Paxton concilians: A New and of Pseudamine Apogonid (Teleostei: Percoidei) from Northwestern Australia: The Sister Group of the Enigmatic Gymnapogon Author(s): Carole C. Baldwin and G. David Johnson Source: Copeia, Vol. 1999, No. 4 (Dec. 17, 1999), pp. 1050-1071 Published by: American Society of Ichthyologists and Herpetologists Stable URL: http://www.jstor.org/stable/1447980 Accessed: 02/06/2010 15:32

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=asih.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

American Society of Ichthyologists and Herpetologists is collaborating with JSTOR to digitize, preserve and extend access to Copeia.

http://www.jstor.org Copeia, 1999(4), pp. 1050-1071

Paxton concilians:A New Genus and Species of Pseudamine Apogonid (Teleostei: Percoidei) from Northwestern Australia: The Sister Group of the Enigmatic Gymnapogon

CAROLE C. BALDWIN AND G. DAVID JOHNSON

Paxtonconcilians is described as a new genus and species of pseudamineapogonid from four specimens trawled off northwesternAustralia. It differs most conspicu- ously from all other apogonids in having a continuous dorsal fin but shares with other members of the family certain specializationsof the dorsal gill arches, last spinous dorsal fin pterygiophore,and cutaneous sensory system. We recognize two subfamiliesof apogonids, the Apogoninaeand Pseudaminae,and diagnosethe latter on the basis of numerous derived features. We propose that Paxtonbe added as a fourth member of the Pseudaminae(with Gymnapogon,Pseudamia, and Pseudamiops) and provide strong evidence that Paxtonis the sister group of Gymnapogon,a taxon whose phylogeneticposition withinthe Apogonidaehas never been well understood.

N his comprehensive study of the osteology well-developed preopercular spine, the parhy- of cardinalfishes, Fraser (1972) followed pural fused to the lower hypural plate, and a Smith (1954) in recognizing the subfamily Pseu- single postcleithrum (Smith, 1954; Fraser, daminae for Pseudamia Bleeker, 1865, Pseuda- 1972). The Pseudaminae have not been diag- miopsSmith, 1954, and GymnapogonRegan, 1905 nosed cladistically, and some pseudamine spec- (including HenicichthysTanaka, 1915, Australa- imens continue to be misplaced in fish collec- and phia Whitley, 1936, Acanthapogon Fowler, tions, often being shelved with eleotrids or go- 1938). Fraser's (1972) classification of the biids. differed from Smith's in his group (1954) only Several years ago, the second author received of Lachneria as a recognition Smith, 1954, syn- from J. R. Paxton (Australian Museum, Sydney) of a close relation- onym Pseudamiops.Although four specimens of an unidentified percoid fish between Pseudamia and was ob- ship apogonids collected by trawl off northwestern Australia vious to Bleeker (1865), of relationships Gym- that were believed to represent an undescribed napogon have not always been so evident. As not- genus of grammistin serranid (Gloerfelt-Tarp ed by Fraser (1972, and references therein), the and Kailola, 1984; Sainsbury et al., 1984; Paxton genus has been placed in or associated with six et al., 1989). Johnson's [unpubl. abstract, 70th families: , Gymnapogonidae, Gobi- Annual Meeting of the American Society of Ich- idae, Henicichthyidae, Pomatomidae, and Tri- thyologists and Herpetologists (ASIH), 1990] chodontidae. Citing features of external anato- conclusion that the specimens represent a new my and osteology, Regan (1940) argued that genus of pseudamine was Gymnapogonis not closely related to apogonid completely pomatom- examination of ids (which at the time included but unexpected upon superficial Scombrops) them: unlike all other the north- rather to apogonids (formerly cheilodipterids). apogonids, western Australian have a continuous Smith (1954:776) first recognized that ". .. Gym- specimens dorsal fin. In the of the napogon Regan, 1905 ..., is clearly and closely process identifying spec- a number of lower was related to Pseudamia Bleeker. . . ." Fraser (1972: imens, large percoids which resulted in the dis- 32-33) also recognized similarities between examined, ultimately of derived characters useful in Gymnapogonand Pseudamia/Pseudamiopsbut not- covery diagnos- the Here we ed nevertheless that the "genus is not intimately ing Apogonidae (Johnson, 1993). describe the related to apogonids or other percoid fishes." external anatomy and osteology of Among the characters of pseudamines that have Paxton concilians,compare the new genus to oth- elicited questions regarding their alignment er apogonids, diagnose the Pseudaminae on the with the typical shallow-water cardinalfishes are basis of derived characters, provide evidence of a weakly scaled body (scales deciduous or ab- a pseudamine affinity for Paxton, and document sent), canine teeth in the upper and lower jaws, with derived characters a sister-group relation- a reduced number of supraneurals and epurals, ship between Paxton and Gymnapogon.A phylo- and an elongate neurocranium (Fraser, 1972). genetic analysis that includes all pseudamine Gymnapogonis further specialized in having a species is in progress.

? 1999 by the American Society of Ichthyologists and Herpetologists BALDWIN AND JOHNSON-NEW PSEUDAMINE APOGONID GENUS 1051

MATERIALS AND METHODS tinuous dorsal fin, dorsal fin rays VI,19; anal fin Institutional abbreviations follow Leviton et rays 1,15-16; caudal peduncle short and deep; entire of obscured skin; al. (1985). Terminology and abbreviations of margin preopercle by structures in the caudal skeleton follow distal margin of dorsal fin with band of dark Fujita neurocranium with a (1990), the intermusculars Patterson and John- pigment; "postfrontal" bone; fifth and sixth infraorbitals son (1995), and the pelvic girdle Stiassny and widely sepa- rated from rest of circumorbital series; third Moore (1992). The osteological description of a fifth P concilians is based primarily on examination epibranchial lacking toothplate; hypural fused to and of a single cleared-and-stained specimen, upper hypural plate; pelvic girdle anterior and CSIRO H.2254-1. Drawings were made with the lacking posterior processes. aid of a camera lucida. Scale bars in all illustra- concilians, new tions represent 1 mm. Typespecies.-Paxton species. In hypothesizing character polarity for the name is the surname of Apogonidae, numerous lower percoids were ex- Etymology.-The generic our friend and R. who amined (see Material Examined). In polarizing colleague John Paxton, the of the new We characters of the Pseudaminae, other - provided specimens species. are to honor with this and ids were considered the first outgroup and the pleased John gesture to him with a reminder same selection of lower percoids the second. provide good-natured that can't a fish its cover." Autapomorphic characters of Paxton were iden- "you judge by tified using other pseudamines as the first out- group and apogonines as the second, based on Paxton concilians, new species the hypotheses of apogonid relationships pro- Figures 1-2, Table 1 posed herein. Comparative cleared-and-stained apogonid Grammistidae gen. and sp. nov. Gloerfelt-Tarp and percoid material examined included the and Kailola, (1984):326. following: Acropomatidae-Acropoma sp. USNM 287444, 2 specimens; Synagrops bella USNM Grammistidae gen. and sp. nov. Sainsbury et al. 218918, 1; Apogonidae-Apogon lateralisUSNM (1984):338. 169600, 1; A. moluccensisUSNM 213380, 1; Apo- gonichthys ocellatus USNM 318066, 1; Apogonops Grammistidae gen. and sp. nov. Paxton et al. anomalous USNM 287447, 1; Archamiazosteropho- (1989):516. ra USNM 218801, 2; Astrapogonstellatus USNM 339373, 1; Cercamiacladara USNM 341821, 1; C. Holotype.-CSIRO H.1067-1. A 63.6-mm SL spec- eremiaUSNM 334713, 1; Cheilodipterusquinqueli- imen collected off northwestern Australia, neatus USNM 218806, 2; variegata northeast of Monte Bello Islands, 20007.2'S, USNM 218804, 2; aprion USNM 115051.7'E to 20007.1'S, 115050.0'E, 63-65 m, 212197, 1; Gymnapogonannona USNM 205671, Frank and Bryce demersal trawl, FRV SOELA,9 1; G. philippinus USNM 268313, 4; 262171, 4; G. October 1987. urospilotusUSNM 288032, 1; Gymnapogonsp. no data, 3; octospina USNM 203779, 1; Paratypes.-All from northwestern Australia: conklini USNM 327788, 9; Pseudamia CSIRO H.2254-1, 49.0 mm SL (cleared and gelatinosa USNM 218802, 1; P hayashii USNM stained), northeast of Monte Bello Islands and USNM 268333, 1; Pseudamiops gracilicauda northwest of Dampier Archipelago, 20o13.8'S, 324809, 1; P pellucidus USNM 268330, 2; Rhab- 116013.4'E to 20012.9'S, 116013.9'E, 46-47 m, damia USNM gracilis 213041, 2; Centropomi- Frank and Bryce demersal trawl, FRV SOELA,30 USNM dae-Centropomus parallelus 306578, 1; September 1987; AMS 1.22672001, 75.5 mm SL, USNM Epigonidae-Brinkmanella elongata northeast of Monte Bello Islands, 19055'S, USNM Incertae 206944, 1; Epigonus sp. 218803, 1; 115059'E to 19054'S, 115057'E, 80 m, demersal USNM 306589, 1; cedis--Howella sp. Polyprioni- trawl, FRV COURAGEOUS,21 July 1979; WAM americanus USNM dae-Polyprion 269542, 1; P.27219035, 66.5 mm SL, Hummock Island, Scombropidae-Scombrops boopsUSNM 049933, Abrolhos, 28048'S, 113003'E, D. Heald, 22 No- 1; Serranidae-Plectropomus maculatus USNM vember 1980. 218818, 1; Symphysanodontidae-Symphysano- don berryiUSNM 269465, 1. Description.-Counts and measurements of ho- lotype, followed parenthetically by those of par- Paxton, nov. gen. atypes if different from holotype: Dorsal fin rays Diagnosis.-A pseudamine apogonid uniquely VI,19 (one paratype with 15 dorsal soft rays, but possessing each of the following features: a con- pterygiophores of posterior end of fin abnor- 1052 COPEIA, 1999, NO. 4

Fig. 1. Holotype of Paxtonconcilians, CSIRO H.1067-1, 63.6 mm SL. Photographby Don Hurlbert,National Museum of NaturalHistory. mal in shape and configuration and thus this on first arch 11, 1 on upper limb, 10 on lower count presumably anomalous); anal fin rays 1,15 limb; all paratypes also with 1-3 rudimentary (two paratypes with 1,16); pectoral fin rays 18 rakers on upper limb or divided between upper (one paratype with 17 on right side); principal and lower limbs; branchiostegal rays 7; verte- caudal fin rays 9 + 8, branched; procurrent cau- brae 10 + 14 = 24; no supraneurals. dal fin rays 10 + 9, unbranched; total gillrakers

TABLE1. PROPORTIONALMEASUREMENTS OFTYPE SPEC- IMENSOF Paxton EXPRESSEDAS A PERCENTAGEOF THE A STANDARDLENGTH.

