http://orcid.org/0000-0002-4915-8349
Figures
Abstract
Background
Body shape can be both a consequence and cause of a species’ evolution and ecology. There are many examples of phenotypes associated with specific ecological niches, likely as a result of specific selective regimes. A classic example of this is the phenotypic change associated with colonization of caves, including body and limb elongation. However, studies explicitly testing for differences in body shape between cave-dwelling and non-cave-dwelling lineages have been limited and so the role of the cave environment in determining morphological characteristics is still not completely understood. Here we examine variation in body shape among 405 individuals representing 20 species in the salamander genus Eurycea (Plethodontidae) and select outgroups exhibiting great diversity in morphology, ecological niche, and life history.
Results
After analyzing morphometric data in a phylogenetic context using phylogenetic MANOVA and examination of the phylomorphospace, we found significant differences in body shape among cave-dwelling and non-cave-dwelling species and between aquatic and terrestrial species. Notably, limb elongation and reduced body and tail size characterized cave-dwelling species. Terrestrial species also exhibited elongation of the limbs and digits. We also observed differences in shape variance among paedomorphic and biphasic species. Our results suggest that the functional limitations imposed by habitat and life history played a key role in the evolution of body shape in this group in the context of their phylogenetic history.
Citation: Edgington HA, Taylor DR (2019) Ecological contributions to body shape evolution in salamanders of the genus Eurycea (Plethodontidae). PLoS ONE 14(5): e0216754. https://doi.org/10.1371/journal.pone.0216754
Editor: Stefan Lötters, Universitat Trier, GERMANY
Received: January 16, 2019; Accepted: April 26, 2019; Published: May 15, 2019
Copyright: © 2019 Edgington, Taylor. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the manuscript and its Supporting Information files.
Funding: The authors received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Body shape is a key part of morphological variation among vertebrates, with impacts on function and ecology [1–2]. Variation in shape may be a result of environmental effects, structural or functional constraints, adaptive differentiation, or shared phylogenetic history [3–5]. There are many examples of body shape divergence that have been attributed completely to adaptation to ecological circumstance [6–8], differences of function (e.g. the use of limbs for running across open ground versus clinging to rocky outcrops) [9–12], or a combination of the two [13–14]. Often, patterns of morphological variation are shaped by shared phylogenetic history [15–16], which may influence variation in function or behavior [17]. Understanding the causes of variation in body shape is important for understanding how it may impact a species’ evolutionary trajectory: for example, increased fitness from the evolution of a certain body shape may prevent divergence from that shape, whereas similarity due to shared evolutionary history may not limit future changes in morphology.
Elongation of body or limb is a specific axis of morphological variation that has long been included in a suite of traits associated with cave-dwelling species [18–24]. Cave-dwelling taxa are of particular interest because of their dramatic morphological and physiological changes, the simplicity of the selection regime within the cave habitat, and the resulting parallel evolution of cave-associated traits [25–26]. These traits, known collectively as troglomorphy [19], include other features such as regression of eyes, depigmentation, enhanced extra-optic sensory structures, and reduced metabolism [26]. Troglomorphic traits result from both a relaxation of selection pressures formerly experienced in the ancestral surface habitats, and as a result of directional selection experienced within the cave environment [27]. Though most cases of troglomorphic elongation have been studied in invertebrates due to their relative abundance (e.g. 25,19), studies of cave vertebrates, and salamanders in particular, also associate elongation with cave-dwelling [20,28–29].
In this study we examined the impact of habitat and life history on body shape and size by comparing morphological measurements among (1) aquatic and terrestrial and (2) cave-restricted and non-cave-restricted species in the salamander genus Eurycea. This group is well suited to studies of phenotypic evolution because of the extensive morphological and ecological variation represented therein: it exhibited exceptionally high rates of both size and shape evolution when compared with other plethodontids [30], and inhabits most known ecological niches available to salamanders. In addition, the independent colonization of caves and of aquatic habitats by multiple lineages of Eurycea [31] presents a natural experiment in potential ecological roles in morphological variation.
