© Zoological Institute, St.Petersburg, 2002

A remarkable new genus of Carangidae fron1 the Upper Paleocene of Turkmenistan (Osteichthyes: Perciformes)

A.M. Prokofiev

Prokofiev,A.M. 2002. A remarkable new genus ofCarangidae fromthe Upper Paleocene of Turkmenistan (Osteichthyes: Perciformes).Zoosystematica Rossica, 11(1): 219-228.

A new genus and species of Carangidae, Uylyaichthys eugeniae gen. et sp. n., is de­ scribed from the Upper Paleocene (middle part of the Danata Formation) of Turkmenistan. It belongs to the subfamily Trachinotinae s. str. and differs from all known members of the family in the combination of low vertebral count (21 or possibly 22), parasphenoid strongly curved mesially, and unique predorsal formula0/0+0+0/l/. Uylyaichthys is con­ sidered to be a sister taxon for Trachinotus + Paratrachinotus, and Trachinotinae s. str. regarded as a sister group of Scomberoidinae. The genus Lichia is excluded from Trachinotinae s. str. and considered to be a sister taxon ofScomberoidinae + Trachinotinae s. str.

A.M. Prokofiev, Department of Fishes and Fish-like Vertebrates, Palaeontological Insti­ tute, Russian Academy ofSciences, Profwyuznaya ul. 123, Moscow 117997, Russia.

Introduction Material and methods

Daniltshenko ( 1968) described the Upper Paleo­ Specimens described herein were collected by cene ichthyofauna of Uylya-Kushlyuk locality Dr. E.K. Sytchevskaya from the middle part of (Turkmenistan) and established 20 species. Danata Formation in Uylya-Kushlyuk locality. Among them, Daniltshenko described two forms, Uylya-Kushlyuk locality is situated 2 km NE of which he placed into the family Carangidae (Ar­ Uylya-Kushlyuk village, SW Turkmenistan, on chaeus oblongus and Trachicaranxtersus). Later, the western slope of the Kopetdagh Mountains Trachicaranx was transferredto the familyApo­ (38°38'N, 55°48'E). Danata Formation is com­ lectidae (Bannikov & F edotov, 1984), while Ar­ posed of the Upper Paleocene to the Middle chaeus oblongus now represents the earliest spe­ Eocene (Grossheim & Korobkov, 1975; Solun, cies undoubtedly belonging to Carangidae. 1975); fish-bearing layer of mottled clays lies at Carangid fishesare a widely distributed and com­ the base of the middle part of the Danata Forma­ mon group of Percoidei, which is relatively abun­ tion and has total thickness of 9 m. This layer is dant in fossils since Middle Eocene. now dated as terminal Upper Paleocene (Upper Archaeus is suggested to be one of the most Tanetien) (Muzylev, 1994). generalised genera of the Carangidae, but not The following fossil and recent (alcohol preserved, ancestral to other representatives of the family radiographed and partially dissected) species of Caran­ (Bannikov, 1990). Therefore, the divergent evo­ gidae and other percoid families from the collections of Palaeontological Institute, Moscow (PIN), Zoological lution of the Carangidae started earlier than Up­ Institute, St. Petersburg (ZISP), and Zoological Museum per Paleocene. In the extensive collection from of the Moscow University (ZMMU) are examined for Uylya-Kushlyuk locality, we founda remarkable comparison with the new genus and phylogenetic discus­ undescribed carangid, which has some resem­ sion (the number of specimens examined is given in pa­ blance to the subfamily Trachinotinae and is renthesis; t indicates fossil species): APOLECTJDAE: Apolectus niger (Bloch), ZISP, no. 38126, Tonkin Gulf important for elucidation of the trachinotine (2); ZMMU, no. 11188, Tonkin Gulf (2); tTrachicaranx phylogeny. This specimen is described in the tersus Daniltshenko, PIN, no. 2179-98, Upper Paleocene, present paper as a new genus and species, Uylya­ Turkmenistan (I; holotype). CARANGIDAE: Alectis ichthys eugeniae. ciliaris (Bloch), ZISP, no. 1222, Japan (I); A. indicus (Ruppe II), ZISP, no. 41599, Hainan I. (I); t Archaeus 220 A.M. Prokofiev: A new Paleocene genus o/Carangidae • ZOOSYST. ROSSICA Vol. l l

