Epinephelus, Bloch, 1793

Ma, Ka Yan & Craig, Matthew T., 2018, An Inconvenient Monophyly: An Update on the Taxonomy of the Groupers (Epinephelidae), Copeia 106 (3), pp. 443-456 : 448-449

Epinephelus .

—The inclusion of the monotypic Anyperodon and Cromileptes within Epinephelus is somewhat surprising although consistent with all previous molecular studies. Anyperodon shares 11 dorsal-fin spines and the absence of trisegmental pterygiophores with Epinephelus . However, it has been retained in a monotypic genus due to the absence of palatine teeth and its ‘‘...elongate and markedly compressed head and body...’’ (Heemstra and Randall, 1993). Heemstra and Randall (1993) noted that ‘‘...there are some other elongate groupers, but none of these are as compressed Anyperodon .’’ ( Fig. 6 ). This is a curious statement as the body width for Anyperodon is 2.3–2.8 times in its depth, while the range of body width for species of Epinephelus is 1.8–2.8 in their depth (Heemstra and Randall, 1993; Craig et al., 2011). Additionally, the body depth in Anyperodon is contained 3.1– 3.7 times in SL, while that for species of Epinephelus ranges from 2.3–3.7 in SL (Heemstra and Randall, 1993; Craig et al., 2011). Finally, the head length of Anyperodon is contained 2.3–2.5 in SL, while in species of Epinephelus the head length is contained 2.1–2.8 times in SL (Heemstra and Randall, 1993; Craig et al., 2011). Thus it appears that Anyperodon is no more elongated or compressed than any other species of Epinephelus . The absence of palatine teeth is thus a curious autapomorphic specialization in the species that is worthy of future investigation; however, it is insufficient to exclude it from Epinephelus .

Cromileptes has a head shape (severely depressed anteriorly and elevated posteriorly) that is unique among groupers ( Fig. 7 ). It shares epineural ribs on vertebrae 1–10, two supraneural bones, and no trisegmental pterygiophores with Epinephelus . Despite its curious cranial morphology, there is no indication that this species is anything other than a highly specialized member of the genus Epinephelus . Similar cases of extreme morphological specializations emerging within a clade are not uncommon in fishes. For example, species of the genus Gomphosus possess a remarkably elongated snout and are nested deeply within the genus Thalassoma (Bernardi et al., 2002) . Similar to Paranthias x Cephalopholis , Gomphosus varius is known to hybridize with Thalassoma lunare (Randall and Allen, 2004) . In captivity, C. altivelis has successfully been hybridized with Epinephelus lanceolatus ; however, it is unknown if this would ever happen in the wild.

The large majority of species of Epinephelus are recovered in two sister clades and form the crown of the grouper tree (clades E and F, Figs. 1 , 2 ). Morphological differentiation among these two sister clades is lacking; however, they do exhibit subtle differences in ecology. In clade E, the majority of the species are relatively large (maximum total length 75 to 250 cm), and most are restricted to coral reefs. On the other hand, Clade F contains species that are relatively small, having maximum total length,75 cm (Heemstra and Randall, 1993; Craig et al., 2011), and most are widely distributed, inhabiting both coral and non-coral reef environments. Ecological factors have likely played a major role in driving the divergence and radiation of these two species-rich clades of groupers as 88% of cladogenic events have been shown to be non-allopatric (Ma et al., 2016). It is also important to note that the ‘‘reticulated-grouper complex’’ (Heemstra and Randall, 1993) is polyphyletic. The nine reticulated-grouper species are scattered across three distinct clades along with ‘‘non-reticulated’’ groupers.

Figures

FIG. 6. Epinephelus goreensis (top; photograph by Klaus Fiedler) and E. leucogrammicus (bottom; photograph © Robert Myers).

FIG. 7. Epinephelus altivelis juvenile (top) and adult (bottom). Photographs © Robert Myers.

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.