Building a Body Shape Morphospace of Teleostean Fishes

Abstract

We present a dataset that quantifies body shape in three dimensions across the teleost phylogeny. Built by a team of researchers measuring easy-to-identify, functionally relevant traits on specimens at the Smithsonian National Museum of Natural History it contains data on 16,609 specimens from 6144 species across 394 families. Using phylogenetic comparative methods to analyze the dataset we describe the teleostean body shape morphospace and identify families with extraordinary rates of morphological evolution. Using log shape ratios, our preferred method of body-size correction, revealed that fish width is the primary axis of morphological evolution across teleosts, describing a continuum from narrow-bodied laterally compressed flatfishes to wide-bodied dorsoventrally flattened anglerfishes. Elongation is the secondary axis of morphological variation and occurs within the more narrow-bodied forms. This result highlights the importance of collecting shape on three dimensions when working across teleosts. Our analyses also uncovered the fastest rates of shape evolution within a clade formed by notothenioids and scorpaeniforms, which primarily thrive in cold waters and/or have benthic habits, along with freshwater elephantfishes, which as their name suggests, have a novel head and body shape. This unprecedented dataset of teleostean body shapes will enable the investigation of the factors that regulate shape diversification. Biomechanical principles, which relate body shape to performance and ecology, are one promising avenue for future research.

Feeding ecology has a stronger evolutionary influence on functional morphology than on body mass in mammals

“…The first method involves standardizing the linear jaw measurements by transforming them to log-shape ratios (Mosimann 1970). The log-shape ratios are calculated by dividing raw measurements by the geometric mean of all 12 linear measurements (a proxy for jaw size), and then log 10 -transforming the resulting ratio (Claude 2013;Price et al 2019). A benefit of this method is that it preserves variation associated with allometry.…”

“…For this method, the JAPr angle was treated in the same manner as linear measurements. Because the PGLS regression model is fit while simultaneously estimating Pagel's lambda via maximum likelihood (Revell 2010), a benefit of this method is that it incorporates phylogenetic relationships and phylogenetic signal into the size correction process (Revell 2009;Price et al 2019). However, drawbacks include the removal of variation associated with allometry (Mosimann 1970) and the elimination of jaw length as a trait that can be used in subsequent analyses.…”

Effects of taphonomic deformation on geometric morphometric analysis of fossils: a study using the dicynodontDiictodon feliceps(Therapsida, Anomodontia)

“…The discriminatory power of geometric morphometrics allows for fine-scale resolution of shape differences between organisms or their parts. Although initially developed in the study of extant organisms, geometric morphometrics has also been used extensively to study fossil taxa, including representatives of all major vertebrate groups (Botha & Angielczyk, 2007;Deeming & Mayr, 2018;Pérez-Ben, Báez & Schoch, 2019;Price et al, 2019;Felice et al, 2019). These techniques have been used for biomechanical modeling (Pierce, Angielczyk & Rayfield, 2008;Polly et al, 2016) and to quantify the evolution of morphological disparity (Brusatte et al, 2012;Lungmus & Angielczyk, 2019), evolutionary rates (Adams, 2014), and ecological adaptions (Grossnickle & Newham, 2016).…”

Regional‐scale aquifer hydrogeology as a driver of phylogeographic structure in the Neotropical catfish Rhamdia guatemalensis (Siluriformes: Heptapteridae) from cenotes of the Yucatán Peninsula, Mexico

“…Subsequently, a regression of shape on the natural logarithm of centroid size was carried out to account for variation in body morphology due to size, or allometry. Residuals of the regression were then used as allometrically adjusted data (e.g.Arroyave et al., 2019; Price et al., 2019). Lastly, shape data were averaged by cenote, and primary axes of variation were determined with principal component analysis (PCA) with the plotTangentSpace function in geomorph .…”

“…The resulting residuals were then averaged by cenote and used as allometrically adjusted trait data (e.g. Price et al., 2019). The second method for size adjustments consisted of dividing each linear variable by either standard length or head length, the latter for smaller measurements restricted to the cranial region.…”