Skip to main content
Dryad logo

Data from: Periodic environmental disturbance drives repeated ecomorphological diversification in an adaptive radiation of Antarctic fishes

Citation

Parker, Elyse et al. (2022), Data from: Periodic environmental disturbance drives repeated ecomorphological diversification in an adaptive radiation of Antarctic fishes, Dryad, Dataset, https://doi.org/10.5061/dryad.bk3j9kdbg

Abstract

The ecological theory of adaptive radiation has profoundly shaped our conceptualization of the rules that govern diversification. However, while many radiations follow classic early burst patterns of diversification as they fill ecological space, the longer-term fates of these radiations depend on many factors, such as climatic stability. In systems with periodic disturbances, species-rich clades can contain nested adaptive radiations of subclades with their own distinct diversification histories, and how adaptive radiation theory applies in these cases is less clear. Here, we investigated patterns of ecological and phenotypic diversification within two iterative adaptive radiations of cryonotothenioid fishes in Antarctica’s Southern Ocean: crocodile icefishes and notoperches. For both clades, we observe evidence of repeated diversification into disparate regions of trait space between closely related taxa and into overlapping regions of trait space between distantly related taxa. We additionally find little evidence that patterns of ecological divergence are correlated with evolution of morphological disparity, suggesting that these axes of divergence may not be tightly linked. Finally, we reveal evidence of repeated convergence in sympatry that suggests niche complementarity. These findings reflect the dynamic history of Antarctic marine habitats, and may guide hypotheses of diversification dynamics in environments characterized by periodic disturbance.

Usage Notes

Descriptions of code scripts:

procD-pgls.R

R script used to test for correlations among Procrustes body shape coordinates and major axes of ecological disparity using phylogenetic generalized least squares (PGLS) regression. These analyses were performed using the procD.pgls function implemented in the R package geomorph (Baken et al. 2021; Adams et al. 2021). 

PGLS.R

R script used to test for relationships among phenotypic (body size and PC axes 1-3 of body shape variation) and ecological (buoyancy, depth, and PC axes 1-3 of diet variation) variables using phylogenetic generalized least squares (PGLS) regression. These analyses were performed using the gls function implemented in the R package nlme (Pinheiro et al. 2017).

DTT.R

R script used to calculate relative subclade disparity through time (DTT; Harmon et al. 2003) in the traits of mean percentage buoyancy, mean depth, and maximum body size. These analyses were performed using the ddt function implemented in the R package geiger2 (Pennell et al. 2014). 

PGLMM.R

R script used to test the relationships among range overlap, clade age, and ecomorphological diversification for all pairwise comparisons of icefishes and notoperches using phylogenetic generalized linear mixed models (PGLMMs). These analyses were performed using the MCMCglmm function implemented in the R package MCMCglmm (Hadfield 2010). 

Descriptions of files:

all_chan.tps

Landmark data for all icefish (Channichthyidae) species analyzed in this study. This file includes the coordinates of 28 landmarks (24 fixed landmarks and 4 sliding semi-landmarks) placed on a total of 45 specimens. The start of landmark information for each specimen is indicated by "LM=" and ends with the scale factor, labeled "SCALE=". The variable "LM" corresponds to the number of landmarks placed on each specimen, "IMAGE" refers to the file name of the specimen image on which landmarks were placed, "ID" corresponds to the species identity of each specimen image, and "SCALE" refers to the scale adjustment applied to each photo in tpsDig. 

all_trem.tps

Landmark data for all notoperch (Trematominae) species analyzed in this study. This file includes the coordinates of 28 landmarks (24 fixed landmarks and 4 sliding semi-landmarks) placed on a total of 125 specimens. The start of landmark information for each specimen is indicated by "LM=" and ends with the scale factor, labeled "SCALE=". The variable "LM" corresponds to the number of landmarks placed on each specimen, "IMAGE" refers to the file name of the specimen image on which landmarks were placed, "ID" corresponds to the species identity of each specimen image, and "SCALE" refers to the scale adjustment applied to each photo in tpsDig. 

chan_classifiers.csv

A 45 x 2 matrix containing information on species identity for each icefish specimen included in the landmark dataset. This file is used as input in the procD-pgls.R script.

chan_diet.csv

Diet data for all icefish (Channichthyidae) species analyzed in this study. For each species listed in this file, diet is represented by mean frequencies of occurrence of 39 different prey categories identified from 34 studies that performed stomach content analyses across both icefishes and notoperches. For each prey item, we calculated the mean prey frequency of occurrence weighted by the number of individual observations in each study for each prey item. Included in this file is the total number of specimens examined across all studies for each focal species. 

chan_PairwiseTraits.csv

Data matrix including data on sympatry, time since common ancestry, and Euclidean distances in phenotypic and ecological traits for all possible pairwise comparisons of icefish species. This is the input data file used for fitting phylogenetic generalized linear mixed models to test the relationships among range overlap, clade age, and ecomorphological divergence. The column label "lineage 1" refers to the "focal" species, and "lineage 2" refers to the species being compared to "lineage 1." The column "sympatry" indicates whether the species pair of interest is sympatric (1) or allopatric (0), and the column "age" indicates age of the most recent common ancestor shared by the species pair of interest. Columns labeled "dietPC1", "dietPC2", "dietPC3", "shapePC1", "shapePC2", "shapePC3", and "size" are the Euclidean distances in each of these traits for all pairwise comparisons of species.

