Trait measurements
The plant genotypes used in this study were previously shown to harbour high genetic variation in direct and indirect defence against the strawberry leaf beetle (Weber et al. 2020a; Weber et al.2020b), as evidenced by the differential performance and preference of both the herbivore and its specialist endoparasitoid, Asecodes parviclava Thompson (Hymenoptera: Eulophidae), on diets of varying genotypic composition. Several primary and secondary compounds were ultimately identified as markers associated with either direct or indirect defence (Weber et al. 2020a). We included here the five compounds most strongly associated with defence: two carbohydrates (dehydroascorbic acid and myo-inositol), and three phenolics (catechin, dihydroxybenzoic acid, and shikimic acid). Details of leaf sampling and GC/TOF-MS metabolomic profiling are presented in Weber et al.(2020a), which was conducted concurrently for the same plants used in this study. Owing to capacity limitations experienced with metabolomic profiling, defence-related traits could only be quantified for 27 of the 81 genotypes. However, selection of these 27 genotypes was made such that the full spectrum of genetic variation in direct and indirect defence was still represented (following Weber et al. 2020a and Weber et al. 2020b). The four attractive traits examined were plant size, total flower number, flower frost tolerance, and inflorescence density. Frost damaged flowers show a complete blackening of the receptacle at the centre of the flower, and in this way could be expected to impact floral attractiveness to pollinators and florivores. Details of how each trait was quantification are provided in Eganet al. (2018).