HABITAT SELECTION OF Marmosops incanus (LUND, 1841) (DIDELPHIMORPHIA, DIDELPHIDAE) IN THE SERRA DOS ÓRGÃOS NATIONAL PARK, ATLANTIC FOREST, SOUTHEAST BRAZIL

Abstract

Habitat selection is one of the mechanisms responsible for allowing species coexistence and, therefore, local richness of communities. This study investigates microhabitat selection by the opossum, Marmosops incanus, in a forested area in the Brazilian Atlantic Forest. We measured microhabitat variables related to vegetation structure and inferred M. incanus abundance as a measure of habitat use. Eight microhabitat variables were measured in each trap station: litter, herbaceous and woody stems, rock outcropping, canopy cover, foliar vertical obstruction at three heights (FOV), and number of fallen trunks. Based on a correlation matrix, we selected four microhabitat variables (litter, rock outcropping cover, FOV, and number of fallen trunks) and performed a Poisson regression using the selected variables and M. incanus abundance. Litter and rock outcropping cover were positively related to abundance, thus representing the habitat characteristics selected by the species. We infer that litter represents a source of food items, since it can support an abundant fauna of macroinvertebrates, especially arthropods. Rock outcroppings can be used as shelter and protection against predators. Other studies have recorded M. incanus using habitat in a unique way, either by changing the frequency of ground/understory use, or habitats with high/few herbaceous and litter cover, as the habitat varied from restinga to fragmented forests. Different habitat selection patterns were observed in areas where M. incanus and M. paulensis are sympatric, or Gracilinanus microtarsus, both of which have comparable size and ecology. We suppose the selection patterns recorded in this study are a result of the absence of a congener species in the study area. It may allow M. incanus to use more frequently areas where near-ground strata are clear and highly covered with litter. It remains to be evaluated if habitat simplification leads to behavior homogenization.