Ecological Archives E096-222-A2

Lauren M. Smith and Heather L. Reynolds. 2015. Plant–soil feedbacks shift from negative to positive with decreasing light in forest understory species. Ecology 96:2523–2532. http://dx.doi.org/10.1890/14-2150.1

Appendix B. The role of Phase I biomass in driving the light × PSF interaction.

In addition to the full model used to test the light-feedback interaction, several models were constructed and compared in order to tease apart biomass during the conditioning phase (Phase I) vs. other light-mediated effects in driving the light-feedback relationship.

To understand how the size of Phase I plants influenced feedback, aboveground biomass was added to the original model as a covariate (Table B1). Adding aboveground biomass did not change the significance of the Source × Response or Light × Source × Response interactions in the model.  Phase 1 aboveground biomass was not a significant predictor of Phase II responses (p = 0.28), indicating that other light-related factors drove responses.

Understanding the relationship between root biomass in Phase I feedback was limited by the fact that not all species’ root biomass could be measured in Phase I. We analyzed two submodels designed to evaluate the importance of root biomass in driving feedback for the four species where these data were available (Asarum, Conoclinium, Euonymus, and Lindera). Table B2 shows the results for Submodel 1, where the factors included in the model were identical to those in the full model. The significance of the Light × Source × Response interaction is lost for this subset of data, likely because three of the four species tend to show neutral or nonlinear feedback patterns across light treatments. Table B3 shows Submodel 2, where root biomass was substituted for light as a factor in the model. The Below Phase I × Source × Response interaction becomes significant, indicating that Phase I root biomass interacts significantly with feedback for these four species, and therefore could be an important driver of the light × feedback relationship reflected in the full model. We note, however, that even for this data subset, Submodel 1 (AIC = 306.9) fit the data better than Submodel 2 (AIC = 713.5), indicating that Phase 1 root biomass alone did not explain Phase II responses as well as light treatment.

Table B1. Results of mixed-effects linear model for Phase II whole-plant biomass with Phase I aboveground biomass included as a factor.

Light = Low, medium, or high light availability. Source = the identity of the species used to condition the soil inoculum during Phase I of the experiment. Response = the identity of the species measured in Phase II. Initial Weight = starting mass for species grown from cuttings. Phase I Above Biomass = the aboveground biomass of the Source plant used to condition inoculum during Phase I. Block and its interactions with all factors were included as random effects. Bold text indicates significance at the p = 0.05 level.

Table B2. Results of Submodel 1.

Light = Low, medium, or high light availability. Source = the identity of the species used to condition the soil inoculum during Phase I of the experiment. Response = the identity of the species measured in Phase II. Initial Weight = starting mass for species grown from cuttings. Block and its interactions with all factors were included as random effects.

Table B3. Results of submodel 2.

Below Phase I = root biomass measured in source plants used to condition soil in Phase I of the experiment. Source = the identity of the species used to condition the soil inoculum during Phase I of the experiment. Response = the identity of the species measured in Phase II. Initial Weight = starting mass for species grown from cuttings. Block and its interactions with all factors were included as random effects.

[Back to E096-222]