Complete Rainforest Elevational Gradient Reveals Unusual Diversity Patterns of Non‐Volant Mammals in New Guinea

ABSTRACT Aim We address a critical gap in the elevational community ecology of tropical non‐volant mammals in the Australian and Oceanian zoogeographic realms. Specifically, we document alpha and beta diversity, environmental predictors and community composition of individual clades in relation to their ecology and evolutionary history along an extensive elevational gradient, for the first time in this region. Location Complete rainforest elevational gradient and subalpine zone of Mount Wilhelm (45–3700 masl), Papua New Guinea. Methods We conducted extensive sampling using traps (> 21,800 trap‐nights) and collaboration with indigenous hunters across nine localities at 500 m intervals along the gradient, from lowland rainforest to subalpine grassland. The marsupial and rodent species were identified using morphological and genetic data. The community variables were correlated with key environmental variables. Results In total, we recorded 61 marsupial and rodent species. We reveal an unusual pattern for tropical mammals: species richness increased linearly with elevation within continuous forest, then declined sharply in the open subalpine habitat. This contrasts with the mid‐elevation richness peak, evident in regional mammal fauna and also common along elevation gradients globally. Community composition was primarily associated with temperature, with a sharp faunal turnover at 1200 masl dividing the assemblages into distinct lowland and highland communities. Main Conclusions The novel elevational pattern challenges prevailing models and underscores the importance of integrating multiple sampling techniques including collaboration with indigenous hunters, where possible, to accurately capture the full extent of community composition. We highlight the ecological and conservation values of examining complete elevational gradients in tropical ecosystems, as a single Mt. Wilhelm transect encompasses 80% of the regional diversity of non‐volant mammals. We suggest that any long rainforest elevational gradient is likely a high conservation priority.