Fatores e processos estruturadores das comunidades de formigas (Hymenoptera: formicidae) de árvores urbanas: uma perspectiva de metacomunidades

Abstract

The environmental traits, dispersal ability and species interactions combine with each other to determine the distribution and abundance of species of local community within habitat patches in such a space-temporal dynamics typical of metacommunities. Urban ants have been studied over the last decades in Brazil, mainly in hospitals, residences and green areas. Hypotheses about the importance of urban vegetation traits and chronic environmental stress on determining the structure of arboreal ant metacommunity were tested in 29 street sections (100 m) in the city of Dourados, located in the state of Mato Grosso do Sul. Four trees on the sidewalks were sampled with arboreal pitfalls and active search for ant colonies above 5 m tall. We found 35 ant species (or morphospecies), 14 of them (40%) only in pitfalls, five (14%) only in tree colonies and 16 (46%) in both methods. Camponotus landolti occured in 29% of the registers and was present at 28 of the sections, but only one colony of this species was found. Dorymyrmex spurius and Pheidole megacephala followed this same pattern, with high frequency in pitfalls and few colonies found. Brachymyrmex sp1 was frequently found in both methods. To evaluate which environmental variables explain the variations on the ant metacommunity structure in different scales, pitfalls data were analyzed under two approachs, one where all pitfalls formed random samples formed by four traps and the other where pitfalls were maintained associated with the street sections. On the two approaches, data matrices of ants and plants were ordinated using NMDS, the variables were reduced by PCA and MANOVA was used to correlate the explanatory variables and the ant community. On both cases, the coherence of the incidence matrix did not differ from the chance. The habitat variables and the plant composition did not explain the structure of arboreal ant community in both approaches, such as time and space variables did in the second approach. We suggest that stochastic factors imposed by city inhabitants, like tree pruning without planning, should be more important to the processes of colonization and extinction of species, determining the composition of local arboreal ant communities.