JOHNSON: PERCOIDEI 



465 



with W. D. Anderson). Callanthiids share a flat nasal organ 

 without laminae, a lateral line that runs along the base of the 

 dorsal fin, ending near its terminus or continuing along the 

 dorsolateral margin of the caudal peduncle, and a midlaleral 

 row of modified scales that bear a series of pits and/or grooves. 

 The larvae of these two genera appear dissimilar (Fig. 255E, F), 

 but specimens of Grammatonotus smaller than 1 3 mm are un- 

 known. Stiginatonolus (based on a small, now lost specimen) 

 was reported to have three opercular spines, and probably rep- 

 resents a larval or juvenile anthiine serranid. The family Gram- 

 matidae, as considered here, contains only Gramma and Li- 

 pogramma. 



Carangidae, Coryphaenidae, Echeneididae, Rachycentridae and 

 Nematistiidae. — See discussion on utility of larval morphology. 



Coracinidae, Drepanidae and Ephippididae. — The family 

 Ephippididae, as defined here, contains the following genera: 

 Chaetodipterus. Ephippus. Parapsetttis. Platax. Proteracanthus. 

 Rhinoprenes and Tripterodon. Ephippidids exhibit considerable 

 diversity in several features that are more commonly conser- 

 vative among percoids, such as scale morphology and the struc- 

 ture and arrangement of median fin supports and predorsal 

 bones. Nonetheless, monophyly of the family is supported by 

 shared specializations of the gill arches that include reduction 

 or absence of the basihyal, absence of the interarcual cartilage, 

 a relatively large first pharynogobranchial and, most notably, a 

 peculiar comblike series of large blunt rakers loosely associated 

 with the anterior margin of the broadened first epibranchial. 

 Springer (1982; pers. comm.), following some previous authors 

 (Jordan, 1923; Golvan, 1965) included Parapsettus in the Scor- 

 pididae. Rhinoprenes was previously treated as a monotypic 

 family, possibly related to the Scatophagidae (Munro, 1967), 

 and Proteracanthus as a girellid (Norman, 1966). Although 

 Drepane may be related to the ephippidids, it does not share 

 the branchial specializations described above, and lacking fur- 

 ther evidence of a direct relationship, I treat it separately. Based 

 on other features of the gill arches a close relationship between 

 Drepane and Coraciniis seems likely. In both genera the basihyal 

 is embedded in thick connective tissue and is tightly bound 

 along the anteroventrally sloping median junction of the hy- 

 pohyals. In addition, an unusual moveable articulation between 

 the hypohyals and the anterior ceratohyal allows for dorsoven- 

 tral rotation of the ceratohyal. Pending further investigation of 

 the possible affinities of these two genera, I retain them as mono- 

 typic families. Larval morphology could provide important in- 

 formation in resolving the relationships among the five ephip- 

 pidid genera, Drepane and Coraciniis, but to date, only the 

 larvae of Chaetodipterus have been described (Fig. 256G). 



Elassoma— In an extensive comparison of the acoustico-lat- 

 eralis system of the Centrarchidae, Branson and Moore (1962) 

 placed the pygmy sunfishes, genus Elassoma. in a separate fam- 

 ily, based on over 20 "major characteristics." These include 

 numerous reductions in the laterosensor>' system (e.g., absence 

 of a lateral-line canal on the body, absence of all infraorbitals 

 except the lacrimal, absence of the mandibular and angular 

 lateralis canals, etc.), presence of numerous free neuromasts of 

 a distinctive form, rudimentary olfactory organ, gill membranes 

 broadly united across the isthmus, rounded caudal fin, and cy- 

 cloid scales. To these, I add the following reductive features of 

 Elassoma, not shared by the Centrarchidae: basisphenoid ab- 

 sent; endopterygoid absent; ectopterygoid absent or fused to 



Table 1 1 9. List of the Families and incertae sedis Genera of the 

 Suborder Percoidei. * Families with a single genus. 