Holotype Paratypes CSIRO AMS WAM H.1067-01 1.22672001 P.27219035 Standardlength (mm) 63.6 75.5 66.5 Body depth 27 28 28 Body width 12 13 14 Head length 40 41 41 Snout length 7.2 7.3 9.0 Orbit diameter 6.3 6.4 6.5 Interorbitalwidth 7.7 7.4 8.1 Caudal peduncle depth 10 12 12 Caudal peduncle length 11 11 12 B Predorsallength 42 40 41 Preanus length 63 61 57 Length of first dorsal spine 3.6 4.4 4.5 Length of longest dorsal spine 5.8 7.5 7.6 Length of first dorsal ray 9.0 10 11 Length of anal spine 4.1 5.6 6.5 Fig. 2. Details of head morphology in (A) Paxton Length of first anal ray 7.2 8.1 9.8 concilians,Paratype, AMS P.27219035, 66.5 mm SL; Length of longest anal and (B) Gymnapogonsp., USNM 262166, 32 mm SL. ray 16 14 15 Lines of sensory papillae drawn as a composite from Length of caudal fin 24 22 24 left and sides of right illustratedspecimen and from Length of pectoral fin 20 18 20 examination of other available when nec- specimens Length of pelvic fin 15 14 13 essary. BALDWIN AND JOHNSON-NEW PSEUDAMINE APOGONID GENUS 1053

Body elongate, depth at origin of dorsal fin Pelvic fin short, 6.6 (7.1-7.8) in SL, reaching 3.7 (3.5-3.6) in SL, and compressed, width 2.2 only several millimeters beyond base of pectoral (2.0-2.2) in depth. Head length 2.5 in SL, orbit fin. diameter 6.3 in head. Snout short, rounded, Color in life: body bright yellow; all fins pale length 5.5 (4.9-5.7) in head. Interorbital flat, except dorsal margin of dorsal fin with wide the least width 5.1 (5.0-5.4) in head. Caudal pe- band of dark brown pigment, and dorsal mar- duncle short, length 9.1 (8.7-9.1) in SL; caudal gin of caudal fin dusky; abdomen and cheek sil- peduncle depth slightly less than length in ho- very. Color of holotype in alcohol: entire body lotype, 9.6 in SL (depth slightly greater than pale tan except dorsal fin with band of dark pig- length in paratypes, 8.4-8.5 in SL). ment along outer margin of entire fin; gut Mouth oblique, anteriorly forming an angle dusky. Dense patch of spots preceding dorsal fin of about 200 with the horizontal, but angle in- and extending along anterior edge of mem- creasing posteriorly as mouth curves ventrally. brane of first dorsal fin spine; 17th dorsal fin Lower jaw projecting slightly beyond upper jaw; ray with tiny bit of pigment distally (none in upper jaw long, extending posteriorly to vertical paratypes); last two dorsal fin soft rays unpig- well beyond posterior edge of orbit. Upper mar- mented (last with small bit of pigment on pos- gin of maxilla obscured by lacrimal for part of terior margin of membrane, about half way up its length. Upper and lower jaws each with length of ray). Many scattered spots on dorsal about 3-5 rows of small conical teeth; premax- portion of caudal peduncle, extending onto illa with several larger, slightly recurved canine dorsal margin of anterior part of caudal fin. teeth anteriorly, and dentary also with several Sparse pigment on distal ends of uppermost half enlarged canine teeth, the largest about principal caudal fin rays (absent in WAM para- surface. No way back along the tooth-bearing type). A few scattered spots on ventral portion supramaxilla. of caudal peduncle and posteriormost ventral Anterior nostril in short tube with small very procurrent rays (both absent in WAM para- flap posteriorly; anterior nostril located just pos- type). terior to upper lip. Posterior nostril situated at anterior of above center edge orbit, just slightly can be from of nostril ovoid and the Comparisons.-Paxton distinguished eye; posterior large, all other on the basis of its continu- axis one-half di- apogonids greatest approximately pupil ous dorsal fin and in fewer dorsal and ameter. recessed; cornea thick and having Eye slightly anal fin and vs whitish. spines (VI I, respectively, VII-IX and II, respectively, in other members of the Opercle with wide fleshy posterior margin family), more dorsal fin soft rays (usually 19 vs that ends posterodorsally in pointed flap. Mar- 7-13), and usually more anal fin soft rays [15 or gin of preopercle covered by skin, not exposed; 16 vs 7-13, except the Archamiawith a long, fleshy, bluntly pointed flap extending apogonine from ventral limb. 13-18 (Fraser, 1972)]. Additionally, the caudal posteriorly preopercular in Paxton is short and sometimes Head and body without scales but with nu- peduncle very than versus 1.3-2.0 merous papillar neuromasts. Larger papillae deeper long approximately times as as in most as forming rows that radiate from orbit. Rows of long deep apogonines, much as twice as as in Pseudamia papillae generally crossing other rows at about long deep et and Re- right angles, and thus a cross-hatch pattern vis- (Randall al., 1985) Gymnapogon(e.g., and three times as ible on many areas of head and body. gan, 1905; Lachner, 1953), as in Dorsal fin originating above posterior tip of long deep Pseudamiops (Lachner, 1953; the entire of the opercular flap. Spinous and soft dorsal fins con- Smith, 1954). Finally, margin is obscured skin in tinuous, a gap between ultimate and penulti- preopercle by Paxton (Fig. mate dorsal fin spines lacking. Spinous dorsal 2A), whereas the entire margin is exposed in fin low, slightly lower than soft dorsal fin; first apogonines, Pseudamia, and Pseudamiops, and dorsal fin spine 11 (9.3-10) in head, longest the dorsal margin of the preopercle is exposed dorsal fin spine, the fifth, 6.8 (5.3-5.4) in head; in Gymnapogon(Fig. 2B). first dorsal fin soft ray 4.4 (3.8-3.9) in head. Paxton is most similar to Gymnapogon.Both First anal fin spine 5.5 (6.3-7.3) in head; longest lack scales, have a bluntly pointed fleshy flap anal fin soft ray (usually 9th-llth) 2.5 (2.6-2.9) extending posteriorly from the ventral margin in head. Caudal fin truncate (slightly round), of the preopercle (Fig. 2), and have little body greatest length 4.2 (4.2-4.5) in SL. Pectoral fin pigment. The Atlantic apogonine genus Phaeop- falling well short of anus, reaching vertical tyx also has a large fleshy preopercular flap (Fra- through approximately the sixth dorsal fin ser and Robins, 1970), but it is not present in spine; greatest pectoral fin length 5.1 (5.1-5.4). other apogonines or pseudamines (some apo- 1054 COPEIA, 1999, NO. 4 gonines have an unossified ventral preopercular moid, ventrally by the parasphenoid, and pos- margin; see Fraser and Struhsaker, 1991). teriorly almost entirely by the dorsal limb of the parasphenoid, although the sphenotic and pter- Etymology.-The specific name concilians is de- osphenoid form the posterodorsal corner. Pro- rived from the Latin concilio, to unite separate otic excluded from orbit, the dorsal limb of the parts into a whole, in reference to the single parasphenoid articulating directly with the pter- dorsal fin of the new species. osphenoid. Several pairs of bones of the post- orbital/otic region of the neurocranium artic- Remarks.-Paxton concilians is known only from ulating with one another through interdigitat- the northwestern coast of Western Australia. All ing sutures: sphenotic (ventral margin) and specimens were taken by trawl at depths of 46 prootic (anterodorsal margin), epioccipital (an- m or greater, whereas other pseudamines typi- terior margin) and pterotic (posterodorsal mar- cally inhabit crevices and caves in shallow la- gin), and exoccipital (anterior margin) and goons, estuaries, , tidal reefs, or coral pterotic (posterior margin). Intercalar elongate, rubble and usually are taken with ichthyocides. overlying portions of exoccipital, pterotic, and Possibly, small specimens of the new species live prootic; intercalar forming part of otic bulla, more cryptically and at shallower depths, as is but intercalary facet not on otic bulla. No high the case with Pseudamia nigra. Allen (1992) de- frontal or supraoccipital crests. scribed P nigra from 28 specimens, 13.5-49.6 A flat, triangular bone overlying posterior mm SL, all taken at depths less than 2 m and end of sphenotic, anterior end of pterotic, and at a locality approximately 8 km upstream from ventral tip of parietal; this bone clearly separate a river mouth; however, we have recently iden- from pterotic. C. Patterson (BMNH) examined tified two larger specimens (67 and 69 mm SL) the neurocranium of Paxton and suggested as P nigra (CSIRO H.3384-02 and H.3473-01) (pers. comm.) that the unidentified bone most that were taken by trawl at 27 m and 50 m, re- closely resembles the postfrontal of Taverne spectively, off the coast of northern or north- (1974), or intertemporal of Le Danois (1967) eastern Australia. Similarly, we have identified and Kershaw (1970). See Taverne (1974, figs. an 80-mm SL specimen of Pseudamia taken by 31-32) for an illustration of the postfrontal in trawl at 40-45 m in the Philippines as P hayashii Pantadon. It is not possible to tell from radio- whereas all 55 of whether a is in the (USNM 270295), specimens P. graphs "postfrontal" present hayashii listed in the original description of the holotype and other paratypes of Paxton. species are smaller (16.5-62 mm SL) and were collected with rotenone from caves or crevices Comparisons.-Fraser (1972) did not identify any or beneath ledges of corals in lagoonal and ex- feature of the neurocranium unique to apogon- posed outer reefs (Randall et al., 1985). Gym- ids, but he noted several distinctive characters napogon also apparently exhibits the same of the pseudamine braincase: the shape [Fraser depth/size relationship; for example, the larg- did not specify the diagnostic aspect of shape, est specimen of the genus in collections at but presumably he meant the elongate nature USNM (G. philippinus, 73 mm SL, USNM of the neurocranium in pseudamines; we deter- 332337) was taken in a trawl deployed to 80.5 mined that the length of the pseudamine neu- m. Other Gymnapogonin USNM collections, in- rocranium is about four times the greatest cluding many type specimens of the various spe- depth (Fig. 3A), vs length about 2-2.5 times cies, are smaller and were taken in shallow areas greatest depth in other apogonids (Fig. 3C)]; of coral debris. absence of a basisphenoid [present in many apogonines, but also reduced or absent in Glos- Osteology of Paxton samia, , , (Bentu- viaichthys), Neamia, ,, Sipham- Neurocranium.-Neurocranium (Fig. 3) elon- ia, Fowleria, and sometimes Phaeoptyx];interca- gate, length about four times greatest depth lary facet not included on otic bulla [also not (through otic bulla). Vomer with one very large included on otic bulla in Glossamia, , canine tooth on anterior margin (two in other and Apogon () among apogonines; see Fra- paratypes, and one large and one small in ho- ser, 1972, 1973]; and parasphenoid not separat- lotype) and a semicircular patch of smaller ed from pterosphenoid by prootic (unique to teeth posteriorly; dorsal margin of vomer flat pseudamines). Paxton has all of those pseuda- and separated from mesethmoid by large eth- mine features and an additional one not noted moid cartilage. Orbit narrow and bordered dor- by Fraser: the dorsal aspect of the vomer is near- sally by the frontal, anteriorly by the ethmoid ly flat and separated from the mesethmoid by a cartilage and posterodorsal tip of lateral eth- large ethmoid cartilage (Fig. 3A); in most other BALDWIN AND JOHNSON-NEW PSEUDAMINE APOGONID GENUS 1055

