tree species composition in the Purus-Madeira interfluve, central Amazonia. PLoS ONE 7, http://dx.doi.org/10.1371/journal.pone.0048559 (2012). 11. ter Steege, et al. Estimating the global conservation status of more than 15,000 Amazonian tree species. Sci. Adv. 1, http://dx.doi.org/10.1126/sciadv.1500936 (2015). 12. Heckenberger, M. J. Amazonia 1492: pristine forest or cultural parkland? Science 301, 1710-1714 (2003). 13. Schmidt, M. J. et al. Dark earths and the human built landscape in Amazonia: a widespread pattern of anthrosol formation. J. Archaeol. Sci. 42, 152–165 (2014). 14. Clement, C. R. 1492 and the loss of Amazonian crop genetic resources. I. The relation between domestication and human population decline. Economic Botany. 53, 188–202 (1999) 15. Stahl, P. W. Interpreting interfluvial landscape transformations in the pre-Columbian Amazon. The Holocene 25, 1598-1603 (2015). 16. Prestes-Carneiro, G., Béarez, P., Bailon, S., Py-Daniel, A. R. & Neves, E. G. Subsistence fishery at Hatahara (750–1230 CE), a pre-Columbian central Amazonian village. J. Archaeo. Sci. Reports, http://dx.doi.org/10.1016/j.jasrep.2015.10.033 (2015). 17. Shepard, G. H. & Ramirez, H. “Made in Brazil”: Human dispersal of the Brazil Nut (Bertholletia excelsa, Lecythidaceae) in ancient Amazonia. Economic Botany 65, 44-65 (2011). 18. Wright, R. M. (ed.) História Indígena e do Indigenismo no Alto Rio Negro. (Mercado de Letras, Campinas, Brazil, 2005). 19. Sanford, R. L., Saldarriaga, J., Clark, K. E., Uhl, C. & Herrera, R. Amazon rain-forest fires. Science 227, 53–55 (1985). 20. McMichael, C. H., Correa-Metrio, A. & Bush, M. B. Pre-Columbian fire regimes in lowland tropical rainforests of southeastern Peru. Palaeogeography, Palaeoclimatology, Palaeoecology 342, 73-83 (2012). 21. Holman, E. W. et al. Automated dating of the world’s language families based on lexical similarity. Current Anthropology 52, 841-875 (2011). 22. Lima, A. J. N. et al. Allometric models for estimating above-and below-ground biomass in Amazonian forests at São Gabriel da Cachoeira in the upper Rio Negro, Brazil. Forest. Ecol. Manag. 277, 163-172 (2012). 23. Saldarriaga, J. G., West, D. C., Tharp, M. L. & Uhl, C. Long-term chronosequence of forest succession in the upper rio Negro of Colombia and Venezuela. Journal of Ecology 76, 938-958 (1988). 24. Poorter, L. et al. Biomass resilience of Neotropical secondary forests. Nature 530, 211-214 (2016). 25. Mcmichael, C. et al. Phytolith assemblages along a gradient of ancient human disturbance in western Amazonia. Front. Ecol. Evol. 3, 141 (2015). 26. McClenachan, L., Cooper, A. B., McKenzie, M. G. & Drew, J. A. The importance of surprising results and best practices in historical ecology. BioScience 65, 932-939 (2015). 27. Soares-Filho, B. S. et al. Modelling conservation in the Amazon basin. Nature 440, 520-523 (2006). 28. Nepstad, D. et al. Frontier governance in Amazonia. Science 295, 629–631 (2002). 29. Peres, C. A., Emilio, T., Schietti, J., Desmoulière, S. J. & Levi, T. Dispersal limitation induces long-term biomass collapse in overhunted Amazonian forests. PNAS 113, 892-897 (2016). 30. Terborgh, J. & Peres, C.A. The problem of people in parks In: Making Parks Work, Strategies for Preserving Tropical Nature. (eds. Terborgh, J., Schaik, C. von, Davenport, L. & Madhu, R.) 307-319 (Island Press, Washington DC, USA, 2002).

15

Acknowledgements: We thank the Baniwa communities of Bobope (Bela Vista), Ttdzealinomana (Santa Marta), Hizdapada (Mauá Cachoeira), Komalhipani (Tucumã Rupitá) and Owhika (Santa Rosa) for their hospitality and patience during field work, the Instituto Socioambiental for logistical support in São Gabriel da Cachoeira, the Federação das Organizações Indígenas do Rio Negro and the Organização Indígena da Bacia do Içana for authorizing the study and providing assistance when necessary, Armindo Baniwa (our local research collaborator) for assistance thoughout the field work, the Baniwa field assistants Gerunso, Paulo Filho, André, Plínio, Pedro, Samuel, Wilson, Francisco, Charles, Carlos, Januário, Armando, Gilberto and Názário, Laura Oliveira of INPA’s Thematic Laboratory of Plants and Soils for assistance with soil analyses, Bruce Nelson, INPA, and André Arruda, Greenpeace, for assistance and elaboration of maps, respectively, the Conselho Nacional de Desenvolvimento Científico and Tecnológico (CNPq) for JFM’s scholarship and research fellowships for CRC, FRCC and GHS. Field research was funded by National Geographic Society grant 9785-15.

