Evolutionary trajectories of multiple defense traits across phylogenetic and geographic scales in Vitis
Data files
May 08, 2025 version files 6.28 KB
-
README.md
2.07 KB
-
species_means_final.csv
4.21 KB
Abstract
The processes driving defense trait correlations may vary within and between species based on ecological or environmental contexts. However, most studies of plant defense theory fail to address this potential for shifts in trait correlations across scales. In this work, we tested for correlations between multiple defensive traits (secondary chemistry, carbon to nitrogen ratio, domatia, leaf toughness, trichomes, and pearl bodies) across a common garden of twenty-one Vitis species and eighteen genotypes of the species Vitis riparia to identify when and where patterns of defense trait evolution persist or break down across biological scales. Additionally, we asked whether Vitis defense trait investment correlates with environmental variables as predicted by plant defense theory, using environmental metrics for each Vitis species and V. riparia genotype from the GBIF and WorldClim databases. We tested for correlations between defense trait investment, herbivore palatability, and environmental variables using phylogenetically informed models. Beyond a few likely physiological exceptions, we observed a lack of significant correlations between defense traits at both intra- and interspecific scales, indicating that these traits evolve independently of each other in Vitis rather than forming predictable defense syndromes. We did find that investment in carbon:nitrogen (at both scales) and pearl bodies increases with proximity to the equator, demonstrating support for plant defense theory’s prediction of higher investment in defenses at more equatorial environments for some, but not all, defense traits. Overall, our results challenge commonly held hypotheses about plant defense evolution, namely the concept of syndromes, by demonstrating that strong correlations between defense traits are not the prevailing pattern both across and within Vitis species. Our work also provides the first comprehensive evaluation of the evolutionary divergence in approaches that Vitis, a genus with significant agricultural value, have evolved to defend themselves against herbivores.
Dataset DOI: 10.5061/dryad.4qrfj6qp1
Description of the data and file structure
Defense metrics were collected from grape vines grown in a common environment. Location information was aggregated from the Global Biodiversity Information Facility, and bioclimate metrics for these locations were downloaded from WorldClim. Full experimental methods can be found in the accompanying manuscript.
Files and variables
File: species_means.csv
Description: all values are mean trait values/abiotic parameters for the respective species.
Variables
- species: grape species name
- genotype: internal genotype code for distinguishing Vitis riparia genotypes
- trichome_density: a 1-9 score of the density of trichomes on the underside of the leaf; no units
- domatia_index: a composite metric of the radius of the domatium multiplied by the density of the hairs of the domatium; no units
- domatia_density: a 1-9 score of the density of trichomes in the domatium area; not units
- pearl_bodies_per_cm2: the density of pearlbodies per unit of area on the underside of the leaf; measured as count per square centimeter (pb/cm^2)
- toughness: the force needed to puncture the leaf as measured by a force penetrometer; units are gram-force (gf)
- mass_consumed: mass of leaf tissue consumed by a generalist herbivore in a standardized amount of time; measured in milligrams (mg)
- CN_ratio: ratio of carbon to nitrogen in the leaf tissue; no units
- chemical_abun: abundance of metabolites in the leaf tissue, measured as the sum of the areas under the metabolite peaks on spectra and standardized using an internal standard
- chemical_rich: richness of metabolites in the leaf tissue, measured as the count of peaks on the sample's spectrum
Code/software
All analyses were performed using the R programming language, version 4.4.1. R packages used in the analysis of this data included:
corrplot
phytools
nlme
