Surveys and analyses of anatomical characters have allowed researchers to describe a wealth of anatomical features and contribute to our evolutionary understanding of fishes for centuries. However, most of these studies have focused on specific lineages or families rather than the broader evolutionary relationships. As such, there has been a lack of progress inferring higher-level relationships among percomorphs. With the use of large-scale DNA-based methods in multiple studies over the past two decades, the backbone of the phylogeny of fishes is becoming increasingly understood. Taking this DNA-based phylogenetic backbone into account, we have the opportunity to integrate discrete morphological characters and DNA sequence data to test earlier topologies and provide new and improved hypotheses of relationships. The carangiform fishes, which include approximately 1,100 species in 29–34 families, were initially recovered as a clade in DNA-based studies. Subsequent to its initial recovery, many molecular phylogenies have been published assessing carangiform relationships, but these studies present a conflicting array of hypotheses on the intrarelationships of this clade. In addition to this diversity of hypotheses, no studies have explicitly diagnosed the clade or its major subgroups from a morphological perspective or conducted a simultaneous analysis to put forth synapomorphies for relationships across the Carangiformes using a combination of molecular and morphological data. In this study, we performed combined analyses of new and previously identified discrete morphological characters and new and previously published genome-scale data to characterize the evolutionary history and anatomical variation within this clade of fishes. Our novel morphological dataset included 201 hard and soft tissue characters, and it was combined with a novel dataset of 463 ultraconserved element loci. Our combined analysis of these data resulted in a monophyletic Carangiformes, with a series of subclades nested within. We put forth a series of subordinal names based on the recovered branching pattern, morphological character evidence, and relative stability in large-scale studies. These suborders are the Centropomoidei, which includes Centropomidae, Lactariidae, Latidae, and Sphyraenidae; Polynemoidei, which includes Polynemidae and the infraorder Pleuronectoideo; Toxotoidei, which includes Leptobramidae and Toxotidae; Nematistioidei, which includes Nematistiidae; and Menoidei, which includes Menidae and Xiphioidea. Furthermore, we highlight and discuss morphological characters that support the relationships between two or more lineages of carangiform fishes. Finally, we highlight patterns of morphological convergence among some carangiform fishes and their previously hypothesized sister lineages.
FOR centuries, studies on the evolution of fishes were based on surveys and analyses of anatomical characters. Comparisons among wet and dry skeletons (e.g., Olney et al., 1993; Holcroft and Wiley, 2015), and surveys of characters through different visualization techniques, such as scanning electron microscopy and histology (e.g., Webb, 1989a; Ghedotti et al., 2018), x-ray computed tomography (e.g., Schaefer, 2003; Webb et al., 2006; Schwarzhans et al., 2018), and magnetic resonance imaging (e.g., Chakrabarty et al., 2011; Graham et al., 2014), have helped identify a wealth of anatomical features that have facilitated our interpretation of fish evolution (e.g., Potthoff et al., 1986; Springer and Johnson, 2004; Hilton et al., 2015). These techniques have helped us discover, differentiate, and assess the homology and phylogenetic significance of particular anatomical features (e.g., Johnson, 1975; Gemballa and Britz, 1998), were critical for identifying characters that suggested novel placements of taxa within the broader phylogeny of fishes (e.g., Rosen and Parenti, 1981; Johnson and Patterson, 1993; Stiassny, 1993), allowed researchers to assess the intrarelationships of lineages of fishes hypothesized to be closely related (e.g., Parenti, 1981; Baldwin and Johnson, 1996; Harold and Weitzman, 1996), or aided the search for the sister group of well-established clades (e.g., Gill and Mooi, 1993; Johnson and Brothers, 1993). Despite this breadth of studies, relatively few explicit anatomical studies have focused on the broader evolutionary relationships of fishes (exceptions include: Johnson and Patterson, 1993; Patterson and Johnson, 1995; Springer and Orrell, 2004), at least compared to the large number of broad-scale DNA-based studies (e.g., Chen et al., 2003; Miya et al., 2003; Smith and Wheeler, 2006; Near et al., 2012; Betancur-R. et al., 2013a; Smith et al., 2016). The limited taxonomic scope common with anatomical phylogenetic studies likely results from the difficult and time-consuming effort needed to examine and distinguish homologous morphological characters across a wide diversity of taxa. The overwhelming diversity of taxa and striking anatomical convergences among the perch-like fishes (Percomorpha), in particular, has played a substantial role in our delayed inference of the phylogenetic relationships of the clade (Johnson, 1984, 1993; Smith, 2010). Two of the most important papers in the systematics of percomorph fishes are Rosen (1973) and Johnson and Patterson (1993), who spent more of their writing delimiting the Percomorpha rather than resolving relationships within the species group.