Revision of the generic classification of the hawkfishes (Cirrhitidae), with descriptions of three new genera. Author John E. Randall text Zootaxa 2001 12 1 12 http://www.zoobank.org/urn:lsid:zoobank.org:pub:E3F3B85F-4DB1-442C-9F05-61F90D5806F2 journal article z00012p001 [[ Family Cirrhitidae ]] The perciform fish family Cirrhitidae is characterized by having 14 pectoral rays, the lower 5 to 7 unbranched and enlarged with the membranes deeply incised; a single dorsal fin of X spines and 11 to 17 soft rays, the fin notched between the spinous and soft portions; membrane at dorsal spine tips with one to several cirri; anal rays III,5-7; 15 principal caudal rays; pelvic rays I,5; opercular spines 2; a fringe of cirri on posterior edge of anterior nostril; margin of preopercle serrate, though often only on upper half (number of serrae increase with age); scales cycloid; gill membranes broadly joined, with a free fold across throat; no swimbladder; 6 branchiostegal rays; and 26 vertebrae. Except for the presence or absence of palatine teeth and except for the reduced size of the canines in the monotypic genus Oxycirrhites , the dentition is essentially the same for all species: an outer row of small incurved canines, the largest in upper jaw at the front, and the largest in the lower jaw as 2 to 4 recurved canines nearly half way back in jaw; a broad dense inner band of villiform teeth at front of jaws, continuing progressively narrower posteriorly in upper jaw; a narrow V-shaped band of small teeth on the vomer. The Cirrhitidae is primarily an Indo-Pacific family, with only three of the 33 species occurring in the Atlantic, and only three in the eastern Pacific (two of which are also wideranging in the Indo-Pacific). Most are found in shallow water on coral reefs or rocky substrata, often in areas exposed to surge. When in a surgy area, they use their thickened lower pectoral rays to wedge themselves in place. Cyprinocirrhites polyactis and Oxycirrhites typus are exceptions in their usual occurrence in 20-100 m. Cirrhitids feed mainly on benthic crustaceans, occasionally on small fishes. C. polyactis feeds well above the substratum on zooplankton, and O. typus makes short forays into the water column for the larger animals of the zooplankton. At least some of the species (and likely all) are protogynous hermaphrodites (Sadovy & Donaldson 1995). The generic classification of the Cirrhitidae has a long and confused history which is summarized here: Guenther (1860) recognized eight genera in the family, only three of which remain in the Cirrhitidae today: Cirrhitus Lacepede (which Guenther and most early authors misspelled Cirrhites ), Cirrhitichthys Bleeker , and Oxycirrhites Bleeker . The other five genera are now classified in the families Cheilodactylidae and Chironemidae. Gill (1862) proposed the genera Cirrhitopsis for Cirrhites aureus Temminck and Schlegel from Japan and Amblycirrhitus for Cirrhites fasciatus Cuvier (1829) for which the type locality was given as Pondichery , India. Castelnau (1873) described Neocirrhites armatus as a new genus and species from northeastern Australia. Bleeker (1875) treated five genera in the family for the East Indian region, adding Amblycirrhitus Gill and Paracirrhites Bleeker . Steindachner & Doederlein (1884) described Paracirrhites japonicus as a new genus and species from Japan. Noting that their Paracirrhites was preoccupied by Paracirrhites Bleeker , Jordan in Jordan & Herre (1907) proposed the replacement name Isobuna . Jenkins (1903) described Cirrhitoidea bimacula as a new genus and species from the Hawaiian Islands, unaware that Amblycirrhitus was available for bimacula . Tanaka (1917) created the genus Cyprinocirrhites for Cirrhitichthys polyactis Bleeker , surprisingly not named earlier. Mowbray in Breder (1927) described Pseudocirrhites pinos as a new genus and species from the Isle of Pines, Cuba. Fowler (1938) proposed the new genus Acanthocirrhitus for Cirrhitus oxycephalus Bleeker . Realizing that Cirrhitus fasciatus Cuvier is preoccupied by C. fasciatus Bennett (1828) , he provided a replacement name, Amblycirrhitus indicus . Randall (2001) has shown that Cuvier’s C. fasciatus is a specimen of