Contribution of authors: JFM, GHS and CRC conceived of the project; JFM collected field data and carried out laboratory analyses; AFBB accompanied all phases of field research, translated and facilated community contacts and participated in site selection and identificaion; JFM and FRCC analysed the data; JFM wrote the initial draft; GHS translated and adapted this draft into English; JFM, GHS, CRC and FRCC revised successive drafts.

Author Information: The authors declare no competing financial interests. Readers are welcome to comment on the online version of the paper. Correspondence and requests for materials should be addressed to JFM (demoraes.franco@gmail.com).

16

Figure 1 - Gradients of floristic composition ordered by NMDS. a, ordered by axis 1; b, ordered by axis 2. The 46 most abundant species were used for the analysis and represent 72% of total abundance (Extended Data Tab. 3). Black columns represent plots in ancestral forests (N = 12) and red columns in old growth forests (N = 4). The size of the rectangular boxes within columns represents the relative abundance of species in their respective plots. Blue rectangular boxes represent managed species. Ancestral and old growth forests occur throughout the floristic gradient related to sand (a), but they are grouped differently in terms of management (b), because ancestral forests have greater abundance of managed species (Fig. 2). Two species were impossible to identify because of damage to the field voucher specimen during transport. These species have an asterisk in the figure and their indigenous names are given.

17

Figure 2 - Comparison between ancestral and old growth forests in terms of relative abundance of managed species. In ancestral forests, N = 12 and median = 16.7. In old growth forests, N = 4 and median = 9.0. Ancestral forests have greater abundance of managed species than old growth forests (P = 0.030).

18

Figure 3 - Concentration of charcoal in the soil at different depths in ancestral and old growth forests. Black dots represent samples located in ancestral forests and grey dots in old growth forests. In the 0-20 cm layer, N = 11 in ancestral forests and N = 4 in old growth forests; in the 20-60 cm layer, N = 13 and N = 4, respectively; in the 80-100 cm layer, N = 6 and N = 1, respectively. Two plots located in ancestral forests were not included in the 0-20 and 20-60 cm analyses because the charcoal was damaged. Two extra samples of charcoal were included in the analysis of the 20-60 cm layer from ancestral forests that were not inventoried.

19

Figure 4 - Projection of the relative abundance of managed species in the Baniwa ancestral territory. Red dots represent plots sampled in ancestral forests (N = 12) and yellow dots in old growth forests (N = 4). Green dots represent current Baniwa communities (Source: Instituto Socioambiental – ISA, unpublished data). Areas in blue represent forests with a mean abundance of 14.7% of managed species and are ≤ 750 m from watercourses. The average of abundance of managed species in ancestral forests was 21.5% (max. 57%, mín. 5.4%) and in old growth forests was 7.9% (max. 10.9%, min. 2.1%). The unweighted arithmetic mean (21.5 + 7.9 / 2) was 14.7%, since the region contains a mosaic of ancestral and old growth forests.

20

Methods.

Study area. Fieldwork was conducted in campinarana forests of the middle

Içana River basin, a tributary of the upper Negro River, in the municipality of São

Gabriel da Cachoeira, Amazonas state, in northwestern Brazilian Amazonia. The

region is located on the Guiana Shield, a Precambrian formation of granitic and

granulite rocks31, and has predominantly sandy, oligotrophic, acidic soils (podzols),

with vegetation ranging from closed canopy forests (campinaranas) on less sandy

soils to open forests (campinas) on very sandy soils32. The annual temperature and

precipitation in the region are 25 °C and 3.500 mm, respectively, and the driest

quarter corresponds to January-March, which has an average of 706.3 mm33. The

source of the Içana River is located in Colombia, but most of the river's 580 km

length is in Brazil. The Içana River basin is inhabited by the Curipaco (upper Içana)

and Baniwa (middle and lower Içana) ethnic groups.

Baniwas. The Baniwa ethnic group belongs to the Arawak linguistic trunk. The

people of this ethnic group call themselves Newiki in their own language and their

social organization is divided into phratries (clans who claim descendants from

common ancestors), such as the Hohodene, Walipere-dakeenai and Dzawinai34. Each phratry claims an ancestral territory encompassing one or more lakes and

streams. When a phratry intends to use resources from the territories of others, they

generally request permission35. Many Baniwa myths and legends describe clan wars

due to conflicts over resources and territory36. Their traditional system of resource

management consists of hunting, fishing, gathering of forest products and shifting

cultivation. The ancestors of the Baniwa, however, may have depended less on

agriculture and more on fishing and gathering19. In the past, Baniwa villages

consisted of multi-family longhouses (malocas) that contained 40-60 people36-38, and

were located mainly away from major rivers along secondary streams18,35,39. Due to

the influence of Salesian missionaries, the military, and the creation of the

municipality of São Gabriel da Cachoeira, the Baniwa progressively abandoned their

longhouses in the interior of the forest and settled in village communities on the

banks of the main rivers of the region, the Içana and Ayari18,39. The old village sites

were abandoned and after centuries have turned into stately, apparently

pristineforests, but upon closer investigation reveal their human legacy. Some

studies consider the Içana River basin as a candidate for the ancient Arawak

21

homeland9 and language reconstructions identify Arawak horticultural crop names

going back 4100 years21. The oldest pottery and charcoal found in the upper Rio

Negro are dated to 3750 and 6260 years ago, respectively, and were found in the

area of San Carlos19, southern Venezuela, about 140 km from the middle Içana

River, which supports the inference of ancient occupation along the Içana River and

its surroundings.

