Effect of biotic neighborhood on traits is more constrained for leaf-level relative to plant-level traits

Background/Question/Methods

Trait-based ecology has been described as the “holy grail” offering the opportunity to predict changes in composition and functioning of ecosystems in response to global change from relatively simple information about the traits of the species present in a community.  However, our ability to make straightforward predictions can be significantly hampered if species’ traits vary as a function of their environment, particularly if species vary in their response to the environment (i.e. there is a species*environment interaction in trait values).  Previous work with wetland plants has shown significant species*environment interactions in response to variation in water and nitrogen availability, even among leaf-level traits that are thought to be tightly constrained.  We asked whether responses to biotic environment are similarly plastic by measuring leaf-level (SLA) and plant-level (total biomass, RGR and root:shoot) traits on individuals of four species grown at 3 densities (0, 2, or 4 competitors)  and varying competitor identity in a fully-factorial design.  We predicted that both density and competitor ID would interact with focal plant ID such that functionally similar species would have a greater effect on traits.  We also predicted that plant-level traits would show greater variation than leaf-level traits.

Results/Conclusions

As we predicted, the leaf-level trait was unaffected by either density or competitor identity – the only significant predictor of SLA was focal species identity.  For the plant-level traits, in addition to focal species identity, density as at least marginally significant for all three traits and the interactions between density and focal species identity  and between focal species and competitor species identity were significant for total biomass and root:shoot.  Interestingly, there was no main effect of competitor species identity on any of the traits, suggesting that, at least among this set of species, the traits of competitors are relatively unimportant in determining traits of focal species.  Taken together, these results show that the biotic environment exerts much weaker control on leaf-level traits than resource availability.  Furthermore, even for plant-level traits, variability in traits is primarily a function of the density of neighbors rather than the identity of the neighbors.  This suggests that predictions based on traits should be relatively insensitive to shifts in community composition.