This work addresses a number of issues with our current knowledge of the evolution of body elongation as it relates to habitat occupancy: First, we analyzed the relationship between habitat and trait evolution in the context of phylogenetic relatedness. Though it is important to consider trait evolution in the context of patterns of relatedness in order to avoid bias [32–33], troglomorphic elongation has not been assessed in this group using phylogenetically based statistical methods to our knowledge. Past comparisons that find differences in shape among cave-dwelling and non-cave-dwelling populations [20,28–29] have focused mainly on the Texas clade of cave-dwelling and non-cave dwelling Eurycea, which are entirely aquatic, and none have included phylogenetic context. By studying a broader taxonomic, ecological and morphological sampling and analyzing trait differences in a phylogenetic context, we were able to compare species in a variety of habitats, providing a greater insight into the relationship between ecology, phylogeny, and morphology.
Materials and methods
Morphometric data collection
In December 2013 and February 2015 we took photographs of 405 preserved specimens representing 20 species of Eurycea in the herpetology collections of the American Museum of Natural History (New York City, New York) and the Smithsonian Institution National Museum of Natural History (Washington, D.C.). Photographs included three angles (dorsal, ventral, and lateral views), and a size standard. Because sexual size dimorphism is minor relative to individual size variance in salamanders [34], we did not attempt to collect data on sex from these specimens. We measured nine morphometric traits from these photographs using the image processing software ImageJ (NIH). These traits include: head width, forelimb length, forelimb width, body width at its widest between the forelimbs and hindlimbs, hindlimb length, hindlimb width, the length of the fourth back digit, tail length, and snout-vent length (SVL). Where the tail tip was missing or undergoing regrowth we did not measure tail length and left it as missing data. One person performed all of the digital processing to avoid among-researcher error in measurement (data can be found in S1 Table). Using information from [35], [36], and [34], we recorded whether each species is aquatic or non-aquatic and restricted to caves or not, and recorded whether each species exhibits obligate paedomorphosis, facultative paedomorphosis, or obligate metamorphosis (Table 1).
Phylogenetic reconstruction
Phylogenetic variance-covariance among species was estimated using a previously published phylogenetic reconstruction [37]. We obtained results of a BEAST species tree reconstruction including all of our sampled species, estimated the consensus tree which included clades represented in greater than 90% of trees and computed branch lengths using the package ape v5.2 [38], and estimated the variance-covariance matrix from this tree using the package geiger v2.0.6 [39]. This consensus tree was visualized using FigTree v1.4.0 (Fig 1; [40]).
Tree was obtained from [37]. Phylogenetic history was reconstructed using three mitochondrial (Co1, Cytb, and ND2) and four nuclear (BDNF, Pomc, RAG1, and Slc8a3) genes using the Bayesian software BEAST 2.4. Annotations represent habitat and life history characteristics of each species.
Statistical analysis
All statistical analyses were performed using R v3.1.2 [41] interfaced through RStudio v0.98.1091 (RStudio, Inc.). We transformed our morphological measurements using log-shape ratios [42] and first tested for evidence of phylogenetic signal influencing these morphological traits using the R package phylocurve v2.0.9 [43]. Since relatedness among species may impinge on the independence of these data, we analyzed them in a phylogenetic context. We tested for significant differences in body shape measurements using a permutational phylogenetic MANOVA with the R package RRPP v0.4.0 [44,45]. We used the Pillai statistic to test for associations between the set of log-shape ratios computed for the measured body shape variables and life history/habitat, including as independent variables cave/non-cave, aquatic/terrestrial, and paedomorphic/facultatively paedomorphic/biphasic. Data were visualized using the R package ggplot2 [46].
