Data from: Strong bat predation and weak environmental constraints predict longer moth tails
Data files
Apr 02, 2025 version files 1.25 GB
-
DatasetS1_Rubin_etal_ProcB2025.xlsx
11.97 KB
-
DatasetS2_Rubin_etal_ProcB2025.xlsx
15.40 KB
-
README.md
6.83 KB
-
Supplementary_Archive_1.zip
53.57 KB
-
Supplementary_Archive_4.zip
52.30 MB
-
Supplementary_Archive2.zip
1.02 GB
-
Supplementary_Archive3.zip
178.12 MB
Abstract
Elaborate traits evolve via intense selective pressure, overpowering ecological constraints. Hindwing tails that thwart bat attack have repeatedly originated in moon moths (Saturniidae), with longer tails having greater anti-predator effect. Here, we take a macroevolutionary approach to evaluate the evolutionary balance between predation pressure and possible limiting environmental factors on tail elongation. To trace the evolution of tail length across time and space, we inferred a time-calibrated phylogeny of the entirely tailed moth group (Actias + Argema) and performed ancestral state reconstruction and biogeographical analyses. We generated metrics of predation via estimates of bat abundance from nearly 200 custom-built species distribution models and environmental metrics via estimates of bioclimatic variables associated with individual moth observations. To access community science data, we developed a novel method for measuring wing lengths from unscaled photos. Integrating these data into phylogenetically-informed mixed models, we find a positive association between bat predation pressure and moth tail length and body size, and a negative association between environmental factors and these morphological traits. Regions with more insectivorous bats and more consistent temperatures tend to host longer-tailed moths. Our study provides insight into tradeoffs between biotic selective pressures and abiotic constraints that shape elaborate traits across the tree of life.
Description of the data and file structure
There are 4 supplementary archives and 2 supplementary datasets:
Supplementary_Archive1 : Statistical scripts and data sheets
- Data sheets containing all moths analyzed in this study and their associated trait measurements and lat/long coordinates; Definitions for Rubin_etal_2025_fulldata; variables as follows:
- Picture_ID = picture used for measurements. This corresponds with the names of the pictures that can be found in Supplementary_Archive2
- Species = full scientific name;
- genus = genus;
- spp = specific epithet;
- decimalLatitude, decimalLongitude = lat/long coordinates taken from GBIF or the specimen’s museum data;
- Ant_length = Antenna length in cm (if the photo had a scale bar) or pixels (if the photo did not have a scale bar);
- FW_length = Right forewing length in cm (if the photo had a scale bar) or pixels (if the photo did not have a scale bar). This is a proxy for body size;
- HW_length = Right hindwing length in cm (if the photo had a scale bar) or pixels (if the photo did not have a scale bar);
- no_scale = binary code where a 0 means the photo had a scale bar and 1 means the photo did not;
- mean_spp_Ant = species mean for antenna length in cm (taken from scaled photos);
- Adj_Ant_length = Antenna length /mean species antenna length;
- Adj_FW_length = Forewing length/(antenna length/mean species antenna length);
- Adj_HW_length = Hindwing length/(antenna length/mean species antenna length);
- batRichness = number of species of large, insectivorous bats at the given lat/long coordinate. Estimated from custom-built species distribution models;
- BatPop = estimated abundance of bats at the given lat/long coordinate;
- mean_annualTemp = Annual mean temperature from Worldclim in degreesC *10 units (BIO1);
- mean_TempSeasonality = standard deviation of mean annual temperature from Worldclim (BIO4);
- mean_precip = mean annual precipitation in mm from Worldclim (BIO12);
- median_LGP: median length of growing period (in days out of a 365 day year) extracted from Food and Agriculture Organization of the UN and is determined by soil temperature and available moisture, accounting for transpiration.
