Morphology

Osteological features of the new species and related taxa were examined using radiographs, specimens cleared and stained for bone and cartilage (C&S) using a protocol based on Taylor and Van Dyke (1985), and dry skeletal preparations. Comparative materials are listed in appendix 1. Light organs were either isolated from each taxon (except those represented only by type specimens) or examined in situ for rare species, to permit detailed morphological comparisons of the LOS. Morphometric measurements were recorded to the nearest 0.1 mm using dial calipers. Standard length (SL) is used throughout. Vertebral counts exclude the ural centrum ( =  last half-centrum). Following Hubbs and Lagler (1949), the first caudal vertebra is here defined as the first vertebra bearing a definite hemal spine. Vertebral and fin spine/ray counts were obtained from radiographs. The terminal dorsal-fin and anal-fin rays, which are branched to the base of the fin, are counted as a single element. Pored scales in the lateral line are counted in series from the dorsal margin of the gill opening to the caudal flexure. Scale counts should be interpreted as approximations, due to high intraspecific variability, irregular arrangement, the deciduous nature of ponyfish scales in preservation, and because small scale size and the degree to which scales are embedded make accurate counts problematic.

Institutional Abbreviations

Institutional abbreviations follow Leviton et al. (1985):

Phylogeny Reconstruction

For the phylogenetic analysis, 6160 equally weighted nucleotide characters (based on the implied alignment [Wheeler, 2003]) from the seven mitochondrial (16S, COI, ND4, ND5, tRNA-His, tRNA-Ser, tRNA-Leu) and two nuclear genes (28S, histone H3) used by Sparks et al. (2005), and 17 morphological features of the LOS (see table 1 and appendix 2), were simultaneously analyzed under the optimality criterion of parsimony with all transformations given equal weight. Two taxa not included in the analysis of Sparks et al. (2005), Photoplagios klunzingeri and P. moretoniensis, are included here on the basis of morphological features of the LOS only, as no tissue sample suitable for molecular studies could be acquired for either taxon. The parsimony analysis for 103 terminals was conducted using direct optimization (Wheeler, 1996) as implemented in the program POY (Wheeler et al., 2003). All methods for sequence acquisition and phylogeny reconstruction are presented in Sparks et al. (2005), as are GenBank accession numbers for included taxa. The length of the resulting implied alignment (Wheeler, 2003) was verified in NONA (Goloboff, 1998). Patterns of character evolution were examined using NONA in conjunction with WinClada (Nixon, 2000). Only unambiguous morphological transformations common to all most-parsimonious dichotomized trees are used to diagnose clades (Goloboff, 1995).

Table 1

Clade designations refer to those recovered in Sparks et al. (2005). Character descriptions are listed in appendix 2. Inapplicable characters are designated by (–).

Phylogenetic Analysis

Simultaneous analysis of the 17 morphological transformations corresponding to features of the LOS (table 1) and the nine gene fragments used by Sparks et al. (2005), comprising a total of 6177 characters (2660 parsimony informative), resulted in 120 equally most-parsimonious trees with lengths of 20,080 steps, a consistency index of 0.30, and a retention index of 0.55 (when uninformative characters are retained). Given that relationships outside of Photoplagios are identical to those presented by Sparks et al. (2005), a strict consensus topology of these optimal trees restricted to Photoplagios is presented (fig. 1). All unambiguously optimized anatomical features of the LOS are indicated on this topology with numerations that match the character descriptions in table 1 and appendix 2.