Frontal Parietalcipital Mesethmoid

Epioccipital 0 Itcl 000 .terosphenoid i 00 0 C, 0 0 00 Su Intercalar ••---i Parasphenoid Basioccipital Prootic Exoccipital

LateralE hoid ShntcSphenotic PsfotlP~oiP stfrontal Pterotic SLateralEthmoid Vomer

C.

...... co

. . S...... o )~o

Fig. 3. Neurocranium.(A), (B) Paxtonconcilians, CSIRO H.2254-1, 49.0 mm SL, left lateralview and dorsal view, respectively;(C) Phaeoptyxpigmentaria, 53.1 mm SL (from Fraser,1972:plate 36). apogonids, there is a dorsal projection on the pect of the vomer is not flat as it is in pseuda- vomer that articulates with the mesethmoid mines). Gymnapogonexhibits variation in this (e.g., as in Phaeoptyx;,Fig. 3C) or nearly does, in character. In G. annona (Fraser, 1972, plate 42) which case it is separated from the mesethmoid and G. urospilotus, the vomer is connected to by a small area of the ethmoid cartilage. Fraser some extent to the mesethmoid, whereas there (1972, plate 33) illustrated the apogonine Si- is no bony connection in G. philippinus, and the phamia with the vomer separate from the mes- condition is nearly identical to that of Paxton, ethmoid by cartilage (although the dorsal as- Pseudamia, and Pseudamiops.The intercalar also 1056 COPEIA, 1999, NO. 4

q9 106 tilage; anteriorly, edges of nasal bone curled dorsally forming shallow canal for passage of branch of cephalic lateral line.

Nasal Comparisons.-The shape of the nasal bone in Paxton, Gymnapogon, and Pseudamia is nearly identical, each having a somewhat broad lami- nar flange extending ventrally from the poste- rior half of the bone. The nasal is not so shaped in other in Lacrimal apogonids, including Pseudamiops, which it is a simple tube with an opening at 103 each end. 102

Fig. 4. Circumorbitaland nasal bones of Paxton Jaws, suspensorium, opercular series.-Premaxilla concilians,CSIRO H. 2254-1,49.0 mm SL, right lateral and dentary armed with numerous small coni- = view. IOn nth infraorbital. cal teeth and several larger canine teeth (Fig. 5A). Supramaxilla lacking. Hyomandibular ar- with but not me- seems to be more elongate in pseudamines, al- ticulating ventrally symplectic with two though it is quite long in some apogonines (see tapterygoid. Metapterygoid ventrally Fraser, 1972). We found no features of the neu- directed rami separated by laminar bone and bound to two directed rocranium unique to Paxton and any subgroup synchondrally dorsally rami of also laminar of pseudamines, and the presence of a postfron- quadrate separated by bone. Between the anterior and tal appears to be autapomorphic for Paxton. posterior syn- chondral joints, quadrate and metapterygoid Circumorbitalbones.-This series (Fig. 4) com- loosely sutured with interdigitating struts of prising six bones, the lacrimal much larger than laminar bone. the others. Third infraorbital tiny and not bear- Ectopterygoid and endopterygoid without ing a subocular shelf. Only the first and second teeth, palatine with long series of small conical infraorbitals contiguous; the third close to the teeth. Endopterygoid lying above dorsal limb of second and fourth but not in contact with them; ectopterygoid and giving rise posteriorly to a the fifth small, poorly ossified, and greatly sep- spine that extends toward the metapterygoid. arated from fourth below and sixth above. Edg- Opercle bearing a single spine on posterior es of all infraorbitals entire. margin; opercle poorly ossified dorsally and giv- ing rise posterodorsally to a long bony filament; Comparisons.-Paxton shares with other apogon- ventral part of opercle overlapping subopercle, ids and many percoids the presence of six in- the posterior ramus of which is also prolonged fraorbitals (epigonids, for example, have seven into a thin, posteriorly directed, bony filament. or eight; Fraser, 1972). Paxton, Pseudamia, and Preopercle with a single blunt spine on poste- Pseudamiopslack a subocular shelf on the third rior margin at a point about one-third the infraorbital. A subocular shelf is present in most length of the bone from its ventral terminus. apogonines but lacking in, for example, Rhab- Interopercle large, weakly ossified ventrally, and damia (Verulux), R. (Bentuviaichthys),Apogon (Zo- joined by a long ligament (aproximately half ramia)-(Fraser, 1972). Gymnapogonis the only the length of the long axis of the interopercle) pseudamine with a subocular shelf, but it has a to the retroarticular. small shelf compared to the well-developed shelf of most apogonines (and lower percoids). Comparisons.-Paxton (Fig. 5A) shares with most Paxton differs from all other apogonids in hav- other apogonids and some percoids a single ing the fifth and sixth infraorbitals widely sep- spine on the posterior opercular margin (e.g., arated from the rest of the series. as in Gymnapogon,Fig. 5B); opercle with deep indentations on either side of the spine in some Nasal bone.-A simple, relatively flat bone (Fig. taxa (e.g., as in Pseudamiops,Fig. 5C), creating 4) lying above dorsolateral surface of rostral car- the impression of three opercular spines. Paxton

Fig. 5. Jaws, suspensorium, and opercular series, right lateral view. (A) Paxton concilians, CSIRO H.2254-1, 49.0 mm SL; (B) Gymnapogonphilippinus, USNM 268314, 35.0 mm SL; (C) Pseudamiopspellucidus, USNM 268330, 37.0 mm SL. BALDWIN AND JOHNSON--NEW PSEUDAMINE APOGONID GENUS 1057

A Preopercle

OperclePm

Ectopterygoid Subopercl S p eMetapterygoid Endopterygoid

Interopercle Plain - ]E'~al~atine ;~ '