Phylomorphospace plot
We also performed a principal component analysis (PCA) on the transformed body shape measurements using the function prcomp() in the R package stats [41], centering and scaling the data. Studying principal components, a common strategy in morphometric analyses [37,47–50], allows us to analyze statistically uncorrelated variables, control for the effects of size and individual variation, and reduce the number of variables. We first interrogated the first three principal components for phylogenetic signal using the function phylosig() in the phytools package [51]. In order to support our statistical models with a visual interpretation of these data we produced a phylomorphospace plot of the first three principal components. This projects a phylogenetic tree into two-dimensional morphological space, which provides an intuitive way to identify clustering of discrete traits and convergent evolution. We used the function phylomorphospace() in the R package phytools [51].
Results
Phylogenetic MANOVA
Comparison with a null star phylogeny revealed that body shape exhibits significant phylogenetic signal (K = 0.268, p < 0.0001; S1 Fig). Phylogenetic MANOVA indicates significant impacts of habitat but not life history on body shape in this group. Body shape was significantly different between cave and non-cave species, as well as between aquatic and terrestrial species (Table 2A). Differences among obligately and facultatively paedomorphic and biphasic species were trending towards significance. Generally, non-cave species exhibited increased tail length, narrowing of the head, and limb reduction compared with cave species. Terrestrial species exhibited limb and digit elongation relative to body size when compared with aquatic species. Though differences were not statistically significant, paedomorphic species tended to have wider heads and more elongated bodies and limbs than biphasic species (Table 2B).
a) MANOVA test statistics comparing measurements among habitat and life history categories. b) Regression coefficients show the direction and effect size of each relationship.
Principal component analysis
Principal component analysis (PCA) was performed on transformed morphometric data from 20 species of Eurycea (Fig 2). The first three principal components accounted for 69% of the cumulative variance (Table 3). PC1, which accounted for 27% of the total variance, represented reduced head width and tail length, and shorter, fatter limbs relative to SVL. PC2 (26% of total variance) represented increased head and body width, longer tails, and shorter limbs relative to SVL. PC3 (16% of total variance) represented general elongation- an increase in SVL and tail length relative to decreased head and body width, and reduction of limb size. To ensure our interpretations of the directionality of the principal components was correct we visualized the relationships between each principal component and its strongest loading trait (S1 Fig). We found that phylogenetic relatedness significantly influences the distribution of principal component 3 (Table 4).
PC1 represents wider, shorter limbs, longer torsos relative to tail length, and narrower heads. PC2 represents wider heads and bodies and longer tails relative to SVL, and shorter limbs and digits. PC3 is representative of an increase in body and tail length, and decreases in limb size and head width.
Log-shape ratios were used to perform a principal components analysis, after centering and scaling the data. Eigenvectors are reported here, together with the proportion of variance assigned to each component. The first three principal components were used in phylomorphospace visualizations.
Σ2 estimates the rate of evolution for each PC. Phylogenetic relatedness was included in linear models for those PCs with significant phylogenetic signal.
Visualizations of morphological space
We used phylomorphospace plots to visualize the variance in each of the principal components exhibiting significant associations with habitat in our analyses (Fig 3). We see not only the segregation of biphasic/paedomorphic species along the PC1 and PC3 axes, as uncovered in the glmm results, and the segregation of cave/non-cave species along the PC3 axes, but also some interesting patterns not observed with our linear models. We observed apparent differences in variance among groups in our phylomorphospace plots, which were confirmed by the results of multiple Breusch Pagan tests for heteroscedasticity (Table 5). Most consistently, paedomorphic species had significantly more variance in their morphology than biphasic species. This is most visually apparent in the comparison of PC1 and PC2 (Fig 2), where paedomorphic species have markedly long branches and span the entire morphospace. Aquatic and terrestrial species also differed in the variance they exhibit, but inconsistently. We also see interesting morphological clustering among unrelated species, as the case of the Hydromantes in the PC1/PC2 comparison: clustering of Hydromantes brunus with H. genei and H. italicus with H. platycephalus (Fig 4) indicates shared morphological features despite closer relationships, sympatry, and shared ecological requirements of H. brunus with H. platycephalus and H. genei with H. italicus.
Clustering similar to glmm results can be observed: segregation between paedomorphic and metamorphic species along the PC1 axes, and between both paedomorphic and metamorphic species and cave and non-cave species along the PC3 axes.
