An Inconvenient Monophyly: An Update on the Taxonomy of the Groupers (Epinephelidae) Author Ma, Ka Yan Simon F. S. Li Marine Science Laboratory, School of Life Sciences, The Chinese University of Hong Kong, Sha Tin, 999077, Hong Kong SAR. Author Craig, Matthew T. 2 National Oceanic and Atmospheric Administration, National Marine Fisheries Service, Southwest Fisheries Science Center, 8901 La Jolla Shores Drive, La Jolla, California 92037 matthew.craig@noaa.gov. text Copeia 2018 2018-10-31 106 3 443 456 http://dx.doi.org/10.1643/ci-18-055 journal article 298900 10.1643/ci-18-055 4919bfd6-3d35-48b9-9421-c852fa35abab 12536069 Cephalopholis . —As with all previous molecular phylogenies where they have been sampled, Aethaloperca , Cephalopholis , Gracila , and Paranthias form a clade and morphological data suggest that they are close allies (e.g., Randall, 1964; Smith-Vaniz et al., 1988). The basis for the exclusion of Aethaloperca , Gracila , and Paranthias from Cephalopholis in past treatments (e.g., Heemstra and Randall, 1993) was solely based on apomorphic characters. However, given that they are deeply nested within Cephalopholis , these characters become autapomorphic, and their usefulness to exclude them from Cephalopholis is rendered null. Smith (1954) elevated the sub-genus Aethaloperca for the species Cephalopholis albomarginatus and Perca rogaa . Randall (1964) removed C. albomarginatus from Aethaloperca and placed it into the new genus Gracila (discussed below). Aethaloperca rogaa was considered to be appropriately placed in a monotypic genus based on the shape of its cranial crests, its deeper body, the steeper profile of its head, the shape of its dorsal fin, and the presence of asymmetrical pectoral fins (Randall, 1964; Smith-Vaniz et al., 1988; Heemstra and Randall, 1993). While these characters may seem to indicate a generic difference, they are not all unique and some are shared with other species of Cephalopholis . For example, the presence of a steep head profile is present in C. igarashiensis and C. sonnerati ( Fig. 3 ). Its body depth (2.1–2.4 times in standard length [SL]; Heemstra and Randall, 1993) is well within the range of other species of Cephalopholis (2.0–3.2 times in SL; Heemstra and Randall, 1993; Craig et al., 2011). The remaining external characters (shape of dorsal and pectoral fins) are the only obvious morphological differences between this monotypic genus and Cephalopholis . As will be discussed below, there are several synapomorphic characters that unite Aethaloperca with Cephalopholis and there is thus little reason to exclude Aethaloperca from Cephalopholis given its nested position within the latter based on these autapomorphic specializations. Following Smith’s (1954) inclusion of Cephalopholis albomarginatus in Aethaloperca , Randall (1964) removed the species and erected the new genus Gracila . While noting that the species showed affinities to Cephalopholis , Randall (1964) highlighted differences in external morphology including a shorter head and an emarginate caudal fin. He also noted that Gracila displayed a different mode of life (semi-pelagic) from other species of Cephalopholis ( Fig. 4 ). Katayama (1974) described and placed a second species within Gracila , G. okinawae ( ¼ C. polleni ), based on the presence of a truncate tail. The head length of Gracila (2.9–3.2 in SL; Heemstra and Randall, 1993; Craig et al., 2011) is well within the range of other species of Cephalopholis (2.2–3.1 in SL; Heemstra and Randall, 1993; Craig et al., 2011). The emarginated caudal fin of G. albomarginata is emarginate only in juveniles, while in adults it is only weakly emarginated and is closer to truncate. A nearly emarginate caudal fin is also present in C. igarashiensis , and a truncate caudal fin is present in C. polleni (although in some specimens the caudal is rounded; Fig. 4 ). These three species also form a clade ( Figs. 1 , 2 ). The semi-pelagic behavior of this species is more-or-less restricted to juveniles; adults display a roving behavior but are still closely associated with the substrate. Given the sum-total of the synapomorphies which are present in this clade, both morphological (see below) and molecular, excluding Gracila from Cephalopholis is unwarranted. The most surprising of the species included in the monophyletic Cephalopholis are those of the genus Paranthias . While rare among groupers, their semi-pelagic behavior is shared with Triso and juveniles of Gracila , and their distinct morphological features such as a deeply forked tail, high gill raker counts, and small teeth, appear to be associated with a transition to the water column ( Fig. 5 ). Species of Paranthias , while clearly semi-pelagic, behave very much like other groupers in that when provided with a negative stimulus, they retreat into reef crevices rather than swim away, clearly reflecting their preference for the reef. It should also be noted that Cephalopholis fulva and Paranthias furcifer are well known to hybridize in the wild (the hybrid being originally described as the monotypic Menephorus punctiferus ; Smith, 1966; Bostrom et al., 2002), and the two species of Paranthias form the sister group to C. fulva ( Figs. 1 , 2 ). FIG. 1. Majority rule consensus cladogram of maximum likelihood analysis, reconstructed using RAxML with 1000 bootstrap replicates following Ma et al. (2016). Numbers at nodes represent bootstrap support values (%). FIG. 2. Bayesian inference phylogram reconstructed using MrBayes following Ma et al. (2016). Numbers on branches represent posterior probabilities. FIG. 3. Cephalopholis sonnerati (top) and C. rogaa (bottom). Photographs © Robert Myers. Evolutionary transitions involving radical changes to morphology within a clade are not unique to groupers. For example, the fusiliers (Caesioninae) are deeply nested within the larger snapper family ( Lutjanidae ) despite having clearly different external adaptations to mid-water habitat (Miller and Cribb, 2007). The semi-pelagic boga ( Haemulon vittatum ) was long regarded as belonging to the monotypic genus Inermia ; however, molecular evidence clearly indicates its placement within Haemulon (Rocha et al., 2008) . While certainly unexpected, the placement of the two species of Paranthias nested within Cephalopholis is clearly warranted given the morphological and molecular synapomorphies that unite this clade. Despite their seemingly different morphologies, these four genera share a number of characters, including the presence of nine dorsal fin spines (shared only with Hyporthodus acanthistius ), the ninth of which is formed indirectly from a soft ray (Kendall, 1979; Leis, 1986), and the presence of trisegmental pterygiophores in the dorsal fin (bisegmental in remaining epinephelines; Heemstra and Randall, 1993). Cephalopholis , Gracila , and Paranthias also share epineural ribs on vertebrae 1–9 ( Aethaloperca , Epinephelus , and Mycteroperca have epineurals on vertebrae 1–10; Baldwin and Johnson, 1993; Heemstra and Randall, 1993).