Symplectic 0 S~Maxilla Quadrate Premaxilla

Interoperculo-mandibular ligament

Articular Dentary - Anguloarticular

C 1058 COPEIA, 1999, NO. 4 also shares with other apogonids and many low- er percoids the extension of the subopercle pos- teriorly via a slender bony filament (Fig. 5A-C). Posterior With other pseudamines, Paxton shares en- Ceratohyal larged canine teeth in the lower jaw and at least some slightly enlarged teeth on the premaxilla (Fig. 5A-C). Among apogonines, slightly en- larged or canine teeth are found in Cheilodipte- Rhabdamia rus, Coranthus, (Bentuviaichthys), AnteriorCeratohyal Lachneratus, and Apogon affinis (Fraser, 1972, 1973; Fraser and Struhsaker, 1991), the relation- 6. Lower arch of Paxton ships of which to pseudamines are unclear. Fig. hyoid concilians, CSIROH.2254-1, 49.0 mm SL, medial view. The opercular series and suspensorium of right Paxton (Fig. 5A) and Gymnapogon(Fig. 5B) are remarkably similar. Those genera are unique among apogonids in having a long bony strut Comparisons.-Paxton shares with other apogon- from the dorsal part of the opercle and in hav- ids and many percoids the presence of seven ing a spine on the posterior margin of the preo- branchiostegals (one fewer on the anterior cer- percle. The preopercular spine is blunt in Pax- atohyal than in, e.g., beryciforms) and the ab- ton, slightly larger and more pointed in Gym- sence of a perforated anterior ceratohyal (ex- napogon, but it is located in the same position cept in Holapogon; Fraser, 1973). Paxton differs at a point approximately one-third the vertical from all apogonids (except Glossamiaaccording length of the preopercle from the ventral end to Fraser, 1972) in having a small bit of bony in both genera (the spine is visible externally in suturing between the medial surfaces of the an- Gymnapogon,Fig. 2B, unexposed in Paxton, Fig. terior and posterior ceratohyals. 2A). The posterior preopercular margin in apo- gonines may be smooth or serrate (Fraser, Gill arches.-Gill arches comprising three ossi- 1972), and Cercamiamay have 2-5 short spines fied basibranchials, the third greatly exceeding (Randall and Smith, 1988; Hayashi, 1991), but the first and second in length, a small fourth a single spine in the position described above basibranchial cartilage, three hypobranchials, for Gymnapogonand Paxton is unique to those five ceratobranchials, four epibranchials, two genera. A striking similarity between Paxton and pharyngobranchials, a fourth upper pharyngeal Gymnapogonis the articulation of the quadrate toothplate, and an interarcual cartilage between and metapterygoid. Rather than articulating at the dorsal portion of the first and second arches a single synchondral joint as in other apogonids (Fig. 7). First and fourth pharyngobranchials (e.g., as in Pseudamiops,Fig. 5C), the quadrate lacking. Gillrakers (distinguished from gill and metapterygoid articulate at two separate plates or rudimentary rakers by being longer synchondral joints, one anteriorly and one pos- than wide) present on the first hypobranchial, teriorly. Between these anterior and posterior first and second ceratobranchials, and first and joints is bone, sometimes with weakly (some spe- sometimes second epibranchials; gill plates cies of Gymnapogon)or strongly (Paxton) devel- present on all three hypobranchials, first oped struts that form an interdigitating suture. through fourth ceratobranchials, and first Gymnapogonalso shares with Paxton a similarly through third epibranchials (Fig. 7; gillrakers shaped endopterygoid characterized by a pos- and plates on epibranchials not illustrated). teriorly directed spine and a significantly longer First epibranchial terminating distally as a interoperculohyoid ligament than in other apo- long thin cartilage but no separate first phar- gonids [this ligament 40-50% the length of the yngobranchial present (Fig. 8A). A very long, long axis of the interopercle in Paxton and Gym- slender interarcual cartilage extending from un- napogon, < 25% in other apogonids (Fig. 5A- cinate process of first epibranchial to anterior C)]. tip of second pharyngobranchial. Second phar- yngobranchial with a single cartilaginous con- Lower hyoid arch.-Anterior and posterior cera- dyle that articulates with the interarcual carti- tohyals sutured to one another on medial sur- lage (Fig. 8A). No direct connection between face by several bony interdigitating struts; ante- second pharyngobranchial and second epibran- rior ceratohyal supporting five branchiostegals, chial (Fig. 8A-B), the latter medially articulat- posterior ceratohyal two; anterior ceratohyal ing only with a laterally directed cartilaginous not perforated (McAllister, 1968), and upper condyle of the third pharyngobranchial (Fig. margin of the bone nearly straight (Fig. 6). 8A). Medial end of third epibranchial greatly BALDWIN AND JOHNSON-NEW PSEUDAMINE APOGONID GENUS 1059

cc? !'0'6 C

Hypobranch~ials 1-3

'Cl "Soo ;4 _44,. "' X~ a?000,lo~I~??` ?x:-? sa~igf~?c ~ ~ '00

Fig. 7. Branchialskeleton of Paxtonconcilians, CSIRO H.2254-1, 49.0 mm SL, dorsalview, dorsal gill arches removed from right side. expanded where it articulates with a very large gonines have a second cartilage-tipped process third pharyngobranchial. Fourth epibranchial that terminates in close proximity to the carti- slender (Figs. 7, 8A). Second and third phar- laginous head of the third pharyngobranchial yngobranchials bearing well-developed fused (e.g., Apogonichthys,Fig. 8E, and Pseudamia Fig. toothplates ventrally, the third pharyngobran- 8D). chial very large (Fig. 8A). A small fourth upper Fraser (1972) noted that the first pharyngo- but fourth pharyngeal toothplate present, phar- branchial is poorly ossified in Gymnapogonand Fifth cerato- yngobranchial cartilage lacking. lacking in Apogonichthys,Fowleria, and Astrapo- branchial of small bearing long patch teeth, gon. In our specimens, the first pharyngobran- about seven rows of teeth at width greatest (Fig. chial is lacking in Gymnapogon (Fig. 8C) but 7). Epibranchial toothplates lacking. present and ossified in . In Apogonichthys(Fig. 8E) and Fowleria (Fig. 8F), Comparisons.-The dorsal gill arches of Paxton there is usually no separate first pharyngobran- resemble those of Gymnapogonin lacking a first chial (one specimen of Fowleriaexamined has a pharyngobranchial and in having the medial separate first pharyngobranchial on one side), end of the first epibranchial terminating as an but within an elongate blue-stained cartilage ex- attenuated cartilage (Fig. 8A,C). Additionally, tending from the tip of the first epibranchial in the interarcual cartilage is very long and slender those taxa, there is a more blue-stained in both genera (also somewhat in darkly elongate nubbin in and a red- some apogonines, e.g., Apogonichthys,Fig. 8E, cartilaginous Apogonichthys stained ossified rod in Fowleria. relative to the very short interarcual cartilage in Presumably these structures are the first other apogonids, e.g., Pseudamia, Fig. 8D), and developing phar- and thus in and gillrakers are present on the first and second yngobranchials, Apogonichthys Fowleria the first arches (vs on only the first in all other apogon- pharyngobranchial apparently ids examined except on first and second arches develops within the cartilaginous tip of the first in , Astrapogon, and ). Finally, epibranchial rather than separately as is typical Paxton and Gymnapogonhave a single cartilage- for perciforms. A more comprehensive survey tipped process at the distal end of the second of the ontogeny of dorsal gill arches in apogon- pharyngobranchial that articulates with the in- ids is needed but is beyond the scope of this terarcual cartilage (Fig. 8A,C). In addition to study. Of relevance here is that there is no evi- this process, other pseudamines and many apo- dence of development or ossification of a first 1060 COPEIA, 1999, NO. 4 pharyngobranchial within the cartilage extend- pterygiophore bearing soft ray in supernumer- ing from the first epibranchial in Paxton and ary association); Figure 9. Gymnapogon(Fig. 8A,C), and we interpret the Dorsal fin rays modally VI,19 (see Description condition in those genera as distinct from the above), last spine not separated from penulti- conditions in Apogonichthysand Fowleria;that is, mate spine by a gap externally. First dorsal fin Paxton and Gymnapogon uniquely lack a first pterygiophore bearing one spine in supernu- pharyngobranchial. Paxton is unique among merary association. Each spine-bearing proxi- apogonids in lacking a toothplate fused to the mal-middle pterygiophore lacking posteriorly third epibranchial (Fig. 8A). projecting shelf for articulation with its serially associated distal radial and thus separated from Postcranial axial skeleton.-Twenty-four vertebrae radial by large gap (Fig. 9). Sixth dorsal fin pter- (10 + 14), usually eight pairs of ribs (on verte- ygiophore lacking serially associated spine, but brae 3 through 10), and eight pairs of epineural a small, free cartilaginous distal radial present. bones (Fig. 9); holotype with seven pairs of ribs, A very small, free cartilaginous element in an- the series ending on vertebra 9; epineurals orig- terior portion of spinous dorsal fin preceding inating from neural arches on first and second distal radial serially associated with first ptery- vertebrae, but origin descending to rib on third giophore. and fourth vertebrae; remaining epineurals free Anal fin comprising one spine and 15 or 16 from axial skeleton. Origin of ribs descending soft rays. One supernumerary spine present on from dorsolateral portion of centrum on third first pterygiophore (Fig. 10). Cartilaginous stay and fourth vertebrae to ventrolateral region of present at posterior base of both dorsal fin and centrum on fifth, and to medial surface of par- anal fin. apophysis on the sixth through ninth; rib on 10th vertebra free from axial skeleton, originat- Comparisons.-Paxton (Fig. 9) shares with most ing ventral to ventral tip of parapophysis. Ribs apogonids (e.g., Gymnapogon,Fig. 11A) the ab- preformed in cartilage, as indicated by their car- sence of a spine serially associated with the sixth ventral All neural arches tilaginous tips. fused dorsal fin pterygiophore [this spine present as to and left and sides of all neural centra, right a nubbin or tiny spine in some apogonines, e.g., fused. spines Apogon, Apogonichthys, Astrapogon (Fig. 11B), Fowleria, Holapogon, and Vincentia; fully devel- Comparisons.-Fraser (1972) considered a re- oped in e.g., Neamia (Fraser, 1972, 1973; this duced number (2-7) of epineurals (his epi- study)]. The missing spine in Paxton does not pleurals) as a character uniting pseudamines, result in an external gap in the dorsal fin as it but Paxton has eight, as do many apogonines, does in other apogonids for two reasons: (1) the and several apogonine genera have seven. Cer- sixth and seventh pterygiophores in Paxton are camia and Pseudamiops appear to be unique much closer to one another than in, for exam- among apogonids in having only two pairs of ple, Gymnapogon,where a wide gap between epineural bones (Fraser, 1972; Randall and those pterygiophores internally corresponds to Smith, 1988). a wide gap between the sixth and seventh dorsal fin spines externally (Fig. 11A); and (2) the Median fins.-Dorsal fin not preceded by supra- sixth and seventh pterygiophores lack long pos- neurals. "Predorsal pattern" of Ahlstrom et al. terior shelves, such as those in, for example, As- (1976): //1/1+1/1/1/1/I+s/s+s/ etc., (where trapogon (Fig. 11B). Although the main bodies / = neural spine, 1 = pterygiophore bearing of the sixth and seventh pterygiophores are in spine in supernumerary association, I = ptery- close proximity in Astrapogon and many other giophore with no supernumerary element, s = apogonids, the long posterior shelves on the