Figure 1

Intrarelationships of Photoplagios, illustrating the placement of Photoplagios antongil and two additional congeners (P. klunzingeri and P. moretoniensis) that were not included in the phylogenetic study of Sparks et al. (2005). Unambiguously optimized anatomical features of the light-organ system (LOS) are indicated on the topology. Photoplagios is diagnosed by the presence in males of (character number is followed by state and indicated on the topology): (1) Character 3:1: Dorsolateral lobes of the light organ that are hypertrophied and extend posteriorly into the gas bladder (extensively in P. elongatus and P. rivulatus, less so in the remaining congeners). (2) Character 6:1: Lateral clearing of the silvery lining of the gas bladder, which is directly correlated with the species-specific transparent lateral flank patches or stripes (e.g., Sparks et al., 2005: figs. 4 and 5). (3) Character 8:1: Lateral luminescence via transparent flank patches or stripes in males. Within Photoplagios, the sister-group relationship between P. klunzingeri and P. leuciscus is supported by Character 16:1: The presence of a translucent triangular patch located ventral to the lateral midline and at mid-flank, which does not extend anteriorly to the pectoral-fin base (fig. 5). Additional putatively synapomorphic features supporting this sister-group relationship within Photoplagios include a highly speckled light organ, a unique pigmentation pattern (densely spotted and speckled with thin irregular lines) on the upper flanks, and a characteristically elongate second dorsal-fin spine. Although in the current analysis there is no synapomorphy supporting a sister-group relationship between P. moretoniensis and P. stercorarius, the shared possession of Character 10:1, a translucent flank stripe in males comprised of numerous, serially arranged, mid-lateral windows, which may be discrete or overlapping, is unique to these two species and may indicate a close relationship between them. Additional putatively synapomorphic features shared by P. moretoniensis and P. stercorarius include a unique pigmentation pattern on the upper flanks comprising large, sparse blotches and irregular lines, with a prominent black stripe just ventral to the dorsal fin, and lateral clearing of the silvery gas bladder lining in males that does not extend the length of the chamber. *Note: Photoplagios sp. “Sri Lanka” was identified as P. lineolatus in the phylogeny presented by Sparks et al. (2005: fig. 2) (see Discussion).

In this reconstruction Photoplagios is monophyletic and is supported by three unambiguously optimized features of the LOS (figs. 1 and 2). Both P. klunzingeri and P. moretoniensis, included here on the basis of anatomical features of the LOS only, are recovered as members of Photoplagios. Members of Photoplagios that exhibit an expansive translucent triangular flank patch (P. elongatus, P. rivulatus, P. leuciscus, and P. klunzingeri) are monophyletic, whereas the intrageneric relationships of Photoplagios with a translucent midlateral flank stripe (P. stercorarius, P. moretoniensis, and P. antongil, n. sp.) are unresolved. Photoplagios klunzingeri is recovered as the sister taxon to P. leuciscus, whereas the intrageneric placement of P. moretoniensis remains unresolved. The intrageneric placement of the new species is also unresolved in this reconstruction.

Figure 2

Diagrammatic illustration of the light organ and associated internal features of the LOS in representative members of Photoplagios. (A) Photoplagios elongatus male, illustrating strongly sexually dimorphic internal features of the LOS. The dorsal lobes of the light organ are greatly enlarged in males and extend well into the gas bladder. (B) Photoplagios antongil male, illustrating sexually dimorphic internal features of the LOS. The dorsal lobes of the light organ are moderately enlarged in males and extend only slightly into the gas bladder chamber. Note variation in light organ pigmentation and the distribution of melanophores. Abbreviations: ES, esophagus; G, gut; GB, gas bladder; LO, light organ. Heavy dashed lines indicate the gas bladder, and gray shading indicates the extent of lateral clearing of the silvery, guanine-lined gas bladder chamber. Black spots indicate the distribution of melanophores on the light organ and gas bladder lining.

Systematic Account

Figure 3

Photoplagios antongil, holotype, AMNH 236544, adult male, 75.0 mm SL; Madagascar: Maroansetra market.

Figure 4

Photoplagios antongil, paratype, AMNH 236545, adult female, 81.8 mm SL; Madagascar: Maroansetra market.

