Search for a command to run...
Abstract Understanding forest regeneration processes is increasingly important as disturbed forests spread rapidly in tropical landscapes. While successional dynamics are relatively well‐understood for plants, they remain poorly known for tropical insects, particularly in montane rainforests and tree canopies. We studied the effect of montane forest succession on arboreal ant communities in New Guinea near their natural elevational limit (1800 m a.s.l.). We censused ant species and nests in replicated 0.1‐ha plots of primary, old and young secondary forests, destined to be felled for gardens by the local landowners. Overall, 1249 felled trees were dissected for all ant‐associated microhabitats. We used rarefaction‐based models manipulating vegetation structure and multivariate randomizations to assess the effects of tree density, tree size, and nesting microhabitats on the ant communities. We expected increased ant species diversity, microhabitat specialization, and species turnover among trees throughout the succession, with distinct community composition among the stages. In contrast to our expectations, ant species composition and diversity per plot did not change during succession, and species similarity among trees was significantly higher only in old secondary forest. The number of ant species per tree increased with forest age, but trees of similar size hosted only slightly more species in primary than secondary forests. Many ant species were associated with particular nesting microhabitats; however, the same microhabitat types occurred in all stages and thus did not generate successional trends in ant communities. Secondary succession had an unexpectedly low impact on tree‐dwelling ant communities in this montane ecosystem, in contrast with our previous findings from lowland rainforests. These results highlight the need to study successional processes along entire elevational gradients, as montane taxa may react differently to vegetation changes than their lowland counterparts.