Statistical Analysis
Data was assessed for normality, and non-normal data was transformed
using natural logarithms or square root transformations when necessary.
If data transformations were unsuccessful, non-parametric tests were
used to determine significance. A one-way ANOVA followed by Tukey’spost hoc test was used for caterpillar RGR and plant nutrient
content analysis. Caterpillar percentage leaf consumption area was
compared using paired T-tests between the treatments for the two-choice
assays. For the three-caterpillar choice assay, the percentage leaf area
consumed was compared using the non-parametric Friedman’s test, followed
by Dunn’s post-hoc test. Caterpillar first finishes and
establishes were compared using Chi-square tests with 1:1 choice
assumption. Differences in the average eggs laid on no-salt-treated and
salt-treated plants were compared using a two-way ANOVA (repeated
measures model). Plant defense protein levels (PPO and TPI) were
analyzed using a mixed-effects model. Volatile emission differences
under salt treatment were analysed using a permutational multivariate
analysis of variance (PERMANOVA, with salt x herbivory as factors).
Student’s T-test was conducted between no-salt and salt treatments under
herbivory and no herbivory for individual volatile compounds.