Figure 5

Photoplagios leuciscus, illustrating expansive transparent triangular flank patch characteristic of males and how the patch may appear in different lots due to method of preservation and condition of material upon preservation. (A) AMNH 237149, adult male, 95.0 mm SL; northeastern Madagascar: Maroansetra ← market. Specimen with silvery guanine intact. Flank patch difficult to discern unless specimen is examined obliquely. (B) AMS I.22978004, adult male, 95.8 faded. Flank patch is easy to locate as darker triangular area surrounded by silvery integument. (C) USNM 373280, adult male, 79.8 mm SL; Iran. Specimen in which guanine is completely lost in preservative. Triangular flank patch appears black due to concentration of dispersed melanophores (see text for discussion).

Holotype

AMNH 236544, 75.0 mm SL, adult male; Northeastern Madagascar: Antongil Bay: Maroansetra market; J.S. Sparks, W.L. Smith, and K.L. Tang, Mad JSS 29-2003, Nov. 2003.

Paratypes

AMNH 236545, 3 ex., 77.3–81.8 mm SL; data as for holotype. SIO 05-117, 1 ex., 76.0 mm SL; data as for holotype.

Diagnosis

Males of Photoplagios antongil are distinguished from congeners by the presence of a broad and presumably translucent (in life) midlateral stripe, which is darkly pigmented in preservative due to a concentration of melanophores, and by the presence of two darkly pigmented flank patches (presumably also translucent in life) located ventral to the lateral midline, one anteroventral to the pectoral-fin base and another ventral to the lateral midline at about midbody. Photoplagios antongil is further distinguished from P. leuciscus, the only externally similar species occurring in the region, by the absence of a large translucent triangular patch on the flanks (present in male P. leuciscus); upper flank pigmentation consisting of large, sparse spots and blotches (vs. highly speckled with fine lines and small spots); a much shorter, although elongate, second dorsal-fin spine; a straight predorsal profile (vs. weakly S-shaped); absence of black pigment in the pectoral-fin axil; and exposed conical oral dentition in two distinct rows (vs. multiple closely set and indistinct rows of unexposed villiform teeth).

Description

Selected proportional measurements and meristic data presented in table 2. A moderately shallow-bodied and elongate leiognathid. Body laterally compressed. Lateral snout outline mostly straight. Weak preorbital protuberance due to protrusion of both frontal and lateral ethmoid ossifications. Predorsal head profile mostly straight to mildly convex. Nuchal spine slightly protruding and distal tip exposed. Nuchal spine with distinct median keel. Two short and stout postnasal spines ( =  protuberances) present on lateral ethmoid, located posterior to nasal foramina and just rostrodorsal of orbit. Postnasal spines followed posteriorly by well-developed supraorbital ridges that converge posteriorly. Dorsal and ventral body profiles moderately rounded. Dorsal-fin origin located slightly posterior to vertical through pelvic-fin origin. Anal-fin origin located at about level of vertical through the last ( =  8th) dorsal-fin spine or first dorsal-fin ray. Eye large. Caudal peduncle slender and shallow. Mouth small and terminal in position, directed slightly downward when protruded. Posterior margin of maxilla exposed, reaching to level of vertical through anterior margin of orbit. Anterior nasal pore small and round; posterior foramen much larger and crescent-shaped, partially encircling anterior pore. Preopercular margin weakly serrate along ventral and ventrocaudal margins. Vertebral count: 10 precaudal + 13 caudal  =  23. Neural and hemal spines of vertebral centrum PU4 somewhat expanded and bladelike. Twelve or 13 elongate and triangular outer gill rakers arrayed along lower limb ( =  ceratobranchial one) of first gill arch.

Table 2

Morphometric and Meristic Data for Photoplagios antongil, New Species

Fins

Dorsal fin with VIII spines and 16 branched rays. First dorsal-fin spine greatly reduced in length and relatively robust. Second through fourth dorsal-fin spines elongate and robust; second spine longest. Second dorsal-fin spine moderately, but not exceedingly, elongate. Third and fourth dorsal-fin spines serrate along anterior margin and “lock” into groove on posterior margin of preceding spine when erect. Dorsal-fin spines five through eight feeble, shorter than second through fourth spines. Anal fin with III spines and 14 branched rays. First anal-fin spine very short. Second and third anal-fin spines robust and elongate; second spine longest, but not appreciably longer than third. Third anal-fin spine serrate on anterior margin and “locks” into groove on posterior margin of second spine when erect. Spinous dorsal and anal fins with asquamate basal sheath. Pelvic fins short, not reaching first anal-fin spine when adducted (i.e., an appreciable gap present). Eight upper and seven lower branched caudal-fin rays. Seventeen total pectoral-fin rays.

