1. The epiphytic vegetation of 60 stems of aspen (Populus tremula) was studied in a 16-ha forest area near Uppsala, Sweden. The cover of bryophytes and of foliose and fruticose lichens was registered at three heights (50, 100 and 150 cm) around each stem and related to tree factors: bark roughness, bark depth, tree diameter and tree age; to forest conditions: total basal area (approximates tree density including basal area of aspen), light conditions, site quality class and field-layer vegetation type; to bark chemistry: pH, total K, Ca, Mg and Na, and exchangeable K, Ca, Mg and S; and to soil chemistry: pH, total K, Na, Mg, Ca, base saturation and total acidity. 2. Canonical correspondance analysis (CCA, following ter Braak) showed that the most important factors affecting the composition of the epiphytic vegetation were field-layer vegetation type and light conditions, which were correlated with the first axis, and basal area of aspen, soil exchangeable Ca and soil exchangeable Na, which were correlated with the second axis. 3. In multiple regression analyses with individual species as dependent variables, important factors were field-layer vegetation type, bark total Na and soil exchangeable Ca. A high species number was correlated with thick bark and shady conditions. 4. There were significant correlations between bark chemistry and soil chemistry: both exchangeable bark Ca and total bark Ca were significantly correlated with soil pH in multiple regression analyses. Redundancy analysis (RDA, following ter Braak) gave similar results and revealed a positive covariation between bark factors (total Ca, exchangeable Ca, total Na,) and soil factors (pH, soil exchangeable Ca, soil exchangeable Mg). 5. The mechanisms for the connection between soil and bark chemistry are unknown, but several explanations can be postulated, of which an impact from the soil on the bark through transportation of nutrients from the tree roots to the bark seems the most likely. Since bark chemical factors are important for epiphytic species, soil properties might affect the composition of the epiphytic vegetation on Populus tremula. 6. There is a need to investigate further the bark-soil relationships and the mechanisms of possible interactions. If the epiphytic species are indirectly affected by the soil through influences on the bark, this could have consequences for conservation management intended to preserve epiphytic species as well as for the use of bryophytes and lichens as indicators of air pollution.