We performed Brausch Pagan tests for heteroscedasticity on linear models including each principal components and our independent variables in order to examine variance among groups. Results, reported as X2(p value), indicate different ecological and life history groups exhibit more variance in their morphologies depending on the trait. Significant results, following a Holm-Bonferroni correction for multiple comparisons, are in bold.
Discussion
Our goal in these analyses was to examine the impact of habitat and life history on body shape among 26 species of Plethodontid salamanders. Our results indicate that habitat has significantly shaped the morphology of these species, while life history may have also played a role. We also found that examination of some morphological traits requires consideration of the underlying phylogenetic relationships. Specifically, we found that cave species do indeed tend to have more elongated limbs compared with non-cave species, together with shorter tails. We also found that terrestrial species exhibit elongation of limbs compared with aquatic species, and that paedomorphic species tended towards elongation of limbs and torsos and wider heads compared with biphasic species. These results at a broad taxonomic scale, though only marginally significant, reflect recent findings that within Eurycea tynerensis, a facultatively paedomorphic species found in the Ozark Plateau, paedomorphic populations have an increased number of vertebrae compared with biphasic populations [52]. Additionally, it was previously shown in the Eurycea that rates of diversification in the vertebral column was found to be dramatically greater among paedomorphic species than biphasic species, which the authors attribute to the imposition of conflicting selective constraints across ontological stages [37]. The general elongation of paedomorphic species together with the limb elongation in cave species we also observe here may contribute to the perception that cave-dwelling species tend toward general elongation [18–24], as the majority of well-known cave obligates are paedomorphic.
Variation in morphology arises through many different mechanisms including environmental influences, structural or functional constraints, or shared evolutionary history [3–5]. Some taxa show distinct differences among populations due primarily to ecological differences [10,15,53], which may be driven primarily by functional differences in how traits benefit organisms in those habitats (e.g., climbing requires different adaptations than swimming or burrowing [54]). Other taxa exhibit a combination of ecology-driven and phylogeny-driven variation among lineages [54–55]. Somewhat surprising among our observations is the minor role that phylogenetic signal plays in the body shape traits we examine when variation is reduced using principal components- we found significant phylogenetic signal in only principal component three, which represents a mere 16% of variance in these morphological traits. The first two principal components, representing 53% of total variance, exhibited no significant phylogenetic signal and this may suggest that much of body shape has evolved in these species as a result of their habitat and life history, and independent of evolutionary origin. The results of our tests for phylogenetic signal, phylogenetic MANOVA, and phylomorphospace visualizations indicate that variation body size and shape in Eurycea and similar Plethodontids is driven by a complex interaction of ecology, life history, and phylogeny.
The shape differences we observe between cave-dwelling and non-cave-dwelling species allow us to make hypotheses about the functional implications of elongation in this group. While the literature leads us to expect to observe elongation of cave species, we find here that they exhibit general reduction of the torso and head and shortened tails. While this is unexpected, it could be hypothesized that the energetic costs of the cave habitat, in which organisms face cold temperatures, high humidity, and a paucity of resources [25–26] contributes to the general body and tail reduction of cave species in this group. Energetic costs of elongation have been observed in other species [56], and in the relatively extreme cave environment those costs may have resulted in adaptive morphological change toward a more metabolically efficient shape. The increase in limb length in contrast to the reduction in body and tail size exhibited by cave-dwelling species can be hypothesized to result from ambulatory requirements of caves. Limb length has been correlated with running speed in a large study of mammalian species, though it was predicted that reducing costs of locomotion may be a stronger influence on the evolution of limb form than potential speed [57], and has also been found to differ among arboreal and terrestrial species in studies of lizards, which is hypothesized to represent trade-offs between traits benefiting running and climbing [11,58,59]. The need for salamanders to climb rocky walls and cling to crevices may drive the evolution of long limbs in cave-dwellers that we see here.