Fig. 8. Dorsal gill arches. (A) Paxtonconcilians, CSIRO H.2254-1, 49.0 mm SL, left side, ventral view; (B) Paxtonconcilians, CSIRO H.2254-1, 49.0 mm SL, left side, dorsal view, the second epibranchialrotated coun- terclockwise;note that the second pharyngobranchialoverlaps but does not articulatewith distalend of second epibranchial; (C) Gymnapogonphilippinus, USNM 268314, 35.0 mm SL, left side, ventral view; (D) Pseudamia gelatinosa,USNM 218802, 53.0 mm SL, right side (drawingreversed), ventral view; (E) Apogonichthysocellatus, USNM 318066, 36 mm SL, left side, ventralview; (F) Fowleriavariegatus, USNM 218804, 40.5 mm SL, left side, ventralview. Ebn = nth epibranchial,Pbn = nth pharyngobranchial,UP4 = fourth upper pharyngealtoothplate, lac = interarcualcartilage. BALDWIN AND JOHNSON-NEW PSEUDAMINE APOGONID GENUS 1061

Pb3 B A Pb2

UP4 ac -Pb,2

lac

Ebi Eb, Eb1 b b E, b 0 D 000"

. 0 000 01 0 0 0 0

C Pbj

00`3 0 E0 00 0 EEbF

0.0.

0O

E ~.~00 .10 Deveopin DevlopI A b

on

Pb * ?AlDeeloin

Deve lopi\:1Va

JJ? 0o 1062 COPEIA, 1999, NO. 4

Free Distal Radial

Proximal-middleElement

Epineural

iib First Anal-fin Pterygiophore Fig. 9. Anterior portion of axial and dorsal fin skeletons of Paxtonconcilians, CSIRO H.2254-1, 49.0 mm SL, right lateralview.

sixth and seventh pterygiophores result in a dorsal spines nearly subequal and longest (Fig. large gap between the spine in supernumerary 9); in other apogonids, the penultimate spine is association with the sixth pterygiophore (the typically smaller than the preceding and ulti- penultimate spine) and the spine serially asso- mate spines (Fig. 11). ciated with the seventh pterygiophore (the ul- Paxton and other pseudamines have a gap be- timate spine). tween the proximal-middle and distal elements Paxton also differs from other apogonids in of at least some spine-bearing pterygiophores having only six dorsal fin spines, the seventh (Figs. 9, 11A), whereas those elements are con- pterygiophore serially supporting a soft ray rath- tiguous in most apogonines (e.g., Astrapogon, er than a spine. Finally, it is unique within the Fig. 11B). In Paxton, the proximal-middle ele- family in having the fourth through last (sixth) ments of all spine-bearing pterygiophores are separated from their distal radials by a gap (Fig. 9). The proximal-middle element of each of the anteriormost two or three pterygiophores may be in contact with its distal radial in Gymnapogon (Fig. 11A), Pseudamia, and Pseudamiops,but fur- ther posteriorly, proximal-middle elements are always well separated from their distal radials (when present; distal radials are lost or greatly reduced posteriorly in Pseudamia and Pseuda- miops). Among apogonines, we have observed separation of proximal-middle and distal ele- Anal-fin Pterygiophorements only in Apogonichthysand Fowleria, the gaps occurring on the fourth through sixth pterygiophores. Paxton shares with Gymnapogonthe absence of Anal-finSpine supraneurals. Fraser (1972) noted that Astrapo- gon lacks supraneurals, but in our cleared-and- stained specimen of A. stellatus, a single ossified Fig. 10. Anterior elements of anal fin of Paxton supraneural is present (Fig. 1iB), suggesting concilians,CSIRO H.2254-1, 49.0 mm SL, right lateral that there may be some variation within the ge- view. nus. Supraneurals also are lacking in Cercamia BALDWIN AND JOHNSON-NEW PSEUDAMINE APOGONID GENUS 1063

A DistnlRadial

Proximal-middleElement

Supraneural 00{

Fig. 11. Anteriorelements of dorsal fin, right lateralview. (A) Gymnapogonphilippinus, USNM 268314, 35.0 mm SL; (B) Astrapogonstellatus, USNM 339373, 41.0 mm SL.

(Randall and Smith, 1988; Hayashi, 1991) and tospina) have both anal fin spines in supernu- Lachneratus (Fraser and Struhsaker, 1991). merary association with the first pterygiophore, Paxton shares with all apogonids a reduction but in those taxa, the first pterygiophore is a in number of anal fin spines from the primitive large, apparently compound structure presum- percoid complement of three; however, it is ably resulting from fusion of the first and sec- unique among apogonids in having a single ond pterygiophores. Paxton also has more anal anal fin spine. (Allen, 1992, reported 1,10 for fin soft rays than most apogonids [15-16 vs. 7- Pseudamia nigra, but specimens examined in this 13 in all apogonid genera except Archamiawith study have 11,9.) Most apogonids, including oth- 12-18 (Fraser, 1972)]. er pseudamines, have two anal fin spines, one in supernumerary association with the first pter- Caudal skeleton.-Principal caudal fin rays 17 (9 ygiophore, the second associated serially. Al- + 8), all branched, segmented, and articulating though the first anal fin element in Paxton is a with hypural plates (Fig. 12A). Procurrent rays spine in supernumerary association with the 10 + 9, all unbranched and articulating with first pterygiophore, the serially associated ele- epurals, haemal spine of second preural cen- ment is a soft ray (Fig. 10). Johnson (1984) not- trum (PU2), or caudal cartilages; procurrent ed that the presence of two anal fin spines in rays not articulating with neural and haemal apogonids and epigonids is not indicative of a spines of third preural centrum (PU3), which close relationship between those families be- are enveloped distally by cartilage. Posterior- cause in epigonids the two spines are both in most three or four procurrent rays of upper and supernumerary association with the first ptery- lower caudal fin lobes segmented. Contralateral giophore (i.e., the third spine of primitive per- hemitrichs of procurrent rays not fused to one coids fails to transform from a ray); in apogon- another. Two epurals and two hypural plates, ids, only the first spine is in supernumerary as- the lower plate autogenous and comprising the sociation with the first pterygiophore (i.e., the fused parhypural + hypurals 1 and 2, the upper first supernumerary spine of primitive percoids fused to terminal centrum and including fused has been lost). We note that a few apogonids hypurals 3 and 4 (and 5?). A separate hypural (e.g., Apogon moluccensis,A. lateralis, Neamia oc- 5 lacking. Uroneurals absent. The interneural- 1064 COPEIA, 1999, NO. 4

of fourth centrum A spine cartilages preural and fused CieP2 cPeP (PU4) PU3 well developed, distally, Epural and enveloping distal third of neural spine of CinPu3 PU3. An additional tiny round interneural-spine cartilage present along posterior edge of neural spine of PUMat posteroventral tip of fused third ~NS and fourth interneural-spine cartilages. A single H3+H4+ postepural 2 cartilage and two interepural 2 car- H5+UC o~ - tilages present. Interhaemal-spine cartilages of PU4 and PU3 large, separate but abutting one Ph+H+2 another distally and enveloping distal third of haemal spine of PU3. One posthaemal-spine car- tilage and two interhaemal-spine cartilages of Cih4 PU2. o/ 0 cihPu2 CphPu C Comparisons.-Paxton (Fig. 12A) and other pseu- damines (Fig. 12B,C) have two epurals; all other SCPeP2 B CPeP Epural Rhabdamia and EpuralC apogonids except (Verulux) s some have three (Fraser, 1972). Among apogonids, Paxton (Fig. 12A) and Gym- napogon (Fig. 12B) are unique in having (1) the parhypural fused with hypurals 1 and 2, and (2) the upper but not lower hypural plate fused to the centrum. Those also have H 3+H4+UC urostylar genera the neural and haemal spines of PU3 shortened and surrounded distally by cartilage; in Pseuda- , •v:.Ph+HI+H2 CinPu, mia (Fig. 12C) and most other apogonids, the neural and haemal spines of PU3 are not short- ened or surrounded distally by cartilage. (Pseu- damiops has an intermediate condition in which the spines are not noticeably shortened, but they are surrounded distally by a thin layer of cartilage.) Paxton is unique among apogonids in ciPu CihPu lacking a separate hypural 5.