oblongus Daniltshenko, PIN, no. 2179-94(l; holotype), 4141, Mediterranean Sea(3). POMATOMIDAE: t Ledne­ no. 2179-95( 1; paratype), and 3 uncatalogued specimens, via oligocenica (Smirnov), PIN, no. 4773, Upper all from Upper Paleocene of Turkmenistan; Caranx Oligocene - Lower Miocene, Apsheron Peninsula, Azer­ melampygus Cuvier, ZMMU, no. 5337, Timor I., Indo­ baijan (6); Pomatomus saltator (Linnaeus), ZISP, no. nesia (!); tCaranx sp., PIN, uncatalogued, Lower 4541, Black Sea offKaradag (3 I), PIN, uncatalogued, Oligocene, Gumista River, West Caucasus(2); Caran­ market specimens( 4). RACHYCENTRIDAE:Rachycent­ goides armatus(Forskal), ZISP, no. 3621, Philippines( 1 ); rum canadum(Linnaeus), ZISP, no. 41069, Hainan I.(I). C.ferdau(Forskal), ZISP, no. 2545, Red Sea(!); ZMMU, SCIAENIDAE: Sciaena umbra Linnaeus, ZISP, no. no. 12252, Indian Ocean(3); Chloroscombrus chrysurus 32306, Black Sea off Sevastopol (I); Umbrina cirrosa (Linnaeus), ZISP, no. 37645, West AfricaoffKonakri (1), (Linnaeus}, ZISP, no. 1817, Black Sea (I). SPARIDAE: ° ZMMU, no, 12621, Gulf of Mexico, 28°20'N, 90 00'E Diplodus annu/aris (Linnaeus), ZISP, no. 22670, Black (3); C. orqueta Jordan & Gilbert, ZMMU, no. 13640, Sea offSevastopol (6); D. sargus (Linnaeus), ZISP, no. Pacific Ocean offPeru (5) ; Chorinemus /ysan(Forskal), 912, Nizza(I); Pagel/us erythrinus(Linnaeus), ZISP, no. ZISP, no. 39473, Hainan I. (2), ZMMU, no. 12380, In­ 14283, Biscay Gulf(I); Puntazzo puntazzo (Gmelin), dian Ocean (I); Decapterus lajang Bleeker, ZISP, no. ZISP, no. 32298, Black Sea offSudak (I); Sarpa salpa 37680, WesternAustralia (I); ZMMU, no. 12036, Indian (Linnaeus), ZISP, no. 2740, Cadix (I). In order to evalu­ Ocean (4); D. maruadsi(Temminck & Schlegel), ZISP, ate phylogenetic relationships cladistic approach formu­ no. 42104, Tonkin Gulf (2); Gnathanodon speciosus lated by Hennig(1966) adopted. (Forskal), ZISP, no. 41515, Hainan I. (I); Hemicaranx Abbreviation: SL ·- standard length. sechurae(Hildebrand), ZMMU, no. 13504, PacificOcean off Peru (3); Lichia amia (Linnaeus), ZISP, no. 6815, Canaries(l); ZMMU, no. 9920, Senegal(I), no. 5307, Order PERCIFORMES Naples(2); Naucrates ductor(Linnaeus), ZISP, no. 426, Madeira(!); ZMMU, no. 4913, Naples(I); O/igoplites Family CARANGIDAE Rafinesque, 1815 saurus (Bloch & Schneider), ZMMU, no. 11583, Cuba (2); Paronasignata (Jenyns), ZISP, no. 42649, Rio de la Genus Uylyaichthys gen. n. Plata off Montevideo (I); t Scomberoides spinosus (Smimov), PIN, uncatalogued, Middle Miocene, Cher­ naya Rechka, Azerbaijan(2); Scyris a/exandrinus(Geof­ Type species. Uylyaichthys eugeniae sp. n. froy Saint-Hilaire), ZISP, no. 41019, West Africa off Diagnosis. Body deeply fusiform, with dorsal Dakar; t Selar cf. quassus Bannikov, PIN, nos. 4773/128, profile slightly more convex than ventral one; 130-131, Upper Oligocene - Lower Miocene, Apsheron maximum body depth contained nearly 2.6 times Peninsula, Azerbaijan(3); Se/amidesleptolepis (Cuvier), in SL. Head moderately long, with high fronto­ ZISP, no. 38147, Tonkin Gulf(!); ZMMU, no. 10857, TonkinGulf (I); Selene vomer(Linnaeus), ZISP, no. 1360, supraoccipital crest, anteriorly extended forward Rio de Janeiro(!); S. (Mitchill), ZMMU, no. to orbit. Mouth oblique; angle between mouth setapinnis ° 10479, West Africa, off Dakar (l); Serio/a dumerili rim and longitudinal body axis about 40 . Pre­ (Risso), ZMMU, no. 4950, Naples(2); S. quinqueradiata maxillary with very small conical teeth in sev­ Temminck & Schlegel, ZISP, no. 4757, Yokohama, Ja­ eral rows. Supramaxillary present. Rear of max­ pan (3); Trachinotus bail/oni Lacepede, ZISP, no. 624, Amboina (l); T. blochii (Lacepedc), ZMMU, no. 155, illary extends only to anterior third of orbit. Samoa(!); T. ovatus(Linnaeus), ZISP, no. 3258, Bahia Parasphenoid slender in lateral view and strongly (2), no. 11450, Naples(l); T. paitensis Cuvier, ZMMU, curved mesially. Posttemporal with equidimen­ no. 18695, near Callao, Peru (1 ); T. russe/li Cuvier & sional dorsal and ventral branches. Pelvic bones Valenciennes, ZISP, no. 23559, Kominato, Amami­ without postpelvic processes forming "apical Oshima, Japan (I}; ZMMU, no. 153, Atlantic? (I); fork". Vertebrae21 or 22, of which 9 abdominal Trachinotus sp., ZMMU, no. 9872, off St. Louis (I); Trachurus trachurus(Linnaeus), ZISP, no. 2829, locality and 12 or 13 caudal. Inferior vertebral foramina unknown(!); Tjaponicus(Temminck & Schlegel), ZISP, absent. Ribs long and strong, extending nearly no. 31376, Japan (I); T. mediterraneus (Steindachner), to ventral margin of body; all of them immedi­ ZMMU, no. 20449, Kerch Strait(!). CORYPHAEN­ ately attached to lateral sides of vertebral centra. IDAE: Linnaeus, ZISP, no. 42244, Coryphaenahippurus Caudal-fin support with three epurals and sepa­ Sea of Japan (2). LACTARIIDAE: Lactarius lactarius (Bloch & Schneider), ZISP, no. 36050, Hainan I. (3). rate fifth hypural. Dorsal fin with 7 spines and LEIOGNATHIDAE: Gazza minuta(Bloch), ZMMU, no. about 24 soft-rays; spines posteriorly increasing 5325, Manila (I); Leiognathus bindus (Valenciennes), in length. Predorsal formula (Ahlstrom et al., ZISP, no. 41136, Singapore(2); L. brevirostris, ZMMU, 1976) 0/0+0+0/l/. Anterior projection of first no. 10950, Tonkin Gulf(2); L. equula(Forskal), ZISP, supraneural touching apex of supraoccipital crest. no. 37457, Hainan I. (I); ZMMU, no. 5383, Cebu, Moluccas (I); t Leiognathoides(?) altapinna (Weiler), Anal fin with 3 spines and about 16 soft-rays. PIN, uncatalogued, Lower Oligocene, Ljubizhna River, No traces of anterior lobes in soft portions of Ukrainian Carpathians(!); tL. minutus(Daniltshenko), dorsal and anal finspresent. Ventro-analdistance PIN, nos. 4773/108-111, Upper Oligocene- Lower Mio­ much smaller than anal-fin base length. First cene, Apsheron Peninsula, Azerbaijan (4). MENIDAE: anal-fin pterygiophore firmly articulated with Mene macu/ata (Bloch), ZISP, no. 37107, Tonkin Gulf (4); ZMMU, no. 12090, Indian Ocean(4); tM. triangulum firsthaemal spine. Three anal-finpterygiophores Daniltshenko, PIN, uncatalogued, Upper Paleocene, placed in first interhaemal space. No detached Turkmenistan (4). NEMATISTIIDAE: Nematistius pec­ or semidetached finlets present. Scales small, tora/is Gill, ZISP, no. 2753, South California(I). POMA­ cycloid; lateral body scutes absent. CENTRIDAE: Chromis chromis(Linnaeus), ZISP, no. Included species. Type-species only. ZOOSYST. ROSSICA Vol. 11 • A.M Prokofiev: A new Paleocene genus ofCarangidae 221 Comparison. The new genus is distinguished and most ribs borneon lateral sides of vertebral from the all known Carangidae by the presence centra. In addition to the number of vertebrae of 2 l ( or possibly 22) total vertebrae, of which 9 and parasphenoid shape, Lichia is also distinct abdominal and 12 ( or 13) caudal ( vs. total verte­ from Uylyaichthys in the rear of maxillary ex­ bral count of 24-27 in other genera, of which 10 or tended behind orbit (vs. to anterior third of or­ 11 abdominal); predorsal formula0/0+o+-0/1/; and bit), and mouth rim nearly straight (vs. strongly parasphenoid distinctly curved (vs. nearly oblique); presence of enlarged conical teeth in straight in other genera). It differs from the outer row on jaws; presence of lobes in anterior Seriolinae and Archaeinae in the deep body with portions of softdorsal and anal fins; presence of high fronto-supraoccipital crest; jaws with fine postpelvic processes and inferior vertebral fo­ teeth; strongly oblique mouth; firm articulation ramina; and absence of separate fifth hypural. between first haemal spine and first anal-fin Range. Upper Paleocene of Turkmenistan. pterygiophore; dorsal-fin spines increasing in Etymology. The generic name is formed from length posteriorly; anterior projection of first Uylya-Kushlyuk locality, and ichthys (Greek, for supraneural touching supraoccipital crest; and fish). from Seriolinae also in the ventro-anal distance much smaller than anal-fin base length (vs. op­ Uylyaichthys eugeniae sp. n. posite condition in Seriolinae). Uylyaichthys is (Figs 1-3) clearly distinguished fromthe Caranginae by the dorsal--fin spines gradually increasing in length Holotype. PIN, no. 4782-80, complete skeleton, single posteriorly, anterior projection of first supra­ plate, Turkmenistan,2 km northeastofUylya-Kushlyuk village, Upper Paleocene, middle part of Danata Forma­ neural touching supraoccipital crest and absence tion, leg. E.K. Sytchevskaya. oflateral body scutes. The new genus differsfrom Other material examined. PIN, no. 4782-37, fragment the Scomberoidinae in the strongly oblique of middle part of body, single plate, same locality and mouth; reduced number of caudal vertebrae ( 12 age as forholotype. or possibly 13 vs. 16-17); supraneurals 2 and 3 Description. Body laterally compressed, deep­ not divided by neural spine; and anterior projec­ ly fusiformin shape, with slender caudal pedun­ tion of first supraneural touching supraoccipital cle. Upper profile of body more convex than crest. The most closely related subfamilyTrachi­ lower one. Maximum body depth (across verti­ notinae s. str. shares the following similarities cal of anal-finorigin) contained 2.6 times in SL. within Uylyaichthys: deep body with high fronto­ Head moderately long, 3.3 times in SL; upper supraoccipital crest extending forwardto the or­ profile smoothly curved; mouth moderate. Hori­ bit; jaw dentition reduced; increased number of zontal diameter of round orbit contained 3.6 times supraneurals, with concomitant reduction of dor­ in head length. Snout length contained 1.3 times sal spines in number; anterior projection of first in orbit diameter. Mouth cleftoblique; angle be­ supraneural touching supraoccipital crest; dor­ tween mouth rim and longitudinal axis of body sal-fin spines increasing in length posteriorly; about 40°. and most ribs (but all in Uylyaichthys) borneon Skull. Limits of individual bones of neurocra­ lateral sides of vertebral centra. However, two nium not clear, but frontals apparently wide, formerly known genera of the Trachinotinae s. smooth, except closely spaced, nearly trans­ str. (Trachinotus and Paratrachinotus) differ versely arranged crests on posterior margin. from Uylyaichthys in vertebral formula I 0 + 24; Fronto-supraoccipital crest very high, anteriorly absence of supramaxillary; mouth rim nearly extended before the orbit, with straight poste­ straight (vs. strongly oblique); nearly straight rior margin; its height contained 2.5 times in head parasphenoid (vs. distinctly curved mesially); length. Mesethmoideum nearly triangular, plate­ presence of!obes in anterior portions of soft dor­ like in lateral view. Lateral faces of sphenotic sal and anal fins; supraneurals 2 and 3 divided and pterotic bearing distinct crests. Parasphenoid by neural spine; and absence of separate fifth relatively slender in lateral view, distinctly hypural. Uylyaichthys also differs from Trachi­ curved mesially, exposed in lower part of orbit. notus in the equally developed dorsal and ven­ Basisphenoid well developed. Shaft of hyo­ tral branches of posttemporal; and from Para­ mandibular long, slender, oriented nearly verti­ trachinotus in the shaftsof three (vs. seven) anal­ cally. Ventral end of hyomandibular shaft re­ fin pterygiophores placed in first interhaemal moved fromposterior end of quadrate. Quadrate space, and in much longer ribs. The genus Lichia triangular, with slightly concave upper margin was formerly included in the Trachinotinae and well developed articularfacet for lower jaw. (Smith-Vaniz, 1984; Gushiken, 1988; Bannikov, Symplectic small, rod-like. Ectopterygoid nar­ 1990; but see discussion below), but it lacks any row, semi-lunar; endo- and metapterygoids trachinotine characters mentioned above, except broad, plate-like, with not clearly seen limits. dorsal-finspines increasing in length posteriorly Lower jaw articulation positioned under mid- 222 A.M Prokofiev: A new Paleocene genus ofCarangidae • ZOOSYST. ROSSICA Vol. 11