Acanthoclinidae 



Acropomatidae 



Ambassidae 



Aplodactylidae 



Apogonidae 



Arripididae* 



Banjosidae* 



Balhyclupeidae* 



Bramidae 



Caesionidae 



Caesioscorpis 



Callanlhiidae 



Carangidae 



Caristiidae* 



Cenlracanthidae 



Centrarchidae 



Cenlrogenysidae* 



Centropomidae 



Cepolidae 



Chaetodontidae 



Cheilodactylidae 



Chironemidae 



Cirrhitidae 



Congrogadidae 



Coracinidae* 



Coryphaenidae* 



Datnioides 



Dinolestidae* 



Dmoperca 



Drepanidae* 



Echeneididae 



Emmelichthyidae 



Enoplosidae* 



Ephippididae 



Epigonidae 



Gerreidae 



Giganthiidae* 



Girellidae 



Glaucosomatidae* 



Grammatidae 



Haemulidae 



Hapalogenys 



Hemiliiljanus 



Howella 



Inermiidae 



Kuhliidae* 



Kyphosidae 



Lactariidae* 



Lateolahrax 



Latrididae 



Leiognathidae 



Leplobramidae* 



Lethrinidae 



Lobotidae* 



Lutjanidae 



Malacanthidae 



Menidae* 



Microcanthidae 



Monodactylidae* 



Moronidae 



Mullidae 



Nandidae 



Nematistiidae* 



Nemipteridae 



Neoscorpis 



Opistognathidae 



Oplegnathidae* 



Ostracoberycidae* 



Parascorpididae* 



Pempherididae 



Pentacerotidae 



Percichthyidae 



Percidae 



Plesiopidae 



Pomacanthidae 



Pomatomidae* 



Polypnon 



Priacanthidae 



Pseudochromidae 



Rachycentridae* 



Scatophagidae 



Sciaenidae 



Scombropidae* 



Scorpididae 



Serranidae 



Sillaginidae 



Simperca 



Sparidae 



Stereolepis 



Symphysanodon 



Terapondiae 



Toxotidae* 



palatine; palatine with a single notch-like articulation with eth- 

 moid cartilage; predorsals usually absent, a single bone present 

 in some (vs. 3-7 in centrarchids); branchiostegals 5 (vs. 6-7); 

 principal caudal rays 6-7 -I- 7-8 (vs. 9 + 8); hypurals 1-2 and 

 3-4-5 fused. 



Branson and Moore (1962) concluded that "either the elas- 

 somids diverged from the centrarchid stock early in the history 

 of the group or they have entirely different affinities." Subse- 

 quent classifications (Greenwood et al., 1966; Nelson, 1976) 

 have continued to treat Elassoma as a subfamily of the Cen- 

 trarchidae, presumably accepting the conclusion of Eaton (1953, 

 1956) that Elassoma is a neotenous centrarchid, with most of 

 its distinctive features having arisen through paedomorphosis. 

 Weitzman and Fink (1983) attributed similar reductions in the 

 laterosensory system of small characids to paedomorphosis and 

 suggested that these characters may be quite labile. These and 

 other osteological reductions similar to those of Elassoma are 

 found in other small fishes such as gobioids (Springer, 1983) 

 and cyprinodontoids (Parenti, 1981), but I know of no such 

 extreme examples among small percoids. 



That the reductive specializations of Elassoma actually rep- 

 resent character states of earlier developmental stages of cen- 

 trarchids has never been clearly demonstrated or even ade- 

 quately investigated, and comparative studies of the osteological 

 development of these fishes would be necessary to answer this 

 question. However, a crucial point, that seems to have been 

 overlooked, is the absence of any other evidence suggesting a 

 close relationship between Elassoma and the Centrarchidae. 

 Although I know of no morphological specialization that defines 

 the family, all centrarchids exhibit a similar mode of nest-build- 



