The loss of hollow-bearing trees in production forest areas can have large impacts on animal populations that rely on them for shelter. This study facilitates the selection of appropriate trees for retention by examining the proportion and type of trees that were used by vertebrate fauna in mature wet and dry Eucalyptus obliqua forest in Tasmania. Felled trees were searched for hollows and secondary evidence of use by fauna. Classification Trees and Bayesian logistic regression modelling were used to examine the site and tree attributes that best explained the use of a tree by fauna. We did two separate analyses: one using attributes expected to be causally related to hollow use, and a second using attributes that might be correlated with hollow use and could be easily assessed in standing trees. In all, 28% of hollow-bearing trees examined showed evidence of use, which is at the lower end of the scale found in other areas of Australia. The variables most strongly related to the use of a tree were hollow abundance, tree size and senescence. Random Forest modelling indicated that the likelihood of a hollow being used increased with hollow size, particularly hollow depth. Counting the number of hollows in standing trees was the best way to identify a tree that is likely to be used by fauna and this was particularly important for younger and healthier trees. It was recommended that trees to be retained should contain at least one large hollow. It was estimated that 8–15 trees per hectare were used by hollow-using fauna in mature wet and dry E. obliqua forest in Tasmania.