The 400-500 wolves currently living in the Apennine range of peninsular Italy are slowly recolonizing the Alps and are expected to move northward. A nationwide management plan for the Italian wolf population is being prepared, and a zoning system with connecting corridors has been suggested. We developed a large-scale probabilistic model of wolf distribution as a contribution to the planning process. Thirteen environmental variables related to wolf needs and human Presence were analyzed in 12 well-studied wolf territories and in 100 areas where the species has been absent for the Past 25 years. These two areas were used as a training set in a discriminant analysis to evaluate Potential wolf presence throughout the entire country. We used the Mahalanobis distance statistic ns an index of environmental quality, calculated as the distance from the arrange environmental conditions of the wolf territories. Based on the Mahalanobis distance statistics, we constructed an actual and potential spatial distribution of the wolf for all of peninsula Italy. The jackknife procedure trans used to assess the stability of the distance model and showed good confidence in our model (coefficient of variation less than or equal to 13%). Distance from the wolf territories' centroid ns an index of environmental quality for the wolf was tested using 287 locations where wolves have been found dead in the past 25 years as a consequence of human action (poison, shotgun car accidents). A useful contribution to conservation Planning resulted from comparing the frequency distribution of the Mahalanobis distance of the dead wolf locations with the percentage? of study area within each distance class. This showed how the number of wolf casualties would greatly decrease with Protection of only a minor part of the study area and indicated the usefulness of our approach for evaluation of other conservation options, such as core areas and corridor identification.