Dentition

Two distinct rows of closely set, elongate and recurved conical teeth present in both upper and lower jaws. Upper jaw teeth somewhat larger than those of lower jaw, particularly anteriorly. Lips not fleshy, and teeth exposed.

Squamation

Body scales cycloid, but not remarkably small. Head and opercular region asquamate. Chest fully scaled, except along ventral midline. Scales conspicuous and extend anteriorly ventral to opercle to about anterior margin of chest. Lateral line arched and complete. Pored scales in lateral line number approximately 60 to 63. Scales difficult to count posteriorly due to high proportion missing. Pores well developed. Pelvic axillary scale well developed and elongate. All fins asquamate, except for caudal fin, which bears several rows of scales of reduced size, particularly centrally on fin.

Pigmentation in Preservative

Body ground coloration yellowish-olive above lateral midline and creamy pale yellow below. Iridescent silvery patches present to varying degree ventral to lateral midline (although much of guanine lost in preservation). Opercle, subopercle, and interopercle iridescent and silvery. Midlateral stripe present in males. Stripe broad and darker than flank regions above and below due to concentration of melanophores, and presumably translucent in life (compare figs. 3 and 4, illustrating an adult male and female, respectively). Stripe begins posterodorsal to pectoral-fin base and extends to about anterior margin of caudal peduncle. Two dark gray to charcoal, and also presumably translucent in life, patches present on flank in males, one anteroventral to pectoral-fin base in all specimens and another ventral to lateral midline at about midbody in only some specimens (compare figs. 3 and 4). Based on comparisons to the translucent flank patches in congeners (e.g., P. leuciscus, P. klunzingeri, P. stercorarius, and P. moretoniensis) in various states of preservation, in specimens in which the silvery guanine layer has faded in preservation, these flank patches are generally very obvious due to a concentration of melanophores causing them to appear blackish (see fig. 5). Thus, it is reasonable to conclude that the blackish midlateral stripe and two lower flank patches in the new species, in which the silvery guanine on the integument is mostly lost in preservation in all specimens, are also translucent in life.

Pigmentation pattern above lateral midline characteristically blotchy and mottled. Blotches large, variable in size, and sparsely arranged. Longitudinal series of blackish spots arrayed directly above midlateral stripe. Pores of lateral line scales edged dorsally and ventrally with melanophores; in combination with concentration of blackish blotches present along lateral line, melanophores create an arching and irregular stripe. Body ventral to lateral midline peppered with melanophores. Cheek and gular region pale yellow; some silvery iridescence on cheek. Head above orbit and nape grayish to grayish brown. Snout with large black patch above upper lip. Lips pale yellow. Two distinct black blotches present anterior to orbit; dorsal blotch covering postnasal spines. Line of black pigment ventral to eye. Chest and belly gray, silvery, or pale yellow; silvery along ventral midline. Caudal peduncle silvery with blackish saddle on dorsal margin. Base of caudal fin iridescent and silvery. Dorsal and anal fins hyaline to pale white, with a concentration of black pigment distally on dorsal-fin spines. Pectoral and pelvic fins pale yellow. Black stripe present on body ventral to dorsal-fin base; more prominent posteriorly. Caudal fin yellowish proximal to base, white distally. Membrane of caudal fin with concentration of black pigment, creating series of thin black radiating lines.