Sampling design. Fieldwork was conducted between September and

November 2015. An initial meeting with Baniwa representatives was held to explain

research objectives and request permission to carry out the reserach in their territory.

With advice from Baniwa representatives and researchers, one Portuguese-speaking

Baniwa man with extensive knowledge about forest management agreed to assist

with fieldwork and facilitate contact with the villages andtranslation. Based on

recommendations from these Baniwa representatives and researchers , five

communities were selected for fieldwork in ancestral territories: Bobope (or Bela

Vista), Hidzapada (or Mauá Cachoeira), Komalhipani (or Tucumã Rupitá), Owhika (or

Santa Rosa) and Ttdzealinomana (or Santa Marta) In all communities, the first step

was participatory mapping of the territory with indigenous informants who had

extensive knowledge of the landscape, its management and its history of occupation.

The activity generated maps that identified current and ancient villages and

swiddens, sacred sites, and small streams. These old village sites and swiddens are

now considered ancestral forests. Based on the maps, the logistics required to

access areas, the availability of indigenous men to assist us and the estimated age of

the areas, 12 ancestral forests and 4 old growth forests were selected. Logistical

difficulties limited the number of old growth sites we could visit. The age of the

selected areas was estimated by interviewing elderly informants between 60 and 84

years, who remembered the ancient dwellings as places that "grandparents of

grandparents" reported to be inhabited in the past. This information was passed on

through generations and many of the older people remember finding houseposts

(Minquartia guianensis trunks that support the maloca structures) in these locations.

Thus we estimate these sites were abandoned at least 200 years ago. The areas of

old growth forests are places that the indigenous informants recognize as never have

been inhabited. They are therefore immemorial forests in terms of management

22

though not necessarily virgin forests, as they may have been used in pre-Columbian

times.

Data collection. Two indigenous helpers from each community that had

extensive knowledge about the names of trees helped to collect data in the areas of

their respective communities. In each area predetermined by the participatory map, a

plot of 10 x 72 m (0.072 ha) was established randomly to inventory the trees ≥ 10 cm

diameter at breast height and measure their circumference at breast height (130 cm).

The indigenous name and use of each tree was provided by the helpers. A list of

plants likely to be found in the area was drafted prior to the field work, each with its

Baniwa and scientific names40-43. The list was used in the field as a guide to

botanically collect species that had not been identified in previous studies. All trees

identified were collected and taken to the National Research Institute for Amazonia

(INPA, campus Manaus, AM, Brazil) for expert identification. The species identified in

this study are deposited in the Herbarium of the Federal Institute of Education,

Science and Technology of Amazonas (EAFM / IFAM, Manaus, AM, Brazil). The

species identified in previous studies are deposited in three herbaria: EAFM (São

Gabriel da Cachoeira and Manaus) and INPA (Manaus). Two soil samples were

collected in each plot using a post-hole digger, one at the beginning and one at the

end. Each sample was divided in two subsamples: 0-20 (discounting the litter layer

and surface roots) and 20-60 cm. When logistically possible, the 80-100 cm layer

was also collected. Two additional ancestral forests that were not botanically

inventoried were sampled for charcoal in the 20-60 and 80-100 cm layers. In each of

these extra ancestral forests, two holes separated by 70 m were dug and soil was

collected. The coordinates of all areas were recorded at the midpoint of each plot

with a Garmin eTrex H GPS. After fieldwork in each community, semistructured

interviews about landscape management, historic occupation of the region,

mythology and plant uses were conducted with 1 or 2 adults aged 30 to 84 years,

totalling seven interviews for the study. Based on the interviews and information

collected about useful plants during the inventories, we created a subcategory of

useful species called "managed species" that encompasses species of edible fruit

managed in several ways. Baniwa management occurs through planting in gardens

and swiddens, with periodic weeding and burning of refuse, protection of certain

species against fire, seed dispersal along forest trails, clearing undergrowth around

23

trees, elimination of competing plants that do not interest them, among other

practices. Fourteen species were considered "managed species": Attalea maripa;

Couma sp.; Dacryodes sp.; Euterpe precatoria; Humiria balsamifera; Inga sp. 1; Inga

sp. 2; Oenocarpus bacaba; Oenocarpus bataua; Poraqueiba sericea; Pourouma

cucura; Pouteria ucuqui; Tapirira guianensis; Theobroma subincanum. All managed

species were cited as sometimes planted in gardens and swiddens; during the

subsequent fallow periods they are then subject to the other management practices

to varying degrees. As management information is current, we cannot know whether

other species were historically managed, which would also affect their distribution in

the forest. According to our informants, other Baniwa communities not involved in the

research manage other species. The distances of the plots to the Içana River, to the

villages of the clans that “own” each ancestral territory and to nearest watercourse

were calculated using the "ruler" tool in Google Earth 7.1.5. The distances travelled

during fieldwork to reach the plots from village were calculated with the "route" tool of

the GPS.

