Mosaic patterns of homoplasy accompany the parallel evolution of suspensory adaptations in the forelimb of tree sloths (Folivora: Xenarthra)

Abstract

We examine how derived functional signal and phylogenetic inheritance interact in the forelimb of tree sloths, to understand the relative contribution of each in the evolution of a novel morphobehavioural suite. Molecular and craniodental data demonstrate that extant tree sloths evolved suspensory behaviours and associated morphologies from a non-suspensory ancestor independently of one another, making them a useful model system. We find that convergence in univariate traits is expressed mosaically, although the signal is largely functional. Three-dimensional analyses suggest there is greater conservatism of gross morphology in more proximal bones than in more distal elements. Convergence in some univariate scapular traits is independent of the gross morphology of the scapula itself, demonstrating that functionally relevant morphologies were mapped on to a more conserved scapular shape. Our results suggest that morphological homoplasy is expressed in a mosaic manner. The relationship between homoplasy and trait integration may be more nuanced than previously thought, even within a single adaptive system.

Integrative Approach Uncovers New Patterns of Ecomorphological Convergence in Slow Arboreal Xenarthrans

“…The suite of investigated functional features may crucially bias the detection of convergence, since this evolutionary process potentially follows mosaic patterns. Indeed, only some traits might converge even within the same anatomical element (Spear and Williams 2020). Moreover, different scales of anatomical detail of the same element, relying on different structures and mechanisms, could evolve through different convergence patterns (Watanabe et al 2020).…”

Challenges and advances in measuring phenotypic convergence

“…Clade comparisons are enabled by 1) excluding pairwise comparisons between within-clade lineages (e.g., two flying squirrel species) and 2) weighting of Ct scores and p -values based on the number of pairwise comparisons between focal clades (see Supplemental Methods). Further, Ct -measures can be measured using single traits ( C -measures only permitted measures of multivariate distances, although they were adapted for univariate analyses in some studies; Spear and Williams 2020, Law 2022), and we updated the C 4 (now Ct 4) calculation to better match the original description of that measure. See the Supplemental Methods for additional information on these updates.…”

“…By default, the original C -measures do not support investigation of convergence in a single trait (although see Spear and Williams, 2020; Law, 2022). To circumvent this limitation we have added code to the convrat .…”

“…Measuring convergence of single traits. By default, the original C -measures do not support investigation of convergence in a single trait (although see Spear and Williams, 2020; Law, 2022). To circumvent this limitation we have added code to the convrat.t function which appends an invariant trait (with value zero) to datasets consisting of a single trait.…”

“…Novel methods for identifying and measuring convergence have recently been developed (Mahler et al 2013, Arbuckle et al 2014, Ingram and Mahler 2013, Stayton 2015A, Speed and Arbuckle 2017, Castiglione et al 2019), and these methods are often accompanied by statistical tests for comparing the measured convergence to that which is expected from a null model-fitting hypothesis or random data permutations. This has increased the accessibility of quantitative tests for phenotypic convergence, leading to a flood of recent studies on that topic (e.g., Friedman et al 2016, Zelditch et al 2017, Baliga and Mehta 2018, Da Silva et al 2018, Arbour and Zanno 2020, Button and Zanno 2020, Grossnickle et al 2020, Martinez et al 2020, Serio et al 2020, Rincon-Sandoval et al 2020, Spear and Williams 2020, Baumgart et al 2021, Huie et al 2021, Rovinsky et al 2021, Tamagnini et al 2021, Alfieri et al 2022, Bennion et al 2022, Canale et al 2022, Law 2022).…”