Skip to main content
Dryad logo

The effects of foraging ecology and allometry on avian skull shape vary across levels of phylogeny


Natale, Rossy; Slater, Graham (2022), The effects of foraging ecology and allometry on avian skull shape vary across levels of phylogeny, Dryad, Dataset,


Avian skull shape diversity is classically thought to result from selection for structures that are well-adapted for distinct ecological functions, but recent work has suggested that allometry is the dominant contributor to avian morphological diversity. If true this hypothesis would overturn much conventional wisdom regarding the importance of form-function relationships in adaptive radiations, but it is possible that these results are biased by the low taxonomic levels of the clades that have been studied. Using 3D morphometric data from the skulls of a relatively old and ecologically diverse order of birds, the Charadriiformes (shorebirds and relatives), we found that foraging ecology explains more than two-thirds of the variation in skull shape across the clade. However, we also found support for the hypothesis that skull allometry evolves, contributing more to shape variation at the level of the family than the order. Allometry may provide an important source of shape variation on which selection can act over short time scales, but its potential to evolve complicates generalizations between clades. Foraging ecology remains a better predictor of avian skull shape over macroevolutionary time scales. 

Usage Notes

Three notes on using this Dryad repository:

1. The three-dimensional surface scans from which the landmark data are collected are available on We are unable to include these on Dryad due to institutional policies. 

2. All markdown files have an accompanying HTML file with the same title. These HTML files contain both the code used in analyses as well as their output. We have included these since depending on the computer the R markdown files are run on, the code may take a substantial amount of time and/or may crash R. The HTML files therefore provide an alternative way to walk through the analyses step by step. 

3. After running Rmarkdown document 1 ("1_ Foraging_Ecology_and_Shape.Rmd"), the processed data is written out as several.Rdata objects. We have included this to save time and exclude running of redundant code. There are four Rdata objects ("Charadriiform_GeoMorph_Data_breeding", "Charadriiform_GeoMorph_Data_Nonbreeding", "slidcoordinates", and "trimmedtree") that are included in this repository. If you chose to run all of the code included here, you will re-generate and overwrite these files with identical versions. They are included here in case a user wants to only run portions of the code. "Chardriiform_GeoMorph_Data Breeding.Rdata" and "Charadriiform_GeoMorph_Data_Nonbreeding.Rdata" contain the information on each species' morphology, body mass, and foraging ecology, along with a phylogenetic tree of all the species in the analyses. The data in both of these files are in the correct 'geomorph data frame (gdf)' format for use in the analyses here. The only difference between these two .Rdata objects is that one ("Chardriiform_GeoMorph_Data Breeding.Rdata") contains the breeding season specific foraging data and the other ("Charadriiform_GeoMorph_Data_Nonbreeding.Rdata") contains the Nonbreeding season specific foraging data. The 'slidcoordinates.Rdata" contains the non-Procrustes aligned coordinates for use in the 'bodymasscentroids.Rmd' exploratory analyses. The 'slid' in the name refers to the fact that the semilandmarks have been slid in this dataset WITHOUT Procrustes aligning the coordinates to adjust for differences in skull size between species. The "trimmedtree.Rdata" file is the phylogenetic tree pruned down to represent only the species used in the analyses here, as opposed to all species in the previously published tree used in our analyses.


This Dryad repository contains the following:

  1. A folder titled 'landmarks'--contains species averaged 3D landmark coordinates, one CSV file per species. This folder needs to be unzipped prior to running the code- please ensure you save the collection of CSV's in a folder entitled 'Landmarks'. These landmark coordinates were collected from 3dimensional surface scans of Charadriiform skulls. We collected up to two sets of landmarks from different museum specimens of the same species and averaged these landmark sets to get the species-averaged landmark set available in these CSV files. Due to damage on some specimens (i.e. missing or damaged bones) and the limitations of surface scanning (e.g. unable to capture portions of the skull that are heavily shadowed), we estimated or mirrored landmarks on many specimens. A full list of the specimens used and notations of any of these estimation or mirroring procedures used are available in the online supplement in table S5.
  2. "Species_Ecology_Mass_Data.csv" -- a CSV file that contains all the information on each species' foraging ecology and body mass. The CSV file has 262 rows, one for each species. For each species, there are 9 columns, each storing a different variable: 1) Genus, the genus each species belongs to, included solely for exploratory purposes and are not directly used in any analyses, 2) Family, the family each species belongs to, 3)Species, a shortened name for each species that includes the genus initial and species, used in the plotting/labeling certain results, 4) Body weight, the body mass of each species in grams, taken from the handbook of the birds of the world, 5)Mechanism breeding, the method of foraging used by that species in the breeding season, 6) Sediment breeding, the sediment in which foraging occurs during the breeding season, 7) Mechanism non-breeding, the method of foraging used in the non-breeding season, 8) Sediment non-breeding, the sediment in which foraging occurs during the nonbreeding season, and 9) Broader_Guild, one of ten broader foraging categories each species was placed into for use in a supplementary analyses run at the end of "1_ Foraging_Ecology_and_Shape.Rmd" to provide a comparison to the main analyses presented in our paper.
  3. 'Untrimmed_Tree.tre'-- the untrimmed phylogenetic tree, this tree contains more species than were included in our analyses. Prior to analyses, our code runs through dropping tips from the tree that we did not have data for and contains code to substitute in synonymous taxonomic names so the tree, morphological & ecological data all have identical species names. 
  4. Three Rmarkdown files that run through the analyses presented in our paper. These correspond to three separate, but related, questions we ask in our manuscript:
    1.  "1_ Foraging_Ecology_and_Shape" -- investigates the relationship between foraging ecology (both broadly and finely defined), body mass, and skull shape. This Rmarkdown file will also run through a series of data processing steps and will write out the processed data as .Rdata files so the user does not need to re-run the data processing steps in the second Rmarkdown document. 
    2.  "2_Testing_for_the_Evolution_of_Allometry" -- investigates support for a model of common or evolving allometries across different Charadriiform families. This document will load in the processed data from the first Rmarkdown document, then will run through a comparison of different models (one with and one without an interaction between family and body mass) of allometry across the Charadriiform order.  
    3.  "3-_Variability_of_Allometry_Across_Families" -- this code will then run through some family-level analyses on four Charadriiform families with large enough sample sizes to investigate smaller scale patterns. In each family, the explanatory power of body mass and foraging ecology on skull shape is assessed and the shape changes associated with body size are visualized. 
  5. "bodymasscentroids.Rmd" -- this is an additional Rmarkdown file that walks through our exploration of how body mass and centroid size vary with one another. These results are presented in the online supplement to the main manuscript. 
  6. "Breeding_Analyses_Repeat.Rmd"-- this is an additional Rmarkdown file that walks through the analyses run in the files under D EXCEPT these analyses use the breeding season-specific foraging ecology instead of the non-breeding season-specific foraging ecology. These results are presented in our online supplement and the code does NOT contain specific details on what is being done given that it is identical to code run in the markdown documents under D. This file does contain some guiding notes on which analyses are being done in what portions, but we refer users to the main R markdown documents under D for a thorough walkthrough of our analyses.