Epural Pectoralgirdle.-Extrascapulars lacking. Posttem- poral forked anteriorly, upper limb articulating with lower limb with intercalar. Pos- H4 epioccipital, terior margin of posttemporal mostly smooth, H3 ventralmost portion overlying dorsal tip of su- pracleithrum (Fig. 13A). Supracleithrum a sim- mI+H2 Ph ple elongate bone overlying anterodorsal part of cleithrum. Cleithrum with dorsally directed projection at anterodorsal corner. Cleithrum fan-shaped posterodorsally and narrowing ven- where it covers anterior of CihPu4 trally margins scap- Cih ula and coracoid. Scapula approximately square 3 and bearing a central fossa and posterodorsal facet for articulation of first pectoral fin ray. Fig. 12. Caudal skeleton. (A) Paxton concilians, Coracoid roughly rectangular, ventral half com- CSIROH.2254-1, 49.0 mm SL, left side (drawingre- prising two diverging bony stems of approxi- USNM versed); (B) Gymnapogonphilippinus, 268314, mately equal length that are flanked on all sides 33.0 mm SL, side; (C) Pseudamia right gelatinosa, by laminar bone; one stem projecting antero- USNM 218802, 53.0 mm SL, right side. CiePn = in- terepuraln cartilage,CihPun = interhaemalspine car- tilage of nth preural vertebra,CinPun = interneural spine cartilageof nth preuralvertebra, CPePn = pos- tepural n cartilage, CphPun = posthaemal spine car- = haemal spine, NS = neural spine, Ph = parhypural, tilage of nth preuralvertebrae, Hn = nth hypural,HS PUn = nth preuralvertebra, UC = urostylarcentrum. BALDWIN AND JOHNSON--NEW PSEUDAMINE APOGONID GENUS 1065

A Posttemporal B

Supraclei

Scapula Cleithrum

Coracoid Postcleithrum

- D~z

.60 ?

Fig. 13. Pectoral girdle, left side. (A) Paxton concilians,CSIRO H.2254-1, 49.0 mm SL; (B) Gymnapogon urospilotus,USNM 288032, 35.5 mm SL; (C) Pseudamiopspellucidus, USNM 268330, 37.0 mm SL; (D) Fowleria variegatus,USNM 218804, 46.0 mm SL. 1066 COPEIA, 1999, NO. 4

and with cleithrum a ventrally articulating (at CentralProcess point approximately one-third the length of cleithrum from its ventral tip), other stem di- rected ventrally and tipped in cartilage. Cora- coid separated dorsally from scapula by carti- lage that continues ventrally along posterior margin of coracoid. Scapula and coracoid sup- ExternalDorsal Win porting four large radials. Pectoral fin compris- ing 18 (rarely 17) rays, lateral and medial heads of each, except dorsalmost one and ventralmost two, embracing small cartilaginous distal radial. Dorsalmost ray embracing ovoid cartilage, the MedialProcess ontogenetic origin of which Johnson and Broth- ers (1993:463) described as equivocal. As noted h those in this carti- A authors, teleosts, a 0 by typically a lage eventually fuses to the base of the medial 0 0 000 half of the first ray, ossifying as a facet that ar- ticulates with the scapular condyle. Single, slen- der, elongate postcleithrum present.

