Habitat loss and fragmentation are major threats to primate populations globally. The Endangered golden monkeyCercopithecus mitis kandtiis only found in two small forest fragments: the Virunga massif in Rwanda, Uganda and the Democratic Republic of the Congo, and the Gishwati Forest in Rwanda. Little is known about the abundance and distribution of this subspecies, or threats to its survival. During 2007–2018, we collected data along 893.7 km of line transects and 354.2 km of recce trails in Volcanoes National Park and in Gishwati–Mukura National Park to estimate golden monkey density and examine any threats. In Volcanoes National Park, golden monkeys were found almost exclusively in the bamboo zone, and in Gishwati–Mukura National Park they occurred only in the remnant tropical montane Gishwati Forest. In Volcanoes National Park, density was estimated to be 7.89 (95% CI: 3.85–16.19), 5.41 (2.64–11.08), and 5.47 (3.68–8.14) groups per km²in 2007, 2011 and 2017–2018, respectively. This corresponds to a total of 4,331 individuals (95% CI: 2,723–5,938) in 2007, 4,487 (2,903–6,071) in 2011 and 4,626 (4,165–5,088) in 2017–2018. In Gishwati Forest, group density averaged 1.98 (95% CI: 1.27–3.16) per km², corresponding to 172 (95% CI: 154–190) individuals in 2017–2018. Survey results from Volcanoes National Park suggest that the golden monkey population has been stable during 2007–2018. Limited habitat, illegal activities such as harvesting of bamboo and firewood, and the presence of feral dogs, threaten the golden monkey in Rwanda and require continued monitoring. The development of a conservation action plan is a priority to protect this species.

The soundscape is a complex arrangement of sounds originating from animals and the environment. It is considered a reliable proxy for ecosystem niche structure at the community level. Acoustic communities of anuran species include advertising males, which compete in acoustic space for conspecific females. Stochastic niche theory predicts that all local niches are occupied, and the acoustic community is species-saturated. Acoustic niches, which include the spectral and temporal call structure and diel and seasonal patterns of call activity, are of similar breadth with small overlap. We tested these predictions in four communities inhabiting pristine wetlands at 2546–3188 m a.s.l. in the Volcanoes National Park, Rwanda. We sampled 74 days of hourly 5 min recordings of the local soundscape (September 2019–March 2020) using passive automated monitoring devices (Songmeter SM4). We identified species based on the advertisement call features and measured call activity as calls per minute. The communities included 4–6, species depending on wetland structure, with a shared stock of three species (Hyperolius castaneus, H. glandicolor, Leptopelis kivuensis). Independent of elevation, niche breadth for call features was similar among species and overlap reduced by partitioning the frequency range used. The diel and seasonal niche breadth of specific call activity varied according to the local temperature regime at different altitudes representing the variable part of the acoustic niche. We conclude that communities are indeed species-saturated and acoustic niches differ primarily by the fixed call features remaining locally adaptable by the modulation of the call activity pattern, corroborating the predictions of the stochastic niche theory.