Together our results suggest that the evolution of body shape is largely influenced by life history and habitat in these Plethodontids, along with their phylogenetic relatedness to a lesser degree. Similar evolution due to ecological niche occupancy has been found in other systems, such as the convergent reductions in bone size in freshwater threespine sticklebacks [8] and repeated elongation within families of reef fish [60]. As many of these species are of conservation concern due to their endemism in miniscule ranges, it is important to note that habitat preservation in these cases is critical not only for the maintenance of species diversity but also morphological diversity.
Supporting information
S1 Fig. Correlations between principal components and shape measurements.
https://doi.org/10.1371/journal.pone.0216754.s001
(PDF)
S1 Table. Raw measurement data for all traits for each individual in this study.
https://doi.org/10.1371/journal.pone.0216754.s002
(DOCX)
Acknowledgments
We would like to acknowledge the contributions of the staff at the American Museum of Natural History, New York NY, including but not limited to David Kizirian and David Dickey as well as those at the National Museum of Natural History, Washington D.C., including but not limited to Jeremy Jacobs. Thanks also to Julia Allen, who performed all of the digital image processing, and to John G. Phillips for his invaluable feedback on the manuscript.
References
- 1.
Carroll RL. Patterns and Processes of Vertebrate Evolution (Vol. 2). Cambridge, UK: Cambridge University Press. 1997.
- 2. Collar DC, Reynaga CM, Ward AB, Mehta RS. A revised metric for quantifying body shape in vertebrates. Zoology. 2013;116:246–257. pmid:23746908
- 3. Blomberg SP, Garland T, Ives AR. Testing for phylogenetic signal in comparative data: behavioral traits are more labile. Evolution. 2003;57:717–745. pmid:12778543
- 4.
Gould SJ. The Structure of Evolutionary Theory. Cambridge, MA: Harvard University Press. 2002.
- 5. Losos JB. Seeing the Forest for the Trees: The Limitations of Phylogenies in Comparative Biology. Amer Nat. 2011;177(6):709–727.
- 6. Schluter D, Clifford E, Nemethy M, McKinnon JS. Parallel Evolution and Inheritance of Quantitative Traits. Amer Nat. 2004;163(6):809–822.
- 7. Shine R. Sexual Differences in Morphology and Niche Utilization in an Aquatic Snake, Acrochordus arafurae. Oecologia. 1986;69(2):260–267. pmid:28311369
- 8. Walker JA, Bell MA. Net evolutionary trajectories of body shape evolution within a microgeographic radiation of threespine sticklebacks (Gasterosteus aculeatus). J Zoology. 2000;252(3):293–302.
- 9. Cunha C, Bastir M, Coelho MM, Doadrio I. Body shape evolution among ploidy levels of the Squalius alburnoides hybrid complex (Teleostei, Cyprinidae). J Evol Biol. 2009;22:718–728. pmid:19320794
- 10. Kamiya T. Morphological and Ethological Adaptations of Ostracoda to Microhabitats in Zostera Beds. Dev Palaeontology Stratigraphy. 2011;11:303–318.
- 11. Losos JB. The Evolution of Form and Function: Morphology and Locomotor Performance in West Indian Anolis Lizards. Evolution. 1990;44(5):1189–1203. pmid:28563896
- 12. Walker JA. Ecological morphology of lacustrine threespine stickleback Gasterosteus aculeatus L. (Gasterosteidae) body shape. Biol J Linnean Soc. 1997;61(1):3–50.
- 13. Klingenberg CP, Barluenga M, Meyer A. Body shape variation in cichlid fishes of the Amphilophus citrinellus species complex. Biol J Linnean Soc. 2003;80:397–408.
- 14. Wikramanayake ED. Ecomorphology and Biogeography of a Tropical Stream Fish Assemblage: Evolution of Assemblage Structure. Ecology. 1990;71:1756–1764.
- 15. Álvarez A, Perez SI, Verzi DH. Ecological and phylogenetic dimensions of cranial shape diversification in South American caviomorph rodents (Rodentia: Hystricomorpha). Biol J Linnean Soc. 2013;110: 898–913.