Light-organ System (los)

Sexual dimorphism of the light organ and associated internal structures is evident, but not striking as in some other members of the genus (e.g., P. elongatus and P. rivulatus [Dunlap and McFall-Ngai, 1984; Sparks et al., 2005]). The light organ of males in general is enlarged compared to similarly sized conspecific females. The dorsal lobes of the light organ in particular are enlarged in males, but not greatly, and extend into gas bladder only slightly at the light-organ window (fig. 2). The light organ itself is highly spotted, with prominent black melanophores on a yellow background, lending a leopard skin pattern to much of the light organ. Based on the limited material available, the light organ of females appears in general to be smaller and not as densely spotted, but more solidly black. There is a broad lateral clearing of the gas bladder lining, extending almost the entire length of the gas bladder chamber, and the chamber is silvered laterally only at the anterior end near the light-organ window. The gas bladder is peppered with melanophores anteriorly near the light-organ window and also caudoventrally, but there are no melanophores present in the guanine-free ( =  clear) region of the gas bladder lining.

The material available is not ideally preserved for reliably sexing some of the specimens (i.e., the gonads are poorly preserved) or for interpreting the extent of flank pigmentation, translucence, or silvering in life. Therefore, until additional material becomes available, I am reluctant to comment further on external sexual dimorphism of the LOS apart from noting that in specimens of Photoplagios (and Photopectoralis) in which the silvery guanine is lost in preservative, as it is in the type series of P. antongil, regions that were translucent in life generally appear blackish due to a concentration of dispersed melanophores (fig. 5B and C). It would appear that these concentrated melanophores, which are generally aggregated and quite small in well-preserved specimens in which the guanine layer is intact (fig. 5A), can be rapidly dispersed to occlude the transparent flank patches and stripes so as to inhibit the emission of luminescence from the light organ.

Distribution

Known only from market specimens purchased in northeastern Madagascar (Maroansetra market). The Malagasy fishermen who supply the market work locally in Antongil Bay, and it can safely be assumed that is where the type series was collected. Presumably the species is more widely distributed; however, it is not represented in collections made elsewhere in Madagascar that are available to the author. In their guide to the commercial fishes of Madagascar, Bauchot and Bianchi (1984) list Leiognathus lineolatus as present in the region. The fish referred to as L. lineolatus may correspond to the new species; however, no specimens from the study of Bauchot and Bianchi (1984) appear to have been deposited at MNHN.

Etymology

Named in reference to the type locality, Antongil Bay, in northeastern Madagascar. The specific epithet, antongil, is used as a noun in apposition.

Comparisons

When originally collected, specimens now assigned to the new species, P. antongil, were thought to represent rather uncharacteristic P. leuciscus females due to similarities in overall pigmentation pattern and body shape (fig. 5). The absence of an expansive triangular patch on the flanks in P. antongil males also led them to be confused initially with P. leuciscus females. Upon closer examination in the laboratory, however, it was obvious that two distinct species were represented in these collections. This hypothesis was corroborated by an analysis combining nucleotide characters from both mitochondrial and nuclear genes and anatomical features of the LOS, in which P. leuciscus was recovered as the sister taxon to P. klunzingeri and shown not to be conspecific with P. antongil (fig. 1).

Sparks et al. (2005: fig. 2) recovered P. antongil as the sister taxon to specimens collected in Sri Lanka that they referred to as P. lineolatus (here referred to as Photoplagios sp. “Sri Lanka”; see below). However, with the inclusion of two additional species of Photoplagios in the current analysis, the intrageneric relationships of P. antongil are unresolved (fig. 1). Unfortunately, Sparks et al. (2005) were able to obtain only a tissue sample of the Sri Lankan material (identified by Seishi Kimura, Mie University, Japan). Likewise, I have been unable to examine preserved specimens of the Sri Lankan material and cannot comment on morphological features that distinguish them from P. antongil. With regard to the molecular data set examined by Sparks et al. (2005) and reanalyzed herein, molecular divergence between P. antongil and the Sri Lankan specimens referred to as P. lineolatus is about 4% (uncorrected p distance) across the four genes sequenced, which is similar to divergences recovered between other species within the genus.

