Intraspecific trait variation shows the predominant influence of functional divergence on species co-occurrence
Data files
Dec 06, 2025 version files 414.97 KB
-
Code_1._calculating_multivariate_trait_means_and_variances_at_each_spatial_level.R
6.66 KB
-
Code_2._exploration_of_species_means_and_variances_for_transformations.R
3.75 KB
-
Code_3._trait_differences_among_and_within_species.R
47.47 KB
-
Code_4._trait_differences_among_and_within_species_by_trait_axis.R
34.75 KB
-
mean_trait_values_by_species_for_each_trait_axis.csv
7.59 KB
-
metadata_and_metacode.csv
5.07 KB
-
raw_data.csv
176.17 KB
-
README.md
5.76 KB
-
species_abundances_by_plot.csv
29.38 KB
-
trait_mean_and_variance_by_plot_for_each_species_and_trait_axis.csv
97.72 KB
-
traits_nonflowering_species.csv
664 B
Abstract
Intraspecific trait variation (ITV) has been recognized as an important source of variation in communities. We separated among-population ITV (differences in mean trait values among populations) from within-population ITV (trait variation within populations) to test three theoretical predictions: i) communities should show greater among-species trait diversity when among-population ITV is accounted for due to trait means shifting away from neighbouring species, ii) the increased differences among species should be greatest in species poor communities, and iii) within-population trait variation should also be greatest in species-poor communities.
We tested these predictions in an experimental old-field metacommunity, measuring four functional traits known to be important for plant function in 1842 individuals across 33 species and 84 patches that varied in size (area), species diversity, and composition.
We found that when among-population ITV was accounted for, species diverged in their average trait values (nearest neighbour distances and functional dispersion), consistent with theory that trait differences among species promote coexistence. This divergence was slightly greater in species-poor communities. Within-population ITV showed a different pattern than predicted, with the highest within-population ITV occurring in large patches and in communities where species were more functionally differentiated.
Overall, among-population ITV revealed the influence of ecological interactions on trait divergence among species. Failing to incorporate ITV underestimates among-species trait differences and may miss opportunities to detect trait-based assembly and coexistence mechanisms. Nonetheless, estimates that ignore ITV are strongly correlated with those that include ITV, suggesting that the relative differences in functional diversity among communities can be estimated without ITV. Within-population ITV showed less consistent results, suggesting that other factors may have a stronger influence within populations.
Synthesis. Theory predicts that variation within species may alter species coexistence and the functional diversity of communities. We found support for the hypothesis that species show higher divergence in their average traits when ITV is accounted for. However, we found little support for the hypothesis that ITV within populations is greater in patches with few species. These results suggest that variation among populations moderates species interactions, facilitating species co-occurrence and shaping community assembly.
Dataset DOI: 10.5061/dryad.cz8w9gjj8
Description of the data and file structure
This data was collected for our study examining how the distribution of functional traits within a species changes across and within populations in response to neighbouring species. We measured four functional traits across thirty-three species in an experimental old-field meta-community at the Koffler Scientific Reserve in Newmarket, Ontario.
Files and variables
File: traits_nonflowering_species.csv
Description: traits used for species found in our study patches but with no reproductive individuals to provide traits this year. Trait data taken from Forsyth and Gilbert (Journal of Ecology. 2021;109:793ñ805) with transformations done as needed.
Variables
- species: species binomial name
- log.height: ln height (cm)
- log.sla: ln sla (cm2/g)
- LR_veg_rep_mass: ln (vegetative mass / reproductive mass) unitless
- sqrt.LDMC: sqrt (leaf_drymass / leaf_wet_mass) unitless
File: species_abundances_by_plot.csv
Description: Counts (population size) of all species in each patch used in the experiment. Grass population sizes were estimated using conversions of percent cover estimates. Methods following Forsyth and Gilbert (Journal of Ecology. 2021;109:793ñ805).
Variables
- block: block in experimental design (Fig. S1)
- size: size of one side of square patch (equal to square root of area in metres)
- plot: patch designator. Each block * size * plot is a unique patch within the experiment. Note that not all plots (patches) were used - see Methods for patch selection.
- species: species binomial name
- abundance: number of individuals in patch
- rel.abundance: number of individuals of species / number of individuals across all species in patch
File: metadata_and_metacode.csv
Description: metadata and metacode for raw data and code files
Variables
- metadata and metacode:
File: raw_data.csv
Description: Individual-level measurements of traits used in this study
Variables
- species: species binomial name
- block: block in experimental design (Fig. S1)
- size: size of one side of square patch (equal to square root of area in metres)
- plot: patch designator. Each block * size * plot is a unique patch within the experiment. Note that not all plots (patches) were used - see Methods for patch selection.
- Individual: individual plant of a given species in a single patch
- height: plant height along stem in cm
- leaf_area: area of leaf before drying in cm2
- leaf_mass: leaf wet mass (mass prior to drying) in grams
- stem_vol: not used in study. We estimated stem volume using water displacement, but discarded this trait as we were not confident in the accuracy of measurements
- plant_mass: dry mass of entire plant in grams
- leaf_drymass: dry mass of leaf used for leaf_area measurement (above), in grams
- repro_drymass: drymass of reproductive parts including flower heads (grams)
- other_drymass: some individuals were separated by leaves, stems, etc. before drying. All measures in grams
- veg_drymass: some individuals were separated by leaves, stems, etc. before drying. All measures in grams
- stem_drymass: some individuals were separated by leaves, stems, etc. before drying. All measures in grams
File: Code_4._trait_differences_among_and_within_species_by_trait_axis.R
Description: Calculating NND, Svar, Fdis, Feve and within-population ITV measures for all traits separately and associated statistical tests
File: mean_trait_values_by_species_for_each_trait_axis.csv
Description: description: metapopulation mean trait values following PCA. Data file generated from data files above in 'Code 1'
Variables
- species: species binomial name
- trait_axis: Principal components analysis axis
- site_PCA_score: metapopulation mean trait value (denotes species score on the trait axis identified)
File: Code_2._exploration_of_species_means_and_variances_for_transformations.R
Description: data exploration to determine appropriate transformations of traits
File: trait_mean_and_variance_by_plot_for_each_species_and_trait_axis.csv
Description: Patch-level trait means and variances for each species and each trait axis. Data file generated from data files above in 'Code 1'
Variables
- species: Species binomial name
- block: block in experimental design (Fig. S1)
- size: size of one side of square patch (equal to square root of area in metres)
- plot: patch designator. Each block * size * plot is a unique patch within the experiment. Note that not all plots (patches) were used - see Methods for patch selection
- trait_axis: Principal components analysis axis
- plot.PCA.score: Patch-level mean
- plot.level.variance: Patch-level variance
- num.ind: number of individuals sampled in the patch
File: Code_1._calculating_multivariate_trait_means_and_variances_at_each_spatial_level.R
Description: data processing to determine metapopulation means, patch means and patch variances of traits and PCA rotation
File: Code_3._trait_differences_among_and_within_species_with_figure_saving_code.R
Description: calculating NND, Svar, Fdis, Feve and within-population ITV measures for all traits simultaneously and associated statistical tests
Code/software
Coding for this project was done in R version 4.2.2 (R Core Team, 2022) and RStudio (Posit Team, 2023). These files can be viewed in R, RStudio, or any text editing software. Data files are attached as .csv.