dle of orbit; lower jaw rela­ tively slender, imperfectlypre­ served. Limits of dentary and articular as well as dentition on former bone not clear. Alveo­ lar arm ofpremaxillary slender, long, bearing several rows of very small conical teeth. As­ cending process of premaxil­ lary moderate; smaller separate articular process present. Max­ illary greatly expanded posteriorly, with small su­ pramaxillary on posterodorsal border. Bones of opercular series coarse, flat, imperfectly pre­ served. However, preopercle possibly relatively narrow, with smooth postr�rior border, and opercle approximately triangu­ lar. Hyoid and branchial arches not clear; number ofbranchio­ stegal rays unknown. Gill--rak­ ers slender, moderately long. Paired finsand their girdles. Forked posttcmporal, with slender, equally sized dorsal and ventral branches situated just above level of vertebral column. Supracleithrum rela­ tively narrow, flat. Cleithrum curved, broadened in middle part, incompletely preserved. Pectoral finrelatively short, not falciform, originating slightly closer to ventral body profile than to vertebral column. Pec­ toral-fin subequal in length to five abdominal vertebrae. There are about 18 pectoral-fin rays. Pelvic fins poorly pre­ served, possibly short. Pelvic bones elongate triangular, with­ out postpelvic processes. Vertebral column. There are 21 vertebrae, 12 of these cau­ dal. However, third preural ver­ tebra with two neural spines, which corresponds to distortion of development of vertebral column during ontogeny. Length of caudal part of verte­ bral column contained 2.3 times in SL. Vertebral centra nearly quadrate, with distinct longitudinal crests on their sides. Posterior three abdomi­ Fig. 1. Uylyaichthys eugeniae gen. et sp. n., holotype. nal vertebrae with triangular ZOOSYST.ROSSI CA Vol. 11 A.M• Prokofiev:A new Paleocene genus of Carangidae 223