- 16. Stayton CT. Morphological evolution of the lizard skull: A geometric morphometrics survey. J Morphology. 2005;263(1):47–59.
- 17. Bergmann PJ, Irschick DJ. Alternate pathways of body shape evolution translate into common patterns of locomotor evolution in two clades of lizards. Evolution. 2010;64:1569–1582. pmid:20050911
- 18. Brandon RA. North American troglobitic salamanders: some aspects of modification in cave habitats, with special reference to Gyrinophilus palleucus. Bull of the Nat Speleological Soc. 1971;33:1–21.
- 19. Christiansen K. Convergence and parallelism in cave Entomobryinae. Evolution. 1961;15(3):288–301.
- 20. Mitchell RW, Reddell JR. Eurycea tridentifera, a new species of troglobitic salamander from Texas and a reclassification of Typhlomolge. Texas J Sci. 1965;17(1):12–27.
- 21. Sket B. Can we agree on an ecological classification of subterranean animals? J Nat Hist. 2008;42(21–22):1549–1563.
- 22.
Weber A. Amphibia. In: Gunn J, editor. Encyclopedia of caves and karst science. Taylor and Francis; 2004. pp. 125–128.
- 23.
White WB, Culver DC. Encyclopedia of Caves. San Diego: Academic Press; 2012.
- 24.
Wilkens H, Culver DC, Humphreys WF. Subterranean ecosystems. Elsevier; 2000.
- 25. Barr T, Holsinger J. Speciation in cave faunas. Annu Rev Ecol Syst. 1985;16:313–337.
- 26.
Culver D. Cave life: Evolution and Ecology. Cambridge, MA: Harvard University Press; 1982.
- 27. Pipan T, Culver DC. Convergence and divergence in the subterranean realm: a reassessment. Biol J Linnean Soc. 2012;107(1):1–14.
- 28. Bendik NF, Meik JM, Gluesenkamp AG, Roelke CE, Chippindale PT. Biogeography, phylogeny, and morphological evolution in central Texas cave and spring salamanders. BMC Evol Biol. 2013;13(1):201.
- 29. Wiens JJ, Chippindale PT, Hillis DM. When are phylogenetic analyses misled by convergence? a case study in Texas cave salamanders. Syst Biol. 2003;52:501–514. pmid:12857641
- 30. Adams DC, Berns CM, Kozak KH, Wiens JJ. Are rates of species diversification correlated with rates of morphological evolution? Proc R Soc Lond [Biol]. 2009;rspb-2009.
- 31. Bonett RM, Steffen MA, Lambert SM, Wiens JJ, Chippindale PT. Evolution of paedomorphosis in plethodontid salamanders: ecological correlates and re-evaluation of metamorphosis. Evolution. 2014;68:466–482. pmid:24102140
- 32. Brooks DR, McLennan DA, Carpenter JM, Weller SG, Coddington JA. Syst, Eco, and Beh. BioScience. 1995;45:687–695.
- 33. Felsenstein J. Phylogenies and the Comparative Method. Amer Nat. 1985;125:1–15.
- 34.
Petranka J. Salamanders of the United States and Canada. Washington.D.C.: Smithsonian Institution Press. 1998.
- 35.
IUCN. The IUCN Red List of Threatened Species. Version 2015–1. <http://www.iucnredlist.org>. Downloaded in February, 2015.
- 36.
AmphibiaWeb: Information on amphibian biology and conservation [web application]. AmphibiaWeb, Berkeley, California. http://amphibiaweb.org. Accessed in February, 2015.
- 37. Bonett RM, Blair AL. Evidence for complex life cycle constraints on salamander body form diversification. Proc Nat Academy Sci. 2017; pmid:28851828
- 38. Paradis E., Schliep K. 2018. ape 5.0: an environment for modern phylogenetics and evolutionary analyses in R. Bioinformatics 2018;526–528.
- 39. Harmon L.J., Weir J.T., Brock C.D., Glor R.E., and Challenger W. GEIGER: investigating evolutionary radiations. Bioinformatics 2008;24:129–131. pmid:18006550
- 40.
