Herbivore pressure and constitutive and induced defense
Assessed by S. littoralis larval weight gain, constitutive defense (-1 × larval weight gain on healthy plants) was positively correlated with herbivore pressure (assessed as described above) across all 12 non-native plant species (χ² = 0.00, P = 0.020), whereas there was no correlation between constitutive defense and herbivore pressure across the 12 native congeners (χ² = 0.32, P = 0.306, Figure 2A). Induced defense [-1 × (larval weight gain on previously damaged plants - larval weight gain on healthy plants)] was negatively correlated with herbivore pressure for non-native plant species (χ² = -0.69, P = 0.014), whereas there was no correlation between induced defense and herbivore pressure for native congeners (χ² = -0.49, P = 0.106) (Figure 2B). Overall, constitutive defense of non-native plant species was 38.7% lower than that of native congeners (χ² = 4.28, P = 0.039) (Figure 2A), while induced defense of non-native plant species was 55.3% higher than that of native congeners (χ² = 5.56, P = 0.018) (Figure 2B).
For chemical defenses assessed by phenolics, results were similar to that assessed by larval weight gain. Constitutive defense (phenolics in healthy plants) (χ² = 0.76, P = 0.004) and induced defense (phenolics in previously attacked plants - phenolics in non-attacked plants) (χ² = -0.84, P < 0.001) were positively and negatively correlated with herbivore pressure, respectively, across all 12 non-native plant species (Figure 2C and 2D). Whereas, for native congeners, defense assessed by phenolic content showed no correlation between constitutive defense and herbivore pressure (χ² = -0.20, P = 0.525) or between induced defense and herbivore pressure (χ² = -0.18, P = 0.571, Figure 2C and 2D). Overall, constitutive defense of non-native plant species, measured as phenolic content, was equal to that of native congeners (χ² = 0.02, P = 0.885, Figure 2C), whereas induced defense was higher than that of native congeners (χ² = 163.20, P < 0.001, Figure 2D). However, for defense assessed by triterpenoids, there was no relationship between constitutive defense and herbivore pressure (non-natives, χ² = -0.08, P = 0.807; natives, χ² = -0.20, P = 0.527), or between induced defense and herbivore pressure (non-natives, χ² = -0.07, P = 0.820; natives, χ² = 0.01, P = 0.970) for non-native plant species and native plant species (Figure 2E and 2F). Constitutive defense (χ² = 0.03, P = 0.852) and induced defense (χ² = 2.15, P = 0.143) of non-native plant species, measured as triterpenoids, was equal to that of native congeners (Figure 2E and 2F).
Across all herbivory treatments (leaves from un-attacked and attacked plants) and plant species (non-natives and natives), larval weight gain was negatively correlated with phenolic content (r = -0.29, P = 0.047, Figure 3A) and with triterpenoid content (r = -0.29, P = 0.044, Figure 3B).