Of the other ponyfish species present in the Malagasy and Mascarene region, P. antongil is most similar to, and has likely been confused in the past with, P. leuciscus. The new species is easily distinguished from P. leuciscus, however, by the presence of a broad and presumably translucent midlateral stripe in males, which is darkly pigmented in preservative due to a concentration of melanophores; absence of a large translucent triangular patch on the flanks (present in P. leuciscus males; fig. 5); pigmentation pattern on the upper flanks (larger, sparse spotting and mottling vs. highly speckled and with thin lines in P. leuciscus; figs. 3–Figure 45); a much shorter, although still elongate, second dorsal-fin spine; a straight predorsal profile (vs. weakly S-shaped); absence of black pigment in the pectoral-fin axil; and exposed conical oral dentition in two distinct rows (vs. multiple indistinct rows of unexposed villiform teeth).

Taxonomy of Equula lineolata Valenciennes

Although Leiognathus lineolatus (Valenciennes, in Cuvier and Valenciennes, 1835) is considered to be a common species distributed throughout the Indo-Pacific (references in Eschmeyer, 2005; Froese and Pauly, 2005), its taxonomic status is problematic and warrants discussion. In his systematic review of Leiognathidae, James (1975) considered Leiognathus lineolatus to be a valid species with a wide geographic distribution; however, he noted that the original description of Valenciennes “lacks many details of body proportions and colour”. He further commented that “the only important characters mentioned are the second dorsal spine is more than half [the] height of the body and the back with many small vertical lines and spots.” Both of these features, however, occur together in other leiognathid species (e.g., P. leuciscus and P. klunzingeri) and are, therefore, insufficient to distinguish L. lineolatus. James's (1975: 163) reliance on these homoplasious features is exemplified by his admittedly questionable synonymy of Leiognathus vermiculatus Fowler, 1904 with L. lineolatus. The holotype of L. vermiculatus (ANSP 27525), however, is clearly referable to Photoplagios leuciscus, due to the presence of an expansive triangular flank patch and other characteristic features, a synonymy that Jones (1985: 590–591) later proposed.

In a subsequent revision of Australian ponyfishes, Jones (1985) noted that based on an examination of the type series (or a portion thereof; see below) the name Leiognathus lineolatus (Valenciennes, in Cuvier and Valenciennes, 1835) seems traditionally to have been incorrectly used. Among the type specimens examined (those “collected by Quoy and Gaimard from Java”), Jones found a number of different species to be represented, none of which corresponded to the original description of Valenciennes (although this seems to have been overlooked by Jones as exemplified by the designation of lectotypes from among this material). According to Jones, “MNHN 6739 consists of one specimen of L. elongatus …, two specimens of Gazza minuta …, and one badly damaged specimen that could not be identified …; MNHN 6738 consists of two specimens in fair condition, which represent L. bindus.” Without further explanation, Jones (1985: 575–576) then designated the “two specimens registered as MNHN 6738 as lectotypes of E. lineolata and regard[ed] this species then as a junior synonym of L. bindus.” Jones (1985) presented no justification as to why the specimens identified as L. bindus were designated lectotypes of E. lineolata (e.g., Gazza minuta [Scomber minutus Bloch, 1795] is an older name), not to mention that it was (and is) in violation of Article 74 of the International Code of Zoological Nomenclature (1961, 1999) to designate more than one lectotype. As discussed in detail below, it is clear from even a cursory reading of Valenciennes's description of E. lineolata that he was describing a fish not easily confused with the deep-bodied L. bindus (fig. 6), which incidentally Valenciennes described earlier in that same publication, stating that “its proportions resemble much those of” Equula ensifera ( =  Leiognathus equulus), a conspicuously deep-bodied and disk-shaped leiognathid. As noted by Eschmeyer (2005), Bauchot and Desoutter (1989: 21) subsequently restricted the lectotype designation to a single specimen (the larger specimen of 46 mm SL); therefore, they designated the lectotype.

Figure 6

Photopectoralis (L.) bindus. (A) Photopectoralis bindus male in preservation, UMMZ uncat. (PVD 00-01/18a), 66.5 mm SL; Philippines. (B) Illustration of Bindoo karah reproduced from Russell (1803: vol. 1, 50, pl. 64), upon which Valenciennes, in Cuvier and Valenciennes, 1835 based his description of Equula bindus.