Figs 2, 3. Uylyaichthys eugeniae gen. et sp. n.: 2, reconstruction of skeleton (based mostly on PIN, no. 4782-80 and partiallyon PIN, no. 4782-37); 3, caudal-finskeleton. Abbreviations: epu, epurals; hp, hypurals; hpu2, haemal spine of second prcural vertebra;npu2, neural spine of second preural vertebra; npu3, neural spines of third preural vertebra;ph, parhypural; pu2, second preural vertebra; pu3, third preural vertebra; tc, terminalcentrum (Pul+Ul); un, uroneural. parapophyses gradually increasing in length Hypurals I to 4 fusedinto two plates. Small fifth posteriorly. Six anterior neural spines in abdomi­ hypural present. Neural spine of second preural nal region moderately inclined caudally and vertebra very short. broadened basally; neural spines in middle part Dorsal and anal fins. Dorsal and anal fins long­ of vertebral column (to fourth caudal vertebra) based, with continuous spinous and softportions. slender and nearly vertical; remaining neural Last dorsal-finray situated on same vertical with spines inclined caudally, much more on poste­ last anal-finray. Both dorsal- and anal-finptery­ rior caudal vertebrae. Inclination of haemal giophores with broad anterior and rather small spines also distinctly increasing posteriorly. In­ posterior plate-like extensions. Dorsal fin of 7 feriorvertebral foramina absent. There are strong, spines and 24 soft-rays. Dorsal-fin spines rela­ long, posteroventrally inclined pleural ribs ex­ tively short, increasing in length posteriorly. Last tending nearly to ventral profileof body. All pleu­ dorsal spine the longest, contained nearly 4.4 ral ribs borneon lateral sides of vertebral centra. times in maximum body depth. Soft portion of Caudal fin and skeleton. Caudal fin deeply dorsal fin originates above eight abdominal ver­ emarginate, its length contained 1.2 times in head tebra. Soft dorsal-fin rays relatively short, not length. There are 1 7 principal and 10 + 8 procur­ forming separate lobe anteriorly and gradually rent caudal-fin rays. Caudal-fin rays slightly decreasing in length posteriorly. Longest dorsal­ overlap hypurals, parhypural and epurals. Cau­ finray slightly longer than longest dorsal spine; dal-fin support includes three separate hypural anterior dorsal ray 2.3 times as long as the last plates, parhypural, stegural, and three epurals. one. First dorsal-fin pterygiophore placed in 224 A.M. Prokofiev: A new Paleocene genus ofCarangidae • ZOOSYST. ROSSICA Vol. 11

Seriolinae Archaeinae

Scomberoidinae

9

4 10 Paratrachinotus Trachinotus

I l !ylyaichthy.1 3 I* 5 6(11] Li,-hia

2 [ I I) 12 Caranginae A B

Fig. 4. Cladogram showing monophyletic groups ofCarangidae based on characters discussed in_ the text. Characters_ in brackets are possibly convergent. Node A indicates presence of gap between last two anal spmes; node B md1cates presence of firm articulation between firstanal-fin pterygiophore and first haemal spine.

