Code and data for the manuscript "The evolution of trait variance creates a tension between species diversity and functional diversity"

System requirements

• Operating systems: CentOS 7, Ubuntu 19.10, Ubuntu 20.04, MacOS 10.13, Windows 10.
• Software dependencies: R (has been tested with R version 4.1.2).
• Required R packages:
  • deSolve (version 1.28): integrating differential equations
  • mvtnorm (version 1.1-2): multivariate normal distributions
  • tidyverse (version 1.3.1): efficient data manipulation and plotting
  • ks (version 1.13.4): multidimensional kernel density estimation
• Required non-standard hardware: none.
• Typical installation time on a normal desktop computer: no appreciable time if R is already installed. Otherwise, it is the installation time of R and the above three packages.

Repository contents

The directory Galapagos_snails has the files:

• snaildata.csv: the raw snail data, in comma-separated value format
• readme.txt: short description of the data file snaildata.csv
• vegzonetotals.csv: the number of distinct host plant species found on the humid and arid zones of each island, in comma-separated value format
• snail_analysis.R: code to obtain and plot species- and functional diversity in the snail dataset
• snail_analysis_kernsmooth.R: code to obtain and plot species- and functional diversity in the snail dataset, using kernel density estimates instead of fitting multivariate Gaussian distributions to estimate functional diversity
• snail_distance_dep.R: code to obtain and plot species- and functional diversity in the snail dataset, with distance-dependent diversity metric

The directory Model_simulation has the files:

• model.R: code that integrates the model equations and displays their results, for a single parameterization
• solve_eqs.R: functions used for numerically solving the model equations, and to re-format the final results in a tidy table
• plot.R: functions for plotting the results
• diversity.R: functions to obtain species- and functional diversity
• run_replicates.R: script to help directly replicate the numerical results

The directory No_variance_evoluton_null_model has the same files as Model_simulation. Everything is identical, except that trait variances do not evolve: the right-hand sides of the differential equations for the entries of the genetic covariance matrix are all set to zero. (Note: this implementation is not optimized for efficiency. Instead, it emphasizes that the sole difference between this null model and the original model is the lack of trait variance evolution.)

The directory No_interaction_null_model has the files:

• diversity.R: functions to obtain species- and functional diversity
• no_interaction.R: script to perform all null model simulations, as well as obtain species- and functional diversity calculations for each

The directory General_1D_model is like Model_simulations, but it only works for one-dimensional trait spaces. In return, it implements a fully general interaction structure between phenotypes, instead of restricting analysis to Gaussian interaction kernels. The code here implements two alternative interaction structures: one that is asymmetric, and one that is nonlocal (hierarchical competition with a competition-mortality tradeoff).

The directory Videos has four short videos of the model's dynamics through time:

• mov_1D_S2.mp4: eco-evolutionary dynamics in a one-dimensional trait space, initialized with two species.
• mov_1D_S8.mp4: eco-evolutionary dynamics in a one-dimensional trait space, initialized with eight species.
• mov_2D_S3.mp4: eco-evolutionary dynamics in a two-dimensional trait space, initialized with three species.
• mov_2D_S25.mp4: eco-evolutionary dynamics in a two-dimensional trait space, initialized with twenty-five species.