- Definitions for Rubin_etal2025_Calibration_data.csv are the same
- Picture_ID = picture used for measurements. This corresponds with the names of the pictures that can be found in Supplementary_Archive2
- genus = genus
- species = full scientific name
- specificEpithet = specific epithet
- Thorax_width = distance between the anterior forewing joint in cm; NA’s are used when this distance is not measurable because the forewing positioning does not allow for accurate measurement
- Ant_length_cm = antenna length in cm, from most distal tip of antenna to the junction with the head
- FW_length_cm = distance from most distal tip of forewing (right, if possible) to the middle of the forewing hinge joint on the thorax
- HW_length_cm = distance from the most distal tip of hindwing (right, if possible) to the middle of the hindwing hinge joint on the thorax
- Mean_ant_cm = average male antenna length for the given species (from scaled photos)
- Stdev_ant_cm = standard deviation of male antenna length for the given species (from scaled photos)
- Ant_pixels = length of antenna measured in the same way as above, but without using scale bar
- FW_pixels = length of forewing measured in the same way as above, but without using scale bar
- HW_pixels = length of hindwing measured in the same way as above, but without using scale bar
- adj_Ant = adjusted antenna length; Ant_pixels / Mean_ant_cm
- adj_FW = adjusted forewing length; FW_pixels / adj_Ant
-
adj_HW = adjusted hindwing length; HW_pixels / adj_Ant
Note: outgroups (Anthereaea, Syntherata, Saturnia, Opodiphthera) do not have adjusted measurements (i.e., value is NA for adj_Ant, adj_FW, adj_HW) because we were only interested in creating these wing length estimates from photos without a scale bar for our ingroup species
- Rubin_etal_2025_ActiasTree.nex Dated Bayesian tree of the tailed moon moth clade (+outgroups). See text for more information.
- R scripts for the phylogenetically-informed models (Rubin_etal_ProcB2025_pglmm-models.R) and comparative trait analyses (Rubin_etal2025_ASR&CompTraitMapping.R)
Supplementary_Archive2 : All photos
- A zipped folder containing all the specimen photos we analyzed
Supplementary_Archive3 : Tree inference
- A zipped folder containing all cleaned loci in FASTA format (Loci.zip)
- An amino acid super matrix (FcC_supermatrix.fas) and partition file (FcC_supermatrix_partition.txt) generated by FASconCAT
- The best ML tree files, with Actias_fullTree.BEAST* being the ML tree that we subsequently used as our starting tree for our Bayesian analysis and IQTREE* being the files pertaining to the ML tree we built using the MFP+Merge function in IQ-TREE (see main text for more details).
- The starting tree (from Actias_fullTree.BEAST files above) modified to align with the timing of the four calibration points from Kawahara et al. 2019 (see main text for more details) = StartingTreeForBEAST.nex.
- All .xml input files for running the four Bayesian trees in BEAST (Yule, Birth-Death : fixed or classic operator)
- MCMC trees from each inferred Bayesian model (MCMC.trees) as well as the maximum likelihood results file for each inferred tree (.mle.result.log); name components: 1CL = 1 clock, BD = birth-death model, Yule = Yule model, Classicop = classic operator mix (allowing BEAST to infer topology), Fixed = tree topology constrained to the topology of the maximum likelihood input tree), Uniform = uniform distribution = distributions used at the calibration nodes, 200milgen = tree inference run for 200 million generations. See main text and supplement for more details.
- A summary species tree generated from ASTRAL (Astral_speciestree.tre)
Supplementary_Archive4 : Ancestral range estimation
- Input dated tree using Nexus format (*.txt)
- Dispersal multipliers and geographic regions input files (*.txt)
- R script (*.R) used to run the BioGeoBEARS analysis on UF’s higher-performance computing cluster
- Output files (resfn, *.Rdata)
Supplementary Dataset S1:
Sequence information for phylogenetic inference
Supplementary Dataset S2:
Target insectivorous bat list (sheet1) and bat species list for which we were able to build SDMs (sheet2)
Sharing/Access information
All data was generated for this project and more information can be found in the main text.
To access species distribution modeling scripts and outputs, please got to: 10.5281/zenodo.8117743