The sustainable use of the forest has been suggested as one of the only alternatives to reduce deforestation, and conserve the biodiversity and the traditional populations in the Amazon. Brazil nut is one of the main extractive products of the Amazon region, and much has been discussed about the impacts of seed harvesting for the future of Brazil-nut populations. The aims of this study were to evaluate the sustainability of Brazil-nut harvesting by the Kayapó Indigenous people of three villages in Southeastern Amazonia and the impacts of intentional and unintentional Brazil-nut management practices by them in Brazil-nut groves of the Kayapó Indigenous Land (IL), as well as provide critical ecological information to guide the management and conservation of Brazil nut in the Kayapó IL and other Amazonian regions. The following questions were investigated: (1) What is the Brazil-nut-seed stock and the proportions of seeds collected by the Kayapó of three villages at landscape, grove and Brazil-nut-tree scales? (2) Do the number of years of harvest and the amount of seeds harvested by the Kayapó affect seedling recruitment in Brazil-nut groves? (3) Do the Kayapó disperse Brazil nuts, intentionally or unintentionally? (4) Does hunting by the Kayapó affect seed removal and seedling recruitment in Brazil-nut groves? (5) What is the magnitude of spatial variability of B. excelsa population parameters in Kayapó IL? The Kayapó Indians from the villages of A'Ukre, Moikarakô and Kikretum have harvested seeds selectively and non-intensively at all analyzed scales: landscape, groves and trees. The abundance of Brazil- nut groves in the Kayapó IL is high and only about 30% of them are exploited by the Kayapó. The proportion of seeds collected by the Kayapó between 2008 and 2010 ranged from 7.2% to 43% of the seed stock of harvested groves, which represents 2.5 to 12.7% of the total stock of seeds in the territories of the three villages. Contrary to expectations, the number of years of harvest was positively related to the density of seedlings within Brazil-nut groves. The Kayapó disperse Brazil-nut seeds unintentionally along trails and the density of seedlings was significantly higher near (28.6 ind./ha ± 7.9 SD) than away from trails (1.2 ± 2.1). In addition, the Kayapó plant Brazil-nut seeds and seedlings in their villages and gardens. Brazil-nut seed removal in harvested Brazil-nut groves, which are also hunted by the Kayapó, was lower than in unharvested and unhunted groves, and seed removal by animals was negatively related to Brazil-nut seedling density in groves. This indicates that seed dispersal by the Kayapó associated with the hunting of Brazil-nut predators and/or dispersers favors seedling regeneration in harvested groves and the sustainability of seed harvesting. Mean growth and mortality of Brazil-nut seedlings, size of groves, density of adult individuals, and age structure varied significantly among groves in Kayapó IL, and even between groves within the territory of each village, indicating that simple and non-spatially-explicit demographic models are not adequate to evaluate impacts of human activities on B. excelsa populations. We conclude that the non-intensive seed harvesting conducted by the Kayapó, associated with direct and indirect, intentional and unintentional, management practices, favors seedling recruitment inside and outside groves in the Kayapó IL. As most groves in the territories of the studied villages are not harvested, there is the possibility of expansion of commercial production of Brazil nuts by the Kayapó, as well as the rotation of groves for harvesting. Demographic models aiming to evaluate the sustainability of Brazil-nut seed harvesting in the Kayapó IL and in other Amazonian regions should incorporate the spatial variability in population dynamics of B. excelsa, which is considerably high at meso and large scales.