Laboratory analysis of soil and charcoal. Soil samples were taken to INPA,

Manaus, for analysis at the Thematic Laboratory of Plants and Soil. The samples

were air dried, broken up, sieved through a 2 mm mesh, and homogenized. The

volume of each sample was measured using a 10³ cm³ glass cube, where the width

(10 cm) and length (10 cm) were multiplied by the height of the soil sample (in cm) in

the cube. In each sample, pieces of charcoal of at least 1 mm were removed with

tweezers and weighed. Chemical (pH [H2O], P, K, Ca, Mg, Al) and physical (total

amount of sand) analyses were performed in the 0-20 cm layer using the methods of

EMBRAPA44. Two charcoal samples were discarded because they had fallen to the

ground during the analysis.

Data analyses. The averages of edaphic variables per plot [pH; P (mg / kg),

Ca, Mg, K and Al (cmol / kg), charcoal (mg / cm3) and sand (%)] were determined for

the 0-20 cm layer (Extended Data. 2). The charcoal in the soil was also determined

for the 20-60 and 80-100 cm layers. The abundance of trees of each species and the

total basal area per plot, originally measured in 0.072 ha plots, were extrapolated to 1

ha. The density of species per plot was not extrapolated because the species / area

relationship is not linear. The relative richness, abundance and basal area of

managed species were calculated as the percentage of managed species in terms of

24

the total richness, abundance and basal area of each parcel, respectively (Data

Extended Tab. 1). The Ecological Importance Value (EIV) of the species was

calculated for ancestral and old growth forests (Fig. 3). The EIV is an index

composed of the sum of frequency, dominance and density45 that indicates the

importance of a particular species in the floristic community and its use can facilitate

the interpretation of past landscape management in studies that seek to understand

pre-Columbian forest enrichment46,47. Differences between ancestral and old growth

forests in soil, charcoal and floristic community structure variables were tested with

Student’s t test when data were normally distributed and with the Mann-Whitney test

when data were non-normal. A simple linear regression was used to examine

relationships between: the distance to the Içana River and the relative richness,

abundance and basal area of managed species; the distance to the nearest

watercourse and the relative richness, abundance and basal area of managed

species; the distance to the Içana River and the charcoal concentration (both 0-20

and 20-60 cm depth); and the charcoal concentration (0-20 depth) and soil pH soil.

The dimensionality of the floristic composition was reduced with non-metric

multidimensional scaling (NMDS)48 based on the relative abundances of the species

per plot and using the Bray-Curtis dissimilarity49. Ordination in the first dimension

explained 67% of the variation in floristic composition with stress of 0.27, considered

inadequate for analysis of community ecology48. The ordination in two dimensions

explained 83% of the variation in floristic composition with stress 0.14, considered

satisfactory. Additional dimensions contributed little to increase the percentage of

variation explained. The ordinations were used to examine patterns presented by the

most abundant species and highlight the contribution of relative abundance of

species to distinguish plots. The NMDS axes, which describe floristic composition,

were used as dependent variables in a multivariate analysis of covariance

(MANCOVA) to test the relationship between floristic composition and soil variables

and presence/absence of management. An exploratory analysis was done to select

the soil variables with independent effects on each axis of floristic composition

variation. The final MANCOVA model included the proportion of sand and soil pH as

quantitative independent variables and the occurrence or not of management

(ancestral vs. old growth forests) as a categorical independent variable. The two

NMDS axes were used as dependent variables. The relative abundance of managed

25

species was projected for the forests of Baniwa ancestral territory in the Içana River

basin35,39 in areas up to 750 m from any watercourse and was based on the average

values of the abundances of managed species of both ancestral and old growth

forests, assuming both forest types to have similar areas in the Içana River. All

analyses were performed using SPSS 17.0 software50. The maps were produced in

ArqGIS 10.4 software51. All ancient village sites and swiddens identified during the

participatory mapping exercise were georeferenced and are shown on the map in

Extended Data Fig. 1. Landsat images in the Extended Data Fig. 1 (coloured boxes)

are from 1985.

Ethical aspects. Because the research was done in indigenous territory and

involves traditional knowledge, we required authorization by Brazilian federal

authorities: National Indian Foundation (Fundação Nacional do Índio - FUNAI) - N°.

