1. Previous investigations of the role of pathogens as drivers of community dynamics in forests have mostly focused on interspecific interactions between canopy trees and seedlings via soil feedbacks. However, feedbacks can also occur directly between canopy trees and seedlings when spores of foliar and stem pathogens fall onto seedlings regenerating underneath the canopy.2. We studied pathogen spillover between canopy trees and conspecific and heterospecific regeneration in the Pinus nigra-Diplodia sapinea pathosystem. We sampled 70 pine stands distributed across a temperature gradient of 9.5-13.7°C(mean annual temperature). In each stand, we linked spore load and spore survival in the canopy with pathogen biomass and disease severity in regenerating seedlings. The density of regenerating seedlings and the health status of conspecific and heterospecific seedlings were also measured. The strength of canopy-understorey interactions was correlated with both climatic and stand variables.3. The most severe symptoms of disease in regenerating seedlings were found in the warmest stands. Structural equation models suggested that disease severity in seedlings was mediated by an increased spore spillover from the canopy.Temperature also increased the pathogen survival rate within seedlings, further contributing to disease severity. Interestingly, disease severity in the canopy did not correlate with spore load in the canopy, suggesting that adult trees were spilling pathogens onto the understorey regardless of their health status.Pathogen spillover increased in more open stands. Greater disease severity in seedlings correlated with stands with a higher density of oak seedlings. 4. Synthesis. These results suggest an understudied mechanism of the Janzen-Connell theory, according to which, canopy-understorey negative feedbacks could be driven by canopy pathogens. Our data also suggest that temperature shapes canopy-understorey interactions, which implies that under a warming climate scenario, an acceleration of conspecific negative feedbacks could be expected, with implications for forest regeneration dynamics.

No abstract