Aim Community‐level assessments of how biodiversity responds to disturbance within forest habitats are often biased towards terrestrial‐based surveys. However, recent research suggests that arboreal communities of several indicator groups (ants, amphibians, beetles and butterflies) are more susceptible to human disturbance than their terrestrial counterparts, but what about wildlife at higher trophic levels? We assess responses to disturbance, from forest floor to canopy, of a key conservation flagship group: medium to large rainforest mammals. Location The Manu Biosphere Reserve, southeast Peru. Methods We deploy both arboreal and terrestrial camera traps to directly compare the response of arboreal and terrestrial mammal communities to rainforest degradation and disturbance. Results We show that the arboreal mammal community is more susceptible to habitat disturbance than the terrestrial community. Furthermore, the largest‐bodied arboreal species, which are major seed dispersers, showed the greatest negative response to forest disturbance. The strongest predictors of occupancy probability for arboreal communities were focal tree connectivity and canopy cover, whereas surrounding forest loss and canopy height were there strongest predictors of terrestrial community occupancy, although these also had effects similar in size and magnitude on the arboreal community. Main conclusions Conservation conclusions drawn from camera‐based studies focused on the terrestrial realm likely underestimate the impact of rainforest degradation to arboreal communities and on arboreal rainforest biodiversity in general. We highlight the importance of implementing arboreal research methods, capable of investigating conservation implications of anthropogenic disturbance across all vertical strata, for accurate conservation assessments and improving rainforest management and restoration strategies.

The sleeping site behavior of Ateline primates has been of interest since the 1980s, yet limited focus has been given to their influence upon other rainforest species. Here, we use a combination of arboreal and terrestrial camera traps, and dung beetle pitfall traps, to characterize spider monkey sleeping site use and quantify the impact of their associated latrines on terrestrial vertebrate and dung beetle activity. We also characterize the physical characteristics of the sleeping sites and the floristic and soil composition of latrines beneath them. Spider monkey activity at sleeping sites peaked at dawn and dusk and group composition varied by sex of the adults detected. The habitat-use of terrestrial fauna (vertebrates and dung beetles) differed between latrine sites and non-latrine controls, underpinned by species-specific changes in the relative abundance of several seed-dispersing species (such as paca and great curassow). Seedling density was higher in latrines than in non-latrine controls. Although most soil properties were similar between latrines and controls, potassium and manganese concentrations were different. These results suggest that spider monkey sleeping site fidelity leads to a hotspot of ecological activity in latrines and downstream impacts on rainforest floristic composition and diversity.

Article impact statement: Human activities reduce the presence of large herbivores and predators, affecting ecosystem function, even in well-conserved forests.

Accurately assessing how biodiversity responds in the Anthropocene is vital. To do so, a number of indicator taxa are commonly used to monitor human-impacted forests and the subsequent recovery of their biodiversity. This makes monitoring more economically feasible, yet only valuable if the responses observed truly reflect the status of biodiversity. Many challenges exist for getting this monitoring right, including choosing the most effective indicators and ultimately choosing the most appropriate methods to capture trends. We have reason to believe that the methods currently used to assess humanimpacted tropical forest might be misrepresenting trends related to the degree of impact of disturbance to biodiversity and to the value of secondary forests for biodiversity conservation. Using recent case studies that assessed butterflies, we challenge the paradigm that fruit-baited butterfly traps are the best method for assessing human-impacted tropical forests, and that their use solely along the forest floor is underestimating the impacts to biodiversity in tropical forests. We suggest that alternative or additional methods could provide a more representative picture of the overall butterfly biodiversity responses to human-impacted tropical forests and that similar assessments of other groups and methods should be carried out.