0270 / GAB / CR-RIO NEGRO / 2015; National Council for Ethics in Research with

Human Subjects (Conselho Nacional de Ética em Pesquisa com Seres Humanos -

CONEP) - CAAE N°. 45373015.2.0000.0006. The Federation of Indigenous

Organizations of the Negro River (Federação das Organizações Indígenas do Rio

Negro - FOIRN) and the Indigenous Organization of the Içana River Basin

(Organização Indígena da Bacia do rio Içana - OIBI) authorized research in their

territory via FOIRN protocol n° 20. At each village the study was explained to village

leaders and members, who authorized research through their representatives. All

informal and semi-structured interviews were authorized by each informant

immediately before the interview, after they were informed that they were not

required to participate. The collections of plants and soil were carried out with the

permission of indigenous authorities - FOIRN and OIBI.

References: 31. Latrubesse, E. M. & Franzinelli, E. The late quaternary evolution of the Negro river, Amazon, Brazil: Implications for island and floodplain formation in large anabranching tropical systems. Geomorphology 70, 372-397 (2005). 32. Anderson, A. B. White-sand vegetation of brazilian Amazonia. Biotropica 13, 199-210 (1981). 33. Instituto Nacional de Meteorologia - INMET. Clima. http://www.inmet.gov.br/portal/ (2009). 34. Wright, R. M. (ed) Cosmos, Self and History in Baniwa Religion: For Those Unborn. (University of Texas Press, Austin, USA, 1998). 35. (Moraes, J. F., personal observations during field work) 36. Cornelio, J. M. et al. Waferinaipe Ianheke: A Sabedoria dos Nossos Antepassados. Histórias dos Hohodene e dos Walipere-Dakenai do Rio Aiari (ed. Wright R. M.) (Acira / Foirn, São Gabriel da Cachoeira, Brazil, 1999). 37. Nimuendajú, C. Reconhecimento dos rios Içána, Ayarí e Uaupés. Journal de la Société des Américanistes 39, 125-182 (1950).

26

38. Koch-Grunberg, T. Dois Anos Entre os Indígenas: Viagens no Noroeste do Brasil -1903/1905 (ed. Pinto, R. F.) (EDUA, Manaus, Brazil, 2005). 39. Andrello, G. & Wright, R. Baniwa: History of occupation. http://pib.socioambiental.org/en/povo/baniwa/1557 (2002). 40. Silva, A. L. No rastro da roça: ecologia, extrativismo e manejo de arumã (Ischnosiphon spp., Marantaceae) em capoeiras dos índios Baniwa do Içana, alto rio Negro. MSc. Thesis, Ecology Program, Instituto Nacional de Pesquisas da Amazônia / Universidade Federal do Amazonas, Manaus (2004). 41. Abraão, M. B., Nelson, B. W., Baniwa, J. C., Yu, D. W. & Shepard Jr, G. H. Ethnobotanical ground‐truthing: indigenous knowledge, floristic inventories and satellite imagery in the upper rio Negro, Brazil. Journal of Biogeography 35, 2237-2248 (2008). 42. Abraão, M. B., Shepard Jr, G. H. & Nelson, B. W. Baniwa vegetation classification in the white-sand campinarana habitat of the northwest Amazon, Brazil. In Landscape Ethnoecology: Concepts of Biotic and Physical Space (eds. Johnson, L. M. & Hunn, E. S.) 83-115 (Berghahn Books, New York, USA, 2010). 43. Stropp, J., van der Sleen, P., Quesada, C. A. & ter Steege, H. Herbivory and habitat association of tree seedlings in lowland evergreen rainforest on white-sand and terra-firme in the upper rio Negro. Plant Ecology & Diversity 7, 255-265 (2014). 44. Donagema, G. K., Campos, D. D., Calderano, S. B., Teixeira, W. G. & Viana, J. H. M. (eds) Manual de Métodos de Análise de Solos, 2° Edição. (Embrapa Solos, Rio de Janeiro, Brazil, 2011). 45. Mueller-Dombois, D. & Ellemberg, H. (eds) Measuring species quantities. In Aims and Methods of Vegetation Ecology 67-92 (John Wiley & Sons, New York, USA, 1974). 46. Balée, W. Indigenous transformation of Amazonian forests: an example from Maranhão, Brazil. L'Homme 33, 231-254 (1993). 47. Erickson, C. L. & Balée, W. The historical ecology of a complex landscape in Bolivia. In Time and Complexity in Historical Ecology: Studies in the Neotropical Lowlands (eds. Balée, W. & Erickson, C. L.) 187-233 (Columbia University Press, New York, USA, 2006). 48. McCune, B. & Grace, J. B. (eds.) Nonmetric multidimensional scaling. In Analysis of Ecological Communities 125-142 (MjM Software Design, Oregon, USA, 2002). 49. Faith, D.P., Minchin, P.R. & Belbin, L. Compositional dissimilarity as a robust measure of ecological distance. Vegetatio 69, 57-68 (1987). 50. Field, A. (ed.) Discovering statistics using SPSS: Third Edition (SAGE Publications, London, England, 2009). 51. Johnston, K., Ver Hoef, J. M., Krivoruchko, K., Lucas N. (eds.) Using ArcGIS Geostatistical Analyst (ESRI, Redlands, USA, 2001).

