We tested which factors explain the distribution pattern of the epiphytic moss Neckera pennata on three spatial scales using the framework of generalized linear models. First, we tested which factors explained its occurrence in forest stands in a 2500 ha landscape. At this scale, we also tested the effect of the historic landscape structure. We recorded its occurrence in all suitable stands. The occurrence probability increased with increasing present quantity of Acer platanoides, and with increasing present and past quantity of Fraxinus excelsior. The probability also increased with increasing connectivity to occupied stands. However, the connectivity to stands present in 1977 (recorded from infra-red aerial photographs) explained more of the variation. This suggests that the regional metapopulation size of N. pennata has decreased during the past decades, and that its present distribution pattern reflects the age of the remaining stands, and the distribution of past dispersal sources in the landscape. Second, we tested which factors explained the occurrence and abundance on individual trees in three forest stands. Neckera pennata mainly occurred on Acer and Fraxinus stems. The most important variable in explaining occurrence probability was connectivity to surrounding occupied trees, which probably reflects the restricted dispersal range in this species. The abundance on occupied trees was also explained by this variable. The occurrence probability and abundance also increased with increasing tree diameter, probably reflecting the time that a tree has been available for colonization and the time since colonization, respectively. The occurrence probability and abundance furthermore decreased on strongly leaning (and deteriorating) trees. The occurrence probability increased with increasing bark roughness, probably reflecting increasing suitability regarding bark chemistry and moisture. Third, we tested its vertical distribution on occupied trees. The main distribution was below 1.6 m.