interspace between third and fourthneural spines, mum body depth 38. 75; minimum depth of cau­ with strong process projecting anteriorly. dal peduncle 10.0; caudal peduncle length 12.5; There are foursupraneurals, each with distinct, head length 30.0; length of caudal part of verte­ anteriorly directed processes; the process on first bral column 43.75; first predorsal distance 38.2; supraneural touching apex of supraoccipital crest. second predorsal distance 50.6; preanal distance Predorsal formula (Ahlstrom et al., 1976) 0/ 61.8; ventro-anal distance 22.5; length of dor­ 0+0+0/1/. sal-finbase 58.75; length of anal-finbase 26.25; Anal fin originates slightly posterior to soft caudal-fin length 23.75; in percentage of head dorsal-fin origin. Ventro-anal distance much length: snout length 18.75; horizontal orbital di­ smaller than anal-fin base length. Anterior part ameter 31.25; maxillary length 25.0; maxillary of anal finpoorly preserved in holotype but well depth 10.4. observable in additional specimen (PIN no. 4782- Range. Known only from the type locality. 37). There are 3 spines and about 16 soft anal­ Etymology. The species is named in honour of fin rays. There is distinct gap between last two Eugenia K. Sytchevskaya (PIN), who collected anal-fin spines. First anal-fin pterygiophore the specimens. broadened ventrally, firmlyarticulated with first haemal spine. Other anal pterygiophores similar Relationships of Uylyaichthys in shape to corresponding dorsal-fin pterygio­ phores. Shafts of three anal-fin pterygiophores According to the most recent revision (Ban­ situated in first interhaemal space. nikov, 1990), the familyCarangidae is composed Scales. There are small, uniform,cycloid scales of six subfamilies: Seriolinae, Archaeinae, Vome­ poorly preserved on whole body. Diameter of sin­ ropsinae, Scomberoidinae, Trachinotinac, and gle scale from middle part of body much less Caranginae, among which the Vomeropsinae than length of single vertebra. Lateral line not possess a number of peculiar characters (Ban­ seen; lateral line scutes absent. nikov, 1984, 1990) justifying family rank, pos­ Measurements. SL of holotype 80 mm, but sibly without direct relationships with the additional specimen (PIN, no. 4782-37) much Carangidae. The monophyly of Carangidae is larger (estimated SL about 250 mm). Measure­ confirmed by the presence of a relatively wide ments of holotype in percentage of SL: maxi- gap between last two anal-fin spines (Smith- ZOOSYST. ROSSICA Vol. 11• A.M Prokofiev: A new Paleocene genus ofCarangidae 225 Vaniz,1984; Gushiken, 1988). However, this gap last two pairs, are attached directly to lateral sides is also present in the recent apolectid Apolectus of vertebral centra, though parapophyses are de­ niger, while absent in the primitive Paleocene veloped fromthird or fourthvertebra, what seems apolectid Trachicaranx. Possibly, this suggests to be apomorphic(Smith-Vaniz, 1984; Gushiken, parallel development of this character in the 1988). In Uylyaichthys, all pleural ribs seem to Carangidae and Apolectidae. Among Carangidae, be borne on lateral sides of vertebral centra and the Seriolinae and Archaeinae seem to be the parapophyses are found from eighth vertebra. most primitive, judging fromthe feeblearticula­ Attachment of most ribs to lateral sides of verte­ tion between firstanal-fin pterygiophore and first bral centra is derived forScomberoidinae, Trachi­ haemal spine. Most authors (Regan, 1909; Su­ notinae s. str. and Lichia, and such an attach­ zuki, 1962; Bannikov, 1986, 1990) noted the ment forall ribs is derived for Uylyaichthys. This Seriolinae as ancestral forother carangids on the character is unknown for Paratrachinotus. basis of the followingprimitive characters: body 2. The longest dorsalfin spines situated in elongately fusiform, not strongly compressed middle offin (0)- dorsal-fin spinesare increas­ laterally, with low supraoccipital crest; presence ing in length posteriorly (I). In most Carangidae, of supramaxillary; absence of body scutes; un­ the longest dorsal spines occur in the middle of paired fins low, without lobes; anal-fin base first dorsal finas in many generalised percoids, shorter than ventro-anal distance; feeblearticu­ while in Scomberoidinae (except forScombero­ lation between first anal-fin pterygiophore and ides spinosus), Trachinotinae s. str. and Lichia first haemal spine; unconsolidated caudal-fin dorsal spines are gradually increasing in length support; etc. However, Smith-Vaniz(1984) and posteriorly. This state seems to be derived. Scom­ Gushiken (1988) reported the Seriolinae ( men­ beroides spinosus has peculiar, fan-shaped spin­ tioned as Naucratini) as a sister group of the ous dorsal fin(see also Bannikov, 1990, p. 23), Caranginae (mentioned as Carangini), and the which is possibly a secondary specialisation. Naucratini + Carangini as a sister group forother 3. The number ofsupraneurals is three (0) - carangids on the basis of attachment of anterior number ofsupraneurals increased to 4-6 with portion of m. adductor mandibularis to subor­ concomitant decrease of dorsal-fin spines in bital shelf in the Naucratini + Carangini. This number (1). Most of the generalised percoids opinion has been criticised by Bannikov (1987, have three supraneurals (Johnson, 1983, 1984; 1990). Archaeinae is distinguished from the Tominaga, 1986), as well as most of the Caran­ Seriolinae in the ventro-anal distance less than gidae, including Lichia. Only Serio/a quinque­ anal-finbase length, which seems to be advanced radiata occassionally has four supraneurals and is characteristic for all Carangidae, except (Bannikov, 1990). In the Scomberoidinae (ex­ the Seriolinae (Bannikov, 1990). The presence cept Scomberoides)and Trachinotinae s. str., the of firmarticulation between first anal-fin ptery­ number of supraneurals increased to 4-6 with giophore and first haemal spine occurring in concomitant decrease of dorsal-fin spines in Uylyaichthys exclude the latter from the Serio­ number, which seems to be derived. In Scombero­ linae or Archaeinae. Interrelationships of Uylya­ ides, the number of supraneurals is usually three ichthys and other advanced carangids with firm (Smith-Vaniz & Staiger, 1973; Bannikov, 1990; articulation between firstanal-fin pterygiophore pers. observ.); this is probably a reversal or the and first haemal spine are given in the present atavistic condition. discussion. 