27

Extended Data Figure 1 - Map of the locations of old village sites with their associated managed areas (ancestral forests) in the Baniwa ancestral territory identified by informants of the five villages that participated in the study. Red dots represent ancestral forests that have been uninhabited for more than 200 years. Lower right and lower centre are rescaled from the black boxes in the main map for detail. Currently the Baniwa live along the banks of Içana and Ayari Rivers (Fig. 4) and practice non-intensive shifting cultivation with annual clearance of new swiddens. These swiddens are visible as yellow to light green areas in the 1985 satellite images obtained from Google Earth that are colour coded to the wine, green and blue boxes in the main map.

28

Extended Data Figure 2 - Ten species with greatest ecological importance values (EIV). a, in ancestral forests; b, in old growth forests. Bars in grey represent managed species. EIV = (relative frequency + relative dominance + relative density) / 3. The four managed species in a represent, together, 18.7% of the total ecological importance in ancestral forests.

29

Extended Data Figure 3 – Comparison between ancestral and old growth forests in terms of four soil characteristics. a, soil pH (ancestral forests, mean ± standard error = 4.77 ± 0.12; old growth forests, m±se = 4.38 ± 0.09; p = 0.052); b, proportion of sand (74.23 ± 3.31; 66.69 ± 6.46; p = 0.286); c, Effective Cation Exchange Capacity - ECEC (1.21 ± 0.13; 1.59 ± 0.18; p = 0.143); d, phosphorus concentration (1.93 ± 0.37; 1.31 ± 0.43; p = 0.384). N = 12 in ancestral forests e N = 4 in old growth forests.

30

Extended Data Figure 4 – Soil pH in relation to the charcoal concentration in soil (0-20 cm). Black dots represent plots in ancestral forests and red dots in old growth forests. The black line is the simple linear regression soil pH = 4.19 + 0.068 * charcoal concentration (R² = 0.67; P < 0.001). N = 15 (one plot in ancestral forests was not included because the charcoal was damaged). Charcoal concentration and soil pH in 0-20 cm layer were higher in ancestral forests (Fig. 3 and Extended Data Fig. 3a).

31

Extended Data Figure 5 – Charcoal concentration in soil in relation to the distance to the Içana River. a, concentration in the 0-20 cm layer (N = 15, R² = 0.081, P = 0.305); b, concentration in 20-60 cm layer (N = 17, R² = 0.023, P = 0.559). Black dots represent samples in ancestral forests and red dots in old growth forests. One plot in an ancestral forests was not included in a, because charcoal of the 0-20 cm layer was damaged. One plot in an ancestral forest was not included in b, because charcoal of the 20-60 cm layer was damaged. Two extra samples of charcoal were included in the analysis of b and belong to collections made in ancestral forests that were not inventoried.

32

Extended Data Figure 6 - Comparison between ancestral and old growth forests in terms of structural characteristics. a, species density; b, total basal area; c, total abundance. In b, the variability in relation to the mean was 3.11 times greater in ancestral forests than in old growth forests (Coefficient of variation = 25.84 % vs 8.30 %, respectively), which shows that management increases the variance of basal area and hardly decreases the total basal area. None of the comparisons are significant at p = 0.05. N = 12 in ancestral forests and N = 4 in old growth forests.

33

To be continued...

34

Extended Data Figure 7 – Relative richness, abundance and basal area of managed species in relation to the distance to the Içana River and nearest watercourse. a, relative richness in relation to the distance of Içana River (R² = 0.126; p = 0.177); b, relative abundance in relation to the distance of Içana River (R² = 0.023; p = 0.573); c, relative basal area in relation to the distance of Içana River (R² = 0.008; p = 0.740); d, relative richness in relation to the nearest watercourse (R² = 0.057; p = 0.373); e, relative abundance in relation to the nearest watercourse (R² = 0.061; p = 0.356); f, relative basal area in relation to the nearest watercourse (R² < 0.001; p = 0.993). Black dots (N = 12) and red dots (N = 4) represent plots in ancestral and old growth forests, respectively.

35

Extended Data Table 1 - Summary of structural floristic data in 16 plots of 0.086 ha. Plots 1-12 are in ancestral forests and 13-16 in old growth forests.