Rambaut A, Drummond A. FigTree: Tree figure drawing tool, v1. 4.2. Institute of Evolutionary Biology, University of Edinburgh. 2012.
- 41.
R Core Team R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/. 2016.
- 42. Claude J. Log-shape ratios, Procrustes superimposition, elliptic Fourier analysis: three worked examples in R. Hystrix. 2013;94.
- 43. Goolsby E.W. Phylogenetic comparative methods for evaluating the evolutionary history of function-valued traits. Systematic Biology. 2015;64(4):568–578. pmid:25672902
- 44.
Collyer, M. L., Adams, D.C. RRPP: Linear Model Evaluation with Randomized Residuals in a Permutation Procedure.https://CRAN.R-project.org/package=RRPP. 2019.
- 45. Collyer M. L., Adams D.C. RRPP: RRPP: An R package for fitting linear models to high-dimensional data using residual randomization. Methods in Ecology and Evolution. 2018;9(2):1772–1779.
- 46.
Wickham H. ggplot2: elegant graphics for data analysis. New York:Springer. 2009.
- 47. Kuris AM, Brody MS. Use of principal components analysis to describe the snail shell resource for hermit crabs. J Exp Marine Biol Ecol. 1976;22:69–77.
- 48. Zink RM, Dittmann DL, Klicka J, Blackwell-Rago RC. Evolutionary patterns of morphometrics, allozymes, and mitochondrial DNA in Thrashers (genus Toxostoma). The Auk. 1999;116(4):1021–1038.
- 49. Polly PD, Lawing AM, Fabre A, Goswami A. Phylogenetic principal components analysis and geometric morphometrics. Hystrix, Italian J Mammology. 2013;24(1):33–41.
- 50. Ndungu NN, Kiatoko N, Ciosi M, Salifu D, Nyansera D, Masiga D, Raina SK. Identification of stingless bees (Hymenoptera: Apidae) in Kenya using morphometrics and DNA barcoding. J Apicultural Res. 2017;56(4):341–353.
- 51. Revell LJ. phytools: An R package for phylogenetic comparative biology (and other things). Methods Ecol. Evol. 2012;3:217–223.
- 52. Bonett RM, Phillips JG, Ledbetter NM, Martin SD, Lehman L. Rapid phenotypic evolution following shifts in life cycle complexity. Proc. R. Soc. B. 2018;285:20172304. pmid:29343600
- 53. Clabaut C, Bunje PME, Salzburger W, Meyer A. Geometric morphometric analyses provide evidence for the adaptive character of the Tanganyikan cichlid fish radiations. Evolution. 2007;61(3):560–578. pmid:17348920
- 54. Blankers T, Adams DC, Wiens JJ. Ecological radiation with limited morphological diversification in salamanders. J Evol Biol. 2012;25:634–646. pmid:22268991
- 55. Jockusch EL. Geographic variation and phenotypic plasticity of number of trunk vertebrae in slender salamanders, Batrachoseps (Caudata: Plethodontidae). Evolution. 1997;51(6):1966–1982. pmid:28565126
- 56. Brown JH, Lasiewski RC. Metabolism of weasels: the cost of being long and thin. Ecology. 1972;53(5):939–43.
- 57. Christiansen PER. Locomotion in terrestrial mammals: the influence of body mass, limb length and bone proportions on speed. Zool J Linnean Soc. 2002;136(4):685–714.
- 58. Arnold EN. Structural niche, limb morphology and locomotion in lacertid lizards (Squamata: Lacertidae), a preliminary survey. Bull br Mus Nat Hist Zool. 1998;64:63–89.
- 59. Sinervo B, Losos JB. Walking the tight rope: arboreal sprint performance among Sceloporus occidentalis lizard population. Ecology. 1991;72(4):1225–1233.
- 60. Claverie T, Wainwright PC. A Morphospace for Reef Fishes: Elongation Is the Dominant Axis of Body Shape Evolution. PLoS ONE 2014;9:e112732 pmid:25409027