chan_ranges.csv

Geographic range data for all icefish (Channichthyidae) species analyzed in this study. Species are coded as present (1) or absent (0) from each of four geographic regions: (1) the Antarctic Peninsula and nearby islands, including the South Shetland and South Orkney Islands; (2) South Georgia, Bouvet, and the South Sandwich Islands; (3) Kerguelen Island; and (4) the High Latitude Antarctic, defined as the region of marine habitats found at latitudes higher than 66 degrees South.

chan_sympatry_matrix.csv

Matrix classifying each pairwise comparison of icefish (Channichthyidae) species as sympatric (1) or allopatric (0). A given pair of species was considered sympatric if they were both present (1) in at least one of the areas listed in the geographic range data file ("chan_ranges.csv"). Otherwise, the pair was considered allopatric. 

ice-trem_datafile

Compilation of all trait data for all icefish and notoperch species examined in this study. This file includes data on maximum body size ("MaxSize", represented by Total Length, cm), mean depth distribution ("MeanDepth", given in meters), mean percentage buoyancy ("Buoyancy"), the first three PC axes of body shape variation ("BodyShapePC1", "BodyShapePC2", and "BodyShapePC3"), and the first three PC axes of diet variation ("DietPC1", "DietPC2", and "DietPC3"). Data on maximum body size were obtained from Eastman (2019) and references therein. Mean depths of occurrence were calculated from trawl records for adult specimens supplied by U.S. Antarctic Marine Living Resources Program Antarctic finfish surveys conducted between 1998 and 2012. Values representing mean percentage buoyancy were obtained from Near et al. (2012). PC axes 1-3 of body shape variation were obtained from a Principal Components Analysis of Procrustes body shape coordinates. PC axes 1-3 of diet variation were obtained from a Principal Components Analysis of the weighted mean frequency of prey occurrence data.

noto_sliders.txt

Matrix indicating the positions of the 4 sliding semi-landmarks placed on images of icefish and notoperch specimens. The column label "slide" indicates the number of the sliding semi-landmarks; "before" indicates the number of the fixed landmark placed immediately before the sliding landmark; and "after" indicates the number of the fixed landmark placed immediately after the sliding landmark.

notothenioid_timetree.tre

Time-calibrated phylogeny for notothenioids inferred using BEAST v2.4.7. This tree was inferred based on the DNA alignment from Near et al. (2018) that includes 104,709 loci captured from 80 species of notothenioids, including our 14 focal crocodile icefish species and our 16 focal notoperch species.

trem_classifiers.csv

A 125 x 2 matrix containing information on species identity for each notoperch specimen included in the landmark dataset. This file is used as input in the procD-pgls.R script.

trem_diets.csv

Diet data for all notoperch (Trematominae) species analyzed in this study. For each species listed in this file, diet is represented by mean frequencies of occurrence of 39 different prey categories identified from 34 studies that performed stomach content analyses across both icefishes and notoperches. For each prey item, we calculated the mean prey frequency of occurrence weighted by the number of individual observations in each study for each prey item. Included in this file is the total number of specimens examined across all studies for each focal species.

trem_PairwiseTraits.csv

Data matrix including data on sympatry, time since common ancestry, and Euclidean distances in phenotypic and ecological traits for all possible pairwise comparisons of notoperch species. This is the input data file used for fitting phylogenetic generalized linear mixed models to test the relationships among range overlap, clade age, and ecomorphological divergence. The column label "lineage 1" refers to the "focal" species, and "lineage 2" refers to the species being compared to "lineage 1." The column "sympatry" indicates whether the species pair of interest is sympatric (1) or allopatric (0), and the column "age" indicates age of the most recent common ancestor shared by the species pair of interest. Columns labeled "buo", "depth", "dietPC1", "dietPC2", "dietPC3", "shapePC1", "shapePC2", "shapePC3", and "size" are the Euclidean distances in each of these traits for all pairwise comparisons of species.

trem_ranges.csv

Geographic range data for all notoperches (Trematominae) species analyzed in this study. Species are coded as present (1) or absent (0) from each of four geographic regions: (1) the Antarctic Peninsula and nearby islands, including the South Shetland and South Orkney Islands; (2) South Georgia, Bouvet, and the South Sandwich Islands; (3) Kerguelen Island; and (4) the High Latitude Antarctic, defined as the region of marine habitats found at latitudes higher than 66 degrees South.

trem_sympatry_matrix.csv

Matrix classifying each pairwise comparison of notoperch (Trematominae) species as sympatric (1) or allopatric (0). A given pair of species was considered sympatric if they were both present (1) in at least one of the areas listed in the geographic range data file ("trem_ranges.csv"). Otherwise, the pair was considered allopatric.

Funding