Plot Total density of

species

Total abundance of species

Total basal area of species

Relative richness of managed species (%)

Relative abundance of managed species

(%)

Relative basal area of managed

species (%) 1 43 158 1.97 7.70 17.83 18.54 2 47 191 2.30 10.64 14.66 54.96 3 81 263 3.60 11.54 9.96 19.12 4 63 223 3.35 8.06 5.41 27.19 5 43 187 2.56 5.13 17.49 30.74 6 59 221 4.03 19.30 38.81 53.77 7 52 187 2.88 17.65 22.40 39.36 8 64 196 2.22 11.11 15.90 53.80 9 47 227 2.42 8.70 9.73 27.14 10 61 216 4.21 15.79 37.44 59.19 11 41 247 3.78 14.63 57.09 71.73 12 82 215 2.48 10.98 10.70 22.14 13 80 212 2.32 8.75 9.91 14.84 14 46 195 2.61 4.44 2.08 2.38 15 69 233 2.39 10.77 10.48 11.94 16 59 269 2.79 12.28 8.21 2.87

The total density of species is the number of tree species sampled in a given plot; total abundance of species is

the total number of individuals sampled in the plot; basal area of species is given in m2 0.086 ha-1; relative

richness, abundance and basal area of managed species is the proportion of managed species in each plot in

terms of total density, abundance and basal area, respectively, and are given as percentages. Species about which

the Baniwa informants were unsure about management were not included in estimated of relative richness,

abundance or basal area. Mean ± sd of number, abundance and basal área (given in m2) of not identified species

by indigenous per plot was, respectively, 1.75 ± 1.57, 1.81 ± 1.68 e 0.01 ± 0.02.

36

Extended Data Table 2 - Summary of the location of plots, soil analysis, charcoal and distances (km).

Plot Village that owns the site

Forest type

Lat. Long. sand P ECEC pH Char 0-20

Char 20-60

Char 80-100

Dist. to Içana

Dist. to village

Dist. travelled

Dist. to nearest watercourse

1 Ttdzealinomana AF 1.25368 -68.2509 89.30 3.45 0.58 5.74 16.26 _ 1.23 1.01 1.45 2.18 0.30 2 Hizdapada AF 1.34068 -68.4416 79.43 1.49 1.00 4.47 3.05 1.81 _ 0.82 1.23 1.10 0.74 3 Hizdapada AF 1.33222 -68.4503 77.04 5.54 1.28 4.37 6.22 3.35 _ 2.02 1.94 2.02 0.26 4 Komalhipani AF 1.33091 -68.4026 77.94 1.62 1.26 4.41 6.68 5.22 0.34 4.12 7.86 12.16 0.04 5 Ttdzealinomana AF 1.26470 -68.2536 80.50 0.52 1.98 4.78 13.17 4.21 _ 3.32 3.73 5.40 0.20 6 Bobope AF 1.34572 -68.3423 70.56 0.28 0.60 5.06 _ 0.50 _ 12.72 19.05 31.08 0.06 7 Komalhipani AF 1.32407 -68.4018 70.26 2.17 1.56 4.59 10.04 2.90 0.30 3.23 6.91 11.09 0.17 8 Hizdapada AF 1.34446 -68.4430 52.36 1.46 1.84 4.59 4.98 1.51 _ 0.15 2.37 0.15 0.16 9 Owhika AF 1.27501 -68.1903 81.15 2.45 1.19 5.35 13.32 6.48 _ 1.56 3.51 7.13 0.06 10 Komalhipani AF 1.33408 -68.4037 73.30 3.15 1.32 4.47 2.33 0.43 0.36 5.10 8.77 14.03 0.02 11 Owhika AF 1.29174 -68.1536 53.26 1.19 1.16 4.84 10.96 4.25 _ 0.07 3.88 0.07 0.07 12 Bobope AF 1.33342 -68.3424 85.69 0.84 0.78 4.58 2.57 6.69 _ 11.26 16.50 26.90 0.06 13 Hizdapada OGF 1.33283 -68.4509 55.16 1.26 1.24 4.16 1.66 1.55 _ 2.18 2.13 2.18 0.23 14 Ttdzealinomana OGF 1.26469 -68.2535 84.93 0.22 1.66 4.55 0.47 3.14 _ 3.32 3.70 5.40 0.20 15 Komalhipani OGF 1.32076 -68.4012 65.89 2.31 2.08 4.51 5.17 5.03 0.30 2.24 5.90 9.36 0.27 16 Hizdapada OGF 1.34240 -68.4400 60.78 1.44 1.39 4.28 3.24 1.16 _ 0.09 1.65 0.09 0.10

*E1 Bobope AF 1.33373 -68.3426 _ _ _ _ _ 0.59 0.18 11.39 16.64 27.02 0.04 *E2 Komalhipani AF 1.34238 -68.3943 _ _ _ _ _ 0.31 0.26 6.54 9.75 10.83 0.08

A summary of number of plot (plot), village that owns the site, forest types (AF, ancestral florest; OGF, old growth forest), latitude

(Lat.) and longitude (Long.), proportion of sand in soil (%), concentration of phosphorus in soil (mg kg-1), Efective cation exchange

capacity (ECEC, cmols kg-1), soil pH, concentration of charcoal in soil (Char 0-20 cm; Char 20-60 cm; Char 80-100 cm; in mg cm3),

straight line distance from the plot to the Içana River (Dist. to Içana), straight line distance to the owner village (Dist. to village),

distance travelled from village to reach the plot (Dist. travelled), and straight line distance to the nearest watercourse (Dist. to

watercourse) for each plot is given. Two charcoal samples were discarded due to problems during analysis (sample 0-20 cm/plot 6 and

20-60 cm/plot 1). The symbol ‘-‘ means no data. Charcoal samples were collected at two additional ancestral forests in the 20-60 and

80-100 cm layers: Extra 1 (*E1) and Extra 2 (*E2).