According to Jones (1985: 575), however, only the type series of Equula lineolata “collected by Quoy and Gaimard from Java” was examined, whereas in the original description of E. lineolata, Valenciennes explicitly stated that specimens were “taken in Java by Mrs. Quoy and Gaimard, and in Antjer, in the strait of the Sonde, by Mr. Raynaud.” Eschmeyer (2005) stated that he was not able to locate the Antjer specimens. Likewise, in their type catalog of fishes in the MNHN, Bauchot and Desoutter (1989: 21) noted “the brought back specimens of Antjer (strait of the Sonde) were not found”. I have been unable to locate the Antjer specimens as well, and they are presumed lost. Jones (1985) apparently was unaware of the additional Antjer specimens, as no reference to this material was included. It is important to note that Valenciennes made no mention of the number of specimens he examined for the original description of E. lineolata, only the collection localities (thus, it is possible that the specimens subsequently assigned MNHN 6738 and 6739 were not part of the type series examined by Valenciennes [i.e., MNHN catalog numbers were not listed at that time], and represented other ponyfish species collected at those localities). It is also possible that the Antjer specimens would correspond to Valenciennes's original description of E. lineolata. Unless they are located, however, this correspondence can never be established.

Regardless, it appears as though Jones (1985) did not formulate these taxonomic decisions regarding the status of Equula lineolata in consultation with the original description of that species; otherwise it would have been obvious that whatever taxon Valenciennes was describing, it was not L. bindus, a very deep-bodied and essentially disc-shaped ponyfish (fig. 6). In his opening comment to the description of E. lineolata (referring to the previously described taxon in that publication, E. oblonga), Valenciennes described E. lineolata as being similar in form and in the presence of a midlateral stripe to E. oblonga. I have examined the holotype of E. oblonga from Timor (MNHN A-6754, 61.8 mm SL; fig. 7), which is in relatively good shape, and I find the species to be quite distinct from L. bindus, a species described as having a “rhombic body” (Russell, 1803: vol. 1, 50). (Valenciennes, in Cuvier and Valenciennes, 1835 described E. bindus based on Bindoo karah Russell, 1803 [vol. 1, 50, pl. 64 ; reproduced in fig. 6B].) Equula oblonga (MNHN A-6754, holotype), as its name suggests, is a very elongate ponyfish (“its height is three times and a half in its overall length”) that exhibits a midlateral stripe (“side line”), a feature lacking in L. bindus (Valenciennes, in Cuvier and Valenciennes, 1835). According to Jones (1985: 575), body depths as a percentage of standard length for the two specimens from the putative syntypic series of E. lineolata that were identified as L. bindus and designated as lectotypes of E. lineolata were 51% and 48%. Evidently, Valenciennes was describing a much more shallow-bodied species than L. bindus; something similar in body shape to E. oblonga (fig. 7).

Figure 7

Equula oblonga, holotype, MNHN A-6754, 61.8 mm SL; Timor, Indonesia.

Therefore, there is justification under Article 74.2 of the Code (“lectotype found not to have been a syntype”) for rejecting the lectotype designation(s) of Jones (1985) and Bauchot and Desoutter (1989: 21) for specimens attributable to E. bindus, not Valenciennes's E. lineolata. Moreover, in the absence of type material corresponding to Valenciennes's original description of Equula lineolata, the status of this taxon remains uncertain. Valenciennes's original description is not adequate to distinguish the species from congeners (i.e., no diagnostic features were presented), and Equula lineolata must be considered a nomen dubium of uncertain placement beyond the family level (i.e., there is no way of knowing exactly what Valenciennes described).

