The savannah enclaves (i.e. patches) in the southern Brazilian Amazonia are among the most threatened and poorly surveyed sites in Amazonia. As part of an extensive mammal survey, we set camera traps in three of these savannah enclaves. We obtained 23 independent records of pampas deer Ozotoceros bezoarticus, a medium sized Neotropical cervid that is strongly associated with open habitats and categorized as Vulnerable on the Brazilian Red List of threatened species. These savannah enclaves with confirmed populations of pampas deer lie outside the species’ previously presumed historical range and are at least 350 km from any known extant population. Together, these savannah enclaves add c. 4,000 km2 to the pampas deer's currently known range. The small pampas deer populations in these enclaves are probably isolated by a matrix of Amazon forest, raising questions about spatial genetic structure and meta-population dynamics, and making them vulnerable to local extinction. We highlight the need for further studies, particularly genetic, to assess the conservation status of these populations, the results of which could potentially inform management decisions in other areas of the heavily fragmented range of this species.

The Roosevelt–Rondon Expedition marmoset Mico marcai was first collected in 1914 and all information on this primate previously came from three skins brought back by this expedition. As a result, M. marcai is categorized as Data Deficient on the IUCN Red List. As the presumed range of M. marcai lies on the path of the advancing arc of deforestation in Brazil, the collection of relevant data to assess the conservation status of this Amazonian species is of some urgency. Here we present the first field data on the distribution and population size of, and threats to, M. marcai, to reassess the species’ conservation status. During 2012–2015 we surveyed the species in the Marmelos–Aripuanã interfluve, and estimated its density using distance sampling. We also used spatial predictive modelling to estimate forest loss within the species range under two deforestation scenarios. We found the marmoset in 13 localities and estimated its extent of occurrence to be 31,073 km2. We estimated the species’ density to be 8.31 individuals/km2 and extrapolated this to estimate a total population of 258,218 individuals (CI 150,705–441,860). Under a business-as-usual deforestation scenario, c. 10,000 km2 of forest, comprising 33% of the species’ range, would be lost in three marmoset generations (c. 18 years), and we, therefore, recommend that M. marcai be categorized as Vulnerable on the IUCN Red List based on criterion A3c. Other Amazonian marmosets require similar reassessment as their ranges also fall in the path of the arc of deforestation.