Results from linear mixed-effects models for plant and aphid performance as depending on host plant growth conditions.

<p>Given are separate models for tansy biomass, the log response ratio (plant biomass LogRR) of tansy biomass infested with aphids compared with control plants, aphid numbers, and aphid load (aphid number per unit plant biomass) from a greenhouse experiment using <i>Macrosiphoniella tanacetaria</i> on <i>Tanacetum vulgare</i> in competition with <i>Solidago canadensis</i>. Minimum adequate models are presented together with terms removed from the model given in brackets. Superscripts give the order in which terms have been removed from the model starting with highest order interactions based on least significance. Significant terms in the final models are given in bold. A random effect for 5 blocks which was included in all models is not shown. °In the model for tansy biomass, aphid presence (AP) was included as explanatory variable, while in the model for aphid number, tansy biomass (TB) was used as a covariate. The biomass of the control plants not infected with aphids but exposed to the same other treatments was used as covariate because this biomass was not confounded with the aphid effect of reducing plant biomass. *indicates data was log transformed.