Comparisons.-Paxton (Fig. 13A) shares with Fig. 14. Pelvic girdle of Paxtonconcilians, CSIRO 49.0 mm SL. pseudamines (e.g., Gymnapogon and Pseuda- H.2254-1, miops, Fig. 13B-C) a ventrally directed stem on the posteroventral portion of the coracoid. In margins of central processes between bases of other apogonids (e.g., Fowleria,Fig. 13D), there pelvic fin spines. is no distinct coracoid stem posteroventrally, al- there be a or though may posteriorly postero- shares with other The location of the Comparisons.-Paxton pseuda- dorsally pointed projection. mines a reduced external ventral and coracoid articulation with the cleithrum also very wing thus the pelvic girdle has little three-dimension- supports a pseudamine affinity for Paxton. In al form. Paxton shares with Gymnapogonthe lat- most apogonids, the coracoid articulates at the erally directed spine near the posterior end of ventral tip of the cleithrum or slightly more dor- the external dorsal wing. Paxton is unique sally (up to a point approximately one-quarter among apogonids in lacking anterior and pos- the length of the cleithrum from the ventral tip, terior processes and in having the median wish- Fig. 13D), whereas in Paxton and pseudamines, bone-shaped cartilage, a feature we have not ob- the coracoid articulates at a point one-third or served in any other fishes. more the length of the cleithrum from the ven- tral tip (Fig. 13A-C). Paxton and Gymnapogon are unique among apogonids in having a single RELATIONSHIPS postcleithrum. In his comparative osteological treatment of the Apogonidae, Fraser (1972) recognized Pelvic girdle.-Pelvic girdle comprising paired three subfamilies, the (apogonids basipterygia, each supporting on its posterior proper, i.e., primarily shallow-water reef forms), margin one spine and five soft rays (Fig. 14). Pseudaminae (small to minute elongate fishes The two halves of girdle in close proximity pos- often with translucent or transparent bodies teriorly, where medial processes approach one with slimy or gelatinous skin, feeble or no squa- another at ventral midline; bony interdigitating mation, conspicuous intersecting lines of papil- suture as described by Stiassny and Moore lae on the head and body, etc.), and Epigoninae (1992:217) for higher percomorphs lacking. (relatively deep-water forms that differ from Central processes well separated from one an- apogonines and pseudamines in having, e.g., other anteriorly. Small laterally directed spine two pairs of uroneurals and 25 vertebrae). Fra- present on external dorsal wing of central pro- ser hypothesized that epigonines are primitive cess near its origin posteriorly. Little three-di- with respect to apogonines and pseudamines, mensional structure to girdle because external and Johnson (1984) subsequently removed epi- ventral wing almost imperceptible. Anterior and gonines from the Apogonidae. Johnson sug- posterior processes lacking. Median wishbone- gested that Brinkmanella, Epigonus, Florenciella, shaped cartilage lying just posterior to posterior Rosenblattia,and Sphyraenopsconstitute a distinct BALDWIN AND JOHNSON-NEW PSEUDAMINE APOGONID GENUS 1067 family, the Epigonidae, characterized by several swimbladder with anterodorsal oval (see Fraser, specializations of the jaws, suspensorium, and 1972, plate 1). The number of anal fin spines infraorbitals. Johnson (1993) provided further has increased and decreased a number of times evidence that epigonids are not closely related in the Percoidei from the primitive condition of to apogonids by identifying synapomorphies of three, but the presence of two spines is never- apogonines + pseudamines that are lacking in theless uncommon in the group (Johnson, epigonids: (1) absence of a direct articulation 1984, table 120). Furthermore, some groups between the second epibranchial and second that have two spines (e.g., epigonids, sciaenids, pharyngobranchial (Fig. 8); (2) expanded third Kurtus) have both spines in supernumerary as- epibranchial and narrow fourth epibranchial sociation with the first pterygiophore rather (Fig. 8); (3) small fourth upper pharyngeal than one (Johnson, 1984, but see note about toothplate lacking a fourth pharyngobranchial anal fin spines in comparisons section of Me- cartilage (Fig. 8); (4) very short distal radial as- dian Fins above). Nearly all apogonines and sociated with last dorsal fin spine, the serially pseudamines for which Fraser (1972) examined associated proximal-middle element nearly con- swimbladders have the oval at the anterodorsal tacting last spine (Fig. 1 lA-B); and (5) horizon- end, whereas epigonids have the oval postero- tal and vertical rows of superficial neuromasts dorsally. Fraser (1972) considered the latter to on head and body (Fig. 2; see also Smith, 1954, be the primitive state for his Apogonidae, but fig. 3; Hayashi, 1991, fig. 11). further investigation of percoid swimbladder Although rare among percoids examined, morphology is needed to corroborate this hy- some of those features are present elsewhere: pothesis. A vertebral count of 10 + 14 and ab- Epigonus is superficially similar to apogonids in sence of an axillary scale at the base of the pec- having a relatively large third epibranchial and toral fin are common percoid features and thus slender fourth; Scombropshas a reduced distal of little value in defining the Apogonidae. radial associated with the last dorsal fin spine; Despite lacking one of the apogonid hall- Pempheris and Glaucosoma lack an articulation marks, separate dorsal fins, Paxton is clearly an between the second epibranchial and second apogonid, as it shares with other members of pharyngobranchial (Johnson, 1993); and Kur- the family most of the diagnostic features listed tus has most of the diagnostic features of apo- above (Paxton is specialized with respect to oth- gonids and may be closely related to the family er apogonids in having a single anal fin spine). (Johnson, 1993). We only tentatively include Furthermore, the differences in dorsal fin con- the presence of horizontal and vertical rows of figuration between Paxton and other apogonids neuromasts as a synapomorphy of apogonids are minimal, and the continuity of the dorsal because ongoing work by L. R. Parenti and J. fin in Paxton is clearly secondarily derived. Like Song (USNM, pers. comm.) suggests that simi- most other apogonids, Paxton lacks a spine se- lar linear patterns are found in more teleosts rially associated with the sixth dorsal fin ptery- than previously recognized. Identification of the giophore (Fig. 9). An external gap in the fin closest relatives of apogonids is needed to test fails to form because the sixth and seventh dor- the polarity of all proposed apogonid synapo- sal fin pterygiophores are closely spaced, and morphies, but they appear to be derived relative both pterygiophores lack long posterior projec- to the general percoid conditions: (1) second tions (see discussion under Median Fins above). epibranchial and second pharyngobranchial ar- We do not consider separate dorsal fins as a syn- ticulate synchondrally, (2) the third and fourth apomorphy of apogonids because many lower epibranchials are not greatly enlarged or re- percoids have this feature, and it is unclear how duced, (3) a small fourth pharyngobranchial many times it has evolved. cartilage is associated with the fourth upper Within the Apogonidae, Paxton superficially pharyngeal toothplate, (4) the distal radial as- bears little resemblance to any described genus, sociated with the last dorsal fin spine is elon- but reductive features such as absence of scales gate, and (5) there are no conspicuous rows of and supraneurals suggest it is related to the sensory papillae on the head and body. Fraser pseudamine genera Pseudamia, Pseudamiops,and (1972) noted that apogonines and pseudamines Gymnapogon.None of those genera is defined (Apogonidae herein) share two anal fin spines, on the basis of derived features, and a phylo- a swimbladder with an anterodorsal oval, 10 + genetic analysis of all pseudamine species is 14 vertebrae, and no developed axillary scale at needed to better understand generic limits and the base of the pectoral fin. Of those, we ten- relationships within the group. Such a study is tatively consider two as synapomorphic for apo- in progress, and results will be reported later. gonids: (6) two anal fin spines, one supernu- Of relevance to the present study are the follow- merary on the first pterygiophore (Fig. 10); (7) ing: Is the subfamily Pseudaminae a natural as- 1068 COPEIA, 1999, NO. 4 semblage? If so, is Paxton a pseudamine? And, and variation in most of those characters is com- if so, what are its affinities within the group? mon within the Percoidei. Among apogonines, Our investigation supports monophyly of Fra- the basisphenoid is reduced or absent in Apo- ser's (1972) Pseudaminae and Johnson's (un- gonichthys, Astrapogon, Foa, Fowleria, Neamia, publ. abstract, 70th Annual ASIH Meeting, Phaeoptyx,Rhabdamia (Bentuviaichthys),Siphamia, 1990) conclusions that Paxton is a member of and Vincentia; Phaeoptyx, Rhabdamia, Siphamia, that group and the sister group of Gymnapogon. and some subgenera of Apogon have one or two From our preceding comparisons of the oste- supraneurals, and supraneurals are apparently ology of Paxton with that of other apogonids lacking in Astrapogonalutus (Fraser, 1972), Cer- and from our survey of the pertinent literature, camia (Hayashi, 1991; this study), and Lachner- we hypothesize that the following characters are atus (Fraser and Struhsaker, 1991); Rhabdamia derived for the Pseudaminae with respect to (Verulux) and some species of Siphamia have two apogonines and lower percoids: (1) neurocra- epurals; and Apogon affinis, ,Cor- nium elongate (length approximately four anthus, Rhabdamia (Bentuviaichthys), and Lach- times depth; Fig. 3); (2) parasphenoid contig- neratus have some slightly enlarged or canine uous with pterosphenoid (Fig. 3); (3) basisphe- teeth in the upper and lower jaws. Nevertheless, noid absent dorsal of vomer (Fig. 3); (4) aspect Fraser (1972) hypothesized that primitive apo- flat and from meseth- nearly widely separated gonine states for those four characters are pres- moid ethmoid by large cartilage (Fig. 3); (5) ence of a basisphenoid, three supraneurals, and lower with canine teeth upper jaws (Fig. 5); three epurals, and jaws with bands of villiform two or fewer one or more (6) supraneurals; (7) teeth, and Fraser (1973) found those presum- dorsal fin with spinous pterygiophores proxi- ably primitive features in Holapogon, a genus he mal-middle and distal elements from separated considered to be "near the ancestor which gave one another a 9, 11); (8) two by gap (Figs. epur- rise to all the living Apogoninae" (Fraser 1973: als 12); (9) coracoid with (Fig. cartilage-tipped 1). Johnson (1984) noted that three epurals ventrally directed stem on posteroventral por- and three to be coracoid with supraneurals appear primitive tion (Fig. 13); (10) articulating states for A cladistic reconstruction of cleithrum at a one-third the of the percoids. point length is needed to further in- cleithrum or more from its ventral apogonine phylogeny tip (Fig. 13); states of all with little three-dimensional vestigate primitive apogonine pseu- (11) pelvic girdle damine characters. form, the external ventral wing of the basipter- Randall and Smith (1988) noted that Cerca- ygium small (Fig. 15); and (12) scales cycloid mia resembles Pseudamiops in having only two and deciduous or lacking. pairs of epipleural ribs (epineural bones), and Primitive apogonid states for those features Hayashi (1991) illustrated the caudal skeleton appear to be (1) a deeper neurocranium of C. eremiawith two epurals, a condition that (length approximately 2-2.5 times greatest we believe, as noted above, be a depth); (2) parasphenoid separated from pter- may synapo- of However, a cleared- osphenoid by prootic; (3) basisphenoid present; morphy pseudamines. and-stained of C. eremiaexamined in (4) dorsal aspect of vomer with dorsal projec- specimen this has two ossified and another tion that meets the mesethmoid or nearly does; study epurals small, to those. (5) canine teeth lacking; (6) three supraneur- cartilaginous epural posterior Cercamiacladara has three ossified and als; (7) proximal-middle elements of all dorsal epurals, thus Randall and Smith's of fin pterygiophores with posterior shelf that ar- (1988) description the as characterized three ticulates with distal radial; (8) three epurals; (9) genus being by epur- als to be correct. A reduced number of posteroventral portion of coracoid without ven- appears tral stem; (10) coracoid articulates with cleith- epineural