Question: In Amazonian moist forest, four questions arose: 1. Do tree species differ in their susceptibility to lianas? 2. What host tree traits (branch-free bole height, growth rate, bark type, leaf length and adult stature) are correlated with the susceptibility of tree species to lianas infesting the trunk and the crown? 3. To what extent do spatial variables (proximity to liana-infested trees and the light environment of the tree crown) affect the likelihood of liana infestation? 4. Are spatial variables or tree traits relatively more important in influencing the susceptibility of trees to lianas? We address all questions separately for trunk and crown infestation.

Location: Tambopata Nature Reserve, Peru.

Methods: We collected information on liana infestation, tree morphological traits, growth, light-environment and position for 3675 trees in seven 1-ha permanent sample plots. We separated trunk from crown infestation and used correlation and logistic regression analyses for tree species and individual tree-level analyses, respectively.

Results: Half of all trees were colonised by at least one liana. Of 41 relatively common dicot tree species, at least five have significantly greater and three significantly lower crown infestation rates than expected by chance. Trunk and crown infestation are influenced by different host traits – trunk infestation was only affected by bark type, while crown infestation is reduced when trees are fast-growing, tall, have low-density wood, long branch-free boles and long leaves. The likelihood of both trunk and crown infestation increases for trees growing in close proximity to another liana-infested tree, but is invariant with the light environment of tree crowns.

Conclusion: Crown and trunk infestation have not been properly distinguished before; it is important to do so as the factors determining the different modes of infestation differ fundamentally. The association between crown infestation and tree traits suggests that increases in liana dominance in Amazonian forests could cause changes in forest composition, including favouring faster growing tree species with low density wood, potentially reducing the carbon stored by mature forests.

Nomenclature: We used current nomenclature at the time of analysis as provided at TROPICOS  http://www.tropicos.org/