EN
The factors influencing the presence/absence of the edible dormouse Glis glis (Linnaeus, 1766) in forest fragments in Mediterranean central Italy were investigated by performing repeated nocturnal surveys in 38 woodlots. The various woodlots were different from each other in terms of area, degree of isolation, and floristic and structural characteristics. The effects of different types of forest management and human activities on dormouse distribution were also analysed. Dormice occurred in all the surveyed high forests, while no sign of their presence was detected in coppices with a rotation cycle short (<18 years). There were significant differences between coppices and high forests, but not between grazed and ungrazed woodlots, and rotational and non-rotational forests. Correspondence analysis corroborated the empirical evidence, showing a close association between presence of dormice and high forests, and, to a lesser degree, rotational forests. Concerning environmental parameters, favourable woods were those with high trees and with low stem number, while the specific tree assemblages proved to be scarcely important. Wood area was more important than isolation in predicting presence or absence of this rodent, as the highest probability was associated with woodlots of 40-50 ha area and more. These results did not provide evidences about the importance of the amount of residual habitat in the 10 km2 surrounding the woodlot and the presence of hedgerows as ecological corridors.