the Atlantic Amblycirrhitus pinos (Mowbray) , and the type locality of India for fasciatus is a locality error for the Atlantic. Schultz (1943) mistakenly relegated Cyprinocirrhites and Neocirrhites to the synonymy of Cirrhitichthys . He described Hughichthys as a new genus for Cirrhites melanotus Guenther , later shown to be a synonym of Neocirrhites armatus Castelnau . He then placed Cirrhitoidea in the synonymy of Paracirrhites and classified bimacula as a species of Paracirrhites . He described hubbardi as a new species in the genus Amblycirrhitus . Watanabe (1949) described Serranocirrhitus latus as a new genus and species of cirrhitid fish from Okinawa. In a review of western Indian Ocean Cirrhitidae , Smith (1951) created the genus Cirrhitops for Cirrhites fasciatus Bennett . He mistakenly listed Amblycirrhitus hubbardi Schultz as a synonym of fasciatus . He also described Gymnocirrhites as a new genus, with Cirrhites arcatus Cuvier as the type species, adding that it is allied to Paracirrhites but distinct on the basis of lacking scales on the gill membrane and on the snout before the nostrils. Schultz in Schultz & collaborators (1960) recognized 13 genera in the Cirrhitidae , including Isobuna , Serranocirrhitus , Cirrhitoidea , Hughichthys (though indicating that Neocirrhites “appears to be related”), and Gymnocirrhites . He obviously had some doubt of the validity of Gymnocirrhites , stating that Smith (1951) “attaches too much value to the scaly gill membranes.” He noted that a specimen of G. arcatus 48 mm in SL lacks scales on the gill membranes, but a specimen of 100 mm SL has them. He left bimacula in the genus Cirrhitoidea and described a new species, sexfasciata , in this genus. Randall (1963) reviewed the family, recognizing 10 genera and 34 species. He discussed Serranocirrhitus in a footnote, noting the following noncirrhitid characters: all the pectoral rays unbranched and none thickened, no teeth on the vomer, ctenoid scales, and the configuration of a pomacentrid. He included Isobuna but stated that it differs from other cirrhitids in having 3 spines on the opercle instead of 2, ctenoid scales, and only 2 rows of scales above the lateral line. He recognized Neocirrhites as a valid genus with Hughichthys a synonym, placed Gymnocirrhites in the synonymy of Paracirrhites , described Isocirrhitus as a new genus for Schultz’ sexfasciata , put hubbardi in Cirrhitops , referred Cirrhitopsis to the synonymy of Cirrhitichthys , and Pseudocirrhites to Amblycirrhitus . Randall & Heemstra (1978) reclassified the genera Serranocirrhitus and Isobuna in the subfamily Anthiinae of the Serranidae. That these two genera should have been placed in the Cirrhitidae is not as serious an error as it might seem. Jordan & Evermann (1898) wrote with respect to the cirrhitoid fishes, “This family should apparently be placed among the Percoidea near the Serranidae....This group agrees with the Percoidea in most respects, the chief external difference lying in the form of the pectorals,...” They quoted Guenther ’s description of the skeleton of Paracirrhites forsteri , adding, “Dr. Boulenger finds that the skeleton has much in common (with the Serranidae)”. In 1969 the author collected 32 specimens of an undescribed species of cirrhitid from Easter Island that appeared to represent a new genus. The following year he collected two more specimens from Pitcairn Island. Unaware of this material, Lavenberg & Yanez (1972) described the species as Cirrhitus wilhelmi from one specimen from Easter Island. Kunio Amaoka requested specimens of the various genera of cirrhitid fishes from the Bishop Museum for a graduate student who planned a phylogenetic study of the family. Specimens were sent, including ones from Easter Island. The student did not complete the study, and the specimens were recently returned by Dr. Amaoka. The present study has confirmed that the Easter Island - Pitcairn specimens of wilhelmi represent a new genus. The investigation of other cirrhitids currently placed in Cirrhitus has revealed two more species that clearly require placement in new monotypic genera. The three new genera are described below after the revised generic key of the family.