bones is apparently independently in rum near its ventral tip; (11) external ventral derived Cercamiaand Pseudamiops,as our ex- wing of basipterygium well developed, promot- amination of C. eremia and C. cladara with re- ing three-dimensional nature of pelvic girdle; spect to the diagnostic features of pseudamines and (12) scales usually ctenoid, cycloid or weak- listed above suggests that Cercamiais not a mem- ly ctenoid (but not deciduous) in several apo- ber of the Pseudaminae. Cercamialacks all pseu- gonine genera. damine characters except for two, a reduction Four of the putative synapomorphies of pseu- in number of supraneural bones (Cercamialacks damines recognized herein, absence of a basi- them) and absence of a basisphenoid. Further sphenoid, two or fewer supraneurals, two epur- study is needed to hypothesize whether the als, and canine teeth in the jaws, are only ten- presence of those features in Cercamiaand other tatively considered derived for the Pseudaminae apogonines is independent of that in pseuda- because they occur in a number of apogonines, mines or if the Apogoninae are paraphyletic, BALDWIN AND JOHNSON-NEW PSEUDAMINE APOGONID GENUS 1069 with some genera being more closely related to trum; and (14) there are two postcleithra. Fra- the Pseudaminae. ser (1972, plate 12G) depicted Pseudamiopspel- Paxton has all of the characters listed above lucidus without supraneurals, but all specimens as diagnostic for the Pseudaminae. Within the that we examined have one or two small carti- subfamily, Paxton shares numerous specializa- laginous supraneural elements. tions with Gymnapogon,represented in our study Although obviously its sister taxon, Paxton dif- by cleared-and-stained specimens of G. anona, fers notably from Gymnapogonin some aspects G. philippinus, and G. urospilotus, an intact syn- of its morphology. Autapomorphies of Paxton type of the type species, G. japonicus Regan, (with respect to other pseudamines and apo- 1905, and the holotype of G. africanus Smith, gonines) include the following: (1) dorsal fin 1954. From our preceding comparisons of Pax- continuous (Figs. 3, 9); (2) six dorsal fin spines ton with other pseudamines, we propose that (Fig. 9); (3) fourth through sixth dorsal fin the following features are derived within the spines approximately subequal and longest (Fig. Pseudaminae and represent synapomorphies of 9); (4) one anal fin spine (Fig. 10); (5) entire Paxton and Gymnapogon: (1) Bluntly pointed, margin of preopercle covered by skin (Fig. 2A); fleshy flap extending posteriorly from ventral (6) postfrontal bone present (Fig. 3); (7) small margin of preopercle (Fig. 1; see also Smith, bit of bony suturing present between anterior 1954, fig. 3); (2) scales lacking; (3) dorsal edge and posterior ceratohyals (Fig. 6); (8) fifth and of opercle with long posteriorly directed bony sixth infraorbitals widely separated from rest of strut (Fig. 5); (4) preopercle with well-devel- series (Fig. 4); (9) third epibranchial toothplate oped spine at angle (Fig. 5); (5) articulation of lacking (Fig. 8); (10) a separate fifth hypural quadrate and metapterygoid characterized by lacking (Fig. 12); (11) anterior and posterior anterior and posterior synchondral joints sepa- pelvic-girdle processes lacking (Fig. 14); and rated by laminar bone that sometimes compris- (12) an autogenous wishbone-shaped cartilage es interdigitating bony struts (Fig. 5); (6) en- present between proximal bases of left and right dopterygoid with posteroventrally directed pelvic fins (Fig. 14). spine (Fig. 5); (7) interoperculo-mandibular lig- Other apogonids typically have (1) separate ament long, 40-50% of the length of the long dorsal fins; (2) seven to nine dorsal fin spines; axis of the interopercle (Fig. 5); (8) first phar- (3) the penultimate dorsal fin spine much yngobranchial lacking, and distal end of first smaller than the ultimate; (4) two anal fin epibranchial with long thin cartilaginous tip spines; (5) the entire margin of preopercle ex- (Fig. 8); (9) interarcual cartilage long and slen- posed (dorsal limb of margin exposed in Gym- der (Fig. 8); (10) gillrakers present on first and napogon, Fig. 2B); (6) no postfrontal bone; (7) second arches (Fig. 7); (11) supraneurals lack- no suturing between anterior and posterior cer- ing (Figs. 9, 11); (12) parhypural fused with hy- atohyals (except in Glossamia;see Fraser, 1972); purals 1 + 2 (Fig. 12); (13) upper hypural plate (8) fifth and sixth infraorbitals in series with fused to urostylar centrum (Fig. 12); and (14) other infraorbitals; (9) a toothplate fused to the a single postcleithrum (Fig. 13). third epibranchial; (10) fifth hypural autoge- Primitively in pseudamines (1) there is no nous; (11) anterior and posterior processes of fleshy flap on the margin of the preopercle; (2) pelvic girdle present; and (12) no autogenous the body is weakly scaled; (3) there is no long cartilage between bases of pelvic fins. bony strut extending posteriorly from the op- ercle; (4) there is no spine at the angle of the SUMMARY preopercle; (5) the quadrate and metaptery- goid articulate via a single synchondral joint; Four specimens of an unidentified percoid (6) the endopterygoid lacks a posteroventrally fish from northwestern Australia are described directed spine; (7) the interoperculo-mandibu- as a new genus and species of pseudamine apo- lar ligament is short, < 25% of the length of gonid, Paxton concilians. The pseudamine genus the long axis of the interopercle; (8) the sus- Gymnapogonhas been referred to as "puzzling" pensory pharyngobranchial is present and ossi- and "aberrant" (Smith, 1954:775) and "not in- fied, and the medial end of the first epibran- timately related to apogonids or other percoid chial has a small cartilaginous tip; (9) the inter- fishes" (Fraser, 1972:32). Our investigation arcual cartilage is short; (10) gillrakers are pres- clearly supports a sister-group relationship be- ent on the first arch only; (11) the dorsal fin is tween Gymnapogonand Paxton and suggests that preceded by one or two supraneurals; (12) the those genera, along with Pseudamia and Pseu- parhypural is autogenous; (13) neither the up- damiops, form the well-defined subfamily Pseu- per nor lower (Pseudamia) or only the lower hy- daminae, diagnosed here on the basis of 12 de- pural plate (Pseudamiops) is fused to the cen- rived characters. Although diagnostic features 1070 COPEIA, 1999, NO. 4 of Paxton are numerous, most of the characters fishes of southern Indonesia and northwestern- Australia. Tien Wah previously used to diagnose Gymnapogon(a large Press, Singapore. M. 1991. of Cercamiaeremia, preopercular spine, absence of scales, absence HAYASHI, Redescription (: from with com- of ectopterygoid teeth, parhypural fused to low- Apogonidae) Japan, ments of the osteological characters. Sci. Rept. Yo- er hypural plate, and a single postcleithrum; kosuka City Mus. 39:35-44. Fraser, 1972) are also present in Paxton. Further G. D. 1984. Percoidei: and of is need- JOHNSON, development investigation Gymnapogonmonophyly relationships, p. 464-498. In: Ontogeny and system- that be useful in ed. One feature may diagnos- atics of fishes. H. G. Moser, W. J. Richards, D. M. ing the genus is a subocular shelf on the third Cohen, M. P. Fahay, A. W. Kendall Jr., and S. L. infraorbital, which is lacking in other pseuda- Richardson (eds.). Spec. Publ. No. 1, American So- mines but present in most apogonines and low- ciety of Ichthyologists and Herpetologists, Allen er percoids (and thus possibly a reversal in Gym- Press, Lawrence, KS. 1993. and napogon). We anticipate that our ongoing work --. Percomorph phylogeny: progress on will resolve problems. Bull. Mar. Sci. 52:3-28. pseudamine phylogeny questions B. Schindler- about the of all JOHNSON,G. D., ANDE. BROTHERS.1993. monophyly pseudamine genera a as well as corroborative evidence for ia: paedomorphic goby (Teleostei: Gobioidei). provide hy- Ibid. 52:441-471. of outlined herein. potheses relationships KERSHAW,D. R.. 1970. The cranial osteology of the "butterfly fish," Pantodon buchholziPeters. Zool. J. ACKNOWLEDGMENTS Linn. Soc. 49:5-19. LACHNER,E. A. 1953. Family Apogonidae: cardinal We thank J. R. Paxton, G. R. Allen, A. Gra- fishes, p. 412-498. In: Fishes of the Marshall and ham, P. Last, M. E. Anderson, D. Didier, and W. Marianas Islands. L. P. Schultz (ed.). Bull. U.S. Nat. Eschmeyer for loans of specimens or photo- Mus. 202. graphs of specimens. We are grateful to K. Dar- LE DANOIS,Y. 1967. Quelques figures descriptives de l'anatomie de Pantodon buchholzi Peters. Bull. row for rendering the inked versions of most illustrations. I.F.A.N., T. XXIX, ser. A. 3:1051-1096. LEVITON,A. E., R. H. GIBBSJR., E. HEALAND C. E. DAWSON.1985. Standards in herpetology and ich- LITERATURECITED thyology. Part 1. Standard symbolic codes for insti- tutional resource collections in herpetology and AHLSTROM,E. H., L. BUTLER,AND B. Y. SUMIDA. J. . Copeia 1985:802-832. 1976. stromateoid fishes (Pisces, Percifor- Pelagic MCALLISTER,D. E. 1968. Evolution ofbranchiostegals mes) of the eastern Pacific: kinds, distributions, and classification of teleostome fishes. Bull. Nat. and life histories and observations on five of early Mus. Can. 221, Biol. Ser. 77:I-xiv, 1-239. these from the northwest Atlantic. Bull. Mar. Sci. PATTERSON,C., AND G. D. 1995. The inter- 26:285-402. JOHNSON. muscular bones and ligaments of teleostean fishes. ALLEN, G. R. 1992. A new species of cardinalfish northern Rec. W. Smiths. Contr. Zool. 559, Washington, DC. (Apogonidae) from Australia. D. F. G. AND E. Aust. Mus. 15:697-701. PAXTON,J. R., HOESE, R. ALLEN, J. HANLEY.1989. of Australia. BLEEKER,P. 1865. Enumeration des de Zoological catalogue especes pois- Vol. 7. Government sons actuellement connues de l'ile d'Amboine. Australian Publishing Service, Canberra, Australian Capital Territory, Australia. Ned. Tijdschr. Dierk. 2:270-293. RANDALL, E., ANDC. L. SMITH.1988. Two new FRASER,T. H. 1972. Comparative osteology of the J. spe- cies and a new of cardinalfishes (Percifor- shallow water cardinal fishes (Perciformes: Apogon- genus mes: from South Pacific Ocean. idae) with reference to the systematics and evolu- Apogonidae) Rapa, Am. Mus. Novit. 2926:1-9. tion of the family. J. L. B. Smith Inst. Ichthyol., Ich- , E. A. LACHNER, AND T. H. FRASER.1985. A thyol. Bull. 24: v + 105, 44 pls. revision of the Indo-Pacific fish --. 1973. Evolutionary significance of Holapogon, apogonid genus with of three new a new genus of cardinal fishes (Apogonidae), with Pseudamia, descriptions species. Indo-Pacific Fishes No. Hono- a redescription of its type-species, Apogon maximus. 6, Bishop Museum, HI. Spec. Publ. J. L. B. Smith Inst. Ichthyol. 10:1-7. lulu, C. T. 1905. On a fishes , ANDC. R. ROBINS.1970. A new Atlantic genus REGAN, collection of from the Sea of cardinalfishes with comments on some species inland of Japan made by Mr. R. Gordon Smith. from the Gulf of Guinea. Stud. Topical Oceanogr. Ann. Mag. Nat. Hist. 15:17-26. 4:302-315. - . 1940. The perciform genera Gymnapogonand , ANDP.J. STRUHSAKER.1991. A new genus and Nannatherina. Copeia 1940:173-175. species of cardinalfish (Apogonidae) from the SAINSBURY, K. J., P. J. KAILOLA, AND G. G. LEYLAND. Indo-West Pacific, with a key to apogonine genera. 1984. Continental shelf fishes of northern and Copeia 1991:718-722. north-western Australia. Clouston and Hall and Pe- FUJITA,K. 1990. The caudal skeleton of teleostean ter Pownall Fisheries Information Service, Canber- fishes. Tokai Univ. Press, Tokyo, Japan. ra, Australian Capital Territory, Australia. GLOERFELT-TARP,T. ANDP. J. KAILOLA. 1984. Trawled SMITH,J. L. B. 1954. Apogonid fishes of the subfamily BALDWIN AND JOHNSON-NEW PSEUDAMINE APOGONID GENUS 1071

Pseudamiinae from southeast Africa. Ann. Mag. Belgique, Memoires de la Classe des Sciences, Col- Nat. Hist., ser. 12, 7:775-797, 1 pl., 3 figs. lection in-8?-2e serie, T.XLII-Fascicule 6 et dernier STIASSNY,M. L. J., ANDJ. A. MOORE.1992. A review 1978. of the pelvic girdle of acanthomorph fishes, with of inter- comments on hypotheses acanthomorph DIVISION OF FISHES, MRC 159, NATIONALMUSE- Zool. Linn. Soc. 104:209-242. relationships. J. UM OF NATURALHISTORY, SMITHSONIAN INSTI- TAVERNE, L. 1974. et Osteologie, phylogenese syste- TUTION, WASHINGTON DC 20560. E-mail: matique des Teleosteens fossiles et actuels du su- (CCB) [email protected]. Send re- per-ordre des Osteoglossomorphes. Deuxieme par- to CCB. Submitted: 23 Nov. tie. Osteologie des genres Phareodus,Phareoides, Bry- print requests chaetus, Musperia, Pantodon, Singida, Notopterus,Xe- 1998. Accepted: 1 Feb. 1999. Section editor: nomystus et Papyrocranus. Academie Royale De S. A. Schaefer.