4. The anterior projection of the first supra­ Characters listed below arc used in the clado­ neural does not touch the apex ofsupraoccipital gram (Fig. 4) showing relationships of Uylya­ crest (0) - the anterior projection of the first ichthys. Only characters available in the fossil supraneural is touching the apex of supra­ material are included in the present study. Plesio­ occipital crest (]). In most carangids and other morphic characters are noted as (0), and apo­ percoids examined, excluding Trachinotus, morphic ones noted as(!). Paratrachinotus, and Uylyaichthys, the anterior 1. Pleural ribs are borne in pits 011 vertebral projection of the firstsupraneural does not touch centra or in parapophyses, when parapophyses the apex of supraoccipital crest. The only other aredeveloped (0)- most pleural ribs, except the examined percoid fish with such a touch is the last two pairs, are attached directly to lateral sparid Pagellus erythrinus(ZISP, no. 14283). sides of vertebral centra (1) - all pleural ribs Since there are no other synapomorphies between f are attached to lateral sides o vertebral centra the Sparidae and Carangidae, this character is (1 *). In most Carangidae and other percoids stud­ considered independently derived forthe Trachi­ ied, pleural ribs are borne in pits on vertebral notinae s. str. centra or in parapophyses, when parapophyses 5. The number of vertebrae is 24 or more (0) - are developed, while in Scomberoidinae, Trachi­ the number of vertebrae is 21 or 22 (]). In all notus and Lichia most pleural ribs, except the the Carangidae, except Uylyaichthys, the total 226 A.M. Prokofiev: A new Paleocene genus of Carangidae • ZOOSYST. ROSSICA Vol. 11 number of vertebrae is 24-27. In Uylyaichthys, 10. Ventral branch of posttemporal not reduced the total number of vertebrae is reduced to 21 or (OJ - ventral branch of posttemporal greatly re­ possibly 22. The number of abdominal vertebrae duced (IJ. In all Carangidae and related caran­ is also reduced in Uylyaichthys (9 vs. 10 or 11 in goid families, except Trachinotus, dorsal and all other Carangidae). This state seems to be ventral branches ofposttemporal are equidimen­ apomorphic. Among related families, similar sional. Among the Trachinotinae s .. str., Uylya-­ number of vertebrae (9 + 13-14 = 22-23) also ichthys possesses this primitive state. Structure occurs in the Menidae, but this character un­ ofposttemporal is not known forParatrachinotus doubtedly was independently derived in Uylya­ (Blot, 1969). ichthys and Menidae. 11. Inferior vertebral foramina absent (OJ - 6. Parasphenoid relatively straight in outline inferior vertebral foramina present (IJ. Inferior (0J-parasphenoid strongly curvedmesially (JJ. vertebral foramina are present within Percoidei In all the Carangidae excluding Uylyaichthys, only in Lichia and most of the Caranginae ( ex­ and in the other studied familiesof percoids, the cluding Selar, Trachurus, and several Caran­ parasphenoid is relatively straight in outline, goides). This is an advanced character for lichia while in Uylyaichthys it is strongly curved me­ and the Caranginae, but it is possibly independ­ sially. This is considered to be a unique derived ently derived, because Lichia has no other syn­ character of Uylyaichthys. apomorphies with the Caranginae and possesses 7. Supramaxillary present (OJ - supramaxil­ derived states of characters I and 2, which sug­ lary absent (JJ. The presence of supramaxillary gests it to be sister group for Scomberoidinae + is a generalised character forpercoids (Patterson, Trachinotinae s. str. 1964; Johnson, 1984; Sasaki, 1989). The loss of 12. Lateral body scutes absent (OJ - lateral supramaxillary has undoubtedly occurred along body scutes present (IJ. Members of the sub­ a number ofpercoid lineages. Among the Caran­ family Caranginae have variously developed gidae, supramaxillary is lacking in scombero­ body scutes covering the lateral line, while in idine Oligoplites and trachinotine Paratrachi­ other carangids, including Uylyaichthys, such notus and Trachinotus.As other synapomorphies scutes are absent. Presence of scutes is a unique between Oligoplitesand Trachinotinae s. str. are character of the Caranginae, and it is considered absent, we consider this character independently to be a synapomorphy ofthis subfamily (see also derived in Oligoplites and both Trachinotus + Gushiken, 1988). Paratrachinotus; this indicates relatively ad­ Two other advanced characters of the Trachi­ vanced status of these genera. notinae s. str. not included presently in the clado­ 8. Anterior soft dorsal and anal rays not ex­ gram: body is deeply fusiformor extremely deep tended and not forming distinct lobes (OJ - ante­ (maximum depth/SL ratio 2.7 or less vs. 2.8 or rior soft dorsal and anal rays more or less ex­ more) and correspondingly very high fronto­ tended and forming anterior lobes (JJ. In the supraoccipital crest extended anterior to orbit. most generalised Seriolinae and Archaeinae, the These characters possibly independently derived softdorsal and anal finslack anterior lobes. Elon­ in many phyletic lineages (e.g., in Parona from gation ofanterior softrays ofunpaired finsseems Scomberoidinae, and in many Caranginae) and to be advanced, but it is foundin various phyletic seem to have no important phyletic value but lineages of the Carangidae. Among the Trachi­ indicating relatively advanced status. Several de-­ notinae s. str., both Trachinotus and Paratrachi­ rived characters occurring in Trachinotus (para­ notus have falciform anterior lobes, while their sphenoid expanded into broad, flattened plate closest relative Uylyaichthys lacks such lobes. posteriorly; upper pharyngeals enlarged and cov­ This probably suggests that this character is in­ ered with blunt, molariform teeth; and presence dependently derived in the Paratrachinotus­ oflateral processes on ventral surface of the ba­ Trachinotus clade and other phyletic lineages of sioccipital) are not observable in two fossil tra­ the Carangidae and it indicates relatively ad­ chinotine genera. Also Uylyaichthys has sepa­ vanced status of the Paratrachinotus-Trachi­ rate fifth hypural in caudal skeleton, which oc­ notus clade. curs in the Carangidae only in the Eocene Serio/a 9. Thenumber ofanal-fin pterygiophores in the prisca (Blot, 1969). This condition seems to be firstinterhaemal space is three (OJ - the number of very primitive but possibly has no important anal-fin pterygiophores in the