Both Jones (1985: 603–605) and Woodland et al. (2001: 2817, pl. III) included undescribed species in their regional treatments of ponyfishes that warrant comment regarding the status of Equula lineolata and specimens traditionally ascribed to this taxon. The species referred to as “Leiognathus sp.” by Jones (1985) closely resembles Photoplagios leuciscus, with males of this putatively undescribed species possessing a single expansive triangular flank patch. According to Jones (1985: 590), however, the latter species is distinguished from the former by a fully (vs. partly) scaled chest. The presence of a translucent triangular flank patch readily distinguishes “Leiognathus sp.” of Jones (1985) from P. antongil, which instead possesses a broad darkly pigmented midlateral stripe. However, the species referred to as “Leiognathus sp. 1” by Woodland et al. (2001), at least on the basis of the included color plate (pl. III), very closely resembles the overall body shape and pigmentation pattern of Photoplagios antongil. From the illustrations presented by Woodland et al. (2001), P. antongil can distinguished from “Leiognathus sp. 1” by the absence of two distinct and large dark blotches on the lower flanks, the absence of a darkly pigmented midlateral stripe, and assuming that a male is illustrated by Woodland et al. (2001), by a shorter second dorsal-fin spine. The chest is completely scaled in P. antongil, except along the ventral midline, whereas Woodland et al. (2001: 2817) reported that the anterior portion of the chest (“below head”) in “Leiognathus sp. 1” is asquamate.

Woodland et al. (2001: 2817) listed “?Leiognathus lineolatus” under “Frequent misidentifications” for their “Leiognathus sp. 1” and included Madagascar within its range. It is puzzling then that under FAO (Food and Agriculture Organization of the United Nations) names ( =  common names), Woodland et al. (2001: 2817) listed “Ornate ponyfish” as the English name for this undescribed species, which is the common name of Leiognathus lineolatus in the literature (in fact, it is the common name used by Woodland on the FishBase website for L. lineolatus [considered there a valid species] [Froese and Pauly, 2005]). Woodland et al. (2001) did not include L. lineolatus in their Western Central Pacific FAO contribution on leiognathids, they did not place the species in synonymy with any included species, and they offered no comment regarding its exclusion (as mentioned, the name only appears as a questionable misidentification of an undescribed species referred to as “Leiognathus sp. 1” [Woodland et al., 2001: 2817]).

According to references cited in FishBase (Froese and Pauly, 2005), L. lineolatus is reportedly widespread, with a range extending from “Durban, South Africa, Madagascar and Pakistan to the Philippines, north to southern Japan and south to northeast Australia.” Yet, the species was not mentioned (not even in synonymy) in the most recent FAO guide covering this region (Woodland et al., 2001). It is unclear whether (and, if so, why) Woodland et al. (2001) followed the synonymy of Equula lineolata with Leiognathus bindus as proposed by Jones (1985) in the FAO publication (although E. lineolata is not listed as a junior synonym of L. bindus in that contribution) but not on the Fishbase website.

Despite a wealth of discussion and disagreement regarding the identity of Leiognathus lineolatus (e.g., Fowler, 1904; James, 1975; Jones, 1985; Woodland et al., 2001), the name continues to be applied without justification to a number of presumably distinct species throughout the Indo-Pacific. Recent researchers seem to have overlooked the troubling fact that there are no extant types that conform to Valenciennes's original description, which on its own is insufficient to diagnose the species or distinguish it from congeners. Jones (1985) was correct in asserting that the name Leiognathus lineolatus (Valenciennes, in Cuvier and Valenciennes, 1835) seems traditionally to have been incorrectly, or at least unjustifiably, used. Jones's (1985) conclusion was based on an absence of specimens in the extant type series that corresponded to what taxonomists had traditionally ascribed to Leiognathus lineolatus, regardless of Valenciennes's description (although it remains unclear as to why Jones synonymized E. lineolata with L. bindus). My interpretation relies on a more literal reading of Valenciennes's original description of Equula lineolata and asserts that in the absence of type material that corresponds to this description, as well as the lack of any features in the original description that serve to diagnose E. lineolata or distinguish the species from congeners, there is no justifiable alternative but to recognize Equula lineolata as a nomen dubium of uncertain placement beyond the family level.