Syst. Biol. 00(0):1–18, 2025 © The Author(s) 2025. Published by Oxford University Press on behalf of the Society of Systematic Biologists. All rights reserved. For commercial re-use, please contact reprints@oup.com for reprints and translation rights for reprints. All other permissions can be obtained through our RightsLink service via the Permissions link on the article page on our site-for further information please contact journals.permissions@oup.com https://doi.org/10.1093/sysbio/syaf058 Advance Access publication August 27, 2025 New Perspectives in Phylogenetic Support Assessment: Using the Relative Contradiction Index to Investigate the Phylogenetic Controversies in Crocodylia Paul Aubier * , Valentin Rineau , Jorge Cubo , Stéphane Jouve Centre de Recherche en Paléontologie—Paris (CR2P, UMR 7207), Sorbonne Université, Muséum National d’Histoire Naturelle, CNRS, Paris 75005, France * Correspondence to be sent to: Centre de Recherche en Paléontologie—Paris (CR2P, UMR 7207), Sorbonne Université, Muséum National d’Histoire Naturelle, CNRS, 4 place Jussieu, Paris 75005, France; E-mail: paul.aubier@gmail.com Received 5 November 2023; reviews returned 26 July 2025; accepted 4 August 2025 Associate Editor: Benoit Dayrat Abstract.— Numerous tools have been developed since the advent of phylogenetic methods to assess tree robustness. Iden- tifying the degree of contradiction in a phylogenetic matrix, as well as the specific contribution of each taxon and character, is essential for estimating its reliability. In parsimony-based phylogenetic inferences, classically used by paleontologists, a phylogeny results from the interaction of all the characters used in the analysis. Consequently, the support initially provided by the characters in the matrix may differ from that after optimization in the final tree, severing the link between the phylo- genetic content of the matrix and that of the final tree. Thus, all methods aimed at measuring support only do so indirectly and the impact of individual characters or taxa can only be assessed after the analysis. Three-taxon analysis (3ta) is a phy- logenetic method that can circumvent these issues by precisely measuring the support of targeted characters and/or taxa directly from the phylogenetic matrix. In 3ta, characters are coded as trees and decomposed into three-taxon statements (3ts). The analysis searches for the largest set of non-contradicting 3ts to compute the optimal classification. Because the analysis is a compatibility procedure, not an optimization procedure, character supports on the tree are independent from one another. This enables direct assessment of support from the matrix, providing meaningful insights into the topology of the optimal trees. Moreover, the decomposition of characters into 3ts allows for precise quantification of the impact of the characters/taxa in the results. In this study, focusing on Crocodylia (a subject of ongoing debate over recent decades), we use 3ta to measure the support of specific characters and/or taxa in the recently published matrix of Rio and Mannion (Rio J.P., Mannion P.D. 2021. Phylogenetic analysis of a new morphological dataset elucidates the evolutionary history of Crocodylia and resolves the long-standing gharial problem. PeerJ 9:e12094). This conflict revolves around two competing hypotheses—Longirostres and Brevirostres—supporting a different placement of the Gavialoidea clade. We also introduce here the Relative Contradiction Index (RCI) to evaluate node support, a metric that reflects the degree of contradiction in a matrix between competing cladistic hypotheses, ranging from 0.5 (maximum contradiction) to 1 (no contradiction). We show that although the Longirostres hypothesis is best-supported, it is strongly challenged by the Brevirostres hypothesis (RCI = 0.62). Furthermore, we find that Tomistominae provides 61% of the supporting evidence for the Longirostres hypothesis, such that, when removed, the matrix supports the Brevirostres hypothesis. Individual tomistomines’ contributions vary only from 2% to 7% of the total support to the Longirostres hypothesis. Finally, we show that characters correlated to longirostry only provide a fraction (22%) of the total support to the Longirostres hypothesis. Thus, our method can quantify the impact of specific characters or taxa on a phylogenetic result. This should prove very useful to phylogeneticists, especially when dealing with incomplete material such as fossils. [Crocodylia; phylogenetics; support; three-taxon analysis.] Extant biodiversity represents only a tiny fraction of the total amount of species that ever existed. In addition, only a small proportion of the extinct species have been preserved in the fossil record. As a result, paleontolog- ical data are essential for reconstructing the long-term history of life on Earth. In this context, one of the ma- jor challenges lies in effectively handling fossil data that are inherently incomplete. Phylogenetic reconstruction of extinct taxa plays a fundamental role in understand- ing the history and diversification of life on Earth. Phy- logenetic methods, however, are designed to produce trees that display relationships between taxa regardless of the input data. As a result, any particular tree is inher- ently incapable of revealing how informative or reliable the underlying data actually are. For this reason, phylo- genies are routinely accompanied by support measures that aim to assess the degree of confidence we can place in the inferred relationships. In parsimony, the simplest estimate of grouping support is the branch length (i.e., the number of transformation steps; Bremer 1994). How- ever, not all steps contribute equally to support. For ex- ample, homologous character states generally provide stronger support than homoplastic ones (Bremer 1994), so relying solely on branch lengths can be misleading. To address this issue, Bremer (1988, 1994) proposed the Branch Support (BS) index, which is based on the num- ber of extra steps required to collapse a branch, rather than on its raw length. Other support indices, such as the Bootstrap (Felsenstein 1985) and Jackknife (Farris et al. 1996), are based on disrupting the data set. The Boot- strap randomly overweights or suppresses characters, whereas the Jackknife randomly deletes a subset of char- acters. Despite their differences, BS, Bootstrap, and Jack- knife all measure global support. Although they assign 1 Downloaded from https://academic.oup.com/sysbio/advance-article/doi/10.1093/sysbio/syaf058/8242158 by SSB Member Access user on 03 October 2025