first interhaemal phylogenetic significance; it suggests close po­ space is seven (JJ. In most carangids and other sition of Uylyaichthys with generalisedancestor percoid families studied, the number of anal-fin of the subfamily Trachinotinae s. str. pterygiophores in the first interhaemal space is The genus Lichia possesses supraneurals 2 and three, while in Paratrachinotus it is described as 3 enclosed in a single intemeural space, while in seven (Blot, 1969). This is undoubtedly a unique the Scomberoidinae and Trachinotinae s. str. (ex­ apomorphic character of Paratrachinotus. cluding Uylyaichthys) supraneurals 2 and 3 are ZOOSYST. ROSSICA Vol. 11 • A.M. Prokofiev: A new Paleocene genus ofCarangidae 227 divided by the neural spine. Tominaga (1986) their monophyletic origin. We place both sub­ listed primitive predorsal formula as 0/0/0+2/, families in a monophyletic lineage, to which which occurs among the Carangidae in the lichia is a sister group, on the basis of the de­ Seriolinae. Bannikov (1990, p. 15, fig. 2) fig­ rived state of character 3 occurring in the ured predorsal formula 0/0+0/1 + 1/ forArchaeus Scomberoidinae and Trachinotinae s. str. Both oblongus, but personal examination ofa para type Scomberoidinae + Trachinotinae s. str. and lichia (PIN, no. 2179-95) of this species with un­ have derived state of characters l and 2, which distorted predorsal bones has found it is clearly shows its monophyletic origin. Trachinotinae s. 0/0/0+ 1+ 1/. Both Scomberoidinae and Trachinot­ str. are a sister group ofScomberoidinae, and the inae s. str.· have derived state of character 3, al­ trachinotine Uylyaichthys is a sister group of though their predorsal formula is variable, but Trachinotus-Paratrachinotus clade. in all of them, except for Uylyaichthys, the sec­ ond and third supraneurals are divided by neural Acknowledgments spine as in the generalised Seriolinac andArchae­ inae. Predorsal formula 0/0+0+0/l/ of Uylyaich­ I am grateful to Dr. Eugenia K. Sytchevskaya (PIN) forcollection and loaning of material; to Drs. A. V. Balush­ thys is aberrant and independently derived. In kin, V.V. Fedorov, GA. Volkova, O.S. Voskoboinikova, Lichia, as in the Caranginae, supraneurals 2 and A.M. Naseka and M.V. Nazarkin (ZISP), for providing 3 are enclosed in a single intemeural space, and preserved material and radiographs for comparison, dis­ its predorsal formulais 0/0+0/2/, 0/0+0/2+ 1/, or cussion, and kind help during my visits in ZISP; to Drs. 0/0+0/l +1/, which suggests relatively advanced Yu.I. Sazonov and I.A. Verigina (ZMMU), and A.F. status but possibly has no important phyletic sig­ Bannikov (PIN) for providing material for comparison nificance,because this character is independently and some assistance. derived in various phyletic lines of Percoidei. lichia possesses inferior vertebral foramina as References in most of the Caranginae, but it is possibly a Ahlstrom, E.H., Butler, J.L. & Sumida, B.Y. 1976. result of parallelism, since other synapomorphies Pelagic stromateoid fishes (Pisces, Perciformes) of between the Caranginae and Lichia are not the eastern Pacific: kinds, distribution and early life known. The only known apomorphic state of histories and observations on five of these from the Lichia is the very irregular, sinuous lateral line northwest Atlantic. Bull. mar. Sci., 26(3): 285-402. (Smith-Vaniz & Staiger, 1973), which is not ob­ Bannikov, A.F. 1984. The new subfamily of carangid fishes. In: Studi e ricerchesui giacimenti terziari di servable in fossiltrachinotine genera. The genus Bolca, 4: 319-321. lichia was formerly included in the Bannikov, A.F. 1986: On the systematic position, com­ Trachinotinae (Smith-Vaniz, 1984; Gushiken, position and origin of the family Carangidae. Vopr. 1988; Bannikov, 1990), however, on the basis of Ikhtiol., 26(6): 883-889. (In Russian). the present study it seems to be doubtful.As noted Bannikov,A.F. 1987. On phylogenetic relationships ofCara­ nginac. Paleontol. Zh., 1987(3): 48-59. (In Russian). previously (Smith-Vaniz & Staiger, 1973), Lichia Bannikov, A.F. 1990. Fossil carangids and apolectids of probably represents a separate subfamily. Smith­ the USSR. Trudy paleontol. Inst. Akad. Nauk SSSR, Vaniz& Staiger (1973) noted that the Scombero­ 244: l-l08. (In Russian). idinae, Trachinotus and Lichia are distinguish­ Bannikov, A.F. & Fedotov, V.F.1984. On fossilApolect­ able as a group fromall other genera of the Caran­ idae (Perciformes). Vopr. lkhtiol., 24(3): 500-501. (In gidae at least on the basis of superficialsimilari­ Russian). Blot, J. 1969. Les poissons fossilesdu Monte Bolca, clas­ ties, which have little or no phylogenetic signifi­ sesjusqu'ici dans Jes families des Carangidae, Men­ cance (Smith-Vaniz & Staiger, 1973, p. 244), but idae, Ephippidae, Scatophagidae. Mus. civ. Stor. nat. the monophyletic origin of the Scomberoidinae, Verona, Mem.fuori, Ser. 2, 1: 1-525. Trachinotusand Lichia was proved later (Smith­ Daniltshenko, P.G 1968. Fishes of the Upper Paleocene Vaniz, 1984; Gushiken, 1988). We found no of Turkmenia. In: Ocherki po filogenii i sistematike iskopaemykh ryb i beschelyustnykh [Studies on synapomorphies of the Trachinotinae s. str. and phylogeny and systematics of fossilPisces and Agna­ Lichia, but the Scomberoidinae, Trachinotinae tha]: I 13-156. Moscow: Nauka. (In Russian). s. str. and Lichia represent a monophyletic group Grossheim, V.A. & Korobkov, I.A. (eds). 1975. Strati­ of the Carangidae. grafiya SSSR: Paleogenovaya sistema [Stratigraphy of the USSR: Paleogene System]. 524 p. Moscow: The Scomberoidinae possess a number of Nedra. (In Russian). unique specialisations (Smith-Vaniz & Staiger, Gushiken, S. 1988. Phylogenetic relationships of the 1973; Smith-Vaniz, 1984; Mok, 1986; Gushiken, perciform genera of the family Carangidae. Jap. J 1988; Bannikov, 1990), which undoubtedly dem­ Ichthyol., 34( 4 ): 443-461. onstrates their monophyletic origin. Most authors Hennig, W. 1966. Phylogenetic systematics. 263 p. Ur­ (Suzuki, 1962; Smith-Vaniz & Staiger, 1973; bana: Univ. Illinois Press. Johnson, G.D. 1983. Niphon spinosus: a primitive Smith-Vaniz, 1984; Gushiken, 1988; Bannikov, epinepheline serranid, with comments on the mono­ 1990) usually reported the Scomberoidinae and phyly and intrarelationships of the Serranidae. Trachinotinae as closely related groups or proved Copeia, 1983(3): 777-787. 228 A.M. Prokofiev: A new Paleocene genus ofCarangidae • ZOOSYST. ROSSICA Vol. 11