37

Extended Data Table 3 - Total abundance of the 46 most abundant species in the 16 plots sampled. Species Total abundance Proportion (%) Number of plots Virola calophylla 362 10.52 14 Dacryodes sp. 136 3.95 13 Vochysia sp. 122 3.55 13 Oenocarpus bacaba 118 3.43 9 Poraqueiba sericea 114 3.31 11 Symphonia globulifera 112 3.26 15 Remijia sp. 94 2.73 10 Licania sclerophylla 80 2.33 14 Guatteria sp. 78 2.27 15 Eperua purpurea 71 2.06 6 Humiria balsamifera 68 1.98 7 Gupia glabra 62 1.80 14 Astronium sp. 50 1.45 11 Compsoneura ulei 49 1.42 8 Iriartella setigera 47 1.37 10 Eschweilera sp. 43 1.25 13 Protium sp. 42 1.22 10 Cyatheales sp. 1 40 1.16 10 Ryania speciosa 40 1.16 8 Abarema jupumba 39 1.13 7 Ocotea sp. 1 38 1.10 13 Tapirira guianensis 38 1.10 8 Miconia sp. 1 36 1.05 8 Euterpe precatoria 35 1.02 7 Myrcia sp. 34 0.99 12 Swartzia argentea 34 0.99 10 Jacaranda copaia 34 0.99 10 Swartzia tomentifera 33 0.96 7 Endlicheria sp. 32 0.93 9 Remijia amazonica 29 0.84 8 Socratea exorriza 29 0.84 6 Oenocarpus bataua 28 0.81 8 Arecaceae sp. 1 25 0.73 7 Clathrotropis macrocarpa 25 0.73 6 Miconia alata 24 0.70 4 Miconia punctata 24 0.70 7 *Potaropa 24 0.70 5 Rinorea sp. 24 0.70 8 Inga sp. 1 23 0.67 12 Helicostylis sp. 22 0.64 6 Ocotea sp. 2 22 0.64 11 Lauraceae sp. 1 21 0.61 8 *Hemapitako 21 0.61 5 Brosimum rubecemus 21 0.61 9 Anthodiscus sp. 19 0.55 8 Pouteria ucuqui 18 0.52 6 Total 2480 72.09 _

The 46 most common species from a total of 277 species and 3440 individuals are shown. For

each species, the total abundance of sampled individuals (abundance), the proportion of this

abundance in relation to the total number of individuals sampled (proportion in percentage), and

the number of plots in which the species occurred (number of plots) is given. The voucher

38

specimens of two species were damaged during transport, making their identification impossible;

these species are identified in the table by an asterisk (*) and their indigenous names.

39

Conclusões

• A composição de espécies arbóreas, de florestas de terra firme da bacia do rio Içana, alto rio Negro, está relacionada à quantidade de areia no solo e ao manejo indígena.

• Florestas ancestrais com mais de 200 anos possuem composição de espécies arbóreas diferente de florestas antigas imemoriais devido ao manejo indígena que aumentou a abundância de espécies manejadas e tornou o pH do solo menos ácido.

• Florestas ancestrais com mais de 200 anos possuem densidade de espécies arbóreas por área, abundância total de árvores e área basal total de árvores semelhantes às de florestas antigas imemoriais.

• Estruturalmente (frequência relativa, dominância relativa e densidade relativa) florestas ancestrais, com mais de 200 anos, são dominadas principalmente por espécies manejadas e florestas antigas imemoriais, da mesma região, são dominadas por poucas espécies não manejadas.

• A densidade de espécies por área, a abundância total de indivíduos e a área basal total de espécies arbóreas manejadas não diminuem com a distância ao rio Içana quando dentro de 750 metros de quaisquer cursos d’água.

• Em florestas de terra firme da bacia do rio Içana, alto rio Negro, o pH do solo é relacionado a quantidade de carvão no solo e ambos são maiores em florestas culturais.

• Florestas ancestrais e antigas imemoriais diferem quanto a quantidade de carvão no solo até 20 cm, porém não diferem quanto a quantidade de carvão no solo entre 20 e 60 cm.

• Não há relação entre a quantidade de carvão no solo, até 20 cm e entre 20 e 60 cm, e a distância ao rio Içana em florestas de terra firme localizadas na bacia do rio Içana, quando dentro de 750 metros de quaisquer cursos d’água.

40

Apêndice:

41

42

MAPA DESENHADO POR INFORMANTES DA COMUNIDADE DE BELA VISTA

43

MAPA DESENHADO POR INFORMANTES DA COMUNIDADE DE TUCUMÃ-RUPITÁ

44

MAPA DESENHADO POR INFORMANTES DA COMUNIDADE DE MAUÁ-CACHOEIRA

45

MAPA DESENHADO POR INFORMANTES DA COMUNIDADE DE SANTA ROSA

46