Environmental and Spatial Determinants of Arboreal Ant Assemblages in a Highly Fragmented Forest Landscape of the Southern Amazon
forest fragmentation; microhabitat; dispersal limitation, Formicidae
This study investigated how environmental and spatial attributes influence arboreal ant assemblages in a highly fragmented landscape of the Southern Amazon. Ants were collected using arboreal pitfall traps installed on living trees, allowing standardized sampling of the canopy stratum. Three matrices were analyzed: worker frequency, number of species per tree, and species composition. Predictor variables included fragment size, presence of moss, bark texture, and canopy connectivity. Worker frequency and number of species were evaluated using GLMs, while species composition was assessed through dbRDA and MRM. Worker frequency responded to structural attributes of fragments and host trees, being higher in smaller fragments and reduced on trees with smooth bark. The number of species was influenced by the interaction between fragment size and the number of trees containing moss, and it decreased in fragments with a higher proportion of smooth-barked trees, highlighting the role of microhabitat structure. In contrast, species composition was unrelated to any environmental predictor and was explained exclusively by geographic distance among fragments, indicating strong distance decay and suggesting dispersal limitation. Only Camponotus femoratus showed a significant association with trees bearing epiphytes, consistent with its status as a species strictly associated with Ant Gardens, a mutualistic system characterized by the construction of arboreal nests and the cultivation of specific epiphytes. Overall, the results show that worker frequency and number of species respond to local environmental gradients, whereas species composition is dominated by spatial processes, revealing the combined influence of microhabitat features and landscape structure on arboreal ant assemblages.