Johnson, GD. 1984. Percoidei: development and rela­ Sasaki, K. 1989. Phylogeny of the family Sciaenidae, tionships.Amer. Soc. lchthyol. llerpetol., Spee. Pub!., with notes on its zoogeography (Teleostei, Perci­ no. l: 464-498. formes). Mem. Fae. Fish. Hokkaido Univ., 36(1/2): Johnson, GD. & Patterson, C. 1993. Percomorph phylo­ 1-137. geny: a survey of acanthomorphs and a new proposal. Smith-Vaniz, W.F. 1984. Carangidae; relationships. Bull. mar. Sci., 52(1): 554-626. Amer. Soc. lchthyol. Herpetol., Spee. Pub/., no. l: Mok, H.K. 1986. Myology, osteology and phylogeny of 522-530. carangid fishes. In: lndo-Pacificfish biology: Proc. Smith-Vaniz, W.F. & Staiger, J.C. 1973. Comparative 2nd lnt. Conf lndo-Pacificfishes: 950-951. Tokyo: revision of Scomberoides, Oligoplites, Parona, and Ichthyol. Soc. Japan. Hypacanthus with comments on the phylogenetic Muzylev, N.G 1994. Succession of geological events in position of Campogramma (Pisces: Carangidae). Early Paleogene of Peritethys. In: Biostratigrafiya Proc. Calif A cad. Sci.( 4 ), 39( I 3): I 85-256. neftegazonosnykhbasseinov: 1 Mezhdunar. Simp., S.­ Solun, V.I. (ed.) 1975. Paleogen Turkmenii: Biostrati­ Peterburg, 3-10 dek. 1994: Tezisy dokladov [Bio­ graficheskoeraschlenenie i korre/yatsiya [The Paleo­ stratigraphy ofoil and gas basins. Proc. I st Int. Symp., gene of Turkmenia: Biostratigraphical zonation and St.Petersburg, Dec. 3-10, 1994]: 67-68. St.Petersburg. correlation]. 195 p. Ashkhabad: Ylym. (In Russian). (In Russian). Suzuki, K. 1962. Anatomical and taxonomical studies Patterson, C. 1964. A review of Mesozoic acantho­ on the carangid fishesof Japan. Rep. Fae. Fish., Pref pterygian fishes, with special reference to those of Univ. Mie, 4: 43-232. the English Chalk. Philos. Trans. roy. Soc. London Tominaga, Y. 1986. The relationships of the familyGlau­ (8), 247(739): 213-482. cosomatidae and Pempherididae. In: !ndo-Pacificfish Regan, C.T. 1909. On the anatomy and classificationof the biology: Proc. 2ndlnt. Conf lndo-Pacificfishes:595- scombroid fishes. Ann. Mag. nat. Hist. (8), 3: 66-75. 599. Tokyo: Jchthyol. Soc. Japan. Received 2 August 2001